Merge branch 'rc-4.5.0'

Update ChangeLog for 4.5.0.
Add casts to CONF_HANDLE_T_U().
2026-07-22 20:53:07 +00:00 · 2017-02-28 19:09:23 -08:00 · 2017-02-28 16:21:05 -08:00 · 2017-02-28 16:20:44 -08:00 · 2017-02-28 15:06:27 -08:00 · 2017-02-28 14:25:06 -08:00
144 changed files with 11231 additions and 4470 deletions
--- a/.appveyor.yml
+++ b/.appveyor.yml
@@ -0,0 +1,42 @@
+version: '{build}'
+
+environment:
+  matrix:
+  - MSYSTEM: MINGW64
+    CPU: x86_64
+    MSVC: amd64
+  - MSYSTEM: MINGW32
+    CPU: i686
+    MSVC: x86
+  - MSYSTEM: MINGW64
+    CPU: x86_64
+  - MSYSTEM: MINGW32
+    CPU: i686
+  - MSYSTEM: MINGW64
+    CPU: x86_64
+    MSVC: amd64
+    CONFIG_FLAGS: --enable-debug
+  - MSYSTEM: MINGW32
+    CPU: i686
+    MSVC: x86
+    CONFIG_FLAGS: --enable-debug
+  - MSYSTEM: MINGW64
+    CPU: x86_64
+    CONFIG_FLAGS: --enable-debug
+  - MSYSTEM: MINGW32
+    CPU: i686
+    CONFIG_FLAGS: --enable-debug
+
+install:
+  - set PATH=c:\msys64\%MSYSTEM%\bin;c:\msys64\usr\bin;%PATH%
+  - if defined MSVC call "c:\Program Files (x86)\Microsoft Visual Studio 14.0\VC\vcvarsall.bat" %MSVC%
+  - if defined MSVC pacman --noconfirm -Rsc mingw-w64-%CPU%-gcc gcc
+  - pacman --noconfirm -Suy mingw-w64-%CPU%-make
+
+build_script:
+  - bash -c "autoconf"
+  - bash -c "./configure $CONFIG_FLAGS"
+  - mingw32-make -j3
+  - file lib/jemalloc.dll
+  - mingw32-make -j3 tests
+  - mingw32-make -k check
--- a/.travis.yml
+++ b/.travis.yml
@@ -0,0 +1,95 @@
+language: generic
+
+matrix:
+  include:
+    - os: linux
+      env: CC=gcc COMPILER_FLAGS="" CONFIGURE_FLAGS=""
+    - os: osx
+      env: CC=gcc COMPILER_FLAGS="" CONFIGURE_FLAGS=""
+    - os: linux
+      env: CC=clang COMPILER_FLAGS="" CONFIGURE_FLAGS=""
+    - os: linux
+      env: CC=gcc COMPILER_FLAGS="-m32" CONFIGURE_FLAGS=""
+      addons:
+        apt:
+          packages:
+            - gcc-multilib
+    - os: linux
+      env: CC=gcc COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-debug"
+    - os: linux
+      env: CC=gcc COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-prof"
+    - os: linux
+      env: CC=gcc COMPILER_FLAGS="" CONFIGURE_FLAGS="--disable-stats"
+    - os: linux
+      env: CC=gcc COMPILER_FLAGS="" CONFIGURE_FLAGS="--disable-tcache"
+    - os: osx
+      env: CC=clang COMPILER_FLAGS="" CONFIGURE_FLAGS=""
+    - os: osx
+      env: CC=gcc COMPILER_FLAGS="-m32" CONFIGURE_FLAGS=""
+    - os: osx
+      env: CC=gcc COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-debug"
+    - os: osx
+      env: CC=gcc COMPILER_FLAGS="" CONFIGURE_FLAGS="--disable-stats"
+    - os: osx
+      env: CC=gcc COMPILER_FLAGS="" CONFIGURE_FLAGS="--disable-tcache"
+    - os: linux
+      env: CC=clang COMPILER_FLAGS="-m32" CONFIGURE_FLAGS=""
+      addons:
+        apt:
+          packages:
+            - gcc-multilib
+    - os: linux
+      env: CC=clang COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-debug"
+    - os: linux
+      env: CC=clang COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-prof"
+    - os: linux
+      env: CC=clang COMPILER_FLAGS="" CONFIGURE_FLAGS="--disable-stats"
+    - os: linux
+      env: CC=clang COMPILER_FLAGS="" CONFIGURE_FLAGS="--disable-tcache"
+    - os: linux
+      env: CC=gcc COMPILER_FLAGS="-m32" CONFIGURE_FLAGS="--enable-debug"
+      addons:
+        apt:
+          packages:
+            - gcc-multilib
+    - os: linux
+      env: CC=gcc COMPILER_FLAGS="-m32" CONFIGURE_FLAGS="--enable-prof"
+      addons:
+        apt:
+          packages:
+            - gcc-multilib
+    - os: linux
+      env: CC=gcc COMPILER_FLAGS="-m32" CONFIGURE_FLAGS="--disable-stats"
+      addons:
+        apt:
+          packages:
+            - gcc-multilib
+    - os: linux
+      env: CC=gcc COMPILER_FLAGS="-m32" CONFIGURE_FLAGS="--disable-tcache"
+      addons:
+        apt:
+          packages:
+            - gcc-multilib
+    - os: linux
+      env: CC=gcc COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-debug --enable-prof"
+    - os: linux
+      env: CC=gcc COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-debug --disable-stats"
+    - os: linux
+      env: CC=gcc COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-debug --disable-tcache"
+    - os: linux
+      env: CC=gcc COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-prof --disable-stats"
+    - os: linux
+      env: CC=gcc COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-prof --disable-tcache"
+    - os: linux
+      env: CC=gcc COMPILER_FLAGS="" CONFIGURE_FLAGS="--disable-stats --disable-tcache"
+
+
+before_script:
+  - autoconf
+  - ./configure ${COMPILER_FLAGS:+       CC="$CC $COMPILER_FLAGS" }       $CONFIGURE_FLAGS
+  - make -j3
+  - make -j3 tests
+
+script:
+  - make check
+
--- a/4
+++ b/4
@@ -1,10 +1,10 @@
 Unless otherwise specified, files in the jemalloc source distribution are
 subject to the following license:
 --------------------------------------------------------------------------------
-Copyright (C) 2002-2015 Jason Evans <jasone@canonware.com>.
+Copyright (C) 2002-2016 Jason Evans <jasone@canonware.com>.
 All rights reserved.
 Copyright (C) 2007-2012 Mozilla Foundation.  All rights reserved.
-Copyright (C) 2009-2015 Facebook, Inc.  All rights reserved.
+Copyright (C) 2009-2016 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:
--- a/155
+++ b/155
@@ -4,6 +4,155 @@ brevity.  Much more detail can be found in the git revision history:

    https://github.com/jemalloc/jemalloc

+* 4.5.0 (February 28, 2017)
+
+  This is the first release to benefit from much broader continuous integration
+  testing, thanks to @davidtgoldblatt.  Had we had this testing infrastructure
+  in place for prior releases, it would have caught all of the most serious
+  regressions fixed by this release.
+
+  New features:
+  - Add --disable-thp and the opt.thp to provide opt-out mechanisms for
+    transparent huge page integration.  (@jasone)
+  - Update zone allocator integration to work with macOS 10.12.  (@glandium)
+  - Restructure *CFLAGS configuration, so that CFLAGS behaves typically, and
+    EXTRA_CFLAGS provides a way to specify e.g. -Werror during building, but not
+    during configuration.  (@jasone, @ronawho)
+
+  Bug fixes:
+  - Fix DSS (sbrk(2)-based) allocation.  This regression was first released in
+    4.3.0.  (@jasone)
+  - Handle race in per size class utilization computation.  This functionality
+    was first released in 4.0.0.  (@interwq)
+  - Fix lock order reversal during gdump.  (@jasone)
+  - Fix-refactor tcache synchronization.  This regression was first released in
+    4.0.0.  (@jasone)
+  - Fix various JSON-formatted malloc_stats_print() bugs.  This functionality
+    was first released in 4.3.0.  (@jasone)
+  - Fix huge-aligned allocation.  This regression was first released in 4.4.0.
+    (@jasone)
+  - When transparent huge page integration is enabled, detect what state pages
+    start in according to the kernel's current operating mode, and only convert
+    arena chunks to non-huge during purging if that is not their initial state.
+    This functionality was first released in 4.4.0.  (@jasone)
+  - Fix lg_chunk clamping for the --enable-cache-oblivious --disable-fill case.
+    This regression was first released in 4.0.0.  (@jasone, @428desmo)
+  - Properly detect sparc64 when building for Linux.  (@glaubitz)
+
+* 4.4.0 (December 3, 2016)
+
+  New features:
+  - Add configure support for *-*-linux-android.  (@cferris1000, @jasone)
+  - Add the --disable-syscall configure option, for use on systems that place
+    security-motivated limitations on syscall(2).  (@jasone)
+  - Add support for Debian GNU/kFreeBSD.  (@thesam)
+
+  Optimizations:
+  - Add extent serial numbers and use them where appropriate as a sort key that
+    is higher priority than address, so that the allocation policy prefers older
+    extents.  This tends to improve locality (decrease fragmentation) when
+    memory grows downward.  (@jasone)
+  - Refactor madvise(2) configuration so that MADV_FREE is detected and utilized
+    on Linux 4.5 and newer.  (@jasone)
+  - Mark partially purged arena chunks as non-huge-page.  This improves
+    interaction with Linux's transparent huge page functionality.  (@jasone)
+
+  Bug fixes:
+  - Fix size class computations for edge conditions involving extremely large
+    allocations.  This regression was first released in 4.0.0.  (@jasone,
+    @ingvarha)
+  - Remove overly restrictive assertions related to the cactive statistic.  This
+    regression was first released in 4.1.0.  (@jasone)
+  - Implement a more reliable detection scheme for os_unfair_lock on macOS.
+    (@jszakmeister)
+
+* 4.3.1 (November 7, 2016)
+
+  Bug fixes:
+  - Fix a severe virtual memory leak.  This regression was first released in
+    4.3.0.  (@interwq, @jasone)
+  - Refactor atomic and prng APIs to restore support for 32-bit platforms that
+    use pre-C11 toolchains, e.g. FreeBSD's mips.  (@jasone)
+
+* 4.3.0 (November 4, 2016)
+
+  This is the first release that passes the test suite for multiple Windows
+  configurations, thanks in large part to @glandium setting up continuous
+  integration via AppVeyor (and Travis CI for Linux and OS X).
+
+  New features:
+  - Add "J" (JSON) support to malloc_stats_print().  (@jasone)
+  - Add Cray compiler support.  (@ronawho)
+
+  Optimizations:
+  - Add/use adaptive spinning for bootstrapping and radix tree node
+    initialization.  (@jasone)
+
+  Bug fixes:
+  - Fix large allocation to search starting in the optimal size class heap,
+    which can substantially reduce virtual memory churn and fragmentation.  This
+    regression was first released in 4.0.0.  (@mjp41, @jasone)
+  - Fix stats.arenas.<i>.nthreads accounting.  (@interwq)
+  - Fix and simplify decay-based purging.  (@jasone)
+  - Make DSS (sbrk(2)-related) operations lockless, which resolves potential
+    deadlocks during thread exit.  (@jasone)
+  - Fix over-sized allocation of radix tree leaf nodes.  (@mjp41, @ogaun,
+    @jasone)
+  - Fix over-sized allocation of arena_t (plus associated stats) data
+    structures.  (@jasone, @interwq)
+  - Fix EXTRA_CFLAGS to not affect configuration.  (@jasone)
+  - Fix a Valgrind integration bug.  (@ronawho)
+  - Disallow 0x5a junk filling when running in Valgrind.  (@jasone)
+  - Fix a file descriptor leak on Linux.  This regression was first released in
+    4.2.0.  (@vsarunas, @jasone)
+  - Fix static linking of jemalloc with glibc.  (@djwatson)
+  - Use syscall(2) rather than {open,read,close}(2) during boot on Linux.  This
+    works around other libraries' system call wrappers performing reentrant
+    allocation.  (@kspinka, @Whissi, @jasone)
+  - Fix OS X default zone replacement to work with OS X 10.12.  (@glandium,
+    @jasone)
+  - Fix cached memory management to avoid needless commit/decommit operations
+    during purging, which resolves permanent virtual memory map fragmentation
+    issues on Windows.  (@mjp41, @jasone)
+  - Fix TSD fetches to avoid (recursive) allocation.  This is relevant to
+    non-TLS and Windows configurations.  (@jasone)
+  - Fix malloc_conf overriding to work on Windows.  (@jasone)
+  - Forcibly disable lazy-lock on Windows (was forcibly *enabled*).  (@jasone)
+
+* 4.2.1 (June 8, 2016)
+
+  Bug fixes:
+  - Fix bootstrapping issues for configurations that require allocation during
+    tsd initialization (e.g. --disable-tls).  (@cferris1000, @jasone)
+  - Fix gettimeofday() version of nstime_update().  (@ronawho)
+  - Fix Valgrind regressions in calloc() and chunk_alloc_wrapper().  (@ronawho)
+  - Fix potential VM map fragmentation regression.  (@jasone)
+  - Fix opt_zero-triggered in-place huge reallocation zeroing.  (@jasone)
+  - Fix heap profiling context leaks in reallocation edge cases.  (@jasone)
+
+* 4.2.0 (May 12, 2016)
+
+  New features:
+  - Add the arena.<i>.reset mallctl, which makes it possible to discard all of
+    an arena's allocations in a single operation.  (@jasone)
+  - Add the stats.retained and stats.arenas.<i>.retained statistics.  (@jasone)
+  - Add the --with-version configure option.  (@jasone)
+  - Support --with-lg-page values larger than actual page size.  (@jasone)
+
+  Optimizations:
+  - Use pairing heaps rather than red-black trees for various hot data
+    structures.  (@djwatson, @jasone)
+  - Streamline fast paths of rtree operations.  (@jasone)
+  - Optimize the fast paths of calloc() and [m,d,sd]allocx().  (@jasone)
+  - Decommit unused virtual memory if the OS does not overcommit.  (@jasone)
+  - Specify MAP_NORESERVE on Linux if [heuristic] overcommit is active, in order
+    to avoid unfortunate interactions during fork(2).  (@jasone)
+
+  Bug fixes:
+  - Fix chunk accounting related to triggering gdump profiles.  (@jasone)
+  - Link against librt for clock_gettime(2) if glibc < 2.17.  (@jasone)
+  - Scale leak report summary according to sampling probability.  (@jasone)
+
 * 4.1.1 (May 3, 2016)

  This bugfix release resolves a variety of mostly minor issues, though the
@@ -21,7 +170,7 @@ brevity.  Much more detail can be found in the git revision history:
    enabled and active.  (@jasone)
  - Fix various chunk leaks in OOM code paths.  (@jasone)
  - Fix malloc_stats_print() to print opt.narenas correctly.  (@jasone)
-  - Fix MSVC-specific build/test issues.  (@rustyx, yuslepukhin)
+  - Fix MSVC-specific build/test issues.  (@rustyx, @yuslepukhin)
  - Fix a variety of test failures that were due to test fragility rather than
    core bugs.  (@jasone)

@@ -80,14 +229,14 @@ brevity.  Much more detail can be found in the git revision history:
  Bug fixes:
  - Fix stats.cactive accounting regression.  (@rustyx, @jasone)
  - Handle unaligned keys in hash().  This caused problems for some ARM systems.
-    (@jasone, Christopher Ferris)
+    (@jasone, @cferris1000)
  - Refactor arenas array.  In addition to fixing a fork-related deadlock, this
    makes arena lookups faster and simpler.  (@jasone)
  - Move retained memory allocation out of the default chunk allocation
    function, to a location that gets executed even if the application installs
    a custom chunk allocation function.  This resolves a virtual memory leak.
    (@buchgr)
-  - Fix a potential tsd cleanup leak.  (Christopher Ferris, @jasone)
+  - Fix a potential tsd cleanup leak.  (@cferris1000, @jasone)
  - Fix run quantization.  In practice this bug had no impact unless
    applications requested memory with alignment exceeding one page.
    (@jasone, @djwatson)
--- a/42
+++ b/42
@@ -35,6 +35,10 @@ any of the following arguments (not a definitive list) to 'configure':
    will cause files to be installed into /usr/local/include, /usr/local/lib,
    and /usr/local/man.

+--with-version=<major>.<minor>.<bugfix>-<nrev>-g<gid>
+    Use the specified version string rather than trying to generate one (if in
+    a git repository) or use existing the VERSION file (if present).
+
 --with-rpath=<colon-separated-rpath>
    Embed one or more library paths, so that libjemalloc can find the libraries
    it is linked to.  This works only on ELF-based systems.
@@ -153,6 +157,13 @@ any of the following arguments (not a definitive list) to 'configure':
    released in bulk, thus reducing the total number of mutex operations.  See
    the "opt.tcache" option for usage details.

+--disable-thp
+    Disable transparent huge page (THP) integration.  On systems with THP
+    support, THPs are explicitly disabled as a side effect of unused dirty page
+    purging for chunks that back small and/or large allocations, because such
+    chunks typically comprise active, unused dirty, and untouched clean
+    pages.
+
 --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.
@@ -202,6 +213,11 @@ any of the following arguments (not a definitive list) to 'configure':
    most extreme case increases physical memory usage for the 16 KiB size class
    to 20 KiB.

+--disable-syscall
+    Disable use of syscall(2) rather than {open,read,write,close}(2).  This is
+    intended as a workaround for systems that place security limitations on
+    syscall(2).
+
 --with-xslroot=<path>
    Specify where to find DocBook XSL stylesheets when building the
    documentation.
@@ -297,17 +313,16 @@ The following environment variables (not a definitive list) impact configure's
 behavior:

 CFLAGS="?"
-    Pass these flags to the compiler.  You probably shouldn't define this unless
-    you know what you are doing.  (Use EXTRA_CFLAGS instead.)
+    Pass these flags to the C compiler.  Any flags set by the configure script
+    are prepended, which means explicitly set flags generally take precedence.
+    Take care when specifying flags such as -Werror, because configure tests may
+    be affected in undesirable ways.

 EXTRA_CFLAGS="?"
-    Append these flags to CFLAGS.  This makes it possible to add flags such as
-    -Werror, while allowing the configure script to determine what other flags
-    are appropriate for the specified configuration.
-
-    The configure script specifically checks whether an optimization flag (-O*)
-    is specified in EXTRA_CFLAGS, and refrains from specifying an optimization
-    level if it finds that one has already been specified.
+    Append these flags to CFLAGS, without passing them to the compiler during
+    configuration.  This makes it possible to add flags such as -Werror, while
+    allowing the configure script to determine what other flags are appropriate
+    for the specified configuration.

 CPPFLAGS="?"
    Pass these flags to the C preprocessor.  Note that CFLAGS is not passed to
@@ -323,6 +338,15 @@ LDFLAGS="?"
 PATH="?"
    'configure' uses this to find programs.

+In some cases it may be necessary to work around configuration results that do
+not match reality.  For example, Linux 4.5 added support for the MADV_FREE flag
+to madvise(2), which can cause problems if building on a host with MADV_FREE
+support and deploying to a target without.  To work around this, use a cache
+file to override the relevant configuration variable defined in configure.ac,
+e.g.:
+
+    echo "je_cv_madv_free=no" > config.cache && ./configure -C
+
 === Advanced compilation =======================================================

 To build only parts of jemalloc, use the following targets:
--- a/Makefile.in
+++ b/Makefile.in
@@ -24,11 +24,13 @@ abs_objroot := @abs_objroot@

 # Build parameters.
 CPPFLAGS := @CPPFLAGS@ -I$(srcroot)include -I$(objroot)include
-CFLAGS := @CFLAGS@
+CONFIGURE_CFLAGS := @CONFIGURE_CFLAGS@
+SPECIFIED_CFLAGS := @SPECIFIED_CFLAGS@
+EXTRA_CFLAGS := @EXTRA_CFLAGS@
+CFLAGS := $(strip $(CONFIGURE_CFLAGS) $(SPECIFIED_CFLAGS) $(EXTRA_CFLAGS))
 LDFLAGS := @LDFLAGS@
 EXTRA_LDFLAGS := @EXTRA_LDFLAGS@
 LIBS := @LIBS@
-TESTLIBS := @TESTLIBS@
 RPATH_EXTRA := @RPATH_EXTRA@
 SO := @so@
 IMPORTLIB := @importlib@
@@ -53,15 +55,19 @@ enable_prof := @enable_prof@
 enable_valgrind := @enable_valgrind@
 enable_zone_allocator := @enable_zone_allocator@
 MALLOC_CONF := @JEMALLOC_CPREFIX@MALLOC_CONF
+link_whole_archive := @link_whole_archive@
 DSO_LDFLAGS = @DSO_LDFLAGS@
 SOREV = @SOREV@
 PIC_CFLAGS = @PIC_CFLAGS@
 CTARGET = @CTARGET@
 LDTARGET = @LDTARGET@
+TEST_LD_MODE = @TEST_LD_MODE@
 MKLIB = @MKLIB@
 AR = @AR@
 ARFLAGS = @ARFLAGS@
 CC_MM = @CC_MM@
+LM := @LM@
+INSTALL = @INSTALL@

 ifeq (macho, $(ABI))
 TEST_LIBRARY_PATH := DYLD_FALLBACK_LIBRARY_PATH="$(objroot)lib"
@@ -100,10 +106,12 @@ C_SRCS := $(srcroot)src/jemalloc.c \
 	$(srcroot)src/quarantine.c \
 	$(srcroot)src/rtree.c \
 	$(srcroot)src/stats.c \
+	$(srcroot)src/spin.c \
 	$(srcroot)src/tcache.c \
 	$(srcroot)src/ticker.c \
 	$(srcroot)src/tsd.c \
-	$(srcroot)src/util.c
+	$(srcroot)src/util.c \
+	$(srcroot)src/witness.c
 ifeq ($(enable_valgrind), 1)
 C_SRCS += $(srcroot)src/valgrind.c
 endif
@@ -122,6 +130,11 @@ DSOS := $(objroot)lib/$(LIBJEMALLOC).$(SOREV)
 ifneq ($(SOREV),$(SO))
 DSOS += $(objroot)lib/$(LIBJEMALLOC).$(SO)
 endif
+ifeq (1, $(link_whole_archive))
+LJEMALLOC := -Wl,--whole-archive -L$(objroot)lib -l$(LIBJEMALLOC) -Wl,--no-whole-archive
+else
+LJEMALLOC := $(objroot)lib/$(LIBJEMALLOC).$(IMPORTLIB)
+endif
 PC := $(objroot)jemalloc.pc
 MAN3 := $(objroot)doc/jemalloc$(install_suffix).3
 DOCS_XML := $(objroot)doc/jemalloc$(install_suffix).xml
@@ -133,11 +146,19 @@ C_TESTLIB_SRCS := $(srcroot)test/src/btalloc.c $(srcroot)test/src/btalloc_0.c \
 	$(srcroot)test/src/mtx.c $(srcroot)test/src/mq.c \
 	$(srcroot)test/src/SFMT.c $(srcroot)test/src/test.c \
 	$(srcroot)test/src/thd.c $(srcroot)test/src/timer.c
+ifeq (1, $(link_whole_archive))
+C_UTIL_INTEGRATION_SRCS :=
+else
 C_UTIL_INTEGRATION_SRCS := $(srcroot)src/nstime.c $(srcroot)src/util.c
-TESTS_UNIT := $(srcroot)test/unit/atomic.c \
+endif
+TESTS_UNIT := \
+	$(srcroot)test/unit/a0.c \
+	$(srcroot)test/unit/arena_reset.c \
+	$(srcroot)test/unit/atomic.c \
 	$(srcroot)test/unit/bitmap.c \
 	$(srcroot)test/unit/ckh.c \
 	$(srcroot)test/unit/decay.c \
+	$(srcroot)test/unit/extent_quantize.c \
 	$(srcroot)test/unit/fork.c \
 	$(srcroot)test/unit/hash.c \
 	$(srcroot)test/unit/junk.c \
@@ -148,6 +169,9 @@ TESTS_UNIT := $(srcroot)test/unit/atomic.c \
 	$(srcroot)test/unit/math.c \
 	$(srcroot)test/unit/mq.c \
 	$(srcroot)test/unit/mtx.c \
+	$(srcroot)test/unit/pack.c \
+	$(srcroot)test/unit/pages.c \
+	$(srcroot)test/unit/ph.c \
 	$(srcroot)test/unit/prng.c \
 	$(srcroot)test/unit/prof_accum.c \
 	$(srcroot)test/unit/prof_active.c \
@@ -165,10 +189,12 @@ TESTS_UNIT := $(srcroot)test/unit/atomic.c \
 	$(srcroot)test/unit/size_classes.c \
 	$(srcroot)test/unit/smoothstep.c \
 	$(srcroot)test/unit/stats.c \
+	$(srcroot)test/unit/stats_print.c \
 	$(srcroot)test/unit/ticker.c \
 	$(srcroot)test/unit/nstime.c \
 	$(srcroot)test/unit/tsd.c \
 	$(srcroot)test/unit/util.c \
+	$(srcroot)test/unit/witness.c \
 	$(srcroot)test/unit/zero.c
 TESTS_INTEGRATION := $(srcroot)test/integration/aligned_alloc.c \
 	$(srcroot)test/integration/allocated.c \
@@ -290,69 +316,69 @@ $(STATIC_LIBS):

 $(objroot)test/unit/%$(EXE): $(objroot)test/unit/%.$(O) $(TESTS_UNIT_LINK_OBJS) $(C_JET_OBJS) $(C_TESTLIB_UNIT_OBJS)
 	@mkdir -p $(@D)
-	$(CC) $(LDTARGET) $(filter %.$(O),$^) $(call RPATH,$(objroot)lib) $(LDFLAGS) $(filter-out -lm,$(LIBS)) -lm $(TESTLIBS) $(EXTRA_LDFLAGS)
+	$(CC) $(LDTARGET) $(filter %.$(O),$^) $(call RPATH,$(objroot)lib) $(LDFLAGS) $(filter-out -lm,$(LIBS)) $(LM) $(EXTRA_LDFLAGS)

 $(objroot)test/integration/%$(EXE): $(objroot)test/integration/%.$(O) $(C_TESTLIB_INTEGRATION_OBJS) $(C_UTIL_INTEGRATION_OBJS) $(objroot)lib/$(LIBJEMALLOC).$(IMPORTLIB)
 	@mkdir -p $(@D)
-	$(CC) $(LDTARGET) $(filter %.$(O),$^) $(call RPATH,$(objroot)lib) $(objroot)lib/$(LIBJEMALLOC).$(IMPORTLIB) $(LDFLAGS) $(filter-out -lm,$(filter -lpthread,$(LIBS))) -lm $(TESTLIBS) $(EXTRA_LDFLAGS)
+	$(CC) $(TEST_LD_MODE) $(LDTARGET) $(filter %.$(O),$^) $(call RPATH,$(objroot)lib) $(LJEMALLOC) $(LDFLAGS) $(filter-out -lm,$(filter -lrt -lpthread,$(LIBS))) $(LM) $(EXTRA_LDFLAGS)

 $(objroot)test/stress/%$(EXE): $(objroot)test/stress/%.$(O) $(C_JET_OBJS) $(C_TESTLIB_STRESS_OBJS) $(objroot)lib/$(LIBJEMALLOC).$(IMPORTLIB)
 	@mkdir -p $(@D)
-	$(CC) $(LDTARGET) $(filter %.$(O),$^) $(call RPATH,$(objroot)lib) $(objroot)lib/$(LIBJEMALLOC).$(IMPORTLIB) $(LDFLAGS) $(filter-out -lm,$(LIBS)) -lm $(TESTLIBS) $(EXTRA_LDFLAGS)
+	$(CC) $(TEST_LD_MODE) $(LDTARGET) $(filter %.$(O),$^) $(call RPATH,$(objroot)lib) $(objroot)lib/$(LIBJEMALLOC).$(IMPORTLIB) $(LDFLAGS) $(filter-out -lm,$(LIBS)) $(LM) $(EXTRA_LDFLAGS)

 build_lib_shared: $(DSOS)
 build_lib_static: $(STATIC_LIBS)
 build_lib: build_lib_shared build_lib_static

 install_bin:
-	install -d $(BINDIR)
+	$(INSTALL) -d $(BINDIR)
 	@for b in $(BINS); do \
-	echo "install -m 755 $$b $(BINDIR)"; \
-	install -m 755 $$b $(BINDIR); \
+	echo "$(INSTALL) -m 755 $$b $(BINDIR)"; \
+	$(INSTALL) -m 755 $$b $(BINDIR); \
 done

 install_include:
-	install -d $(INCLUDEDIR)/jemalloc
+	$(INSTALL) -d $(INCLUDEDIR)/jemalloc
 	@for h in $(C_HDRS); do \
-	echo "install -m 644 $$h $(INCLUDEDIR)/jemalloc"; \
-	install -m 644 $$h $(INCLUDEDIR)/jemalloc; \
+	echo "$(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)
+	$(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)
+	$(INSTALL) -d $(LIBDIR)
 	@for l in $(STATIC_LIBS); do \
-	echo "install -m 755 $$l $(LIBDIR)"; \
-	install -m 755 $$l $(LIBDIR); \
+	echo "$(INSTALL) -m 755 $$l $(LIBDIR)"; \
+	$(INSTALL) -m 755 $$l $(LIBDIR); \
 done

 install_lib_pc: $(PC)
-	install -d $(LIBDIR)/pkgconfig
+	$(INSTALL) -d $(LIBDIR)/pkgconfig
 	@for l in $(PC); do \
-	echo "install -m 644 $$l $(LIBDIR)/pkgconfig"; \
-	install -m 644 $$l $(LIBDIR)/pkgconfig; \
+	echo "$(INSTALL) -m 644 $$l $(LIBDIR)/pkgconfig"; \
+	$(INSTALL) -m 644 $$l $(LIBDIR)/pkgconfig; \
 done

 install_lib: install_lib_shared install_lib_static install_lib_pc

 install_doc_html:
-	install -d $(DATADIR)/doc/jemalloc$(install_suffix)
+	$(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); \
+	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
+	$(INSTALL) -d $(MANDIR)/man3
 	@for d in $(DOCS_MAN3); do \
-	echo "install -m 644 $$d $(MANDIR)/man3"; \
-	install -m 644 $$d $(MANDIR)/man3; \
+	echo "$(INSTALL) -m 644 $$d $(MANDIR)/man3"; \
+	$(INSTALL) -m 644 $$d $(MANDIR)/man3; \
 done

 install_doc: install_doc_html install_doc_man
--- a/2
+++ b/2
@@ -17,4 +17,4 @@ jemalloc.

 The ChangeLog file contains a brief summary of changes for each release.

-URL: http://www.canonware.com/jemalloc/
+URL: http://jemalloc.net/
--- a/build-aux/config.guess
+++ b/build-aux/config.guess
@@ -1,8 +1,8 @@
 #! /bin/sh
 # Attempt to guess a canonical system name.
-#   Copyright 1992-2014 Free Software Foundation, Inc.
+#   Copyright 1992-2016 Free Software Foundation, Inc.

-timestamp='2014-03-23'
+timestamp='2016-10-02'

 # This file is free software; you can redistribute it and/or modify it
 # under the terms of the GNU General Public License as published by
@@ -24,12 +24,12 @@ timestamp='2014-03-23'
 # program.  This Exception is an additional permission under section 7
 # of the GNU General Public License, version 3 ("GPLv3").
 #
-# Originally written by Per Bothner.
+# Originally written by Per Bothner; maintained since 2000 by Ben Elliston.
 #
 # You can get the latest version of this script from:
-# http://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.guess;hb=HEAD
+# http://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.guess
 #
-# Please send patches with a ChangeLog entry to config-patches@gnu.org.
+# Please send patches to <config-patches@gnu.org>.


 me=`echo "$0" | sed -e 's,.*/,,'`
@@ -50,7 +50,7 @@ version="\
 GNU config.guess ($timestamp)

 Originally written by Per Bothner.
-Copyright 1992-2014 Free Software Foundation, Inc.
+Copyright 1992-2016 Free Software Foundation, Inc.

 This is free software; see the source for copying conditions.  There is NO
 warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE."
@@ -168,19 +168,29 @@ case "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" in
 	# Note: NetBSD doesn't particularly care about the vendor
 	# portion of the name.  We always set it to "unknown".
 	sysctl="sysctl -n hw.machine_arch"
-	UNAME_MACHINE_ARCH=`(/sbin/$sysctl 2>/dev/null || \
-	    /usr/sbin/$sysctl 2>/dev/null || echo unknown)`
+	UNAME_MACHINE_ARCH=`(uname -p 2>/dev/null || \
+	    /sbin/$sysctl 2>/dev/null || \
+	    /usr/sbin/$sysctl 2>/dev/null || \
+	    echo unknown)`
 	case "${UNAME_MACHINE_ARCH}" in
 	    armeb) machine=armeb-unknown ;;
 	    arm*) machine=arm-unknown ;;
 	    sh3el) machine=shl-unknown ;;
 	    sh3eb) machine=sh-unknown ;;
 	    sh5el) machine=sh5le-unknown ;;
+	    earmv*)
+		arch=`echo ${UNAME_MACHINE_ARCH} | sed -e 's,^e\(armv[0-9]\).*$,\1,'`
+		endian=`echo ${UNAME_MACHINE_ARCH} | sed -ne 's,^.*\(eb\)$,\1,p'`
+		machine=${arch}${endian}-unknown
+		;;
 	    *) machine=${UNAME_MACHINE_ARCH}-unknown ;;
 	esac
 	# The Operating System including object format, if it has switched
-	# to ELF recently, or will in the future.
+	# to ELF recently (or will in the future) and ABI.
 	case "${UNAME_MACHINE_ARCH}" in
+	    earm*)
+		os=netbsdelf
+		;;
 	    arm*|i386|m68k|ns32k|sh3*|sparc|vax)
 		eval $set_cc_for_build
 		if echo __ELF__ | $CC_FOR_BUILD -E - 2>/dev/null \
@@ -197,6 +207,13 @@ case "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" in
 		os=netbsd
 		;;
 	esac
+	# Determine ABI tags.
+	case "${UNAME_MACHINE_ARCH}" in
+	    earm*)
+		expr='s/^earmv[0-9]/-eabi/;s/eb$//'
+		abi=`echo ${UNAME_MACHINE_ARCH} | sed -e "$expr"`
+		;;
+	esac
 	# The OS release
 	# Debian GNU/NetBSD machines have a different userland, and
 	# thus, need a distinct triplet. However, they do not need
@@ -207,13 +224,13 @@ case "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" in
 		release='-gnu'
 		;;
 	    *)
-		release=`echo ${UNAME_RELEASE}|sed -e 's/[-_].*/\./'`
+		release=`echo ${UNAME_RELEASE} | sed -e 's/[-_].*//' | cut -d. -f1,2`
 		;;
 	esac
 	# Since CPU_TYPE-MANUFACTURER-KERNEL-OPERATING_SYSTEM:
 	# contains redundant information, the shorter form:
 	# CPU_TYPE-MANUFACTURER-OPERATING_SYSTEM is used.
-	echo "${machine}-${os}${release}"
+	echo "${machine}-${os}${release}${abi}"
 	exit ;;
    *:Bitrig:*:*)
 	UNAME_MACHINE_ARCH=`arch | sed 's/Bitrig.//'`
@@ -223,6 +240,10 @@ case "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" in
 	UNAME_MACHINE_ARCH=`arch | sed 's/OpenBSD.//'`
 	echo ${UNAME_MACHINE_ARCH}-unknown-openbsd${UNAME_RELEASE}
 	exit ;;
+    *:LibertyBSD:*:*)
+	UNAME_MACHINE_ARCH=`arch | sed 's/^.*BSD\.//'`
+	echo ${UNAME_MACHINE_ARCH}-unknown-libertybsd${UNAME_RELEASE}
+	exit ;;
    *:ekkoBSD:*:*)
 	echo ${UNAME_MACHINE}-unknown-ekkobsd${UNAME_RELEASE}
 	exit ;;
@@ -235,6 +256,9 @@ case "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" in
    *:MirBSD:*:*)
 	echo ${UNAME_MACHINE}-unknown-mirbsd${UNAME_RELEASE}
 	exit ;;
+    *:Sortix:*:*)
+	echo ${UNAME_MACHINE}-unknown-sortix
+	exit ;;
    alpha:OSF1:*:*)
 	case $UNAME_RELEASE in
 	*4.0)
@@ -251,42 +275,42 @@ case "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" in
 	ALPHA_CPU_TYPE=`/usr/sbin/psrinfo -v | sed -n -e 's/^  The alpha \(.*\) processor.*$/\1/p' | head -n 1`
 	case "$ALPHA_CPU_TYPE" in
 	    "EV4 (21064)")
-		UNAME_MACHINE="alpha" ;;
+		UNAME_MACHINE=alpha ;;
 	    "EV4.5 (21064)")
-		UNAME_MACHINE="alpha" ;;
+		UNAME_MACHINE=alpha ;;
 	    "LCA4 (21066/21068)")
-		UNAME_MACHINE="alpha" ;;
+		UNAME_MACHINE=alpha ;;
 	    "EV5 (21164)")
-		UNAME_MACHINE="alphaev5" ;;
+		UNAME_MACHINE=alphaev5 ;;
 	    "EV5.6 (21164A)")
-		UNAME_MACHINE="alphaev56" ;;
+		UNAME_MACHINE=alphaev56 ;;
 	    "EV5.6 (21164PC)")
-		UNAME_MACHINE="alphapca56" ;;
+		UNAME_MACHINE=alphapca56 ;;
 	    "EV5.7 (21164PC)")
-		UNAME_MACHINE="alphapca57" ;;
+		UNAME_MACHINE=alphapca57 ;;
 	    "EV6 (21264)")
-		UNAME_MACHINE="alphaev6" ;;
+		UNAME_MACHINE=alphaev6 ;;
 	    "EV6.7 (21264A)")
-		UNAME_MACHINE="alphaev67" ;;
+		UNAME_MACHINE=alphaev67 ;;
 	    "EV6.8CB (21264C)")
-		UNAME_MACHINE="alphaev68" ;;
+		UNAME_MACHINE=alphaev68 ;;
 	    "EV6.8AL (21264B)")
-		UNAME_MACHINE="alphaev68" ;;
+		UNAME_MACHINE=alphaev68 ;;
 	    "EV6.8CX (21264D)")
-		UNAME_MACHINE="alphaev68" ;;
+		UNAME_MACHINE=alphaev68 ;;
 	    "EV6.9A (21264/EV69A)")
-		UNAME_MACHINE="alphaev69" ;;
+		UNAME_MACHINE=alphaev69 ;;
 	    "EV7 (21364)")
-		UNAME_MACHINE="alphaev7" ;;
+		UNAME_MACHINE=alphaev7 ;;
 	    "EV7.9 (21364A)")
-		UNAME_MACHINE="alphaev79" ;;
+		UNAME_MACHINE=alphaev79 ;;
 	esac
 	# A Pn.n version is a patched version.
 	# A Vn.n version is a released version.
 	# A Tn.n version is a released field test version.
 	# A Xn.n version is an unreleased experimental baselevel.
 	# 1.2 uses "1.2" for uname -r.
-	echo ${UNAME_MACHINE}-dec-osf`echo ${UNAME_RELEASE} | sed -e 's/^[PVTX]//' | tr 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' 'abcdefghijklmnopqrstuvwxyz'`
+	echo ${UNAME_MACHINE}-dec-osf`echo ${UNAME_RELEASE} | sed -e 's/^[PVTX]//' | tr ABCDEFGHIJKLMNOPQRSTUVWXYZ abcdefghijklmnopqrstuvwxyz`
 	# Reset EXIT trap before exiting to avoid spurious non-zero exit code.
 	exitcode=$?
 	trap '' 0
@@ -359,16 +383,16 @@ case "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" in
 	exit ;;
    i86pc:SunOS:5.*:* | i86xen:SunOS:5.*:*)
 	eval $set_cc_for_build
-	SUN_ARCH="i386"
+	SUN_ARCH=i386
 	# If there is a compiler, see if it is configured for 64-bit objects.
 	# Note that the Sun cc does not turn __LP64__ into 1 like gcc does.
 	# This test works for both compilers.
-	if [ "$CC_FOR_BUILD" != 'no_compiler_found' ]; then
+	if [ "$CC_FOR_BUILD" != no_compiler_found ]; then
 	    if (echo '#ifdef __amd64'; echo IS_64BIT_ARCH; echo '#endif') | \
-		(CCOPTS= $CC_FOR_BUILD -E - 2>/dev/null) | \
+		(CCOPTS="" $CC_FOR_BUILD -E - 2>/dev/null) | \
 		grep IS_64BIT_ARCH >/dev/null
 	    then
-		SUN_ARCH="x86_64"
+		SUN_ARCH=x86_64
 	    fi
 	fi
 	echo ${SUN_ARCH}-pc-solaris2`echo ${UNAME_RELEASE}|sed -e 's/[^.]*//'`
@@ -393,7 +417,7 @@ case "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" in
 	exit ;;
    sun*:*:4.2BSD:*)
 	UNAME_RELEASE=`(sed 1q /etc/motd | awk '{print substr($5,1,3)}') 2>/dev/null`
-	test "x${UNAME_RELEASE}" = "x" && UNAME_RELEASE=3
+	test "x${UNAME_RELEASE}" = x && UNAME_RELEASE=3
 	case "`/bin/arch`" in
 	    sun3)
 		echo m68k-sun-sunos${UNAME_RELEASE}
@@ -579,8 +603,9 @@ EOF
 	else
 		IBM_ARCH=powerpc
 	fi
-	if [ -x /usr/bin/oslevel ] ; then
-		IBM_REV=`/usr/bin/oslevel`
+	if [ -x /usr/bin/lslpp ] ; then
+		IBM_REV=`/usr/bin/lslpp -Lqc bos.rte.libc |
+			   awk -F: '{ print $3 }' | sed s/[0-9]*$/0/`
 	else
 		IBM_REV=${UNAME_VERSION}.${UNAME_RELEASE}
 	fi
@@ -617,13 +642,13 @@ EOF
 		    sc_cpu_version=`/usr/bin/getconf SC_CPU_VERSION 2>/dev/null`
 		    sc_kernel_bits=`/usr/bin/getconf SC_KERNEL_BITS 2>/dev/null`
 		    case "${sc_cpu_version}" in
-		      523) HP_ARCH="hppa1.0" ;; # CPU_PA_RISC1_0
-		      528) HP_ARCH="hppa1.1" ;; # CPU_PA_RISC1_1
+		      523) HP_ARCH=hppa1.0 ;; # CPU_PA_RISC1_0
+		      528) HP_ARCH=hppa1.1 ;; # CPU_PA_RISC1_1
 		      532)                      # CPU_PA_RISC2_0
 			case "${sc_kernel_bits}" in
-			  32) HP_ARCH="hppa2.0n" ;;
-			  64) HP_ARCH="hppa2.0w" ;;
-			  '') HP_ARCH="hppa2.0" ;;   # HP-UX 10.20
+			  32) HP_ARCH=hppa2.0n ;;
+			  64) HP_ARCH=hppa2.0w ;;
+			  '') HP_ARCH=hppa2.0 ;;   # HP-UX 10.20
 			esac ;;
 		    esac
 		fi
@@ -662,11 +687,11 @@ EOF
 		    exit (0);
 		}
 EOF
-		    (CCOPTS= $CC_FOR_BUILD -o $dummy $dummy.c 2>/dev/null) && HP_ARCH=`$dummy`
+		    (CCOPTS="" $CC_FOR_BUILD -o $dummy $dummy.c 2>/dev/null) && HP_ARCH=`$dummy`
 		    test -z "$HP_ARCH" && HP_ARCH=hppa
 		fi ;;
 	esac
-	if [ ${HP_ARCH} = "hppa2.0w" ]
+	if [ ${HP_ARCH} = hppa2.0w ]
 	then
 	    eval $set_cc_for_build

@@ -679,12 +704,12 @@ EOF
 	    # $ CC_FOR_BUILD="cc +DA2.0w" ./config.guess
 	    # => hppa64-hp-hpux11.23

-	    if echo __LP64__ | (CCOPTS= $CC_FOR_BUILD -E - 2>/dev/null) |
+	    if echo __LP64__ | (CCOPTS="" $CC_FOR_BUILD -E - 2>/dev/null) |
 		grep -q __LP64__
 	    then
-		HP_ARCH="hppa2.0w"
+		HP_ARCH=hppa2.0w
 	    else
-		HP_ARCH="hppa64"
+		HP_ARCH=hppa64
 	    fi
 	fi
 	echo ${HP_ARCH}-hp-hpux${HPUX_REV}
@@ -789,14 +814,14 @@ EOF
 	echo craynv-cray-unicosmp${UNAME_RELEASE} | sed -e 's/\.[^.]*$/.X/'
 	exit ;;
    F30[01]:UNIX_System_V:*:* | F700:UNIX_System_V:*:*)
-	FUJITSU_PROC=`uname -m | tr 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' 'abcdefghijklmnopqrstuvwxyz'`
-	FUJITSU_SYS=`uname -p | tr 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' 'abcdefghijklmnopqrstuvwxyz' | sed -e 's/\///'`
+	FUJITSU_PROC=`uname -m | tr ABCDEFGHIJKLMNOPQRSTUVWXYZ abcdefghijklmnopqrstuvwxyz`
+	FUJITSU_SYS=`uname -p | tr ABCDEFGHIJKLMNOPQRSTUVWXYZ abcdefghijklmnopqrstuvwxyz | sed -e 's/\///'`
 	FUJITSU_REL=`echo ${UNAME_RELEASE} | sed -e 's/ /_/'`
 	echo "${FUJITSU_PROC}-fujitsu-${FUJITSU_SYS}${FUJITSU_REL}"
 	exit ;;
    5000:UNIX_System_V:4.*:*)
-	FUJITSU_SYS=`uname -p | tr 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' 'abcdefghijklmnopqrstuvwxyz' | sed -e 's/\///'`
-	FUJITSU_REL=`echo ${UNAME_RELEASE} | tr 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' 'abcdefghijklmnopqrstuvwxyz' | sed -e 's/ /_/'`
+	FUJITSU_SYS=`uname -p | tr ABCDEFGHIJKLMNOPQRSTUVWXYZ abcdefghijklmnopqrstuvwxyz | sed -e 's/\///'`
+	FUJITSU_REL=`echo ${UNAME_RELEASE} | tr ABCDEFGHIJKLMNOPQRSTUVWXYZ abcdefghijklmnopqrstuvwxyz | sed -e 's/ /_/'`
 	echo "sparc-fujitsu-${FUJITSU_SYS}${FUJITSU_REL}"
 	exit ;;
    i*86:BSD/386:*:* | i*86:BSD/OS:*:* | *:Ascend\ Embedded/OS:*:*)
@@ -878,7 +903,7 @@ EOF
 	exit ;;
    *:GNU/*:*:*)
 	# other systems with GNU libc and userland
-	echo ${UNAME_MACHINE}-unknown-`echo ${UNAME_SYSTEM} | sed 's,^[^/]*/,,' | tr '[A-Z]' '[a-z]'``echo ${UNAME_RELEASE}|sed -e 's/[-(].*//'`-${LIBC}
+	echo ${UNAME_MACHINE}-unknown-`echo ${UNAME_SYSTEM} | sed 's,^[^/]*/,,' | tr "[:upper:]" "[:lower:]"``echo ${UNAME_RELEASE}|sed -e 's/[-(].*//'`-${LIBC}
 	exit ;;
    i*86:Minix:*:*)
 	echo ${UNAME_MACHINE}-pc-minix
@@ -901,7 +926,7 @@ EOF
 	  EV68*) UNAME_MACHINE=alphaev68 ;;
 	esac
 	objdump --private-headers /bin/sh | grep -q ld.so.1
-	if test "$?" = 0 ; then LIBC="gnulibc1" ; fi
+	if test "$?" = 0 ; then LIBC=gnulibc1 ; fi
 	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
 	exit ;;
    arc:Linux:*:* | arceb:Linux:*:*)
@@ -932,6 +957,9 @@ EOF
    crisv32:Linux:*:*)
 	echo ${UNAME_MACHINE}-axis-linux-${LIBC}
 	exit ;;
+    e2k:Linux:*:*)
+	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
+	exit ;;
    frv:Linux:*:*)
 	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
 	exit ;;
@@ -944,6 +972,9 @@ EOF
    ia64:Linux:*:*)
 	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
 	exit ;;
+    k1om:Linux:*:*)
+	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
+	exit ;;
    m32r*:Linux:*:*)
 	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
 	exit ;;
@@ -969,6 +1000,9 @@ EOF
 	eval `$CC_FOR_BUILD -E $dummy.c 2>/dev/null | grep '^CPU'`
 	test x"${CPU}" != x && { echo "${CPU}-unknown-linux-${LIBC}"; exit; }
 	;;
+    mips64el:Linux:*:*)
+	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
+	exit ;;
    openrisc*:Linux:*:*)
 	echo or1k-unknown-linux-${LIBC}
 	exit ;;
@@ -1001,6 +1035,9 @@ EOF
    ppcle:Linux:*:*)
 	echo powerpcle-unknown-linux-${LIBC}
 	exit ;;
+    riscv32:Linux:*:* | riscv64:Linux:*:*)
+	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
+	exit ;;
    s390:Linux:*:* | s390x:Linux:*:*)
 	echo ${UNAME_MACHINE}-ibm-linux-${LIBC}
 	exit ;;
@@ -1020,7 +1057,7 @@ EOF
 	echo ${UNAME_MACHINE}-dec-linux-${LIBC}
 	exit ;;
    x86_64:Linux:*:*)
-	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
+	echo ${UNAME_MACHINE}-pc-linux-${LIBC}
 	exit ;;
    xtensa*:Linux:*:*)
 	echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
@@ -1099,7 +1136,7 @@ EOF
 	# uname -m prints for DJGPP always 'pc', but it prints nothing about
 	# the processor, so we play safe by assuming i586.
 	# Note: whatever this is, it MUST be the same as what config.sub
-	# prints for the "djgpp" host, or else GDB configury will decide that
+	# prints for the "djgpp" host, or else GDB configure will decide that
 	# this is a cross-build.
 	echo i586-pc-msdosdjgpp
 	exit ;;
@@ -1248,6 +1285,9 @@ EOF
    SX-8R:SUPER-UX:*:*)
 	echo sx8r-nec-superux${UNAME_RELEASE}
 	exit ;;
+    SX-ACE:SUPER-UX:*:*)
+	echo sxace-nec-superux${UNAME_RELEASE}
+	exit ;;
    Power*:Rhapsody:*:*)
 	echo powerpc-apple-rhapsody${UNAME_RELEASE}
 	exit ;;
@@ -1261,9 +1301,9 @@ EOF
 	    UNAME_PROCESSOR=powerpc
 	fi
 	if test `echo "$UNAME_RELEASE" | sed -e 's/\..*//'` -le 10 ; then
-	    if [ "$CC_FOR_BUILD" != 'no_compiler_found' ]; then
+	    if [ "$CC_FOR_BUILD" != no_compiler_found ]; then
 		if (echo '#ifdef __LP64__'; echo IS_64BIT_ARCH; echo '#endif') | \
-		    (CCOPTS= $CC_FOR_BUILD -E - 2>/dev/null) | \
+		    (CCOPTS="" $CC_FOR_BUILD -E - 2>/dev/null) | \
 		    grep IS_64BIT_ARCH >/dev/null
 		then
 		    case $UNAME_PROCESSOR in
@@ -1285,7 +1325,7 @@ EOF
 	exit ;;
    *:procnto*:*:* | *:QNX:[0123456789]*:*)
 	UNAME_PROCESSOR=`uname -p`
-	if test "$UNAME_PROCESSOR" = "x86"; then
+	if test "$UNAME_PROCESSOR" = x86; then
 		UNAME_PROCESSOR=i386
 		UNAME_MACHINE=pc
 	fi
@@ -1316,7 +1356,7 @@ EOF
 	# "uname -m" is not consistent, so use $cputype instead. 386
 	# is converted to i386 for consistency with other x86
 	# operating systems.
-	if test "$cputype" = "386"; then
+	if test "$cputype" = 386; then
 	    UNAME_MACHINE=i386
 	else
 	    UNAME_MACHINE="$cputype"
@@ -1358,7 +1398,7 @@ EOF
 	echo i386-pc-xenix
 	exit ;;
    i*86:skyos:*:*)
-	echo ${UNAME_MACHINE}-pc-skyos`echo ${UNAME_RELEASE}` | sed -e 's/ .*$//'
+	echo ${UNAME_MACHINE}-pc-skyos`echo ${UNAME_RELEASE} | sed -e 's/ .*$//'`
 	exit ;;
    i*86:rdos:*:*)
 	echo ${UNAME_MACHINE}-pc-rdos
@@ -1369,23 +1409,25 @@ EOF
    x86_64:VMkernel:*:*)
 	echo ${UNAME_MACHINE}-unknown-esx
 	exit ;;
+    amd64:Isilon\ OneFS:*:*)
+	echo x86_64-unknown-onefs
+	exit ;;
 esac

 cat >&2 <<EOF
 $0: unable to guess system type

-This script, last modified $timestamp, has failed to recognize
-the operating system you are using. It is advised that you
-download the most up to date version of the config scripts from
+This script (version $timestamp), has failed to recognize the
+operating system you are using. If your script is old, overwrite
+config.guess and config.sub with the latest versions from:

-  http://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.guess;hb=HEAD
+  http://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.guess
 and
-  http://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.sub;hb=HEAD
+  http://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.sub

-If the version you run ($0) is already up to date, please
-send the following data and any information you think might be
-pertinent to <config-patches@gnu.org> in order to provide the needed
-information to handle your system.
+If $0 has already been updated, send the following data and any
+information you think might be pertinent to config-patches@gnu.org to
+provide the necessary information to handle your system.

 config.guess timestamp = $timestamp

--- a/build-aux/config.sub
+++ b/build-aux/config.sub
@@ -1,8 +1,8 @@
 #! /bin/sh
 # Configuration validation subroutine script.
-#   Copyright 1992-2014 Free Software Foundation, Inc.
+#   Copyright 1992-2016 Free Software Foundation, Inc.

-timestamp='2014-05-01'
+timestamp='2016-11-04'

 # This file is free software; you can redistribute it and/or modify it
 # under the terms of the GNU General Public License as published by
@@ -25,7 +25,7 @@ timestamp='2014-05-01'
 # of the GNU General Public License, version 3 ("GPLv3").


-# Please send patches with a ChangeLog entry to config-patches@gnu.org.
+# Please send patches to <config-patches@gnu.org>.
 #
 # Configuration subroutine to validate and canonicalize a configuration type.
 # Supply the specified configuration type as an argument.
@@ -33,7 +33,7 @@ timestamp='2014-05-01'
 # 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
+# http://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.sub

 # This file is supposed to be the same for all GNU packages
 # and recognize all the CPU types, system types and aliases
@@ -53,8 +53,7 @@ timestamp='2014-05-01'
 me=`echo "$0" | sed -e 's,.*/,,'`

 usage="\
-Usage: $0 [OPTION] CPU-MFR-OPSYS
-       $0 [OPTION] ALIAS
+Usage: $0 [OPTION] CPU-MFR-OPSYS or ALIAS

 Canonicalize a configuration name.

@@ -68,7 +67,7 @@ Report bugs and patches to <config-patches@gnu.org>."
 version="\
 GNU config.sub ($timestamp)

-Copyright 1992-2014 Free Software Foundation, Inc.
+Copyright 1992-2016 Free Software Foundation, Inc.

 This is free software; see the source for copying conditions.  There is NO
 warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE."
@@ -117,8 +116,8 @@ maybe_os=`echo $1 | sed 's/^\(.*\)-\([^-]*-[^-]*\)$/\2/'`
 case $maybe_os in
  nto-qnx* | linux-gnu* | linux-android* | linux-dietlibc | linux-newlib* | \
  linux-musl* | linux-uclibc* | uclinux-uclibc* | uclinux-gnu* | kfreebsd*-gnu* | \
-  knetbsd*-gnu* | netbsd*-gnu* | \
-  kopensolaris*-gnu* | \
+  knetbsd*-gnu* | netbsd*-gnu* | netbsd*-eabi* | \
+  kopensolaris*-gnu* | cloudabi*-eabi* | \
  storm-chaos* | os2-emx* | rtmk-nova*)
    os=-$maybe_os
    basic_machine=`echo $1 | sed 's/^\(.*\)-\([^-]*-[^-]*\)$/\1/'`
@@ -255,12 +254,13 @@ case $basic_machine in
 	| arc | arceb \
 	| arm | arm[bl]e | arme[lb] | armv[2-8] | armv[3-8][lb] | armv7[arm] \
 	| avr | avr32 \
+	| ba \
 	| be32 | be64 \
 	| bfin \
 	| c4x | c8051 | clipper \
 	| d10v | d30v | dlx | dsp16xx \
-	| epiphany \
-	| fido | fr30 | frv \
+	| e2k | epiphany \
+	| fido | fr30 | frv | ft32 \
 	| h8300 | h8500 | hppa | hppa1.[01] | hppa2.0 | hppa2.0[nw] | hppa64 \
 	| hexagon \
 	| i370 | i860 | i960 | ia64 \
@@ -301,10 +301,12 @@ case $basic_machine in
 	| open8 | or1k | or1knd | or32 \
 	| pdp10 | pdp11 | pj | pjl \
 	| powerpc | powerpc64 | powerpc64le | powerpcle \
+	| pru \
 	| pyramid \
+	| riscv32 | riscv64 \
 	| 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 \
+	| sh | sh[1234] | sh[24]a | sh[24]aeb | sh[23]e | sh[234]eb | sheb | shbe | shle | sh[1234]le | sh3ele \
 	| sh64 | sh64le \
 	| sparc | sparc64 | sparc64b | sparc64v | sparc86x | sparclet | sparclite \
 	| sparcv8 | sparcv9 | sparcv9b | sparcv9v \
@@ -312,6 +314,7 @@ case $basic_machine in
 	| tahoe | tic4x | tic54x | tic55x | tic6x | tic80 | tron \
 	| ubicom32 \
 	| v850 | v850e | v850e1 | v850e2 | v850es | v850e2v3 \
+	| visium \
 	| we32k \
 	| x86 | xc16x | xstormy16 | xtensa \
 	| z8k | z80)
@@ -326,6 +329,9 @@ case $basic_machine in
 	c6x)
 		basic_machine=tic6x-unknown
 		;;
+	leon|leon[3-9])
+		basic_machine=sparc-$basic_machine
+		;;
 	m6811 | m68hc11 | m6812 | m68hc12 | m68hcs12x | nvptx | picochip)
 		basic_machine=$basic_machine-unknown
 		os=-none
@@ -371,12 +377,13 @@ case $basic_machine in
 	| alphapca5[67]-* | alpha64pca5[67]-* | arc-* | arceb-* \
 	| arm-*  | armbe-* | armle-* | armeb-* | armv*-* \
 	| avr-* | avr32-* \
+	| ba-* \
 	| be32-* | be64-* \
 	| bfin-* | bs2000-* \
 	| c[123]* | c30-* | [cjt]90-* | c4x-* \
 	| c8051-* | clipper-* | craynv-* | cydra-* \
 	| d10v-* | d30v-* | dlx-* \
-	| elxsi-* \
+	| e2k-* | elxsi-* \
 	| f30[01]-* | f700-* | fido-* | fr30-* | frv-* | fx80-* \
 	| h8300-* | h8500-* \
 	| hppa-* | hppa1.[01]-* | hppa2.0-* | hppa2.0[nw]-* | hppa64-* \
@@ -422,13 +429,15 @@ case $basic_machine in
 	| orion-* \
 	| pdp10-* | pdp11-* | pj-* | pjl-* | pn-* | power-* \
 	| powerpc-* | powerpc64-* | powerpc64le-* | powerpcle-* \
+	| pru-* \
 	| pyramid-* \
+	| riscv32-* | riscv64-* \
 	| 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-* | sparc64b-* | sparc64v-* | sparc86x-* | sparclet-* \
 	| sparclite-* \
-	| sparcv8-* | sparcv9-* | sparcv9b-* | sparcv9v-* | sv1-* | sx?-* \
+	| sparcv8-* | sparcv9-* | sparcv9b-* | sparcv9v-* | sv1-* | sx*-* \
 	| tahoe-* \
 	| tic30-* | tic4x-* | tic54x-* | tic55x-* | tic6x-* | tic80-* \
 	| tile*-* \
@@ -436,6 +445,7 @@ case $basic_machine in
 	| ubicom32-* \
 	| v850-* | v850e-* | v850e1-* | v850es-* | v850e2-* | v850e2v3-* \
 	| vax-* \
+	| visium-* \
 	| we32k-* \
 	| x86-* | x86_64-* | xc16x-* | xps100-* \
 	| xstormy16-* | xtensa*-* \
@@ -512,6 +522,9 @@ case $basic_machine in
 		basic_machine=i386-pc
 		os=-aros
 		;;
+	asmjs)
+		basic_machine=asmjs-unknown
+		;;
 	aux)
 		basic_machine=m68k-apple
 		os=-aux
@@ -632,6 +645,14 @@ case $basic_machine in
 		basic_machine=m68k-bull
 		os=-sysv3
 		;;
+	e500v[12])
+		basic_machine=powerpc-unknown
+		os=$os"spe"
+		;;
+	e500v[12]-*)
+		basic_machine=powerpc-`echo $basic_machine | sed 's/^[^-]*-//'`
+		os=$os"spe"
+		;;
 	ebmon29k)
 		basic_machine=a29k-amd
 		os=-ebmon
@@ -773,6 +794,9 @@ case $basic_machine in
 		basic_machine=m68k-isi
 		os=-sysv
 		;;
+	leon-*|leon[3-9]-*)
+		basic_machine=sparc-`echo $basic_machine | sed 's/-.*//'`
+		;;
 	m68knommu)
 		basic_machine=m68k-unknown
 		os=-linux
@@ -828,6 +852,10 @@ case $basic_machine in
 		basic_machine=powerpc-unknown
 		os=-morphos
 		;;
+	moxiebox)
+		basic_machine=moxie-unknown
+		os=-moxiebox
+		;;
 	msdos)
 		basic_machine=i386-pc
 		os=-msdos
@@ -1004,7 +1032,7 @@ case $basic_machine in
 	ppc-* | ppcbe-*)
 		basic_machine=powerpc-`echo $basic_machine | sed 's/^[^-]*-//'`
 		;;
-	ppcle | powerpclittle | ppc-le | powerpc-little)
+	ppcle | powerpclittle)
 		basic_machine=powerpcle-unknown
 		;;
 	ppcle-* | powerpclittle-*)
@@ -1014,7 +1042,7 @@ case $basic_machine in
 		;;
 	ppc64-*) basic_machine=powerpc64-`echo $basic_machine | sed 's/^[^-]*-//'`
 		;;
-	ppc64le | powerpc64little | ppc64-le | powerpc64-little)
+	ppc64le | powerpc64little)
 		basic_machine=powerpc64le-unknown
 		;;
 	ppc64le-* | powerpc64little-*)
@@ -1360,27 +1388,28 @@ case $os in
 	      | -hpux* | -unos* | -osf* | -luna* | -dgux* | -auroraux* | -solaris* \
 	      | -sym* | -kopensolaris* | -plan9* \
 	      | -amigaos* | -amigados* | -msdos* | -newsos* | -unicos* | -aof* \
-	      | -aos* | -aros* \
+	      | -aos* | -aros* | -cloudabi* | -sortix* \
 	      | -nindy* | -vxsim* | -vxworks* | -ebmon* | -hms* | -mvs* \
 	      | -clix* | -riscos* | -uniplus* | -iris* | -rtu* | -xenix* \
 	      | -hiux* | -386bsd* | -knetbsd* | -mirbsd* | -netbsd* \
-	      | -bitrig* | -openbsd* | -solidbsd* \
+	      | -bitrig* | -openbsd* | -solidbsd* | -libertybsd* \
 	      | -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* | -cegcc* \
 	      | -cygwin* | -msys* | -pe* | -psos* | -moss* | -proelf* | -rtems* \
-	      | -mingw32* | -mingw64* | -linux-gnu* | -linux-android* \
+	      | -midipix* | -mingw32* | -mingw64* | -linux-gnu* | -linux-android* \
 	      | -linux-newlib* | -linux-musl* | -linux-uclibc* \
-	      | -uxpv* | -beos* | -mpeix* | -udk* \
+	      | -uxpv* | -beos* | -mpeix* | -udk* | -moxiebox* \
 	      | -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* \
-	      | -skyos* | -haiku* | -rdos* | -toppers* | -drops* | -es* | -tirtos*)
+	      | -skyos* | -haiku* | -rdos* | -toppers* | -drops* | -es* \
+	      | -onefs* | -tirtos* | -phoenix* | -fuchsia*)
 	# Remember, each alternative MUST END IN *, to match a version number.
 		;;
 	-qnx*)
@@ -1404,9 +1433,6 @@ case $os in
 	-mac*)
 		os=`echo $os | sed -e 's|mac|macos|'`
 		;;
-	# Apple iOS
-	-ios*)
-		;;
 	-linux-dietlibc)
 		os=-linux-dietlibc
 		;;
@@ -1515,6 +1541,8 @@ case $os in
 		;;
 	-nacl*)
 		;;
+	-ios)
+		;;
 	-none)
 		;;
 	*)
--- a/configure.ac
+++ b/configure.ac
@@ -6,29 +6,66 @@ AC_CONFIG_AUX_DIR([build-aux])
 dnl ============================================================================
 dnl Custom macro definitions.

-dnl JE_CFLAGS_APPEND(cflag)
-AC_DEFUN([JE_CFLAGS_APPEND],
+dnl JE_CONCAT_VVV(r, a, b)
+dnl 
+dnl Set $r to the concatenation of $a and $b, with a space separating them iff
+dnl both $a and $b are non-emty.
+AC_DEFUN([JE_CONCAT_VVV],
+if test "x[$]{$2}" = "x" -o "x[$]{$3}" = "x" ; then
+  $1="[$]{$2}[$]{$3}"
+else
+  $1="[$]{$2} [$]{$3}"
+fi
+)
+
+dnl JE_APPEND_VS(a, b)
+dnl 
+dnl Set $a to the concatenation of $a and b, with a space separating them iff
+dnl both $a and b are non-empty.
+AC_DEFUN([JE_APPEND_VS],
+  T_APPEND_V=$2
+  JE_CONCAT_VVV($1, $1, T_APPEND_V)
+)
+
+CONFIGURE_CFLAGS=
+SPECIFIED_CFLAGS="${CFLAGS}"
+dnl JE_CFLAGS_ADD(cflag)
+dnl 
+dnl CFLAGS is the concatenation of CONFIGURE_CFLAGS and SPECIFIED_CFLAGS
+dnl (ignoring EXTRA_CFLAGS, which does not impact configure tests.  This macro
+dnl appends to CONFIGURE_CFLAGS and regenerates CFLAGS.
+AC_DEFUN([JE_CFLAGS_ADD],
 [
 AC_MSG_CHECKING([whether compiler supports $1])
-TCFLAGS="${CFLAGS}"
-if test "x${CFLAGS}" = "x" ; then
-  CFLAGS="$1"
-else
-  CFLAGS="${CFLAGS} $1"
-fi
+T_CONFIGURE_CFLAGS="${CONFIGURE_CFLAGS}"
+JE_APPEND_VS(CONFIGURE_CFLAGS, $1)
+JE_CONCAT_VVV(CFLAGS, CONFIGURE_CFLAGS, SPECIFIED_CFLAGS)
 AC_COMPILE_IFELSE([AC_LANG_PROGRAM(
 [[
 ]], [[
    return 0;
 ]])],
-              [je_cv_cflags_appended=$1]
+              [je_cv_cflags_added=$1]
              AC_MSG_RESULT([yes]),
-              [je_cv_cflags_appended=]
+              [je_cv_cflags_added=]
              AC_MSG_RESULT([no])
-              [CFLAGS="${TCFLAGS}"]
+              [CONFIGURE_CFLAGS="${T_CONFIGURE_CFLAGS}"]
 )
+JE_CONCAT_VVV(CFLAGS, CONFIGURE_CFLAGS, SPECIFIED_CFLAGS)
 ])

+dnl JE_CFLAGS_SAVE()
+dnl JE_CFLAGS_RESTORE()
+dnl 
+dnl Save/restore CFLAGS.  Nesting is not supported.
+AC_DEFUN([JE_CFLAGS_SAVE],
+SAVED_CONFIGURE_CFLAGS="${CONFIGURE_CFLAGS}"
+)
+AC_DEFUN([JE_CFLAGS_RESTORE],
+CONFIGURE_CFLAGS="${SAVED_CONFIGURE_CFLAGS}"
+JE_CONCAT_VVV(CFLAGS, CONFIGURE_CFLAGS, SPECIFIED_CFLAGS)
+)
+
 dnl JE_COMPILABLE(label, hcode, mcode, rvar)
 dnl 
 dnl Use AC_LINK_IFELSE() rather than AC_COMPILE_IFELSE() so that linker errors
@@ -118,6 +155,7 @@ dnl If CFLAGS isn't defined, set CFLAGS to something reasonable.  Otherwise,
 dnl just prevent autoconf from molesting CFLAGS.
 CFLAGS=$CFLAGS
 AC_PROG_CC
+
 if test "x$GCC" != "xyes" ; then
  AC_CACHE_CHECK([whether compiler is MSVC],
                 [je_cv_msvc],
@@ -131,31 +169,82 @@ if test "x$GCC" != "xyes" ; then
                               [je_cv_msvc=no])])
 fi

-if test "x$CFLAGS" = "x" ; then
-  no_CFLAGS="yes"
-  if test "x$GCC" = "xyes" ; then
-    JE_CFLAGS_APPEND([-std=gnu99])
-    if test "x$je_cv_cflags_appended" = "x-std=gnu99" ; then
+dnl check if a cray prgenv wrapper compiler is being used
+je_cv_cray_prgenv_wrapper=""
+if test "x${PE_ENV}" != "x" ; then
+  case "${CC}" in
+    CC|cc)
+	je_cv_cray_prgenv_wrapper="yes"
+	;;
+    *)
+       ;;
+  esac
+fi
+
+AC_CACHE_CHECK([whether compiler is cray],
+              [je_cv_cray],
+              [AC_COMPILE_IFELSE([AC_LANG_PROGRAM([],
+                                                  [
+#ifndef _CRAYC
+  int fail[-1];
+#endif
+])],
+                            [je_cv_cray=yes],
+                            [je_cv_cray=no])])
+
+if test "x${je_cv_cray}" = "xyes" ; then
+  AC_CACHE_CHECK([whether cray compiler version is 8.4],
+                [je_cv_cray_84],
+                [AC_COMPILE_IFELSE([AC_LANG_PROGRAM([],
+                                                      [
+#if !(_RELEASE_MAJOR == 8 && _RELEASE_MINOR == 4)
+  int fail[-1];
+#endif
+])],
+                              [je_cv_cray_84=yes],
+                              [je_cv_cray_84=no])])
+fi
+
+if test "x$GCC" = "xyes" ; then
+  JE_CFLAGS_ADD([-std=gnu11])
+  if test "x$je_cv_cflags_added" = "x-std=gnu11" ; then
+    AC_DEFINE_UNQUOTED([JEMALLOC_HAS_RESTRICT])
+  else
+    JE_CFLAGS_ADD([-std=gnu99])
+    if test "x$je_cv_cflags_added" = "x-std=gnu99" ; then
      AC_DEFINE_UNQUOTED([JEMALLOC_HAS_RESTRICT])
    fi
-    JE_CFLAGS_APPEND([-Wall])
-    JE_CFLAGS_APPEND([-Werror=declaration-after-statement])
-    JE_CFLAGS_APPEND([-Wshorten-64-to-32])
-    JE_CFLAGS_APPEND([-pipe])
-    JE_CFLAGS_APPEND([-g3])
-  elif test "x$je_cv_msvc" = "xyes" ; then
-    CC="$CC -nologo"
-    JE_CFLAGS_APPEND([-Zi])
-    JE_CFLAGS_APPEND([-MT])
-    JE_CFLAGS_APPEND([-W3])
-    JE_CFLAGS_APPEND([-FS])
-    CPPFLAGS="$CPPFLAGS -I${srcdir}/include/msvc_compat"
+  fi
+  JE_CFLAGS_ADD([-Wall])
+  JE_CFLAGS_ADD([-Werror=declaration-after-statement])
+  JE_CFLAGS_ADD([-Wshorten-64-to-32])
+  JE_CFLAGS_ADD([-Wsign-compare])
+  JE_CFLAGS_ADD([-pipe])
+  JE_CFLAGS_ADD([-g3])
+elif test "x$je_cv_msvc" = "xyes" ; then
+  CC="$CC -nologo"
+  JE_CFLAGS_ADD([-Zi])
+  JE_CFLAGS_ADD([-MT])
+  JE_CFLAGS_ADD([-W3])
+  JE_CFLAGS_ADD([-FS])
+  JE_APPEND_VS(CPPFLAGS, -I${srcdir}/include/msvc_compat)
+fi
+if test "x$je_cv_cray" = "xyes" ; then
+  dnl cray compiler 8.4 has an inlining bug
+  if test "x$je_cv_cray_84" = "xyes" ; then
+    JE_CFLAGS_ADD([-hipa2])
+    JE_CFLAGS_ADD([-hnognu])
+  fi
+  if test "x$enable_cc_silence" != "xno" ; then
+    dnl ignore unreachable code warning
+    JE_CFLAGS_ADD([-hnomessage=128])
+    dnl ignore redefinition of "malloc", "free", etc warning
+    JE_CFLAGS_ADD([-hnomessage=1357])
  fi
 fi
-dnl Append EXTRA_CFLAGS to CFLAGS, if defined.
-if test "x$EXTRA_CFLAGS" != "x" ; then
-  JE_CFLAGS_APPEND([$EXTRA_CFLAGS])
-fi
+AC_SUBST([CONFIGURE_CFLAGS])
+AC_SUBST([SPECIFIED_CFLAGS])
+AC_SUBST([EXTRA_CFLAGS])
 AC_PROG_CPP

 AC_C_BIGENDIAN([ac_cv_big_endian=1], [ac_cv_big_endian=0])
@@ -164,7 +253,7 @@ if test "x${ac_cv_big_endian}" = "x1" ; then
 fi

 if test "x${je_cv_msvc}" = "xyes" -a "x${ac_cv_header_inttypes_h}" = "xno"; then
-  CPPFLAGS="$CPPFLAGS -I${srcdir}/include/msvc_compat/C99"
+  JE_APPEND_VS(CPPFLAGS, -I${srcdir}/include/msvc_compat/C99)
 fi

 if test "x${je_cv_msvc}" = "xyes" ; then
@@ -262,17 +351,27 @@ o="$ac_objext"
 a="a"
 exe="$ac_exeext"
 libprefix="lib"
+link_whole_archive="0"
 DSO_LDFLAGS='-shared -Wl,-soname,$(@F)'
 RPATH='-Wl,-rpath,$(1)'
 SOREV="${so}.${rev}"
 PIC_CFLAGS='-fPIC -DPIC'
 CTARGET='-o $@'
 LDTARGET='-o $@'
+TEST_LD_MODE=
 EXTRA_LDFLAGS=
 ARFLAGS='crus'
 AROUT=' $@'
 CC_MM=1

+if test "x$je_cv_cray_prgenv_wrapper" = "xyes" ; then
+  TEST_LD_MODE='-dynamic'
+fi
+
+if test "x${je_cv_cray}" = "xyes" ; then
+  CC_MM=
+fi
+
 AN_MAKEVAR([AR], [AC_PROG_AR])
 AN_PROGRAM([ar], [AC_PROG_AR])
 AC_DEFUN([AC_PROG_AR], [AC_CHECK_TOOL(AR, ar, :)])
@@ -289,9 +388,7 @@ default_munmap="1"
 maps_coalesce="1"
 case "${host}" in
  *-*-darwin* | *-*-ios*)
-	CFLAGS="$CFLAGS"
 	abi="macho"
-	AC_DEFINE([JEMALLOC_PURGE_MADVISE_FREE], [ ])
 	RPATH=""
 	LD_PRELOAD_VAR="DYLD_INSERT_LIBRARIES"
 	so="dylib"
@@ -302,33 +399,37 @@ case "${host}" in
 	sbrk_deprecated="1"
 	;;
  *-*-freebsd*)
-	CFLAGS="$CFLAGS"
 	abi="elf"
-	AC_DEFINE([JEMALLOC_PURGE_MADVISE_FREE], [ ])
+	AC_DEFINE([JEMALLOC_SYSCTL_VM_OVERCOMMIT], [ ])
 	force_lazy_lock="1"
 	;;
  *-*-dragonfly*)
-	CFLAGS="$CFLAGS"
 	abi="elf"
-	AC_DEFINE([JEMALLOC_PURGE_MADVISE_FREE], [ ])
 	;;
  *-*-openbsd*)
-	CFLAGS="$CFLAGS"
 	abi="elf"
-	AC_DEFINE([JEMALLOC_PURGE_MADVISE_FREE], [ ])
 	force_tls="0"
 	;;
  *-*-bitrig*)
-	CFLAGS="$CFLAGS"
 	abi="elf"
-	AC_DEFINE([JEMALLOC_PURGE_MADVISE_FREE], [ ])
 	;;
-  *-*-linux*)
-	CFLAGS="$CFLAGS"
-	CPPFLAGS="$CPPFLAGS -D_GNU_SOURCE"
+  *-*-linux-android)
+	dnl syscall(2) and secure_getenv(3) are exposed by _GNU_SOURCE.
+	JE_APPEND_VS(CPPFLAGS, -D_GNU_SOURCE)
 	abi="elf"
 	AC_DEFINE([JEMALLOC_HAS_ALLOCA_H])
-	AC_DEFINE([JEMALLOC_PURGE_MADVISE_DONTNEED], [ ])
+	AC_DEFINE([JEMALLOC_PROC_SYS_VM_OVERCOMMIT_MEMORY], [ ])
+	AC_DEFINE([JEMALLOC_THREADED_INIT], [ ])
+	AC_DEFINE([JEMALLOC_C11ATOMICS])
+	force_tls="0"
+	default_munmap="0"
+	;;
+  *-*-linux* | *-*-kfreebsd*)
+	dnl syscall(2) and secure_getenv(3) are exposed by _GNU_SOURCE.
+	JE_APPEND_VS(CPPFLAGS, -D_GNU_SOURCE)
+	abi="elf"
+	AC_DEFINE([JEMALLOC_HAS_ALLOCA_H])
+	AC_DEFINE([JEMALLOC_PROC_SYS_VM_OVERCOMMIT_MEMORY], [ ])
 	AC_DEFINE([JEMALLOC_THREADED_INIT], [ ])
 	AC_DEFINE([JEMALLOC_USE_CXX_THROW], [ ])
 	default_munmap="0"
@@ -342,19 +443,16 @@ case "${host}" in
 #error aout
 #endif
 ]])],
-                          [CFLAGS="$CFLAGS"; abi="elf"],
+                          [abi="elf"],
                          [abi="aout"])
 	AC_MSG_RESULT([$abi])
-	AC_DEFINE([JEMALLOC_PURGE_MADVISE_FREE], [ ])
 	;;
  *-*-solaris2*)
-	CFLAGS="$CFLAGS"
 	abi="elf"
-	AC_DEFINE([JEMALLOC_PURGE_MADVISE_FREE], [ ])
 	RPATH='-Wl,-R,$(1)'
 	dnl Solaris needs this for sigwait().
-	CPPFLAGS="$CPPFLAGS -D_POSIX_PTHREAD_SEMANTICS"
-	LIBS="$LIBS -lposix4 -lsocket -lnsl"
+	JE_APPEND_VS(CPPFLAGS, -D_POSIX_PTHREAD_SEMANTICS)
+	JE_APPEND_VS(LIBS, -lposix4 -lsocket -lnsl)
 	;;
  *-ibm-aix*)
 	if "$LG_SIZEOF_PTR" = "8"; then
@@ -369,7 +467,6 @@ case "${host}" in
  *-*-mingw* | *-*-cygwin*)
 	abi="pecoff"
 	force_tls="0"
-	force_lazy_lock="1"
 	maps_coalesce="0"
 	RPATH=""
 	so="dll"
@@ -386,6 +483,7 @@ case "${host}" in
        else
 	  importlib="${so}"
 	  DSO_LDFLAGS="-shared"
+	  link_whole_archive="1"
 	fi
 	a="lib"
 	libprefix=""
@@ -423,17 +521,28 @@ AC_SUBST([o])
 AC_SUBST([a])
 AC_SUBST([exe])
 AC_SUBST([libprefix])
+AC_SUBST([link_whole_archive])
 AC_SUBST([DSO_LDFLAGS])
 AC_SUBST([EXTRA_LDFLAGS])
 AC_SUBST([SOREV])
 AC_SUBST([PIC_CFLAGS])
 AC_SUBST([CTARGET])
 AC_SUBST([LDTARGET])
+AC_SUBST([TEST_LD_MODE])
 AC_SUBST([MKLIB])
 AC_SUBST([ARFLAGS])
 AC_SUBST([AROUT])
 AC_SUBST([CC_MM])

+dnl Determine whether libm must be linked to use e.g. log(3).
+AC_SEARCH_LIBS([log], [m], , [AC_MSG_ERROR([Missing math functions])])
+if test "x$ac_cv_search_log" != "xnone required" ; then
+  LM="$ac_cv_search_log"
+else
+  LM=
+fi
+AC_SUBST(LM)
+
 JE_COMPILABLE([__attribute__ syntax],
              [static __attribute__((unused)) void foo(void){}],
              [],
@@ -441,18 +550,19 @@ JE_COMPILABLE([__attribute__ syntax],
 if test "x${je_cv_attribute}" = "xyes" ; then
  AC_DEFINE([JEMALLOC_HAVE_ATTR], [ ])
  if test "x${GCC}" = "xyes" -a "x${abi}" = "xelf"; then
-    JE_CFLAGS_APPEND([-fvisibility=hidden])
+    JE_CFLAGS_ADD([-fvisibility=hidden])
  fi
 fi
 dnl Check for tls_model attribute support (clang 3.0 still lacks support).
-SAVED_CFLAGS="${CFLAGS}"
-JE_CFLAGS_APPEND([-Werror])
+JE_CFLAGS_SAVE()
+JE_CFLAGS_ADD([-Werror])
+JE_CFLAGS_ADD([-herror_on_warning])
 JE_COMPILABLE([tls_model attribute], [],
              [static __thread int
               __attribute__((tls_model("initial-exec"), unused)) foo;
               foo = 0;],
              [je_cv_tls_model])
-CFLAGS="${SAVED_CFLAGS}"
+JE_CFLAGS_RESTORE()
 if test "x${je_cv_tls_model}" = "xyes" ; then
  AC_DEFINE([JEMALLOC_TLS_MODEL],
            [__attribute__((tls_model("initial-exec")))])
@@ -460,32 +570,35 @@ else
  AC_DEFINE([JEMALLOC_TLS_MODEL], [ ])
 fi
 dnl Check for alloc_size attribute support.
-SAVED_CFLAGS="${CFLAGS}"
-JE_CFLAGS_APPEND([-Werror])
+JE_CFLAGS_SAVE()
+JE_CFLAGS_ADD([-Werror])
+JE_CFLAGS_ADD([-herror_on_warning])
 JE_COMPILABLE([alloc_size attribute], [#include <stdlib.h>],
              [void *foo(size_t size) __attribute__((alloc_size(1)));],
              [je_cv_alloc_size])
-CFLAGS="${SAVED_CFLAGS}"
+JE_CFLAGS_RESTORE()
 if test "x${je_cv_alloc_size}" = "xyes" ; then
  AC_DEFINE([JEMALLOC_HAVE_ATTR_ALLOC_SIZE], [ ])
 fi
 dnl Check for format(gnu_printf, ...) attribute support.
-SAVED_CFLAGS="${CFLAGS}"
-JE_CFLAGS_APPEND([-Werror])
+JE_CFLAGS_SAVE()
+JE_CFLAGS_ADD([-Werror])
+JE_CFLAGS_ADD([-herror_on_warning])
 JE_COMPILABLE([format(gnu_printf, ...) attribute], [#include <stdlib.h>],
              [void *foo(const char *format, ...) __attribute__((format(gnu_printf, 1, 2)));],
              [je_cv_format_gnu_printf])
-CFLAGS="${SAVED_CFLAGS}"
+JE_CFLAGS_RESTORE()
 if test "x${je_cv_format_gnu_printf}" = "xyes" ; then
  AC_DEFINE([JEMALLOC_HAVE_ATTR_FORMAT_GNU_PRINTF], [ ])
 fi
 dnl Check for format(printf, ...) attribute support.
-SAVED_CFLAGS="${CFLAGS}"
-JE_CFLAGS_APPEND([-Werror])
+JE_CFLAGS_SAVE()
+JE_CFLAGS_ADD([-Werror])
+JE_CFLAGS_ADD([-herror_on_warning])
 JE_COMPILABLE([format(printf, ...) attribute], [#include <stdlib.h>],
              [void *foo(const char *format, ...) __attribute__((format(printf, 1, 2)));],
              [je_cv_format_printf])
-CFLAGS="${SAVED_CFLAGS}"
+JE_CFLAGS_RESTORE()
 if test "x${je_cv_format_printf}" = "xyes" ; then
  AC_DEFINE([JEMALLOC_HAVE_ATTR_FORMAT_PRINTF], [ ])
 fi
@@ -547,9 +660,9 @@ if test "x$enable_code_coverage" = "x1" ; then
  deoptimize="no"
  echo "$CFLAGS $EXTRA_CFLAGS" | grep '\-O' >/dev/null || deoptimize="yes"
  if test "x${deoptimize}" = "xyes" ; then
-    JE_CFLAGS_APPEND([-O0])
+    JE_CFLAGS_ADD([-O0])
  fi
-  JE_CFLAGS_APPEND([-fprofile-arcs -ftest-coverage])
+  JE_CFLAGS_ADD([-fprofile-arcs -ftest-coverage])
  EXTRA_LDFLAGS="$EXTRA_LDFLAGS -fprofile-arcs -ftest-coverage"
  AC_DEFINE([JEMALLOC_CODE_COVERAGE], [ ])
 fi
@@ -739,19 +852,14 @@ if test "x$enable_ivsalloc" = "x1" ; then
 fi

 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 "$CFLAGS $EXTRA_CFLAGS" | grep '\-O' >/dev/null || optimize="yes"
-  if test "x${optimize}" = "xyes" ; then
-    if test "x$GCC" = "xyes" ; then
-      JE_CFLAGS_APPEND([-O3])
-      JE_CFLAGS_APPEND([-funroll-loops])
-    elif test "x$je_cv_msvc" = "xyes" ; then
-      JE_CFLAGS_APPEND([-O2])
-    else
-      JE_CFLAGS_APPEND([-O])
-    fi
+if test "x$enable_debug" = "x0" ; then
+  if test "x$GCC" = "xyes" ; then
+    JE_CFLAGS_ADD([-O3])
+    JE_CFLAGS_ADD([-funroll-loops])
+  elif test "x$je_cv_msvc" = "xyes" ; then
+    JE_CFLAGS_ADD([-O2])
+  else
+    JE_CFLAGS_ADD([-O])
  fi
 fi

@@ -815,10 +923,10 @@ fi,
 if test "x$backtrace_method" = "x" -a "x$enable_prof_libunwind" = "x1" ; then
  AC_CHECK_HEADERS([libunwind.h], , [enable_prof_libunwind="0"])
  if test "x$LUNWIND" = "x-lunwind" ; then
-    AC_CHECK_LIB([unwind], [unw_backtrace], [LIBS="$LIBS $LUNWIND"],
+    AC_CHECK_LIB([unwind], [unw_backtrace], [JE_APPEND_VS(LIBS, $LUNWIND)],
                 [enable_prof_libunwind="0"])
  else
-    LIBS="$LIBS $LUNWIND"
+    JE_APPEND_VS(LIBS, $LUNWIND)
  fi
  if test "x${enable_prof_libunwind}" = "x1" ; then
    backtrace_method="libunwind"
@@ -840,7 +948,7 @@ fi
 if test "x$backtrace_method" = "x" -a "x$enable_prof_libgcc" = "x1" \
     -a "x$GCC" = "xyes" ; then
  AC_CHECK_HEADERS([unwind.h], , [enable_prof_libgcc="0"])
-  AC_CHECK_LIB([gcc], [_Unwind_Backtrace], [LIBS="$LIBS -lgcc"], [enable_prof_libgcc="0"])
+  AC_CHECK_LIB([gcc], [_Unwind_Backtrace], [JE_APPEND_VS(LIBS, -lgcc)], [enable_prof_libgcc="0"])
  if test "x${enable_prof_libgcc}" = "x1" ; then
    backtrace_method="libgcc"
    AC_DEFINE([JEMALLOC_PROF_LIBGCC], [ ])
@@ -862,7 +970,7 @@ fi
 )
 if test "x$backtrace_method" = "x" -a "x$enable_prof_gcc" = "x1" \
     -a "x$GCC" = "xyes" ; then
-  JE_CFLAGS_APPEND([-fno-omit-frame-pointer])
+  JE_CFLAGS_ADD([-fno-omit-frame-pointer])
  backtrace_method="gcc intrinsics"
  AC_DEFINE([JEMALLOC_PROF_GCC], [ ])
 else
@@ -876,10 +984,8 @@ fi
 AC_MSG_CHECKING([configured backtracing method])
 AC_MSG_RESULT([$backtrace_method])
 if test "x$enable_prof" = "x1" ; then
-  if test "x$abi" != "xpecoff"; then
-    dnl Heap profiling uses the log(3) function.
-    LIBS="$LIBS -lm"
-  fi
+  dnl Heap profiling uses the log(3) function.
+  JE_APPEND_VS(LIBS, $LM)

  AC_DEFINE([JEMALLOC_PROF], [ ])
 fi
@@ -1047,6 +1153,23 @@ if test "x$enable_cache_oblivious" = "x1" ; then
 fi
 AC_SUBST([enable_cache_oblivious])

+
+
+JE_COMPILABLE([a program using __builtin_unreachable], [
+void foo (void) {
+  __builtin_unreachable();
+}
+], [
+	{
+		foo();
+	}
+], [je_cv_gcc_builtin_unreachable])
+if test "x${je_cv_gcc_builtin_unreachable}" = "xyes" ; then
+  AC_DEFINE([JEMALLOC_INTERNAL_UNREACHABLE], [__builtin_unreachable])
+else
+  AC_DEFINE([JEMALLOC_INTERNAL_UNREACHABLE], [abort])
+fi
+
 dnl ============================================================================
 dnl Check for  __builtin_ffsl(), then ffsl(3), and fail if neither are found.
 dnl One of those two functions should (theoretically) exist on all platforms
@@ -1172,27 +1295,36 @@ dnl ============================================================================
 dnl jemalloc configuration.
 dnl 

-dnl Set VERSION if source directory is inside a git repository.
-if test "x`test ! \"${srcroot}\" && cd \"${srcroot}\"; git rev-parse --is-inside-work-tree 2>/dev/null`" = "xtrue" ; then
-  dnl Pattern globs aren't powerful enough to match both single- and
-  dnl double-digit version numbers, so iterate over patterns to support up to
-  dnl version 99.99.99 without any accidental matches.
-  rm -f "${objroot}VERSION"
-  for pattern in ['[0-9].[0-9].[0-9]' '[0-9].[0-9].[0-9][0-9]' \
-                 '[0-9].[0-9][0-9].[0-9]' '[0-9].[0-9][0-9].[0-9][0-9]' \
-                 '[0-9][0-9].[0-9].[0-9]' '[0-9][0-9].[0-9].[0-9][0-9]' \
-                 '[0-9][0-9].[0-9][0-9].[0-9]' \
-                 '[0-9][0-9].[0-9][0-9].[0-9][0-9]']; do
-    if test ! -e "${objroot}VERSION" ; then
-      (test ! "${srcroot}" && cd "${srcroot}"; git describe --long --abbrev=40 --match="${pattern}") > "${objroot}VERSION.tmp" 2>/dev/null
-      if test $? -eq 0 ; then
-        mv "${objroot}VERSION.tmp" "${objroot}VERSION"
-        break
-      fi
+AC_ARG_WITH([version],
+  [AS_HELP_STRING([--with-version=<major>.<minor>.<bugfix>-<nrev>-g<gid>],
+   [Version string])],
+  [
+    echo "${with_version}" | grep ['^[0-9]\+\.[0-9]\+\.[0-9]\+-[0-9]\+-g[0-9a-f]\+$'] 2>&1 1>/dev/null
+    if test $? -ne 0 ; then
+      AC_MSG_ERROR([${with_version} does not match <major>.<minor>.<bugfix>-<nrev>-g<gid>])
    fi
-  done
-fi
-rm -f "${objroot}VERSION.tmp"
+    echo "$with_version" > "${objroot}VERSION"
+  ], [
+    dnl Set VERSION if source directory is inside a git repository.
+    if test "x`test ! \"${srcroot}\" && cd \"${srcroot}\"; git rev-parse --is-inside-work-tree 2>/dev/null`" = "xtrue" ; then
+      dnl Pattern globs aren't powerful enough to match both single- and
+      dnl double-digit version numbers, so iterate over patterns to support up
+      dnl to version 99.99.99 without any accidental matches.
+      for pattern in ['[0-9].[0-9].[0-9]' '[0-9].[0-9].[0-9][0-9]' \
+                     '[0-9].[0-9][0-9].[0-9]' '[0-9].[0-9][0-9].[0-9][0-9]' \
+                     '[0-9][0-9].[0-9].[0-9]' '[0-9][0-9].[0-9].[0-9][0-9]' \
+                     '[0-9][0-9].[0-9][0-9].[0-9]' \
+                     '[0-9][0-9].[0-9][0-9].[0-9][0-9]']; do
+        (test ! "${srcroot}" && cd "${srcroot}"; git describe --long --abbrev=40 --match="${pattern}") > "${objroot}VERSION.tmp" 2>/dev/null
+        if test $? -eq 0 ; then
+          mv "${objroot}VERSION.tmp" "${objroot}VERSION"
+          break
+        fi
+      done
+    fi
+    rm -f "${objroot}VERSION.tmp"
+  ])
+
 if test ! -e "${objroot}VERSION" ; then
  if test ! -e "${srcroot}VERSION" ; then
    AC_MSG_RESULT(
@@ -1222,20 +1354,104 @@ if test "x$abi" != "xpecoff" ; then
  AC_CHECK_HEADERS([pthread.h], , [AC_MSG_ERROR([pthread.h is missing])])
  dnl Some systems may embed pthreads functionality in libc; check for libpthread
  dnl first, but try libc too before failing.
-  AC_CHECK_LIB([pthread], [pthread_create], [LIBS="$LIBS -lpthread"],
+  AC_CHECK_LIB([pthread], [pthread_create], [JE_APPEND_VS(LIBS, -lpthread)],
               [AC_SEARCH_LIBS([pthread_create], , ,
                               AC_MSG_ERROR([libpthread is missing]))])
+  JE_COMPILABLE([pthread_atfork(3)], [
+#include <pthread.h>
+], [
+  pthread_atfork((void *)0, (void *)0, (void *)0);
+], [je_cv_pthread_atfork])
+  if test "x${je_cv_pthread_atfork}" = "xyes" ; then
+    AC_DEFINE([JEMALLOC_HAVE_PTHREAD_ATFORK], [ ])
+  fi
 fi

-CPPFLAGS="$CPPFLAGS -D_REENTRANT"
+JE_APPEND_VS(CPPFLAGS, -D_REENTRANT)

-dnl Check whether clock_gettime(2) is in libc or librt.  This function is only
-dnl used in test code, so save the result to TESTLIBS to avoid poluting LIBS.
-SAVED_LIBS="${LIBS}"
-LIBS=
-AC_SEARCH_LIBS([clock_gettime], [rt], [TESTLIBS="${LIBS}"])
-AC_SUBST([TESTLIBS])
-LIBS="${SAVED_LIBS}"
+dnl Check whether clock_gettime(2) is in libc or librt.
+AC_SEARCH_LIBS([clock_gettime], [rt])
+
+dnl Cray wrapper compiler often adds `-lrt` when using `-static`. Check with
+dnl `-dynamic` as well in case a user tries to dynamically link in jemalloc
+if test "x$je_cv_cray_prgenv_wrapper" = "xyes" ; then
+  if test "$ac_cv_search_clock_gettime" != "-lrt"; then
+    JE_CFLAGS_SAVE()
+
+    unset ac_cv_search_clock_gettime
+    JE_CFLAGS_ADD([-dynamic])
+    AC_SEARCH_LIBS([clock_gettime], [rt])
+
+    JE_CFLAGS_RESTORE()
+  fi
+fi
+
+dnl check for CLOCK_MONOTONIC_COARSE (Linux-specific).
+JE_COMPILABLE([clock_gettime(CLOCK_MONOTONIC_COARSE, ...)], [
+#include <time.h>
+], [
+	struct timespec ts;
+
+	clock_gettime(CLOCK_MONOTONIC_COARSE, &ts);
+], [je_cv_clock_monotonic_coarse])
+if test "x${je_cv_clock_monotonic_coarse}" = "xyes" ; then
+  AC_DEFINE([JEMALLOC_HAVE_CLOCK_MONOTONIC_COARSE])
+fi
+
+dnl check for CLOCK_MONOTONIC.
+JE_COMPILABLE([clock_gettime(CLOCK_MONOTONIC, ...)], [
+#include <unistd.h>
+#include <time.h>
+], [
+	struct timespec ts;
+
+	clock_gettime(CLOCK_MONOTONIC, &ts);
+#if !defined(_POSIX_MONOTONIC_CLOCK) || _POSIX_MONOTONIC_CLOCK < 0
+#  error _POSIX_MONOTONIC_CLOCK missing/invalid
+#endif
+], [je_cv_clock_monotonic])
+if test "x${je_cv_clock_monotonic}" = "xyes" ; then
+  AC_DEFINE([JEMALLOC_HAVE_CLOCK_MONOTONIC])
+fi
+
+dnl Check for mach_absolute_time().
+JE_COMPILABLE([mach_absolute_time()], [
+#include <mach/mach_time.h>
+], [
+	mach_absolute_time();
+], [je_cv_mach_absolute_time])
+if test "x${je_cv_mach_absolute_time}" = "xyes" ; then
+  AC_DEFINE([JEMALLOC_HAVE_MACH_ABSOLUTE_TIME])
+fi
+
+dnl Use syscall(2) (if available) by default.
+AC_ARG_ENABLE([syscall],
+  [AS_HELP_STRING([--disable-syscall], [Disable use of syscall(2)])],
+[if test "x$enable_syscall" = "xno" ; then
+  enable_syscall="0"
+else
+  enable_syscall="1"
+fi
+],
+[enable_syscall="1"]
+)
+if test "x$enable_syscall" = "x1" ; then
+  dnl Check if syscall(2) is usable.  Treat warnings as errors, so that e.g. OS
+  dnl X 10.12's deprecation warning prevents use.
+  JE_CFLAGS_SAVE()
+  JE_CFLAGS_ADD([-Werror])
+  JE_COMPILABLE([syscall(2)], [
+#include <sys/syscall.h>
+#include <unistd.h>
+], [
+	syscall(SYS_write, 2, "hello", 5);
+],
+                [je_cv_syscall])
+  JE_CFLAGS_RESTORE()
+  if test "x$je_cv_syscall" = "xyes" ; then
+    AC_DEFINE([JEMALLOC_USE_SYSCALL], [ ])
+  fi
+fi

 dnl Check if the GNU-specific secure_getenv function exists.
 AC_CHECK_FUNC([secure_getenv],
@@ -1291,21 +1507,27 @@ fi
 ],
 [enable_lazy_lock=""]
 )
-if test "x$enable_lazy_lock" = "x" -a "x${force_lazy_lock}" = "x1" ; then
-  AC_MSG_RESULT([Forcing lazy-lock to avoid allocator/threading bootstrap issues])
-  enable_lazy_lock="1"
+if test "x${enable_lazy_lock}" = "x" ; then
+  if test "x${force_lazy_lock}" = "x1" ; then
+    AC_MSG_RESULT([Forcing lazy-lock to avoid allocator/threading bootstrap issues])
+    enable_lazy_lock="1"
+  else
+    enable_lazy_lock="0"
+  fi
+fi
+if test "x${enable_lazy_lock}" = "x1" -a "x${abi}" = "xpecoff" ; then
+  AC_MSG_RESULT([Forcing no lazy-lock because thread creation monitoring is unimplemented])
+  enable_lazy_lock="0"
 fi
 if test "x$enable_lazy_lock" = "x1" ; then
  if test "x$abi" != "xpecoff" ; then
    AC_CHECK_HEADERS([dlfcn.h], , [AC_MSG_ERROR([dlfcn.h is missing])])
    AC_CHECK_FUNC([dlsym], [],
-      [AC_CHECK_LIB([dl], [dlsym], [LIBS="$LIBS -ldl"],
+      [AC_CHECK_LIB([dl], [dlsym], [JE_APPEND_VS(LIBS, -ldl)],
                    [AC_MSG_ERROR([libdl is missing])])
      ])
  fi
  AC_DEFINE([JEMALLOC_LAZY_LOCK], [ ])
-else
-  enable_lazy_lock="0"
 fi
 AC_SUBST([enable_lazy_lock])

@@ -1428,14 +1650,64 @@ dnl Check for madvise(2).
 JE_COMPILABLE([madvise(2)], [
 #include <sys/mman.h>
 ], [
-	{
-		madvise((void *)0, 0, 0);
-	}
+	madvise((void *)0, 0, 0);
 ], [je_cv_madvise])
 if test "x${je_cv_madvise}" = "xyes" ; then
  AC_DEFINE([JEMALLOC_HAVE_MADVISE], [ ])
+
+  dnl Check for madvise(..., MADV_FREE).
+  JE_COMPILABLE([madvise(..., MADV_FREE)], [
+#include <sys/mman.h>
+], [
+	madvise((void *)0, 0, MADV_FREE);
+], [je_cv_madv_free])
+  if test "x${je_cv_madv_free}" = "xyes" ; then
+    AC_DEFINE([JEMALLOC_PURGE_MADVISE_FREE], [ ])
+  fi
+
+  dnl Check for madvise(..., MADV_DONTNEED).
+  JE_COMPILABLE([madvise(..., MADV_DONTNEED)], [
+#include <sys/mman.h>
+], [
+	madvise((void *)0, 0, MADV_DONTNEED);
+], [je_cv_madv_dontneed])
+  if test "x${je_cv_madv_dontneed}" = "xyes" ; then
+    AC_DEFINE([JEMALLOC_PURGE_MADVISE_DONTNEED], [ ])
+  fi
+
+  dnl Check for madvise(..., MADV_[NO]HUGEPAGE).
+  JE_COMPILABLE([madvise(..., MADV_[[NO]]HUGEPAGE)], [
+#include <sys/mman.h>
+], [
+	madvise((void *)0, 0, MADV_HUGEPAGE);
+	madvise((void *)0, 0, MADV_NOHUGEPAGE);
+], [je_cv_thp])
+  if test "x${je_cv_thp}" = "xyes" ; then
+    AC_DEFINE([JEMALLOC_HAVE_MADVISE_HUGE], [ ])
+  fi
 fi

+dnl Enable transparent huge page support by default.
+AC_ARG_ENABLE([thp],
+  [AS_HELP_STRING([--disable-thp],
+                  [Disable transparent huge page supprot])],
+[if test "x$enable_thp" = "xno" -o "x${je_cv_thp}" != "xyes" ; then
+  enable_thp="0"
+else
+  enable_thp="1"
+fi
+],
+[if test "x${je_cv_thp}" = "xyes" ; then
+  enable_thp="1"
+else
+  enable_thp="0"
+fi
+])
+if test "x$enable_thp" = "x1" ; then
+  AC_DEFINE([JEMALLOC_THP], [ ])
+fi
+AC_SUBST([enable_thp])
+
 dnl ============================================================================
 dnl Check whether __sync_{add,sub}_and_fetch() are available despite
 dnl __GCC_HAVE_SYNC_COMPARE_AND_SWAP_n macros being undefined.
@@ -1493,6 +1765,25 @@ if test "x${je_cv_builtin_clz}" = "xyes" ; then
  AC_DEFINE([JEMALLOC_HAVE_BUILTIN_CLZ], [ ])
 fi

+dnl ============================================================================
+dnl Check for os_unfair_lock operations as provided on Darwin.
+
+JE_COMPILABLE([Darwin os_unfair_lock_*()], [
+#include <os/lock.h>
+#include <AvailabilityMacros.h>
+], [
+	#if MAC_OS_X_VERSION_MIN_REQUIRED < 101200
+	#error "os_unfair_lock is not supported"
+	#else
+	os_unfair_lock lock = OS_UNFAIR_LOCK_INIT;
+	os_unfair_lock_lock(&lock);
+	os_unfair_lock_unlock(&lock);
+	#endif
+], [je_cv_os_unfair_lock])
+if test "x${je_cv_os_unfair_lock}" = "xyes" ; then
+  AC_DEFINE([JEMALLOC_OS_UNFAIR_LOCK], [ ])
+fi
+
 dnl ============================================================================
 dnl Check for spinlock(3) operations as provided on Darwin.

@@ -1532,37 +1823,6 @@ if test "x${enable_zone_allocator}" = "x1" ; then
    AC_MSG_ERROR([--enable-zone-allocator is only supported on Darwin])
  fi
  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_DEFUN([JE_ZONE_PROGRAM],
-    [AC_LANG_PROGRAM(
-      [#include <malloc/malloc.h>],
-      [static int foo[[sizeof($1) $2 sizeof(void *) * $3 ? 1 : -1]]]
-    )])
-
-  AC_COMPILE_IFELSE([JE_ZONE_PROGRAM(malloc_zone_t,==,14)],[JEMALLOC_ZONE_VERSION=3],[
-  AC_COMPILE_IFELSE([JE_ZONE_PROGRAM(malloc_zone_t,==,15)],[JEMALLOC_ZONE_VERSION=5],[
-  AC_COMPILE_IFELSE([JE_ZONE_PROGRAM(malloc_zone_t,==,16)],[
-    AC_COMPILE_IFELSE([JE_ZONE_PROGRAM(malloc_introspection_t,==,9)],[JEMALLOC_ZONE_VERSION=6],[
-    AC_COMPILE_IFELSE([JE_ZONE_PROGRAM(malloc_introspection_t,==,13)],[JEMALLOC_ZONE_VERSION=7],[JEMALLOC_ZONE_VERSION=]
-  )])],[
-  AC_COMPILE_IFELSE([JE_ZONE_PROGRAM(malloc_zone_t,==,17)],[JEMALLOC_ZONE_VERSION=8],[
-  AC_COMPILE_IFELSE([JE_ZONE_PROGRAM(malloc_zone_t,>,17)],[JEMALLOC_ZONE_VERSION=9],[JEMALLOC_ZONE_VERSION=]
-  )])])])])
-  if test "x${JEMALLOC_ZONE_VERSION}" = "x"; then
-    AC_MSG_RESULT([unsupported])
-    AC_MSG_ERROR([Unsupported malloc zone version])
-  fi
-  if test "${JEMALLOC_ZONE_VERSION}" = 9; then
-    JEMALLOC_ZONE_VERSION=8
-    AC_MSG_RESULT([> 8])
-  else
-    AC_MSG_RESULT([$JEMALLOC_ZONE_VERSION])
-  fi
-  AC_DEFINE_UNQUOTED(JEMALLOC_ZONE_VERSION, [$JEMALLOC_ZONE_VERSION])
 fi

 dnl ============================================================================
@@ -1736,12 +1996,13 @@ AC_MSG_RESULT([library revision   : ${rev}])
 AC_MSG_RESULT([])
 AC_MSG_RESULT([CONFIG             : ${CONFIG}])
 AC_MSG_RESULT([CC                 : ${CC}])
-AC_MSG_RESULT([CFLAGS             : ${CFLAGS}])
+AC_MSG_RESULT([CONFIGURE_CFLAGS   : ${CONFIGURE_CFLAGS}])
+AC_MSG_RESULT([SPECIFIED_CFLAGS   : ${SPECIFIED_CFLAGS}])
+AC_MSG_RESULT([EXTRA_CFLAGS       : ${EXTRA_CFLAGS}])
 AC_MSG_RESULT([CPPFLAGS           : ${CPPFLAGS}])
 AC_MSG_RESULT([LDFLAGS            : ${LDFLAGS}])
 AC_MSG_RESULT([EXTRA_LDFLAGS      : ${EXTRA_LDFLAGS}])
 AC_MSG_RESULT([LIBS               : ${LIBS}])
-AC_MSG_RESULT([TESTLIBS           : ${TESTLIBS}])
 AC_MSG_RESULT([RPATH_EXTRA        : ${RPATH_EXTRA}])
 AC_MSG_RESULT([])
 AC_MSG_RESULT([XSLTPROC           : ${XSLTPROC}])
@@ -1774,6 +2035,7 @@ AC_MSG_RESULT([prof-libunwind     : ${enable_prof_libunwind}])
 AC_MSG_RESULT([prof-libgcc        : ${enable_prof_libgcc}])
 AC_MSG_RESULT([prof-gcc           : ${enable_prof_gcc}])
 AC_MSG_RESULT([tcache             : ${enable_tcache}])
+AC_MSG_RESULT([thp                : ${enable_thp}])
 AC_MSG_RESULT([fill               : ${enable_fill}])
 AC_MSG_RESULT([utrace             : ${enable_utrace}])
 AC_MSG_RESULT([valgrind           : ${enable_valgrind}])
--- a/doc/html.xsl.in
+++ b/doc/html.xsl.in
@@ -1,4 +1,5 @@
 <xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">
  <xsl:import href="@XSLROOT@/html/docbook.xsl"/>
  <xsl:import href="@abs_srcroot@doc/stylesheet.xsl"/>
+  <xsl:output method="xml" encoding="utf-8"/>
 </xsl:stylesheet>
--- a/doc/jemalloc.xml.in
+++ b/doc/jemalloc.xml.in
@@ -52,7 +52,7 @@
    <title>LIBRARY</title>
    <para>This manual describes jemalloc @jemalloc_version@.  More information
    can be found at the <ulink
-    url="http://www.canonware.com/jemalloc/">jemalloc website</ulink>.</para>
+    url="http://jemalloc.net/">jemalloc website</ulink>.</para>
  </refsect1>
  <refsynopsisdiv>
    <title>SYNOPSIS</title>
@@ -180,20 +180,20 @@
    <refsect2>
      <title>Standard API</title>

-      <para>The <function>malloc<parameter/></function> function allocates
+      <para>The <function>malloc()</function> function allocates
      <parameter>size</parameter> bytes of uninitialized memory.  The allocated
      space is suitably aligned (after possible pointer coercion) for storage
      of any type of object.</para>

-      <para>The <function>calloc<parameter/></function> function allocates
+      <para>The <function>calloc()</function> function allocates
      space for <parameter>number</parameter> objects, each
      <parameter>size</parameter> bytes in length.  The result is identical to
-      calling <function>malloc<parameter/></function> with an argument of
+      calling <function>malloc()</function> with an argument of
      <parameter>number</parameter> * <parameter>size</parameter>, with the
      exception that the allocated memory is explicitly initialized to zero
      bytes.</para>

-      <para>The <function>posix_memalign<parameter/></function> function
+      <para>The <function>posix_memalign()</function> function
      allocates <parameter>size</parameter> bytes of memory such that the
      allocation's base address is a multiple of
      <parameter>alignment</parameter>, and returns the allocation in the value
@@ -201,7 +201,7 @@
      <parameter>alignment</parameter> must be a power of 2 at least as large as
      <code language="C">sizeof(<type>void *</type>)</code>.</para>

-      <para>The <function>aligned_alloc<parameter/></function> function
+      <para>The <function>aligned_alloc()</function> function
      allocates <parameter>size</parameter> bytes of memory such that the
      allocation's base address is a multiple of
      <parameter>alignment</parameter>.  The requested
@@ -209,7 +209,7 @@
      undefined if <parameter>size</parameter> is not an integral multiple of
      <parameter>alignment</parameter>.</para>

-      <para>The <function>realloc<parameter/></function> function changes the
+      <para>The <function>realloc()</function> function changes the
      size of the previously allocated memory referenced by
      <parameter>ptr</parameter> to <parameter>size</parameter> bytes.  The
      contents of the memory are unchanged up to the lesser of the new and old
@@ -217,26 +217,26 @@
      portion of the memory are undefined.  Upon success, the memory referenced
      by <parameter>ptr</parameter> is freed and a pointer to the newly
      allocated memory is returned.  Note that
-      <function>realloc<parameter/></function> may move the memory allocation,
+      <function>realloc()</function> may move the memory allocation,
      resulting in a different return value than <parameter>ptr</parameter>.
      If <parameter>ptr</parameter> is <constant>NULL</constant>, the
-      <function>realloc<parameter/></function> function behaves identically to
-      <function>malloc<parameter/></function> for the specified size.</para>
+      <function>realloc()</function> function behaves identically to
+      <function>malloc()</function> for the specified size.</para>

-      <para>The <function>free<parameter/></function> function causes the
+      <para>The <function>free()</function> function causes the
      allocated memory referenced by <parameter>ptr</parameter> to be made
      available for future allocations.  If <parameter>ptr</parameter> is
      <constant>NULL</constant>, no action occurs.</para>
    </refsect2>
    <refsect2>
      <title>Non-standard API</title>
-      <para>The <function>mallocx<parameter/></function>,
-      <function>rallocx<parameter/></function>,
-      <function>xallocx<parameter/></function>,
-      <function>sallocx<parameter/></function>,
-      <function>dallocx<parameter/></function>,
-      <function>sdallocx<parameter/></function>, and
-      <function>nallocx<parameter/></function> functions all have a
+      <para>The <function>mallocx()</function>,
+      <function>rallocx()</function>,
+      <function>xallocx()</function>,
+      <function>sallocx()</function>,
+      <function>dallocx()</function>,
+      <function>sdallocx()</function>, and
+      <function>nallocx()</function> functions all have a
      <parameter>flags</parameter> argument that can be used to specify
      options.  The functions only check the options that are contextually
      relevant.  Use bitwise or (<code language="C">|</code>) operations to
@@ -307,19 +307,19 @@
        </variablelist>
      </para>

-      <para>The <function>mallocx<parameter/></function> function allocates at
+      <para>The <function>mallocx()</function> function allocates at
      least <parameter>size</parameter> bytes of memory, and returns a pointer
      to the base address of the allocation.  Behavior is undefined if
      <parameter>size</parameter> is <constant>0</constant>.</para>

-      <para>The <function>rallocx<parameter/></function> function resizes the
+      <para>The <function>rallocx()</function> function resizes the
      allocation at <parameter>ptr</parameter> to be at least
      <parameter>size</parameter> bytes, and returns a pointer to the base
      address of the resulting allocation, which may or may not have moved from
      its original location.  Behavior is undefined if
      <parameter>size</parameter> is <constant>0</constant>.</para>

-      <para>The <function>xallocx<parameter/></function> function resizes the
+      <para>The <function>xallocx()</function> function resizes the
      allocation at <parameter>ptr</parameter> in place to be at least
      <parameter>size</parameter> bytes, and returns the real size of the
      allocation.  If <parameter>extra</parameter> is non-zero, an attempt is
@@ -332,32 +332,32 @@
      language="C">(<parameter>size</parameter> + <parameter>extra</parameter>
      &gt; <constant>SIZE_T_MAX</constant>)</code>.</para>

-      <para>The <function>sallocx<parameter/></function> function returns the
+      <para>The <function>sallocx()</function> function returns the
      real size of the allocation at <parameter>ptr</parameter>.</para>

-      <para>The <function>dallocx<parameter/></function> function causes the
+      <para>The <function>dallocx()</function> function causes the
      memory referenced by <parameter>ptr</parameter> to be made available for
      future allocations.</para>

-      <para>The <function>sdallocx<parameter/></function> function is an
-      extension of <function>dallocx<parameter/></function> with a
+      <para>The <function>sdallocx()</function> function is an
+      extension of <function>dallocx()</function> with a
      <parameter>size</parameter> parameter to allow the caller to pass in the
      allocation size as an optimization.  The minimum valid input size is the
      original requested size of the allocation, and the maximum valid input
      size is the corresponding value returned by
-      <function>nallocx<parameter/></function> or
-      <function>sallocx<parameter/></function>.</para>
+      <function>nallocx()</function> or
+      <function>sallocx()</function>.</para>

-      <para>The <function>nallocx<parameter/></function> function allocates no
+      <para>The <function>nallocx()</function> function allocates no
      memory, but it performs the same size computation as the
-      <function>mallocx<parameter/></function> function, and returns the real
+      <function>mallocx()</function> function, and returns the real
      size of the allocation that would result from the equivalent
-      <function>mallocx<parameter/></function> function call, or
+      <function>mallocx()</function> function call, or
      <constant>0</constant> if the inputs exceed the maximum supported size
      class and/or alignment.  Behavior is undefined if
      <parameter>size</parameter> is <constant>0</constant>.</para>

-      <para>The <function>mallctl<parameter/></function> function provides a
+      <para>The <function>mallctl()</function> function provides a
      general interface for introspecting the memory allocator, as well as
      setting modifiable parameters and triggering actions.  The
      period-separated <parameter>name</parameter> argument specifies a
@@ -372,12 +372,12 @@
      <parameter>newlen</parameter>; otherwise pass <constant>NULL</constant>
      and <constant>0</constant>.</para>

-      <para>The <function>mallctlnametomib<parameter/></function> function
+      <para>The <function>mallctlnametomib()</function> function
      provides a way to avoid repeated name lookups for applications that
      repeatedly query the same portion of the namespace, by translating a name
-      to a &ldquo;Management Information Base&rdquo; (MIB) that can be passed
-      repeatedly to <function>mallctlbymib<parameter/></function>.  Upon
-      successful return from <function>mallctlnametomib<parameter/></function>,
+      to a <quote>Management Information Base</quote> (MIB) that can be passed
+      repeatedly to <function>mallctlbymib()</function>.  Upon
+      successful return from <function>mallctlnametomib()</function>,
      <parameter>mibp</parameter> contains an array of
      <parameter>*miblenp</parameter> integers, where
      <parameter>*miblenp</parameter> is the lesser of the number of components
@@ -406,43 +406,44 @@ for (i = 0; i < nbins; i++) {

 	mib[2] = i;
 	len = sizeof(bin_size);
-	mallctlbymib(mib, miblen, &bin_size, &len, NULL, 0);
+	mallctlbymib(mib, miblen, (void *)&bin_size, &len, NULL, 0);
 	/* Do something with bin_size... */
 }]]></programlisting></para>

-      <para>The <function>malloc_stats_print<parameter/></function> function
-      writes human-readable summary statistics via the
-      <parameter>write_cb</parameter> callback function pointer and
-      <parameter>cbopaque</parameter> data passed to
-      <parameter>write_cb</parameter>, or
-      <function>malloc_message<parameter/></function> if
-      <parameter>write_cb</parameter> is <constant>NULL</constant>.  This
-      function can be called repeatedly.  General information that never
-      changes during execution can be omitted by specifying "g" as a character
+      <para>The <function>malloc_stats_print()</function> function writes
+      summary statistics via the <parameter>write_cb</parameter> callback
+      function pointer and <parameter>cbopaque</parameter> data passed to
+      <parameter>write_cb</parameter>, or <function>malloc_message()</function>
+      if <parameter>write_cb</parameter> is <constant>NULL</constant>.  The
+      statistics are presented in human-readable form unless <quote>J</quote> is
+      specified as a character within the <parameter>opts</parameter> string, in
+      which case the statistics are presented in <ulink
+      url="http://www.json.org/">JSON format</ulink>.  This function can be
+      called repeatedly.  General information that never changes during
+      execution can be omitted by specifying <quote>g</quote> as a character
      within the <parameter>opts</parameter> string.  Note that
-      <function>malloc_message<parameter/></function> uses the
-      <function>mallctl*<parameter/></function> functions internally, so
-      inconsistent statistics can be reported if multiple threads use these
-      functions simultaneously.  If <option>--enable-stats</option> is
-      specified during configuration, &ldquo;m&rdquo; and &ldquo;a&rdquo; can
-      be specified to omit merged arena and per arena statistics, respectively;
-      &ldquo;b&rdquo;, &ldquo;l&rdquo;, and &ldquo;h&rdquo; can be specified to
-      omit per size class statistics for bins, large objects, and huge objects,
-      respectively.  Unrecognized characters are silently ignored.  Note that
-      thread caching may prevent some statistics from being completely up to
-      date, since extra locking would be required to merge counters that track
-      thread cache operations.
-      </para>
+      <function>malloc_message()</function> uses the
+      <function>mallctl*()</function> functions internally, so inconsistent
+      statistics can be reported if multiple threads use these functions
+      simultaneously.  If <option>--enable-stats</option> is specified during
+      configuration, <quote>m</quote> and <quote>a</quote> can be specified to
+      omit merged arena and per arena statistics, respectively;
+      <quote>b</quote>, <quote>l</quote>, and <quote>h</quote> can be specified
+      to omit per size class statistics for bins, large objects, and huge
+      objects, respectively.  Unrecognized characters are silently ignored.
+      Note that thread caching may prevent some statistics from being completely
+      up to date, since extra locking would be required to merge counters that
+      track thread cache operations.</para>

-      <para>The <function>malloc_usable_size<parameter/></function> function
+      <para>The <function>malloc_usable_size()</function> function
      returns the usable size of the allocation pointed to by
      <parameter>ptr</parameter>.  The return value may be larger than the size
      that was requested during allocation.  The
-      <function>malloc_usable_size<parameter/></function> function is not a
-      mechanism for in-place <function>realloc<parameter/></function>; rather
+      <function>malloc_usable_size()</function> function is not a
+      mechanism for in-place <function>realloc()</function>; rather
      it is provided solely as a tool for introspection purposes.  Any
      discrepancy between the requested allocation size and the size reported
-      by <function>malloc_usable_size<parameter/></function> should not be
+      by <function>malloc_usable_size()</function> should not be
      depended on, since such behavior is entirely implementation-dependent.
      </para>
    </refsect2>
@@ -455,12 +456,12 @@ for (i = 0; i < nbins; i++) {

    <para>The string specified via <option>--with-malloc-conf</option>, the
    string pointed to by the global variable <varname>malloc_conf</varname>, the
-    &ldquo;name&rdquo; of the file referenced by the symbolic link named
+    <quote>name</quote> of the file referenced by the symbolic link named
    <filename class="symlink">/etc/malloc.conf</filename>, and the value of the
    environment variable <envar>MALLOC_CONF</envar>, will be interpreted, in
    that order, from left to right as options.  Note that
    <varname>malloc_conf</varname> may be read before
-    <function>main<parameter/></function> is entered, so the declaration of
+    <function>main()</function> is entered, so the declaration of
    <varname>malloc_conf</varname> should specify an initializer that contains
    the final value to be read by jemalloc.  <option>--with-malloc-conf</option>
    and <varname>malloc_conf</varname> are compile-time mechanisms, whereas
@@ -540,8 +541,8 @@ for (i = 0; i < nbins; i++) {
    are smaller than four times the page size, large size classes are smaller
    than the chunk size (see the <link
    linkend="opt.lg_chunk"><mallctl>opt.lg_chunk</mallctl></link> option), and
-    huge size classes extend from the chunk size up to one size class less than
-    the full address space size.</para>
+    huge size classes extend from the chunk size up to the largest size class
+    that does not exceed <constant>PTRDIFF_MAX</constant>.</para>

    <para>Allocations are packed tightly together, which can be an issue for
    multi-threaded applications.  If you need to assure that allocations do not
@@ -549,14 +550,14 @@ for (i = 0; i < nbins; i++) {
    nearest multiple of the cacheline size, or specify cacheline alignment when
    allocating.</para>

-    <para>The <function>realloc<parameter/></function>,
-    <function>rallocx<parameter/></function>, and
-    <function>xallocx<parameter/></function> functions may resize allocations
+    <para>The <function>realloc()</function>,
+    <function>rallocx()</function>, and
+    <function>xallocx()</function> functions may resize allocations
    without moving them under limited circumstances.  Unlike the
-    <function>*allocx<parameter/></function> API, the standard API does not
+    <function>*allocx()</function> API, the standard API does not
    officially round up the usable size of an allocation to the nearest size
    class, so technically it is necessary to call
-    <function>realloc<parameter/></function> to grow e.g. a 9-byte allocation to
+    <function>realloc()</function> to grow e.g. a 9-byte allocation to
    16 bytes, or shrink a 16-byte allocation to 9 bytes.  Growth and shrinkage
    trivially succeeds in place as long as the pre-size and post-size both round
    up to the same size class.  No other API guarantees are made regarding
@@ -659,7 +660,7 @@ for (i = 0; i < nbins; i++) {
          <entry>[1280 KiB, 1536 KiB, 1792 KiB]</entry>
        </row>
        <row>
-          <entry morerows="6">Huge</entry>
+          <entry morerows="8">Huge</entry>
          <entry>256 KiB</entry>
          <entry>[2 MiB]</entry>
        </row>
@@ -687,6 +688,14 @@ for (i = 0; i < nbins; i++) {
          <entry>...</entry>
          <entry>...</entry>
        </row>
+        <row>
+          <entry>512 PiB</entry>
+          <entry>[2560 PiB, 3 EiB, 3584 PiB, 4 EiB]</entry>
+        </row>
+        <row>
+          <entry>1 EiB</entry>
+          <entry>[5 EiB, 6 EiB, 7 EiB]</entry>
+        </row>
      </tbody>
      </tgroup>
    </table>
@@ -694,7 +703,7 @@ for (i = 0; i < nbins; i++) {
  <refsect1 id="mallctl_namespace">
    <title>MALLCTL NAMESPACE</title>
    <para>The following names are defined in the namespace accessible via the
-    <function>mallctl*<parameter/></function> functions.  Value types are
+    <function>mallctl*()</function> functions.  Value types are
    specified in parentheses, their readable/writable statuses are encoded as
    <literal>rw</literal>, <literal>r-</literal>, <literal>-w</literal>, or
    <literal>--</literal>, and required build configuration flags follow, if
@@ -725,7 +734,7 @@ for (i = 0; i < nbins; i++) {
          <literal>rw</literal>
        </term>
        <listitem><para>If a value is passed in, refresh the data from which
-        the <function>mallctl*<parameter/></function> functions report values,
+        the <function>mallctl*()</function> functions report values,
        and increment the epoch.  Return the current epoch.  This is useful for
        detecting whether another thread caused a refresh.</para></listitem>
      </varlistentry>
@@ -841,6 +850,17 @@ for (i = 0; i < nbins; i++) {
        during build configuration.</para></listitem>
      </varlistentry>

+      <varlistentry id="config.thp">
+        <term>
+          <mallctl>config.thp</mallctl>
+          (<type>bool</type>)
+          <literal>r-</literal>
+        </term>
+        <listitem><para><option>--disable-thp</option> was not specified
+        during build configuration, and the system supports transparent huge
+        page manipulation.</para></listitem>
+      </varlistentry>
+
      <varlistentry id="config.tls">
        <term>
          <mallctl>config.tls</mallctl>
@@ -909,12 +929,12 @@ for (i = 0; i < nbins; i++) {
        settings are supported if
        <citerefentry><refentrytitle>sbrk</refentrytitle>
        <manvolnum>2</manvolnum></citerefentry> is supported by the operating
-        system: &ldquo;disabled&rdquo;, &ldquo;primary&rdquo;, and
-        &ldquo;secondary&rdquo;; otherwise only &ldquo;disabled&rdquo; is
-        supported.  The default is &ldquo;secondary&rdquo; if
+        system: <quote>disabled</quote>, <quote>primary</quote>, and
+        <quote>secondary</quote>; otherwise only <quote>disabled</quote> is
+        supported.  The default is <quote>secondary</quote> if
        <citerefentry><refentrytitle>sbrk</refentrytitle>
        <manvolnum>2</manvolnum></citerefentry> is supported by the operating
-        system; &ldquo;disabled&rdquo; otherwise.
+        system; <quote>disabled</quote> otherwise.
        </para></listitem>
      </varlistentry>

@@ -1005,19 +1025,19 @@ for (i = 0; i < nbins; i++) {
          <literal>r-</literal>
        </term>
        <listitem><para>Enable/disable statistics printing at exit.  If
-        enabled, the <function>malloc_stats_print<parameter/></function>
+        enabled, the <function>malloc_stats_print()</function>
        function is called at program exit via an
        <citerefentry><refentrytitle>atexit</refentrytitle>
        <manvolnum>3</manvolnum></citerefentry> function.  If
        <option>--enable-stats</option> is specified during configuration, this
        has the potential to cause deadlock for a multi-threaded process that
        exits while one or more threads are executing in the memory allocation
-        functions.  Furthermore, <function>atexit<parameter/></function> may
+        functions.  Furthermore, <function>atexit()</function> may
        allocate memory during application initialization and then deadlock
        internally when jemalloc in turn calls
-        <function>atexit<parameter/></function>, so this option is not
+        <function>atexit()</function>, so this option is not
        universally usable (though the application can register its own
-        <function>atexit<parameter/></function> function with equivalent
+        <function>atexit()</function> function with equivalent
        functionality).  Therefore, this option should only be used with care;
        it is primarily intended as a performance tuning aid during application
        development.  This option is disabled by default.</para></listitem>
@@ -1030,15 +1050,16 @@ for (i = 0; i < nbins; i++) {
          <literal>r-</literal>
          [<option>--enable-fill</option>]
        </term>
-        <listitem><para>Junk filling.  If set to "alloc", each byte of
-        uninitialized allocated memory will be initialized to
-        <literal>0xa5</literal>.  If set to "free", all deallocated memory will
-        be initialized to <literal>0x5a</literal>.  If set to "true", both
-        allocated and deallocated memory will be initialized, and if set to
-        "false", junk filling be disabled entirely.  This is intended for
-        debugging and will impact performance negatively.  This option is
-        "false" by default unless <option>--enable-debug</option> is specified
-        during configuration, in which case it is "true" by default unless
+        <listitem><para>Junk filling.  If set to <quote>alloc</quote>, each byte
+        of uninitialized allocated memory will be initialized to
+        <literal>0xa5</literal>.  If set to <quote>free</quote>, all deallocated
+        memory will be initialized to <literal>0x5a</literal>.  If set to
+        <quote>true</quote>, both allocated and deallocated memory will be
+        initialized, and if set to <quote>false</quote>, junk filling be
+        disabled entirely.  This is intended for debugging and will impact
+        performance negatively.  This option is <quote>false</quote> by default
+        unless <option>--enable-debug</option> is specified during
+        configuration, in which case it is <quote>true</quote> by default unless
        running inside <ulink
        url="http://valgrind.org/">Valgrind</ulink>.</para></listitem>
      </varlistentry>
@@ -1093,8 +1114,8 @@ for (i = 0; i < nbins; i++) {
        <listitem><para>Zero filling enabled/disabled.  If enabled, each byte
        of uninitialized allocated memory will be initialized to 0.  Note that
        this initialization only happens once for each byte, so
-        <function>realloc<parameter/></function> and
-        <function>rallocx<parameter/></function> calls do not zero memory that
+        <function>realloc()</function> and
+        <function>rallocx()</function> calls do not zero memory that
        was previously allocated.  This is intended for debugging and will
        impact performance negatively.  This option is disabled by default.
        </para></listitem>
@@ -1152,6 +1173,21 @@ malloc_conf = "xmalloc:true";]]></programlisting>
        forcefully disabled.</para></listitem>
      </varlistentry>

+      <varlistentry id="opt.thp">
+        <term>
+          <mallctl>opt.thp</mallctl>
+          (<type>bool</type>)
+          <literal>r-</literal>
+          [<option>--enable-thp</option>]
+        </term>
+        <listitem><para>Transparent huge page (THP) integration
+        enabled/disabled.  When enabled, THPs are explicitly disabled as a side
+        effect of unused dirty page purging for chunks that back small and/or
+        large allocations, because such chunks typically comprise active,
+        unused dirty, and untouched clean pages.  This option is enabled by
+        default.</para></listitem>
+      </varlistentry>
+
      <varlistentry id="opt.lg_tcache_max">
        <term>
          <mallctl>opt.lg_tcache_max</mallctl>
@@ -1317,11 +1353,11 @@ malloc_conf = "xmalloc:true";]]></programlisting>
        <filename>&lt;prefix&gt;.&lt;pid&gt;.&lt;seq&gt;.f.heap</filename>,
        where <literal>&lt;prefix&gt;</literal> is controlled by the <link
        linkend="opt.prof_prefix"><mallctl>opt.prof_prefix</mallctl></link>
-        option.  Note that <function>atexit<parameter/></function> may allocate
+        option.  Note that <function>atexit()</function> may allocate
        memory during application initialization and then deadlock internally
-        when jemalloc in turn calls <function>atexit<parameter/></function>, so
+        when jemalloc in turn calls <function>atexit()</function>, so
        this option is not universally usable (though the application can
-        register its own <function>atexit<parameter/></function> function with
+        register its own <function>atexit()</function> function with
        equivalent functionality).  This option is disabled by
        default.</para></listitem>
      </varlistentry>
@@ -1380,7 +1416,7 @@ malloc_conf = "xmalloc:true";]]></programlisting>
        <link
        linkend="thread.allocated"><mallctl>thread.allocated</mallctl></link>
        mallctl.  This is useful for avoiding the overhead of repeated
-        <function>mallctl*<parameter/></function> calls.</para></listitem>
+        <function>mallctl*()</function> calls.</para></listitem>
      </varlistentry>

      <varlistentry id="thread.deallocated">
@@ -1407,7 +1443,7 @@ malloc_conf = "xmalloc:true";]]></programlisting>
        <link
        linkend="thread.deallocated"><mallctl>thread.deallocated</mallctl></link>
        mallctl.  This is useful for avoiding the overhead of repeated
-        <function>mallctl*<parameter/></function> calls.</para></listitem>
+        <function>mallctl*()</function> calls.</para></listitem>
      </varlistentry>

      <varlistentry id="thread.tcache.enabled">
@@ -1550,6 +1586,23 @@ malloc_conf = "xmalloc:true";]]></programlisting>
        details.</para></listitem>
      </varlistentry>

+      <varlistentry id="arena.i.reset">
+        <term>
+          <mallctl>arena.&lt;i&gt;.reset</mallctl>
+          (<type>void</type>)
+          <literal>--</literal>
+        </term>
+        <listitem><para>Discard all of the arena's extant allocations.  This
+        interface can only be used with arenas created via <link
+        linkend="arenas.extend"><mallctl>arenas.extend</mallctl></link>.  None
+        of the arena's discarded/cached allocations may accessed afterward.  As
+        part of this requirement, all thread caches which were used to
+        allocate/deallocate in conjunction with the arena must be flushed
+        beforehand.  This interface cannot be used if running inside Valgrind,
+        nor if the <link linkend="opt.quarantine">quarantine</link> size is
+        non-zero.</para></listitem>
+      </varlistentry>
+
      <varlistentry id="arena.i.dss">
        <term>
          <mallctl>arena.&lt;i&gt;.dss</mallctl>
@@ -2161,6 +2214,25 @@ typedef struct {
        linkend="stats.resident"><mallctl>stats.resident</mallctl></link>.</para></listitem>
      </varlistentry>

+      <varlistentry id="stats.retained">
+        <term>
+          <mallctl>stats.retained</mallctl>
+          (<type>size_t</type>)
+          <literal>r-</literal>
+          [<option>--enable-stats</option>]
+        </term>
+        <listitem><para>Total number of bytes in virtual memory mappings that
+        were retained rather than being returned to the operating system via
+        e.g. <citerefentry><refentrytitle>munmap</refentrytitle>
+        <manvolnum>2</manvolnum></citerefentry>.  Retained virtual memory is
+        typically untouched, decommitted, or purged, so it has no strongly
+        associated physical memory (see <link
+        linkend="arena.i.chunk_hooks">chunk hooks</link> for details).  Retained
+        memory is excluded from mapped memory statistics, e.g. <link
+        linkend="stats.mapped"><mallctl>stats.mapped</mallctl></link>.
+        </para></listitem>
+      </varlistentry>
+
      <varlistentry id="stats.arenas.i.dss">
        <term>
          <mallctl>stats.arenas.&lt;i&gt;.dss</mallctl>
@@ -2241,6 +2313,18 @@ typedef struct {
        <listitem><para>Number of mapped bytes.</para></listitem>
      </varlistentry>

+      <varlistentry id="stats.arenas.i.retained">
+        <term>
+          <mallctl>stats.arenas.&lt;i&gt;.retained</mallctl>
+          (<type>size_t</type>)
+          <literal>r-</literal>
+          [<option>--enable-stats</option>]
+        </term>
+        <listitem><para>Number of retained bytes.  See <link
+        linkend="stats.retained"><mallctl>stats.retained</mallctl></link> for
+        details.</para></listitem>
+      </varlistentry>
+
      <varlistentry id="stats.arenas.i.metadata.mapped">
        <term>
          <mallctl>stats.arenas.&lt;i&gt;.metadata.mapped</mallctl>
@@ -2678,7 +2762,7 @@ MAPPED_LIBRARIES:
    of run-time assertions that catch application errors such as double-free,
    write-after-free, etc.</para>

-    <para>Programs often accidentally depend on &ldquo;uninitialized&rdquo;
+    <para>Programs often accidentally depend on <quote>uninitialized</quote>
    memory actually being filled with zero bytes.  Junk filling
    (see the <link linkend="opt.junk"><mallctl>opt.junk</mallctl></link>
    option) tends to expose such bugs in the form of obviously incorrect
@@ -2707,29 +2791,29 @@ MAPPED_LIBRARIES:
    to override the function which emits the text strings forming the errors
    and warnings if for some reason the <constant>STDERR_FILENO</constant> file
    descriptor is not suitable for this.
-    <function>malloc_message<parameter/></function> takes the
+    <function>malloc_message()</function> takes the
    <parameter>cbopaque</parameter> pointer argument that is
    <constant>NULL</constant> unless overridden by the arguments in a call to
-    <function>malloc_stats_print<parameter/></function>, followed by a string
+    <function>malloc_stats_print()</function>, followed by a string
    pointer.  Please note that doing anything which tries to allocate memory in
    this function is likely to result in a crash or deadlock.</para>

    <para>All messages are prefixed by
-    &ldquo;<computeroutput>&lt;jemalloc&gt;: </computeroutput>&rdquo;.</para>
+    <quote><computeroutput>&lt;jemalloc&gt;: </computeroutput></quote>.</para>
  </refsect1>
  <refsect1 id="return_values">
    <title>RETURN VALUES</title>
    <refsect2>
      <title>Standard API</title>
-      <para>The <function>malloc<parameter/></function> and
-      <function>calloc<parameter/></function> functions return a pointer to the
+      <para>The <function>malloc()</function> and
+      <function>calloc()</function> functions return a pointer to the
      allocated memory if successful; otherwise a <constant>NULL</constant>
      pointer is returned and <varname>errno</varname> is set to
      <errorname>ENOMEM</errorname>.</para>

-      <para>The <function>posix_memalign<parameter/></function> function
+      <para>The <function>posix_memalign()</function> function
      returns the value 0 if successful; otherwise it returns an error value.
-      The <function>posix_memalign<parameter/></function> function will fail
+      The <function>posix_memalign()</function> function will fail
      if:
        <variablelist>
          <varlistentry>
@@ -2748,11 +2832,11 @@ MAPPED_LIBRARIES:
        </variablelist>
      </para>

-      <para>The <function>aligned_alloc<parameter/></function> function returns
+      <para>The <function>aligned_alloc()</function> function returns
      a pointer to the allocated memory if successful; otherwise a
      <constant>NULL</constant> pointer is returned and
      <varname>errno</varname> is set.  The
-      <function>aligned_alloc<parameter/></function> function will fail if:
+      <function>aligned_alloc()</function> function will fail if:
        <variablelist>
          <varlistentry>
            <term><errorname>EINVAL</errorname></term>
@@ -2769,44 +2853,44 @@ MAPPED_LIBRARIES:
        </variablelist>
      </para>

-      <para>The <function>realloc<parameter/></function> function returns a
+      <para>The <function>realloc()</function> function returns a
      pointer, possibly identical to <parameter>ptr</parameter>, to the
      allocated memory if successful; otherwise a <constant>NULL</constant>
      pointer is returned, and <varname>errno</varname> is set to
      <errorname>ENOMEM</errorname> if the error was the result of an
-      allocation failure.  The <function>realloc<parameter/></function>
+      allocation failure.  The <function>realloc()</function>
      function always leaves the original buffer intact when an error occurs.
      </para>

-      <para>The <function>free<parameter/></function> function returns no
+      <para>The <function>free()</function> function returns no
      value.</para>
    </refsect2>
    <refsect2>
      <title>Non-standard API</title>
-      <para>The <function>mallocx<parameter/></function> and
-      <function>rallocx<parameter/></function> functions return a pointer to
+      <para>The <function>mallocx()</function> and
+      <function>rallocx()</function> functions return a pointer to
      the allocated memory if successful; otherwise a <constant>NULL</constant>
      pointer is returned to indicate insufficient contiguous memory was
      available to service the allocation request.  </para>

-      <para>The <function>xallocx<parameter/></function> function returns the
+      <para>The <function>xallocx()</function> function returns the
      real size of the resulting resized allocation pointed to by
      <parameter>ptr</parameter>, which is a value less than
      <parameter>size</parameter> if the allocation could not be adequately
      grown in place.  </para>

-      <para>The <function>sallocx<parameter/></function> function returns the
+      <para>The <function>sallocx()</function> function returns the
      real size of the allocation pointed to by <parameter>ptr</parameter>.
      </para>

-      <para>The <function>nallocx<parameter/></function> returns the real size
+      <para>The <function>nallocx()</function> returns the real size
      that would result from a successful equivalent
-      <function>mallocx<parameter/></function> function call, or zero if
+      <function>mallocx()</function> function call, or zero if
      insufficient memory is available to perform the size computation.  </para>

-      <para>The <function>mallctl<parameter/></function>,
-      <function>mallctlnametomib<parameter/></function>, and
-      <function>mallctlbymib<parameter/></function> functions return 0 on
+      <para>The <function>mallctl()</function>,
+      <function>mallctlnametomib()</function>, and
+      <function>mallctlbymib()</function> functions return 0 on
      success; otherwise they return an error value.  The functions will fail
      if:
        <variablelist>
@@ -2842,13 +2926,13 @@ MAPPED_LIBRARIES:
            <term><errorname>EFAULT</errorname></term>

            <listitem><para>An interface with side effects failed in some way
-            not directly related to <function>mallctl*<parameter/></function>
+            not directly related to <function>mallctl*()</function>
            read/write processing.</para></listitem>
          </varlistentry>
        </variablelist>
      </para>

-      <para>The <function>malloc_usable_size<parameter/></function> function
+      <para>The <function>malloc_usable_size()</function> function
      returns the usable size of the allocation pointed to by
      <parameter>ptr</parameter>.  </para>
    </refsect2>
@@ -2896,13 +2980,13 @@ malloc_conf = "lg_chunk:24";]]></programlisting></para>
  </refsect1>
  <refsect1 id="standards">
    <title>STANDARDS</title>
-    <para>The <function>malloc<parameter/></function>,
-    <function>calloc<parameter/></function>,
-    <function>realloc<parameter/></function>, and
-    <function>free<parameter/></function> functions conform to ISO/IEC
-    9899:1990 (&ldquo;ISO C90&rdquo;).</para>
+    <para>The <function>malloc()</function>,
+    <function>calloc()</function>,
+    <function>realloc()</function>, and
+    <function>free()</function> functions conform to ISO/IEC
+    9899:1990 (<quote>ISO C90</quote>).</para>

-    <para>The <function>posix_memalign<parameter/></function> function conforms
-    to IEEE Std 1003.1-2001 (&ldquo;POSIX.1&rdquo;).</para>
+    <para>The <function>posix_memalign()</function> function conforms
+    to IEEE Std 1003.1-2001 (<quote>POSIX.1</quote>).</para>
  </refsect1>
 </refentry>
--- a/doc/stylesheet.xsl
+++ b/doc/stylesheet.xsl
@@ -1,7 +1,10 @@
 <xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">
  <xsl:param name="funcsynopsis.style">ansi</xsl:param>
-  <xsl:param name="function.parens" select="1"/>
+  <xsl:param name="function.parens" select="0"/>
+  <xsl:template match="function">
+    <xsl:call-template name="inline.monoseq"/>
+  </xsl:template>
  <xsl:template match="mallctl">
-    "<xsl:call-template name="inline.monoseq"/>"
+    <quote><xsl:call-template name="inline.monoseq"/></quote>
  </xsl:template>
 </xsl:stylesheet>
--- a/include/jemalloc/internal/arena.h
+++ b/include/jemalloc/internal/arena.h
@@ -36,11 +36,13 @@ typedef enum {
 #define	DECAY_NTICKS_PER_UPDATE	1000

 typedef struct arena_runs_dirty_link_s arena_runs_dirty_link_t;
+typedef struct arena_avail_links_s arena_avail_links_t;
 typedef struct arena_run_s arena_run_t;
 typedef struct arena_chunk_map_bits_s arena_chunk_map_bits_t;
 typedef struct arena_chunk_map_misc_s arena_chunk_map_misc_t;
 typedef struct arena_chunk_s arena_chunk_t;
 typedef struct arena_bin_info_s arena_bin_info_t;
+typedef struct arena_decay_s arena_decay_t;
 typedef struct arena_bin_s arena_bin_t;
 typedef struct arena_s arena_t;
 typedef struct arena_tdata_s arena_tdata_t;
@@ -153,13 +155,13 @@ struct arena_runs_dirty_link_s {
 */
 struct arena_chunk_map_misc_s {
 	/*
-	 * Linkage for run trees.  There are two disjoint uses:
+	 * Linkage for run heaps.  There are two disjoint uses:
 	 *
-	 * 1) arena_t's runs_avail tree.
+	 * 1) arena_t's runs_avail heaps.
 	 * 2) arena_run_t conceptually uses this linkage for in-use non-full
 	 *    runs, rather than directly embedding linkage.
 	 */
-	rb_node(arena_chunk_map_misc_t)		rb_link;
+	phn(arena_chunk_map_misc_t)		ph_link;

 	union {
 		/* Linkage for list of dirty runs. */
@@ -175,7 +177,7 @@ struct arena_chunk_map_misc_s {
 		arena_run_t			run;
 	};
 };
-typedef rb_tree(arena_chunk_map_misc_t) arena_run_tree_t;
+typedef ph(arena_chunk_map_misc_t) arena_run_heap_t;
 #endif /* JEMALLOC_ARENA_STRUCTS_A */

 #ifdef JEMALLOC_ARENA_STRUCTS_B
@@ -188,6 +190,14 @@ struct arena_chunk_s {
 	 */
 	extent_node_t		node;

+	/*
+	 * True if memory could be backed by transparent huge pages.  This is
+	 * only directly relevant to Linux, since it is the only supported
+	 * platform on which jemalloc interacts with explicit transparent huge
+	 * page controls.
+	 */
+	bool			hugepage;
+
 	/*
 	 * 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
@@ -256,6 +266,49 @@ struct arena_bin_info_s {
 	uint32_t		reg0_offset;
 };

+struct arena_decay_s {
+	/*
+	 * Approximate time in seconds from the creation of a set of unused
+	 * dirty pages until an equivalent set of unused dirty pages is purged
+	 * and/or reused.
+	 */
+	ssize_t			time;
+	/* time / SMOOTHSTEP_NSTEPS. */
+	nstime_t		interval;
+	/*
+	 * Time at which the current decay interval logically started.  We do
+	 * not actually advance to a new epoch until sometime after it starts
+	 * because of scheduling and computation delays, and it is even possible
+	 * to completely skip epochs.  In all cases, during epoch advancement we
+	 * merge all relevant activity into the most recently recorded epoch.
+	 */
+	nstime_t		epoch;
+	/* Deadline randomness generator. */
+	uint64_t		jitter_state;
+	/*
+	 * Deadline for current epoch.  This is the sum of interval and per
+	 * epoch jitter which is a uniform random variable in [0..interval).
+	 * Epochs always advance by precise multiples of interval, but we
+	 * randomize the deadline to reduce the likelihood of arenas purging in
+	 * lockstep.
+	 */
+	nstime_t		deadline;
+	/*
+	 * Number of dirty pages at beginning of current epoch.  During epoch
+	 * advancement we use the delta between arena->decay.ndirty and
+	 * arena->ndirty to determine how many dirty pages, if any, were
+	 * generated.
+	 */
+	size_t			ndirty;
+	/*
+	 * Trailing log of how many unused dirty pages were generated during
+	 * each of the past SMOOTHSTEP_NSTEPS decay epochs, where the last
+	 * element is the most recent epoch.  Corresponding epoch times are
+	 * relative to epoch.
+	 */
+	size_t			backlog[SMOOTHSTEP_NSTEPS];
+};
+
 struct arena_bin_s {
 	/*
 	 * All operations on runcur, runs, and stats require that lock be
@@ -272,13 +325,13 @@ struct arena_bin_s {
 	arena_run_t		*runcur;

 	/*
-	 * Tree of non-full runs.  This tree is used when looking for an
+	 * Heap of non-full runs.  This heap 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;
+	arena_run_heap_t	runs;

 	/* Bin statistics. */
 	malloc_bin_stats_t	stats;
@@ -289,10 +342,18 @@ struct arena_s {
 	unsigned		ind;

 	/*
-	 * Number of threads currently assigned to this arena.  This field is
-	 * synchronized via atomic operations.
+	 * Number of threads currently assigned to this arena, synchronized via
+	 * atomic operations.  Each thread has two distinct assignments, one for
+	 * application-serving allocation, and the other for internal metadata
+	 * allocation.  Internal metadata must not be allocated from arenas
+	 * created via the arenas.extend mallctl, because the arena.<i>.reset
+	 * mallctl indiscriminately discards all allocations for the affected
+	 * arena.
+	 *
+	 *   0: Application allocation.
+	 *   1: Internal metadata allocation.
 	 */
-	unsigned		nthreads;
+	unsigned		nthreads[2];

 	/*
 	 * There are three classes of arena operations from a locking
@@ -317,10 +378,16 @@ struct arena_s {
 	 * PRNG state for cache index randomization of large allocation base
 	 * pointers.
 	 */
-	uint64_t		offset_state;
+	size_t			offset_state;

 	dss_prec_t		dss_prec;

+	/* Extant arena chunks. */
+	ql_head(extent_node_t)	achunks;
+
+	/* Extent serial number generator state. */
+	size_t			extent_sn_next;
+
 	/*
 	 * In order to avoid rapid chunk allocation/deallocation when an arena
 	 * oscillates right on the cusp of needing a new chunk, cache the most
@@ -381,52 +448,8 @@ struct arena_s {
 	arena_runs_dirty_link_t	runs_dirty;
 	extent_node_t		chunks_cache;

-	/*
-	 * Approximate time in seconds from the creation of a set of unused
-	 * dirty pages until an equivalent set of unused dirty pages is purged
-	 * and/or reused.
-	 */
-	ssize_t			decay_time;
-	/* decay_time / SMOOTHSTEP_NSTEPS. */
-	nstime_t		decay_interval;
-	/*
-	 * Time at which the current decay interval logically started.  We do
-	 * not actually advance to a new epoch until sometime after it starts
-	 * because of scheduling and computation delays, and it is even possible
-	 * to completely skip epochs.  In all cases, during epoch advancement we
-	 * merge all relevant activity into the most recently recorded epoch.
-	 */
-	nstime_t		decay_epoch;
-	/* decay_deadline randomness generator. */
-	uint64_t		decay_jitter_state;
-	/*
-	 * Deadline for current epoch.  This is the sum of decay_interval and
-	 * per epoch jitter which is a uniform random variable in
-	 * [0..decay_interval).  Epochs always advance by precise multiples of
-	 * decay_interval, but we randomize the deadline to reduce the
-	 * likelihood of arenas purging in lockstep.
-	 */
-	nstime_t		decay_deadline;
-	/*
-	 * Number of dirty pages at beginning of current epoch.  During epoch
-	 * advancement we use the delta between decay_ndirty and ndirty to
-	 * determine how many dirty pages, if any, were generated, and record
-	 * the result in decay_backlog.
-	 */
-	size_t			decay_ndirty;
-	/*
-	 * Memoized result of arena_decay_backlog_npages_limit() corresponding
-	 * to the current contents of decay_backlog, i.e. the limit on how many
-	 * pages are allowed to exist for the decay epochs.
-	 */
-	size_t			decay_backlog_npages_limit;
-	/*
-	 * Trailing log of how many unused dirty pages were generated during
-	 * each of the past SMOOTHSTEP_NSTEPS decay epochs, where the last
-	 * element is the most recent epoch.  Corresponding epoch times are
-	 * relative to decay_epoch.
-	 */
-	size_t			decay_backlog[SMOOTHSTEP_NSTEPS];
+	/* Decay-based purging state. */
+	arena_decay_t		decay;

 	/* Extant huge allocations. */
 	ql_head(extent_node_t)	huge;
@@ -440,9 +463,9 @@ struct arena_s {
 	 * orderings are needed, which is why there are two trees with the same
 	 * contents.
 	 */
-	extent_tree_t		chunks_szad_cached;
+	extent_tree_t		chunks_szsnad_cached;
 	extent_tree_t		chunks_ad_cached;
-	extent_tree_t		chunks_szad_retained;
+	extent_tree_t		chunks_szsnad_retained;
 	extent_tree_t		chunks_ad_retained;

 	malloc_mutex_t		chunks_mtx;
@@ -457,10 +480,12 @@ struct arena_s {
 	arena_bin_t		bins[NBINS];

 	/*
-	 * Quantized address-ordered trees of this arena's available runs.  The
-	 * trees are used for first-best-fit run allocation.
+	 * Size-segregated address-ordered heaps of this arena's available runs,
+	 * used for first-best-fit run allocation.  Runs are quantized, i.e.
+	 * they reside in the last heap which corresponds to a size class less
+	 * than or equal to the run size.
 	 */
-	arena_run_tree_t	runs_avail[1]; /* Dynamically sized. */
+	arena_run_heap_t	runs_avail[NPSIZES];
 };

 /* Used in conjunction with tsd for fast arena-related context lookup. */
@@ -481,6 +506,7 @@ static const size_t	large_pad =
 #endif
    ;

+extern bool		opt_thp;
 extern purge_mode_t	opt_purge;
 extern const char	*purge_mode_names[];
 extern ssize_t		opt_lg_dirty_mult;
@@ -492,7 +518,6 @@ extern size_t		map_bias; /* Number of arena chunk header pages. */
 extern size_t		map_misc_offset;
 extern size_t		arena_maxrun; /* Max run size for arenas. */
 extern size_t		large_maxclass; /* Max large size class. */
-extern size_t		run_quantize_max; /* Max run_quantize_*() input. */
 extern unsigned		nlclasses; /* Number of large size classes. */
 extern unsigned		nhclasses; /* Number of huge size classes. */

@@ -505,25 +530,28 @@ void	arena_chunk_cache_maybe_insert(arena_t *arena, extent_node_t *node,
    bool cache);
 void	arena_chunk_cache_maybe_remove(arena_t *arena, extent_node_t *node,
    bool cache);
-extent_node_t	*arena_node_alloc(arena_t *arena);
-void	arena_node_dalloc(arena_t *arena, extent_node_t *node);
-void	*arena_chunk_alloc_huge(arena_t *arena, size_t usize, size_t alignment,
-    bool *zero);
-void	arena_chunk_dalloc_huge(arena_t *arena, void *chunk, size_t usize);
-void	arena_chunk_ralloc_huge_similar(arena_t *arena, void *chunk,
-    size_t oldsize, size_t usize);
-void	arena_chunk_ralloc_huge_shrink(arena_t *arena, void *chunk,
-    size_t oldsize, size_t usize);
-bool	arena_chunk_ralloc_huge_expand(arena_t *arena, void *chunk,
-    size_t oldsize, size_t usize, bool *zero);
-ssize_t	arena_lg_dirty_mult_get(arena_t *arena);
-bool	arena_lg_dirty_mult_set(arena_t *arena, ssize_t lg_dirty_mult);
-ssize_t	arena_decay_time_get(arena_t *arena);
-bool	arena_decay_time_set(arena_t *arena, ssize_t decay_time);
-void	arena_maybe_purge(arena_t *arena);
-void	arena_purge(arena_t *arena, bool all);
-void	arena_tcache_fill_small(tsd_t *tsd, arena_t *arena, tcache_bin_t *tbin,
-    szind_t binind, uint64_t prof_accumbytes);
+extent_node_t	*arena_node_alloc(tsdn_t *tsdn, arena_t *arena);
+void	arena_node_dalloc(tsdn_t *tsdn, arena_t *arena, extent_node_t *node);
+void	*arena_chunk_alloc_huge(tsdn_t *tsdn, arena_t *arena, size_t usize,
+    size_t alignment, size_t *sn, bool *zero);
+void	arena_chunk_dalloc_huge(tsdn_t *tsdn, arena_t *arena, void *chunk,
+    size_t usize, size_t sn);
+void	arena_chunk_ralloc_huge_similar(tsdn_t *tsdn, arena_t *arena,
+    void *chunk, size_t oldsize, size_t usize);
+void	arena_chunk_ralloc_huge_shrink(tsdn_t *tsdn, arena_t *arena,
+    void *chunk, size_t oldsize, size_t usize, size_t sn);
+bool	arena_chunk_ralloc_huge_expand(tsdn_t *tsdn, arena_t *arena,
+    void *chunk, size_t oldsize, size_t usize, bool *zero);
+ssize_t	arena_lg_dirty_mult_get(tsdn_t *tsdn, arena_t *arena);
+bool	arena_lg_dirty_mult_set(tsdn_t *tsdn, arena_t *arena,
+    ssize_t lg_dirty_mult);
+ssize_t	arena_decay_time_get(tsdn_t *tsdn, arena_t *arena);
+bool	arena_decay_time_set(tsdn_t *tsdn, arena_t *arena, ssize_t decay_time);
+void	arena_purge(tsdn_t *tsdn, arena_t *arena, bool all);
+void	arena_maybe_purge(tsdn_t *tsdn, arena_t *arena);
+void	arena_reset(tsd_t *tsd, arena_t *arena);
+void	arena_tcache_fill_small(tsdn_t *tsdn, arena_t *arena,
+    tcache_bin_t *tbin, szind_t binind, uint64_t prof_accumbytes);
 void	arena_alloc_junk_small(void *ptr, arena_bin_info_t *bin_info,
    bool zero);
 #ifdef JEMALLOC_JET
@@ -536,17 +564,18 @@ extern arena_dalloc_junk_small_t *arena_dalloc_junk_small;
 void	arena_dalloc_junk_small(void *ptr, arena_bin_info_t *bin_info);
 #endif
 void	arena_quarantine_junk_small(void *ptr, size_t usize);
-void	*arena_malloc_large(tsd_t *tsd, arena_t *arena, szind_t ind, bool zero);
-void	*arena_malloc_hard(tsd_t *tsd, arena_t *arena, size_t size, szind_t ind,
-    bool zero, tcache_t *tcache);
-void	*arena_palloc(tsd_t *tsd, arena_t *arena, size_t usize,
+void	*arena_malloc_large(tsdn_t *tsdn, arena_t *arena, szind_t ind,
+    bool zero);
+void	*arena_malloc_hard(tsdn_t *tsdn, arena_t *arena, size_t size,
+    szind_t ind, bool zero);
+void	*arena_palloc(tsdn_t *tsdn, arena_t *arena, size_t usize,
    size_t alignment, bool zero, tcache_t *tcache);
-void	arena_prof_promoted(const void *ptr, size_t size);
-void	arena_dalloc_bin_junked_locked(arena_t *arena, arena_chunk_t *chunk,
-    void *ptr, arena_chunk_map_bits_t *bitselm);
-void	arena_dalloc_bin(arena_t *arena, arena_chunk_t *chunk, void *ptr,
-    size_t pageind, arena_chunk_map_bits_t *bitselm);
-void	arena_dalloc_small(tsd_t *tsd, arena_t *arena, arena_chunk_t *chunk,
+void	arena_prof_promoted(tsdn_t *tsdn, const void *ptr, size_t size);
+void	arena_dalloc_bin_junked_locked(tsdn_t *tsdn, arena_t *arena,
+    arena_chunk_t *chunk, void *ptr, arena_chunk_map_bits_t *bitselm);
+void	arena_dalloc_bin(tsdn_t *tsdn, arena_t *arena, arena_chunk_t *chunk,
+    void *ptr, size_t pageind, arena_chunk_map_bits_t *bitselm);
+void	arena_dalloc_small(tsdn_t *tsdn, arena_t *arena, arena_chunk_t *chunk,
    void *ptr, size_t pageind);
 #ifdef JEMALLOC_JET
 typedef void (arena_dalloc_junk_large_t)(void *, size_t);
@@ -554,70 +583,81 @@ extern arena_dalloc_junk_large_t *arena_dalloc_junk_large;
 #else
 void	arena_dalloc_junk_large(void *ptr, size_t usize);
 #endif
-void	arena_dalloc_large_junked_locked(arena_t *arena, arena_chunk_t *chunk,
-    void *ptr);
-void	arena_dalloc_large(tsd_t *tsd, arena_t *arena, arena_chunk_t *chunk,
+void	arena_dalloc_large_junked_locked(tsdn_t *tsdn, arena_t *arena,
+    arena_chunk_t *chunk, void *ptr);
+void	arena_dalloc_large(tsdn_t *tsdn, arena_t *arena, arena_chunk_t *chunk,
    void *ptr);
 #ifdef JEMALLOC_JET
 typedef void (arena_ralloc_junk_large_t)(void *, size_t, size_t);
 extern arena_ralloc_junk_large_t *arena_ralloc_junk_large;
 #endif
-bool	arena_ralloc_no_move(tsd_t *tsd, void *ptr, size_t oldsize, size_t size,
-    size_t extra, bool zero);
+bool	arena_ralloc_no_move(tsdn_t *tsdn, void *ptr, size_t oldsize,
+    size_t size, size_t extra, bool zero);
 void	*arena_ralloc(tsd_t *tsd, arena_t *arena, void *ptr, size_t oldsize,
    size_t size, size_t alignment, bool zero, tcache_t *tcache);
-dss_prec_t	arena_dss_prec_get(arena_t *arena);
-bool	arena_dss_prec_set(arena_t *arena, dss_prec_t dss_prec);
+dss_prec_t	arena_dss_prec_get(tsdn_t *tsdn, arena_t *arena);
+bool	arena_dss_prec_set(tsdn_t *tsdn, arena_t *arena, dss_prec_t dss_prec);
 ssize_t	arena_lg_dirty_mult_default_get(void);
 bool	arena_lg_dirty_mult_default_set(ssize_t lg_dirty_mult);
 ssize_t	arena_decay_time_default_get(void);
 bool	arena_decay_time_default_set(ssize_t decay_time);
-void	arena_basic_stats_merge(arena_t *arena, unsigned *nthreads,
+void	arena_basic_stats_merge(tsdn_t *tsdn, arena_t *arena,
+    unsigned *nthreads, const char **dss, ssize_t *lg_dirty_mult,
+    ssize_t *decay_time, size_t *nactive, size_t *ndirty);
+void	arena_stats_merge(tsdn_t *tsdn, arena_t *arena, unsigned *nthreads,
    const char **dss, ssize_t *lg_dirty_mult, ssize_t *decay_time,
-    size_t *nactive, size_t *ndirty);
-void	arena_stats_merge(arena_t *arena, unsigned *nthreads, const char **dss,
-    ssize_t *lg_dirty_mult, ssize_t *decay_time, size_t *nactive,
-    size_t *ndirty, arena_stats_t *astats, malloc_bin_stats_t *bstats,
-    malloc_large_stats_t *lstats, malloc_huge_stats_t *hstats);
-unsigned	arena_nthreads_get(arena_t *arena);
-void	arena_nthreads_inc(arena_t *arena);
-void	arena_nthreads_dec(arena_t *arena);
-arena_t	*arena_new(unsigned ind);
-bool	arena_boot(void);
-void	arena_prefork0(arena_t *arena);
-void	arena_prefork1(arena_t *arena);
-void	arena_prefork2(arena_t *arena);
-void	arena_prefork3(arena_t *arena);
-void	arena_postfork_parent(arena_t *arena);
-void	arena_postfork_child(arena_t *arena);
+    size_t *nactive, size_t *ndirty, arena_stats_t *astats,
+    malloc_bin_stats_t *bstats, malloc_large_stats_t *lstats,
+    malloc_huge_stats_t *hstats);
+unsigned	arena_nthreads_get(arena_t *arena, bool internal);
+void	arena_nthreads_inc(arena_t *arena, bool internal);
+void	arena_nthreads_dec(arena_t *arena, bool internal);
+size_t	arena_extent_sn_next(arena_t *arena);
+arena_t	*arena_new(tsdn_t *tsdn, unsigned ind);
+void	arena_boot(void);
+void	arena_prefork0(tsdn_t *tsdn, arena_t *arena);
+void	arena_prefork1(tsdn_t *tsdn, arena_t *arena);
+void	arena_prefork2(tsdn_t *tsdn, arena_t *arena);
+void	arena_prefork3(tsdn_t *tsdn, arena_t *arena);
+void	arena_postfork_parent(tsdn_t *tsdn, arena_t *arena);
+void	arena_postfork_child(tsdn_t *tsdn, arena_t *arena);

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

 #ifndef JEMALLOC_ENABLE_INLINE
-arena_chunk_map_bits_t	*arena_bitselm_get(arena_chunk_t *chunk,
+arena_chunk_map_bits_t	*arena_bitselm_get_mutable(arena_chunk_t *chunk,
    size_t pageind);
-arena_chunk_map_misc_t	*arena_miscelm_get(arena_chunk_t *chunk,
+const arena_chunk_map_bits_t	*arena_bitselm_get_const(
+    const arena_chunk_t *chunk, size_t pageind);
+arena_chunk_map_misc_t	*arena_miscelm_get_mutable(arena_chunk_t *chunk,
    size_t pageind);
+const arena_chunk_map_misc_t	*arena_miscelm_get_const(
+    const arena_chunk_t *chunk, size_t pageind);
 size_t	arena_miscelm_to_pageind(const arena_chunk_map_misc_t *miscelm);
-void	*arena_miscelm_to_rpages(arena_chunk_map_misc_t *miscelm);
+void	*arena_miscelm_to_rpages(const arena_chunk_map_misc_t *miscelm);
 arena_chunk_map_misc_t	*arena_rd_to_miscelm(arena_runs_dirty_link_t *rd);
 arena_chunk_map_misc_t	*arena_run_to_miscelm(arena_run_t *run);
-size_t	*arena_mapbitsp_get(arena_chunk_t *chunk, size_t pageind);
-size_t	arena_mapbitsp_read(size_t *mapbitsp);
-size_t	arena_mapbits_get(arena_chunk_t *chunk, size_t pageind);
-size_t	arena_mapbits_size_decode(size_t mapbits);
-size_t	arena_mapbits_unallocated_size_get(arena_chunk_t *chunk,
+size_t	*arena_mapbitsp_get_mutable(arena_chunk_t *chunk, size_t pageind);
+const size_t	*arena_mapbitsp_get_const(const 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);
-szind_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_decommitted_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);
+size_t	arena_mapbitsp_read(const size_t *mapbitsp);
+size_t	arena_mapbits_get(const arena_chunk_t *chunk, size_t pageind);
+size_t	arena_mapbits_size_decode(size_t mapbits);
+size_t	arena_mapbits_unallocated_size_get(const arena_chunk_t *chunk,
+    size_t pageind);
+size_t	arena_mapbits_large_size_get(const arena_chunk_t *chunk,
+    size_t pageind);
+size_t	arena_mapbits_small_runind_get(const arena_chunk_t *chunk,
+    size_t pageind);
+szind_t	arena_mapbits_binind_get(const arena_chunk_t *chunk, size_t pageind);
+size_t	arena_mapbits_dirty_get(const arena_chunk_t *chunk, size_t pageind);
+size_t	arena_mapbits_unzeroed_get(const arena_chunk_t *chunk, size_t pageind);
+size_t	arena_mapbits_decommitted_get(const arena_chunk_t *chunk,
+    size_t pageind);
+size_t	arena_mapbits_large_get(const arena_chunk_t *chunk, size_t pageind);
+size_t	arena_mapbits_allocated_get(const arena_chunk_t *chunk, size_t pageind);
 void	arena_mapbitsp_write(size_t *mapbitsp, size_t mapbits);
 size_t	arena_mapbits_size_encode(size_t size);
 void	arena_mapbits_unallocated_set(arena_chunk_t *chunk, size_t pageind,
@@ -637,29 +677,31 @@ void	arena_metadata_allocated_sub(arena_t *arena, size_t size);
 size_t	arena_metadata_allocated_get(arena_t *arena);
 bool	arena_prof_accum_impl(arena_t *arena, uint64_t accumbytes);
 bool	arena_prof_accum_locked(arena_t *arena, uint64_t accumbytes);
-bool	arena_prof_accum(arena_t *arena, uint64_t accumbytes);
+bool	arena_prof_accum(tsdn_t *tsdn, arena_t *arena, uint64_t accumbytes);
 szind_t	arena_ptr_small_binind_get(const void *ptr, size_t mapbits);
 szind_t	arena_bin_index(arena_t *arena, arena_bin_t *bin);
 size_t	arena_run_regind(arena_run_t *run, arena_bin_info_t *bin_info,
    const void *ptr);
-prof_tctx_t	*arena_prof_tctx_get(const void *ptr);
-void	arena_prof_tctx_set(const void *ptr, size_t usize, prof_tctx_t *tctx);
-void	arena_prof_tctx_reset(const void *ptr, size_t usize,
+prof_tctx_t	*arena_prof_tctx_get(tsdn_t *tsdn, const void *ptr);
+void	arena_prof_tctx_set(tsdn_t *tsdn, const void *ptr, size_t usize,
+    prof_tctx_t *tctx);
+void	arena_prof_tctx_reset(tsdn_t *tsdn, const void *ptr, size_t usize,
    const void *old_ptr, prof_tctx_t *old_tctx);
-void	arena_decay_ticks(tsd_t *tsd, arena_t *arena, unsigned nticks);
-void	arena_decay_tick(tsd_t *tsd, arena_t *arena);
-void	*arena_malloc(tsd_t *tsd, arena_t *arena, size_t size, szind_t ind,
+void	arena_decay_ticks(tsdn_t *tsdn, arena_t *arena, unsigned nticks);
+void	arena_decay_tick(tsdn_t *tsdn, arena_t *arena);
+void	*arena_malloc(tsdn_t *tsdn, arena_t *arena, size_t size, szind_t ind,
    bool zero, tcache_t *tcache, bool slow_path);
 arena_t	*arena_aalloc(const void *ptr);
-size_t	arena_salloc(const void *ptr, bool demote);
-void	arena_dalloc(tsd_t *tsd, void *ptr, tcache_t *tcache, bool slow_path);
-void	arena_sdalloc(tsd_t *tsd, void *ptr, size_t size, tcache_t *tcache);
+size_t	arena_salloc(tsdn_t *tsdn, const void *ptr, bool demote);
+void	arena_dalloc(tsdn_t *tsdn, void *ptr, tcache_t *tcache, bool slow_path);
+void	arena_sdalloc(tsdn_t *tsdn, void *ptr, size_t size, tcache_t *tcache,
+    bool slow_path);
 #endif

 #if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_ARENA_C_))
 #  ifdef JEMALLOC_ARENA_INLINE_A
 JEMALLOC_ALWAYS_INLINE arena_chunk_map_bits_t *
-arena_bitselm_get(arena_chunk_t *chunk, size_t pageind)
+arena_bitselm_get_mutable(arena_chunk_t *chunk, size_t pageind)
 {

 	assert(pageind >= map_bias);
@@ -668,8 +710,15 @@ arena_bitselm_get(arena_chunk_t *chunk, size_t pageind)
 	return (&chunk->map_bits[pageind-map_bias]);
 }

+JEMALLOC_ALWAYS_INLINE const arena_chunk_map_bits_t *
+arena_bitselm_get_const(const arena_chunk_t *chunk, size_t pageind)
+{
+
+	return (arena_bitselm_get_mutable((arena_chunk_t *)chunk, pageind));
+}
+
 JEMALLOC_ALWAYS_INLINE arena_chunk_map_misc_t *
-arena_miscelm_get(arena_chunk_t *chunk, size_t pageind)
+arena_miscelm_get_mutable(arena_chunk_t *chunk, size_t pageind)
 {

 	assert(pageind >= map_bias);
@@ -679,6 +728,13 @@ arena_miscelm_get(arena_chunk_t *chunk, size_t pageind)
 	    (uintptr_t)map_misc_offset) + pageind-map_bias);
 }

+JEMALLOC_ALWAYS_INLINE const arena_chunk_map_misc_t *
+arena_miscelm_get_const(const arena_chunk_t *chunk, size_t pageind)
+{
+
+	return (arena_miscelm_get_mutable((arena_chunk_t *)chunk, pageind));
+}
+
 JEMALLOC_ALWAYS_INLINE size_t
 arena_miscelm_to_pageind(const arena_chunk_map_misc_t *miscelm)
 {
@@ -693,7 +749,7 @@ arena_miscelm_to_pageind(const arena_chunk_map_misc_t *miscelm)
 }

 JEMALLOC_ALWAYS_INLINE void *
-arena_miscelm_to_rpages(arena_chunk_map_misc_t *miscelm)
+arena_miscelm_to_rpages(const arena_chunk_map_misc_t *miscelm)
 {
 	arena_chunk_t *chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(miscelm);
 	size_t pageind = arena_miscelm_to_pageind(miscelm);
@@ -726,24 +782,31 @@ arena_run_to_miscelm(arena_run_t *run)
 }

 JEMALLOC_ALWAYS_INLINE size_t *
-arena_mapbitsp_get(arena_chunk_t *chunk, size_t pageind)
+arena_mapbitsp_get_mutable(arena_chunk_t *chunk, size_t pageind)
 {

-	return (&arena_bitselm_get(chunk, pageind)->bits);
+	return (&arena_bitselm_get_mutable(chunk, pageind)->bits);
+}
+
+JEMALLOC_ALWAYS_INLINE const size_t *
+arena_mapbitsp_get_const(const arena_chunk_t *chunk, size_t pageind)
+{
+
+	return (arena_mapbitsp_get_mutable((arena_chunk_t *)chunk, pageind));
 }

 JEMALLOC_ALWAYS_INLINE size_t
-arena_mapbitsp_read(size_t *mapbitsp)
+arena_mapbitsp_read(const size_t *mapbitsp)
 {

 	return (*mapbitsp);
 }

 JEMALLOC_ALWAYS_INLINE size_t
-arena_mapbits_get(arena_chunk_t *chunk, size_t pageind)
+arena_mapbits_get(const arena_chunk_t *chunk, size_t pageind)
 {

-	return (arena_mapbitsp_read(arena_mapbitsp_get(chunk, pageind)));
+	return (arena_mapbitsp_read(arena_mapbitsp_get_const(chunk, pageind)));
 }

 JEMALLOC_ALWAYS_INLINE size_t
@@ -763,7 +826,7 @@ arena_mapbits_size_decode(size_t mapbits)
 }

 JEMALLOC_ALWAYS_INLINE size_t
-arena_mapbits_unallocated_size_get(arena_chunk_t *chunk, size_t pageind)
+arena_mapbits_unallocated_size_get(const arena_chunk_t *chunk, size_t pageind)
 {
 	size_t mapbits;

@@ -773,7 +836,7 @@ arena_mapbits_unallocated_size_get(arena_chunk_t *chunk, size_t pageind)
 }

 JEMALLOC_ALWAYS_INLINE size_t
-arena_mapbits_large_size_get(arena_chunk_t *chunk, size_t pageind)
+arena_mapbits_large_size_get(const arena_chunk_t *chunk, size_t pageind)
 {
 	size_t mapbits;

@@ -784,7 +847,7 @@ arena_mapbits_large_size_get(arena_chunk_t *chunk, size_t pageind)
 }

 JEMALLOC_ALWAYS_INLINE size_t
-arena_mapbits_small_runind_get(arena_chunk_t *chunk, size_t pageind)
+arena_mapbits_small_runind_get(const arena_chunk_t *chunk, size_t pageind)
 {
 	size_t mapbits;

@@ -795,7 +858,7 @@ arena_mapbits_small_runind_get(arena_chunk_t *chunk, size_t pageind)
 }

 JEMALLOC_ALWAYS_INLINE szind_t
-arena_mapbits_binind_get(arena_chunk_t *chunk, size_t pageind)
+arena_mapbits_binind_get(const arena_chunk_t *chunk, size_t pageind)
 {
 	size_t mapbits;
 	szind_t binind;
@@ -807,7 +870,7 @@ arena_mapbits_binind_get(arena_chunk_t *chunk, size_t pageind)
 }

 JEMALLOC_ALWAYS_INLINE size_t
-arena_mapbits_dirty_get(arena_chunk_t *chunk, size_t pageind)
+arena_mapbits_dirty_get(const arena_chunk_t *chunk, size_t pageind)
 {
 	size_t mapbits;

@@ -818,7 +881,7 @@ arena_mapbits_dirty_get(arena_chunk_t *chunk, size_t pageind)
 }

 JEMALLOC_ALWAYS_INLINE size_t
-arena_mapbits_unzeroed_get(arena_chunk_t *chunk, size_t pageind)
+arena_mapbits_unzeroed_get(const arena_chunk_t *chunk, size_t pageind)
 {
 	size_t mapbits;

@@ -829,7 +892,7 @@ arena_mapbits_unzeroed_get(arena_chunk_t *chunk, size_t pageind)
 }

 JEMALLOC_ALWAYS_INLINE size_t
-arena_mapbits_decommitted_get(arena_chunk_t *chunk, size_t pageind)
+arena_mapbits_decommitted_get(const arena_chunk_t *chunk, size_t pageind)
 {
 	size_t mapbits;

@@ -840,7 +903,7 @@ arena_mapbits_decommitted_get(arena_chunk_t *chunk, size_t pageind)
 }

 JEMALLOC_ALWAYS_INLINE size_t
-arena_mapbits_large_get(arena_chunk_t *chunk, size_t pageind)
+arena_mapbits_large_get(const arena_chunk_t *chunk, size_t pageind)
 {
 	size_t mapbits;

@@ -849,7 +912,7 @@ arena_mapbits_large_get(arena_chunk_t *chunk, size_t pageind)
 }

 JEMALLOC_ALWAYS_INLINE size_t
-arena_mapbits_allocated_get(arena_chunk_t *chunk, size_t pageind)
+arena_mapbits_allocated_get(const arena_chunk_t *chunk, size_t pageind)
 {
 	size_t mapbits;

@@ -885,7 +948,7 @@ JEMALLOC_ALWAYS_INLINE void
 arena_mapbits_unallocated_set(arena_chunk_t *chunk, size_t pageind, size_t size,
    size_t flags)
 {
-	size_t *mapbitsp = arena_mapbitsp_get(chunk, pageind);
+	size_t *mapbitsp = arena_mapbitsp_get_mutable(chunk, pageind);

 	assert((size & PAGE_MASK) == 0);
 	assert((flags & CHUNK_MAP_FLAGS_MASK) == flags);
@@ -899,7 +962,7 @@ JEMALLOC_ALWAYS_INLINE void
 arena_mapbits_unallocated_size_set(arena_chunk_t *chunk, size_t pageind,
    size_t size)
 {
-	size_t *mapbitsp = arena_mapbitsp_get(chunk, pageind);
+	size_t *mapbitsp = arena_mapbitsp_get_mutable(chunk, pageind);
 	size_t mapbits = arena_mapbitsp_read(mapbitsp);

 	assert((size & PAGE_MASK) == 0);
@@ -911,7 +974,7 @@ arena_mapbits_unallocated_size_set(arena_chunk_t *chunk, size_t pageind,
 JEMALLOC_ALWAYS_INLINE void
 arena_mapbits_internal_set(arena_chunk_t *chunk, size_t pageind, size_t flags)
 {
-	size_t *mapbitsp = arena_mapbitsp_get(chunk, pageind);
+	size_t *mapbitsp = arena_mapbitsp_get_mutable(chunk, pageind);

 	assert((flags & CHUNK_MAP_UNZEROED) == flags);
 	arena_mapbitsp_write(mapbitsp, flags);
@@ -921,7 +984,7 @@ JEMALLOC_ALWAYS_INLINE void
 arena_mapbits_large_set(arena_chunk_t *chunk, size_t pageind, size_t size,
    size_t flags)
 {
-	size_t *mapbitsp = arena_mapbitsp_get(chunk, pageind);
+	size_t *mapbitsp = arena_mapbitsp_get_mutable(chunk, pageind);

 	assert((size & PAGE_MASK) == 0);
 	assert((flags & CHUNK_MAP_FLAGS_MASK) == flags);
@@ -936,7 +999,7 @@ JEMALLOC_ALWAYS_INLINE void
 arena_mapbits_large_binind_set(arena_chunk_t *chunk, size_t pageind,
    szind_t binind)
 {
-	size_t *mapbitsp = arena_mapbitsp_get(chunk, pageind);
+	size_t *mapbitsp = arena_mapbitsp_get_mutable(chunk, pageind);
 	size_t mapbits = arena_mapbitsp_read(mapbitsp);

 	assert(binind <= BININD_INVALID);
@@ -950,7 +1013,7 @@ JEMALLOC_ALWAYS_INLINE void
 arena_mapbits_small_set(arena_chunk_t *chunk, size_t pageind, size_t runind,
    szind_t binind, size_t flags)
 {
-	size_t *mapbitsp = arena_mapbitsp_get(chunk, pageind);
+	size_t *mapbitsp = arena_mapbitsp_get_mutable(chunk, pageind);

 	assert(binind < BININD_INVALID);
 	assert(pageind - runind >= map_bias);
@@ -1007,7 +1070,7 @@ arena_prof_accum_locked(arena_t *arena, uint64_t accumbytes)
 }

 JEMALLOC_INLINE bool
-arena_prof_accum(arena_t *arena, uint64_t accumbytes)
+arena_prof_accum(tsdn_t *tsdn, arena_t *arena, uint64_t accumbytes)
 {

 	cassert(config_prof);
@@ -1018,9 +1081,9 @@ arena_prof_accum(arena_t *arena, uint64_t accumbytes)
 	{
 		bool ret;

-		malloc_mutex_lock(&arena->lock);
+		malloc_mutex_lock(tsdn, &arena->lock);
 		ret = arena_prof_accum_impl(arena, accumbytes);
-		malloc_mutex_unlock(&arena->lock);
+		malloc_mutex_unlock(tsdn, &arena->lock);
 		return (ret);
 	}
 }
@@ -1038,12 +1101,12 @@ arena_ptr_small_binind_get(const void *ptr, size_t mapbits)
 		size_t pageind;
 		size_t actual_mapbits;
 		size_t rpages_ind;
-		arena_run_t *run;
+		const arena_run_t *run;
 		arena_bin_t *bin;
 		szind_t run_binind, actual_binind;
 		arena_bin_info_t *bin_info;
-		arena_chunk_map_misc_t *miscelm;
-		void *rpages;
+		const arena_chunk_map_misc_t *miscelm;
+		const void *rpages;

 		assert(binind != BININD_INVALID);
 		assert(binind < NBINS);
@@ -1056,7 +1119,7 @@ arena_ptr_small_binind_get(const void *ptr, size_t mapbits)
 		assert(arena_mapbits_allocated_get(chunk, pageind) != 0);
 		rpages_ind = pageind - arena_mapbits_small_runind_get(chunk,
 		    pageind);
-		miscelm = arena_miscelm_get(chunk, rpages_ind);
+		miscelm = arena_miscelm_get_const(chunk, rpages_ind);
 		run = &miscelm->run;
 		run_binind = run->binind;
 		bin = &arena->bins[run_binind];
@@ -1156,7 +1219,7 @@ arena_run_regind(arena_run_t *run, arena_bin_info_t *bin_info, const void *ptr)
 }

 JEMALLOC_INLINE prof_tctx_t *
-arena_prof_tctx_get(const void *ptr)
+arena_prof_tctx_get(tsdn_t *tsdn, const void *ptr)
 {
 	prof_tctx_t *ret;
 	arena_chunk_t *chunk;
@@ -1172,18 +1235,19 @@ arena_prof_tctx_get(const void *ptr)
 		if (likely((mapbits & CHUNK_MAP_LARGE) == 0))
 			ret = (prof_tctx_t *)(uintptr_t)1U;
 		else {
-			arena_chunk_map_misc_t *elm = arena_miscelm_get(chunk,
-			    pageind);
+			arena_chunk_map_misc_t *elm =
+			    arena_miscelm_get_mutable(chunk, pageind);
 			ret = atomic_read_p(&elm->prof_tctx_pun);
 		}
 	} else
-		ret = huge_prof_tctx_get(ptr);
+		ret = huge_prof_tctx_get(tsdn, ptr);

 	return (ret);
 }

 JEMALLOC_INLINE void
-arena_prof_tctx_set(const void *ptr, size_t usize, prof_tctx_t *tctx)
+arena_prof_tctx_set(tsdn_t *tsdn, const void *ptr, size_t usize,
+    prof_tctx_t *tctx)
 {
 	arena_chunk_t *chunk;

@@ -1202,7 +1266,7 @@ arena_prof_tctx_set(const void *ptr, size_t usize, prof_tctx_t *tctx)

 			assert(arena_mapbits_large_get(chunk, pageind) != 0);

-			elm = arena_miscelm_get(chunk, pageind);
+			elm = arena_miscelm_get_mutable(chunk, pageind);
 			atomic_write_p(&elm->prof_tctx_pun, tctx);
 		} else {
 			/*
@@ -1214,12 +1278,12 @@ arena_prof_tctx_set(const void *ptr, size_t usize, prof_tctx_t *tctx)
 			assert(arena_mapbits_large_get(chunk, pageind) == 0);
 		}
 	} else
-		huge_prof_tctx_set(ptr, tctx);
+		huge_prof_tctx_set(tsdn, ptr, tctx);
 }

 JEMALLOC_INLINE void
-arena_prof_tctx_reset(const void *ptr, size_t usize, const void *old_ptr,
-    prof_tctx_t *old_tctx)
+arena_prof_tctx_reset(tsdn_t *tsdn, const void *ptr, size_t usize,
+    const void *old_ptr, prof_tctx_t *old_tctx)
 {

 	cassert(config_prof);
@@ -1238,56 +1302,59 @@ arena_prof_tctx_reset(const void *ptr, size_t usize, const void *old_ptr,
 			    0);
 			assert(arena_mapbits_large_get(chunk, pageind) != 0);

-			elm = arena_miscelm_get(chunk, pageind);
+			elm = arena_miscelm_get_mutable(chunk, pageind);
 			atomic_write_p(&elm->prof_tctx_pun,
 			    (prof_tctx_t *)(uintptr_t)1U);
 		} else
-			huge_prof_tctx_reset(ptr);
+			huge_prof_tctx_reset(tsdn, ptr);
 	}
 }

 JEMALLOC_ALWAYS_INLINE void
-arena_decay_ticks(tsd_t *tsd, arena_t *arena, unsigned nticks)
+arena_decay_ticks(tsdn_t *tsdn, arena_t *arena, unsigned nticks)
 {
+	tsd_t *tsd;
 	ticker_t *decay_ticker;

-	if (unlikely(tsd == NULL))
+	if (unlikely(tsdn_null(tsdn)))
 		return;
+	tsd = tsdn_tsd(tsdn);
 	decay_ticker = decay_ticker_get(tsd, arena->ind);
 	if (unlikely(decay_ticker == NULL))
 		return;
 	if (unlikely(ticker_ticks(decay_ticker, nticks)))
-		arena_purge(arena, false);
+		arena_purge(tsdn, arena, false);
 }

 JEMALLOC_ALWAYS_INLINE void
-arena_decay_tick(tsd_t *tsd, arena_t *arena)
+arena_decay_tick(tsdn_t *tsdn, arena_t *arena)
 {

-	arena_decay_ticks(tsd, arena, 1);
+	arena_decay_ticks(tsdn, arena, 1);
 }

 JEMALLOC_ALWAYS_INLINE void *
-arena_malloc(tsd_t *tsd, arena_t *arena, size_t size, szind_t ind, bool zero,
+arena_malloc(tsdn_t *tsdn, arena_t *arena, size_t size, szind_t ind, bool zero,
    tcache_t *tcache, bool slow_path)
 {

+	assert(!tsdn_null(tsdn) || tcache == NULL);
 	assert(size != 0);

 	if (likely(tcache != NULL)) {
 		if (likely(size <= SMALL_MAXCLASS)) {
-			return (tcache_alloc_small(tsd, arena, tcache, size,
-			    ind, zero, slow_path));
+			return (tcache_alloc_small(tsdn_tsd(tsdn), arena,
+			    tcache, size, ind, zero, slow_path));
 		}
 		if (likely(size <= tcache_maxclass)) {
-			return (tcache_alloc_large(tsd, arena, tcache, size,
-			    ind, zero, slow_path));
+			return (tcache_alloc_large(tsdn_tsd(tsdn), arena,
+			    tcache, size, ind, zero, slow_path));
 		}
 		/* (size > tcache_maxclass) case falls through. */
 		assert(size > tcache_maxclass);
 	}

-	return (arena_malloc_hard(tsd, arena, size, ind, zero, tcache));
+	return (arena_malloc_hard(tsdn, arena, size, ind, zero));
 }

 JEMALLOC_ALWAYS_INLINE arena_t *
@@ -1304,7 +1371,7 @@ arena_aalloc(const void *ptr)

 /* Return the size of the allocation pointed to by ptr. */
 JEMALLOC_ALWAYS_INLINE size_t
-arena_salloc(const void *ptr, bool demote)
+arena_salloc(tsdn_t *tsdn, const void *ptr, bool demote)
 {
 	size_t ret;
 	arena_chunk_t *chunk;
@@ -1347,17 +1414,18 @@ arena_salloc(const void *ptr, bool demote)
 			ret = index2size(binind);
 		}
 	} else
-		ret = huge_salloc(ptr);
+		ret = huge_salloc(tsdn, ptr);

 	return (ret);
 }

 JEMALLOC_ALWAYS_INLINE void
-arena_dalloc(tsd_t *tsd, void *ptr, tcache_t *tcache, bool slow_path)
+arena_dalloc(tsdn_t *tsdn, void *ptr, tcache_t *tcache, bool slow_path)
 {
 	arena_chunk_t *chunk;
 	size_t pageind, mapbits;

+	assert(!tsdn_null(tsdn) || tcache == NULL);
 	assert(ptr != NULL);

 	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
@@ -1370,11 +1438,12 @@ arena_dalloc(tsd_t *tsd, void *ptr, tcache_t *tcache, bool slow_path)
 			if (likely(tcache != NULL)) {
 				szind_t binind = arena_ptr_small_binind_get(ptr,
 				    mapbits);
-				tcache_dalloc_small(tsd, tcache, ptr, binind,
-				    slow_path);
+				tcache_dalloc_small(tsdn_tsd(tsdn), tcache, ptr,
+				    binind, slow_path);
 			} else {
-				arena_dalloc_small(tsd, extent_node_arena_get(
-				    &chunk->node), chunk, ptr, pageind);
+				arena_dalloc_small(tsdn,
+				    extent_node_arena_get(&chunk->node), chunk,
+				    ptr, pageind);
 			}
 		} else {
 			size_t size = arena_mapbits_large_size_get(chunk,
@@ -1385,22 +1454,26 @@ arena_dalloc(tsd_t *tsd, void *ptr, tcache_t *tcache, bool slow_path)

 			if (likely(tcache != NULL) && size - large_pad <=
 			    tcache_maxclass) {
-				tcache_dalloc_large(tsd, tcache, ptr, size -
-				    large_pad, slow_path);
+				tcache_dalloc_large(tsdn_tsd(tsdn), tcache, ptr,
+				    size - large_pad, slow_path);
 			} else {
-				arena_dalloc_large(tsd, extent_node_arena_get(
-				    &chunk->node), chunk, ptr);
+				arena_dalloc_large(tsdn,
+				    extent_node_arena_get(&chunk->node), chunk,
+				    ptr);
 			}
 		}
 	} else
-		huge_dalloc(tsd, ptr, tcache);
+		huge_dalloc(tsdn, ptr);
 }

 JEMALLOC_ALWAYS_INLINE void
-arena_sdalloc(tsd_t *tsd, void *ptr, size_t size, tcache_t *tcache)
+arena_sdalloc(tsdn_t *tsdn, void *ptr, size_t size, tcache_t *tcache,
+    bool slow_path)
 {
 	arena_chunk_t *chunk;

+	assert(!tsdn_null(tsdn) || tcache == NULL);
+
 	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
 	if (likely(chunk != ptr)) {
 		if (config_prof && opt_prof) {
@@ -1417,34 +1490,36 @@ arena_sdalloc(tsd_t *tsd, void *ptr, size_t size, tcache_t *tcache)
 				    pageind) - large_pad;
 			}
 		}
-		assert(s2u(size) == s2u(arena_salloc(ptr, false)));
+		assert(s2u(size) == s2u(arena_salloc(tsdn, ptr, false)));

 		if (likely(size <= SMALL_MAXCLASS)) {
 			/* Small allocation. */
 			if (likely(tcache != NULL)) {
 				szind_t binind = size2index(size);
-				tcache_dalloc_small(tsd, tcache, ptr, binind,
-				    true);
+				tcache_dalloc_small(tsdn_tsd(tsdn), tcache, ptr,
+				    binind, slow_path);
 			} else {
 				size_t pageind = ((uintptr_t)ptr -
 				    (uintptr_t)chunk) >> LG_PAGE;
-				arena_dalloc_small(tsd, extent_node_arena_get(
-				    &chunk->node), chunk, ptr, pageind);
+				arena_dalloc_small(tsdn,
+				    extent_node_arena_get(&chunk->node), chunk,
+				    ptr, pageind);
 			}
 		} else {
 			assert(config_cache_oblivious || ((uintptr_t)ptr &
 			    PAGE_MASK) == 0);

 			if (likely(tcache != NULL) && size <= tcache_maxclass) {
-				tcache_dalloc_large(tsd, tcache, ptr, size,
-				    true);
+				tcache_dalloc_large(tsdn_tsd(tsdn), tcache, ptr,
+				    size, slow_path);
 			} else {
-				arena_dalloc_large(tsd, extent_node_arena_get(
-				    &chunk->node), chunk, ptr);
+				arena_dalloc_large(tsdn,
+				    extent_node_arena_get(&chunk->node), chunk,
+				    ptr);
 			}
 		}
 	} else
-		huge_dalloc(tsd, ptr, tcache);
+		huge_dalloc(tsdn, ptr);
 }
 #  endif /* JEMALLOC_ARENA_INLINE_B */
 #endif
--- a/include/jemalloc/internal/base.h
+++ b/include/jemalloc/internal/base.h
@@ -9,12 +9,13 @@
 /******************************************************************************/
 #ifdef JEMALLOC_H_EXTERNS

-void	*base_alloc(size_t size);
-void	base_stats_get(size_t *allocated, size_t *resident, size_t *mapped);
+void	*base_alloc(tsdn_t *tsdn, size_t size);
+void	base_stats_get(tsdn_t *tsdn, size_t *allocated, size_t *resident,
+    size_t *mapped);
 bool	base_boot(void);
-void	base_prefork(void);
-void	base_postfork_parent(void);
-void	base_postfork_child(void);
+void	base_prefork(tsdn_t *tsdn);
+void	base_postfork_parent(tsdn_t *tsdn);
+void	base_postfork_child(tsdn_t *tsdn);

 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
--- a/include/jemalloc/internal/bitmap.h
+++ b/include/jemalloc/internal/bitmap.h
@@ -17,8 +17,8 @@ typedef unsigned long bitmap_t;

 /*
 * Do some analysis on how big the bitmap is before we use a tree.  For a brute
- * force linear search, if we would have to call ffsl more than 2^3 times, use a
- * tree instead.
+ * force linear search, if we would have to call ffs_lu() more than 2^3 times,
+ * use a tree instead.
 */
 #if LG_BITMAP_MAXBITS - LG_BITMAP_GROUP_NBITS > 3
 #  define USE_TREE
@@ -223,7 +223,7 @@ bitmap_sfu(bitmap_t *bitmap, const bitmap_info_t *binfo)
 		i++;
 		g = bitmap[i];
 	}
-	bit = (bit - 1) + (i << LG_BITMAP_GROUP_NBITS);
+	bit = (i << LG_BITMAP_GROUP_NBITS) + (bit - 1);
 #endif
 	bitmap_set(bitmap, binfo, bit);
 	return (bit);
--- a/include/jemalloc/internal/chunk.h
+++ b/include/jemalloc/internal/chunk.h
@@ -48,28 +48,30 @@ extern size_t		chunk_npages;

 extern const chunk_hooks_t	chunk_hooks_default;

-chunk_hooks_t	chunk_hooks_get(arena_t *arena);
-chunk_hooks_t	chunk_hooks_set(arena_t *arena,
+chunk_hooks_t	chunk_hooks_get(tsdn_t *tsdn, arena_t *arena);
+chunk_hooks_t	chunk_hooks_set(tsdn_t *tsdn, arena_t *arena,
    const chunk_hooks_t *chunk_hooks);

-bool	chunk_register(const void *chunk, const extent_node_t *node);
+bool	chunk_register(const void *chunk, const extent_node_t *node,
+    bool *gdump);
 void	chunk_deregister(const void *chunk, const extent_node_t *node);
 void	*chunk_alloc_base(size_t size);
-void	*chunk_alloc_cache(arena_t *arena, chunk_hooks_t *chunk_hooks,
-    void *new_addr, size_t size, size_t alignment, bool *zero,
-    bool dalloc_node);
-void	*chunk_alloc_wrapper(arena_t *arena, chunk_hooks_t *chunk_hooks,
-    void *new_addr, size_t size, size_t alignment, bool *zero, bool *commit);
-void	chunk_dalloc_cache(arena_t *arena, chunk_hooks_t *chunk_hooks,
-    void *chunk, size_t size, bool committed);
-void	chunk_dalloc_wrapper(arena_t *arena, chunk_hooks_t *chunk_hooks,
-    void *chunk, size_t size, bool zeroed, bool committed);
-bool	chunk_purge_wrapper(arena_t *arena, chunk_hooks_t *chunk_hooks,
-    void *chunk, size_t size, size_t offset, size_t length);
+void	*chunk_alloc_cache(tsdn_t *tsdn, arena_t *arena,
+    chunk_hooks_t *chunk_hooks, void *new_addr, size_t size, size_t alignment,
+    size_t *sn, bool *zero, bool *commit, bool dalloc_node);
+void	*chunk_alloc_wrapper(tsdn_t *tsdn, arena_t *arena,
+    chunk_hooks_t *chunk_hooks, void *new_addr, size_t size, size_t alignment,
+    size_t *sn, bool *zero, bool *commit);
+void	chunk_dalloc_cache(tsdn_t *tsdn, arena_t *arena,
+    chunk_hooks_t *chunk_hooks, void *chunk, size_t size, size_t sn,
+    bool committed);
+void	chunk_dalloc_wrapper(tsdn_t *tsdn, arena_t *arena,
+    chunk_hooks_t *chunk_hooks, void *chunk, size_t size, size_t sn,
+    bool zeroed, bool committed);
+bool	chunk_purge_wrapper(tsdn_t *tsdn, arena_t *arena,
+    chunk_hooks_t *chunk_hooks, void *chunk, size_t size, size_t offset,
+    size_t length);
 bool	chunk_boot(void);
-void	chunk_prefork(void);
-void	chunk_postfork_parent(void);
-void	chunk_postfork_child(void);

 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
--- a/include/jemalloc/internal/chunk_dss.h
+++ b/include/jemalloc/internal/chunk_dss.h
@@ -23,13 +23,11 @@ extern const char *dss_prec_names[];

 dss_prec_t	chunk_dss_prec_get(void);
 bool	chunk_dss_prec_set(dss_prec_t dss_prec);
-void	*chunk_alloc_dss(arena_t *arena, void *new_addr, size_t size,
-    size_t alignment, bool *zero, bool *commit);
+void	*chunk_alloc_dss(tsdn_t *tsdn, arena_t *arena, void *new_addr,
+    size_t size, size_t alignment, bool *zero, bool *commit);
 bool	chunk_in_dss(void *chunk);
-bool	chunk_dss_boot(void);
-void	chunk_dss_prefork(void);
-void	chunk_dss_postfork_parent(void);
-void	chunk_dss_postfork_child(void);
+bool	chunk_dss_mergeable(void *chunk_a, void *chunk_b);
+void	chunk_dss_boot(void);

 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
--- a/include/jemalloc/internal/ctl.h
+++ b/include/jemalloc/internal/ctl.h
@@ -21,13 +21,14 @@ struct ctl_named_node_s {
 	/* 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);
+	int			(*ctl)(tsd_t *, 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);
+	const ctl_named_node_t	*(*index)(tsdn_t *, const size_t *, size_t,
+	    size_t);
 };

 struct ctl_arena_stats_s {
@@ -60,6 +61,7 @@ struct ctl_stats_s {
 	size_t			metadata;
 	size_t			resident;
 	size_t			mapped;
+	size_t			retained;
 	unsigned		narenas;
 	ctl_arena_stats_t	*arenas;	/* (narenas + 1) elements. */
 };
@@ -68,16 +70,17 @@ struct ctl_stats_s {
 /******************************************************************************/
 #ifdef JEMALLOC_H_EXTERNS

-int	ctl_byname(const char *name, void *oldp, size_t *oldlenp, void *newp,
-    size_t newlen);
-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,
+int	ctl_byname(tsd_t *tsd, const char *name, void *oldp, size_t *oldlenp,
    void *newp, size_t newlen);
+int	ctl_nametomib(tsdn_t *tsdn, const char *name, size_t *mibp,
+    size_t *miblenp);
+
+int	ctl_bymib(tsd_t *tsd, 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);
+void	ctl_prefork(tsdn_t *tsdn);
+void	ctl_postfork_parent(tsdn_t *tsdn);
+void	ctl_postfork_child(tsdn_t *tsdn);

 #define	xmallctl(name, oldp, oldlenp, newp, newlen) do {		\
 	if (je_mallctl(name, oldp, oldlenp, newp, newlen)		\
--- a/include/jemalloc/internal/extent.h
+++ b/include/jemalloc/internal/extent.h
@@ -18,6 +18,20 @@ struct extent_node_s {
 	/* Total region size. */
 	size_t			en_size;

+	/*
+	 * Serial number (potentially non-unique).
+	 *
+	 * In principle serial numbers can wrap around on 32-bit systems if
+	 * JEMALLOC_MUNMAP is defined, but as long as comparison functions fall
+	 * back on address comparison for equal serial numbers, stable (if
+	 * imperfect) ordering is maintained.
+	 *
+	 * Serial numbers may not be unique even in the absence of wrap-around,
+	 * e.g. when splitting an extent and assigning the same serial number to
+	 * both resulting adjacent extents.
+	 */
+	size_t			en_sn;
+
 	/*
 	 * The zeroed flag is used by chunk recycling code to track whether
 	 * memory is zero-filled.
@@ -45,10 +59,10 @@ struct extent_node_s {
 	qr(extent_node_t)	cc_link;

 	union {
-		/* Linkage for the size/address-ordered tree. */
-		rb_node(extent_node_t)	szad_link;
+		/* Linkage for the size/sn/address-ordered tree. */
+		rb_node(extent_node_t)	szsnad_link;

-		/* Linkage for arena's huge and node_cache lists. */
+		/* Linkage for arena's achunks, huge, and node_cache lists. */
 		ql_elm(extent_node_t)	ql_link;
 	};

@@ -61,7 +75,12 @@ typedef rb_tree(extent_node_t) extent_tree_t;
 /******************************************************************************/
 #ifdef JEMALLOC_H_EXTERNS

-rb_proto(, extent_tree_szad_, extent_tree_t, extent_node_t)
+#ifdef JEMALLOC_JET
+size_t	extent_size_quantize_floor(size_t size);
+#endif
+size_t	extent_size_quantize_ceil(size_t size);
+
+rb_proto(, extent_tree_szsnad_, extent_tree_t, extent_node_t)

 rb_proto(, extent_tree_ad_, extent_tree_t, extent_node_t)

@@ -73,6 +92,7 @@ rb_proto(, extent_tree_ad_, extent_tree_t, extent_node_t)
 arena_t	*extent_node_arena_get(const extent_node_t *node);
 void	*extent_node_addr_get(const extent_node_t *node);
 size_t	extent_node_size_get(const extent_node_t *node);
+size_t	extent_node_sn_get(const extent_node_t *node);
 bool	extent_node_zeroed_get(const extent_node_t *node);
 bool	extent_node_committed_get(const extent_node_t *node);
 bool	extent_node_achunk_get(const extent_node_t *node);
@@ -80,12 +100,13 @@ prof_tctx_t	*extent_node_prof_tctx_get(const extent_node_t *node);
 void	extent_node_arena_set(extent_node_t *node, arena_t *arena);
 void	extent_node_addr_set(extent_node_t *node, void *addr);
 void	extent_node_size_set(extent_node_t *node, size_t size);
+void	extent_node_sn_set(extent_node_t *node, size_t sn);
 void	extent_node_zeroed_set(extent_node_t *node, bool zeroed);
 void	extent_node_committed_set(extent_node_t *node, bool committed);
 void	extent_node_achunk_set(extent_node_t *node, bool achunk);
 void	extent_node_prof_tctx_set(extent_node_t *node, prof_tctx_t *tctx);
 void	extent_node_init(extent_node_t *node, arena_t *arena, void *addr,
-    size_t size, bool zeroed, bool committed);
+    size_t size, size_t sn, bool zeroed, bool committed);
 void	extent_node_dirty_linkage_init(extent_node_t *node);
 void	extent_node_dirty_insert(extent_node_t *node,
    arena_runs_dirty_link_t *runs_dirty, extent_node_t *chunks_dirty);
@@ -114,6 +135,13 @@ extent_node_size_get(const extent_node_t *node)
 	return (node->en_size);
 }

+JEMALLOC_INLINE size_t
+extent_node_sn_get(const extent_node_t *node)
+{
+
+	return (node->en_sn);
+}
+
 JEMALLOC_INLINE bool
 extent_node_zeroed_get(const extent_node_t *node)
 {
@@ -164,6 +192,13 @@ extent_node_size_set(extent_node_t *node, size_t size)
 	node->en_size = size;
 }

+JEMALLOC_INLINE void
+extent_node_sn_set(extent_node_t *node, size_t sn)
+{
+
+	node->en_sn = sn;
+}
+
 JEMALLOC_INLINE void
 extent_node_zeroed_set(extent_node_t *node, bool zeroed)
 {
@@ -194,12 +229,13 @@ extent_node_prof_tctx_set(extent_node_t *node, prof_tctx_t *tctx)

 JEMALLOC_INLINE void
 extent_node_init(extent_node_t *node, arena_t *arena, void *addr, size_t size,
-    bool zeroed, bool committed)
+    size_t sn, bool zeroed, bool committed)
 {

 	extent_node_arena_set(node, arena);
 	extent_node_addr_set(node, addr);
 	extent_node_size_set(node, size);
+	extent_node_sn_set(node, sn);
 	extent_node_zeroed_set(node, zeroed);
 	extent_node_committed_set(node, committed);
 	extent_node_achunk_set(node, false);
--- a/include/jemalloc/internal/huge.h
+++ b/include/jemalloc/internal/huge.h
@@ -9,11 +9,10 @@
 /******************************************************************************/
 #ifdef JEMALLOC_H_EXTERNS

-void	*huge_malloc(tsd_t *tsd, arena_t *arena, size_t usize, bool zero,
-    tcache_t *tcache);
-void	*huge_palloc(tsd_t *tsd, arena_t *arena, size_t usize, size_t alignment,
-    bool zero, tcache_t *tcache);
-bool	huge_ralloc_no_move(tsd_t *tsd, void *ptr, size_t oldsize,
+void	*huge_malloc(tsdn_t *tsdn, arena_t *arena, size_t usize, bool zero);
+void	*huge_palloc(tsdn_t *tsdn, arena_t *arena, size_t usize,
+    size_t alignment, bool zero);
+bool	huge_ralloc_no_move(tsdn_t *tsdn, void *ptr, size_t oldsize,
    size_t usize_min, size_t usize_max, bool zero);
 void	*huge_ralloc(tsd_t *tsd, arena_t *arena, void *ptr, size_t oldsize,
    size_t usize, size_t alignment, bool zero, tcache_t *tcache);
@@ -21,12 +20,12 @@ void	*huge_ralloc(tsd_t *tsd, arena_t *arena, void *ptr, size_t oldsize,
 typedef void (huge_dalloc_junk_t)(void *, size_t);
 extern huge_dalloc_junk_t *huge_dalloc_junk;
 #endif
-void	huge_dalloc(tsd_t *tsd, void *ptr, tcache_t *tcache);
+void	huge_dalloc(tsdn_t *tsdn, void *ptr);
 arena_t	*huge_aalloc(const void *ptr);
-size_t	huge_salloc(const void *ptr);
-prof_tctx_t	*huge_prof_tctx_get(const void *ptr);
-void	huge_prof_tctx_set(const void *ptr, prof_tctx_t *tctx);
-void	huge_prof_tctx_reset(const void *ptr);
+size_t	huge_salloc(tsdn_t *tsdn, const void *ptr);
+prof_tctx_t	*huge_prof_tctx_get(tsdn_t *tsdn, const void *ptr);
+void	huge_prof_tctx_set(tsdn_t *tsdn, const void *ptr, prof_tctx_t *tctx);
+void	huge_prof_tctx_reset(tsdn_t *tsdn, const void *ptr);

 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
--- a/include/jemalloc/internal/jemalloc_internal.h.in
+++ b/include/jemalloc/internal/jemalloc_internal.h.in
@@ -99,6 +99,13 @@ static const bool config_tcache =
    false
 #endif
    ;
+static const bool config_thp =
+#ifdef JEMALLOC_THP
+    true
+#else
+    false
+#endif
+    ;
 static const bool config_tls =
 #ifdef JEMALLOC_TLS
    true
@@ -158,10 +165,12 @@ static const bool config_cache_oblivious =
 #include <mach/mach_error.h>
 #include <mach/mach_init.h>
 #include <mach/vm_map.h>
-#include <malloc/malloc.h>
 #endif

+#include "jemalloc/internal/ph.h"
+#ifndef __PGI
 #define	RB_COMPACT
+#endif
 #include "jemalloc/internal/rb.h"
 #include "jemalloc/internal/qr.h"
 #include "jemalloc/internal/ql.h"
@@ -184,6 +193,9 @@ static const bool config_cache_oblivious =

 #include "jemalloc/internal/jemalloc_internal_macros.h"

+/* Page size index type. */
+typedef unsigned pszind_t;
+
 /* Size class index type. */
 typedef unsigned szind_t;

@@ -233,7 +245,7 @@ typedef unsigned szind_t;
 #  ifdef __alpha__
 #    define LG_QUANTUM		4
 #  endif
-#  if (defined(__sparc64__) || defined(__sparcv9))
+#  if (defined(__sparc64__) || defined(__sparcv9) || defined(__sparc_v9__))
 #    define LG_QUANTUM		4
 #  endif
 #  if (defined(__amd64__) || defined(__x86_64__) || defined(_M_X64))
@@ -257,6 +269,9 @@ typedef unsigned szind_t;
 #  ifdef __powerpc__
 #    define LG_QUANTUM		4
 #  endif
+#  ifdef __riscv__
+#    define LG_QUANTUM		4
+#  endif
 #  ifdef __s390__
 #    define LG_QUANTUM		4
 #  endif
@@ -328,7 +343,7 @@ typedef unsigned szind_t;

 /* Return the nearest aligned address at or below a. */
 #define	ALIGNMENT_ADDR2BASE(a, alignment)				\
-	((void *)((uintptr_t)(a) & (-(alignment))))
+	((void *)((uintptr_t)(a) & ((~(alignment)) + 1)))

 /* Return the offset between a and the nearest aligned address at or below a. */
 #define	ALIGNMENT_ADDR2OFFSET(a, alignment)				\
@@ -336,7 +351,7 @@ typedef unsigned szind_t;

 /* Return the smallest alignment multiple that is >= s. */
 #define	ALIGNMENT_CEILING(s, alignment)					\
-	(((s) + (alignment - 1)) & (-(alignment)))
+	(((s) + (alignment - 1)) & ((~(alignment)) + 1))

 /* Declare a variable-length array. */
 #if __STDC_VERSION__ < 199901L
@@ -360,6 +375,7 @@ typedef unsigned szind_t;
 #include "jemalloc/internal/valgrind.h"
 #include "jemalloc/internal/util.h"
 #include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/spin.h"
 #include "jemalloc/internal/prng.h"
 #include "jemalloc/internal/ticker.h"
 #include "jemalloc/internal/ckh.h"
@@ -367,6 +383,7 @@ typedef unsigned szind_t;
 #include "jemalloc/internal/smoothstep.h"
 #include "jemalloc/internal/stats.h"
 #include "jemalloc/internal/ctl.h"
+#include "jemalloc/internal/witness.h"
 #include "jemalloc/internal/mutex.h"
 #include "jemalloc/internal/tsd.h"
 #include "jemalloc/internal/mb.h"
@@ -391,6 +408,7 @@ typedef unsigned szind_t;
 #include "jemalloc/internal/valgrind.h"
 #include "jemalloc/internal/util.h"
 #include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/spin.h"
 #include "jemalloc/internal/prng.h"
 #include "jemalloc/internal/ticker.h"
 #include "jemalloc/internal/ckh.h"
@@ -398,6 +416,7 @@ typedef unsigned szind_t;
 #include "jemalloc/internal/smoothstep.h"
 #include "jemalloc/internal/stats.h"
 #include "jemalloc/internal/ctl.h"
+#include "jemalloc/internal/witness.h"
 #include "jemalloc/internal/mutex.h"
 #include "jemalloc/internal/mb.h"
 #include "jemalloc/internal/bitmap.h"
@@ -440,17 +459,25 @@ extern bool	in_valgrind;
 /* Number of CPUs. */
 extern unsigned	ncpus;

+/* Number of arenas used for automatic multiplexing of threads and arenas. */
+extern unsigned	narenas_auto;
+
 /*
 * 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;

+/*
+ * pind2sz_tab encodes the same information as could be computed by
+ * pind2sz_compute().
+ */
+extern size_t const	pind2sz_tab[NPSIZES];
 /*
 * index2size_tab encodes the same information as could be computed (at
 * unacceptable cost in some code paths) by index2size_compute().
 */
-extern size_t const	index2size_tab[NSIZES+1];
+extern size_t const	index2size_tab[NSIZES];
 /*
 * size2index_tab is a compact lookup table that rounds request sizes up to
 * size classes.  In order to reduce cache footprint, the table is compressed,
@@ -458,19 +485,20 @@ extern size_t const	index2size_tab[NSIZES+1];
 */
 extern uint8_t const	size2index_tab[];

+arena_t	*a0get(void);
 void	*a0malloc(size_t size);
 void	a0dalloc(void *ptr);
 void	*bootstrap_malloc(size_t size);
 void	*bootstrap_calloc(size_t num, size_t size);
 void	bootstrap_free(void *ptr);
-arena_t	*arenas_extend(unsigned ind);
 unsigned	narenas_total_get(void);
-arena_t	*arena_init(unsigned ind);
+arena_t	*arena_init(tsdn_t *tsdn, unsigned ind);
 arena_tdata_t	*arena_tdata_get_hard(tsd_t *tsd, unsigned ind);
-arena_t	*arena_choose_hard(tsd_t *tsd);
+arena_t	*arena_choose_hard(tsd_t *tsd, bool internal);
 void	arena_migrate(tsd_t *tsd, unsigned oldind, unsigned newind);
 void	thread_allocated_cleanup(tsd_t *tsd);
 void	thread_deallocated_cleanup(tsd_t *tsd);
+void	iarena_cleanup(tsd_t *tsd);
 void	arena_cleanup(tsd_t *tsd);
 void	arenas_tdata_cleanup(tsd_t *tsd);
 void	narenas_tdata_cleanup(tsd_t *tsd);
@@ -483,6 +511,7 @@ void	jemalloc_postfork_child(void);
 #include "jemalloc/internal/valgrind.h"
 #include "jemalloc/internal/util.h"
 #include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/spin.h"
 #include "jemalloc/internal/prng.h"
 #include "jemalloc/internal/ticker.h"
 #include "jemalloc/internal/ckh.h"
@@ -490,6 +519,7 @@ void	jemalloc_postfork_child(void);
 #include "jemalloc/internal/smoothstep.h"
 #include "jemalloc/internal/stats.h"
 #include "jemalloc/internal/ctl.h"
+#include "jemalloc/internal/witness.h"
 #include "jemalloc/internal/mutex.h"
 #include "jemalloc/internal/mb.h"
 #include "jemalloc/internal/bitmap.h"
@@ -514,6 +544,7 @@ void	jemalloc_postfork_child(void);
 #include "jemalloc/internal/valgrind.h"
 #include "jemalloc/internal/util.h"
 #include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/spin.h"
 #include "jemalloc/internal/prng.h"
 #include "jemalloc/internal/ticker.h"
 #include "jemalloc/internal/ckh.h"
@@ -521,8 +552,9 @@ void	jemalloc_postfork_child(void);
 #include "jemalloc/internal/smoothstep.h"
 #include "jemalloc/internal/stats.h"
 #include "jemalloc/internal/ctl.h"
-#include "jemalloc/internal/mutex.h"
 #include "jemalloc/internal/tsd.h"
+#include "jemalloc/internal/witness.h"
+#include "jemalloc/internal/mutex.h"
 #include "jemalloc/internal/mb.h"
 #include "jemalloc/internal/extent.h"
 #include "jemalloc/internal/base.h"
@@ -532,6 +564,11 @@ void	jemalloc_postfork_child(void);
 #include "jemalloc/internal/huge.h"

 #ifndef JEMALLOC_ENABLE_INLINE
+pszind_t	psz2ind(size_t psz);
+size_t	pind2sz_compute(pszind_t pind);
+size_t	pind2sz_lookup(pszind_t pind);
+size_t	pind2sz(pszind_t pind);
+size_t	psz2u(size_t psz);
 szind_t	size2index_compute(size_t size);
 szind_t	size2index_lookup(size_t size);
 szind_t	size2index(size_t size);
@@ -542,18 +579,100 @@ size_t	s2u_compute(size_t size);
 size_t	s2u_lookup(size_t size);
 size_t	s2u(size_t size);
 size_t	sa2u(size_t size, size_t alignment);
+arena_t	*arena_choose_impl(tsd_t *tsd, arena_t *arena, bool internal);
 arena_t	*arena_choose(tsd_t *tsd, arena_t *arena);
+arena_t	*arena_ichoose(tsd_t *tsd, arena_t *arena);
 arena_tdata_t	*arena_tdata_get(tsd_t *tsd, unsigned ind,
    bool refresh_if_missing);
-arena_t	*arena_get(unsigned ind, bool init_if_missing);
+arena_t	*arena_get(tsdn_t *tsdn, unsigned ind, bool init_if_missing);
 ticker_t	*decay_ticker_get(tsd_t *tsd, unsigned ind);
 #endif

 #if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_C_))
+JEMALLOC_INLINE pszind_t
+psz2ind(size_t psz)
+{
+
+	if (unlikely(psz > HUGE_MAXCLASS))
+		return (NPSIZES);
+	{
+		pszind_t x = lg_floor((psz<<1)-1);
+		pszind_t shift = (x < LG_SIZE_CLASS_GROUP + LG_PAGE) ? 0 : x -
+		    (LG_SIZE_CLASS_GROUP + LG_PAGE);
+		pszind_t grp = shift << LG_SIZE_CLASS_GROUP;
+
+		pszind_t lg_delta = (x < LG_SIZE_CLASS_GROUP + LG_PAGE + 1) ?
+		    LG_PAGE : x - LG_SIZE_CLASS_GROUP - 1;
+
+		size_t delta_inverse_mask = ZI(-1) << lg_delta;
+		pszind_t mod = ((((psz-1) & delta_inverse_mask) >> lg_delta)) &
+		    ((ZU(1) << LG_SIZE_CLASS_GROUP) - 1);
+
+		pszind_t ind = grp + mod;
+		return (ind);
+	}
+}
+
+JEMALLOC_INLINE size_t
+pind2sz_compute(pszind_t pind)
+{
+
+	{
+		size_t grp = pind >> LG_SIZE_CLASS_GROUP;
+		size_t mod = pind & ((ZU(1) << LG_SIZE_CLASS_GROUP) - 1);
+
+		size_t grp_size_mask = ~((!!grp)-1);
+		size_t grp_size = ((ZU(1) << (LG_PAGE +
+		    (LG_SIZE_CLASS_GROUP-1))) << grp) & grp_size_mask;
+
+		size_t shift = (grp == 0) ? 1 : grp;
+		size_t lg_delta = shift + (LG_PAGE-1);
+		size_t mod_size = (mod+1) << lg_delta;
+
+		size_t sz = grp_size + mod_size;
+		return (sz);
+	}
+}
+
+JEMALLOC_INLINE size_t
+pind2sz_lookup(pszind_t pind)
+{
+	size_t ret = (size_t)pind2sz_tab[pind];
+	assert(ret == pind2sz_compute(pind));
+	return (ret);
+}
+
+JEMALLOC_INLINE size_t
+pind2sz(pszind_t pind)
+{
+
+	assert(pind < NPSIZES);
+	return (pind2sz_lookup(pind));
+}
+
+JEMALLOC_INLINE size_t
+psz2u(size_t psz)
+{
+
+	if (unlikely(psz > HUGE_MAXCLASS))
+		return (0);
+	{
+		size_t x = lg_floor((psz<<1)-1);
+		size_t lg_delta = (x < LG_SIZE_CLASS_GROUP + LG_PAGE + 1) ?
+		    LG_PAGE : x - LG_SIZE_CLASS_GROUP - 1;
+		size_t delta = ZU(1) << lg_delta;
+		size_t delta_mask = delta - 1;
+		size_t usize = (psz + delta_mask) & ~delta_mask;
+		return (usize);
+	}
+}
+
 JEMALLOC_INLINE szind_t
 size2index_compute(size_t size)
 {

+	if (unlikely(size > HUGE_MAXCLASS))
+		return (NSIZES);
 #if (NTBINS != 0)
 	if (size <= (ZU(1) << LG_TINY_MAXCLASS)) {
 		szind_t lg_tmin = LG_TINY_MAXCLASS - NTBINS + 1;
@@ -562,9 +681,7 @@ size2index_compute(size_t size)
 	}
 #endif
 	{
-		szind_t x = unlikely(ZI(size) < 0) ? ((size<<1) ?
-		    (ZU(1)<<(LG_SIZEOF_PTR+3)) : ((ZU(1)<<(LG_SIZEOF_PTR+3))-1))
-		    : lg_floor((size<<1)-1);
+		szind_t x = lg_floor((size<<1)-1);
 		szind_t shift = (x < LG_SIZE_CLASS_GROUP + LG_QUANTUM) ? 0 :
 		    x - (LG_SIZE_CLASS_GROUP + LG_QUANTUM);
 		szind_t grp = shift << LG_SIZE_CLASS_GROUP;
@@ -650,6 +767,8 @@ JEMALLOC_ALWAYS_INLINE size_t
 s2u_compute(size_t size)
 {

+	if (unlikely(size > HUGE_MAXCLASS))
+		return (0);
 #if (NTBINS > 0)
 	if (size <= (ZU(1) << LG_TINY_MAXCLASS)) {
 		size_t lg_tmin = LG_TINY_MAXCLASS - NTBINS + 1;
@@ -659,9 +778,7 @@ s2u_compute(size_t size)
 	}
 #endif
 	{
-		size_t x = unlikely(ZI(size) < 0) ? ((size<<1) ?
-		    (ZU(1)<<(LG_SIZEOF_PTR+3)) : ((ZU(1)<<(LG_SIZEOF_PTR+3))-1))
-		    : lg_floor((size<<1)-1);
+		size_t x = lg_floor((size<<1)-1);
 		size_t lg_delta = (x < LG_SIZE_CLASS_GROUP + LG_QUANTUM + 1)
 		    ?  LG_QUANTUM : x - LG_SIZE_CLASS_GROUP - 1;
 		size_t delta = ZU(1) << lg_delta;
@@ -780,19 +897,34 @@ sa2u(size_t size, size_t alignment)

 /* Choose an arena based on a per-thread value. */
 JEMALLOC_INLINE arena_t *
-arena_choose(tsd_t *tsd, arena_t *arena)
+arena_choose_impl(tsd_t *tsd, arena_t *arena, bool internal)
 {
 	arena_t *ret;

 	if (arena != NULL)
 		return (arena);

-	if (unlikely((ret = tsd_arena_get(tsd)) == NULL))
-		ret = arena_choose_hard(tsd);
+	ret = internal ? tsd_iarena_get(tsd) : tsd_arena_get(tsd);
+	if (unlikely(ret == NULL))
+		ret = arena_choose_hard(tsd, internal);

 	return (ret);
 }

+JEMALLOC_INLINE arena_t *
+arena_choose(tsd_t *tsd, arena_t *arena)
+{
+
+	return (arena_choose_impl(tsd, arena, false));
+}
+
+JEMALLOC_INLINE arena_t *
+arena_ichoose(tsd_t *tsd, arena_t *arena)
+{
+
+	return (arena_choose_impl(tsd, arena, true));
+}
+
 JEMALLOC_INLINE arena_tdata_t *
 arena_tdata_get(tsd_t *tsd, unsigned ind, bool refresh_if_missing)
 {
@@ -819,7 +951,7 @@ arena_tdata_get(tsd_t *tsd, unsigned ind, bool refresh_if_missing)
 }

 JEMALLOC_INLINE arena_t *
-arena_get(unsigned ind, bool init_if_missing)
+arena_get(tsdn_t *tsdn, unsigned ind, bool init_if_missing)
 {
 	arena_t *ret;

@@ -829,7 +961,7 @@ arena_get(unsigned ind, bool init_if_missing)
 	if (unlikely(ret == NULL)) {
 		ret = atomic_read_p((void *)&arenas[ind]);
 		if (init_if_missing && unlikely(ret == NULL))
-			ret = arena_init(ind);
+			ret = arena_init(tsdn, ind);
 	}
 	return (ret);
 }
@@ -863,30 +995,27 @@ decay_ticker_get(tsd_t *tsd, unsigned ind)

 #ifndef JEMALLOC_ENABLE_INLINE
 arena_t	*iaalloc(const void *ptr);
-size_t	isalloc(const void *ptr, bool demote);
-void	*iallocztm(tsd_t *tsd, size_t size, szind_t ind, bool zero,
+size_t	isalloc(tsdn_t *tsdn, const void *ptr, bool demote);
+void	*iallocztm(tsdn_t *tsdn, size_t size, szind_t ind, bool zero,
    tcache_t *tcache, bool is_metadata, arena_t *arena, bool slow_path);
-void	*imalloct(tsd_t *tsd, size_t size, szind_t ind, tcache_t *tcache,
-    arena_t *arena);
-void	*imalloc(tsd_t *tsd, size_t size, szind_t ind, bool slow_path);
-void	*icalloct(tsd_t *tsd, size_t size, szind_t ind, tcache_t *tcache,
-    arena_t *arena);
-void	*icalloc(tsd_t *tsd, size_t size, szind_t ind);
-void	*ipallocztm(tsd_t *tsd, size_t usize, size_t alignment, bool zero,
+void	*ialloc(tsd_t *tsd, size_t size, szind_t ind, bool zero,
+    bool slow_path);
+void	*ipallocztm(tsdn_t *tsdn, size_t usize, size_t alignment, bool zero,
    tcache_t *tcache, bool is_metadata, arena_t *arena);
-void	*ipalloct(tsd_t *tsd, size_t usize, size_t alignment, bool zero,
+void	*ipalloct(tsdn_t *tsdn, size_t usize, size_t alignment, bool zero,
    tcache_t *tcache, arena_t *arena);
 void	*ipalloc(tsd_t *tsd, size_t usize, size_t alignment, bool zero);
-size_t	ivsalloc(const void *ptr, bool demote);
+size_t	ivsalloc(tsdn_t *tsdn, const void *ptr, bool demote);
 size_t	u2rz(size_t usize);
-size_t	p2rz(const void *ptr);
-void	idalloctm(tsd_t *tsd, void *ptr, tcache_t *tcache, bool is_metadata,
+size_t	p2rz(tsdn_t *tsdn, const void *ptr);
+void	idalloctm(tsdn_t *tsdn, void *ptr, tcache_t *tcache, bool is_metadata,
    bool slow_path);
-void	idalloct(tsd_t *tsd, void *ptr, tcache_t *tcache);
 void	idalloc(tsd_t *tsd, void *ptr);
 void	iqalloc(tsd_t *tsd, void *ptr, tcache_t *tcache, bool slow_path);
-void	isdalloct(tsd_t *tsd, void *ptr, size_t size, tcache_t *tcache);
-void	isqalloc(tsd_t *tsd, void *ptr, size_t size, tcache_t *tcache);
+void	isdalloct(tsdn_t *tsdn, void *ptr, size_t size, tcache_t *tcache,
+    bool slow_path);
+void	isqalloc(tsd_t *tsd, void *ptr, size_t size, tcache_t *tcache,
+    bool slow_path);
 void	*iralloct_realign(tsd_t *tsd, void *ptr, size_t oldsize, size_t size,
    size_t extra, size_t alignment, bool zero, tcache_t *tcache,
    arena_t *arena);
@@ -894,7 +1023,7 @@ void	*iralloct(tsd_t *tsd, void *ptr, size_t oldsize, size_t size,
    size_t alignment, bool zero, tcache_t *tcache, arena_t *arena);
 void	*iralloc(tsd_t *tsd, void *ptr, size_t oldsize, size_t size,
    size_t alignment, bool zero);
-bool	ixalloc(tsd_t *tsd, void *ptr, size_t oldsize, size_t size,
+bool	ixalloc(tsdn_t *tsdn, void *ptr, size_t oldsize, size_t size,
    size_t extra, size_t alignment, bool zero);
 #endif

@@ -910,102 +1039,85 @@ iaalloc(const void *ptr)

 /*
 * Typical usage:
+ *   tsdn_t *tsdn = [...]
 *   void *ptr = [...]
- *   size_t sz = isalloc(ptr, config_prof);
+ *   size_t sz = isalloc(tsdn, ptr, config_prof);
 */
 JEMALLOC_ALWAYS_INLINE size_t
-isalloc(const void *ptr, bool demote)
+isalloc(tsdn_t *tsdn, const void *ptr, bool demote)
 {

 	assert(ptr != NULL);
 	/* Demotion only makes sense if config_prof is true. */
 	assert(config_prof || !demote);

-	return (arena_salloc(ptr, demote));
+	return (arena_salloc(tsdn, ptr, demote));
 }

 JEMALLOC_ALWAYS_INLINE void *
-iallocztm(tsd_t *tsd, size_t size, szind_t ind, bool zero, tcache_t *tcache,
+iallocztm(tsdn_t *tsdn, size_t size, szind_t ind, bool zero, tcache_t *tcache,
    bool is_metadata, arena_t *arena, bool slow_path)
 {
 	void *ret;

 	assert(size != 0);
+	assert(!is_metadata || tcache == NULL);
+	assert(!is_metadata || arena == NULL || arena->ind < narenas_auto);

-	ret = arena_malloc(tsd, arena, size, ind, zero, tcache, slow_path);
+	ret = arena_malloc(tsdn, arena, size, ind, zero, tcache, slow_path);
 	if (config_stats && is_metadata && likely(ret != NULL)) {
-		arena_metadata_allocated_add(iaalloc(ret), isalloc(ret,
-		    config_prof));
+		arena_metadata_allocated_add(iaalloc(ret),
+		    isalloc(tsdn, ret, config_prof));
 	}
 	return (ret);
 }

 JEMALLOC_ALWAYS_INLINE void *
-imalloct(tsd_t *tsd, size_t size, szind_t ind, tcache_t *tcache, arena_t *arena)
+ialloc(tsd_t *tsd, size_t size, szind_t ind, bool zero, bool slow_path)
 {

-	return (iallocztm(tsd, size, ind, false, tcache, false, arena, true));
+	return (iallocztm(tsd_tsdn(tsd), size, ind, zero, tcache_get(tsd, true),
+	    false, NULL, slow_path));
 }

 JEMALLOC_ALWAYS_INLINE void *
-imalloc(tsd_t *tsd, size_t size, szind_t ind, bool slow_path)
-{
-
-	return (iallocztm(tsd, size, ind, false, tcache_get(tsd, true), false,
-	    NULL, slow_path));
-}
-
-JEMALLOC_ALWAYS_INLINE void *
-icalloct(tsd_t *tsd, size_t size, szind_t ind, tcache_t *tcache, arena_t *arena)
-{
-
-	return (iallocztm(tsd, size, ind, true, tcache, false, arena, true));
-}
-
-JEMALLOC_ALWAYS_INLINE void *
-icalloc(tsd_t *tsd, size_t size, szind_t ind)
-{
-
-	return (iallocztm(tsd, size, ind, true, tcache_get(tsd, true), false,
-	    NULL, true));
-}
-
-JEMALLOC_ALWAYS_INLINE void *
-ipallocztm(tsd_t *tsd, size_t usize, size_t alignment, bool zero,
+ipallocztm(tsdn_t *tsdn, size_t usize, size_t alignment, bool zero,
    tcache_t *tcache, bool is_metadata, arena_t *arena)
 {
 	void *ret;

 	assert(usize != 0);
 	assert(usize == sa2u(usize, alignment));
+	assert(!is_metadata || tcache == NULL);
+	assert(!is_metadata || arena == NULL || arena->ind < narenas_auto);

-	ret = arena_palloc(tsd, arena, usize, alignment, zero, tcache);
+	ret = arena_palloc(tsdn, arena, usize, alignment, zero, tcache);
 	assert(ALIGNMENT_ADDR2BASE(ret, alignment) == ret);
 	if (config_stats && is_metadata && likely(ret != NULL)) {
-		arena_metadata_allocated_add(iaalloc(ret), isalloc(ret,
+		arena_metadata_allocated_add(iaalloc(ret), isalloc(tsdn, ret,
 		    config_prof));
 	}
 	return (ret);
 }

 JEMALLOC_ALWAYS_INLINE void *
-ipalloct(tsd_t *tsd, size_t usize, size_t alignment, bool zero,
+ipalloct(tsdn_t *tsdn, size_t usize, size_t alignment, bool zero,
    tcache_t *tcache, arena_t *arena)
 {

-	return (ipallocztm(tsd, usize, alignment, zero, tcache, false, arena));
+	return (ipallocztm(tsdn, usize, alignment, zero, tcache, false, arena));
 }

 JEMALLOC_ALWAYS_INLINE void *
 ipalloc(tsd_t *tsd, size_t usize, size_t alignment, bool zero)
 {

-	return (ipallocztm(tsd, usize, alignment, zero, tcache_get(tsd, true),
-	    false, NULL));
+	return (ipallocztm(tsd_tsdn(tsd), usize, alignment, zero,
+	    tcache_get(tsd, true), false, NULL));
 }

 JEMALLOC_ALWAYS_INLINE size_t
-ivsalloc(const void *ptr, bool demote)
+ivsalloc(tsdn_t *tsdn, const void *ptr, bool demote)
 {
 	extent_node_t *node;

@@ -1017,7 +1129,7 @@ ivsalloc(const void *ptr, bool demote)
 	assert(extent_node_addr_get(node) == ptr ||
 	    extent_node_achunk_get(node));

-	return (isalloc(ptr, demote));
+	return (isalloc(tsdn, ptr, demote));
 }

 JEMALLOC_INLINE size_t
@@ -1035,39 +1147,34 @@ u2rz(size_t usize)
 }

 JEMALLOC_INLINE size_t
-p2rz(const void *ptr)
+p2rz(tsdn_t *tsdn, const void *ptr)
 {
-	size_t usize = isalloc(ptr, false);
+	size_t usize = isalloc(tsdn, ptr, false);

 	return (u2rz(usize));
 }

 JEMALLOC_ALWAYS_INLINE void
-idalloctm(tsd_t *tsd, void *ptr, tcache_t *tcache, bool is_metadata,
+idalloctm(tsdn_t *tsdn, void *ptr, tcache_t *tcache, bool is_metadata,
    bool slow_path)
 {

 	assert(ptr != NULL);
+	assert(!is_metadata || tcache == NULL);
+	assert(!is_metadata || iaalloc(ptr)->ind < narenas_auto);
 	if (config_stats && is_metadata) {
-		arena_metadata_allocated_sub(iaalloc(ptr), isalloc(ptr,
+		arena_metadata_allocated_sub(iaalloc(ptr), isalloc(tsdn, ptr,
 		    config_prof));
 	}

-	arena_dalloc(tsd, ptr, tcache, slow_path);
-}
-
-JEMALLOC_ALWAYS_INLINE void
-idalloct(tsd_t *tsd, void *ptr, tcache_t *tcache)
-{
-
-	idalloctm(tsd, ptr, tcache, false, true);
+	arena_dalloc(tsdn, ptr, tcache, slow_path);
 }

 JEMALLOC_ALWAYS_INLINE void
 idalloc(tsd_t *tsd, void *ptr)
 {

-	idalloctm(tsd, ptr, tcache_get(tsd, false), false, true);
+	idalloctm(tsd_tsdn(tsd), ptr, tcache_get(tsd, false), false, true);
 }

 JEMALLOC_ALWAYS_INLINE void
@@ -1077,24 +1184,25 @@ iqalloc(tsd_t *tsd, void *ptr, tcache_t *tcache, bool slow_path)
 	if (slow_path && config_fill && unlikely(opt_quarantine))
 		quarantine(tsd, ptr);
 	else
-		idalloctm(tsd, ptr, tcache, false, slow_path);
+		idalloctm(tsd_tsdn(tsd), ptr, tcache, false, slow_path);
 }

 JEMALLOC_ALWAYS_INLINE void
-isdalloct(tsd_t *tsd, void *ptr, size_t size, tcache_t *tcache)
+isdalloct(tsdn_t *tsdn, void *ptr, size_t size, tcache_t *tcache,
+    bool slow_path)
 {

-	arena_sdalloc(tsd, ptr, size, tcache);
+	arena_sdalloc(tsdn, ptr, size, tcache, slow_path);
 }

 JEMALLOC_ALWAYS_INLINE void
-isqalloc(tsd_t *tsd, void *ptr, size_t size, tcache_t *tcache)
+isqalloc(tsd_t *tsd, void *ptr, size_t size, tcache_t *tcache, bool slow_path)
 {

-	if (config_fill && unlikely(opt_quarantine))
+	if (slow_path && config_fill && unlikely(opt_quarantine))
 		quarantine(tsd, ptr);
 	else
-		isdalloct(tsd, ptr, size, tcache);
+		isdalloct(tsd_tsdn(tsd), ptr, size, tcache, slow_path);
 }

 JEMALLOC_ALWAYS_INLINE void *
@@ -1107,7 +1215,7 @@ iralloct_realign(tsd_t *tsd, void *ptr, size_t oldsize, size_t size,
 	usize = sa2u(size + extra, alignment);
 	if (unlikely(usize == 0 || usize > HUGE_MAXCLASS))
 		return (NULL);
-	p = ipalloct(tsd, usize, alignment, zero, tcache, arena);
+	p = ipalloct(tsd_tsdn(tsd), usize, alignment, zero, tcache, arena);
 	if (p == NULL) {
 		if (extra == 0)
 			return (NULL);
@@ -1115,7 +1223,8 @@ iralloct_realign(tsd_t *tsd, void *ptr, size_t oldsize, size_t size,
 		usize = sa2u(size, alignment);
 		if (unlikely(usize == 0 || usize > HUGE_MAXCLASS))
 			return (NULL);
-		p = ipalloct(tsd, usize, alignment, zero, tcache, arena);
+		p = ipalloct(tsd_tsdn(tsd), usize, alignment, zero, tcache,
+		    arena);
 		if (p == NULL)
 			return (NULL);
 	}
@@ -1125,7 +1234,7 @@ iralloct_realign(tsd_t *tsd, void *ptr, size_t oldsize, size_t size,
 	 */
 	copysize = (size < oldsize) ? size : oldsize;
 	memcpy(p, ptr, copysize);
-	isqalloc(tsd, ptr, oldsize, tcache);
+	isqalloc(tsd, ptr, oldsize, tcache, true);
 	return (p);
 }

@@ -1161,7 +1270,7 @@ iralloc(tsd_t *tsd, void *ptr, size_t oldsize, size_t size, size_t alignment,
 }

 JEMALLOC_ALWAYS_INLINE bool
-ixalloc(tsd_t *tsd, void *ptr, size_t oldsize, size_t size, size_t extra,
+ixalloc(tsdn_t *tsdn, void *ptr, size_t oldsize, size_t size, size_t extra,
    size_t alignment, bool zero)
 {

@@ -1174,7 +1283,7 @@ ixalloc(tsd_t *tsd, void *ptr, size_t oldsize, size_t size, size_t extra,
 		return (true);
 	}

-	return (arena_ralloc_no_move(tsd, ptr, oldsize, size, extra, zero));
+	return (arena_ralloc_no_move(tsdn, ptr, oldsize, size, extra, zero));
 }
 #endif

--- a/include/jemalloc/internal/jemalloc_internal_decls.h
+++ b/include/jemalloc/internal/jemalloc_internal_decls.h
@@ -17,8 +17,18 @@
 #    include <sys/uio.h>
 #  endif
 #  include <pthread.h>
+#  ifdef JEMALLOC_OS_UNFAIR_LOCK
+#    include <os/lock.h>
+#  endif
+#  ifdef JEMALLOC_GLIBC_MALLOC_HOOK
+#    include <sched.h>
+#  endif
 #  include <errno.h>
 #  include <sys/time.h>
+#  include <time.h>
+#  ifdef JEMALLOC_HAVE_MACH_ABSOLUTE_TIME
+#    include <mach/mach_time.h>
+#  endif
 #endif
 #include <sys/types.h>

--- a/include/jemalloc/internal/jemalloc_internal_defs.h.in
+++ b/include/jemalloc/internal/jemalloc_internal_defs.h.in
@@ -56,9 +56,9 @@
 #undef JEMALLOC_HAVE_BUILTIN_CLZ

 /*
- * Defined if madvise(2) is available.
+ * Defined if os_unfair_lock_*() functions are available, as provided by Darwin.
 */
-#undef JEMALLOC_HAVE_MADVISE
+#undef JEMALLOC_OS_UNFAIR_LOCK

 /*
 * Defined if OSSpin*() functions are available, as provided by Darwin, and
@@ -66,6 +66,9 @@
 */
 #undef JEMALLOC_OSSPIN

+/* Defined if syscall(2) is usable. */
+#undef JEMALLOC_USE_SYSCALL
+
 /*
 * Defined if secure_getenv(3) is available.
 */
@@ -76,6 +79,24 @@
 */
 #undef JEMALLOC_HAVE_ISSETUGID

+/* Defined if pthread_atfork(3) is available. */
+#undef JEMALLOC_HAVE_PTHREAD_ATFORK
+
+/*
+ * Defined if clock_gettime(CLOCK_MONOTONIC_COARSE, ...) is available.
+ */
+#undef JEMALLOC_HAVE_CLOCK_MONOTONIC_COARSE
+
+/*
+ * Defined if clock_gettime(CLOCK_MONOTONIC, ...) is available.
+ */
+#undef JEMALLOC_HAVE_CLOCK_MONOTONIC
+
+/*
+ * Defined if mach_absolute_time() is available.
+ */
+#undef JEMALLOC_HAVE_MACH_ABSOLUTE_TIME
+
 /*
 * Defined if _malloc_thread_cleanup() exists.  At least in the case of
 * FreeBSD, pthread_key_create() allocates, which if used during malloc
@@ -188,6 +209,12 @@
 /* TLS is used to map arenas and magazine caches to threads. */
 #undef JEMALLOC_TLS

+/*
+ * Used to mark unreachable code to quiet "end of non-void" compiler warnings.
+ * Don't use this directly; instead use unreachable() from util.h
+ */
+#undef JEMALLOC_INTERNAL_UNREACHABLE
+
 /*
 * ffs*() functions to use for bitmapping.  Don't use these directly; instead,
 * use ffs_*() from util.h.
@@ -212,20 +239,39 @@
 * Darwin (OS X) uses zones to work around Mach-O symbol override shortcomings.
 */
 #undef JEMALLOC_ZONE
-#undef JEMALLOC_ZONE_VERSION
+
+/*
+ * Methods for determining whether the OS overcommits.
+ * JEMALLOC_PROC_SYS_VM_OVERCOMMIT_MEMORY: Linux's
+ *                                         /proc/sys/vm.overcommit_memory file.
+ * JEMALLOC_SYSCTL_VM_OVERCOMMIT: FreeBSD's vm.overcommit sysctl.
+ */
+#undef JEMALLOC_SYSCTL_VM_OVERCOMMIT
+#undef JEMALLOC_PROC_SYS_VM_OVERCOMMIT_MEMORY
+
+/* Defined if madvise(2) is available. */
+#undef JEMALLOC_HAVE_MADVISE
+
+/*
+ * Defined if transparent huge pages are supported via the MADV_[NO]HUGEPAGE
+ * arguments to madvise(2).
+ */
+#undef JEMALLOC_HAVE_MADVISE_HUGE

 /*
 * 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.
+ *   madvise(..., MADV_FREE) : This marks pages as being unused, such that they
+ *                             will be discarded rather than swapped out.
+ *   madvise(..., MADV_DONTNEED) : This immediately discards pages, such that
+ *                                 new pages will be demand-zeroed if the
+ *                                 address region is later touched.
 */
-#undef JEMALLOC_PURGE_MADVISE_DONTNEED
 #undef JEMALLOC_PURGE_MADVISE_FREE
+#undef JEMALLOC_PURGE_MADVISE_DONTNEED
+
+/* Defined if transparent huge page support is enabled. */
+#undef JEMALLOC_THP

 /* Define if operating system has alloca.h header. */
 #undef JEMALLOC_HAS_ALLOCA_H
--- a/include/jemalloc/internal/mb.h
+++ b/include/jemalloc/internal/mb.h
@@ -42,7 +42,7 @@ mb_write(void)
 	    : /* Inputs. */
 	    : "memory" /* Clobbers. */
 	    );
-#else
+#  else
 	/*
 	 * This is hopefully enough to keep the compiler from reordering
 	 * instructions around this one.
@@ -52,7 +52,7 @@ mb_write(void)
 	    : /* Inputs. */
 	    : "memory" /* Clobbers. */
 	    );
-#endif
+#  endif
 }
 #elif (defined(__amd64__) || defined(__x86_64__))
 JEMALLOC_INLINE void
@@ -76,7 +76,7 @@ mb_write(void)
 	    : "memory" /* Clobbers. */
 	    );
 }
-#elif defined(__sparc64__)
+#elif defined(__sparc__) && defined(__arch64__)
 JEMALLOC_INLINE void
 mb_write(void)
 {
@@ -104,9 +104,9 @@ mb_write(void)
 {
 	malloc_mutex_t mtx;

-	malloc_mutex_init(&mtx);
-	malloc_mutex_lock(&mtx);
-	malloc_mutex_unlock(&mtx);
+	malloc_mutex_init(&mtx, "mb", WITNESS_RANK_OMIT);
+	malloc_mutex_lock(TSDN_NULL, &mtx);
+	malloc_mutex_unlock(TSDN_NULL, &mtx);
 }
 #endif
 #endif
--- a/include/jemalloc/internal/mutex.h
+++ b/include/jemalloc/internal/mutex.h
@@ -5,18 +5,25 @@ typedef struct malloc_mutex_s malloc_mutex_t;

 #ifdef _WIN32
 #  define MALLOC_MUTEX_INITIALIZER
+#elif (defined(JEMALLOC_OS_UNFAIR_LOCK))
+#  define MALLOC_MUTEX_INITIALIZER					\
+     {OS_UNFAIR_LOCK_INIT, WITNESS_INITIALIZER(WITNESS_RANK_OMIT)}
 #elif (defined(JEMALLOC_OSSPIN))
-#  define MALLOC_MUTEX_INITIALIZER {0}
+#  define MALLOC_MUTEX_INITIALIZER {0, WITNESS_INITIALIZER(WITNESS_RANK_OMIT)}
 #elif (defined(JEMALLOC_MUTEX_INIT_CB))
-#  define MALLOC_MUTEX_INITIALIZER {PTHREAD_MUTEX_INITIALIZER, NULL}
+#  define MALLOC_MUTEX_INITIALIZER					\
+    {PTHREAD_MUTEX_INITIALIZER, NULL, WITNESS_INITIALIZER(WITNESS_RANK_OMIT)}
 #else
 #  if (defined(JEMALLOC_HAVE_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}
+#    define MALLOC_MUTEX_INITIALIZER					\
+       {PTHREAD_ADAPTIVE_MUTEX_INITIALIZER_NP,				\
+        WITNESS_INITIALIZER(WITNESS_RANK_OMIT)}
 #  else
 #    define MALLOC_MUTEX_TYPE PTHREAD_MUTEX_DEFAULT
-#    define MALLOC_MUTEX_INITIALIZER {PTHREAD_MUTEX_INITIALIZER}
+#    define MALLOC_MUTEX_INITIALIZER					\
+       {PTHREAD_MUTEX_INITIALIZER, WITNESS_INITIALIZER(WITNESS_RANK_OMIT)}
 #  endif
 #endif

@@ -31,6 +38,8 @@ struct malloc_mutex_s {
 #  else
 	CRITICAL_SECTION	lock;
 #  endif
+#elif (defined(JEMALLOC_OS_UNFAIR_LOCK))
+	os_unfair_lock		lock;
 #elif (defined(JEMALLOC_OSSPIN))
 	OSSpinLock		lock;
 #elif (defined(JEMALLOC_MUTEX_INIT_CB))
@@ -39,6 +48,7 @@ struct malloc_mutex_s {
 #else
 	pthread_mutex_t		lock;
 #endif
+	witness_t		witness;
 };

 #endif /* JEMALLOC_H_STRUCTS */
@@ -52,26 +62,30 @@ extern bool isthreaded;
 #  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);
+bool	malloc_mutex_init(malloc_mutex_t *mutex, const char *name,
+    witness_rank_t rank);
+void	malloc_mutex_prefork(tsdn_t *tsdn, malloc_mutex_t *mutex);
+void	malloc_mutex_postfork_parent(tsdn_t *tsdn, malloc_mutex_t *mutex);
+void	malloc_mutex_postfork_child(tsdn_t *tsdn, malloc_mutex_t *mutex);
+bool	malloc_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);
+void	malloc_mutex_lock(tsdn_t *tsdn, malloc_mutex_t *mutex);
+void	malloc_mutex_unlock(tsdn_t *tsdn, malloc_mutex_t *mutex);
+void	malloc_mutex_assert_owner(tsdn_t *tsdn, malloc_mutex_t *mutex);
+void	malloc_mutex_assert_not_owner(tsdn_t *tsdn, malloc_mutex_t *mutex);
 #endif

 #if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_MUTEX_C_))
 JEMALLOC_INLINE void
-malloc_mutex_lock(malloc_mutex_t *mutex)
+malloc_mutex_lock(tsdn_t *tsdn, malloc_mutex_t *mutex)
 {

+	witness_assert_not_owner(tsdn, &mutex->witness);
 	if (isthreaded) {
 #ifdef _WIN32
 #  if _WIN32_WINNT >= 0x0600
@@ -79,18 +93,22 @@ malloc_mutex_lock(malloc_mutex_t *mutex)
 #  else
 		EnterCriticalSection(&mutex->lock);
 #  endif
+#elif (defined(JEMALLOC_OS_UNFAIR_LOCK))
+		os_unfair_lock_lock(&mutex->lock);
 #elif (defined(JEMALLOC_OSSPIN))
 		OSSpinLockLock(&mutex->lock);
 #else
 		pthread_mutex_lock(&mutex->lock);
 #endif
 	}
+	witness_lock(tsdn, &mutex->witness);
 }

 JEMALLOC_INLINE void
-malloc_mutex_unlock(malloc_mutex_t *mutex)
+malloc_mutex_unlock(tsdn_t *tsdn, malloc_mutex_t *mutex)
 {

+	witness_unlock(tsdn, &mutex->witness);
 	if (isthreaded) {
 #ifdef _WIN32
 #  if _WIN32_WINNT >= 0x0600
@@ -98,6 +116,8 @@ malloc_mutex_unlock(malloc_mutex_t *mutex)
 #  else
 		LeaveCriticalSection(&mutex->lock);
 #  endif
+#elif (defined(JEMALLOC_OS_UNFAIR_LOCK))
+		os_unfair_lock_unlock(&mutex->lock);
 #elif (defined(JEMALLOC_OSSPIN))
 		OSSpinLockUnlock(&mutex->lock);
 #else
@@ -105,6 +125,20 @@ malloc_mutex_unlock(malloc_mutex_t *mutex)
 #endif
 	}
 }
+
+JEMALLOC_INLINE void
+malloc_mutex_assert_owner(tsdn_t *tsdn, malloc_mutex_t *mutex)
+{
+
+	witness_assert_owner(tsdn, &mutex->witness);
+}
+
+JEMALLOC_INLINE void
+malloc_mutex_assert_not_owner(tsdn_t *tsdn, malloc_mutex_t *mutex)
+{
+
+	witness_assert_not_owner(tsdn, &mutex->witness);
+}
 #endif

 #endif /* JEMALLOC_H_INLINES */
--- a/include/jemalloc/internal/nstime.h
+++ b/include/jemalloc/internal/nstime.h
@@ -1,13 +1,10 @@
 /******************************************************************************/
 #ifdef JEMALLOC_H_TYPES

-#define JEMALLOC_CLOCK_GETTIME defined(_POSIX_MONOTONIC_CLOCK) \
-    && _POSIX_MONOTONIC_CLOCK >= 0
-
 typedef struct nstime_s nstime_t;

 /* Maximum supported number of seconds (~584 years). */
-#define	NSTIME_SEC_MAX	18446744072
+#define	NSTIME_SEC_MAX	KQU(18446744072)

 #endif /* JEMALLOC_H_TYPES */
 /******************************************************************************/
@@ -34,9 +31,12 @@ void	nstime_imultiply(nstime_t *time, uint64_t multiplier);
 void	nstime_idivide(nstime_t *time, uint64_t divisor);
 uint64_t	nstime_divide(const nstime_t *time, const nstime_t *divisor);
 #ifdef JEMALLOC_JET
+typedef bool (nstime_monotonic_t)(void);
+extern nstime_monotonic_t *nstime_monotonic;
 typedef bool (nstime_update_t)(nstime_t *);
 extern nstime_update_t *nstime_update;
 #else
+bool	nstime_monotonic(void);
 bool	nstime_update(nstime_t *time);
 #endif

--- a/include/jemalloc/internal/pages.h
+++ b/include/jemalloc/internal/pages.h
@@ -9,13 +9,16 @@
 /******************************************************************************/
 #ifdef JEMALLOC_H_EXTERNS

-void	*pages_map(void *addr, size_t size);
+void	*pages_map(void *addr, size_t size, bool *commit);
 void	pages_unmap(void *addr, size_t size);
 void	*pages_trim(void *addr, size_t alloc_size, size_t leadsize,
-    size_t size);
+    size_t size, bool *commit);
 bool	pages_commit(void *addr, size_t size);
 bool	pages_decommit(void *addr, size_t size);
 bool	pages_purge(void *addr, size_t size);
+bool	pages_huge(void *addr, size_t size);
+bool	pages_nohuge(void *addr, size_t size);
+void	pages_boot(void);

 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
--- a/include/jemalloc/internal/ph.h
+++ b/include/jemalloc/internal/ph.h
@@ -0,0 +1,345 @@
+/*
+ * A Pairing Heap implementation.
+ *
+ * "The Pairing Heap: A New Form of Self-Adjusting Heap"
+ * https://www.cs.cmu.edu/~sleator/papers/pairing-heaps.pdf
+ *
+ * With auxiliary twopass list, described in a follow on paper.
+ *
+ * "Pairing Heaps: Experiments and Analysis"
+ * http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.106.2988&rep=rep1&type=pdf
+ *
+ *******************************************************************************
+ */
+
+#ifndef PH_H_
+#define	PH_H_
+
+/* Node structure. */
+#define	phn(a_type)							\
+struct {								\
+	a_type	*phn_prev;						\
+	a_type	*phn_next;						\
+	a_type	*phn_lchild;						\
+}
+
+/* Root structure. */
+#define	ph(a_type)							\
+struct {								\
+	a_type	*ph_root;						\
+}
+
+/* Internal utility macros. */
+#define	phn_lchild_get(a_type, a_field, a_phn)				\
+	(a_phn->a_field.phn_lchild)
+#define	phn_lchild_set(a_type, a_field, a_phn, a_lchild) do {		\
+	a_phn->a_field.phn_lchild = a_lchild;				\
+} while (0)
+
+#define	phn_next_get(a_type, a_field, a_phn)				\
+	(a_phn->a_field.phn_next)
+#define	phn_prev_set(a_type, a_field, a_phn, a_prev) do {		\
+	a_phn->a_field.phn_prev = a_prev;				\
+} while (0)
+
+#define	phn_prev_get(a_type, a_field, a_phn)				\
+	(a_phn->a_field.phn_prev)
+#define	phn_next_set(a_type, a_field, a_phn, a_next) do {		\
+	a_phn->a_field.phn_next = a_next;				\
+} while (0)
+
+#define	phn_merge_ordered(a_type, a_field, a_phn0, a_phn1, a_cmp) do {	\
+	a_type *phn0child;						\
+									\
+	assert(a_phn0 != NULL);						\
+	assert(a_phn1 != NULL);						\
+	assert(a_cmp(a_phn0, a_phn1) <= 0);				\
+									\
+	phn_prev_set(a_type, a_field, a_phn1, a_phn0);			\
+	phn0child = phn_lchild_get(a_type, a_field, a_phn0);		\
+	phn_next_set(a_type, a_field, a_phn1, phn0child);		\
+	if (phn0child != NULL)						\
+		phn_prev_set(a_type, a_field, phn0child, a_phn1);	\
+	phn_lchild_set(a_type, a_field, a_phn0, a_phn1);		\
+} while (0)
+
+#define	phn_merge(a_type, a_field, a_phn0, a_phn1, a_cmp, r_phn) do {	\
+	if (a_phn0 == NULL)						\
+		r_phn = a_phn1;						\
+	else if (a_phn1 == NULL)					\
+		r_phn = a_phn0;						\
+	else if (a_cmp(a_phn0, a_phn1) < 0) {				\
+		phn_merge_ordered(a_type, a_field, a_phn0, a_phn1,	\
+		    a_cmp);						\
+		r_phn = a_phn0;						\
+	} else {							\
+		phn_merge_ordered(a_type, a_field, a_phn1, a_phn0,	\
+		    a_cmp);						\
+		r_phn = a_phn1;						\
+	}								\
+} while (0)
+
+#define	ph_merge_siblings(a_type, a_field, a_phn, a_cmp, r_phn) do {	\
+	a_type *head = NULL;						\
+	a_type *tail = NULL;						\
+	a_type *phn0 = a_phn;						\
+	a_type *phn1 = phn_next_get(a_type, a_field, phn0);		\
+									\
+	/*								\
+	 * Multipass merge, wherein the first two elements of a FIFO	\
+	 * are repeatedly merged, and each result is appended to the	\
+	 * singly linked FIFO, until the FIFO contains only a single	\
+	 * element.  We start with a sibling list but no reference to	\
+	 * its tail, so we do a single pass over the sibling list to	\
+	 * populate the FIFO.						\
+	 */								\
+	if (phn1 != NULL) {						\
+		a_type *phnrest = phn_next_get(a_type, a_field, phn1);	\
+		if (phnrest != NULL)					\
+			phn_prev_set(a_type, a_field, phnrest, NULL);	\
+		phn_prev_set(a_type, a_field, phn0, NULL);		\
+		phn_next_set(a_type, a_field, phn0, NULL);		\
+		phn_prev_set(a_type, a_field, phn1, NULL);		\
+		phn_next_set(a_type, a_field, phn1, NULL);		\
+		phn_merge(a_type, a_field, phn0, phn1, a_cmp, phn0);	\
+		head = tail = phn0;					\
+		phn0 = phnrest;						\
+		while (phn0 != NULL) {					\
+			phn1 = phn_next_get(a_type, a_field, phn0);	\
+			if (phn1 != NULL) {				\
+				phnrest = phn_next_get(a_type, a_field,	\
+				    phn1);				\
+				if (phnrest != NULL) {			\
+					phn_prev_set(a_type, a_field,	\
+					    phnrest, NULL);		\
+				}					\
+				phn_prev_set(a_type, a_field, phn0,	\
+				    NULL);				\
+				phn_next_set(a_type, a_field, phn0,	\
+				    NULL);				\
+				phn_prev_set(a_type, a_field, phn1,	\
+				    NULL);				\
+				phn_next_set(a_type, a_field, phn1,	\
+				    NULL);				\
+				phn_merge(a_type, a_field, phn0, phn1,	\
+				    a_cmp, phn0);			\
+				phn_next_set(a_type, a_field, tail,	\
+				    phn0);				\
+				tail = phn0;				\
+				phn0 = phnrest;				\
+			} else {					\
+				phn_next_set(a_type, a_field, tail,	\
+				    phn0);				\
+				tail = phn0;				\
+				phn0 = NULL;				\
+			}						\
+		}							\
+		phn0 = head;						\
+		phn1 = phn_next_get(a_type, a_field, phn0);		\
+		if (phn1 != NULL) {					\
+			while (true) {					\
+				head = phn_next_get(a_type, a_field,	\
+				    phn1);				\
+				assert(phn_prev_get(a_type, a_field,	\
+				    phn0) == NULL);			\
+				phn_next_set(a_type, a_field, phn0,	\
+				    NULL);				\
+				assert(phn_prev_get(a_type, a_field,	\
+				    phn1) == NULL);			\
+				phn_next_set(a_type, a_field, phn1,	\
+				    NULL);				\
+				phn_merge(a_type, a_field, phn0, phn1,	\
+				    a_cmp, phn0);			\
+				if (head == NULL)			\
+					break;				\
+				phn_next_set(a_type, a_field, tail,	\
+				    phn0);				\
+				tail = phn0;				\
+				phn0 = head;				\
+				phn1 = phn_next_get(a_type, a_field,	\
+				    phn0);				\
+			}						\
+		}							\
+	}								\
+	r_phn = phn0;							\
+} while (0)
+
+#define	ph_merge_aux(a_type, a_field, a_ph, a_cmp) do {			\
+	a_type *phn = phn_next_get(a_type, a_field, a_ph->ph_root);	\
+	if (phn != NULL) {						\
+		phn_prev_set(a_type, a_field, a_ph->ph_root, NULL);	\
+		phn_next_set(a_type, a_field, a_ph->ph_root, NULL);	\
+		phn_prev_set(a_type, a_field, phn, NULL);		\
+		ph_merge_siblings(a_type, a_field, phn, a_cmp, phn);	\
+		assert(phn_next_get(a_type, a_field, phn) == NULL);	\
+		phn_merge(a_type, a_field, a_ph->ph_root, phn, a_cmp,	\
+		    a_ph->ph_root);					\
+	}								\
+} while (0)
+
+#define	ph_merge_children(a_type, a_field, a_phn, a_cmp, r_phn) do {	\
+	a_type *lchild = phn_lchild_get(a_type, a_field, a_phn);	\
+	if (lchild == NULL)						\
+		r_phn = NULL;						\
+	else {								\
+		ph_merge_siblings(a_type, a_field, lchild, a_cmp,	\
+		    r_phn);						\
+	}								\
+} while (0)
+
+/*
+ * The ph_proto() macro generates function prototypes that correspond to the
+ * functions generated by an equivalently parameterized call to ph_gen().
+ */
+#define	ph_proto(a_attr, a_prefix, a_ph_type, a_type)			\
+a_attr void	a_prefix##new(a_ph_type *ph);				\
+a_attr bool	a_prefix##empty(a_ph_type *ph);				\
+a_attr a_type	*a_prefix##first(a_ph_type *ph);			\
+a_attr void	a_prefix##insert(a_ph_type *ph, a_type *phn);		\
+a_attr a_type	*a_prefix##remove_first(a_ph_type *ph);			\
+a_attr void	a_prefix##remove(a_ph_type *ph, a_type *phn);
+
+/*
+ * The ph_gen() macro generates a type-specific pairing heap implementation,
+ * based on the above cpp macros.
+ */
+#define	ph_gen(a_attr, a_prefix, a_ph_type, a_type, a_field, a_cmp)	\
+a_attr void								\
+a_prefix##new(a_ph_type *ph)						\
+{									\
+									\
+	memset(ph, 0, sizeof(ph(a_type)));				\
+}									\
+a_attr bool								\
+a_prefix##empty(a_ph_type *ph)						\
+{									\
+									\
+	return (ph->ph_root == NULL);					\
+}									\
+a_attr a_type *								\
+a_prefix##first(a_ph_type *ph)						\
+{									\
+									\
+	if (ph->ph_root == NULL)					\
+		return (NULL);						\
+	ph_merge_aux(a_type, a_field, ph, a_cmp);			\
+	return (ph->ph_root);						\
+}									\
+a_attr void								\
+a_prefix##insert(a_ph_type *ph, a_type *phn)				\
+{									\
+									\
+	memset(&phn->a_field, 0, sizeof(phn(a_type)));			\
+									\
+	/*								\
+	 * Treat the root as an aux list during insertion, and lazily	\
+	 * merge during a_prefix##remove_first().  For elements that	\
+	 * are inserted, then removed via a_prefix##remove() before the	\
+	 * aux list is ever processed, this makes insert/remove		\
+	 * constant-time, whereas eager merging would make insert	\
+	 * O(log n).							\
+	 */								\
+	if (ph->ph_root == NULL)					\
+		ph->ph_root = phn;					\
+	else {								\
+		phn_next_set(a_type, a_field, phn, phn_next_get(a_type,	\
+		    a_field, ph->ph_root));				\
+		if (phn_next_get(a_type, a_field, ph->ph_root) !=	\
+		    NULL) {						\
+			phn_prev_set(a_type, a_field,			\
+			    phn_next_get(a_type, a_field, ph->ph_root),	\
+			    phn);					\
+		}							\
+		phn_prev_set(a_type, a_field, phn, ph->ph_root);	\
+		phn_next_set(a_type, a_field, ph->ph_root, phn);	\
+	}								\
+}									\
+a_attr a_type *								\
+a_prefix##remove_first(a_ph_type *ph)					\
+{									\
+	a_type *ret;							\
+									\
+	if (ph->ph_root == NULL)					\
+		return (NULL);						\
+	ph_merge_aux(a_type, a_field, ph, a_cmp);			\
+									\
+	ret = ph->ph_root;						\
+									\
+	ph_merge_children(a_type, a_field, ph->ph_root, a_cmp,		\
+	    ph->ph_root);						\
+									\
+	return (ret);							\
+}									\
+a_attr void								\
+a_prefix##remove(a_ph_type *ph, a_type *phn)				\
+{									\
+	a_type *replace, *parent;					\
+									\
+	/*								\
+	 * We can delete from aux list without merging it, but we need	\
+	 * to merge if we are dealing with the root node.		\
+	 */								\
+	if (ph->ph_root == phn) {					\
+		ph_merge_aux(a_type, a_field, ph, a_cmp);		\
+		if (ph->ph_root == phn) {				\
+			ph_merge_children(a_type, a_field, ph->ph_root,	\
+			    a_cmp, ph->ph_root);			\
+			return;						\
+		}							\
+	}								\
+									\
+	/* Get parent (if phn is leftmost child) before mutating. */	\
+	if ((parent = phn_prev_get(a_type, a_field, phn)) != NULL) {	\
+		if (phn_lchild_get(a_type, a_field, parent) != phn)	\
+			parent = NULL;					\
+	}								\
+	/* Find a possible replacement node, and link to parent. */	\
+	ph_merge_children(a_type, a_field, phn, a_cmp, replace);	\
+	/* Set next/prev for sibling linked list. */			\
+	if (replace != NULL) {						\
+		if (parent != NULL) {					\
+			phn_prev_set(a_type, a_field, replace, parent);	\
+			phn_lchild_set(a_type, a_field, parent,		\
+			    replace);					\
+		} else {						\
+			phn_prev_set(a_type, a_field, replace,		\
+			    phn_prev_get(a_type, a_field, phn));	\
+			if (phn_prev_get(a_type, a_field, phn) !=	\
+			    NULL) {					\
+				phn_next_set(a_type, a_field,		\
+				    phn_prev_get(a_type, a_field, phn),	\
+				    replace);				\
+			}						\
+		}							\
+		phn_next_set(a_type, a_field, replace,			\
+		    phn_next_get(a_type, a_field, phn));		\
+		if (phn_next_get(a_type, a_field, phn) != NULL) {	\
+			phn_prev_set(a_type, a_field,			\
+			    phn_next_get(a_type, a_field, phn),		\
+			    replace);					\
+		}							\
+	} else {							\
+		if (parent != NULL) {					\
+			a_type *next = phn_next_get(a_type, a_field,	\
+			    phn);					\
+			phn_lchild_set(a_type, a_field, parent, next);	\
+			if (next != NULL) {				\
+				phn_prev_set(a_type, a_field, next,	\
+				    parent);				\
+			}						\
+		} else {						\
+			assert(phn_prev_get(a_type, a_field, phn) !=	\
+			    NULL);					\
+			phn_next_set(a_type, a_field,			\
+			    phn_prev_get(a_type, a_field, phn),		\
+			    phn_next_get(a_type, a_field, phn));	\
+		}							\
+		if (phn_next_get(a_type, a_field, phn) != NULL) {	\
+			phn_prev_set(a_type, a_field,			\
+			    phn_next_get(a_type, a_field, phn),		\
+			    phn_prev_get(a_type, a_field, phn));	\
+		}							\
+	}								\
+}
+
+#endif /* PH_H_ */
--- a/include/jemalloc/internal/private_symbols.txt
+++ b/include/jemalloc/internal/private_symbols.txt
@@ -1,14 +1,17 @@
 a0dalloc
+a0get
 a0malloc
 arena_aalloc
 arena_alloc_junk_small
 arena_basic_stats_merge
 arena_bin_index
 arena_bin_info
-arena_bitselm_get
+arena_bitselm_get_const
+arena_bitselm_get_mutable
 arena_boot
 arena_choose
 arena_choose_hard
+arena_choose_impl
 arena_chunk_alloc_huge
 arena_chunk_cache_maybe_insert
 arena_chunk_cache_maybe_remove
@@ -21,9 +24,7 @@ arena_dalloc
 arena_dalloc_bin
 arena_dalloc_bin_junked_locked
 arena_dalloc_junk_large
-arena_dalloc_junk_large_impl
 arena_dalloc_junk_small
-arena_dalloc_junk_small_impl
 arena_dalloc_large
 arena_dalloc_large_junked_locked
 arena_dalloc_small
@@ -35,7 +36,9 @@ arena_decay_time_get
 arena_decay_time_set
 arena_dss_prec_get
 arena_dss_prec_set
+arena_extent_sn_next
 arena_get
+arena_ichoose
 arena_init
 arena_lg_dirty_mult_default_get
 arena_lg_dirty_mult_default_set
@@ -62,7 +65,8 @@ arena_mapbits_unallocated_set
 arena_mapbits_unallocated_size_get
 arena_mapbits_unallocated_size_set
 arena_mapbits_unzeroed_get
-arena_mapbitsp_get
+arena_mapbitsp_get_const
+arena_mapbitsp_get_mutable
 arena_mapbitsp_read
 arena_mapbitsp_write
 arena_maxrun
@@ -71,7 +75,8 @@ arena_metadata_allocated_add
 arena_metadata_allocated_get
 arena_metadata_allocated_sub
 arena_migrate
-arena_miscelm_get
+arena_miscelm_get_const
+arena_miscelm_get_mutable
 arena_miscelm_to_pageind
 arena_miscelm_to_rpages
 arena_new
@@ -102,6 +107,7 @@ arena_ralloc_junk_large
 arena_ralloc_no_move
 arena_rd_to_miscelm
 arena_redzone_corruption
+arena_reset
 arena_run_regind
 arena_run_to_miscelm
 arena_salloc
@@ -163,20 +169,15 @@ chunk_dalloc_mmap
 chunk_dalloc_wrapper
 chunk_deregister
 chunk_dss_boot
-chunk_dss_postfork_child
-chunk_dss_postfork_parent
+chunk_dss_mergeable
 chunk_dss_prec_get
 chunk_dss_prec_set
-chunk_dss_prefork
 chunk_hooks_default
 chunk_hooks_get
 chunk_hooks_set
 chunk_in_dss
 chunk_lookup
 chunk_npages
-chunk_postfork_child
-chunk_postfork_parent
-chunk_prefork
 chunk_purge_wrapper
 chunk_register
 chunks_rtree
@@ -218,8 +219,12 @@ extent_node_prof_tctx_get
 extent_node_prof_tctx_set
 extent_node_size_get
 extent_node_size_set
+extent_node_sn_get
+extent_node_sn_set
 extent_node_zeroed_get
 extent_node_zeroed_set
+extent_size_quantize_ceil
+extent_size_quantize_floor
 extent_tree_ad_destroy
 extent_tree_ad_destroy_recurse
 extent_tree_ad_empty
@@ -239,25 +244,25 @@ extent_tree_ad_reverse_iter
 extent_tree_ad_reverse_iter_recurse
 extent_tree_ad_reverse_iter_start
 extent_tree_ad_search
-extent_tree_szad_destroy
-extent_tree_szad_destroy_recurse
-extent_tree_szad_empty
-extent_tree_szad_first
-extent_tree_szad_insert
-extent_tree_szad_iter
-extent_tree_szad_iter_recurse
-extent_tree_szad_iter_start
-extent_tree_szad_last
-extent_tree_szad_new
-extent_tree_szad_next
-extent_tree_szad_nsearch
-extent_tree_szad_prev
-extent_tree_szad_psearch
-extent_tree_szad_remove
-extent_tree_szad_reverse_iter
-extent_tree_szad_reverse_iter_recurse
-extent_tree_szad_reverse_iter_start
-extent_tree_szad_search
+extent_tree_szsnad_destroy
+extent_tree_szsnad_destroy_recurse
+extent_tree_szsnad_empty
+extent_tree_szsnad_first
+extent_tree_szsnad_insert
+extent_tree_szsnad_iter
+extent_tree_szsnad_iter_recurse
+extent_tree_szsnad_iter_start
+extent_tree_szsnad_last
+extent_tree_szsnad_new
+extent_tree_szsnad_next
+extent_tree_szsnad_nsearch
+extent_tree_szsnad_prev
+extent_tree_szsnad_psearch
+extent_tree_szsnad_remove
+extent_tree_szsnad_reverse_iter
+extent_tree_szsnad_reverse_iter_recurse
+extent_tree_szsnad_reverse_iter_start
+extent_tree_szsnad_search
 ffs_llu
 ffs_lu
 ffs_u
@@ -287,14 +292,11 @@ huge_ralloc
 huge_ralloc_no_move
 huge_salloc
 iaalloc
+ialloc
 iallocztm
-icalloc
-icalloct
+iarena_cleanup
 idalloc
-idalloct
 idalloctm
-imalloc
-imalloct
 in_valgrind
 index2size
 index2size_compute
@@ -320,6 +322,9 @@ large_maxclass
 lg_floor
 lg_prof_sample
 malloc_cprintf
+malloc_mutex_assert_not_owner
+malloc_mutex_assert_owner
+malloc_mutex_boot
 malloc_mutex_init
 malloc_mutex_lock
 malloc_mutex_postfork_child
@@ -341,7 +346,7 @@ malloc_write
 map_bias
 map_misc_offset
 mb_write
-mutex_boot
+narenas_auto
 narenas_tdata_cleanup
 narenas_total_get
 ncpus
@@ -356,12 +361,12 @@ nstime_idivide
 nstime_imultiply
 nstime_init
 nstime_init2
+nstime_monotonic
 nstime_ns
 nstime_nsec
 nstime_sec
 nstime_subtract
 nstime_update
-nstime_update_impl
 opt_abort
 opt_decay_time
 opt_dss
@@ -387,21 +392,36 @@ opt_quarantine
 opt_redzone
 opt_stats_print
 opt_tcache
+opt_thp
 opt_utrace
 opt_xmalloc
 opt_zero
 p2rz
+pages_boot
 pages_commit
 pages_decommit
+pages_huge
 pages_map
+pages_nohuge
 pages_purge
 pages_trim
 pages_unmap
+pind2sz
+pind2sz_compute
+pind2sz_lookup
+pind2sz_tab
 pow2_ceil_u32
 pow2_ceil_u64
 pow2_ceil_zu
-prng_lg_range
-prng_range
+prng_lg_range_u32
+prng_lg_range_u64
+prng_lg_range_zu
+prng_range_u32
+prng_range_u64
+prng_range_zu
+prng_state_next_u32
+prng_state_next_u64
+prng_state_next_zu
 prof_active
 prof_active_get
 prof_active_get_unlocked
@@ -450,12 +470,13 @@ prof_thread_active_init_set
 prof_thread_active_set
 prof_thread_name_get
 prof_thread_name_set
+psz2ind
+psz2u
 purge_mode_names
 quarantine
 quarantine_alloc_hook
 quarantine_alloc_hook_work
 quarantine_cleanup
-register_zone
 rtree_child_read
 rtree_child_read_hard
 rtree_child_tryread
@@ -473,7 +494,6 @@ rtree_val_read
 rtree_val_write
 run_quantize_ceil
 run_quantize_floor
-run_quantize_max
 s2u
 s2u_compute
 s2u_lookup
@@ -483,6 +503,8 @@ size2index
 size2index_compute
 size2index_lookup
 size2index_tab
+spin_adaptive
+spin_init
 stats_cactive
 stats_cactive_add
 stats_cactive_get
@@ -492,8 +514,6 @@ tcache_alloc_easy
 tcache_alloc_large
 tcache_alloc_small
 tcache_alloc_small_hard
-tcache_arena_associate
-tcache_arena_dissociate
 tcache_arena_reassociate
 tcache_bin_flush_large
 tcache_bin_flush_small
@@ -512,6 +532,9 @@ tcache_flush
 tcache_get
 tcache_get_hard
 tcache_maxclass
+tcache_postfork_child
+tcache_postfork_parent
+tcache_prefork
 tcache_salloc
 tcache_stats_merge
 tcaches
@@ -539,19 +562,25 @@ tsd_boot
 tsd_boot0
 tsd_boot1
 tsd_booted
+tsd_booted_get
 tsd_cleanup
 tsd_cleanup_wrapper
 tsd_fetch
+tsd_fetch_impl
 tsd_get
+tsd_get_allocates
+tsd_iarena_get
+tsd_iarena_set
+tsd_iarenap_get
+tsd_initialized
+tsd_init_check_recursion
+tsd_init_finish
+tsd_init_head
 tsd_narenas_tdata_get
 tsd_narenas_tdata_set
 tsd_narenas_tdatap_get
 tsd_wrapper_get
 tsd_wrapper_set
-tsd_initialized
-tsd_init_check_recursion
-tsd_init_finish
-tsd_init_head
 tsd_nominal
 tsd_prof_tdata_get
 tsd_prof_tdata_set
@@ -574,8 +603,37 @@ tsd_thread_deallocated_set
 tsd_thread_deallocatedp_get
 tsd_tls
 tsd_tsd
+tsd_tsdn
+tsd_witness_fork_get
+tsd_witness_fork_set
+tsd_witness_forkp_get
+tsd_witnesses_get
+tsd_witnesses_set
+tsd_witnessesp_get
+tsdn_fetch
+tsdn_null
+tsdn_tsd
 u2rz
 valgrind_freelike_block
 valgrind_make_mem_defined
 valgrind_make_mem_noaccess
 valgrind_make_mem_undefined
+witness_assert_depth
+witness_assert_depth_to_rank
+witness_assert_lockless
+witness_assert_not_owner
+witness_assert_owner
+witness_depth_error
+witness_fork_cleanup
+witness_init
+witness_lock
+witness_lock_error
+witness_not_owner_error
+witness_owner
+witness_owner_error
+witness_postfork_child
+witness_postfork_parent
+witness_prefork
+witness_unlock
+witnesses_cleanup
+zone_register
--- a/include/jemalloc/internal/prng.h
+++ b/include/jemalloc/internal/prng.h
@@ -19,8 +19,12 @@
 * the next has a cycle of 4, etc.  For this reason, we prefer to use the upper
 * bits.
 */
-#define	PRNG_A	UINT64_C(6364136223846793005)
-#define	PRNG_C	UINT64_C(1442695040888963407)
+
+#define	PRNG_A_32	UINT32_C(1103515241)
+#define	PRNG_C_32	UINT32_C(12347)
+
+#define	PRNG_A_64	UINT64_C(6364136223846793005)
+#define	PRNG_C_64	UINT64_C(1442695040888963407)

 #endif /* JEMALLOC_H_TYPES */
 /******************************************************************************/
@@ -35,28 +39,133 @@
 #ifdef JEMALLOC_H_INLINES

 #ifndef JEMALLOC_ENABLE_INLINE
-uint64_t	prng_lg_range(uint64_t *state, unsigned lg_range);
-uint64_t	prng_range(uint64_t *state, uint64_t range);
+uint32_t	prng_state_next_u32(uint32_t state);
+uint64_t	prng_state_next_u64(uint64_t state);
+size_t	prng_state_next_zu(size_t state);
+
+uint32_t	prng_lg_range_u32(uint32_t *state, unsigned lg_range,
+    bool atomic);
+uint64_t	prng_lg_range_u64(uint64_t *state, unsigned lg_range);
+size_t	prng_lg_range_zu(size_t *state, unsigned lg_range, bool atomic);
+
+uint32_t	prng_range_u32(uint32_t *state, uint32_t range, bool atomic);
+uint64_t	prng_range_u64(uint64_t *state, uint64_t range);
+size_t	prng_range_zu(size_t *state, size_t range, bool atomic);
 #endif

 #if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_PRNG_C_))
-JEMALLOC_ALWAYS_INLINE uint64_t
-prng_lg_range(uint64_t *state, unsigned lg_range)
+JEMALLOC_ALWAYS_INLINE uint32_t
+prng_state_next_u32(uint32_t state)
 {
-	uint64_t ret;
+
+	return ((state * PRNG_A_32) + PRNG_C_32);
+}
+
+JEMALLOC_ALWAYS_INLINE uint64_t
+prng_state_next_u64(uint64_t state)
+{
+
+	return ((state * PRNG_A_64) + PRNG_C_64);
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+prng_state_next_zu(size_t state)
+{
+
+#if LG_SIZEOF_PTR == 2
+	return ((state * PRNG_A_32) + PRNG_C_32);
+#elif LG_SIZEOF_PTR == 3
+	return ((state * PRNG_A_64) + PRNG_C_64);
+#else
+#error Unsupported pointer size
+#endif
+}
+
+JEMALLOC_ALWAYS_INLINE uint32_t
+prng_lg_range_u32(uint32_t *state, unsigned lg_range, bool atomic)
+{
+	uint32_t ret, state1;
+
+	assert(lg_range > 0);
+	assert(lg_range <= 32);
+
+	if (atomic) {
+		uint32_t state0;
+
+		do {
+			state0 = atomic_read_uint32(state);
+			state1 = prng_state_next_u32(state0);
+		} while (atomic_cas_uint32(state, state0, state1));
+	} else {
+		state1 = prng_state_next_u32(*state);
+		*state = state1;
+	}
+	ret = state1 >> (32 - lg_range);
+
+	return (ret);
+}
+
+/* 64-bit atomic operations cannot be supported on all relevant platforms. */
+JEMALLOC_ALWAYS_INLINE uint64_t
+prng_lg_range_u64(uint64_t *state, unsigned lg_range)
+{
+	uint64_t ret, state1;

 	assert(lg_range > 0);
 	assert(lg_range <= 64);

-	ret = (*state * PRNG_A) + PRNG_C;
-	*state = ret;
-	ret >>= (64 - lg_range);
+	state1 = prng_state_next_u64(*state);
+	*state = state1;
+	ret = state1 >> (64 - lg_range);
+
+	return (ret);
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+prng_lg_range_zu(size_t *state, unsigned lg_range, bool atomic)
+{
+	size_t ret, state1;
+
+	assert(lg_range > 0);
+	assert(lg_range <= ZU(1) << (3 + LG_SIZEOF_PTR));
+
+	if (atomic) {
+		size_t state0;
+
+		do {
+			state0 = atomic_read_z(state);
+			state1 = prng_state_next_zu(state0);
+		} while (atomic_cas_z(state, state0, state1));
+	} else {
+		state1 = prng_state_next_zu(*state);
+		*state = state1;
+	}
+	ret = state1 >> ((ZU(1) << (3 + LG_SIZEOF_PTR)) - lg_range);
+
+	return (ret);
+}
+
+JEMALLOC_ALWAYS_INLINE uint32_t
+prng_range_u32(uint32_t *state, uint32_t range, bool atomic)
+{
+	uint32_t ret;
+	unsigned lg_range;
+
+	assert(range > 1);
+
+	/* Compute the ceiling of lg(range). */
+	lg_range = ffs_u32(pow2_ceil_u32(range)) - 1;
+
+	/* Generate a result in [0..range) via repeated trial. */
+	do {
+		ret = prng_lg_range_u32(state, lg_range, atomic);
+	} while (ret >= range);

 	return (ret);
 }

 JEMALLOC_ALWAYS_INLINE uint64_t
-prng_range(uint64_t *state, uint64_t range)
+prng_range_u64(uint64_t *state, uint64_t range)
 {
 	uint64_t ret;
 	unsigned lg_range;
@@ -68,7 +177,26 @@ prng_range(uint64_t *state, uint64_t range)

 	/* Generate a result in [0..range) via repeated trial. */
 	do {
-		ret = prng_lg_range(state, lg_range);
+		ret = prng_lg_range_u64(state, lg_range);
+	} while (ret >= range);
+
+	return (ret);
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+prng_range_zu(size_t *state, size_t range, bool atomic)
+{
+	size_t ret;
+	unsigned lg_range;
+
+	assert(range > 1);
+
+	/* Compute the ceiling of lg(range). */
+	lg_range = ffs_u64(pow2_ceil_u64(range)) - 1;
+
+	/* Generate a result in [0..range) via repeated trial. */
+	do {
+		ret = prng_lg_range_zu(state, lg_range, atomic);
 	} while (ret >= range);

 	return (ret);
--- a/include/jemalloc/internal/prof.h
+++ b/include/jemalloc/internal/prof.h
@@ -281,7 +281,7 @@ extern uint64_t	prof_interval;
 extern size_t	lg_prof_sample;

 void	prof_alloc_rollback(tsd_t *tsd, prof_tctx_t *tctx, bool updated);
-void	prof_malloc_sample_object(const void *ptr, size_t usize,
+void	prof_malloc_sample_object(tsdn_t *tsdn, const void *ptr, size_t usize,
    prof_tctx_t *tctx);
 void	prof_free_sampled_object(tsd_t *tsd, size_t usize, prof_tctx_t *tctx);
 void	bt_init(prof_bt_t *bt, void **vec);
@@ -293,33 +293,33 @@ size_t	prof_bt_count(void);
 const prof_cnt_t *prof_cnt_all(void);
 typedef int (prof_dump_open_t)(bool, const char *);
 extern prof_dump_open_t *prof_dump_open;
-typedef bool (prof_dump_header_t)(bool, const prof_cnt_t *);
+typedef bool (prof_dump_header_t)(tsdn_t *, bool, const prof_cnt_t *);
 extern prof_dump_header_t *prof_dump_header;
 #endif
-void	prof_idump(void);
-bool	prof_mdump(const char *filename);
-void	prof_gdump(void);
+void	prof_idump(tsdn_t *tsdn);
+bool	prof_mdump(tsd_t *tsd, const char *filename);
+void	prof_gdump(tsdn_t *tsdn);
 prof_tdata_t	*prof_tdata_init(tsd_t *tsd);
 prof_tdata_t	*prof_tdata_reinit(tsd_t *tsd, prof_tdata_t *tdata);
 void	prof_reset(tsd_t *tsd, size_t lg_sample);
 void	prof_tdata_cleanup(tsd_t *tsd);
-const char	*prof_thread_name_get(void);
-bool	prof_active_get(void);
-bool	prof_active_set(bool active);
+bool	prof_active_get(tsdn_t *tsdn);
+bool	prof_active_set(tsdn_t *tsdn, bool active);
+const char	*prof_thread_name_get(tsd_t *tsd);
 int	prof_thread_name_set(tsd_t *tsd, const char *thread_name);
-bool	prof_thread_active_get(void);
-bool	prof_thread_active_set(bool active);
-bool	prof_thread_active_init_get(void);
-bool	prof_thread_active_init_set(bool active_init);
-bool	prof_gdump_get(void);
-bool	prof_gdump_set(bool active);
+bool	prof_thread_active_get(tsd_t *tsd);
+bool	prof_thread_active_set(tsd_t *tsd, bool active);
+bool	prof_thread_active_init_get(tsdn_t *tsdn);
+bool	prof_thread_active_init_set(tsdn_t *tsdn, bool active_init);
+bool	prof_gdump_get(tsdn_t *tsdn);
+bool	prof_gdump_set(tsdn_t *tsdn, bool active);
 void	prof_boot0(void);
 void	prof_boot1(void);
-bool	prof_boot2(void);
-void	prof_prefork0(void);
-void	prof_prefork1(void);
-void	prof_postfork_parent(void);
-void	prof_postfork_child(void);
+bool	prof_boot2(tsd_t *tsd);
+void	prof_prefork0(tsdn_t *tsdn);
+void	prof_prefork1(tsdn_t *tsdn);
+void	prof_postfork_parent(tsdn_t *tsdn);
+void	prof_postfork_child(tsdn_t *tsdn);
 void	prof_sample_threshold_update(prof_tdata_t *tdata);

 #endif /* JEMALLOC_H_EXTERNS */
@@ -330,17 +330,17 @@ void	prof_sample_threshold_update(prof_tdata_t *tdata);
 bool	prof_active_get_unlocked(void);
 bool	prof_gdump_get_unlocked(void);
 prof_tdata_t	*prof_tdata_get(tsd_t *tsd, bool create);
+prof_tctx_t	*prof_tctx_get(tsdn_t *tsdn, const void *ptr);
+void	prof_tctx_set(tsdn_t *tsdn, const void *ptr, size_t usize,
+    prof_tctx_t *tctx);
+void	prof_tctx_reset(tsdn_t *tsdn, const void *ptr, size_t usize,
+    const void *old_ptr, prof_tctx_t *tctx);
 bool	prof_sample_accum_update(tsd_t *tsd, size_t usize, bool commit,
    prof_tdata_t **tdata_out);
 prof_tctx_t	*prof_alloc_prep(tsd_t *tsd, size_t usize, bool prof_active,
    bool update);
-prof_tctx_t	*prof_tctx_get(const void *ptr);
-void	prof_tctx_set(const void *ptr, size_t usize, prof_tctx_t *tctx);
-void	prof_tctx_reset(const void *ptr, size_t usize, const void *old_ptr,
+void	prof_malloc(tsdn_t *tsdn, const void *ptr, size_t usize,
    prof_tctx_t *tctx);
-void	prof_malloc_sample_object(const void *ptr, size_t usize,
-    prof_tctx_t *tctx);
-void	prof_malloc(const void *ptr, size_t usize, prof_tctx_t *tctx);
 void	prof_realloc(tsd_t *tsd, const void *ptr, size_t usize,
    prof_tctx_t *tctx, bool prof_active, bool updated, const void *old_ptr,
    size_t old_usize, prof_tctx_t *old_tctx);
@@ -398,34 +398,34 @@ prof_tdata_get(tsd_t *tsd, bool create)
 }

 JEMALLOC_ALWAYS_INLINE prof_tctx_t *
-prof_tctx_get(const void *ptr)
+prof_tctx_get(tsdn_t *tsdn, const void *ptr)
 {

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

-	return (arena_prof_tctx_get(ptr));
+	return (arena_prof_tctx_get(tsdn, ptr));
 }

 JEMALLOC_ALWAYS_INLINE void
-prof_tctx_set(const void *ptr, size_t usize, prof_tctx_t *tctx)
+prof_tctx_set(tsdn_t *tsdn, const void *ptr, size_t usize, prof_tctx_t *tctx)
 {

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

-	arena_prof_tctx_set(ptr, usize, tctx);
+	arena_prof_tctx_set(tsdn, ptr, usize, tctx);
 }

 JEMALLOC_ALWAYS_INLINE void
-prof_tctx_reset(const void *ptr, size_t usize, const void *old_ptr,
+prof_tctx_reset(tsdn_t *tsdn, const void *ptr, size_t usize, const void *old_ptr,
    prof_tctx_t *old_tctx)
 {

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

-	arena_prof_tctx_reset(ptr, usize, old_ptr, old_tctx);
+	arena_prof_tctx_reset(tsdn, ptr, usize, old_ptr, old_tctx);
 }

 JEMALLOC_ALWAYS_INLINE bool
@@ -480,17 +480,17 @@ prof_alloc_prep(tsd_t *tsd, size_t usize, bool prof_active, bool update)
 }

 JEMALLOC_ALWAYS_INLINE void
-prof_malloc(const void *ptr, size_t usize, prof_tctx_t *tctx)
+prof_malloc(tsdn_t *tsdn, const void *ptr, size_t usize, prof_tctx_t *tctx)
 {

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

 	if (unlikely((uintptr_t)tctx > (uintptr_t)1U))
-		prof_malloc_sample_object(ptr, usize, tctx);
+		prof_malloc_sample_object(tsdn, ptr, usize, tctx);
 	else
-		prof_tctx_set(ptr, usize, (prof_tctx_t *)(uintptr_t)1U);
+		prof_tctx_set(tsdn, ptr, usize, (prof_tctx_t *)(uintptr_t)1U);
 }

 JEMALLOC_ALWAYS_INLINE void
@@ -504,7 +504,7 @@ prof_realloc(tsd_t *tsd, const void *ptr, size_t usize, prof_tctx_t *tctx,
 	assert(ptr != NULL || (uintptr_t)tctx <= (uintptr_t)1U);

 	if (prof_active && !updated && ptr != NULL) {
-		assert(usize == isalloc(ptr, true));
+		assert(usize == isalloc(tsd_tsdn(tsd), ptr, true));
 		if (prof_sample_accum_update(tsd, usize, true, NULL)) {
 			/*
 			 * Don't sample.  The usize passed to prof_alloc_prep()
@@ -513,6 +513,7 @@ prof_realloc(tsd_t *tsd, const void *ptr, size_t usize, prof_tctx_t *tctx,
 			 * though its actual usize was insufficient to cross the
 			 * sample threshold.
 			 */
+			prof_alloc_rollback(tsd, tctx, true);
 			tctx = (prof_tctx_t *)(uintptr_t)1U;
 		}
 	}
@@ -521,9 +522,9 @@ prof_realloc(tsd_t *tsd, const void *ptr, size_t usize, prof_tctx_t *tctx,
 	old_sampled = ((uintptr_t)old_tctx > (uintptr_t)1U);

 	if (unlikely(sampled))
-		prof_malloc_sample_object(ptr, usize, tctx);
+		prof_malloc_sample_object(tsd_tsdn(tsd), ptr, usize, tctx);
 	else
-		prof_tctx_reset(ptr, usize, old_ptr, old_tctx);
+		prof_tctx_reset(tsd_tsdn(tsd), ptr, usize, old_ptr, old_tctx);

 	if (unlikely(old_sampled))
 		prof_free_sampled_object(tsd, old_usize, old_tctx);
@@ -532,10 +533,10 @@ prof_realloc(tsd_t *tsd, const void *ptr, size_t usize, prof_tctx_t *tctx,
 JEMALLOC_ALWAYS_INLINE void
 prof_free(tsd_t *tsd, const void *ptr, size_t usize)
 {
-	prof_tctx_t *tctx = prof_tctx_get(ptr);
+	prof_tctx_t *tctx = prof_tctx_get(tsd_tsdn(tsd), ptr);

 	cassert(config_prof);
-	assert(usize == isalloc(ptr, true));
+	assert(usize == isalloc(tsd_tsdn(tsd), ptr, true));

 	if (unlikely((uintptr_t)tctx > (uintptr_t)1U))
 		prof_free_sampled_object(tsd, usize, tctx);
--- a/include/jemalloc/internal/rtree.h
+++ b/include/jemalloc/internal/rtree.h
@@ -15,9 +15,10 @@ typedef struct rtree_s rtree_t;
 * machine address width.
 */
 #define	LG_RTREE_BITS_PER_LEVEL	4
-#define	RTREE_BITS_PER_LEVEL	(ZU(1) << LG_RTREE_BITS_PER_LEVEL)
+#define	RTREE_BITS_PER_LEVEL	(1U << LG_RTREE_BITS_PER_LEVEL)
+/* Maximum rtree height. */
 #define	RTREE_HEIGHT_MAX						\
-    ((ZU(1) << (LG_SIZEOF_PTR+3)) / RTREE_BITS_PER_LEVEL)
+    ((1U << (LG_SIZEOF_PTR+3)) / RTREE_BITS_PER_LEVEL)

 /* Used for two-stage lock-free node initialization. */
 #define	RTREE_NODE_INITIALIZING	((rtree_node_elm_t *)0x1)
@@ -111,22 +112,25 @@ unsigned	rtree_start_level(rtree_t *rtree, uintptr_t key);
 uintptr_t	rtree_subkey(rtree_t *rtree, uintptr_t key, unsigned level);

 bool	rtree_node_valid(rtree_node_elm_t *node);
-rtree_node_elm_t	*rtree_child_tryread(rtree_node_elm_t *elm);
+rtree_node_elm_t	*rtree_child_tryread(rtree_node_elm_t *elm,
+    bool dependent);
 rtree_node_elm_t	*rtree_child_read(rtree_t *rtree, rtree_node_elm_t *elm,
-    unsigned level);
+    unsigned level, bool dependent);
 extent_node_t	*rtree_val_read(rtree_t *rtree, rtree_node_elm_t *elm,
    bool dependent);
 void	rtree_val_write(rtree_t *rtree, rtree_node_elm_t *elm,
    const extent_node_t *val);
-rtree_node_elm_t	*rtree_subtree_tryread(rtree_t *rtree, unsigned level);
-rtree_node_elm_t	*rtree_subtree_read(rtree_t *rtree, unsigned level);
+rtree_node_elm_t	*rtree_subtree_tryread(rtree_t *rtree, unsigned level,
+    bool dependent);
+rtree_node_elm_t	*rtree_subtree_read(rtree_t *rtree, unsigned level,
+    bool dependent);

 extent_node_t	*rtree_get(rtree_t *rtree, uintptr_t key, bool dependent);
 bool	rtree_set(rtree_t *rtree, uintptr_t key, const extent_node_t *val);
 #endif

 #if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_RTREE_C_))
-JEMALLOC_INLINE unsigned
+JEMALLOC_ALWAYS_INLINE unsigned
 rtree_start_level(rtree_t *rtree, uintptr_t key)
 {
 	unsigned start_level;
@@ -140,7 +144,7 @@ rtree_start_level(rtree_t *rtree, uintptr_t key)
 	return (start_level);
 }

-JEMALLOC_INLINE uintptr_t
+JEMALLOC_ALWAYS_INLINE uintptr_t
 rtree_subkey(rtree_t *rtree, uintptr_t key, unsigned level)
 {

@@ -149,37 +153,40 @@ rtree_subkey(rtree_t *rtree, uintptr_t key, unsigned level)
 	    rtree->levels[level].bits) - 1));
 }

-JEMALLOC_INLINE bool
+JEMALLOC_ALWAYS_INLINE bool
 rtree_node_valid(rtree_node_elm_t *node)
 {

 	return ((uintptr_t)node > (uintptr_t)RTREE_NODE_INITIALIZING);
 }

-JEMALLOC_INLINE rtree_node_elm_t *
-rtree_child_tryread(rtree_node_elm_t *elm)
+JEMALLOC_ALWAYS_INLINE rtree_node_elm_t *
+rtree_child_tryread(rtree_node_elm_t *elm, bool dependent)
 {
 	rtree_node_elm_t *child;

 	/* Double-checked read (first read may be stale. */
 	child = elm->child;
-	if (!rtree_node_valid(child))
+	if (!dependent && !rtree_node_valid(child))
 		child = atomic_read_p(&elm->pun);
+	assert(!dependent || child != NULL);
 	return (child);
 }

-JEMALLOC_INLINE rtree_node_elm_t *
-rtree_child_read(rtree_t *rtree, rtree_node_elm_t *elm, unsigned level)
+JEMALLOC_ALWAYS_INLINE rtree_node_elm_t *
+rtree_child_read(rtree_t *rtree, rtree_node_elm_t *elm, unsigned level,
+    bool dependent)
 {
 	rtree_node_elm_t *child;

-	child = rtree_child_tryread(elm);
-	if (unlikely(!rtree_node_valid(child)))
+	child = rtree_child_tryread(elm, dependent);
+	if (!dependent && unlikely(!rtree_node_valid(child)))
 		child = rtree_child_read_hard(rtree, elm, level);
+	assert(!dependent || child != NULL);
 	return (child);
 }

-JEMALLOC_INLINE extent_node_t *
+JEMALLOC_ALWAYS_INLINE extent_node_t *
 rtree_val_read(rtree_t *rtree, rtree_node_elm_t *elm, bool dependent)
 {

@@ -208,54 +215,119 @@ rtree_val_write(rtree_t *rtree, rtree_node_elm_t *elm, const extent_node_t *val)
 	atomic_write_p(&elm->pun, val);
 }

-JEMALLOC_INLINE rtree_node_elm_t *
-rtree_subtree_tryread(rtree_t *rtree, unsigned level)
+JEMALLOC_ALWAYS_INLINE rtree_node_elm_t *
+rtree_subtree_tryread(rtree_t *rtree, unsigned level, bool dependent)
 {
 	rtree_node_elm_t *subtree;

 	/* Double-checked read (first read may be stale. */
 	subtree = rtree->levels[level].subtree;
-	if (!rtree_node_valid(subtree))
+	if (!dependent && unlikely(!rtree_node_valid(subtree)))
 		subtree = atomic_read_p(&rtree->levels[level].subtree_pun);
+	assert(!dependent || subtree != NULL);
 	return (subtree);
 }

-JEMALLOC_INLINE rtree_node_elm_t *
-rtree_subtree_read(rtree_t *rtree, unsigned level)
+JEMALLOC_ALWAYS_INLINE rtree_node_elm_t *
+rtree_subtree_read(rtree_t *rtree, unsigned level, bool dependent)
 {
 	rtree_node_elm_t *subtree;

-	subtree = rtree_subtree_tryread(rtree, level);
-	if (unlikely(!rtree_node_valid(subtree)))
+	subtree = rtree_subtree_tryread(rtree, level, dependent);
+	if (!dependent && unlikely(!rtree_node_valid(subtree)))
 		subtree = rtree_subtree_read_hard(rtree, level);
+	assert(!dependent || subtree != NULL);
 	return (subtree);
 }

-JEMALLOC_INLINE extent_node_t *
+JEMALLOC_ALWAYS_INLINE extent_node_t *
 rtree_get(rtree_t *rtree, uintptr_t key, bool dependent)
 {
 	uintptr_t subkey;
-	unsigned i, start_level;
-	rtree_node_elm_t *node, *child;
+	unsigned start_level;
+	rtree_node_elm_t *node;

 	start_level = rtree_start_level(rtree, key);

-	for (i = start_level, node = rtree_subtree_tryread(rtree, start_level);
-	    /**/; i++, node = child) {
-		if (!dependent && unlikely(!rtree_node_valid(node)))
-			return (NULL);
-		subkey = rtree_subkey(rtree, key, i);
-		if (i == rtree->height - 1) {
-			/*
-			 * node is a leaf, so it contains values rather than
-			 * child pointers.
-			 */
-			return (rtree_val_read(rtree, &node[subkey],
-			    dependent));
-		}
-		assert(i < rtree->height - 1);
-		child = rtree_child_tryread(&node[subkey]);
+	node = rtree_subtree_tryread(rtree, start_level, dependent);
+#define	RTREE_GET_BIAS	(RTREE_HEIGHT_MAX - rtree->height)
+	switch (start_level + RTREE_GET_BIAS) {
+#define	RTREE_GET_SUBTREE(level)					\
+	case level:							\
+		assert(level < (RTREE_HEIGHT_MAX-1));			\
+		if (!dependent && unlikely(!rtree_node_valid(node)))	\
+			return (NULL);					\
+		subkey = rtree_subkey(rtree, key, level -		\
+		    RTREE_GET_BIAS);					\
+		node = rtree_child_tryread(&node[subkey], dependent);	\
+		/* Fall through. */
+#define	RTREE_GET_LEAF(level)						\
+	case level:							\
+		assert(level == (RTREE_HEIGHT_MAX-1));			\
+		if (!dependent && unlikely(!rtree_node_valid(node)))	\
+			return (NULL);					\
+		subkey = rtree_subkey(rtree, key, level -		\
+		    RTREE_GET_BIAS);					\
+		/*							\
+		 * node is a leaf, so it contains values rather than	\
+		 * child pointers.					\
+		 */							\
+		return (rtree_val_read(rtree, &node[subkey],		\
+		    dependent));
+#if RTREE_HEIGHT_MAX > 1
+	RTREE_GET_SUBTREE(0)
+#endif
+#if RTREE_HEIGHT_MAX > 2
+	RTREE_GET_SUBTREE(1)
+#endif
+#if RTREE_HEIGHT_MAX > 3
+	RTREE_GET_SUBTREE(2)
+#endif
+#if RTREE_HEIGHT_MAX > 4
+	RTREE_GET_SUBTREE(3)
+#endif
+#if RTREE_HEIGHT_MAX > 5
+	RTREE_GET_SUBTREE(4)
+#endif
+#if RTREE_HEIGHT_MAX > 6
+	RTREE_GET_SUBTREE(5)
+#endif
+#if RTREE_HEIGHT_MAX > 7
+	RTREE_GET_SUBTREE(6)
+#endif
+#if RTREE_HEIGHT_MAX > 8
+	RTREE_GET_SUBTREE(7)
+#endif
+#if RTREE_HEIGHT_MAX > 9
+	RTREE_GET_SUBTREE(8)
+#endif
+#if RTREE_HEIGHT_MAX > 10
+	RTREE_GET_SUBTREE(9)
+#endif
+#if RTREE_HEIGHT_MAX > 11
+	RTREE_GET_SUBTREE(10)
+#endif
+#if RTREE_HEIGHT_MAX > 12
+	RTREE_GET_SUBTREE(11)
+#endif
+#if RTREE_HEIGHT_MAX > 13
+	RTREE_GET_SUBTREE(12)
+#endif
+#if RTREE_HEIGHT_MAX > 14
+	RTREE_GET_SUBTREE(13)
+#endif
+#if RTREE_HEIGHT_MAX > 15
+	RTREE_GET_SUBTREE(14)
+#endif
+#if RTREE_HEIGHT_MAX > 16
+#  error Unsupported RTREE_HEIGHT_MAX
+#endif
+	RTREE_GET_LEAF(RTREE_HEIGHT_MAX-1)
+#undef RTREE_GET_SUBTREE
+#undef RTREE_GET_LEAF
+	default: not_reached();
 	}
+#undef RTREE_GET_BIAS
 	not_reached();
 }

@@ -268,7 +340,7 @@ rtree_set(rtree_t *rtree, uintptr_t key, const extent_node_t *val)

 	start_level = rtree_start_level(rtree, key);

-	node = rtree_subtree_read(rtree, start_level);
+	node = rtree_subtree_read(rtree, start_level, false);
 	if (node == NULL)
 		return (true);
 	for (i = start_level; /**/; i++, node = child) {
@@ -282,7 +354,7 @@ rtree_set(rtree_t *rtree, uintptr_t key, const extent_node_t *val)
 			return (false);
 		}
 		assert(i + 1 < rtree->height);
-		child = rtree_child_read(rtree, &node[subkey], i);
+		child = rtree_child_read(rtree, &node[subkey], i, false);
 		if (child == NULL)
 			return (true);
 	}
--- a/include/jemalloc/internal/size_classes.sh
+++ b/include/jemalloc/internal/size_classes.sh
@@ -48,6 +48,21 @@ size_class() {
  lg_p=$5
  lg_kmax=$6

+  if [ ${lg_delta} -ge ${lg_p} ] ; then
+    psz="yes"
+  else
+    pow2 ${lg_p}; p=${pow2_result}
+    pow2 ${lg_grp}; grp=${pow2_result}
+    pow2 ${lg_delta}; delta=${pow2_result}
+    sz=$((${grp} + ${delta} * ${ndelta}))
+    npgs=$((${sz} / ${p}))
+    if [ ${sz} -eq $((${npgs} * ${p})) ] ; then
+      psz="yes"
+    else
+      psz="no"
+    fi
+  fi
+
  lg ${ndelta}; lg_ndelta=${lg_result}; pow2 ${lg_ndelta}
  if [ ${pow2_result} -lt ${ndelta} ] ; then
    rem="yes"
@@ -74,14 +89,15 @@ size_class() {
  else
    lg_delta_lookup="no"
  fi
-  printf '    SC(%3d, %6d, %8d, %6d, %3s, %2s) \\\n' ${index} ${lg_grp} ${lg_delta} ${ndelta} ${bin} ${lg_delta_lookup}
+  printf '    SC(%3d, %6d, %8d, %6d, %3s, %3s, %2s) \\\n' ${index} ${lg_grp} ${lg_delta} ${ndelta} ${psz} ${bin} ${lg_delta_lookup}
  # Defined upon return:
-  # - lg_delta_lookup (${lg_delta} or "no")
+  # - psz ("yes" or "no")
  # - bin ("yes" or "no")
+  # - lg_delta_lookup (${lg_delta} or "no")
 }

 sep_line() {
-  echo "                                               \\"
+  echo "                                                         \\"
 }

 size_classes() {
@@ -95,12 +111,13 @@ size_classes() {
  pow2 ${lg_g}; g=${pow2_result}

  echo "#define	SIZE_CLASSES \\"
-  echo "  /* index, lg_grp, lg_delta, ndelta, bin, lg_delta_lookup */ \\"
+  echo "  /* index, lg_grp, lg_delta, ndelta, psz, bin, lg_delta_lookup */ \\"

  ntbins=0
  nlbins=0
  lg_tiny_maxclass='"NA"'
  nbins=0
+  npsizes=0

  # Tiny size classes.
  ndelta=0
@@ -112,6 +129,9 @@ size_classes() {
    if [ ${lg_delta_lookup} != "no" ] ; then
      nlbins=$((${index} + 1))
    fi
+    if [ ${psz} = "yes" ] ; then
+      npsizes=$((${npsizes} + 1))
+    fi
    if [ ${bin} != "no" ] ; then
      nbins=$((${index} + 1))
    fi
@@ -133,11 +153,17 @@ size_classes() {
    index=$((${index} + 1))
    lg_grp=$((${lg_grp} + 1))
    lg_delta=$((${lg_delta} + 1))
+    if [ ${psz} = "yes" ] ; then
+      npsizes=$((${npsizes} + 1))
+    fi
  fi
  while [ ${ndelta} -lt ${g} ] ; do
    size_class ${index} ${lg_grp} ${lg_delta} ${ndelta} ${lg_p} ${lg_kmax}
    index=$((${index} + 1))
    ndelta=$((${ndelta} + 1))
+    if [ ${psz} = "yes" ] ; then
+      npsizes=$((${npsizes} + 1))
+    fi
  done

  # All remaining groups.
@@ -157,6 +183,9 @@ size_classes() {
        # Final written value is correct:
        lookup_maxclass="((((size_t)1) << ${lg_grp}) + (((size_t)${ndelta}) << ${lg_delta}))"
      fi
+      if [ ${psz} = "yes" ] ; then
+        npsizes=$((${npsizes} + 1))
+      fi
      if [ ${bin} != "no" ] ; then
        nbins=$((${index} + 1))
        # Final written value is correct:
@@ -183,6 +212,7 @@ size_classes() {
  # - nlbins
  # - nbins
  # - nsizes
+  # - npsizes
  # - lg_tiny_maxclass
  # - lookup_maxclass
  # - small_maxclass
@@ -200,13 +230,13 @@ cat <<EOF
 * be defined prior to inclusion, and it in turn defines:
 *
 *   LG_SIZE_CLASS_GROUP: Lg of size class count for each size doubling.
- *   SIZE_CLASSES: Complete table of
- *                 SC(index, lg_grp, lg_delta, ndelta, bin, lg_delta_lookup)
- *                 tuples.
+ *   SIZE_CLASSES: Complete table of SC(index, lg_grp, lg_delta, ndelta, psz,
+ *                 bin, lg_delta_lookup) tuples.
 *     index: Size class index.
 *     lg_grp: Lg group base size (no deltas added).
 *     lg_delta: Lg delta to previous size class.
 *     ndelta: Delta multiplier.  size == 1<<lg_grp + ndelta<<lg_delta
+ *     psz: 'yes' if a multiple of the page size, 'no' otherwise.
 *     bin: 'yes' if a small bin size class, 'no' otherwise.
 *     lg_delta_lookup: Same as lg_delta if a lookup table size class, 'no'
 *                      otherwise.
@@ -214,6 +244,7 @@ cat <<EOF
 *   NLBINS: Number of bins supported by the lookup table.
 *   NBINS: Number of small size class bins.
 *   NSIZES: Number of size classes.
+ *   NPSIZES: Number of size classes that are a multiple of (1U << LG_PAGE).
 *   LG_TINY_MAXCLASS: Lg of maximum tiny size class.
 *   LOOKUP_MAXCLASS: Maximum size class included in lookup table.
 *   SMALL_MAXCLASS: Maximum small size class.
@@ -238,6 +269,7 @@ for lg_z in ${lg_zarr} ; do
        echo "#define	NLBINS			${nlbins}"
        echo "#define	NBINS			${nbins}"
        echo "#define	NSIZES			${nsizes}"
+        echo "#define	NPSIZES			${npsizes}"
        echo "#define	LG_TINY_MAXCLASS	${lg_tiny_maxclass}"
        echo "#define	LOOKUP_MAXCLASS		${lookup_maxclass}"
        echo "#define	SMALL_MAXCLASS		${small_maxclass}"
--- a/include/jemalloc/internal/spin.h
+++ b/include/jemalloc/internal/spin.h
@@ -0,0 +1,51 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+typedef struct spin_s spin_t;
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+struct spin_s {
+	unsigned iteration;
+};
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#ifndef JEMALLOC_ENABLE_INLINE
+void	spin_init(spin_t *spin);
+void	spin_adaptive(spin_t *spin);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_SPIN_C_))
+JEMALLOC_INLINE void
+spin_init(spin_t *spin)
+{
+
+	spin->iteration = 0;
+}
+
+JEMALLOC_INLINE void
+spin_adaptive(spin_t *spin)
+{
+	volatile uint64_t i;
+
+	for (i = 0; i < (KQU(1) << spin->iteration); i++)
+		CPU_SPINWAIT;
+
+	if (spin->iteration < 63)
+		spin->iteration++;
+}
+
+#endif
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
+
--- a/include/jemalloc/internal/stats.h
+++ b/include/jemalloc/internal/stats.h
@@ -102,6 +102,14 @@ struct arena_stats_s {
 	/* Number of bytes currently mapped. */
 	size_t		mapped;

+	/*
+	 * Number of bytes currently retained as a side effect of munmap() being
+	 * disabled/bypassed.  Retained bytes are technically mapped (though
+	 * always decommitted or purged), but they are excluded from the mapped
+	 * statistic (above).
+	 */
+	size_t		retained;
+
 	/*
 	 * Total number of purge sweeps, total number of madvise calls made,
 	 * and total pages purged in order to keep dirty unused memory under
@@ -167,25 +175,21 @@ stats_cactive_get(void)
 JEMALLOC_INLINE void
 stats_cactive_add(size_t size)
 {
-	UNUSED size_t cactive;

 	assert(size > 0);
 	assert((size & chunksize_mask) == 0);

-	cactive = atomic_add_z(&stats_cactive, size);
-	assert(cactive - size < cactive);
+	atomic_add_z(&stats_cactive, size);
 }

 JEMALLOC_INLINE void
 stats_cactive_sub(size_t size)
 {
-	UNUSED size_t cactive;

 	assert(size > 0);
 	assert((size & chunksize_mask) == 0);

-	cactive = atomic_sub_z(&stats_cactive, size);
-	assert(cactive + size > cactive);
+	atomic_sub_z(&stats_cactive, size);
 }
 #endif

--- a/include/jemalloc/internal/tcache.h
+++ b/include/jemalloc/internal/tcache.h
@@ -130,27 +130,28 @@ extern size_t	tcache_maxclass;
 */
 extern tcaches_t	*tcaches;

-size_t	tcache_salloc(const void *ptr);
+size_t	tcache_salloc(tsdn_t *tsdn, const void *ptr);
 void	tcache_event_hard(tsd_t *tsd, tcache_t *tcache);
-void	*tcache_alloc_small_hard(tsd_t *tsd, arena_t *arena, tcache_t *tcache,
+void	*tcache_alloc_small_hard(tsdn_t *tsdn, arena_t *arena, tcache_t *tcache,
    tcache_bin_t *tbin, szind_t binind, bool *tcache_success);
 void	tcache_bin_flush_small(tsd_t *tsd, tcache_t *tcache, tcache_bin_t *tbin,
    szind_t binind, unsigned rem);
 void	tcache_bin_flush_large(tsd_t *tsd, tcache_bin_t *tbin, szind_t binind,
    unsigned rem, tcache_t *tcache);
-void	tcache_arena_associate(tcache_t *tcache, arena_t *arena);
-void	tcache_arena_reassociate(tcache_t *tcache, arena_t *oldarena,
-    arena_t *newarena);
-void	tcache_arena_dissociate(tcache_t *tcache, arena_t *arena);
+void	tcache_arena_reassociate(tsdn_t *tsdn, tcache_t *tcache,
+    arena_t *oldarena, arena_t *newarena);
 tcache_t *tcache_get_hard(tsd_t *tsd);
-tcache_t *tcache_create(tsd_t *tsd, arena_t *arena);
+tcache_t *tcache_create(tsdn_t *tsdn, arena_t *arena);
 void	tcache_cleanup(tsd_t *tsd);
 void	tcache_enabled_cleanup(tsd_t *tsd);
-void	tcache_stats_merge(tcache_t *tcache, arena_t *arena);
+void	tcache_stats_merge(tsdn_t *tsdn, tcache_t *tcache, arena_t *arena);
 bool	tcaches_create(tsd_t *tsd, unsigned *r_ind);
 void	tcaches_flush(tsd_t *tsd, unsigned ind);
 void	tcaches_destroy(tsd_t *tsd, unsigned ind);
-bool	tcache_boot(void);
+bool	tcache_boot(tsdn_t *tsdn);
+void tcache_prefork(tsdn_t *tsdn);
+void tcache_postfork_parent(tsdn_t *tsdn);
+void tcache_postfork_child(tsdn_t *tsdn);

 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
@@ -297,8 +298,8 @@ tcache_alloc_small(tsd_t *tsd, arena_t *arena, tcache_t *tcache, size_t size,
 		if (unlikely(arena == NULL))
 			return (NULL);

-		ret = tcache_alloc_small_hard(tsd, arena, tcache, tbin, binind,
-			&tcache_hard_success);
+		ret = tcache_alloc_small_hard(tsd_tsdn(tsd), arena, tcache,
+		    tbin, binind, &tcache_hard_success);
 		if (tcache_hard_success == false)
 			return (NULL);
 	}
@@ -310,7 +311,7 @@ tcache_alloc_small(tsd_t *tsd, arena_t *arena, tcache_t *tcache, size_t size,
 	 */
 	if (config_prof || (slow_path && config_fill) || unlikely(zero)) {
 		usize = index2size(binind);
-		assert(tcache_salloc(ret) == usize);
+		assert(tcache_salloc(tsd_tsdn(tsd), ret) == usize);
 	}

 	if (likely(!zero)) {
@@ -358,7 +359,7 @@ tcache_alloc_large(tsd_t *tsd, arena_t *arena, tcache_t *tcache, size_t size,
 		if (unlikely(arena == NULL))
 			return (NULL);

-		ret = arena_malloc_large(tsd, arena, binind, zero);
+		ret = arena_malloc_large(tsd_tsdn(tsd), arena, binind, zero);
 		if (ret == NULL)
 			return (NULL);
 	} else {
@@ -381,9 +382,10 @@ tcache_alloc_large(tsd_t *tsd, arena_t *arena, tcache_t *tcache, size_t size,
 		}
 		if (likely(!zero)) {
 			if (slow_path && config_fill) {
-				if (unlikely(opt_junk_alloc))
-					memset(ret, 0xa5, usize);
-				else if (unlikely(opt_zero))
+				if (unlikely(opt_junk_alloc)) {
+					memset(ret, JEMALLOC_ALLOC_JUNK,
+					    usize);
+				} else if (unlikely(opt_zero))
 					memset(ret, 0, usize);
 			}
 		} else
@@ -406,7 +408,7 @@ tcache_dalloc_small(tsd_t *tsd, tcache_t *tcache, void *ptr, szind_t binind,
 	tcache_bin_t *tbin;
 	tcache_bin_info_t *tbin_info;

-	assert(tcache_salloc(ptr) <= SMALL_MAXCLASS);
+	assert(tcache_salloc(tsd_tsdn(tsd), ptr) <= SMALL_MAXCLASS);

 	if (slow_path && config_fill && unlikely(opt_junk_free))
 		arena_dalloc_junk_small(ptr, &arena_bin_info[binind]);
@@ -433,8 +435,8 @@ tcache_dalloc_large(tsd_t *tsd, tcache_t *tcache, void *ptr, size_t size,
 	tcache_bin_info_t *tbin_info;

 	assert((size & PAGE_MASK) == 0);
-	assert(tcache_salloc(ptr) > SMALL_MAXCLASS);
-	assert(tcache_salloc(ptr) <= tcache_maxclass);
+	assert(tcache_salloc(tsd_tsdn(tsd), ptr) > SMALL_MAXCLASS);
+	assert(tcache_salloc(tsd_tsdn(tsd), ptr) <= tcache_maxclass);

 	binind = size2index(size);

@@ -458,8 +460,10 @@ JEMALLOC_ALWAYS_INLINE tcache_t *
 tcaches_get(tsd_t *tsd, unsigned ind)
 {
 	tcaches_t *elm = &tcaches[ind];
-	if (unlikely(elm->tcache == NULL))
-		elm->tcache = tcache_create(tsd, arena_choose(tsd, NULL));
+	if (unlikely(elm->tcache == NULL)) {
+		elm->tcache = tcache_create(tsd_tsdn(tsd), arena_choose(tsd,
+		    NULL));
+	}
 	return (elm->tcache);
 }
 #endif
--- a/include/jemalloc/internal/tsd.h
+++ b/include/jemalloc/internal/tsd.h
@@ -13,6 +13,9 @@ typedef struct tsd_init_head_s tsd_init_head_t;
 #endif

 typedef struct tsd_s tsd_t;
+typedef struct tsdn_s tsdn_t;
+
+#define	TSDN_NULL	((tsdn_t *)0)

 typedef enum {
 	tsd_state_uninitialized,
@@ -44,7 +47,8 @@ typedef enum {
 * The result is a set of generated functions, e.g.:
 *
 *   bool example_tsd_boot(void) {...}
- *   example_t *example_tsd_get() {...}
+ *   bool example_tsd_booted_get(void) {...}
+ *   example_t *example_tsd_get(bool init) {...}
 *   void example_tsd_set(example_t *val) {...}
 *
 * Note that all of the functions deal in terms of (a_type *) rather than
@@ -98,8 +102,10 @@ a_attr void								\
 a_name##tsd_boot1(void);						\
 a_attr bool								\
 a_name##tsd_boot(void);							\
+a_attr bool								\
+a_name##tsd_booted_get(void);						\
 a_attr a_type *								\
-a_name##tsd_get(void);							\
+a_name##tsd_get(bool init);						\
 a_attr void								\
 a_name##tsd_set(a_type *val);

@@ -201,9 +207,21 @@ a_name##tsd_boot(void)							\
 									\
 	return (a_name##tsd_boot0());					\
 }									\
+a_attr bool								\
+a_name##tsd_booted_get(void)						\
+{									\
+									\
+	return (a_name##tsd_booted);					\
+}									\
+a_attr bool								\
+a_name##tsd_get_allocates(void)						\
+{									\
+									\
+	return (false);							\
+}									\
 /* Get/set. */								\
 a_attr a_type *								\
-a_name##tsd_get(void)							\
+a_name##tsd_get(bool init)						\
 {									\
 									\
 	assert(a_name##tsd_booted);					\
@@ -246,9 +264,21 @@ a_name##tsd_boot(void)							\
 									\
 	return (a_name##tsd_boot0());					\
 }									\
+a_attr bool								\
+a_name##tsd_booted_get(void)						\
+{									\
+									\
+	return (a_name##tsd_booted);					\
+}									\
+a_attr bool								\
+a_name##tsd_get_allocates(void)						\
+{									\
+									\
+	return (false);							\
+}									\
 /* Get/set. */								\
 a_attr a_type *								\
-a_name##tsd_get(void)							\
+a_name##tsd_get(bool init)						\
 {									\
 									\
 	assert(a_name##tsd_booted);					\
@@ -307,14 +337,14 @@ a_name##tsd_wrapper_set(a_name##tsd_wrapper_t *wrapper)			\
 	}								\
 }									\
 a_attr a_name##tsd_wrapper_t *						\
-a_name##tsd_wrapper_get(void)						\
+a_name##tsd_wrapper_get(bool init)					\
 {									\
 	DWORD error = GetLastError();					\
 	a_name##tsd_wrapper_t *wrapper = (a_name##tsd_wrapper_t *)	\
 	    TlsGetValue(a_name##tsd_tsd);				\
 	SetLastError(error);						\
 									\
-	if (unlikely(wrapper == NULL)) {				\
+	if (init && unlikely(wrapper == NULL)) {			\
 		wrapper = (a_name##tsd_wrapper_t *)			\
 		    malloc_tsd_malloc(sizeof(a_name##tsd_wrapper_t));	\
 		if (wrapper == NULL) {					\
@@ -368,14 +398,28 @@ a_name##tsd_boot(void)							\
 	a_name##tsd_boot1();						\
 	return (false);							\
 }									\
+a_attr bool								\
+a_name##tsd_booted_get(void)						\
+{									\
+									\
+	return (a_name##tsd_booted);					\
+}									\
+a_attr bool								\
+a_name##tsd_get_allocates(void)						\
+{									\
+									\
+	return (true);							\
+}									\
 /* Get/set. */								\
 a_attr a_type *								\
-a_name##tsd_get(void)							\
+a_name##tsd_get(bool init)						\
 {									\
 	a_name##tsd_wrapper_t *wrapper;					\
 									\
 	assert(a_name##tsd_booted);					\
-	wrapper = a_name##tsd_wrapper_get();				\
+	wrapper = a_name##tsd_wrapper_get(init);			\
+	if (a_name##tsd_get_allocates() && !init && wrapper == NULL)	\
+		return (NULL);						\
 	return (&wrapper->val);						\
 }									\
 a_attr void								\
@@ -384,7 +428,7 @@ a_name##tsd_set(a_type *val)						\
 	a_name##tsd_wrapper_t *wrapper;					\
 									\
 	assert(a_name##tsd_booted);					\
-	wrapper = a_name##tsd_wrapper_get();				\
+	wrapper = a_name##tsd_wrapper_get(true);			\
 	wrapper->val = *(val);						\
 	if (a_cleanup != malloc_tsd_no_cleanup)				\
 		wrapper->initialized = true;				\
@@ -428,20 +472,21 @@ a_name##tsd_wrapper_set(a_name##tsd_wrapper_t *wrapper)			\
 	}								\
 }									\
 a_attr a_name##tsd_wrapper_t *						\
-a_name##tsd_wrapper_get(void)						\
+a_name##tsd_wrapper_get(bool init)					\
 {									\
 	a_name##tsd_wrapper_t *wrapper = (a_name##tsd_wrapper_t *)	\
 	    pthread_getspecific(a_name##tsd_tsd);			\
 									\
-	if (unlikely(wrapper == NULL)) {				\
+	if (init && unlikely(wrapper == NULL)) {			\
 		tsd_init_block_t block;					\
-		wrapper = tsd_init_check_recursion(			\
-		    &a_name##tsd_init_head, &block);			\
+		wrapper = (a_name##tsd_wrapper_t *)			\
+		    tsd_init_check_recursion(&a_name##tsd_init_head,	\
+		    &block);						\
 		if (wrapper)						\
 		    return (wrapper);					\
 		wrapper = (a_name##tsd_wrapper_t *)			\
 		    malloc_tsd_malloc(sizeof(a_name##tsd_wrapper_t));	\
-		block.data = wrapper;					\
+		block.data = (void *)wrapper;				\
 		if (wrapper == NULL) {					\
 			malloc_write("<jemalloc>: Error allocating"	\
 			    " TSD for "#a_name"\n");			\
@@ -490,14 +535,28 @@ a_name##tsd_boot(void)							\
 	a_name##tsd_boot1();						\
 	return (false);							\
 }									\
+a_attr bool								\
+a_name##tsd_booted_get(void)						\
+{									\
+									\
+	return (a_name##tsd_booted);					\
+}									\
+a_attr bool								\
+a_name##tsd_get_allocates(void)						\
+{									\
+									\
+	return (true);							\
+}									\
 /* Get/set. */								\
 a_attr a_type *								\
-a_name##tsd_get(void)							\
+a_name##tsd_get(bool init)						\
 {									\
 	a_name##tsd_wrapper_t *wrapper;					\
 									\
 	assert(a_name##tsd_booted);					\
-	wrapper = a_name##tsd_wrapper_get();				\
+	wrapper = a_name##tsd_wrapper_get(init);			\
+	if (a_name##tsd_get_allocates() && !init && wrapper == NULL)	\
+		return (NULL);						\
 	return (&wrapper->val);						\
 }									\
 a_attr void								\
@@ -506,7 +565,7 @@ a_name##tsd_set(a_type *val)						\
 	a_name##tsd_wrapper_t *wrapper;					\
 									\
 	assert(a_name##tsd_booted);					\
-	wrapper = a_name##tsd_wrapper_get();				\
+	wrapper = a_name##tsd_wrapper_get(true);			\
 	wrapper->val = *(val);						\
 	if (a_cleanup != malloc_tsd_no_cleanup)				\
 		wrapper->initialized = true;				\
@@ -536,12 +595,15 @@ struct tsd_init_head_s {
    O(thread_allocated,		uint64_t)				\
    O(thread_deallocated,	uint64_t)				\
    O(prof_tdata,		prof_tdata_t *)				\
+    O(iarena,			arena_t *)				\
    O(arena,			arena_t *)				\
    O(arenas_tdata,		arena_tdata_t *)			\
    O(narenas_tdata,		unsigned)				\
    O(arenas_tdata_bypass,	bool)					\
    O(tcache_enabled,		tcache_enabled_t)			\
    O(quarantine,		quarantine_t *)				\
+    O(witnesses,		witness_list_t)				\
+    O(witness_fork,		bool)					\

 #define	TSD_INITIALIZER {						\
    tsd_state_uninitialized,						\
@@ -551,10 +613,13 @@ struct tsd_init_head_s {
    NULL,								\
    NULL,								\
    NULL,								\
+    NULL,								\
    0,									\
    false,								\
    tcache_enabled_default,						\
-    NULL								\
+    NULL,								\
+    ql_head_initializer(witnesses),					\
+    false								\
 }

 struct tsd_s {
@@ -565,6 +630,15 @@ MALLOC_TSD
 #undef O
 };

+/*
+ * Wrapper around tsd_t that makes it possible to avoid implicit conversion
+ * between tsd_t and tsdn_t, where tsdn_t is "nullable" and has to be
+ * explicitly converted to tsd_t, which is non-nullable.
+ */
+struct tsdn_s {
+	tsd_t	tsd;
+};
+
 static const tsd_t tsd_initializer = TSD_INITIALIZER;

 malloc_tsd_types(, tsd_t)
@@ -577,7 +651,7 @@ void	*malloc_tsd_malloc(size_t size);
 void	malloc_tsd_dalloc(void *wrapper);
 void	malloc_tsd_no_cleanup(void *arg);
 void	malloc_tsd_cleanup_register(bool (*f)(void));
-bool	malloc_tsd_boot0(void);
+tsd_t	*malloc_tsd_boot0(void);
 void	malloc_tsd_boot1(void);
 #if (!defined(JEMALLOC_MALLOC_THREAD_CLEANUP) && !defined(JEMALLOC_TLS) && \
    !defined(_WIN32))
@@ -594,7 +668,9 @@ void	tsd_cleanup(void *arg);
 #ifndef JEMALLOC_ENABLE_INLINE
 malloc_tsd_protos(JEMALLOC_ATTR(unused), , tsd_t)

+tsd_t	*tsd_fetch_impl(bool init);
 tsd_t	*tsd_fetch(void);
+tsdn_t	*tsd_tsdn(tsd_t *tsd);
 bool	tsd_nominal(tsd_t *tsd);
 #define	O(n, t)								\
 t	*tsd_##n##p_get(tsd_t *tsd);					\
@@ -602,6 +678,9 @@ t	tsd_##n##_get(tsd_t *tsd);					\
 void	tsd_##n##_set(tsd_t *tsd, t n);
 MALLOC_TSD
 #undef O
+tsdn_t	*tsdn_fetch(void);
+bool	tsdn_null(const tsdn_t *tsdn);
+tsd_t	*tsdn_tsd(tsdn_t *tsdn);
 #endif

 #if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_TSD_C_))
@@ -609,9 +688,13 @@ malloc_tsd_externs(, tsd_t)
 malloc_tsd_funcs(JEMALLOC_ALWAYS_INLINE, , tsd_t, tsd_initializer, tsd_cleanup)

 JEMALLOC_ALWAYS_INLINE tsd_t *
-tsd_fetch(void)
+tsd_fetch_impl(bool init)
 {
-	tsd_t *tsd = tsd_get();
+	tsd_t *tsd = tsd_get(init);
+
+	if (!init && tsd_get_allocates() && tsd == NULL)
+		return (NULL);
+	assert(tsd != NULL);

 	if (unlikely(tsd->state != tsd_state_nominal)) {
 		if (tsd->state == tsd_state_uninitialized) {
@@ -628,6 +711,20 @@ tsd_fetch(void)
 	return (tsd);
 }

+JEMALLOC_ALWAYS_INLINE tsd_t *
+tsd_fetch(void)
+{
+
+	return (tsd_fetch_impl(true));
+}
+
+JEMALLOC_ALWAYS_INLINE tsdn_t *
+tsd_tsdn(tsd_t *tsd)
+{
+
+	return ((tsdn_t *)tsd);
+}
+
 JEMALLOC_INLINE bool
 tsd_nominal(tsd_t *tsd)
 {
@@ -659,6 +756,32 @@ tsd_##n##_set(tsd_t *tsd, t n)						\
 }
 MALLOC_TSD
 #undef O
+
+JEMALLOC_ALWAYS_INLINE tsdn_t *
+tsdn_fetch(void)
+{
+
+	if (!tsd_booted_get())
+		return (NULL);
+
+	return (tsd_tsdn(tsd_fetch_impl(false)));
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+tsdn_null(const tsdn_t *tsdn)
+{
+
+	return (tsdn == NULL);
+}
+
+JEMALLOC_ALWAYS_INLINE tsd_t *
+tsdn_tsd(tsdn_t *tsdn)
+{
+
+	assert(!tsdn_null(tsdn));
+
+	return (&tsdn->tsd);
+}
 #endif

 #endif /* JEMALLOC_H_INLINES */
--- a/include/jemalloc/internal/util.h
+++ b/include/jemalloc/internal/util.h
@@ -40,6 +40,14 @@
 */
 #define	MALLOC_PRINTF_BUFSIZE	4096

+/* Junk fill patterns. */
+#ifndef JEMALLOC_ALLOC_JUNK
+#  define JEMALLOC_ALLOC_JUNK	((uint8_t)0xa5)
+#endif
+#ifndef JEMALLOC_FREE_JUNK
+#  define JEMALLOC_FREE_JUNK	((uint8_t)0x5a)
+#endif
+
 /*
 * Wrap a cpp argument that contains commas such that it isn't broken up into
 * multiple arguments.
@@ -57,30 +65,20 @@
 #	define JEMALLOC_CC_SILENCE_INIT(v)
 #endif

-#define	JEMALLOC_GNUC_PREREQ(major, minor)				\
-    (!defined(__clang__) &&						\
-    (__GNUC__ > (major) || (__GNUC__ == (major) && __GNUC_MINOR__ >= (minor))))
-#ifndef __has_builtin
-#  define __has_builtin(builtin) (0)
-#endif
-#define	JEMALLOC_CLANG_HAS_BUILTIN(builtin)				\
-    (defined(__clang__) && __has_builtin(builtin))
-
 #ifdef __GNUC__
 #	define likely(x)   __builtin_expect(!!(x), 1)
 #	define unlikely(x) __builtin_expect(!!(x), 0)
-#  if JEMALLOC_GNUC_PREREQ(4, 6) ||					\
-      JEMALLOC_CLANG_HAS_BUILTIN(__builtin_unreachable)
-#	define unreachable() __builtin_unreachable()
-#  else
-#	define unreachable()
-#  endif
 #else
 #	define likely(x)   !!(x)
 #	define unlikely(x) !!(x)
-#	define unreachable()
 #endif

+#if !defined(JEMALLOC_INTERNAL_UNREACHABLE)
+#  error JEMALLOC_INTERNAL_UNREACHABLE should have been defined by configure
+#endif
+
+#define unreachable() JEMALLOC_INTERNAL_UNREACHABLE()
+
 #include "jemalloc/internal/assert.h"

 /* Use to assert a particular configuration, e.g., cassert(config_debug). */
@@ -106,9 +104,9 @@ 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,
+size_t	malloc_vsnprintf(char *str, size_t size, const char *format,
    va_list ap);
-int	malloc_snprintf(char *str, size_t size, const char *format, ...)
+size_t	malloc_snprintf(char *str, size_t size, const char *format, ...)
    JEMALLOC_FORMAT_PRINTF(3, 4);
 void	malloc_vcprintf(void (*write_cb)(void *, const char *), void *cbopaque,
    const char *format, va_list ap);
--- a/include/jemalloc/internal/valgrind.h
+++ b/include/jemalloc/internal/valgrind.h
@@ -30,17 +30,31 @@
 * calls must be embedded in macros rather than in functions so that when
 * Valgrind reports errors, there are no extra stack frames in the backtraces.
 */
-#define	JEMALLOC_VALGRIND_MALLOC(cond, ptr, usize, zero) do {		\
-	if (unlikely(in_valgrind && cond))				\
-		VALGRIND_MALLOCLIKE_BLOCK(ptr, usize, p2rz(ptr), zero);	\
+#define	JEMALLOC_VALGRIND_MALLOC(cond, tsdn, ptr, usize, zero) do {	\
+	if (unlikely(in_valgrind && cond)) {				\
+		VALGRIND_MALLOCLIKE_BLOCK(ptr, usize, p2rz(tsdn, ptr),	\
+		    zero);						\
+	}								\
 } while (0)
-#define	JEMALLOC_VALGRIND_REALLOC(maybe_moved, ptr, usize,		\
-    ptr_maybe_null, old_ptr, old_usize, old_rzsize, old_ptr_maybe_null,	\
-    zero) do {								\
+#define	JEMALLOC_VALGRIND_REALLOC_MOVED_no(ptr, old_ptr)		\
+    (false)
+#define	JEMALLOC_VALGRIND_REALLOC_MOVED_maybe(ptr, old_ptr)		\
+    ((ptr) != (old_ptr))
+#define	JEMALLOC_VALGRIND_REALLOC_PTR_NULL_no(ptr)			\
+    (false)
+#define	JEMALLOC_VALGRIND_REALLOC_PTR_NULL_maybe(ptr)			\
+    (ptr == NULL)
+#define	JEMALLOC_VALGRIND_REALLOC_OLD_PTR_NULL_no(old_ptr)		\
+    (false)
+#define	JEMALLOC_VALGRIND_REALLOC_OLD_PTR_NULL_maybe(old_ptr)		\
+    (old_ptr == NULL)
+#define	JEMALLOC_VALGRIND_REALLOC(moved, tsdn, ptr, usize, ptr_null,	\
+    old_ptr, old_usize, old_rzsize, old_ptr_null, zero) do {		\
 	if (unlikely(in_valgrind)) {					\
-		size_t rzsize = p2rz(ptr);				\
+		size_t rzsize = p2rz(tsdn, ptr);			\
 									\
-		if (!maybe_moved || ptr == old_ptr) {			\
+		if (!JEMALLOC_VALGRIND_REALLOC_MOVED_##moved(ptr,	\
+		    old_ptr)) {						\
 			VALGRIND_RESIZEINPLACE_BLOCK(ptr, old_usize,	\
 			    usize, rzsize);				\
 			if (zero && old_usize < usize) {		\
@@ -49,11 +63,13 @@
 				    old_usize), usize - old_usize);	\
 			}						\
 		} else {						\
-			if (!old_ptr_maybe_null || old_ptr != NULL) {	\
+			if (!JEMALLOC_VALGRIND_REALLOC_OLD_PTR_NULL_##	\
+			    old_ptr_null(old_ptr)) {			\
 				valgrind_freelike_block(old_ptr,	\
 				    old_rzsize);			\
 			}						\
-			if (!ptr_maybe_null || ptr != NULL) {		\
+			if (!JEMALLOC_VALGRIND_REALLOC_PTR_NULL_##	\
+			    ptr_null(ptr)) {				\
 				size_t copy_size = (old_usize < usize)	\
 				    ?  old_usize : usize;		\
 				size_t tail_size = usize - copy_size;	\
@@ -81,8 +97,8 @@
 #define	JEMALLOC_VALGRIND_MAKE_MEM_NOACCESS(ptr, usize) do {} while (0)
 #define	JEMALLOC_VALGRIND_MAKE_MEM_UNDEFINED(ptr, usize) do {} while (0)
 #define	JEMALLOC_VALGRIND_MAKE_MEM_DEFINED(ptr, usize) do {} while (0)
-#define	JEMALLOC_VALGRIND_MALLOC(cond, ptr, usize, zero) do {} while (0)
-#define	JEMALLOC_VALGRIND_REALLOC(maybe_moved, ptr, usize,		\
+#define	JEMALLOC_VALGRIND_MALLOC(cond, tsdn, ptr, usize, zero) do {} while (0)
+#define	JEMALLOC_VALGRIND_REALLOC(maybe_moved, tsdn, ptr, usize,	\
    ptr_maybe_null, old_ptr, old_usize, old_rzsize, old_ptr_maybe_null,	\
    zero) do {} while (0)
 #define	JEMALLOC_VALGRIND_FREE(ptr, rzsize) do {} while (0)
--- a/include/jemalloc/internal/witness.h
+++ b/include/jemalloc/internal/witness.h
@@ -0,0 +1,304 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+typedef struct witness_s witness_t;
+typedef unsigned witness_rank_t;
+typedef ql_head(witness_t) witness_list_t;
+typedef int witness_comp_t (const witness_t *, const witness_t *);
+
+/*
+ * Lock ranks.  Witnesses with rank WITNESS_RANK_OMIT are completely ignored by
+ * the witness machinery.
+ */
+#define	WITNESS_RANK_OMIT		0U
+
+#define WITNESS_RANK_MIN		1U
+
+#define	WITNESS_RANK_INIT		1U
+#define	WITNESS_RANK_CTL		1U
+#define WITNESS_RANK_TCACHES		2U
+#define	WITNESS_RANK_ARENAS		3U
+
+#define	WITNESS_RANK_PROF_DUMP		4U
+#define	WITNESS_RANK_PROF_BT2GCTX	5U
+#define	WITNESS_RANK_PROF_TDATAS	6U
+#define	WITNESS_RANK_PROF_TDATA		7U
+#define	WITNESS_RANK_PROF_GCTX		8U
+
+/*
+ * Used as an argument to witness_assert_depth_to_rank() in order to validate
+ * depth excluding non-core locks with lower ranks.  Since the rank argument to
+ * witness_assert_depth_to_rank() is inclusive rather than exclusive, this
+ * definition can have the same value as the minimally ranked core lock.
+ */
+#define WITNESS_RANK_CORE		9U
+
+#define	WITNESS_RANK_ARENA		9U
+#define	WITNESS_RANK_ARENA_CHUNKS	10U
+#define	WITNESS_RANK_ARENA_NODE_CACHE	11U
+
+#define	WITNESS_RANK_BASE		12U
+
+#define	WITNESS_RANK_LEAF		0xffffffffU
+#define	WITNESS_RANK_ARENA_BIN		WITNESS_RANK_LEAF
+#define	WITNESS_RANK_ARENA_HUGE		WITNESS_RANK_LEAF
+#define	WITNESS_RANK_DSS		WITNESS_RANK_LEAF
+#define	WITNESS_RANK_PROF_ACTIVE	WITNESS_RANK_LEAF
+#define	WITNESS_RANK_PROF_DUMP_SEQ	WITNESS_RANK_LEAF
+#define	WITNESS_RANK_PROF_GDUMP		WITNESS_RANK_LEAF
+#define	WITNESS_RANK_PROF_NEXT_THR_UID	WITNESS_RANK_LEAF
+#define	WITNESS_RANK_PROF_THREAD_ACTIVE_INIT	WITNESS_RANK_LEAF
+
+#define	WITNESS_INITIALIZER(rank) {"initializer", rank, NULL, {NULL, NULL}}
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+struct witness_s {
+	/* Name, used for printing lock order reversal messages. */
+	const char		*name;
+
+	/*
+	 * Witness rank, where 0 is lowest and UINT_MAX is highest.  Witnesses
+	 * must be acquired in order of increasing rank.
+	 */
+	witness_rank_t		rank;
+
+	/*
+	 * If two witnesses are of equal rank and they have the samp comp
+	 * function pointer, it is called as a last attempt to differentiate
+	 * between witnesses of equal rank.
+	 */
+	witness_comp_t		*comp;
+
+	/* Linkage for thread's currently owned locks. */
+	ql_elm(witness_t)	link;
+};
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+void	witness_init(witness_t *witness, const char *name, witness_rank_t rank,
+    witness_comp_t *comp);
+#ifdef JEMALLOC_JET
+typedef void (witness_lock_error_t)(const witness_list_t *, const witness_t *);
+extern witness_lock_error_t *witness_lock_error;
+#else
+void	witness_lock_error(const witness_list_t *witnesses,
+    const witness_t *witness);
+#endif
+#ifdef JEMALLOC_JET
+typedef void (witness_owner_error_t)(const witness_t *);
+extern witness_owner_error_t *witness_owner_error;
+#else
+void	witness_owner_error(const witness_t *witness);
+#endif
+#ifdef JEMALLOC_JET
+typedef void (witness_not_owner_error_t)(const witness_t *);
+extern witness_not_owner_error_t *witness_not_owner_error;
+#else
+void	witness_not_owner_error(const witness_t *witness);
+#endif
+#ifdef JEMALLOC_JET
+typedef void (witness_depth_error_t)(const witness_list_t *,
+    witness_rank_t rank_inclusive, unsigned depth);
+extern witness_depth_error_t *witness_depth_error;
+#else
+void	witness_depth_error(const witness_list_t *witnesses,
+    witness_rank_t rank_inclusive, unsigned depth);
+#endif
+
+void	witnesses_cleanup(tsd_t *tsd);
+void	witness_fork_cleanup(tsd_t *tsd);
+void	witness_prefork(tsd_t *tsd);
+void	witness_postfork_parent(tsd_t *tsd);
+void	witness_postfork_child(tsd_t *tsd);
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#ifndef JEMALLOC_ENABLE_INLINE
+bool	witness_owner(tsd_t *tsd, const witness_t *witness);
+void	witness_assert_owner(tsdn_t *tsdn, const witness_t *witness);
+void	witness_assert_not_owner(tsdn_t *tsdn, const witness_t *witness);
+void witness_assert_depth_to_rank(tsdn_t *tsdn, witness_rank_t rank_inclusive,
+    unsigned depth);
+void witness_assert_depth(tsdn_t *tsdn, unsigned depth);
+void	witness_assert_lockless(tsdn_t *tsdn);
+void	witness_lock(tsdn_t *tsdn, witness_t *witness);
+void	witness_unlock(tsdn_t *tsdn, witness_t *witness);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_MUTEX_C_))
+JEMALLOC_INLINE bool
+witness_owner(tsd_t *tsd, const witness_t *witness)
+{
+	witness_list_t *witnesses;
+	witness_t *w;
+
+	cassert(config_debug);
+
+	witnesses = tsd_witnessesp_get(tsd);
+	ql_foreach(w, witnesses, link) {
+		if (w == witness)
+			return (true);
+	}
+
+	return (false);
+}
+
+JEMALLOC_INLINE void
+witness_assert_owner(tsdn_t *tsdn, const witness_t *witness)
+{
+	tsd_t *tsd;
+
+	if (!config_debug)
+		return;
+
+	if (tsdn_null(tsdn))
+		return;
+	tsd = tsdn_tsd(tsdn);
+	if (witness->rank == WITNESS_RANK_OMIT)
+		return;
+
+	if (witness_owner(tsd, witness))
+		return;
+	witness_owner_error(witness);
+}
+
+JEMALLOC_INLINE void
+witness_assert_not_owner(tsdn_t *tsdn, const witness_t *witness)
+{
+	tsd_t *tsd;
+	witness_list_t *witnesses;
+	witness_t *w;
+
+	if (!config_debug)
+		return;
+
+	if (tsdn_null(tsdn))
+		return;
+	tsd = tsdn_tsd(tsdn);
+	if (witness->rank == WITNESS_RANK_OMIT)
+		return;
+
+	witnesses = tsd_witnessesp_get(tsd);
+	ql_foreach(w, witnesses, link) {
+		if (w == witness)
+			witness_not_owner_error(witness);
+	}
+}
+
+JEMALLOC_INLINE void
+witness_assert_depth_to_rank(tsdn_t *tsdn, witness_rank_t rank_inclusive,
+    unsigned depth) {
+	tsd_t *tsd;
+	unsigned d;
+	witness_list_t *witnesses;
+	witness_t *w;
+
+	if (!config_debug)
+		return;
+
+	if (tsdn_null(tsdn))
+		return;
+	tsd = tsdn_tsd(tsdn);
+
+	d = 0;
+	witnesses = tsd_witnessesp_get(tsd);
+	w = ql_last(witnesses, link);
+	if (w != NULL) {
+		ql_reverse_foreach(w, witnesses, link) {
+			if (w->rank < rank_inclusive) {
+				break;
+			}
+			d++;
+		}
+	}
+	if (d != depth)
+		witness_depth_error(witnesses, rank_inclusive, depth);
+}
+
+JEMALLOC_INLINE void
+witness_assert_depth(tsdn_t *tsdn, unsigned depth) {
+	witness_assert_depth_to_rank(tsdn, WITNESS_RANK_MIN, depth);
+}
+
+JEMALLOC_INLINE void
+witness_assert_lockless(tsdn_t *tsdn) {
+	witness_assert_depth(tsdn, 0);
+}
+
+JEMALLOC_INLINE void
+witness_lock(tsdn_t *tsdn, witness_t *witness)
+{
+	tsd_t *tsd;
+	witness_list_t *witnesses;
+	witness_t *w;
+
+	if (!config_debug)
+		return;
+
+	if (tsdn_null(tsdn))
+		return;
+	tsd = tsdn_tsd(tsdn);
+	if (witness->rank == WITNESS_RANK_OMIT)
+		return;
+
+	witness_assert_not_owner(tsdn, witness);
+
+	witnesses = tsd_witnessesp_get(tsd);
+	w = ql_last(witnesses, link);
+	if (w == NULL) {
+		/* No other locks; do nothing. */
+	} else if (tsd_witness_fork_get(tsd) && w->rank <= witness->rank) {
+		/* Forking, and relaxed ranking satisfied. */
+	} else if (w->rank > witness->rank) {
+		/* Not forking, rank order reversal. */
+		witness_lock_error(witnesses, witness);
+	} else if (w->rank == witness->rank && (w->comp == NULL || w->comp !=
+	    witness->comp || w->comp(w, witness) > 0)) {
+		/*
+		 * Missing/incompatible comparison function, or comparison
+		 * function indicates rank order reversal.
+		 */
+		witness_lock_error(witnesses, witness);
+	}
+
+	ql_elm_new(witness, link);
+	ql_tail_insert(witnesses, witness, link);
+}
+
+JEMALLOC_INLINE void
+witness_unlock(tsdn_t *tsdn, witness_t *witness)
+{
+	tsd_t *tsd;
+	witness_list_t *witnesses;
+
+	if (!config_debug)
+		return;
+
+	if (tsdn_null(tsdn))
+		return;
+	tsd = tsdn_tsd(tsdn);
+	if (witness->rank == WITNESS_RANK_OMIT)
+		return;
+
+	/*
+	 * Check whether owner before removal, rather than relying on
+	 * witness_assert_owner() to abort, so that unit tests can test this
+	 * function's failure mode without causing undefined behavior.
+	 */
+	if (witness_owner(tsd, witness)) {
+		witnesses = tsd_witnessesp_get(tsd);
+		ql_remove(witnesses, witness, link);
+	} else
+		witness_assert_owner(tsdn, witness);
+}
+#endif
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
--- a/include/jemalloc/jemalloc_macros.h.in
+++ b/include/jemalloc/jemalloc_macros.h.in
@@ -13,11 +13,11 @@

 #  define MALLOCX_LG_ALIGN(la)	((int)(la))
 #  if LG_SIZEOF_PTR == 2
-#    define MALLOCX_ALIGN(a)	((int)(ffs(a)-1))
+#    define MALLOCX_ALIGN(a)	((int)(ffs((int)(a))-1))
 #  else
 #    define MALLOCX_ALIGN(a)						\
-       ((int)(((a) < (size_t)INT_MAX) ? ffs((int)(a))-1 :		\
-       ffs((int)((a)>>32))+31))
+       ((int)(((size_t)(a) < (size_t)INT_MAX) ? ffs((int)(a))-1 :	\
+       ffs((int)(((size_t)(a))>>32))+31))
 #  endif
 #  define MALLOCX_ZERO	((int)0x40)
 /*
@@ -29,7 +29,7 @@
 /*
 * Bias arena index bits so that 0 encodes "use an automatically chosen arena".
 */
-#  define MALLOCX_ARENA(a)	((int)(((a)+1) << 20))
+#  define MALLOCX_ARENA(a)	((((int)(a))+1) << 20)

 #if defined(__cplusplus) && defined(JEMALLOC_USE_CXX_THROW)
 #  define JEMALLOC_CXX_THROW throw()
--- a/jemalloc.pc.in
+++ b/jemalloc.pc.in
@@ -6,7 +6,7 @@ install_suffix=@install_suffix@

 Name: jemalloc
 Description: A general purpose malloc(3) implementation that emphasizes fragmentation avoidance and scalable concurrency support.
-URL: http://www.canonware.com/jemalloc
+URL: http://jemalloc.net/
 Version: @jemalloc_version@
 Cflags: -I${includedir}
 Libs: -L${libdir} -ljemalloc${install_suffix}
--- a/msvc/ReadMe.txt
+++ b/msvc/ReadMe.txt
@@ -17,7 +17,7 @@ How to build jemalloc for Windows
   (note: x86/x64 doesn't matter at this point)

 5. Generate header files:
-   sh -c "./autogen.sh CC=cl --enable-lazy-lock=no"
+   sh -c "CC=cl ./autogen.sh"

 6. Now the project can be opened and built in Visual Studio:
   msvc\jemalloc_vc2015.sln
--- a/msvc/projects/vc2015/jemalloc/jemalloc.vcxproj
+++ b/msvc/projects/vc2015/jemalloc/jemalloc.vcxproj
@@ -56,6 +56,7 @@
    <ClInclude Include="..\..\..\..\include\jemalloc\internal\mutex.h" />
    <ClInclude Include="..\..\..\..\include\jemalloc\internal\nstime.h" />
    <ClInclude Include="..\..\..\..\include\jemalloc\internal\pages.h" />
+    <ClInclude Include="..\..\..\..\include\jemalloc\internal\ph.h" />
    <ClInclude Include="..\..\..\..\include\jemalloc\internal\private_namespace.h" />
    <ClInclude Include="..\..\..\..\include\jemalloc\internal\private_unnamespace.h" />
    <ClInclude Include="..\..\..\..\include\jemalloc\internal\prng.h" />
@@ -68,12 +69,14 @@
    <ClInclude Include="..\..\..\..\include\jemalloc\internal\rb.h" />
    <ClInclude Include="..\..\..\..\include\jemalloc\internal\rtree.h" />
    <ClInclude Include="..\..\..\..\include\jemalloc\internal\size_classes.h" />
+    <ClInclude Include="..\..\..\..\include\jemalloc\internal\smoothstep.h" />
+    <ClInclude Include="..\..\..\..\include\jemalloc\internal\spin.h" />
    <ClInclude Include="..\..\..\..\include\jemalloc\internal\stats.h" />
    <ClInclude Include="..\..\..\..\include\jemalloc\internal\tcache.h" />
    <ClInclude Include="..\..\..\..\include\jemalloc\internal\ticker.h" />
    <ClInclude Include="..\..\..\..\include\jemalloc\internal\tsd.h" />
    <ClInclude Include="..\..\..\..\include\jemalloc\internal\util.h" />
-    <ClInclude Include="..\..\..\..\include\jemalloc\internal\valgrind.h" />
+    <ClInclude Include="..\..\..\..\include\jemalloc\internal\witness.h" />
    <ClInclude Include="..\..\..\..\include\jemalloc\jemalloc.h" />
    <ClInclude Include="..\..\..\..\include\jemalloc\jemalloc_defs.h" />
    <ClInclude Include="..\..\..\..\include\jemalloc\jemalloc_macros.h" />
@@ -109,11 +112,13 @@
    <ClCompile Include="..\..\..\..\src\prof.c" />
    <ClCompile Include="..\..\..\..\src\quarantine.c" />
    <ClCompile Include="..\..\..\..\src\rtree.c" />
+    <ClCompile Include="..\..\..\..\src\spin.c" />
    <ClCompile Include="..\..\..\..\src\stats.c" />
    <ClCompile Include="..\..\..\..\src\tcache.c" />
    <ClCompile Include="..\..\..\..\src\ticker.c" />
    <ClCompile Include="..\..\..\..\src\tsd.c" />
    <ClCompile Include="..\..\..\..\src\util.c" />
+    <ClCompile Include="..\..\..\..\src\witness.c" />
  </ItemGroup>
  <PropertyGroup Label="Globals">
    <ProjectGuid>{8D6BB292-9E1C-413D-9F98-4864BDC1514A}</ProjectGuid>
@@ -250,7 +255,7 @@
      <Optimization>Disabled</Optimization>
      <PreprocessorDefinitions>_REENTRANT;_WINDLL;DLLEXPORT;JEMALLOC_DEBUG;_DEBUG;%(PreprocessorDefinitions)</PreprocessorDefinitions>
      <AdditionalIncludeDirectories>..\..\..\..\include;..\..\..\..\include\msvc_compat;%(AdditionalIncludeDirectories)</AdditionalIncludeDirectories>
-      <DisableSpecificWarnings>4090;4146;4244;4267;4334</DisableSpecificWarnings>
+      <DisableSpecificWarnings>4090;4146;4267;4334</DisableSpecificWarnings>
      <ProgramDataBaseFileName>$(OutputPath)$(TargetName).pdb</ProgramDataBaseFileName>
    </ClCompile>
    <Link>
@@ -267,7 +272,7 @@
      <PreprocessorDefinitions>JEMALLOC_DEBUG;_REENTRANT;JEMALLOC_EXPORT=;_DEBUG;_LIB;%(PreprocessorDefinitions)</PreprocessorDefinitions>
      <AdditionalIncludeDirectories>..\..\..\..\include;..\..\..\..\include\msvc_compat;%(AdditionalIncludeDirectories)</AdditionalIncludeDirectories>
      <RuntimeLibrary>MultiThreadedDebug</RuntimeLibrary>
-      <DisableSpecificWarnings>4090;4146;4244;4267;4334</DisableSpecificWarnings>
+      <DisableSpecificWarnings>4090;4146;4267;4334</DisableSpecificWarnings>
      <ProgramDataBaseFileName>$(OutputPath)$(TargetName).pdb</ProgramDataBaseFileName>
    </ClCompile>
    <Link>
@@ -283,7 +288,7 @@
      <Optimization>Disabled</Optimization>
      <PreprocessorDefinitions>_REENTRANT;_WINDLL;DLLEXPORT;JEMALLOC_DEBUG;_DEBUG;%(PreprocessorDefinitions)</PreprocessorDefinitions>
      <AdditionalIncludeDirectories>..\..\..\..\include;..\..\..\..\include\msvc_compat;%(AdditionalIncludeDirectories)</AdditionalIncludeDirectories>
-      <DisableSpecificWarnings>4090;4146;4244;4267;4334</DisableSpecificWarnings>
+      <DisableSpecificWarnings>4090;4146;4267;4334</DisableSpecificWarnings>
      <ProgramDataBaseFileName>$(OutputPath)$(TargetName).pdb</ProgramDataBaseFileName>
    </ClCompile>
    <Link>
@@ -300,8 +305,9 @@
      <PreprocessorDefinitions>JEMALLOC_DEBUG;_REENTRANT;JEMALLOC_EXPORT=;_DEBUG;_LIB;%(PreprocessorDefinitions)</PreprocessorDefinitions>
      <AdditionalIncludeDirectories>..\..\..\..\include;..\..\..\..\include\msvc_compat;%(AdditionalIncludeDirectories)</AdditionalIncludeDirectories>
      <RuntimeLibrary>MultiThreadedDebug</RuntimeLibrary>
-      <DisableSpecificWarnings>4090;4146;4244;4267;4334</DisableSpecificWarnings>
-      <ProgramDataBaseFileName>$(OutputPath)$(TargetName).pdb</ProgramDataBaseFileName>
+      <DisableSpecificWarnings>4090;4146;4267;4334</DisableSpecificWarnings>
+      <DebugInformationFormat>OldStyle</DebugInformationFormat>
+      <MinimalRebuild>false</MinimalRebuild>
    </ClCompile>
    <Link>
      <SubSystem>Windows</SubSystem>
@@ -318,7 +324,7 @@
      <IntrinsicFunctions>true</IntrinsicFunctions>
      <PreprocessorDefinitions>_REENTRANT;_WINDLL;DLLEXPORT;NDEBUG;%(PreprocessorDefinitions)</PreprocessorDefinitions>
      <AdditionalIncludeDirectories>..\..\..\..\include;..\..\..\..\include\msvc_compat;%(AdditionalIncludeDirectories)</AdditionalIncludeDirectories>
-      <DisableSpecificWarnings>4090;4146;4244;4267;4334</DisableSpecificWarnings>
+      <DisableSpecificWarnings>4090;4146;4267;4334</DisableSpecificWarnings>
      <ProgramDataBaseFileName>$(OutputPath)$(TargetName).pdb</ProgramDataBaseFileName>
    </ClCompile>
    <Link>
@@ -339,7 +345,7 @@
      <PreprocessorDefinitions>_REENTRANT;JEMALLOC_EXPORT=;NDEBUG;_LIB;%(PreprocessorDefinitions)</PreprocessorDefinitions>
      <AdditionalIncludeDirectories>..\..\..\..\include;..\..\..\..\include\msvc_compat;%(AdditionalIncludeDirectories)</AdditionalIncludeDirectories>
      <RuntimeLibrary>MultiThreaded</RuntimeLibrary>
-      <DisableSpecificWarnings>4090;4146;4244;4267;4334</DisableSpecificWarnings>
+      <DisableSpecificWarnings>4090;4146;4267;4334</DisableSpecificWarnings>
      <ProgramDataBaseFileName>$(OutputPath)$(TargetName).pdb</ProgramDataBaseFileName>
    </ClCompile>
    <Link>
@@ -359,7 +365,7 @@
      <IntrinsicFunctions>true</IntrinsicFunctions>
      <AdditionalIncludeDirectories>..\..\..\..\include;..\..\..\..\include\msvc_compat;%(AdditionalIncludeDirectories)</AdditionalIncludeDirectories>
      <PreprocessorDefinitions>_REENTRANT;_WINDLL;DLLEXPORT;NDEBUG;%(PreprocessorDefinitions)</PreprocessorDefinitions>
-      <DisableSpecificWarnings>4090;4146;4244;4267;4334</DisableSpecificWarnings>
+      <DisableSpecificWarnings>4090;4146;4267;4334</DisableSpecificWarnings>
      <ProgramDataBaseFileName>$(OutputPath)$(TargetName).pdb</ProgramDataBaseFileName>
    </ClCompile>
    <Link>
@@ -380,8 +386,8 @@
      <PreprocessorDefinitions>_REENTRANT;JEMALLOC_EXPORT=;NDEBUG;_LIB;%(PreprocessorDefinitions)</PreprocessorDefinitions>
      <AdditionalIncludeDirectories>..\..\..\..\include;..\..\..\..\include\msvc_compat;%(AdditionalIncludeDirectories)</AdditionalIncludeDirectories>
      <RuntimeLibrary>MultiThreaded</RuntimeLibrary>
-      <DisableSpecificWarnings>4090;4146;4244;4267;4334</DisableSpecificWarnings>
-      <ProgramDataBaseFileName>$(OutputPath)$(TargetName).pdb</ProgramDataBaseFileName>
+      <DisableSpecificWarnings>4090;4146;4267;4334</DisableSpecificWarnings>
+      <DebugInformationFormat>OldStyle</DebugInformationFormat>
    </ClCompile>
    <Link>
      <SubSystem>Windows</SubSystem>
--- a/msvc/projects/vc2015/jemalloc/jemalloc.vcxproj.filters
+++ b/msvc/projects/vc2015/jemalloc/jemalloc.vcxproj.filters
@@ -107,6 +107,9 @@
    <ClInclude Include="..\..\..\..\include\jemalloc\internal\pages.h">
      <Filter>Header Files\internal</Filter>
    </ClInclude>
+    <ClInclude Include="..\..\..\..\include\jemalloc\internal\ph.h">
+      <Filter>Header Files\internal</Filter>
+    </ClInclude>
    <ClInclude Include="..\..\..\..\include\jemalloc\internal\private_namespace.h">
      <Filter>Header Files\internal</Filter>
    </ClInclude>
@@ -143,6 +146,12 @@
    <ClInclude Include="..\..\..\..\include\jemalloc\internal\size_classes.h">
      <Filter>Header Files\internal</Filter>
    </ClInclude>
+    <ClInclude Include="..\..\..\..\include\jemalloc\internal\smoothstep.h">
+      <Filter>Header Files\internal</Filter>
+    </ClInclude>
+    <ClInclude Include="..\..\..\..\include\jemalloc\internal\spin.h">
+      <Filter>Header Files\internal</Filter>
+    </ClInclude>
    <ClInclude Include="..\..\..\..\include\jemalloc\internal\stats.h">
      <Filter>Header Files\internal</Filter>
    </ClInclude>
@@ -158,7 +167,7 @@
    <ClInclude Include="..\..\..\..\include\jemalloc\internal\util.h">
      <Filter>Header Files\internal</Filter>
    </ClInclude>
-    <ClInclude Include="..\..\..\..\include\jemalloc\internal\valgrind.h">
+    <ClInclude Include="..\..\..\..\include\jemalloc\internal\witness.h">
      <Filter>Header Files\internal</Filter>
    </ClInclude>
    <ClInclude Include="..\..\..\..\include\msvc_compat\strings.h">
@@ -238,6 +247,9 @@
    <ClCompile Include="..\..\..\..\src\rtree.c">
      <Filter>Source Files</Filter>
    </ClCompile>
+    <ClCompile Include="..\..\..\..\src\spin.c">
+      <Filter>Source Files</Filter>
+    </ClCompile>
    <ClCompile Include="..\..\..\..\src\stats.c">
      <Filter>Source Files</Filter>
    </ClCompile>
@@ -253,5 +265,8 @@
    <ClCompile Include="..\..\..\..\src\util.c">
      <Filter>Source Files</Filter>
    </ClCompile>
+    <ClCompile Include="..\..\..\..\src\witness.c">
+      <Filter>Source Files</Filter>
+    </ClCompile>
  </ItemGroup>
-</Project>
+</Project>
--- a/msvc/projects/vc2015/test_threads/test_threads.cpp
+++ b/msvc/projects/vc2015/test_threads/test_threads.cpp
@@ -21,7 +21,7 @@ int test_threads()
  je_malloc_conf = "narenas:3";
  int narenas = 0;
  size_t sz = sizeof(narenas);
-  je_mallctl("opt.narenas", &narenas, &sz, NULL, 0);
+  je_mallctl("opt.narenas", (void *)&narenas, &sz, NULL, 0);
  if (narenas != 3) {
    printf("Error: unexpected number of arenas: %d\n", narenas);
    return 1;
@@ -33,7 +33,7 @@ int test_threads()
  je_malloc_stats_print(NULL, NULL, NULL);
  size_t allocated1;
  size_t sz1 = sizeof(allocated1);
-  je_mallctl("stats.active", &allocated1, &sz1, NULL, 0);
+  je_mallctl("stats.active", (void *)&allocated1, &sz1, NULL, 0);
  printf("\nPress Enter to start threads...\n");
  getchar();
  printf("Starting %d threads x %d x %d iterations...\n", numThreads, numIter1, numIter2);
@@ -78,7 +78,7 @@ int test_threads()
  }
  je_malloc_stats_print(NULL, NULL, NULL);
  size_t allocated2;
-  je_mallctl("stats.active", &allocated2, &sz1, NULL, 0);
+  je_mallctl("stats.active", (void *)&allocated2, &sz1, NULL, 0);
  size_t leaked = allocated2 - allocated1;
  printf("\nDone. Leaked: %zd bytes\n", leaked);
  bool failed = leaked > 65536; // in case C++ runtime allocated something (e.g. iostream locale or facet)
--- a/scripts/gen_travis.py
+++ b/scripts/gen_travis.py
@@ -0,0 +1,85 @@
+#!/usr/bin/env python
+
+from itertools import combinations
+
+travis_template = """\
+language: generic
+
+matrix:
+  include:
+%s
+
+before_script:
+  - autoconf
+  - ./configure ${COMPILER_FLAGS:+ \
+      CC="$CC $COMPILER_FLAGS" } \
+      $CONFIGURE_FLAGS
+  - make -j3
+  - make -j3 tests
+
+script:
+  - make check
+"""
+
+# The 'default' configuration is gcc, on linux, with no compiler or configure
+# flags.  We also test with clang, -m32, --enable-debug, --enable-prof,
+# --disable-stats, and --disable-tcache.  To avoid abusing travis though, we
+# don't test all 2**7 = 128 possible combinations of these; instead, we only
+# test combinations of up to 2 'unusual' settings, under the hope that bugs
+# involving interactions of such settings are rare.
+# things at once, for C(7, 0) + C(7, 1) + C(7, 2) = 29
+MAX_UNUSUAL_OPTIONS = 2
+
+os_default = 'linux'
+os_unusual = 'osx'
+
+compilers_default = 'CC=gcc'
+compilers_unusual = 'CC=clang'
+
+compiler_flag_unusuals = ['-m32']
+
+configure_flag_unusuals = [
+    '--enable-debug', '--enable-prof', '--disable-stats', '--disable-tcache',
+]
+
+all_unusuals = (
+    [os_unusual] + [compilers_unusual] + compiler_flag_unusuals
+    + configure_flag_unusuals
+)
+
+unusual_combinations_to_test = []
+for i in xrange(MAX_UNUSUAL_OPTIONS + 1):
+    unusual_combinations_to_test += combinations(all_unusuals, i)
+
+include_rows = ""
+for unusual_combination in unusual_combinations_to_test:
+    os = os_default
+    if os_unusual in unusual_combination:
+        os = os_unusual
+
+    compilers = compilers_default
+    if compilers_unusual in unusual_combination:
+        compilers = compilers_unusual
+
+    compiler_flags = [
+        x for x in unusual_combination if x in compiler_flag_unusuals]
+
+    configure_flags = [
+        x for x in unusual_combination if x in configure_flag_unusuals]
+
+    # Filter out an unsupported configuration - heap profiling on OS X.
+    if os == 'osx' and '--enable-prof' in configure_flags:
+        continue
+
+    env_string = '{} COMPILER_FLAGS="{}" CONFIGURE_FLAGS="{}"'.format(
+        compilers, " ".join(compiler_flags), " ".join(configure_flags))
+
+    include_rows += '    - os: %s\n' % os
+    include_rows += '      env: %s\n' % env_string
+    if '-m32' in unusual_combination and os == 'linux':
+        include_rows += '      addons:\n'
+	include_rows += '        apt:\n'
+	include_rows += '          packages:\n'
+	include_rows += '            - gcc-multilib\n'
+
+print travis_template % include_rows
--- a/src/arena.c
+++ b/src/arena.c
--- a/src/base.c
+++ b/src/base.c
@@ -5,7 +5,8 @@
 /* Data. */

 static malloc_mutex_t	base_mtx;
-static extent_tree_t	base_avail_szad;
+static size_t		base_extent_sn_next;
+static extent_tree_t	base_avail_szsnad;
 static extent_node_t	*base_nodes;
 static size_t		base_allocated;
 static size_t		base_resident;
@@ -13,12 +14,13 @@ static size_t		base_mapped;

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

-/* base_mtx must be held. */
 static extent_node_t *
-base_node_try_alloc(void)
+base_node_try_alloc(tsdn_t *tsdn)
 {
 	extent_node_t *node;

+	malloc_mutex_assert_owner(tsdn, &base_mtx);
+
 	if (base_nodes == NULL)
 		return (NULL);
 	node = base_nodes;
@@ -27,33 +29,42 @@ base_node_try_alloc(void)
 	return (node);
 }

-/* base_mtx must be held. */
 static void
-base_node_dalloc(extent_node_t *node)
+base_node_dalloc(tsdn_t *tsdn, extent_node_t *node)
 {

+	malloc_mutex_assert_owner(tsdn, &base_mtx);
+
 	JEMALLOC_VALGRIND_MAKE_MEM_UNDEFINED(node, sizeof(extent_node_t));
 	*(extent_node_t **)node = base_nodes;
 	base_nodes = node;
 }

-/* base_mtx must be held. */
+static void
+base_extent_node_init(extent_node_t *node, void *addr, size_t size)
+{
+	size_t sn = atomic_add_z(&base_extent_sn_next, 1) - 1;
+
+	extent_node_init(node, NULL, addr, size, sn, true, true);
+}
+
 static extent_node_t *
-base_chunk_alloc(size_t minsize)
+base_chunk_alloc(tsdn_t *tsdn, size_t minsize)
 {
 	extent_node_t *node;
 	size_t csize, nsize;
 	void *addr;

+	malloc_mutex_assert_owner(tsdn, &base_mtx);
 	assert(minsize != 0);
-	node = base_node_try_alloc();
+	node = base_node_try_alloc(tsdn);
 	/* Allocate enough space to also carve a node out if necessary. */
 	nsize = (node == NULL) ? CACHELINE_CEILING(sizeof(extent_node_t)) : 0;
 	csize = CHUNK_CEILING(minsize + nsize);
 	addr = chunk_alloc_base(csize);
 	if (addr == NULL) {
 		if (node != NULL)
-			base_node_dalloc(node);
+			base_node_dalloc(tsdn, node);
 		return (NULL);
 	}
 	base_mapped += csize;
@@ -66,7 +77,7 @@ base_chunk_alloc(size_t minsize)
 			base_resident += PAGE_CEILING(nsize);
 		}
 	}
-	extent_node_init(node, NULL, addr, csize, true, true);
+	base_extent_node_init(node, addr, csize);
 	return (node);
 }

@@ -76,7 +87,7 @@ base_chunk_alloc(size_t minsize)
 * physical memory usage.
 */
 void *
-base_alloc(size_t size)
+base_alloc(tsdn_t *tsdn, size_t size)
 {
 	void *ret;
 	size_t csize, usize;
@@ -90,15 +101,15 @@ base_alloc(size_t size)
 	csize = CACHELINE_CEILING(size);

 	usize = s2u(csize);
-	extent_node_init(&key, NULL, NULL, usize, false, false);
-	malloc_mutex_lock(&base_mtx);
-	node = extent_tree_szad_nsearch(&base_avail_szad, &key);
+	extent_node_init(&key, NULL, NULL, usize, 0, false, false);
+	malloc_mutex_lock(tsdn, &base_mtx);
+	node = extent_tree_szsnad_nsearch(&base_avail_szsnad, &key);
 	if (node != NULL) {
 		/* Use existing space. */
-		extent_tree_szad_remove(&base_avail_szad, node);
+		extent_tree_szsnad_remove(&base_avail_szsnad, node);
 	} else {
 		/* Try to allocate more space. */
-		node = base_chunk_alloc(csize);
+		node = base_chunk_alloc(tsdn, csize);
 	}
 	if (node == NULL) {
 		ret = NULL;
@@ -109,9 +120,9 @@ base_alloc(size_t size)
 	if (extent_node_size_get(node) > csize) {
 		extent_node_addr_set(node, (void *)((uintptr_t)ret + csize));
 		extent_node_size_set(node, extent_node_size_get(node) - csize);
-		extent_tree_szad_insert(&base_avail_szad, node);
+		extent_tree_szsnad_insert(&base_avail_szsnad, node);
 	} else
-		base_node_dalloc(node);
+		base_node_dalloc(tsdn, node);
 	if (config_stats) {
 		base_allocated += csize;
 		/*
@@ -123,52 +134,54 @@ base_alloc(size_t size)
 	}
 	JEMALLOC_VALGRIND_MAKE_MEM_DEFINED(ret, csize);
 label_return:
-	malloc_mutex_unlock(&base_mtx);
+	malloc_mutex_unlock(tsdn, &base_mtx);
 	return (ret);
 }

 void
-base_stats_get(size_t *allocated, size_t *resident, size_t *mapped)
+base_stats_get(tsdn_t *tsdn, size_t *allocated, size_t *resident,
+    size_t *mapped)
 {

-	malloc_mutex_lock(&base_mtx);
+	malloc_mutex_lock(tsdn, &base_mtx);
 	assert(base_allocated <= base_resident);
 	assert(base_resident <= base_mapped);
 	*allocated = base_allocated;
 	*resident = base_resident;
 	*mapped = base_mapped;
-	malloc_mutex_unlock(&base_mtx);
+	malloc_mutex_unlock(tsdn, &base_mtx);
 }

 bool
 base_boot(void)
 {

-	if (malloc_mutex_init(&base_mtx))
+	if (malloc_mutex_init(&base_mtx, "base", WITNESS_RANK_BASE))
 		return (true);
-	extent_tree_szad_new(&base_avail_szad);
+	base_extent_sn_next = 0;
+	extent_tree_szsnad_new(&base_avail_szsnad);
 	base_nodes = NULL;

 	return (false);
 }

 void
-base_prefork(void)
+base_prefork(tsdn_t *tsdn)
 {

-	malloc_mutex_prefork(&base_mtx);
+	malloc_mutex_prefork(tsdn, &base_mtx);
 }

 void
-base_postfork_parent(void)
+base_postfork_parent(tsdn_t *tsdn)
 {

-	malloc_mutex_postfork_parent(&base_mtx);
+	malloc_mutex_postfork_parent(tsdn, &base_mtx);
 }

 void
-base_postfork_child(void)
+base_postfork_child(tsdn_t *tsdn)
 {

-	malloc_mutex_postfork_child(&base_mtx);
+	malloc_mutex_postfork_child(tsdn, &base_mtx);
 }
--- a/src/bitmap.c
+++ b/src/bitmap.c
@@ -74,15 +74,11 @@ bitmap_init(bitmap_t *bitmap, const bitmap_info_t *binfo)
 void
 bitmap_info_init(bitmap_info_t *binfo, size_t nbits)
 {
-	size_t i;

 	assert(nbits > 0);
 	assert(nbits <= (ZU(1) << LG_BITMAP_MAXBITS));

-	i = nbits >> LG_BITMAP_GROUP_NBITS;
-	if (nbits % BITMAP_GROUP_NBITS != 0)
-		i++;
-	binfo->ngroups = i;
+	binfo->ngroups = BITMAP_BITS2GROUPS(nbits);
 	binfo->nbits = nbits;
 }

@@ -99,9 +95,10 @@ bitmap_init(bitmap_t *bitmap, const bitmap_info_t *binfo)
 	size_t extra;

 	memset(bitmap, 0xffU, bitmap_size(binfo));
-	extra = (binfo->nbits % (binfo->ngroups * BITMAP_GROUP_NBITS));
+	extra = (BITMAP_GROUP_NBITS - (binfo->nbits & BITMAP_GROUP_NBITS_MASK))
+	    & BITMAP_GROUP_NBITS_MASK;
 	if (extra != 0)
-		bitmap[binfo->ngroups - 1] >>= (BITMAP_GROUP_NBITS - extra);
+		bitmap[binfo->ngroups - 1] >>= extra;
 }

 #endif /* USE_TREE */
--- a/src/chunk.c
+++ b/src/chunk.c
@@ -49,9 +49,10 @@ const chunk_hooks_t	chunk_hooks_default = {
 * definition.
 */

-static void	chunk_record(arena_t *arena, chunk_hooks_t *chunk_hooks,
-    extent_tree_t *chunks_szad, extent_tree_t *chunks_ad, bool cache,
-    void *chunk, size_t size, bool zeroed, bool committed);
+static void	chunk_record(tsdn_t *tsdn, arena_t *arena,
+    chunk_hooks_t *chunk_hooks, extent_tree_t *chunks_szsnad,
+    extent_tree_t *chunks_ad, bool cache, void *chunk, size_t size, size_t sn,
+    bool zeroed, bool committed);

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

@@ -63,23 +64,23 @@ chunk_hooks_get_locked(arena_t *arena)
 }

 chunk_hooks_t
-chunk_hooks_get(arena_t *arena)
+chunk_hooks_get(tsdn_t *tsdn, arena_t *arena)
 {
 	chunk_hooks_t chunk_hooks;

-	malloc_mutex_lock(&arena->chunks_mtx);
+	malloc_mutex_lock(tsdn, &arena->chunks_mtx);
 	chunk_hooks = chunk_hooks_get_locked(arena);
-	malloc_mutex_unlock(&arena->chunks_mtx);
+	malloc_mutex_unlock(tsdn, &arena->chunks_mtx);

 	return (chunk_hooks);
 }

 chunk_hooks_t
-chunk_hooks_set(arena_t *arena, const chunk_hooks_t *chunk_hooks)
+chunk_hooks_set(tsdn_t *tsdn, arena_t *arena, const chunk_hooks_t *chunk_hooks)
 {
 	chunk_hooks_t old_chunk_hooks;

-	malloc_mutex_lock(&arena->chunks_mtx);
+	malloc_mutex_lock(tsdn, &arena->chunks_mtx);
 	old_chunk_hooks = arena->chunk_hooks;
 	/*
 	 * Copy each field atomically so that it is impossible for readers to
@@ -104,14 +105,14 @@ chunk_hooks_set(arena_t *arena, const chunk_hooks_t *chunk_hooks)
 	ATOMIC_COPY_HOOK(split);
 	ATOMIC_COPY_HOOK(merge);
 #undef ATOMIC_COPY_HOOK
-	malloc_mutex_unlock(&arena->chunks_mtx);
+	malloc_mutex_unlock(tsdn, &arena->chunks_mtx);

 	return (old_chunk_hooks);
 }

 static void
-chunk_hooks_assure_initialized_impl(arena_t *arena, chunk_hooks_t *chunk_hooks,
-    bool locked)
+chunk_hooks_assure_initialized_impl(tsdn_t *tsdn, arena_t *arena,
+    chunk_hooks_t *chunk_hooks, bool locked)
 {
 	static const chunk_hooks_t uninitialized_hooks =
 	    CHUNK_HOOKS_INITIALIZER;
@@ -119,27 +120,28 @@ chunk_hooks_assure_initialized_impl(arena_t *arena, chunk_hooks_t *chunk_hooks,
 	if (memcmp(chunk_hooks, &uninitialized_hooks, sizeof(chunk_hooks_t)) ==
 	    0) {
 		*chunk_hooks = locked ? chunk_hooks_get_locked(arena) :
-		    chunk_hooks_get(arena);
+		    chunk_hooks_get(tsdn, arena);
 	}
 }

 static void
-chunk_hooks_assure_initialized_locked(arena_t *arena,
+chunk_hooks_assure_initialized_locked(tsdn_t *tsdn, arena_t *arena,
    chunk_hooks_t *chunk_hooks)
 {

-	chunk_hooks_assure_initialized_impl(arena, chunk_hooks, true);
+	chunk_hooks_assure_initialized_impl(tsdn, arena, chunk_hooks, true);
 }

 static void
-chunk_hooks_assure_initialized(arena_t *arena, chunk_hooks_t *chunk_hooks)
+chunk_hooks_assure_initialized(tsdn_t *tsdn, arena_t *arena,
+    chunk_hooks_t *chunk_hooks)
 {

-	chunk_hooks_assure_initialized_impl(arena, chunk_hooks, false);
+	chunk_hooks_assure_initialized_impl(tsdn, arena, chunk_hooks, false);
 }

 bool
-chunk_register(const void *chunk, const extent_node_t *node)
+chunk_register(const void *chunk, const extent_node_t *node, bool *gdump)
 {

 	assert(extent_node_addr_get(node) == chunk);
@@ -158,8 +160,7 @@ chunk_register(const void *chunk, const extent_node_t *node)
 			 */
 			high = atomic_read_z(&highchunks);
 		}
-		if (cur > high && prof_gdump_get_unlocked())
-			prof_gdump();
+		*gdump = (cur > high && prof_gdump_get_unlocked());
 	}

 	return (false);
@@ -181,33 +182,40 @@ chunk_deregister(const void *chunk, const extent_node_t *node)
 }

 /*
- * Do first-best-fit chunk selection, i.e. select the lowest chunk that best
- * fits.
+ * Do first-best-fit chunk selection, i.e. select the oldest/lowest chunk that
+ * best fits.
 */
 static extent_node_t *
-chunk_first_best_fit(arena_t *arena, extent_tree_t *chunks_szad,
-    extent_tree_t *chunks_ad, size_t size)
+chunk_first_best_fit(arena_t *arena, extent_tree_t *chunks_szsnad, size_t size)
 {
+	extent_node_t *node;
+	size_t qsize;
 	extent_node_t key;

 	assert(size == CHUNK_CEILING(size));

-	extent_node_init(&key, arena, NULL, size, false, false);
-	return (extent_tree_szad_nsearch(chunks_szad, &key));
+	qsize = extent_size_quantize_ceil(size);
+	extent_node_init(&key, arena, NULL, qsize, 0, false, false);
+	node = extent_tree_szsnad_nsearch(chunks_szsnad, &key);
+	assert(node == NULL || extent_node_size_get(node) >= size);
+	return node;
 }

 static void *
-chunk_recycle(arena_t *arena, chunk_hooks_t *chunk_hooks,
-    extent_tree_t *chunks_szad, extent_tree_t *chunks_ad, bool cache,
-    void *new_addr, size_t size, size_t alignment, bool *zero, bool *commit,
-    bool dalloc_node)
+chunk_recycle(tsdn_t *tsdn, arena_t *arena, chunk_hooks_t *chunk_hooks,
+    extent_tree_t *chunks_szsnad, extent_tree_t *chunks_ad, bool cache,
+    void *new_addr, size_t size, size_t alignment, size_t *sn, bool *zero,
+    bool *commit, bool dalloc_node)
 {
 	void *ret;
 	extent_node_t *node;
 	size_t alloc_size, leadsize, trailsize;
 	bool zeroed, committed;

+	assert(CHUNK_CEILING(size) == size);
+	assert(alignment > 0);
 	assert(new_addr == NULL || alignment == chunksize);
+	assert(CHUNK_ADDR2BASE(new_addr) == new_addr);
 	/*
 	 * Cached chunks use the node linkage embedded in their headers, in
 	 * which case dalloc_node is true, and new_addr is non-NULL because
@@ -215,24 +223,23 @@ chunk_recycle(arena_t *arena, chunk_hooks_t *chunk_hooks,
 	 */
 	assert(dalloc_node || new_addr != NULL);

-	alloc_size = CHUNK_CEILING(s2u(size + alignment - chunksize));
+	alloc_size = size + CHUNK_CEILING(alignment) - chunksize;
 	/* Beware size_t wrap-around. */
 	if (alloc_size < size)
 		return (NULL);
-	malloc_mutex_lock(&arena->chunks_mtx);
-	chunk_hooks_assure_initialized_locked(arena, chunk_hooks);
+	malloc_mutex_lock(tsdn, &arena->chunks_mtx);
+	chunk_hooks_assure_initialized_locked(tsdn, arena, chunk_hooks);
 	if (new_addr != NULL) {
 		extent_node_t key;
-		extent_node_init(&key, arena, new_addr, alloc_size, false,
+		extent_node_init(&key, arena, new_addr, alloc_size, 0, false,
 		    false);
 		node = extent_tree_ad_search(chunks_ad, &key);
 	} else {
-		node = chunk_first_best_fit(arena, chunks_szad, chunks_ad,
-		    alloc_size);
+		node = chunk_first_best_fit(arena, chunks_szsnad, alloc_size);
 	}
 	if (node == NULL || (new_addr != NULL && extent_node_size_get(node) <
 	    size)) {
-		malloc_mutex_unlock(&arena->chunks_mtx);
+		malloc_mutex_unlock(tsdn, &arena->chunks_mtx);
 		return (NULL);
 	}
 	leadsize = ALIGNMENT_CEILING((uintptr_t)extent_node_addr_get(node),
@@ -241,6 +248,7 @@ chunk_recycle(arena_t *arena, chunk_hooks_t *chunk_hooks,
 	assert(extent_node_size_get(node) >= leadsize + size);
 	trailsize = extent_node_size_get(node) - leadsize - size;
 	ret = (void *)((uintptr_t)extent_node_addr_get(node) + leadsize);
+	*sn = extent_node_sn_get(node);
 	zeroed = extent_node_zeroed_get(node);
 	if (zeroed)
 		*zero = true;
@@ -251,17 +259,17 @@ chunk_recycle(arena_t *arena, chunk_hooks_t *chunk_hooks,
 	if (leadsize != 0 &&
 	    chunk_hooks->split(extent_node_addr_get(node),
 	    extent_node_size_get(node), leadsize, size, false, arena->ind)) {
-		malloc_mutex_unlock(&arena->chunks_mtx);
+		malloc_mutex_unlock(tsdn, &arena->chunks_mtx);
 		return (NULL);
 	}
 	/* Remove node from the tree. */
-	extent_tree_szad_remove(chunks_szad, node);
+	extent_tree_szsnad_remove(chunks_szsnad, node);
 	extent_tree_ad_remove(chunks_ad, node);
 	arena_chunk_cache_maybe_remove(arena, node, cache);
 	if (leadsize != 0) {
 		/* Insert the leading space as a smaller chunk. */
 		extent_node_size_set(node, leadsize);
-		extent_tree_szad_insert(chunks_szad, node);
+		extent_tree_szsnad_insert(chunks_szsnad, node);
 		extent_tree_ad_insert(chunks_ad, node);
 		arena_chunk_cache_maybe_insert(arena, node, cache);
 		node = NULL;
@@ -271,41 +279,42 @@ chunk_recycle(arena_t *arena, chunk_hooks_t *chunk_hooks,
 		if (chunk_hooks->split(ret, size + trailsize, size,
 		    trailsize, false, arena->ind)) {
 			if (dalloc_node && node != NULL)
-				arena_node_dalloc(arena, node);
-			malloc_mutex_unlock(&arena->chunks_mtx);
-			chunk_record(arena, chunk_hooks, chunks_szad, chunks_ad,
-			    cache, ret, size + trailsize, zeroed, committed);
+				arena_node_dalloc(tsdn, arena, node);
+			malloc_mutex_unlock(tsdn, &arena->chunks_mtx);
+			chunk_record(tsdn, arena, chunk_hooks, chunks_szsnad,
+			    chunks_ad, cache, ret, size + trailsize, *sn,
+			    zeroed, committed);
 			return (NULL);
 		}
 		/* Insert the trailing space as a smaller chunk. */
 		if (node == NULL) {
-			node = arena_node_alloc(arena);
+			node = arena_node_alloc(tsdn, arena);
 			if (node == NULL) {
-				malloc_mutex_unlock(&arena->chunks_mtx);
-				chunk_record(arena, chunk_hooks, chunks_szad,
-				    chunks_ad, cache, ret, size + trailsize,
-				    zeroed, committed);
+				malloc_mutex_unlock(tsdn, &arena->chunks_mtx);
+				chunk_record(tsdn, arena, chunk_hooks,
+				    chunks_szsnad, chunks_ad, cache, ret, size
+				    + trailsize, *sn, zeroed, committed);
 				return (NULL);
 			}
 		}
 		extent_node_init(node, arena, (void *)((uintptr_t)(ret) + size),
-		    trailsize, zeroed, committed);
-		extent_tree_szad_insert(chunks_szad, node);
+		    trailsize, *sn, zeroed, committed);
+		extent_tree_szsnad_insert(chunks_szsnad, node);
 		extent_tree_ad_insert(chunks_ad, node);
 		arena_chunk_cache_maybe_insert(arena, node, cache);
 		node = NULL;
 	}
 	if (!committed && chunk_hooks->commit(ret, size, 0, size, arena->ind)) {
-		malloc_mutex_unlock(&arena->chunks_mtx);
-		chunk_record(arena, chunk_hooks, chunks_szad, chunks_ad, cache,
-		    ret, size, zeroed, committed);
+		malloc_mutex_unlock(tsdn, &arena->chunks_mtx);
+		chunk_record(tsdn, arena, chunk_hooks, chunks_szsnad, chunks_ad,
+		    cache, ret, size, *sn, zeroed, committed);
 		return (NULL);
 	}
-	malloc_mutex_unlock(&arena->chunks_mtx);
+	malloc_mutex_unlock(tsdn, &arena->chunks_mtx);

 	assert(dalloc_node || node != NULL);
 	if (dalloc_node && node != NULL)
-		arena_node_dalloc(arena, node);
+		arena_node_dalloc(tsdn, arena, node);
 	if (*zero) {
 		if (!zeroed)
 			memset(ret, 0, size);
@@ -313,10 +322,11 @@ chunk_recycle(arena_t *arena, chunk_hooks_t *chunk_hooks,
 			size_t i;
 			size_t *p = (size_t *)(uintptr_t)ret;

-			JEMALLOC_VALGRIND_MAKE_MEM_DEFINED(ret, size);
 			for (i = 0; i < size / sizeof(size_t); i++)
 				assert(p[i] == 0);
 		}
+		if (config_valgrind)
+			JEMALLOC_VALGRIND_MAKE_MEM_DEFINED(ret, size);
 	}
 	return (ret);
 }
@@ -328,8 +338,8 @@ chunk_recycle(arena_t *arena, chunk_hooks_t *chunk_hooks,
 * them if they are returned.
 */
 static void *
-chunk_alloc_core(arena_t *arena, void *new_addr, size_t size, size_t alignment,
-    bool *zero, bool *commit, dss_prec_t dss_prec)
+chunk_alloc_core(tsdn_t *tsdn, arena_t *arena, void *new_addr, size_t size,
+    size_t alignment, bool *zero, bool *commit, dss_prec_t dss_prec)
 {
 	void *ret;

@@ -340,8 +350,8 @@ chunk_alloc_core(arena_t *arena, void *new_addr, size_t size, size_t alignment,

 	/* "primary" dss. */
 	if (have_dss && dss_prec == dss_prec_primary && (ret =
-	    chunk_alloc_dss(arena, new_addr, size, alignment, zero, commit)) !=
-	    NULL)
+	    chunk_alloc_dss(tsdn, arena, new_addr, size, alignment, zero,
+	    commit)) != NULL)
 		return (ret);
 	/* mmap. */
 	if ((ret = chunk_alloc_mmap(new_addr, size, alignment, zero, commit)) !=
@@ -349,8 +359,8 @@ chunk_alloc_core(arena_t *arena, void *new_addr, size_t size, size_t alignment,
 		return (ret);
 	/* "secondary" dss. */
 	if (have_dss && dss_prec == dss_prec_secondary && (ret =
-	    chunk_alloc_dss(arena, new_addr, size, alignment, zero, commit)) !=
-	    NULL)
+	    chunk_alloc_dss(tsdn, arena, new_addr, size, alignment, zero,
+	    commit)) != NULL)
 		return (ret);

 	/* All strategies for allocation failed. */
@@ -380,35 +390,33 @@ chunk_alloc_base(size_t size)
 }

 void *
-chunk_alloc_cache(arena_t *arena, chunk_hooks_t *chunk_hooks, void *new_addr,
-    size_t size, size_t alignment, bool *zero, bool dalloc_node)
+chunk_alloc_cache(tsdn_t *tsdn, arena_t *arena, chunk_hooks_t *chunk_hooks,
+    void *new_addr, size_t size, size_t alignment, size_t *sn, bool *zero,
+    bool *commit, bool dalloc_node)
 {
 	void *ret;
-	bool commit;

 	assert(size != 0);
 	assert((size & chunksize_mask) == 0);
 	assert(alignment != 0);
 	assert((alignment & chunksize_mask) == 0);

-	commit = true;
-	ret = chunk_recycle(arena, chunk_hooks, &arena->chunks_szad_cached,
-	    &arena->chunks_ad_cached, true, new_addr, size, alignment, zero,
-	    &commit, dalloc_node);
+	ret = chunk_recycle(tsdn, arena, chunk_hooks,
+	    &arena->chunks_szsnad_cached, &arena->chunks_ad_cached, true,
+	    new_addr, size, alignment, sn, zero, commit, dalloc_node);
 	if (ret == NULL)
 		return (NULL);
-	assert(commit);
 	if (config_valgrind)
 		JEMALLOC_VALGRIND_MAKE_MEM_UNDEFINED(ret, size);
 	return (ret);
 }

 static arena_t *
-chunk_arena_get(unsigned arena_ind)
+chunk_arena_get(tsdn_t *tsdn, unsigned arena_ind)
 {
 	arena_t *arena;

-	arena = arena_get(arena_ind, false);
+	arena = arena_get(tsdn, arena_ind, false);
 	/*
 	 * The arena we're allocating on behalf of must have been initialized
 	 * already.
@@ -418,15 +426,13 @@ chunk_arena_get(unsigned arena_ind)
 }

 static void *
-chunk_alloc_default(void *new_addr, size_t size, size_t alignment, bool *zero,
-    bool *commit, unsigned arena_ind)
+chunk_alloc_default_impl(tsdn_t *tsdn, arena_t *arena, void *new_addr,
+    size_t size, size_t alignment, bool *zero, bool *commit)
 {
 	void *ret;
-	arena_t *arena;

-	arena = chunk_arena_get(arena_ind);
-	ret = chunk_alloc_core(arena, new_addr, size, alignment, zero, commit,
-	    arena->dss_prec);
+	ret = chunk_alloc_core(tsdn, arena, new_addr, size, alignment, zero,
+	    commit, arena->dss_prec);
 	if (ret == NULL)
 		return (NULL);
 	if (config_valgrind)
@@ -436,46 +442,79 @@ chunk_alloc_default(void *new_addr, size_t size, size_t alignment, bool *zero,
 }

 static void *
-chunk_alloc_retained(arena_t *arena, chunk_hooks_t *chunk_hooks, void *new_addr,
-    size_t size, size_t alignment, bool *zero, bool *commit)
+chunk_alloc_default(void *new_addr, size_t size, size_t alignment, bool *zero,
+    bool *commit, unsigned arena_ind)
 {
+	tsdn_t *tsdn;
+	arena_t *arena;
+
+	tsdn = tsdn_fetch();
+	arena = chunk_arena_get(tsdn, arena_ind);
+
+	return (chunk_alloc_default_impl(tsdn, arena, new_addr, size, alignment,
+	    zero, commit));
+}
+
+static void *
+chunk_alloc_retained(tsdn_t *tsdn, arena_t *arena, chunk_hooks_t *chunk_hooks,
+    void *new_addr, size_t size, size_t alignment, size_t *sn, bool *zero,
+    bool *commit)
+{
+	void *ret;

 	assert(size != 0);
 	assert((size & chunksize_mask) == 0);
 	assert(alignment != 0);
 	assert((alignment & chunksize_mask) == 0);

-	return (chunk_recycle(arena, chunk_hooks, &arena->chunks_szad_retained,
-	    &arena->chunks_ad_retained, false, new_addr, size, alignment, zero,
-	    commit, true));
+	ret = chunk_recycle(tsdn, arena, chunk_hooks,
+	    &arena->chunks_szsnad_retained, &arena->chunks_ad_retained, false,
+	    new_addr, size, alignment, sn, zero, commit, true);
+
+	if (config_stats && ret != NULL)
+		arena->stats.retained -= size;
+
+	return (ret);
 }

 void *
-chunk_alloc_wrapper(arena_t *arena, chunk_hooks_t *chunk_hooks, void *new_addr,
-    size_t size, size_t alignment, bool *zero, bool *commit)
+chunk_alloc_wrapper(tsdn_t *tsdn, arena_t *arena, chunk_hooks_t *chunk_hooks,
+    void *new_addr, size_t size, size_t alignment, size_t *sn, bool *zero,
+    bool *commit)
 {
 	void *ret;

-	chunk_hooks_assure_initialized(arena, chunk_hooks);
+	chunk_hooks_assure_initialized(tsdn, arena, chunk_hooks);

-	ret = chunk_alloc_retained(arena, chunk_hooks, new_addr, size,
-	    alignment, zero, commit);
+	ret = chunk_alloc_retained(tsdn, arena, chunk_hooks, new_addr, size,
+	    alignment, sn, zero, commit);
 	if (ret == NULL) {
-		ret = chunk_hooks->alloc(new_addr, size, alignment, zero,
-		    commit, arena->ind);
+		if (chunk_hooks->alloc == chunk_alloc_default) {
+			/* Call directly to propagate tsdn. */
+			ret = chunk_alloc_default_impl(tsdn, arena, new_addr,
+			    size, alignment, zero, commit);
+		} else {
+			ret = chunk_hooks->alloc(new_addr, size, alignment,
+			    zero, commit, arena->ind);
+		}
+
 		if (ret == NULL)
 			return (NULL);
+
+		*sn = arena_extent_sn_next(arena);
+
+		if (config_valgrind && chunk_hooks->alloc !=
+		    chunk_alloc_default)
+			JEMALLOC_VALGRIND_MAKE_MEM_UNDEFINED(ret, chunksize);
 	}

-	if (config_valgrind && chunk_hooks->alloc != chunk_alloc_default)
-		JEMALLOC_VALGRIND_MAKE_MEM_UNDEFINED(ret, chunksize);
 	return (ret);
 }

 static void
-chunk_record(arena_t *arena, chunk_hooks_t *chunk_hooks,
-    extent_tree_t *chunks_szad, extent_tree_t *chunks_ad, bool cache,
-    void *chunk, size_t size, bool zeroed, bool committed)
+chunk_record(tsdn_t *tsdn, arena_t *arena, chunk_hooks_t *chunk_hooks,
+    extent_tree_t *chunks_szsnad, extent_tree_t *chunks_ad, bool cache,
+    void *chunk, size_t size, size_t sn, bool zeroed, bool committed)
 {
 	bool unzeroed;
 	extent_node_t *node, *prev;
@@ -485,9 +524,9 @@ chunk_record(arena_t *arena, chunk_hooks_t *chunk_hooks,
 	unzeroed = cache || !zeroed;
 	JEMALLOC_VALGRIND_MAKE_MEM_NOACCESS(chunk, size);

-	malloc_mutex_lock(&arena->chunks_mtx);
-	chunk_hooks_assure_initialized_locked(arena, chunk_hooks);
-	extent_node_init(&key, arena, (void *)((uintptr_t)chunk + size), 0,
+	malloc_mutex_lock(tsdn, &arena->chunks_mtx);
+	chunk_hooks_assure_initialized_locked(tsdn, arena, chunk_hooks);
+	extent_node_init(&key, arena, (void *)((uintptr_t)chunk + size), 0, 0,
 	    false, false);
 	node = extent_tree_ad_nsearch(chunks_ad, &key);
 	/* Try to coalesce forward. */
@@ -499,19 +538,21 @@ chunk_record(arena_t *arena, chunk_hooks_t *chunk_hooks,
 		/*
 		 * Coalesce chunk with the following address range.  This does
 		 * not change the position within chunks_ad, so only
-		 * remove/insert from/into chunks_szad.
+		 * remove/insert from/into chunks_szsnad.
 		 */
-		extent_tree_szad_remove(chunks_szad, node);
+		extent_tree_szsnad_remove(chunks_szsnad, node);
 		arena_chunk_cache_maybe_remove(arena, node, cache);
 		extent_node_addr_set(node, chunk);
 		extent_node_size_set(node, size + extent_node_size_get(node));
+		if (sn < extent_node_sn_get(node))
+			extent_node_sn_set(node, sn);
 		extent_node_zeroed_set(node, extent_node_zeroed_get(node) &&
 		    !unzeroed);
-		extent_tree_szad_insert(chunks_szad, node);
+		extent_tree_szsnad_insert(chunks_szsnad, node);
 		arena_chunk_cache_maybe_insert(arena, node, cache);
 	} else {
 		/* Coalescing forward failed, so insert a new node. */
-		node = arena_node_alloc(arena);
+		node = arena_node_alloc(tsdn, arena);
 		if (node == NULL) {
 			/*
 			 * Node allocation failed, which is an exceedingly
@@ -520,15 +561,15 @@ chunk_record(arena_t *arena, chunk_hooks_t *chunk_hooks,
 			 * a virtual memory leak.
 			 */
 			if (cache) {
-				chunk_purge_wrapper(arena, chunk_hooks, chunk,
-				    size, 0, size);
+				chunk_purge_wrapper(tsdn, arena, chunk_hooks,
+				    chunk, size, 0, size);
 			}
 			goto label_return;
 		}
-		extent_node_init(node, arena, chunk, size, !unzeroed,
+		extent_node_init(node, arena, chunk, size, sn, !unzeroed,
 		    committed);
 		extent_tree_ad_insert(chunks_ad, node);
-		extent_tree_szad_insert(chunks_szad, node);
+		extent_tree_szsnad_insert(chunks_szsnad, node);
 		arena_chunk_cache_maybe_insert(arena, node, cache);
 	}

@@ -542,31 +583,33 @@ chunk_record(arena_t *arena, chunk_hooks_t *chunk_hooks,
 		/*
 		 * 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.
+		 * remove/insert node from/into chunks_szsnad.
 		 */
-		extent_tree_szad_remove(chunks_szad, prev);
+		extent_tree_szsnad_remove(chunks_szsnad, prev);
 		extent_tree_ad_remove(chunks_ad, prev);
 		arena_chunk_cache_maybe_remove(arena, prev, cache);
-		extent_tree_szad_remove(chunks_szad, node);
+		extent_tree_szsnad_remove(chunks_szsnad, node);
 		arena_chunk_cache_maybe_remove(arena, node, cache);
 		extent_node_addr_set(node, extent_node_addr_get(prev));
 		extent_node_size_set(node, extent_node_size_get(prev) +
 		    extent_node_size_get(node));
+		if (extent_node_sn_get(prev) < extent_node_sn_get(node))
+			extent_node_sn_set(node, extent_node_sn_get(prev));
 		extent_node_zeroed_set(node, extent_node_zeroed_get(prev) &&
 		    extent_node_zeroed_get(node));
-		extent_tree_szad_insert(chunks_szad, node);
+		extent_tree_szsnad_insert(chunks_szsnad, node);
 		arena_chunk_cache_maybe_insert(arena, node, cache);

-		arena_node_dalloc(arena, prev);
+		arena_node_dalloc(tsdn, arena, prev);
 	}

 label_return:
-	malloc_mutex_unlock(&arena->chunks_mtx);
+	malloc_mutex_unlock(tsdn, &arena->chunks_mtx);
 }

 void
-chunk_dalloc_cache(arena_t *arena, chunk_hooks_t *chunk_hooks, void *chunk,
-    size_t size, bool committed)
+chunk_dalloc_cache(tsdn_t *tsdn, arena_t *arena, chunk_hooks_t *chunk_hooks,
+    void *chunk, size_t size, size_t sn, bool committed)
 {

 	assert(chunk != NULL);
@@ -574,14 +617,14 @@ chunk_dalloc_cache(arena_t *arena, chunk_hooks_t *chunk_hooks, void *chunk,
 	assert(size != 0);
 	assert((size & chunksize_mask) == 0);

-	chunk_record(arena, chunk_hooks, &arena->chunks_szad_cached,
-	    &arena->chunks_ad_cached, true, chunk, size, false, committed);
-	arena_maybe_purge(arena);
+	chunk_record(tsdn, arena, chunk_hooks, &arena->chunks_szsnad_cached,
+	    &arena->chunks_ad_cached, true, chunk, size, sn, false,
+	    committed);
+	arena_maybe_purge(tsdn, arena);
 }

 static bool
-chunk_dalloc_default(void *chunk, size_t size, bool committed,
-    unsigned arena_ind)
+chunk_dalloc_default_impl(void *chunk, size_t size)
 {

 	if (!have_dss || !chunk_in_dss(chunk))
@@ -589,19 +632,34 @@ chunk_dalloc_default(void *chunk, size_t size, bool committed,
 	return (true);
 }

-void
-chunk_dalloc_wrapper(arena_t *arena, chunk_hooks_t *chunk_hooks, void *chunk,
-    size_t size, bool zeroed, bool committed)
+static bool
+chunk_dalloc_default(void *chunk, size_t size, bool committed,
+    unsigned arena_ind)
 {

+	return (chunk_dalloc_default_impl(chunk, size));
+}
+
+void
+chunk_dalloc_wrapper(tsdn_t *tsdn, arena_t *arena, chunk_hooks_t *chunk_hooks,
+    void *chunk, size_t size, size_t sn, bool zeroed, bool committed)
+{
+	bool err;
+
 	assert(chunk != NULL);
 	assert(CHUNK_ADDR2BASE(chunk) == chunk);
 	assert(size != 0);
 	assert((size & chunksize_mask) == 0);

-	chunk_hooks_assure_initialized(arena, chunk_hooks);
+	chunk_hooks_assure_initialized(tsdn, arena, chunk_hooks);
 	/* Try to deallocate. */
-	if (!chunk_hooks->dalloc(chunk, size, committed, arena->ind))
+	if (chunk_hooks->dalloc == chunk_dalloc_default) {
+		/* Call directly to propagate tsdn. */
+		err = chunk_dalloc_default_impl(chunk, size);
+	} else
+		err = chunk_hooks->dalloc(chunk, size, committed, arena->ind);
+
+	if (!err)
 		return;
 	/* Try to decommit; purge if that fails. */
 	if (committed) {
@@ -610,8 +668,12 @@ chunk_dalloc_wrapper(arena_t *arena, chunk_hooks_t *chunk_hooks, void *chunk,
 	}
 	zeroed = !committed || !chunk_hooks->purge(chunk, size, 0, size,
 	    arena->ind);
-	chunk_record(arena, chunk_hooks, &arena->chunks_szad_retained,
-	    &arena->chunks_ad_retained, false, chunk, size, zeroed, committed);
+	chunk_record(tsdn, arena, chunk_hooks, &arena->chunks_szsnad_retained,
+	    &arena->chunks_ad_retained, false, chunk, size, sn, zeroed,
+	    committed);
+
+	if (config_stats)
+		arena->stats.retained += size;
 }

 static bool
@@ -648,11 +710,11 @@ chunk_purge_default(void *chunk, size_t size, size_t offset, size_t length,
 }

 bool
-chunk_purge_wrapper(arena_t *arena, chunk_hooks_t *chunk_hooks, void *chunk,
-    size_t size, size_t offset, size_t length)
+chunk_purge_wrapper(tsdn_t *tsdn, arena_t *arena, chunk_hooks_t *chunk_hooks,
+    void *chunk, size_t size, size_t offset, size_t length)
 {

-	chunk_hooks_assure_initialized(arena, chunk_hooks);
+	chunk_hooks_assure_initialized(tsdn, arena, chunk_hooks);
 	return (chunk_hooks->purge(chunk, size, offset, length, arena->ind));
 }

@@ -667,23 +729,30 @@ chunk_split_default(void *chunk, size_t size, size_t size_a, size_t size_b,
 }

 static bool
-chunk_merge_default(void *chunk_a, size_t size_a, void *chunk_b, size_t size_b,
-    bool committed, unsigned arena_ind)
+chunk_merge_default_impl(void *chunk_a, void *chunk_b)
 {

 	if (!maps_coalesce)
 		return (true);
-	if (have_dss && chunk_in_dss(chunk_a) != chunk_in_dss(chunk_b))
+	if (have_dss && !chunk_dss_mergeable(chunk_a, chunk_b))
 		return (true);

 	return (false);
 }

+static bool
+chunk_merge_default(void *chunk_a, size_t size_a, void *chunk_b, size_t size_b,
+    bool committed, unsigned arena_ind)
+{
+
+	return (chunk_merge_default_impl(chunk_a, chunk_b));
+}
+
 static rtree_node_elm_t *
 chunks_rtree_node_alloc(size_t nelms)
 {

-	return ((rtree_node_elm_t *)base_alloc(nelms *
+	return ((rtree_node_elm_t *)base_alloc(TSDN_NULL, nelms *
 	    sizeof(rtree_node_elm_t)));
 }

@@ -720,32 +789,11 @@ chunk_boot(void)
 	chunksize_mask = chunksize - 1;
 	chunk_npages = (chunksize >> LG_PAGE);

-	if (have_dss && chunk_dss_boot())
-		return (true);
+	if (have_dss)
+		chunk_dss_boot();
 	if (rtree_new(&chunks_rtree, (unsigned)((ZU(1) << (LG_SIZEOF_PTR+3)) -
 	    opt_lg_chunk), chunks_rtree_node_alloc, NULL))
 		return (true);

 	return (false);
 }
-
-void
-chunk_prefork(void)
-{
-
-	chunk_dss_prefork();
-}
-
-void
-chunk_postfork_parent(void)
-{
-
-	chunk_dss_postfork_parent();
-}
-
-void
-chunk_postfork_child(void)
-{
-
-	chunk_dss_postfork_child();
-}
--- a/src/chunk_dss.c
+++ b/src/chunk_dss.c
@@ -10,20 +10,19 @@ const char	*dss_prec_names[] = {
 	"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.
+ * Current dss precedence default, used when creating new arenas.  NB: This is
+ * stored as unsigned rather than dss_prec_t because in principle there's no
+ * guarantee that sizeof(dss_prec_t) is the same as sizeof(unsigned), and we use
+ * atomic operations to synchronize the setting.
 */
-static malloc_mutex_t	dss_mtx;
+static unsigned		dss_prec_default = (unsigned)DSS_PREC_DEFAULT;

 /* 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. */
+/* Atomic boolean indicating whether the DSS is exhausted. */
+static unsigned		dss_exhausted;
+/* Atomic current upper limit on DSS addresses. */
 static void		*dss_max;

 /******************************************************************************/
@@ -47,9 +46,7 @@ chunk_dss_prec_get(void)

 	if (!have_dss)
 		return (dss_prec_disabled);
-	malloc_mutex_lock(&dss_mtx);
-	ret = dss_prec_default;
-	malloc_mutex_unlock(&dss_mtx);
+	ret = (dss_prec_t)atomic_read_u(&dss_prec_default);
 	return (ret);
 }

@@ -59,15 +56,46 @@ chunk_dss_prec_set(dss_prec_t dss_prec)

 	if (!have_dss)
 		return (dss_prec != dss_prec_disabled);
-	malloc_mutex_lock(&dss_mtx);
-	dss_prec_default = dss_prec;
-	malloc_mutex_unlock(&dss_mtx);
+	atomic_write_u(&dss_prec_default, (unsigned)dss_prec);
 	return (false);
 }

+static void *
+chunk_dss_max_update(void *new_addr)
+{
+	void *max_cur;
+	spin_t spinner;
+
+	/*
+	 * Get the current end of the DSS as max_cur and assure that dss_max is
+	 * up to date.
+	 */
+	spin_init(&spinner);
+	while (true) {
+		void *max_prev = atomic_read_p(&dss_max);
+
+		max_cur = chunk_dss_sbrk(0);
+		if ((uintptr_t)max_prev > (uintptr_t)max_cur) {
+			/*
+			 * Another thread optimistically updated dss_max.  Wait
+			 * for it to finish.
+			 */
+			spin_adaptive(&spinner);
+			continue;
+		}
+		if (!atomic_cas_p(&dss_max, max_prev, max_cur))
+			break;
+	}
+	/* Fixed new_addr can only be supported if it is at the edge of DSS. */
+	if (new_addr != NULL && max_cur != new_addr)
+		return (NULL);
+
+	return (max_cur);
+}
+
 void *
-chunk_alloc_dss(arena_t *arena, void *new_addr, size_t size, size_t alignment,
-    bool *zero, bool *commit)
+chunk_alloc_dss(tsdn_t *tsdn, arena_t *arena, void *new_addr, size_t size,
+    size_t alignment, bool *zero, bool *commit)
 {
 	cassert(have_dss);
 	assert(size > 0 && (size & chunksize_mask) == 0);
@@ -80,63 +108,71 @@ chunk_alloc_dss(arena_t *arena, void *new_addr, size_t size, size_t alignment,
 	if ((intptr_t)size < 0)
 		return (NULL);

-	malloc_mutex_lock(&dss_mtx);
-	if (dss_prev != (void *)-1) {
-
+	if (!atomic_read_u(&dss_exhausted)) {
 		/*
 		 * The loop is necessary to recover from races with other
 		 * threads that are using the DSS for something other than
 		 * malloc.
 		 */
-		do {
-			void *ret, *cpad, *dss_next;
-			size_t gap_size, cpad_size;
+		while (true) {
+			void *ret, *max_cur, *dss_next, *dss_prev;
+			void *gap_addr_chunk, *gap_addr_subchunk;
+			size_t gap_size_chunk, gap_size_subchunk;
 			intptr_t incr;
-			/* Avoid an unnecessary system call. */
-			if (new_addr != NULL && dss_max != new_addr)
-				break;

-			/* Get the current end of the DSS. */
-			dss_max = chunk_dss_sbrk(0);
-
-			/* Make sure the earlier condition still holds. */
-			if (new_addr != NULL && dss_max != new_addr)
-				break;
+			max_cur = chunk_dss_max_update(new_addr);
+			if (max_cur == NULL)
+				goto label_oom;

 			/*
-			 * 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
+			 * Compute how much chunk-aligned gap 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;
+			gap_addr_chunk = (void *)(CHUNK_CEILING(
+			    (uintptr_t)max_cur));
+			ret = (void *)ALIGNMENT_CEILING(
+			    (uintptr_t)gap_addr_chunk, alignment);
+			gap_size_chunk = (uintptr_t)ret -
+			    (uintptr_t)gap_addr_chunk;
+			/*
+			 * Compute the address just past the end of the desired
+			 * allocation space.
+			 */
 			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;
+			if ((uintptr_t)ret < (uintptr_t)max_cur ||
+			    (uintptr_t)dss_next < (uintptr_t)max_cur)
+				goto label_oom; /* Wrap-around. */
+			/* Compute the increment, including subchunk bytes. */
+			gap_addr_subchunk = max_cur;
+			gap_size_subchunk = (uintptr_t)ret -
+			    (uintptr_t)gap_addr_subchunk;
+			incr = gap_size_subchunk + size;
+
+			assert((uintptr_t)max_cur + incr == (uintptr_t)ret +
+			    size);
+
+			/*
+			 * Optimistically update dss_max, and roll back below if
+			 * sbrk() fails.  No other thread will try to extend the
+			 * DSS while dss_max is greater than the current DSS
+			 * max reported by sbrk(0).
+			 */
+			if (atomic_cas_p(&dss_max, max_cur, dss_next))
+				continue;
+
+			/* Try to allocate. */
 			dss_prev = chunk_dss_sbrk(incr);
-			if (dss_prev == dss_max) {
+			if (dss_prev == max_cur) {
 				/* Success. */
-				dss_max = dss_next;
-				malloc_mutex_unlock(&dss_mtx);
-				if (cpad_size != 0) {
+				if (gap_size_chunk != 0) {
 					chunk_hooks_t chunk_hooks =
 					    CHUNK_HOOKS_INITIALIZER;
-					chunk_dalloc_wrapper(arena,
-					    &chunk_hooks, cpad, cpad_size,
-					    false, true);
+					chunk_dalloc_wrapper(tsdn, arena,
+					    &chunk_hooks, gap_addr_chunk,
+					    gap_size_chunk,
+					    arena_extent_sn_next(arena), false,
+					    true);
 				}
 				if (*zero) {
 					JEMALLOC_VALGRIND_MAKE_MEM_UNDEFINED(
@@ -147,68 +183,65 @@ chunk_alloc_dss(arena_t *arena, void *new_addr, size_t size, size_t alignment,
 					*commit = pages_decommit(ret, size);
 				return (ret);
 			}
-		} while (dss_prev != (void *)-1);
-	}
-	malloc_mutex_unlock(&dss_mtx);

+			/*
+			 * Failure, whether due to OOM or a race with a raw
+			 * sbrk() call from outside the allocator.  Try to roll
+			 * back optimistic dss_max update; if rollback fails,
+			 * it's due to another caller of this function having
+			 * succeeded since this invocation started, in which
+			 * case rollback is not necessary.
+			 */
+			atomic_cas_p(&dss_max, dss_next, max_cur);
+			if (dss_prev == (void *)-1) {
+				/* OOM. */
+				atomic_write_u(&dss_exhausted, (unsigned)true);
+				goto label_oom;
+			}
+		}
+	}
+label_oom:
 	return (NULL);
 }

+static bool
+chunk_in_dss_helper(void *chunk, void *max)
+{
+
+	return ((uintptr_t)chunk >= (uintptr_t)dss_base && (uintptr_t)chunk <
+	    (uintptr_t)max);
+}
+
 bool
 chunk_in_dss(void *chunk)
 {
-	bool ret;

 	cassert(have_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);
+	return (chunk_in_dss_helper(chunk, atomic_read_p(&dss_max)));
 }

 bool
+chunk_dss_mergeable(void *chunk_a, void *chunk_b)
+{
+	void *max;
+
+	cassert(have_dss);
+
+	max = atomic_read_p(&dss_max);
+	return (chunk_in_dss_helper(chunk_a, max) ==
+	    chunk_in_dss_helper(chunk_b, max));
+}
+
+void
 chunk_dss_boot(void)
 {

 	cassert(have_dss);

-	if (malloc_mutex_init(&dss_mtx))
-		return (true);
 	dss_base = chunk_dss_sbrk(0);
-	dss_prev = dss_base;
+	dss_exhausted = (unsigned)(dss_base == (void *)-1);
 	dss_max = dss_base;
-
-	return (false);
-}
-
-void
-chunk_dss_prefork(void)
-{
-
-	if (have_dss)
-		malloc_mutex_prefork(&dss_mtx);
-}
-
-void
-chunk_dss_postfork_parent(void)
-{
-
-	if (have_dss)
-		malloc_mutex_postfork_parent(&dss_mtx);
-}
-
-void
-chunk_dss_postfork_child(void)
-{
-
-	if (have_dss)
-		malloc_mutex_postfork_child(&dss_mtx);
 }

 /******************************************************************************/
--- a/src/chunk_mmap.c
+++ b/src/chunk_mmap.c
@@ -16,18 +16,16 @@ chunk_alloc_mmap_slow(size_t size, size_t alignment, bool *zero, bool *commit)
 	do {
 		void *pages;
 		size_t leadsize;
-		pages = pages_map(NULL, alloc_size);
+		pages = pages_map(NULL, alloc_size, commit);
 		if (pages == NULL)
 			return (NULL);
 		leadsize = ALIGNMENT_CEILING((uintptr_t)pages, alignment) -
 		    (uintptr_t)pages;
-		ret = pages_trim(pages, alloc_size, leadsize, size);
+		ret = pages_trim(pages, alloc_size, leadsize, size, commit);
 	} while (ret == NULL);

 	assert(ret != NULL);
 	*zero = true;
-	if (!*commit)
-		*commit = pages_decommit(ret, size);
 	return (ret);
 }

@@ -54,7 +52,7 @@ chunk_alloc_mmap(void *new_addr, size_t size, size_t alignment, bool *zero,
 	assert(alignment != 0);
 	assert((alignment & chunksize_mask) == 0);

-	ret = pages_map(new_addr, size);
+	ret = pages_map(new_addr, size, commit);
 	if (ret == NULL || ret == new_addr)
 		return (ret);
 	assert(new_addr == NULL);
@@ -66,8 +64,6 @@ chunk_alloc_mmap(void *new_addr, size_t size, size_t alignment, bool *zero,

 	assert(ret != NULL);
 	*zero = true;
-	if (!*commit)
-		*commit = pages_decommit(ret, size);
 	return (ret);
 }

--- a/src/ckh.c
+++ b/src/ckh.c
@@ -99,7 +99,8 @@ 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.
 	 */
-	offset = (unsigned)prng_lg_range(&ckh->prng_state, LG_CKH_BUCKET_CELLS);
+	offset = (unsigned)prng_lg_range_u64(&ckh->prng_state,
+	    LG_CKH_BUCKET_CELLS);
 	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))];
@@ -141,7 +142,7 @@ ckh_evict_reloc_insert(ckh_t *ckh, size_t argbucket, void const **argkey,
 		 * were an item for which both hashes indicated the same
 		 * bucket.
 		 */
-		i = (unsigned)prng_lg_range(&ckh->prng_state,
+		i = (unsigned)prng_lg_range_u64(&ckh->prng_state,
 		    LG_CKH_BUCKET_CELLS);
 		cell = &ckh->tab[(bucket << LG_CKH_BUCKET_CELLS) + i];
 		assert(cell->key != NULL);
@@ -270,8 +271,8 @@ ckh_grow(tsd_t *tsd, ckh_t *ckh)
 			ret = true;
 			goto label_return;
 		}
-		tab = (ckhc_t *)ipallocztm(tsd, usize, CACHELINE, true, NULL,
-		    true, NULL);
+		tab = (ckhc_t *)ipallocztm(tsd_tsdn(tsd), usize, CACHELINE,
+		    true, NULL, true, arena_ichoose(tsd, NULL));
 		if (tab == NULL) {
 			ret = true;
 			goto label_return;
@@ -283,12 +284,12 @@ ckh_grow(tsd_t *tsd, ckh_t *ckh)
 		ckh->lg_curbuckets = lg_curcells - LG_CKH_BUCKET_CELLS;

 		if (!ckh_rebuild(ckh, tab)) {
-			idalloctm(tsd, tab, tcache_get(tsd, false), true, true);
+			idalloctm(tsd_tsdn(tsd), tab, NULL, true, true);
 			break;
 		}

 		/* Rebuilding failed, so back out partially rebuilt table. */
-		idalloctm(tsd, ckh->tab, tcache_get(tsd, false), true, true);
+		idalloctm(tsd_tsdn(tsd), ckh->tab, NULL, true, true);
 		ckh->tab = tab;
 		ckh->lg_curbuckets = lg_prevbuckets;
 	}
@@ -314,8 +315,8 @@ ckh_shrink(tsd_t *tsd, ckh_t *ckh)
 	usize = sa2u(sizeof(ckhc_t) << lg_curcells, CACHELINE);
 	if (unlikely(usize == 0 || usize > HUGE_MAXCLASS))
 		return;
-	tab = (ckhc_t *)ipallocztm(tsd, usize, CACHELINE, true, NULL, true,
-	    NULL);
+	tab = (ckhc_t *)ipallocztm(tsd_tsdn(tsd), usize, CACHELINE, true, NULL,
+	    true, arena_ichoose(tsd, NULL));
 	if (tab == NULL) {
 		/*
 		 * An OOM error isn't worth propagating, since it doesn't
@@ -330,7 +331,7 @@ ckh_shrink(tsd_t *tsd, ckh_t *ckh)
 	ckh->lg_curbuckets = lg_curcells - LG_CKH_BUCKET_CELLS;

 	if (!ckh_rebuild(ckh, tab)) {
-		idalloctm(tsd, tab, tcache_get(tsd, false), true, true);
+		idalloctm(tsd_tsdn(tsd), tab, NULL, true, true);
 #ifdef CKH_COUNT
 		ckh->nshrinks++;
 #endif
@@ -338,7 +339,7 @@ ckh_shrink(tsd_t *tsd, ckh_t *ckh)
 	}

 	/* Rebuilding failed, so back out partially rebuilt table. */
-	idalloctm(tsd, ckh->tab, tcache_get(tsd, false), true, true);
+	idalloctm(tsd_tsdn(tsd), ckh->tab, NULL, true, true);
 	ckh->tab = tab;
 	ckh->lg_curbuckets = lg_prevbuckets;
 #ifdef CKH_COUNT
@@ -391,8 +392,8 @@ ckh_new(tsd_t *tsd, ckh_t *ckh, size_t minitems, ckh_hash_t *hash,
 		ret = true;
 		goto label_return;
 	}
-	ckh->tab = (ckhc_t *)ipallocztm(tsd, usize, CACHELINE, true, NULL, true,
-	    NULL);
+	ckh->tab = (ckhc_t *)ipallocztm(tsd_tsdn(tsd), usize, CACHELINE, true,
+	    NULL, true, arena_ichoose(tsd, NULL));
 	if (ckh->tab == NULL) {
 		ret = true;
 		goto label_return;
@@ -421,9 +422,9 @@ ckh_delete(tsd_t *tsd, ckh_t *ckh)
 	    (unsigned long long)ckh->nrelocs);
 #endif

-	idalloctm(tsd, ckh->tab, tcache_get(tsd, false), true, true);
+	idalloctm(tsd_tsdn(tsd), ckh->tab, NULL, true, true);
 	if (config_debug)
-		memset(ckh, 0x5a, sizeof(ckh_t));
+		memset(ckh, JEMALLOC_FREE_JUNK, sizeof(ckh_t));
 }

 size_t
--- a/src/ctl.c
+++ b/src/ctl.c
--- a/src/extent.c
+++ b/src/extent.c
@@ -3,42 +3,64 @@

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

-JEMALLOC_INLINE_C size_t
-extent_quantize(size_t size)
-{
+#ifndef JEMALLOC_JET
+static
+#endif
+size_t
+extent_size_quantize_floor(size_t size) {
+	size_t ret;
+	szind_t ind;

-	/*
-	 * Round down to the nearest chunk size that can actually be requested
-	 * during normal huge allocation.
-	 */
-	return (index2size(size2index(size + 1) - 1));
-}
+	assert(size > 0);

-JEMALLOC_INLINE_C int
-extent_szad_comp(const extent_node_t *a, const extent_node_t *b)
-{
-	int ret;
-	size_t a_qsize = extent_quantize(extent_node_size_get(a));
-	size_t b_qsize = extent_quantize(extent_node_size_get(b));
-
-	/*
-	 * Compare based on quantized size rather than size, in order to sort
-	 * equally useful extents only by address.
-	 */
-	ret = (a_qsize > b_qsize) - (a_qsize < b_qsize);
-	if (ret == 0) {
-		uintptr_t a_addr = (uintptr_t)extent_node_addr_get(a);
-		uintptr_t b_addr = (uintptr_t)extent_node_addr_get(b);
-
-		ret = (a_addr > b_addr) - (a_addr < b_addr);
+	ind = size2index(size + 1);
+	if (ind == 0) {
+		/* Avoid underflow. */
+		return (index2size(0));
 	}
-
+	ret = index2size(ind - 1);
+	assert(ret <= size);
 	return (ret);
 }

-/* Generate red-black tree functions. */
-rb_gen(, extent_tree_szad_, extent_tree_t, extent_node_t, szad_link,
-    extent_szad_comp)
+size_t
+extent_size_quantize_ceil(size_t size) {
+	size_t ret;
+
+	assert(size > 0);
+
+	ret = extent_size_quantize_floor(size);
+	if (ret < size) {
+		/*
+		 * Skip a quantization that may have an adequately large extent,
+		 * because under-sized extents may be mixed in.  This only
+		 * happens when an unusual size is requested, i.e. for aligned
+		 * allocation, and is just one of several places where linear
+		 * search would potentially find sufficiently aligned available
+		 * memory somewhere lower.
+		 */
+		ret = index2size(size2index(ret  + 1));
+	}
+	return ret;
+}
+
+JEMALLOC_INLINE_C int
+extent_sz_comp(const extent_node_t *a, const extent_node_t *b)
+{
+	size_t a_qsize = extent_size_quantize_floor(extent_node_size_get(a));
+	size_t b_qsize = extent_size_quantize_floor(extent_node_size_get(b));
+
+	return ((a_qsize > b_qsize) - (a_qsize < b_qsize));
+}
+
+JEMALLOC_INLINE_C int
+extent_sn_comp(const extent_node_t *a, const extent_node_t *b)
+{
+	size_t a_sn = extent_node_sn_get(a);
+	size_t b_sn = extent_node_sn_get(b);
+
+	return ((a_sn > b_sn) - (a_sn < b_sn));
+}

 JEMALLOC_INLINE_C int
 extent_ad_comp(const extent_node_t *a, const extent_node_t *b)
@@ -49,5 +71,26 @@ extent_ad_comp(const extent_node_t *a, const extent_node_t *b)
 	return ((a_addr > b_addr) - (a_addr < b_addr));
 }

+JEMALLOC_INLINE_C int
+extent_szsnad_comp(const extent_node_t *a, const extent_node_t *b)
+{
+	int ret;
+
+	ret = extent_sz_comp(a, b);
+	if (ret != 0)
+		return (ret);
+
+	ret = extent_sn_comp(a, b);
+	if (ret != 0)
+		return (ret);
+
+	ret = extent_ad_comp(a, b);
+	return (ret);
+}
+
+/* Generate red-black tree functions. */
+rb_gen(, extent_tree_szsnad_, extent_tree_t, extent_node_t, szsnad_link,
+    extent_szsnad_comp)
+
 /* Generate red-black tree functions. */
 rb_gen(, extent_tree_ad_, extent_tree_t, extent_node_t, ad_link, extent_ad_comp)
--- a/src/huge.c
+++ b/src/huge.c
@@ -15,12 +15,21 @@ huge_node_get(const void *ptr)
 }

 static bool
-huge_node_set(const void *ptr, extent_node_t *node)
+huge_node_set(tsdn_t *tsdn, const void *ptr, extent_node_t *node, bool *gdump)
 {

 	assert(extent_node_addr_get(node) == ptr);
 	assert(!extent_node_achunk_get(node));
-	return (chunk_register(ptr, node));
+	return (chunk_register(ptr, node, gdump));
+}
+
+static void
+huge_node_reset(tsdn_t *tsdn, const void *ptr, extent_node_t *node, bool *gdump)
+{
+	bool err;
+
+	err = huge_node_set(tsdn, ptr, node, gdump);
+	assert(!err);
 }

 static void
@@ -31,34 +40,41 @@ huge_node_unset(const void *ptr, const extent_node_t *node)
 }

 void *
-huge_malloc(tsd_t *tsd, arena_t *arena, size_t usize, bool zero,
-    tcache_t *tcache)
+huge_malloc(tsdn_t *tsdn, arena_t *arena, size_t usize, bool zero)
 {

 	assert(usize == s2u(usize));

-	return (huge_palloc(tsd, arena, usize, chunksize, zero, tcache));
+	return (huge_palloc(tsdn, arena, usize, chunksize, zero));
 }

 void *
-huge_palloc(tsd_t *tsd, arena_t *arena, size_t usize, size_t alignment,
-    bool zero, tcache_t *tcache)
+huge_palloc(tsdn_t *tsdn, arena_t *arena, size_t usize, size_t alignment,
+    bool zero)
 {
 	void *ret;
 	size_t ausize;
+	arena_t *iarena;
 	extent_node_t *node;
-	bool is_zeroed;
+	size_t sn;
+	bool is_zeroed, gdump;

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

+	assert(!tsdn_null(tsdn) || arena != NULL);
+	/* prof_gdump() requirement. */
+	witness_assert_depth_to_rank(tsdn, WITNESS_RANK_CORE, 0);
+
 	ausize = sa2u(usize, alignment);
 	if (unlikely(ausize == 0 || ausize > HUGE_MAXCLASS))
 		return (NULL);
 	assert(ausize >= chunksize);

 	/* Allocate an extent node with which to track the chunk. */
-	node = ipallocztm(tsd, CACHELINE_CEILING(sizeof(extent_node_t)),
-	    CACHELINE, false, tcache, true, arena);
+	iarena = (!tsdn_null(tsdn)) ? arena_ichoose(tsdn_tsd(tsdn), NULL) :
+	    a0get();
+	node = ipallocztm(tsdn, CACHELINE_CEILING(sizeof(extent_node_t)),
+	    CACHELINE, false, NULL, true, iarena);
 	if (node == NULL)
 		return (NULL);

@@ -67,34 +83,37 @@ huge_palloc(tsd_t *tsd, arena_t *arena, size_t usize, size_t alignment,
 	 * it is possible to make correct junk/zero fill decisions below.
 	 */
 	is_zeroed = zero;
-	arena = arena_choose(tsd, arena);
-	if (unlikely(arena == NULL) || (ret = arena_chunk_alloc_huge(arena,
-	    usize, alignment, &is_zeroed)) == NULL) {
-		idalloctm(tsd, node, tcache, true, true);
+	if (likely(!tsdn_null(tsdn)))
+		arena = arena_choose(tsdn_tsd(tsdn), arena);
+	if (unlikely(arena == NULL) || (ret = arena_chunk_alloc_huge(tsdn,
+	    arena, usize, alignment, &sn, &is_zeroed)) == NULL) {
+		idalloctm(tsdn, node, NULL, true, true);
 		return (NULL);
 	}

-	extent_node_init(node, arena, ret, usize, is_zeroed, true);
+	extent_node_init(node, arena, ret, usize, sn, is_zeroed, true);

-	if (huge_node_set(ret, node)) {
-		arena_chunk_dalloc_huge(arena, ret, usize);
-		idalloctm(tsd, node, tcache, true, true);
+	if (huge_node_set(tsdn, ret, node, &gdump)) {
+		arena_chunk_dalloc_huge(tsdn, arena, ret, usize, sn);
+		idalloctm(tsdn, node, NULL, true, true);
 		return (NULL);
 	}
+	if (config_prof && opt_prof && gdump)
+		prof_gdump(tsdn);

 	/* Insert node into huge. */
-	malloc_mutex_lock(&arena->huge_mtx);
+	malloc_mutex_lock(tsdn, &arena->huge_mtx);
 	ql_elm_new(node, ql_link);
 	ql_tail_insert(&arena->huge, node, ql_link);
-	malloc_mutex_unlock(&arena->huge_mtx);
+	malloc_mutex_unlock(tsdn, &arena->huge_mtx);

 	if (zero || (config_fill && unlikely(opt_zero))) {
 		if (!is_zeroed)
 			memset(ret, 0, usize);
 	} else if (config_fill && unlikely(opt_junk_alloc))
-		memset(ret, 0xa5, usize);
+		memset(ret, JEMALLOC_ALLOC_JUNK, usize);

-	arena_decay_tick(tsd, arena);
+	arena_decay_tick(tsdn, arena);
 	return (ret);
 }

@@ -112,7 +131,7 @@ huge_dalloc_junk(void *ptr, size_t usize)
 		 * unmapped.
 		 */
 		if (!config_munmap || (have_dss && chunk_in_dss(ptr)))
-			memset(ptr, 0x5a, usize);
+			memset(ptr, JEMALLOC_FREE_JUNK, usize);
 	}
 }
 #ifdef JEMALLOC_JET
@@ -122,14 +141,17 @@ huge_dalloc_junk_t *huge_dalloc_junk = JEMALLOC_N(huge_dalloc_junk_impl);
 #endif

 static void
-huge_ralloc_no_move_similar(void *ptr, size_t oldsize, size_t usize_min,
-    size_t usize_max, bool zero)
+huge_ralloc_no_move_similar(tsdn_t *tsdn, void *ptr, size_t oldsize,
+    size_t usize_min, size_t usize_max, bool zero)
 {
 	size_t usize, usize_next;
 	extent_node_t *node;
 	arena_t *arena;
 	chunk_hooks_t chunk_hooks = CHUNK_HOOKS_INITIALIZER;
-	bool pre_zeroed, post_zeroed;
+	bool pre_zeroed, post_zeroed, gdump;
+
+	/* prof_gdump() requirement. */
+	witness_assert_depth_to_rank(tsdn, WITNESS_RANK_CORE, 0);

 	/* Increase usize to incorporate extra. */
 	for (usize = usize_min; usize < usize_max && (usize_next = s2u(usize+1))
@@ -147,24 +169,31 @@ huge_ralloc_no_move_similar(void *ptr, size_t oldsize, size_t usize_min,
 	if (oldsize > usize) {
 		size_t sdiff = oldsize - usize;
 		if (config_fill && unlikely(opt_junk_free)) {
-			memset((void *)((uintptr_t)ptr + usize), 0x5a, sdiff);
+			memset((void *)((uintptr_t)ptr + usize),
+			    JEMALLOC_FREE_JUNK, sdiff);
 			post_zeroed = false;
 		} else {
-			post_zeroed = !chunk_purge_wrapper(arena, &chunk_hooks,
-			    ptr, CHUNK_CEILING(oldsize), usize, sdiff);
+			post_zeroed = !chunk_purge_wrapper(tsdn, arena,
+			    &chunk_hooks, ptr, CHUNK_CEILING(oldsize), usize,
+			    sdiff);
 		}
 	} else
 		post_zeroed = pre_zeroed;

-	malloc_mutex_lock(&arena->huge_mtx);
+	malloc_mutex_lock(tsdn, &arena->huge_mtx);
 	/* Update the size of the huge allocation. */
+	huge_node_unset(ptr, node);
 	assert(extent_node_size_get(node) != usize);
 	extent_node_size_set(node, usize);
+	huge_node_reset(tsdn, ptr, node, &gdump);
 	/* Update zeroed. */
 	extent_node_zeroed_set(node, post_zeroed);
-	malloc_mutex_unlock(&arena->huge_mtx);
+	malloc_mutex_unlock(tsdn, &arena->huge_mtx);
+	/* gdump without any locks held. */
+	if (config_prof && opt_prof && gdump)
+		prof_gdump(tsdn);

-	arena_chunk_ralloc_huge_similar(arena, ptr, oldsize, usize);
+	arena_chunk_ralloc_huge_similar(tsdn, arena, ptr, oldsize, usize);

 	/* Fill if necessary (growing). */
 	if (oldsize < usize) {
@@ -174,27 +203,30 @@ huge_ralloc_no_move_similar(void *ptr, size_t oldsize, size_t usize_min,
 				    usize - oldsize);
 			}
 		} else if (config_fill && unlikely(opt_junk_alloc)) {
-			memset((void *)((uintptr_t)ptr + oldsize), 0xa5, usize -
-			    oldsize);
+			memset((void *)((uintptr_t)ptr + oldsize),
+			    JEMALLOC_ALLOC_JUNK, usize - oldsize);
 		}
 	}
 }

 static bool
-huge_ralloc_no_move_shrink(void *ptr, size_t oldsize, size_t usize)
+huge_ralloc_no_move_shrink(tsdn_t *tsdn, void *ptr, size_t oldsize,
+    size_t usize)
 {
 	extent_node_t *node;
 	arena_t *arena;
 	chunk_hooks_t chunk_hooks;
 	size_t cdiff;
-	bool pre_zeroed, post_zeroed;
+	bool pre_zeroed, post_zeroed, gdump;

 	node = huge_node_get(ptr);
 	arena = extent_node_arena_get(node);
 	pre_zeroed = extent_node_zeroed_get(node);
-	chunk_hooks = chunk_hooks_get(arena);
+	chunk_hooks = chunk_hooks_get(tsdn, arena);

 	assert(oldsize > usize);
+	/* prof_gdump() requirement. */
+	witness_assert_depth_to_rank(tsdn, WITNESS_RANK_CORE, 0);

 	/* Split excess chunks. */
 	cdiff = CHUNK_CEILING(oldsize) - CHUNK_CEILING(usize);
@@ -209,53 +241,68 @@ huge_ralloc_no_move_shrink(void *ptr, size_t oldsize, size_t usize)
 			    sdiff);
 			post_zeroed = false;
 		} else {
-			post_zeroed = !chunk_purge_wrapper(arena, &chunk_hooks,
-			    CHUNK_ADDR2BASE((uintptr_t)ptr + usize),
-			    CHUNK_CEILING(oldsize),
+			post_zeroed = !chunk_purge_wrapper(tsdn, arena,
+			    &chunk_hooks, CHUNK_ADDR2BASE((uintptr_t)ptr +
+			    usize), CHUNK_CEILING(oldsize),
 			    CHUNK_ADDR2OFFSET((uintptr_t)ptr + usize), sdiff);
 		}
 	} else
 		post_zeroed = pre_zeroed;

-	malloc_mutex_lock(&arena->huge_mtx);
+	malloc_mutex_lock(tsdn, &arena->huge_mtx);
 	/* Update the size of the huge allocation. */
+	huge_node_unset(ptr, node);
 	extent_node_size_set(node, usize);
+	huge_node_reset(tsdn, ptr, node, &gdump);
 	/* Update zeroed. */
 	extent_node_zeroed_set(node, post_zeroed);
-	malloc_mutex_unlock(&arena->huge_mtx);
+	malloc_mutex_unlock(tsdn, &arena->huge_mtx);
+	/* gdump without any locks held. */
+	if (config_prof && opt_prof && gdump)
+		prof_gdump(tsdn);

 	/* Zap the excess chunks. */
-	arena_chunk_ralloc_huge_shrink(arena, ptr, oldsize, usize);
+	arena_chunk_ralloc_huge_shrink(tsdn, arena, ptr, oldsize, usize,
+	    extent_node_sn_get(node));

 	return (false);
 }

 static bool
-huge_ralloc_no_move_expand(void *ptr, size_t oldsize, size_t usize, bool zero) {
+huge_ralloc_no_move_expand(tsdn_t *tsdn, void *ptr, size_t oldsize,
+    size_t usize, bool zero) {
 	extent_node_t *node;
 	arena_t *arena;
-	bool is_zeroed_subchunk, is_zeroed_chunk;
+	bool is_zeroed_subchunk, is_zeroed_chunk, gdump;

 	node = huge_node_get(ptr);
 	arena = extent_node_arena_get(node);
-	malloc_mutex_lock(&arena->huge_mtx);
+	malloc_mutex_lock(tsdn, &arena->huge_mtx);
 	is_zeroed_subchunk = extent_node_zeroed_get(node);
-	malloc_mutex_unlock(&arena->huge_mtx);
+	malloc_mutex_unlock(tsdn, &arena->huge_mtx);
+
+	/* prof_gdump() requirement. */
+	witness_assert_depth_to_rank(tsdn, WITNESS_RANK_CORE, 0);

 	/*
-	 * Copy zero into is_zeroed_chunk and pass the copy to chunk_alloc(), so
-	 * that it is possible to make correct junk/zero fill decisions below.
+	 * Use is_zeroed_chunk to detect whether the trailing memory is zeroed,
+	 * update extent's zeroed field, and zero as necessary.
 	 */
-	is_zeroed_chunk = zero;
-
-	if (arena_chunk_ralloc_huge_expand(arena, ptr, oldsize, usize,
+	is_zeroed_chunk = false;
+	if (arena_chunk_ralloc_huge_expand(tsdn, arena, ptr, oldsize, usize,
 	     &is_zeroed_chunk))
 		return (true);

-	malloc_mutex_lock(&arena->huge_mtx);
-	/* Update the size of the huge allocation. */
+	malloc_mutex_lock(tsdn, &arena->huge_mtx);
+	huge_node_unset(ptr, node);
 	extent_node_size_set(node, usize);
-	malloc_mutex_unlock(&arena->huge_mtx);
+	extent_node_zeroed_set(node, extent_node_zeroed_get(node) &&
+	    is_zeroed_chunk);
+	huge_node_reset(tsdn, ptr, node, &gdump);
+	malloc_mutex_unlock(tsdn, &arena->huge_mtx);
+	/* gdump without any locks held. */
+	if (config_prof && opt_prof && gdump)
+		prof_gdump(tsdn);

 	if (zero || (config_fill && unlikely(opt_zero))) {
 		if (!is_zeroed_subchunk) {
@@ -268,15 +315,15 @@ huge_ralloc_no_move_expand(void *ptr, size_t oldsize, size_t usize, bool zero) {
 			    CHUNK_CEILING(oldsize));
 		}
 	} else if (config_fill && unlikely(opt_junk_alloc)) {
-		memset((void *)((uintptr_t)ptr + oldsize), 0xa5, usize -
-		    oldsize);
+		memset((void *)((uintptr_t)ptr + oldsize), JEMALLOC_ALLOC_JUNK,
+		    usize - oldsize);
 	}

 	return (false);
 }

 bool
-huge_ralloc_no_move(tsd_t *tsd, void *ptr, size_t oldsize, size_t usize_min,
+huge_ralloc_no_move(tsdn_t *tsdn, void *ptr, size_t oldsize, size_t usize_min,
    size_t usize_max, bool zero)
 {

@@ -290,16 +337,16 @@ huge_ralloc_no_move(tsd_t *tsd, void *ptr, size_t oldsize, size_t usize_min,

 	if (CHUNK_CEILING(usize_max) > CHUNK_CEILING(oldsize)) {
 		/* Attempt to expand the allocation in-place. */
-		if (!huge_ralloc_no_move_expand(ptr, oldsize, usize_max,
+		if (!huge_ralloc_no_move_expand(tsdn, ptr, oldsize, usize_max,
 		    zero)) {
-			arena_decay_tick(tsd, huge_aalloc(ptr));
+			arena_decay_tick(tsdn, huge_aalloc(ptr));
 			return (false);
 		}
 		/* Try again, this time with usize_min. */
 		if (usize_min < usize_max && CHUNK_CEILING(usize_min) >
-		    CHUNK_CEILING(oldsize) && huge_ralloc_no_move_expand(ptr,
-		    oldsize, usize_min, zero)) {
-			arena_decay_tick(tsd, huge_aalloc(ptr));
+		    CHUNK_CEILING(oldsize) && huge_ralloc_no_move_expand(tsdn,
+		    ptr, oldsize, usize_min, zero)) {
+			arena_decay_tick(tsdn, huge_aalloc(ptr));
 			return (false);
 		}
 	}
@@ -310,16 +357,17 @@ huge_ralloc_no_move(tsd_t *tsd, void *ptr, size_t oldsize, size_t usize_min,
 	 */
 	if (CHUNK_CEILING(oldsize) >= CHUNK_CEILING(usize_min)
 	    && CHUNK_CEILING(oldsize) <= CHUNK_CEILING(usize_max)) {
-		huge_ralloc_no_move_similar(ptr, oldsize, usize_min, usize_max,
-		    zero);
-		arena_decay_tick(tsd, huge_aalloc(ptr));
+		huge_ralloc_no_move_similar(tsdn, ptr, oldsize, usize_min,
+		    usize_max, zero);
+		arena_decay_tick(tsdn, huge_aalloc(ptr));
 		return (false);
 	}

 	/* Attempt to shrink the allocation in-place. */
 	if (CHUNK_CEILING(oldsize) > CHUNK_CEILING(usize_max)) {
-		if (!huge_ralloc_no_move_shrink(ptr, oldsize, usize_max)) {
-			arena_decay_tick(tsd, huge_aalloc(ptr));
+		if (!huge_ralloc_no_move_shrink(tsdn, ptr, oldsize,
+		    usize_max)) {
+			arena_decay_tick(tsdn, huge_aalloc(ptr));
 			return (false);
 		}
 	}
@@ -327,18 +375,18 @@ huge_ralloc_no_move(tsd_t *tsd, void *ptr, size_t oldsize, size_t usize_min,
 }

 static void *
-huge_ralloc_move_helper(tsd_t *tsd, arena_t *arena, size_t usize,
-    size_t alignment, bool zero, tcache_t *tcache)
+huge_ralloc_move_helper(tsdn_t *tsdn, arena_t *arena, size_t usize,
+    size_t alignment, bool zero)
 {

 	if (alignment <= chunksize)
-		return (huge_malloc(tsd, arena, usize, zero, tcache));
-	return (huge_palloc(tsd, arena, usize, alignment, zero, tcache));
+		return (huge_malloc(tsdn, arena, usize, zero));
+	return (huge_palloc(tsdn, arena, usize, alignment, zero));
 }

 void *
-huge_ralloc(tsd_t *tsd, arena_t *arena, void *ptr, size_t oldsize, size_t usize,
-    size_t alignment, bool zero, tcache_t *tcache)
+huge_ralloc(tsd_t *tsd, arena_t *arena, void *ptr, size_t oldsize,
+    size_t usize, size_t alignment, bool zero, tcache_t *tcache)
 {
 	void *ret;
 	size_t copysize;
@@ -347,7 +395,8 @@ huge_ralloc(tsd_t *tsd, arena_t *arena, void *ptr, size_t oldsize, size_t usize,
 	assert(usize > 0 && usize <= HUGE_MAXCLASS);

 	/* Try to avoid moving the allocation. */
-	if (!huge_ralloc_no_move(tsd, ptr, oldsize, usize, usize, zero))
+	if (!huge_ralloc_no_move(tsd_tsdn(tsd), ptr, oldsize, usize, usize,
+	    zero))
 		return (ptr);

 	/*
@@ -355,19 +404,19 @@ huge_ralloc(tsd_t *tsd, arena_t *arena, void *ptr, size_t oldsize, size_t usize,
 	 * different size class.  In that case, fall back to allocating new
 	 * space and copying.
 	 */
-	ret = huge_ralloc_move_helper(tsd, arena, usize, alignment, zero,
-	    tcache);
+	ret = huge_ralloc_move_helper(tsd_tsdn(tsd), arena, usize, alignment,
+	    zero);
 	if (ret == NULL)
 		return (NULL);

 	copysize = (usize < oldsize) ? usize : oldsize;
 	memcpy(ret, ptr, copysize);
-	isqalloc(tsd, ptr, oldsize, tcache);
+	isqalloc(tsd, ptr, oldsize, tcache, true);
 	return (ret);
 }

 void
-huge_dalloc(tsd_t *tsd, void *ptr, tcache_t *tcache)
+huge_dalloc(tsdn_t *tsdn, void *ptr)
 {
 	extent_node_t *node;
 	arena_t *arena;
@@ -375,17 +424,18 @@ huge_dalloc(tsd_t *tsd, void *ptr, tcache_t *tcache)
 	node = huge_node_get(ptr);
 	arena = extent_node_arena_get(node);
 	huge_node_unset(ptr, node);
-	malloc_mutex_lock(&arena->huge_mtx);
+	malloc_mutex_lock(tsdn, &arena->huge_mtx);
 	ql_remove(&arena->huge, node, ql_link);
-	malloc_mutex_unlock(&arena->huge_mtx);
+	malloc_mutex_unlock(tsdn, &arena->huge_mtx);

 	huge_dalloc_junk(extent_node_addr_get(node),
 	    extent_node_size_get(node));
-	arena_chunk_dalloc_huge(extent_node_arena_get(node),
-	    extent_node_addr_get(node), extent_node_size_get(node));
-	idalloctm(tsd, node, tcache, true, true);
+	arena_chunk_dalloc_huge(tsdn, extent_node_arena_get(node),
+	    extent_node_addr_get(node), extent_node_size_get(node),
+	    extent_node_sn_get(node));
+	idalloctm(tsdn, node, NULL, true, true);

-	arena_decay_tick(tsd, arena);
+	arena_decay_tick(tsdn, arena);
 }

 arena_t *
@@ -396,7 +446,7 @@ huge_aalloc(const void *ptr)
 }

 size_t
-huge_salloc(const void *ptr)
+huge_salloc(tsdn_t *tsdn, const void *ptr)
 {
 	size_t size;
 	extent_node_t *node;
@@ -404,15 +454,15 @@ huge_salloc(const void *ptr)

 	node = huge_node_get(ptr);
 	arena = extent_node_arena_get(node);
-	malloc_mutex_lock(&arena->huge_mtx);
+	malloc_mutex_lock(tsdn, &arena->huge_mtx);
 	size = extent_node_size_get(node);
-	malloc_mutex_unlock(&arena->huge_mtx);
+	malloc_mutex_unlock(tsdn, &arena->huge_mtx);

 	return (size);
 }

 prof_tctx_t *
-huge_prof_tctx_get(const void *ptr)
+huge_prof_tctx_get(tsdn_t *tsdn, const void *ptr)
 {
 	prof_tctx_t *tctx;
 	extent_node_t *node;
@@ -420,29 +470,29 @@ huge_prof_tctx_get(const void *ptr)

 	node = huge_node_get(ptr);
 	arena = extent_node_arena_get(node);
-	malloc_mutex_lock(&arena->huge_mtx);
+	malloc_mutex_lock(tsdn, &arena->huge_mtx);
 	tctx = extent_node_prof_tctx_get(node);
-	malloc_mutex_unlock(&arena->huge_mtx);
+	malloc_mutex_unlock(tsdn, &arena->huge_mtx);

 	return (tctx);
 }

 void
-huge_prof_tctx_set(const void *ptr, prof_tctx_t *tctx)
+huge_prof_tctx_set(tsdn_t *tsdn, const void *ptr, prof_tctx_t *tctx)
 {
 	extent_node_t *node;
 	arena_t *arena;

 	node = huge_node_get(ptr);
 	arena = extent_node_arena_get(node);
-	malloc_mutex_lock(&arena->huge_mtx);
+	malloc_mutex_lock(tsdn, &arena->huge_mtx);
 	extent_node_prof_tctx_set(node, tctx);
-	malloc_mutex_unlock(&arena->huge_mtx);
+	malloc_mutex_unlock(tsdn, &arena->huge_mtx);
 }

 void
-huge_prof_tctx_reset(const void *ptr)
+huge_prof_tctx_reset(tsdn_t *tsdn, const void *ptr)
 {

-	huge_prof_tctx_set(ptr, (prof_tctx_t *)(uintptr_t)1U);
+	huge_prof_tctx_set(tsdn, ptr, (prof_tctx_t *)(uintptr_t)1U);
 }
--- a/src/jemalloc.c
+++ b/src/jemalloc.c
--- a/src/mutex.c
+++ b/src/mutex.c
@@ -69,7 +69,7 @@ JEMALLOC_EXPORT int	_pthread_mutex_init_calloc_cb(pthread_mutex_t *mutex,
 #endif

 bool
-malloc_mutex_init(malloc_mutex_t *mutex)
+malloc_mutex_init(malloc_mutex_t *mutex, const char *name, witness_rank_t rank)
 {

 #ifdef _WIN32
@@ -80,6 +80,8 @@ malloc_mutex_init(malloc_mutex_t *mutex)
 	    _CRT_SPINCOUNT))
 		return (true);
 #  endif
+#elif (defined(JEMALLOC_OS_UNFAIR_LOCK))
+	mutex->lock = OS_UNFAIR_LOCK_INIT;
 #elif (defined(JEMALLOC_OSSPIN))
 	mutex->lock = 0;
 #elif (defined(JEMALLOC_MUTEX_INIT_CB))
@@ -103,31 +105,34 @@ malloc_mutex_init(malloc_mutex_t *mutex)
 	}
 	pthread_mutexattr_destroy(&attr);
 #endif
+	if (config_debug)
+		witness_init(&mutex->witness, name, rank, NULL);
 	return (false);
 }

 void
-malloc_mutex_prefork(malloc_mutex_t *mutex)
+malloc_mutex_prefork(tsdn_t *tsdn, malloc_mutex_t *mutex)
 {

-	malloc_mutex_lock(mutex);
+	malloc_mutex_lock(tsdn, mutex);
 }

 void
-malloc_mutex_postfork_parent(malloc_mutex_t *mutex)
+malloc_mutex_postfork_parent(tsdn_t *tsdn, malloc_mutex_t *mutex)
 {

-	malloc_mutex_unlock(mutex);
+	malloc_mutex_unlock(tsdn, mutex);
 }

 void
-malloc_mutex_postfork_child(malloc_mutex_t *mutex)
+malloc_mutex_postfork_child(tsdn_t *tsdn, malloc_mutex_t *mutex)
 {

 #ifdef JEMALLOC_MUTEX_INIT_CB
-	malloc_mutex_unlock(mutex);
+	malloc_mutex_unlock(tsdn, mutex);
 #else
-	if (malloc_mutex_init(mutex)) {
+	if (malloc_mutex_init(mutex, mutex->witness.name,
+	    mutex->witness.rank)) {
 		malloc_printf("<jemalloc>: Error re-initializing mutex in "
 		    "child\n");
 		if (opt_abort)
@@ -137,7 +142,7 @@ malloc_mutex_postfork_child(malloc_mutex_t *mutex)
 }

 bool
-mutex_boot(void)
+malloc_mutex_boot(void)
 {

 #ifdef JEMALLOC_MUTEX_INIT_CB
--- a/src/nstime.c
+++ b/src/nstime.c
@@ -97,9 +97,79 @@ nstime_divide(const nstime_t *time, const nstime_t *divisor)
 	return (time->ns / divisor->ns);
 }

+#ifdef _WIN32
+#  define NSTIME_MONOTONIC true
+static void
+nstime_get(nstime_t *time)
+{
+	FILETIME ft;
+	uint64_t ticks_100ns;
+
+	GetSystemTimeAsFileTime(&ft);
+	ticks_100ns = (((uint64_t)ft.dwHighDateTime) << 32) | ft.dwLowDateTime;
+
+	nstime_init(time, ticks_100ns * 100);
+}
+#elif JEMALLOC_HAVE_CLOCK_MONOTONIC_COARSE
+#  define NSTIME_MONOTONIC true
+static void
+nstime_get(nstime_t *time)
+{
+	struct timespec ts;
+
+	clock_gettime(CLOCK_MONOTONIC_COARSE, &ts);
+	nstime_init2(time, ts.tv_sec, ts.tv_nsec);
+}
+#elif JEMALLOC_HAVE_CLOCK_MONOTONIC
+#  define NSTIME_MONOTONIC true
+static void
+nstime_get(nstime_t *time)
+{
+	struct timespec ts;
+
+	clock_gettime(CLOCK_MONOTONIC, &ts);
+	nstime_init2(time, ts.tv_sec, ts.tv_nsec);
+}
+#elif JEMALLOC_HAVE_MACH_ABSOLUTE_TIME
+#  define NSTIME_MONOTONIC true
+static void
+nstime_get(nstime_t *time)
+{
+
+	nstime_init(time, mach_absolute_time());
+}
+#else
+#  define NSTIME_MONOTONIC false
+static void
+nstime_get(nstime_t *time)
+{
+	struct timeval tv;
+
+	gettimeofday(&tv, NULL);
+	nstime_init2(time, tv.tv_sec, tv.tv_usec * 1000);
+}
+#endif
+
+#ifdef JEMALLOC_JET
+#undef nstime_monotonic
+#define	nstime_monotonic JEMALLOC_N(n_nstime_monotonic)
+#endif
+bool
+nstime_monotonic(void)
+{
+
+	return (NSTIME_MONOTONIC);
+#undef NSTIME_MONOTONIC
+}
+#ifdef JEMALLOC_JET
+#undef nstime_monotonic
+#define	nstime_monotonic JEMALLOC_N(nstime_monotonic)
+nstime_monotonic_t *nstime_monotonic = JEMALLOC_N(n_nstime_monotonic);
+#endif
+
 #ifdef JEMALLOC_JET
 #undef nstime_update
-#define	nstime_update JEMALLOC_N(nstime_update_impl)
+#define	nstime_update JEMALLOC_N(n_nstime_update)
 #endif
 bool
 nstime_update(nstime_t *time)
@@ -107,31 +177,7 @@ nstime_update(nstime_t *time)
 	nstime_t old_time;

 	nstime_copy(&old_time, time);
-
-#ifdef _WIN32
-	{
-		FILETIME ft;
-		uint64_t ticks;
-		GetSystemTimeAsFileTime(&ft);
-		ticks = (((uint64_t)ft.dwHighDateTime) << 32) |
-		    ft.dwLowDateTime;
-		time->ns = ticks * 100;
-	}
-#elif JEMALLOC_CLOCK_GETTIME
-	{
-		struct timespec ts;
-
-		if (sysconf(_SC_MONOTONIC_CLOCK) > 0)
-			clock_gettime(CLOCK_MONOTONIC, &ts);
-		else
-			clock_gettime(CLOCK_REALTIME, &ts);
-		time->ns = ts.tv_sec * BILLION + ts.tv_nsec;
-	}
-#else
-	struct timeval tv;
-	gettimeofday(&tv, NULL);
-	time->ns = tv.tv_sec * BILLION + tv.tv_usec * 1000;
-#endif
+	nstime_get(time);

 	/* Handle non-monotonic clocks. */
 	if (unlikely(nstime_compare(&old_time, time) > 0)) {
@@ -144,5 +190,5 @@ nstime_update(nstime_t *time)
 #ifdef JEMALLOC_JET
 #undef nstime_update
 #define	nstime_update JEMALLOC_N(nstime_update)
-nstime_update_t *nstime_update = JEMALLOC_N(nstime_update_impl);
+nstime_update_t *nstime_update = JEMALLOC_N(n_nstime_update);
 #endif
--- a/src/pages.c
+++ b/src/pages.c
@@ -1,29 +1,49 @@
 #define	JEMALLOC_PAGES_C_
 #include "jemalloc/internal/jemalloc_internal.h"

+#ifdef JEMALLOC_SYSCTL_VM_OVERCOMMIT
+#include <sys/sysctl.h>
+#endif
+
+/******************************************************************************/
+/* Data. */
+
+#ifndef _WIN32
+#  define PAGES_PROT_COMMIT (PROT_READ | PROT_WRITE)
+#  define PAGES_PROT_DECOMMIT (PROT_NONE)
+static int	mmap_flags;
+#endif
+static bool	os_overcommits;
+
 /******************************************************************************/

 void *
-pages_map(void *addr, size_t size)
+pages_map(void *addr, size_t size, bool *commit)
 {
 	void *ret;

 	assert(size != 0);

+	if (os_overcommits)
+		*commit = true;
+
 #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,
+	ret = VirtualAlloc(addr, size, MEM_RESERVE | (*commit ? MEM_COMMIT : 0),
 	    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);
+	{
+		int prot = *commit ? PAGES_PROT_COMMIT : PAGES_PROT_DECOMMIT;
+
+		ret = mmap(addr, size, prot, mmap_flags, -1, 0);
+	}
 	assert(ret != NULL);

 	if (ret == MAP_FAILED)
@@ -67,7 +87,8 @@ pages_unmap(void *addr, size_t size)
 }

 void *
-pages_trim(void *addr, size_t alloc_size, size_t leadsize, size_t size)
+pages_trim(void *addr, size_t alloc_size, size_t leadsize, size_t size,
+    bool *commit)
 {
 	void *ret = (void *)((uintptr_t)addr + leadsize);

@@ -77,7 +98,7 @@ pages_trim(void *addr, size_t alloc_size, size_t leadsize, size_t size)
 		void *new_addr;

 		pages_unmap(addr, alloc_size);
-		new_addr = pages_map(ret, size);
+		new_addr = pages_map(ret, size, commit);
 		if (new_addr == ret)
 			return (ret);
 		if (new_addr)
@@ -101,17 +122,17 @@ static bool
 pages_commit_impl(void *addr, size_t size, bool commit)
 {

-#ifndef _WIN32
-	/*
-	 * The following decommit/commit implementation is functional, but
-	 * always disabled because it doesn't add value beyong improved
-	 * debugging (at the cost of extra system calls) on systems that
-	 * overcommit.
-	 */
-	if (false) {
-		int prot = commit ? (PROT_READ | PROT_WRITE) : PROT_NONE;
-		void *result = mmap(addr, size, prot, MAP_PRIVATE | MAP_ANON |
-		    MAP_FIXED, -1, 0);
+	if (os_overcommits)
+		return (true);
+
+#ifdef _WIN32
+	return (commit ? (addr != VirtualAlloc(addr, size, MEM_COMMIT,
+	    PAGE_READWRITE)) : (!VirtualFree(addr, size, MEM_DECOMMIT)));
+#else
+	{
+		int prot = commit ? PAGES_PROT_COMMIT : PAGES_PROT_DECOMMIT;
+		void *result = mmap(addr, size, prot, mmap_flags | MAP_FIXED,
+		    -1, 0);
 		if (result == MAP_FAILED)
 			return (true);
 		if (result != addr) {
@@ -125,7 +146,6 @@ pages_commit_impl(void *addr, size_t size, bool commit)
 		return (false);
 	}
 #endif
-	return (true);
 }

 bool
@@ -150,15 +170,16 @@ pages_purge(void *addr, size_t size)
 #ifdef _WIN32
 	VirtualAlloc(addr, size, MEM_RESET, PAGE_READWRITE);
 	unzeroed = true;
-#elif defined(JEMALLOC_HAVE_MADVISE)
-#  ifdef JEMALLOC_PURGE_MADVISE_DONTNEED
-#    define JEMALLOC_MADV_PURGE MADV_DONTNEED
-#    define JEMALLOC_MADV_ZEROS true
-#  elif defined(JEMALLOC_PURGE_MADVISE_FREE)
+#elif (defined(JEMALLOC_PURGE_MADVISE_FREE) || \
+    defined(JEMALLOC_PURGE_MADVISE_DONTNEED))
+#  if defined(JEMALLOC_PURGE_MADVISE_FREE)
 #    define JEMALLOC_MADV_PURGE MADV_FREE
 #    define JEMALLOC_MADV_ZEROS false
+#  elif defined(JEMALLOC_PURGE_MADVISE_DONTNEED)
+#    define JEMALLOC_MADV_PURGE MADV_DONTNEED
+#    define JEMALLOC_MADV_ZEROS true
 #  else
-#    error "No madvise(2) flag defined for purging unused dirty pages."
+#    error No madvise(2) flag defined for purging unused dirty pages
 #  endif
 	int err = madvise(addr, size, JEMALLOC_MADV_PURGE);
 	unzeroed = (!JEMALLOC_MADV_ZEROS || err != 0);
@@ -171,3 +192,111 @@ pages_purge(void *addr, size_t size)
 	return (unzeroed);
 }

+bool
+pages_huge(void *addr, size_t size)
+{
+
+	assert(PAGE_ADDR2BASE(addr) == addr);
+	assert(PAGE_CEILING(size) == size);
+
+#ifdef JEMALLOC_HAVE_MADVISE_HUGE
+	return (madvise(addr, size, MADV_HUGEPAGE) != 0);
+#else
+	return (false);
+#endif
+}
+
+bool
+pages_nohuge(void *addr, size_t size)
+{
+
+	assert(PAGE_ADDR2BASE(addr) == addr);
+	assert(PAGE_CEILING(size) == size);
+
+#ifdef JEMALLOC_HAVE_MADVISE_HUGE
+	return (madvise(addr, size, MADV_NOHUGEPAGE) != 0);
+#else
+	return (false);
+#endif
+}
+
+#ifdef JEMALLOC_SYSCTL_VM_OVERCOMMIT
+static bool
+os_overcommits_sysctl(void)
+{
+	int vm_overcommit;
+	size_t sz;
+
+	sz = sizeof(vm_overcommit);
+	if (sysctlbyname("vm.overcommit", &vm_overcommit, &sz, NULL, 0) != 0)
+		return (false); /* Error. */
+
+	return ((vm_overcommit & 0x3) == 0);
+}
+#endif
+
+#ifdef JEMALLOC_PROC_SYS_VM_OVERCOMMIT_MEMORY
+/*
+ * Use syscall(2) rather than {open,read,close}(2) when possible to avoid
+ * reentry during bootstrapping if another library has interposed system call
+ * wrappers.
+ */
+static bool
+os_overcommits_proc(void)
+{
+	int fd;
+	char buf[1];
+	ssize_t nread;
+
+#if defined(JEMALLOC_USE_SYSCALL) && defined(SYS_open)
+	fd = (int)syscall(SYS_open, "/proc/sys/vm/overcommit_memory", O_RDONLY);
+#else
+	fd = open("/proc/sys/vm/overcommit_memory", O_RDONLY);
+#endif
+	if (fd == -1)
+		return (false); /* Error. */
+
+#if defined(JEMALLOC_USE_SYSCALL) && defined(SYS_read)
+	nread = (ssize_t)syscall(SYS_read, fd, &buf, sizeof(buf));
+#else
+	nread = read(fd, &buf, sizeof(buf));
+#endif
+
+#if defined(JEMALLOC_USE_SYSCALL) && defined(SYS_close)
+	syscall(SYS_close, fd);
+#else
+	close(fd);
+#endif
+
+	if (nread < 1)
+		return (false); /* Error. */
+	/*
+	 * /proc/sys/vm/overcommit_memory meanings:
+	 * 0: Heuristic overcommit.
+	 * 1: Always overcommit.
+	 * 2: Never overcommit.
+	 */
+	return (buf[0] == '0' || buf[0] == '1');
+}
+#endif
+
+void
+pages_boot(void)
+{
+
+#ifndef _WIN32
+	mmap_flags = MAP_PRIVATE | MAP_ANON;
+#endif
+
+#ifdef JEMALLOC_SYSCTL_VM_OVERCOMMIT
+	os_overcommits = os_overcommits_sysctl();
+#elif defined(JEMALLOC_PROC_SYS_VM_OVERCOMMIT_MEMORY)
+	os_overcommits = os_overcommits_proc();
+#  ifdef MAP_NORESERVE
+	if (os_overcommits)
+		mmap_flags |= MAP_NORESERVE;
+#  endif
+#else
+	os_overcommits = false;
+#endif
+}
--- a/src/prof.c
+++ b/src/prof.c
--- a/src/quarantine.c
+++ b/src/quarantine.c
@@ -13,24 +13,22 @@
 /* Function prototypes for non-inline static functions. */

 static quarantine_t	*quarantine_grow(tsd_t *tsd, quarantine_t *quarantine);
-static void	quarantine_drain_one(tsd_t *tsd, quarantine_t *quarantine);
-static void	quarantine_drain(tsd_t *tsd, quarantine_t *quarantine,
+static void	quarantine_drain_one(tsdn_t *tsdn, quarantine_t *quarantine);
+static void	quarantine_drain(tsdn_t *tsdn, quarantine_t *quarantine,
    size_t upper_bound);

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

 static quarantine_t *
-quarantine_init(tsd_t *tsd, size_t lg_maxobjs)
+quarantine_init(tsdn_t *tsdn, size_t lg_maxobjs)
 {
 	quarantine_t *quarantine;
 	size_t size;

-	assert(tsd_nominal(tsd));
-
 	size = offsetof(quarantine_t, objs) + ((ZU(1) << lg_maxobjs) *
 	    sizeof(quarantine_obj_t));
-	quarantine = (quarantine_t *)iallocztm(tsd, size, size2index(size),
-	    false, tcache_get(tsd, true), true, NULL, true);
+	quarantine = (quarantine_t *)iallocztm(tsdn, size, size2index(size),
+	    false, NULL, true, arena_get(TSDN_NULL, 0, true), true);
 	if (quarantine == NULL)
 		return (NULL);
 	quarantine->curbytes = 0;
@@ -49,7 +47,7 @@ quarantine_alloc_hook_work(tsd_t *tsd)
 	if (!tsd_nominal(tsd))
 		return;

-	quarantine = quarantine_init(tsd, LG_MAXOBJS_INIT);
+	quarantine = quarantine_init(tsd_tsdn(tsd), LG_MAXOBJS_INIT);
 	/*
 	 * Check again whether quarantine has been initialized, because
 	 * quarantine_init() may have triggered recursive initialization.
@@ -57,7 +55,7 @@ quarantine_alloc_hook_work(tsd_t *tsd)
 	if (tsd_quarantine_get(tsd) == NULL)
 		tsd_quarantine_set(tsd, quarantine);
 	else
-		idalloctm(tsd, quarantine, tcache_get(tsd, false), true, true);
+		idalloctm(tsd_tsdn(tsd), quarantine, NULL, true, true);
 }

 static quarantine_t *
@@ -65,9 +63,9 @@ quarantine_grow(tsd_t *tsd, quarantine_t *quarantine)
 {
 	quarantine_t *ret;

-	ret = quarantine_init(tsd, quarantine->lg_maxobjs + 1);
+	ret = quarantine_init(tsd_tsdn(tsd), quarantine->lg_maxobjs + 1);
 	if (ret == NULL) {
-		quarantine_drain_one(tsd, quarantine);
+		quarantine_drain_one(tsd_tsdn(tsd), quarantine);
 		return (quarantine);
 	}

@@ -89,18 +87,18 @@ quarantine_grow(tsd_t *tsd, quarantine_t *quarantine)
 		memcpy(&ret->objs[ncopy_a], quarantine->objs, ncopy_b *
 		    sizeof(quarantine_obj_t));
 	}
-	idalloctm(tsd, quarantine, tcache_get(tsd, false), true, true);
+	idalloctm(tsd_tsdn(tsd), quarantine, NULL, true, true);

 	tsd_quarantine_set(tsd, ret);
 	return (ret);
 }

 static void
-quarantine_drain_one(tsd_t *tsd, quarantine_t *quarantine)
+quarantine_drain_one(tsdn_t *tsdn, quarantine_t *quarantine)
 {
 	quarantine_obj_t *obj = &quarantine->objs[quarantine->first];
-	assert(obj->usize == isalloc(obj->ptr, config_prof));
-	idalloctm(tsd, obj->ptr, NULL, false, true);
+	assert(obj->usize == isalloc(tsdn, obj->ptr, config_prof));
+	idalloctm(tsdn, obj->ptr, NULL, false, true);
 	quarantine->curbytes -= obj->usize;
 	quarantine->curobjs--;
 	quarantine->first = (quarantine->first + 1) & ((ZU(1) <<
@@ -108,24 +106,24 @@ quarantine_drain_one(tsd_t *tsd, quarantine_t *quarantine)
 }

 static void
-quarantine_drain(tsd_t *tsd, quarantine_t *quarantine, size_t upper_bound)
+quarantine_drain(tsdn_t *tsdn, quarantine_t *quarantine, size_t upper_bound)
 {

 	while (quarantine->curbytes > upper_bound && quarantine->curobjs > 0)
-		quarantine_drain_one(tsd, quarantine);
+		quarantine_drain_one(tsdn, quarantine);
 }

 void
 quarantine(tsd_t *tsd, void *ptr)
 {
 	quarantine_t *quarantine;
-	size_t usize = isalloc(ptr, config_prof);
+	size_t usize = isalloc(tsd_tsdn(tsd), ptr, config_prof);

 	cassert(config_fill);
 	assert(opt_quarantine);

 	if ((quarantine = tsd_quarantine_get(tsd)) == NULL) {
-		idalloctm(tsd, ptr, NULL, false, true);
+		idalloctm(tsd_tsdn(tsd), ptr, NULL, false, true);
 		return;
 	}
 	/*
@@ -135,7 +133,7 @@ quarantine(tsd_t *tsd, void *ptr)
 	if (quarantine->curbytes + usize > opt_quarantine) {
 		size_t upper_bound = (opt_quarantine >= usize) ? opt_quarantine
 		    - usize : 0;
-		quarantine_drain(tsd, quarantine, upper_bound);
+		quarantine_drain(tsd_tsdn(tsd), quarantine, upper_bound);
 	}
 	/* Grow the quarantine ring buffer if it's full. */
 	if (quarantine->curobjs == (ZU(1) << quarantine->lg_maxobjs))
@@ -160,11 +158,11 @@ quarantine(tsd_t *tsd, void *ptr)
 			    && usize <= SMALL_MAXCLASS)
 				arena_quarantine_junk_small(ptr, usize);
 			else
-				memset(ptr, 0x5a, usize);
+				memset(ptr, JEMALLOC_FREE_JUNK, usize);
 		}
 	} else {
 		assert(quarantine->curbytes == 0);
-		idalloctm(tsd, ptr, NULL, false, true);
+		idalloctm(tsd_tsdn(tsd), ptr, NULL, false, true);
 	}
 }

@@ -178,8 +176,8 @@ quarantine_cleanup(tsd_t *tsd)

 	quarantine = tsd_quarantine_get(tsd);
 	if (quarantine != NULL) {
-		quarantine_drain(tsd, quarantine, 0);
-		idalloctm(tsd, quarantine, tcache_get(tsd, false), true, true);
+		quarantine_drain(tsd_tsdn(tsd), quarantine, 0);
+		idalloctm(tsd_tsdn(tsd), quarantine, NULL, true, true);
 		tsd_quarantine_set(tsd, NULL);
 	}
 }
--- a/src/rtree.c
+++ b/src/rtree.c
@@ -15,6 +15,8 @@ rtree_new(rtree_t *rtree, unsigned bits, rtree_node_alloc_t *alloc,
 {
 	unsigned bits_in_leaf, height, i;

+	assert(RTREE_HEIGHT_MAX == ((ZU(1) << (LG_SIZEOF_PTR+3)) /
+	    RTREE_BITS_PER_LEVEL));
 	assert(bits > 0 && bits <= (sizeof(uintptr_t) << 3));

 	bits_in_leaf = (bits % RTREE_BITS_PER_LEVEL) == 0 ? RTREE_BITS_PER_LEVEL
@@ -94,12 +96,15 @@ rtree_node_init(rtree_t *rtree, unsigned level, rtree_node_elm_t **elmp)
 	rtree_node_elm_t *node;

 	if (atomic_cas_p((void **)elmp, NULL, RTREE_NODE_INITIALIZING)) {
+		spin_t spinner;
+
 		/*
 		 * Another thread is already in the process of initializing.
 		 * Spin-wait until initialization is complete.
 		 */
+		spin_init(&spinner);
 		do {
-			CPU_SPINWAIT;
+			spin_adaptive(&spinner);
 			node = atomic_read_p((void **)elmp);
 		} while (node == RTREE_NODE_INITIALIZING);
 	} else {
@@ -123,5 +128,5 @@ rtree_node_elm_t *
 rtree_child_read_hard(rtree_t *rtree, rtree_node_elm_t *elm, unsigned level)
 {

-	return (rtree_node_init(rtree, level, &elm->child));
+	return (rtree_node_init(rtree, level+1, &elm->child));
 }
--- a/src/spin.c
+++ b/src/spin.c
@@ -0,0 +1,2 @@
+#define	JEMALLOC_SPIN_C_
+#include "jemalloc/internal/jemalloc_internal.h"
--- a/src/stats.c
+++ b/src/stats.c
--- a/src/tcache.c
+++ b/src/tcache.c
@@ -21,12 +21,16 @@ static unsigned		tcaches_past;
 /* Head of singly linked list tracking available tcaches elements. */
 static tcaches_t	*tcaches_avail;

+/* Protects tcaches{,_past,_avail}. */
+static malloc_mutex_t	tcaches_mtx;
+
 /******************************************************************************/

-size_t	tcache_salloc(const void *ptr)
+size_t
+tcache_salloc(tsdn_t *tsdn, const void *ptr)
 {

-	return (arena_salloc(ptr, false));
+	return (arena_salloc(tsdn, ptr, false));
 }

 void
@@ -70,12 +74,12 @@ tcache_event_hard(tsd_t *tsd, tcache_t *tcache)
 }

 void *
-tcache_alloc_small_hard(tsd_t *tsd, arena_t *arena, tcache_t *tcache,
+tcache_alloc_small_hard(tsdn_t *tsdn, arena_t *arena, tcache_t *tcache,
    tcache_bin_t *tbin, szind_t binind, bool *tcache_success)
 {
 	void *ret;

-	arena_tcache_fill_small(tsd, arena, tbin, binind, config_prof ?
+	arena_tcache_fill_small(tsdn, arena, tbin, binind, config_prof ?
 	    tcache->prof_accumbytes : 0);
 	if (config_prof)
 		tcache->prof_accumbytes = 0;
@@ -106,12 +110,13 @@ tcache_bin_flush_small(tsd_t *tsd, tcache_t *tcache, tcache_bin_t *tbin,
 		arena_bin_t *bin = &bin_arena->bins[binind];

 		if (config_prof && bin_arena == arena) {
-			if (arena_prof_accum(arena, tcache->prof_accumbytes))
-				prof_idump();
+			if (arena_prof_accum(tsd_tsdn(tsd), arena,
+			    tcache->prof_accumbytes))
+				prof_idump(tsd_tsdn(tsd));
 			tcache->prof_accumbytes = 0;
 		}

-		malloc_mutex_lock(&bin->lock);
+		malloc_mutex_lock(tsd_tsdn(tsd), &bin->lock);
 		if (config_stats && bin_arena == arena) {
 			assert(!merged_stats);
 			merged_stats = true;
@@ -128,9 +133,9 @@ tcache_bin_flush_small(tsd_t *tsd, tcache_t *tcache, tcache_bin_t *tbin,
 				size_t pageind = ((uintptr_t)ptr -
 				    (uintptr_t)chunk) >> LG_PAGE;
 				arena_chunk_map_bits_t *bitselm =
-				    arena_bitselm_get(chunk, pageind);
-				arena_dalloc_bin_junked_locked(bin_arena, chunk,
-				    ptr, bitselm);
+				    arena_bitselm_get_mutable(chunk, pageind);
+				arena_dalloc_bin_junked_locked(tsd_tsdn(tsd),
+				    bin_arena, chunk, ptr, bitselm);
 			} else {
 				/*
 				 * This object was allocated via a different
@@ -142,8 +147,8 @@ tcache_bin_flush_small(tsd_t *tsd, tcache_t *tcache, tcache_bin_t *tbin,
 				ndeferred++;
 			}
 		}
-		malloc_mutex_unlock(&bin->lock);
-		arena_decay_ticks(tsd, bin_arena, nflush - ndeferred);
+		malloc_mutex_unlock(tsd_tsdn(tsd), &bin->lock);
+		arena_decay_ticks(tsd_tsdn(tsd), bin_arena, nflush - ndeferred);
 	}
 	if (config_stats && !merged_stats) {
 		/*
@@ -151,11 +156,11 @@ tcache_bin_flush_small(tsd_t *tsd, tcache_t *tcache, tcache_bin_t *tbin,
 		 * arena, so the stats didn't get merged.  Manually do so now.
 		 */
 		arena_bin_t *bin = &arena->bins[binind];
-		malloc_mutex_lock(&bin->lock);
+		malloc_mutex_lock(tsd_tsdn(tsd), &bin->lock);
 		bin->stats.nflushes++;
 		bin->stats.nrequests += tbin->tstats.nrequests;
 		tbin->tstats.nrequests = 0;
-		malloc_mutex_unlock(&bin->lock);
+		malloc_mutex_unlock(tsd_tsdn(tsd), &bin->lock);
 	}

 	memmove(tbin->avail - rem, tbin->avail - tbin->ncached, rem *
@@ -188,7 +193,7 @@ tcache_bin_flush_large(tsd_t *tsd, tcache_bin_t *tbin, szind_t binind,

 		if (config_prof)
 			idump = false;
-		malloc_mutex_lock(&locked_arena->lock);
+		malloc_mutex_lock(tsd_tsdn(tsd), &locked_arena->lock);
 		if ((config_prof || config_stats) && locked_arena == arena) {
 			if (config_prof) {
 				idump = arena_prof_accum_locked(arena,
@@ -211,8 +216,8 @@ tcache_bin_flush_large(tsd_t *tsd, tcache_bin_t *tbin, szind_t binind,
 			chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
 			if (extent_node_arena_get(&chunk->node) ==
 			    locked_arena) {
-				arena_dalloc_large_junked_locked(locked_arena,
-				    chunk, ptr);
+				arena_dalloc_large_junked_locked(tsd_tsdn(tsd),
+				    locked_arena, chunk, ptr);
 			} else {
 				/*
 				 * This object was allocated via a different
@@ -224,22 +229,23 @@ tcache_bin_flush_large(tsd_t *tsd, tcache_bin_t *tbin, szind_t binind,
 				ndeferred++;
 			}
 		}
-		malloc_mutex_unlock(&locked_arena->lock);
+		malloc_mutex_unlock(tsd_tsdn(tsd), &locked_arena->lock);
 		if (config_prof && idump)
-			prof_idump();
-		arena_decay_ticks(tsd, locked_arena, nflush - ndeferred);
+			prof_idump(tsd_tsdn(tsd));
+		arena_decay_ticks(tsd_tsdn(tsd), locked_arena, nflush -
+		    ndeferred);
 	}
 	if (config_stats && !merged_stats) {
 		/*
 		 * The flush loop didn't happen to flush to this thread's
 		 * arena, so the stats didn't get merged.  Manually do so now.
 		 */
-		malloc_mutex_lock(&arena->lock);
+		malloc_mutex_lock(tsd_tsdn(tsd), &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);
+		malloc_mutex_unlock(tsd_tsdn(tsd), &arena->lock);
 	}

 	memmove(tbin->avail - rem, tbin->avail - tbin->ncached, rem *
@@ -249,34 +255,26 @@ tcache_bin_flush_large(tsd_t *tsd, tcache_bin_t *tbin, szind_t binind,
 		tbin->low_water = tbin->ncached;
 }

-void
-tcache_arena_associate(tcache_t *tcache, arena_t *arena)
+static void
+tcache_arena_associate(tsdn_t *tsdn, tcache_t *tcache, arena_t *arena)
 {

 	if (config_stats) {
 		/* Link into list of extant tcaches. */
-		malloc_mutex_lock(&arena->lock);
+		malloc_mutex_lock(tsdn, &arena->lock);
 		ql_elm_new(tcache, link);
 		ql_tail_insert(&arena->tcache_ql, tcache, link);
-		malloc_mutex_unlock(&arena->lock);
+		malloc_mutex_unlock(tsdn, &arena->lock);
 	}
 }

-void
-tcache_arena_reassociate(tcache_t *tcache, arena_t *oldarena, arena_t *newarena)
-{
-
-	tcache_arena_dissociate(tcache, oldarena);
-	tcache_arena_associate(tcache, newarena);
-}
-
-void
-tcache_arena_dissociate(tcache_t *tcache, arena_t *arena)
+static void
+tcache_arena_dissociate(tsdn_t *tsdn, tcache_t *tcache, arena_t *arena)
 {

 	if (config_stats) {
 		/* Unlink from list of extant tcaches. */
-		malloc_mutex_lock(&arena->lock);
+		malloc_mutex_lock(tsdn, &arena->lock);
 		if (config_debug) {
 			bool in_ql = false;
 			tcache_t *iter;
@@ -289,11 +287,20 @@ tcache_arena_dissociate(tcache_t *tcache, arena_t *arena)
 			assert(in_ql);
 		}
 		ql_remove(&arena->tcache_ql, tcache, link);
-		tcache_stats_merge(tcache, arena);
-		malloc_mutex_unlock(&arena->lock);
+		tcache_stats_merge(tsdn, tcache, arena);
+		malloc_mutex_unlock(tsdn, &arena->lock);
 	}
 }

+void
+tcache_arena_reassociate(tsdn_t *tsdn, tcache_t *tcache, arena_t *oldarena,
+    arena_t *newarena)
+{
+
+	tcache_arena_dissociate(tsdn, tcache, oldarena);
+	tcache_arena_associate(tsdn, tcache, newarena);
+}
+
 tcache_t *
 tcache_get_hard(tsd_t *tsd)
 {
@@ -307,11 +314,11 @@ tcache_get_hard(tsd_t *tsd)
 	arena = arena_choose(tsd, NULL);
 	if (unlikely(arena == NULL))
 		return (NULL);
-	return (tcache_create(tsd, arena));
+	return (tcache_create(tsd_tsdn(tsd), arena));
 }

 tcache_t *
-tcache_create(tsd_t *tsd, arena_t *arena)
+tcache_create(tsdn_t *tsdn, arena_t *arena)
 {
 	tcache_t *tcache;
 	size_t size, stack_offset;
@@ -325,12 +332,12 @@ tcache_create(tsd_t *tsd, arena_t *arena)
 	/* Avoid false cacheline sharing. */
 	size = sa2u(size, CACHELINE);

-	tcache = ipallocztm(tsd, size, CACHELINE, true, false, true,
-	    arena_get(0, false));
+	tcache = ipallocztm(tsdn, size, CACHELINE, true, NULL, true,
+	    arena_get(TSDN_NULL, 0, true));
 	if (tcache == NULL)
 		return (NULL);

-	tcache_arena_associate(tcache, arena);
+	tcache_arena_associate(tsdn, tcache, arena);

 	ticker_init(&tcache->gc_ticker, TCACHE_GC_INCR);

@@ -357,7 +364,7 @@ tcache_destroy(tsd_t *tsd, tcache_t *tcache)
 	unsigned i;

 	arena = arena_choose(tsd, NULL);
-	tcache_arena_dissociate(tcache, arena);
+	tcache_arena_dissociate(tsd_tsdn(tsd), tcache, arena);

 	for (i = 0; i < NBINS; i++) {
 		tcache_bin_t *tbin = &tcache->tbins[i];
@@ -365,9 +372,9 @@ tcache_destroy(tsd_t *tsd, tcache_t *tcache)

 		if (config_stats && tbin->tstats.nrequests != 0) {
 			arena_bin_t *bin = &arena->bins[i];
-			malloc_mutex_lock(&bin->lock);
+			malloc_mutex_lock(tsd_tsdn(tsd), &bin->lock);
 			bin->stats.nrequests += tbin->tstats.nrequests;
-			malloc_mutex_unlock(&bin->lock);
+			malloc_mutex_unlock(tsd_tsdn(tsd), &bin->lock);
 		}
 	}

@@ -376,19 +383,19 @@ tcache_destroy(tsd_t *tsd, tcache_t *tcache)
 		tcache_bin_flush_large(tsd, tbin, i, 0, tcache);

 		if (config_stats && tbin->tstats.nrequests != 0) {
-			malloc_mutex_lock(&arena->lock);
+			malloc_mutex_lock(tsd_tsdn(tsd), &arena->lock);
 			arena->stats.nrequests_large += tbin->tstats.nrequests;
 			arena->stats.lstats[i - NBINS].nrequests +=
 			    tbin->tstats.nrequests;
-			malloc_mutex_unlock(&arena->lock);
+			malloc_mutex_unlock(tsd_tsdn(tsd), &arena->lock);
 		}
 	}

 	if (config_prof && tcache->prof_accumbytes > 0 &&
-	    arena_prof_accum(arena, tcache->prof_accumbytes))
-		prof_idump();
+	    arena_prof_accum(tsd_tsdn(tsd), arena, tcache->prof_accumbytes))
+		prof_idump(tsd_tsdn(tsd));

-	idalloctm(tsd, tcache, false, true, true);
+	idalloctm(tsd_tsdn(tsd), tcache, NULL, true, true);
 }

 void
@@ -412,21 +419,22 @@ tcache_enabled_cleanup(tsd_t *tsd)
 	/* Do nothing. */
 }

-/* Caller must own arena->lock. */
 void
-tcache_stats_merge(tcache_t *tcache, arena_t *arena)
+tcache_stats_merge(tsdn_t *tsdn, tcache_t *tcache, arena_t *arena)
 {
 	unsigned i;

 	cassert(config_stats);

+	malloc_mutex_assert_owner(tsdn, &arena->lock);
+
 	/* 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);
+		malloc_mutex_lock(tsdn, &bin->lock);
 		bin->stats.nrequests += tbin->tstats.nrequests;
-		malloc_mutex_unlock(&bin->lock);
+		malloc_mutex_unlock(tsdn, &bin->lock);
 		tbin->tstats.nrequests = 0;
 	}

@@ -439,25 +447,56 @@ tcache_stats_merge(tcache_t *tcache, arena_t *arena)
 	}
 }

+static bool
+tcaches_create_prep(tsd_t *tsd) {
+	bool err;
+
+	malloc_mutex_lock(tsd_tsdn(tsd), &tcaches_mtx);
+
+	if (tcaches == NULL) {
+		tcaches = base_alloc(tsd_tsdn(tsd), sizeof(tcache_t *) *
+		    (MALLOCX_TCACHE_MAX+1));
+		if (tcaches == NULL) {
+			err = true;
+			goto label_return;
+		}
+	}
+
+	if (tcaches_avail == NULL && tcaches_past > MALLOCX_TCACHE_MAX) {
+		err = true;
+		goto label_return;
+	}
+
+	err = false;
+label_return:
+	malloc_mutex_unlock(tsd_tsdn(tsd), &tcaches_mtx);
+	return err;
+}
+
 bool
-tcaches_create(tsd_t *tsd, unsigned *r_ind)
-{
+tcaches_create(tsd_t *tsd, unsigned *r_ind) {
+	bool err;
+	arena_t *arena;
 	tcache_t *tcache;
 	tcaches_t *elm;

-	if (tcaches == NULL) {
-		tcaches = base_alloc(sizeof(tcache_t *) *
-		    (MALLOCX_TCACHE_MAX+1));
-		if (tcaches == NULL)
-			return (true);
+	if (tcaches_create_prep(tsd)) {
+		err = true;
+		goto label_return;
 	}

-	if (tcaches_avail == NULL && tcaches_past > MALLOCX_TCACHE_MAX)
-		return (true);
-	tcache = tcache_create(tsd, arena_get(0, false));
-	if (tcache == NULL)
-		return (true);
+	arena = arena_ichoose(tsd, NULL);
+	if (unlikely(arena == NULL)) {
+		err = true;
+		goto label_return;
+	}
+	tcache = tcache_create(tsd_tsdn(tsd), arena);
+	if (tcache == NULL) {
+		err = true;
+		goto label_return;
+	}

+	malloc_mutex_lock(tsd_tsdn(tsd), &tcaches_mtx);
 	if (tcaches_avail != NULL) {
 		elm = tcaches_avail;
 		tcaches_avail = tcaches_avail->next;
@@ -469,56 +508,69 @@ tcaches_create(tsd_t *tsd, unsigned *r_ind)
 		*r_ind = tcaches_past;
 		tcaches_past++;
 	}
+	malloc_mutex_unlock(tsd_tsdn(tsd), &tcaches_mtx);

-	return (false);
+	err = false;
+label_return:
+	malloc_mutex_assert_not_owner(tsd_tsdn(tsd), &tcaches_mtx);
+	return err;
 }

 static void
-tcaches_elm_flush(tsd_t *tsd, tcaches_t *elm)
-{
+tcaches_elm_flush(tsd_t *tsd, tcaches_t *elm) {
+	malloc_mutex_assert_owner(tsd_tsdn(tsd), &tcaches_mtx);

-	if (elm->tcache == NULL)
+	if (elm->tcache == NULL) {
 		return;
+	}
 	tcache_destroy(tsd, elm->tcache);
 	elm->tcache = NULL;
 }

 void
-tcaches_flush(tsd_t *tsd, unsigned ind)
-{
-
+tcaches_flush(tsd_t *tsd, unsigned ind) {
+	malloc_mutex_lock(tsd_tsdn(tsd), &tcaches_mtx);
 	tcaches_elm_flush(tsd, &tcaches[ind]);
+	malloc_mutex_unlock(tsd_tsdn(tsd), &tcaches_mtx);
 }

 void
-tcaches_destroy(tsd_t *tsd, unsigned ind)
-{
-	tcaches_t *elm = &tcaches[ind];
+tcaches_destroy(tsd_t *tsd, unsigned ind) {
+	tcaches_t *elm;
+
+	malloc_mutex_lock(tsd_tsdn(tsd), &tcaches_mtx);
+	elm = &tcaches[ind];
 	tcaches_elm_flush(tsd, elm);
 	elm->next = tcaches_avail;
 	tcaches_avail = elm;
+	malloc_mutex_unlock(tsd_tsdn(tsd), &tcaches_mtx);
 }

 bool
-tcache_boot(void)
-{
+tcache_boot(tsdn_t *tsdn) {
 	unsigned i;

+	cassert(config_tcache);
+
 	/*
 	 * If necessary, clamp opt_lg_tcache_max, now that large_maxclass is
 	 * known.
 	 */
-	if (opt_lg_tcache_max < 0 || (1U << opt_lg_tcache_max) < SMALL_MAXCLASS)
+	if (opt_lg_tcache_max < 0 || (ZU(1) << opt_lg_tcache_max) < SMALL_MAXCLASS)
 		tcache_maxclass = SMALL_MAXCLASS;
-	else if ((1U << opt_lg_tcache_max) > large_maxclass)
+	else if ((ZU(1) << opt_lg_tcache_max) > large_maxclass)
 		tcache_maxclass = large_maxclass;
 	else
-		tcache_maxclass = (1U << opt_lg_tcache_max);
+		tcache_maxclass = (ZU(1) << opt_lg_tcache_max);
+
+	if (malloc_mutex_init(&tcaches_mtx, "tcaches", WITNESS_RANK_TCACHES)) {
+		return true;
+	}

 	nhbins = size2index(tcache_maxclass) + 1;

 	/* Initialize tcache_bin_info. */
-	tcache_bin_info = (tcache_bin_info_t *)base_alloc(nhbins *
+	tcache_bin_info = (tcache_bin_info_t *)base_alloc(tsdn, nhbins *
 	    sizeof(tcache_bin_info_t));
 	if (tcache_bin_info == NULL)
 		return (true);
@@ -544,3 +596,24 @@ tcache_boot(void)

 	return (false);
 }
+
+void
+tcache_prefork(tsdn_t *tsdn) {
+	if (!config_prof && opt_tcache) {
+		malloc_mutex_prefork(tsdn, &tcaches_mtx);
+	}
+}
+
+void
+tcache_postfork_parent(tsdn_t *tsdn) {
+	if (!config_prof && opt_tcache) {
+		malloc_mutex_postfork_parent(tsdn, &tcaches_mtx);
+	}
+}
+
+void
+tcache_postfork_child(tsdn_t *tsdn) {
+	if (!config_prof && opt_tcache) {
+		malloc_mutex_postfork_child(tsdn, &tcaches_mtx);
+	}
+}
--- a/src/tsd.c
+++ b/src/tsd.c
@@ -77,7 +77,7 @@ tsd_cleanup(void *arg)
 		/* Do nothing. */
 		break;
 	case tsd_state_nominal:
-#define O(n, t)								\
+#define	O(n, t)								\
 		n##_cleanup(tsd);
 MALLOC_TSD
 #undef O
@@ -106,15 +106,17 @@ MALLOC_TSD
 	}
 }

-bool
+tsd_t *
 malloc_tsd_boot0(void)
 {
+	tsd_t *tsd;

 	ncleanups = 0;
 	if (tsd_boot0())
-		return (true);
-	*tsd_arenas_tdata_bypassp_get(tsd_fetch()) = true;
-	return (false);
+		return (NULL);
+	tsd = tsd_fetch();
+	*tsd_arenas_tdata_bypassp_get(tsd) = true;
+	return (tsd);
 }

 void
@@ -169,10 +171,10 @@ tsd_init_check_recursion(tsd_init_head_t *head, tsd_init_block_t *block)
 	tsd_init_block_t *iter;

 	/* Check whether this thread has already inserted into the list. */
-	malloc_mutex_lock(&head->lock);
+	malloc_mutex_lock(TSDN_NULL, &head->lock);
 	ql_foreach(iter, &head->blocks, link) {
 		if (iter->thread == self) {
-			malloc_mutex_unlock(&head->lock);
+			malloc_mutex_unlock(TSDN_NULL, &head->lock);
 			return (iter->data);
 		}
 	}
@@ -180,7 +182,7 @@ tsd_init_check_recursion(tsd_init_head_t *head, tsd_init_block_t *block)
 	ql_elm_new(block, link);
 	block->thread = self;
 	ql_tail_insert(&head->blocks, block, link);
-	malloc_mutex_unlock(&head->lock);
+	malloc_mutex_unlock(TSDN_NULL, &head->lock);
 	return (NULL);
 }

@@ -188,8 +190,8 @@ void
 tsd_init_finish(tsd_init_head_t *head, tsd_init_block_t *block)
 {

-	malloc_mutex_lock(&head->lock);
+	malloc_mutex_lock(TSDN_NULL, &head->lock);
 	ql_remove(&head->blocks, block, link);
-	malloc_mutex_unlock(&head->lock);
+	malloc_mutex_unlock(TSDN_NULL, &head->lock);
 }
 #endif
--- a/src/util.c
+++ b/src/util.c
@@ -14,6 +14,7 @@
 		malloc_write("<jemalloc>: Unreachable code reached\n");	\
 		abort();						\
 	}								\
+	unreachable();							\
 } while (0)

 #define	not_implemented() do {						\
@@ -48,7 +49,7 @@ static void
 wrtmessage(void *cbopaque, const char *s)
 {

-#ifdef SYS_write
+#if defined(JEMALLOC_USE_SYSCALL) && defined(SYS_write)
 	/*
 	 * Use syscall(2) rather than write(2) when possible in order to avoid
 	 * the possibility of memory allocation within libc.  This is necessary
@@ -199,7 +200,7 @@ malloc_strtoumax(const char *restrict nptr, char **restrict endptr, int base)
 		p++;
 	}
 	if (neg)
-		ret = -ret;
+		ret = (uintmax_t)(-((intmax_t)ret));

 	if (p == ns) {
 		/* No conversion performed. */
@@ -314,10 +315,9 @@ x2s(uintmax_t x, bool alt_form, bool uppercase, char *s, size_t *slen_p)
 	return (s);
 }

-int
+size_t
 malloc_vsnprintf(char *str, size_t size, const char *format, va_list ap)
 {
-	int ret;
 	size_t i;
 	const char *f;

@@ -408,6 +408,8 @@ malloc_vsnprintf(char *str, size_t size, const char *format, va_list ap)
 			int prec = -1;
 			int width = -1;
 			unsigned char len = '?';
+			char *s;
+			size_t slen;

 			f++;
 			/* Flags. */
@@ -498,8 +500,6 @@ malloc_vsnprintf(char *str, size_t size, const char *format, va_list ap)
 			}
 			/* Conversion specifier. */
 			switch (*f) {
-				char *s;
-				size_t slen;
 			case '%':
 				/* %% */
 				APPEND_C(*f);
@@ -585,21 +585,19 @@ malloc_vsnprintf(char *str, size_t size, const char *format, va_list ap)
 		str[i] = '\0';
 	else
 		str[size - 1] = '\0';
-	assert(i < INT_MAX);
-	ret = (int)i;

 #undef APPEND_C
 #undef APPEND_S
 #undef APPEND_PADDED_S
 #undef GET_ARG_NUMERIC
-	return (ret);
+	return (i);
 }

 JEMALLOC_FORMAT_PRINTF(3, 4)
-int
+size_t
 malloc_snprintf(char *str, size_t size, const char *format, ...)
 {
-	int ret;
+	size_t ret;
 	va_list ap;

 	va_start(ap, format);
--- a/src/witness.c
+++ b/src/witness.c
@@ -0,0 +1,136 @@
+#define	JEMALLOC_WITNESS_C_
+#include "jemalloc/internal/jemalloc_internal.h"
+
+void
+witness_init(witness_t *witness, const char *name, witness_rank_t rank,
+    witness_comp_t *comp)
+{
+
+	witness->name = name;
+	witness->rank = rank;
+	witness->comp = comp;
+}
+
+#ifdef JEMALLOC_JET
+#undef witness_lock_error
+#define	witness_lock_error JEMALLOC_N(n_witness_lock_error)
+#endif
+void
+witness_lock_error(const witness_list_t *witnesses, const witness_t *witness)
+{
+	witness_t *w;
+
+	malloc_printf("<jemalloc>: Lock rank order reversal:");
+	ql_foreach(w, witnesses, link) {
+		malloc_printf(" %s(%u)", w->name, w->rank);
+	}
+	malloc_printf(" %s(%u)\n", witness->name, witness->rank);
+	abort();
+}
+#ifdef JEMALLOC_JET
+#undef witness_lock_error
+#define	witness_lock_error JEMALLOC_N(witness_lock_error)
+witness_lock_error_t *witness_lock_error = JEMALLOC_N(n_witness_lock_error);
+#endif
+
+#ifdef JEMALLOC_JET
+#undef witness_owner_error
+#define	witness_owner_error JEMALLOC_N(n_witness_owner_error)
+#endif
+void
+witness_owner_error(const witness_t *witness)
+{
+
+	malloc_printf("<jemalloc>: Should own %s(%u)\n", witness->name,
+	    witness->rank);
+	abort();
+}
+#ifdef JEMALLOC_JET
+#undef witness_owner_error
+#define	witness_owner_error JEMALLOC_N(witness_owner_error)
+witness_owner_error_t *witness_owner_error = JEMALLOC_N(n_witness_owner_error);
+#endif
+
+#ifdef JEMALLOC_JET
+#undef witness_not_owner_error
+#define	witness_not_owner_error JEMALLOC_N(n_witness_not_owner_error)
+#endif
+void
+witness_not_owner_error(const witness_t *witness)
+{
+
+	malloc_printf("<jemalloc>: Should not own %s(%u)\n", witness->name,
+	    witness->rank);
+	abort();
+}
+#ifdef JEMALLOC_JET
+#undef witness_not_owner_error
+#define	witness_not_owner_error JEMALLOC_N(witness_not_owner_error)
+witness_not_owner_error_t *witness_not_owner_error =
+    JEMALLOC_N(n_witness_not_owner_error);
+#endif
+
+#ifdef JEMALLOC_JET
+#undef witness_depth_error
+#define witness_depth_error JEMALLOC_N(n_witness_depth_error)
+#endif
+void
+witness_depth_error(const witness_list_t *witnesses,
+    witness_rank_t rank_inclusive, unsigned depth) {
+	witness_t *w;
+
+	malloc_printf("<jemalloc>: Should own %u lock%s of rank >= %u:", depth,
+	    (depth != 1) ?  "s" : "", rank_inclusive);
+	ql_foreach(w, witnesses, link) {
+		malloc_printf(" %s(%u)", w->name, w->rank);
+	}
+	malloc_printf("\n");
+	abort();
+}
+#ifdef JEMALLOC_JET
+#undef witness_depth_error
+#define witness_depth_error JEMALLOC_N(witness_depth_error)
+witness_depth_error_t *witness_depth_error = JEMALLOC_N(n_witness_depth_error);
+#endif
+
+void
+witnesses_cleanup(tsd_t *tsd)
+{
+
+	witness_assert_lockless(tsd_tsdn(tsd));
+
+	/* Do nothing. */
+}
+
+void
+witness_fork_cleanup(tsd_t *tsd)
+{
+
+	/* Do nothing. */
+}
+
+void
+witness_prefork(tsd_t *tsd)
+{
+
+	tsd_witness_fork_set(tsd, true);
+}
+
+void
+witness_postfork_parent(tsd_t *tsd)
+{
+
+	tsd_witness_fork_set(tsd, false);
+}
+
+void
+witness_postfork_child(tsd_t *tsd)
+{
+#ifndef JEMALLOC_MUTEX_INIT_CB
+	witness_list_t *witnesses;
+
+	witnesses = tsd_witnessesp_get(tsd);
+	ql_new(witnesses);
+#endif
+	tsd_witness_fork_set(tsd, false);
+}
--- a/src/zone.c
+++ b/src/zone.c
@@ -3,8 +3,77 @@
 #  error "This source file is for zones on Darwin (OS X)."
 #endif

+/* Definitions of the following structs in malloc/malloc.h might be too old
+ * for the built binary to run on newer versions of OSX. So use the newest
+ * possible version of those structs.
+ */
+typedef struct _malloc_zone_t {
+	void *reserved1;
+	void *reserved2;
+	size_t (*size)(struct _malloc_zone_t *, const void *);
+	void *(*malloc)(struct _malloc_zone_t *, size_t);
+	void *(*calloc)(struct _malloc_zone_t *, size_t, size_t);
+	void *(*valloc)(struct _malloc_zone_t *, size_t);
+	void (*free)(struct _malloc_zone_t *, void *);
+	void *(*realloc)(struct _malloc_zone_t *, void *, size_t);
+	void (*destroy)(struct _malloc_zone_t *);
+	const char *zone_name;
+	unsigned (*batch_malloc)(struct _malloc_zone_t *, size_t, void **, unsigned);
+	void (*batch_free)(struct _malloc_zone_t *, void **, unsigned);
+	struct malloc_introspection_t *introspect;
+	unsigned version;
+	void *(*memalign)(struct _malloc_zone_t *, size_t, size_t);
+	void (*free_definite_size)(struct _malloc_zone_t *, void *, size_t);
+	size_t (*pressure_relief)(struct _malloc_zone_t *, size_t);
+} malloc_zone_t;
+
+typedef struct {
+	vm_address_t address;
+	vm_size_t size;
+} vm_range_t;
+
+typedef struct malloc_statistics_t {
+	unsigned blocks_in_use;
+	size_t size_in_use;
+	size_t max_size_in_use;
+	size_t size_allocated;
+} malloc_statistics_t;
+
+typedef kern_return_t memory_reader_t(task_t, vm_address_t, vm_size_t, void **);
+
+typedef void vm_range_recorder_t(task_t, void *, unsigned type, vm_range_t *, unsigned);
+
+typedef struct malloc_introspection_t {
+	kern_return_t (*enumerator)(task_t, void *, unsigned, vm_address_t, memory_reader_t, vm_range_recorder_t);
+	size_t (*good_size)(malloc_zone_t *, size_t);
+	boolean_t (*check)(malloc_zone_t *);
+	void (*print)(malloc_zone_t *, boolean_t);
+	void (*log)(malloc_zone_t *, void *);
+	void (*force_lock)(malloc_zone_t *);
+	void (*force_unlock)(malloc_zone_t *);
+	void (*statistics)(malloc_zone_t *, malloc_statistics_t *);
+	boolean_t (*zone_locked)(malloc_zone_t *);
+	boolean_t (*enable_discharge_checking)(malloc_zone_t *);
+	boolean_t (*disable_discharge_checking)(malloc_zone_t *);
+	void (*discharge)(malloc_zone_t *, void *);
+#ifdef __BLOCKS__
+	void (*enumerate_discharged_pointers)(malloc_zone_t *, void (^)(void *, void *));
+#else
+	void *enumerate_unavailable_without_blocks;
+#endif
+	void (*reinit_lock)(malloc_zone_t *);
+} malloc_introspection_t;
+
+extern kern_return_t malloc_get_all_zones(task_t, memory_reader_t, vm_address_t **, unsigned *);
+
+extern malloc_zone_t *malloc_default_zone(void);
+
+extern void malloc_zone_register(malloc_zone_t *zone);
+
+extern void malloc_zone_unregister(malloc_zone_t *zone);
+
 /*
- * The malloc_default_purgeable_zone function is only available on >= 10.6.
+ * The malloc_default_purgeable_zone() function is only available on >= 10.6.
 * We need to check whether it is present at runtime, thus the weak_import.
 */
 extern malloc_zone_t *malloc_default_purgeable_zone(void)
@@ -13,30 +82,42 @@ JEMALLOC_ATTR(weak_import);
 /******************************************************************************/
 /* Data. */

-static malloc_zone_t zone;
-static struct malloc_introspection_t zone_introspect;
+static malloc_zone_t *default_zone, *purgeable_zone;
+static malloc_zone_t jemalloc_zone;
+static struct malloc_introspection_t jemalloc_zone_introspect;

 /******************************************************************************/
 /* Function prototypes for non-inline static functions. */

-static size_t	zone_size(malloc_zone_t *zone, void *ptr);
+static size_t	zone_size(malloc_zone_t *zone, const 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 >= 5)
 static void	*zone_memalign(malloc_zone_t *zone, size_t alignment,
-#endif
-#if (JEMALLOC_ZONE_VERSION >= 6)
    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 void	zone_destroy(malloc_zone_t *zone);
+static unsigned	zone_batch_malloc(struct _malloc_zone_t *zone, size_t size,
+    void **results, unsigned num_requested);
+static void	zone_batch_free(struct _malloc_zone_t *zone,
+    void **to_be_freed, unsigned num_to_be_freed);
+static size_t	zone_pressure_relief(struct _malloc_zone_t *zone, size_t goal);
 static size_t	zone_good_size(malloc_zone_t *zone, size_t size);
+static kern_return_t	zone_enumerator(task_t task, void *data, unsigned type_mask,
+    vm_address_t zone_address, memory_reader_t reader,
+    vm_range_recorder_t recorder);
+static boolean_t	zone_check(malloc_zone_t *zone);
+static void	zone_print(malloc_zone_t *zone, boolean_t verbose);
+static void	zone_log(malloc_zone_t *zone, void *address);
 static void	zone_force_lock(malloc_zone_t *zone);
 static void	zone_force_unlock(malloc_zone_t *zone);
+static void	zone_statistics(malloc_zone_t *zone,
+    malloc_statistics_t *stats);
+static boolean_t	zone_locked(malloc_zone_t *zone);
+static void	zone_reinit_lock(malloc_zone_t *zone);

 /******************************************************************************/
 /*
@@ -44,7 +125,7 @@ static void	zone_force_unlock(malloc_zone_t *zone);
 */

 static size_t
-zone_size(malloc_zone_t *zone, void *ptr)
+zone_size(malloc_zone_t *zone, const void *ptr)
 {

 	/*
@@ -56,7 +137,7 @@ zone_size(malloc_zone_t *zone, void *ptr)
 	 * not work in practice, we must check all pointers to assure that they
 	 * reside within a mapped chunk before determining size.
 	 */
-	return (ivsalloc(ptr, config_prof));
+	return (ivsalloc(tsdn_fetch(), ptr, config_prof));
 }

 static void *
@@ -87,7 +168,7 @@ static void
 zone_free(malloc_zone_t *zone, void *ptr)
 {

-	if (ivsalloc(ptr, config_prof) != 0) {
+	if (ivsalloc(tsdn_fetch(), ptr, config_prof) != 0) {
 		je_free(ptr);
 		return;
 	}
@@ -99,13 +180,12 @@ static void *
 zone_realloc(malloc_zone_t *zone, void *ptr, size_t size)
 {

-	if (ivsalloc(ptr, config_prof) != 0)
+	if (ivsalloc(tsdn_fetch(), ptr, config_prof) != 0)
 		return (je_realloc(ptr, size));

 	return (realloc(ptr, size));
 }

-#if (JEMALLOC_ZONE_VERSION >= 5)
 static void *
 zone_memalign(malloc_zone_t *zone, size_t alignment, size_t size)
 {
@@ -115,15 +195,13 @@ zone_memalign(malloc_zone_t *zone, size_t alignment, size_t size)

 	return (ret);
 }
-#endif

-#if (JEMALLOC_ZONE_VERSION >= 6)
 static void
 zone_free_definite_size(malloc_zone_t *zone, void *ptr, size_t size)
 {
 	size_t alloc_size;

-	alloc_size = ivsalloc(ptr, config_prof);
+	alloc_size = ivsalloc(tsdn_fetch(), ptr, config_prof);
 	if (alloc_size != 0) {
 		assert(alloc_size == size);
 		je_free(ptr);
@@ -132,15 +210,46 @@ zone_free_definite_size(malloc_zone_t *zone, void *ptr, size_t size)

 	free(ptr);
 }
-#endif

-static void *
+static void
 zone_destroy(malloc_zone_t *zone)
 {

 	/* This function should never be called. */
 	not_reached();
-	return (NULL);
+}
+
+static unsigned
+zone_batch_malloc(struct _malloc_zone_t *zone, size_t size, void **results,
+    unsigned num_requested)
+{
+	unsigned i;
+
+	for (i = 0; i < num_requested; i++) {
+		results[i] = je_malloc(size);
+		if (!results[i])
+			break;
+	}
+
+	return i;
+}
+
+static void
+zone_batch_free(struct _malloc_zone_t *zone, void **to_be_freed,
+    unsigned num_to_be_freed)
+{
+	unsigned i;
+
+	for (i = 0; i < num_to_be_freed; i++) {
+		zone_free(zone, to_be_freed[i]);
+		to_be_freed[i] = NULL;
+	}
+}
+
+static size_t
+zone_pressure_relief(struct _malloc_zone_t *zone, size_t goal)
+{
+	return 0;
 }

 static size_t
@@ -152,6 +261,30 @@ zone_good_size(malloc_zone_t *zone, size_t size)
 	return (s2u(size));
 }

+static kern_return_t
+zone_enumerator(task_t task, void *data, unsigned type_mask,
+    vm_address_t zone_address, memory_reader_t reader,
+    vm_range_recorder_t recorder)
+{
+	return KERN_SUCCESS;
+}
+
+static boolean_t
+zone_check(malloc_zone_t *zone)
+{
+	return true;
+}
+
+static void
+zone_print(malloc_zone_t *zone, boolean_t verbose)
+{
+}
+
+static void
+zone_log(malloc_zone_t *zone, void *address)
+{
+}
+
 static void
 zone_force_lock(malloc_zone_t *zone)
 {
@@ -164,89 +297,119 @@ static void
 zone_force_unlock(malloc_zone_t *zone)
 {

+	/*
+	 * Call jemalloc_postfork_child() rather than
+	 * jemalloc_postfork_parent(), because this function is executed by both
+	 * parent and child.  The parent can tolerate having state
+	 * reinitialized, but the child cannot unlock mutexes that were locked
+	 * by the parent.
+	 */
 	if (isthreaded)
-		jemalloc_postfork_parent();
+		jemalloc_postfork_child();
 }

-JEMALLOC_ATTR(constructor)
-void
-register_zone(void)
+static void
+zone_statistics(malloc_zone_t *zone, malloc_statistics_t *stats)
+{
+	/* We make no effort to actually fill the values */
+	stats->blocks_in_use = 0;
+	stats->size_in_use = 0;
+	stats->max_size_in_use = 0;
+	stats->size_allocated = 0;
+}
+
+static boolean_t
+zone_locked(malloc_zone_t *zone)
+{
+	/* Pretend no lock is being held */
+	return false;
+}
+
+static void
+zone_reinit_lock(malloc_zone_t *zone)
+{
+	/* As of OSX 10.12, this function is only used when force_unlock would
+	 * be used if the zone version were < 9. So just use force_unlock. */
+	zone_force_unlock(zone);
+}
+
+static void
+zone_init(void)
 {

+	jemalloc_zone.size = zone_size;
+	jemalloc_zone.malloc = zone_malloc;
+	jemalloc_zone.calloc = zone_calloc;
+	jemalloc_zone.valloc = zone_valloc;
+	jemalloc_zone.free = zone_free;
+	jemalloc_zone.realloc = zone_realloc;
+	jemalloc_zone.destroy = zone_destroy;
+	jemalloc_zone.zone_name = "jemalloc_zone";
+	jemalloc_zone.batch_malloc = zone_batch_malloc;
+	jemalloc_zone.batch_free = zone_batch_free;
+	jemalloc_zone.introspect = &jemalloc_zone_introspect;
+	jemalloc_zone.version = 9;
+	jemalloc_zone.memalign = zone_memalign;
+	jemalloc_zone.free_definite_size = zone_free_definite_size;
+	jemalloc_zone.pressure_relief = zone_pressure_relief;
+
+	jemalloc_zone_introspect.enumerator = zone_enumerator;
+	jemalloc_zone_introspect.good_size = zone_good_size;
+	jemalloc_zone_introspect.check = zone_check;
+	jemalloc_zone_introspect.print = zone_print;
+	jemalloc_zone_introspect.log = zone_log;
+	jemalloc_zone_introspect.force_lock = zone_force_lock;
+	jemalloc_zone_introspect.force_unlock = zone_force_unlock;
+	jemalloc_zone_introspect.statistics = zone_statistics;
+	jemalloc_zone_introspect.zone_locked = zone_locked;
+	jemalloc_zone_introspect.enable_discharge_checking = NULL;
+	jemalloc_zone_introspect.disable_discharge_checking = NULL;
+	jemalloc_zone_introspect.discharge = NULL;
+#ifdef __BLOCKS__
+	jemalloc_zone_introspect.enumerate_discharged_pointers = NULL;
+#else
+	jemalloc_zone_introspect.enumerate_unavailable_without_blocks = NULL;
+#endif
+	jemalloc_zone_introspect.reinit_lock = zone_reinit_lock;
+}
+
+static malloc_zone_t *
+zone_default_get(void)
+{
+	malloc_zone_t **zones = NULL;
+	unsigned int num_zones = 0;
+
 	/*
-	 * If something else replaced the system default zone allocator, don't
-	 * register jemalloc's.
+	 * On OSX 10.12, malloc_default_zone returns a special zone that is not
+	 * present in the list of registered zones. That zone uses a "lite zone"
+	 * if one is present (apparently enabled when malloc stack logging is
+	 * enabled), or the first registered zone otherwise. In practice this
+	 * means unless malloc stack logging is enabled, the first registered
+	 * zone is the default.  So get the list of zones to get the first one,
+	 * instead of relying on malloc_default_zone.
 	 */
-	malloc_zone_t *default_zone = malloc_default_zone();
-	malloc_zone_t *purgeable_zone = NULL;
-	if (!default_zone->zone_name ||
-	    strcmp(default_zone->zone_name, "DefaultMallocZone") != 0) {
-		return;
+	if (KERN_SUCCESS != malloc_get_all_zones(0, NULL,
+	    (vm_address_t**)&zones, &num_zones)) {
+		/*
+		 * Reset the value in case the failure happened after it was
+		 * set.
+		 */
+		num_zones = 0;
 	}

-	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 >= 5)
-	zone.memalign = zone_memalign;
-#endif
-#if (JEMALLOC_ZONE_VERSION >= 6)
-	zone.free_definite_size = zone_free_definite_size;
-#endif
-#if (JEMALLOC_ZONE_VERSION >= 8)
-	zone.pressure_relief = NULL;
-#endif
+	if (num_zones)
+		return (zones[0]);

-	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
-#if (JEMALLOC_ZONE_VERSION >= 7)
-	zone_introspect.enable_discharge_checking = NULL;
-	zone_introspect.disable_discharge_checking = NULL;
-	zone_introspect.discharge = NULL;
-#ifdef __BLOCKS__
-	zone_introspect.enumerate_discharged_pointers = NULL;
-#else
-	zone_introspect.enumerate_unavailable_without_blocks = NULL;
-#endif
-#endif
+	return (malloc_default_zone());
+}

-	/*
-	 * The default purgeable zone is created lazily by OSX's libc.  It uses
-	 * the default zone when it is created for "small" allocations
-	 * (< 15 KiB), but assumes the default zone is a scalable_zone.  This
-	 * obviously fails when the default zone is the jemalloc zone, so
-	 * malloc_default_purgeable_zone is called beforehand so that the
-	 * default purgeable zone is created when the default zone is still
-	 * a scalable_zone.  As purgeable zones only exist on >= 10.6, we need
-	 * to check for the existence of malloc_default_purgeable_zone() at
-	 * run time.
-	 */
-	if (malloc_default_purgeable_zone != NULL)
-		purgeable_zone = malloc_default_purgeable_zone();
-
-	/* Register the custom zone.  At this point it won't be the default. */
-	malloc_zone_register(&zone);
+/* As written, this function can only promote jemalloc_zone. */
+static void
+zone_promote(void)
+{
+	malloc_zone_t *zone;

 	do {
-		default_zone = malloc_default_zone();
 		/*
 		 * Unregister and reregister the default zone.  On OSX >= 10.6,
 		 * unregistering takes the last registered zone and places it
@@ -257,6 +420,7 @@ register_zone(void)
 		 */
 		malloc_zone_unregister(default_zone);
 		malloc_zone_register(default_zone);
+
 		/*
 		 * On OSX 10.6, having the default purgeable zone appear before
 		 * the default zone makes some things crash because it thinks it
@@ -268,9 +432,47 @@ register_zone(void)
 		 * above, i.e. the default zone.  Registering it again then puts
 		 * it at the end, obviously after the default zone.
 		 */
-		if (purgeable_zone) {
+		if (purgeable_zone != NULL) {
 			malloc_zone_unregister(purgeable_zone);
 			malloc_zone_register(purgeable_zone);
 		}
-	} while (malloc_default_zone() != &zone);
+
+		zone = zone_default_get();
+	} while (zone != &jemalloc_zone);
+}
+
+JEMALLOC_ATTR(constructor)
+void
+zone_register(void)
+{
+
+	/*
+	 * If something else replaced the system default zone allocator, don't
+	 * register jemalloc's.
+	 */
+	default_zone = zone_default_get();
+	if (!default_zone->zone_name || strcmp(default_zone->zone_name,
+	    "DefaultMallocZone") != 0)
+		return;
+
+	/*
+	 * The default purgeable zone is created lazily by OSX's libc.  It uses
+	 * the default zone when it is created for "small" allocations
+	 * (< 15 KiB), but assumes the default zone is a scalable_zone.  This
+	 * obviously fails when the default zone is the jemalloc zone, so
+	 * malloc_default_purgeable_zone() is called beforehand so that the
+	 * default purgeable zone is created when the default zone is still
+	 * a scalable_zone.  As purgeable zones only exist on >= 10.6, we need
+	 * to check for the existence of malloc_default_purgeable_zone() at
+	 * run time.
+	 */
+	purgeable_zone = (malloc_default_purgeable_zone == NULL) ? NULL :
+	    malloc_default_purgeable_zone();
+
+	/* Register the custom zone.  At this point it won't be the default. */
+	zone_init();
+	malloc_zone_register(&jemalloc_zone);
+
+	/* Promote the custom zone to be default. */
+	zone_promote();
 }
--- a/test/include/test/jemalloc_test.h.in
+++ b/test/include/test/jemalloc_test.h.in
@@ -19,39 +19,6 @@
 #  include <pthread.h>
 #endif

-/******************************************************************************/
-/*
- * Define always-enabled assertion macros, so that test assertions execute even
- * if assertions are disabled in the library code.  These definitions must
- * exist prior to including "jemalloc/internal/util.h".
- */
-#define	assert(e) do {							\
-	if (!(e)) {							\
-		malloc_printf(						\
-		    "<jemalloc>: %s:%d: Failed assertion: \"%s\"\n",	\
-		    __FILE__, __LINE__, #e);				\
-		abort();						\
-	}								\
-} while (0)
-
-#define	not_reached() do {						\
-	malloc_printf(							\
-	    "<jemalloc>: %s:%d: Unreachable code reached\n",		\
-	    __FILE__, __LINE__);					\
-	abort();							\
-} while (0)
-
-#define	not_implemented() do {						\
-	malloc_printf("<jemalloc>: %s:%d: Not implemented\n",		\
-	    __FILE__, __LINE__);					\
-	abort();							\
-} while (0)
-
-#define	assert_not_implemented(e) do {					\
-	if (!(e))							\
-		not_implemented();					\
-} while (0)
-
 #include "test/jemalloc_test_defs.h"

 #ifdef JEMALLOC_OSSPIN
@@ -86,6 +53,14 @@
 #  include "jemalloc/internal/jemalloc_internal_defs.h"
 #  include "jemalloc/internal/jemalloc_internal_macros.h"

+static const bool config_debug =
+#ifdef JEMALLOC_DEBUG
+    true
+#else
+    false
+#endif
+    ;
+
 #  define JEMALLOC_N(n) @private_namespace@##n
 #  include "jemalloc/internal/private_namespace.h"

@@ -149,3 +124,40 @@
 #include "test/thd.h"
 #define	MEXP 19937
 #include "test/SFMT.h"
+
+/******************************************************************************/
+/*
+ * Define always-enabled assertion macros, so that test assertions execute even
+ * if assertions are disabled in the library code.
+ */
+#undef assert
+#undef not_reached
+#undef not_implemented
+#undef assert_not_implemented
+
+#define	assert(e) do {							\
+	if (!(e)) {							\
+		malloc_printf(						\
+		    "<jemalloc>: %s:%d: Failed assertion: \"%s\"\n",	\
+		    __FILE__, __LINE__, #e);				\
+		abort();						\
+	}								\
+} while (0)
+
+#define	not_reached() do {						\
+	malloc_printf(							\
+	    "<jemalloc>: %s:%d: Unreachable code reached\n",		\
+	    __FILE__, __LINE__);					\
+	abort();							\
+} while (0)
+
+#define	not_implemented() do {						\
+	malloc_printf("<jemalloc>: %s:%d: Not implemented\n",		\
+	    __FILE__, __LINE__);					\
+	abort();							\
+} while (0)
+
+#define	assert_not_implemented(e) do {					\
+	if (!(e))							\
+		not_implemented();					\
+} while (0)
--- a/test/include/test/mtx.h
+++ b/test/include/test/mtx.h
@@ -8,6 +8,8 @@
 typedef struct {
 #ifdef _WIN32
 	CRITICAL_SECTION	lock;
+#elif (defined(JEMALLOC_OS_UNFAIR_LOCK))
+	os_unfair_lock		lock;
 #elif (defined(JEMALLOC_OSSPIN))
 	OSSpinLock		lock;
 #else
--- a/test/include/test/test.h
+++ b/test/include/test/test.h
@@ -311,6 +311,9 @@ label_test_end:								\
 #define	test(...)							\
 	p_test(__VA_ARGS__, NULL)

+#define	test_no_malloc_init(...)					\
+	p_test_no_malloc_init(__VA_ARGS__, NULL)
+
 #define	test_skip_if(e) do {						\
 	if (e) {							\
 		test_skip("%s:%s:%d: Test skipped: (%s)",		\
@@ -324,6 +327,7 @@ void	test_fail(const char *format, ...) JEMALLOC_FORMAT_PRINTF(1, 2);

 /* For private use by macros. */
 test_status_t	p_test(test_t *t, ...);
+test_status_t	p_test_no_malloc_init(test_t *t, ...);
 void	p_test_init(const char *name);
 void	p_test_fini(void);
 void	p_test_fail(const char *prefix, const char *message);
--- a/test/integration/MALLOCX_ARENA.c
+++ b/test/integration/MALLOCX_ARENA.c
@@ -19,8 +19,8 @@ thd_start(void *arg)
 	size_t sz;

 	sz = sizeof(arena_ind);
-	assert_d_eq(mallctl("arenas.extend", &arena_ind, &sz, NULL, 0), 0,
-	    "Error in arenas.extend");
+	assert_d_eq(mallctl("arenas.extend", (void *)&arena_ind, &sz, NULL, 0),
+	    0, "Error in arenas.extend");

 	if (thread_ind % 4 != 3) {
 		size_t mib[3];
--- a/test/integration/aligned_alloc.c
+++ b/test/integration/aligned_alloc.c
@@ -1,9 +1,20 @@
 #include "test/jemalloc_test.h"

 #define	CHUNK 0x400000
-/* #define MAXALIGN ((size_t)UINT64_C(0x80000000000)) */
-#define	MAXALIGN ((size_t)0x2000000LU)
-#define	NITER 4
+#define	MAXALIGN (((size_t)1) << 23)
+
+/*
+ * On systems which can't merge extents, tests that call this function generate
+ * a lot of dirty memory very quickly.  Purging between cycles mitigates
+ * potential OOM on e.g. 32-bit Windows.
+ */
+static void
+purge(void)
+{
+
+	assert_d_eq(mallctl("arena.0.purge", NULL, NULL, NULL, 0), 0,
+	    "Unexpected mallctl error");
+}

 TEST_BEGIN(test_alignment_errors)
 {
@@ -74,6 +85,7 @@ TEST_END

 TEST_BEGIN(test_alignment_and_size)
 {
+#define	NITER 4
 	size_t alignment, size, total;
 	unsigned i;
 	void *ps[NITER];
@@ -110,7 +122,9 @@ TEST_BEGIN(test_alignment_and_size)
 				}
 			}
 		}
+		purge();
 	}
+#undef NITER
 }
 TEST_END

--- a/test/integration/allocated.c
+++ b/test/integration/allocated.c
@@ -18,14 +18,14 @@ thd_start(void *arg)
 	size_t sz, usize;

 	sz = sizeof(a0);
-	if ((err = mallctl("thread.allocated", &a0, &sz, NULL, 0))) {
+	if ((err = mallctl("thread.allocated", (void *)&a0, &sz, NULL, 0))) {
 		if (err == ENOENT)
 			goto label_ENOENT;
 		test_fail("%s(): Error in mallctl(): %s", __func__,
 		    strerror(err));
 	}
 	sz = sizeof(ap0);
-	if ((err = mallctl("thread.allocatedp", &ap0, &sz, NULL, 0))) {
+	if ((err = mallctl("thread.allocatedp", (void *)&ap0, &sz, NULL, 0))) {
 		if (err == ENOENT)
 			goto label_ENOENT;
 		test_fail("%s(): Error in mallctl(): %s", __func__,
@@ -36,14 +36,15 @@ thd_start(void *arg)
 	    "storage");

 	sz = sizeof(d0);
-	if ((err = mallctl("thread.deallocated", &d0, &sz, NULL, 0))) {
+	if ((err = mallctl("thread.deallocated", (void *)&d0, &sz, NULL, 0))) {
 		if (err == ENOENT)
 			goto label_ENOENT;
 		test_fail("%s(): Error in mallctl(): %s", __func__,
 		    strerror(err));
 	}
 	sz = sizeof(dp0);
-	if ((err = mallctl("thread.deallocatedp", &dp0, &sz, NULL, 0))) {
+	if ((err = mallctl("thread.deallocatedp", (void *)&dp0, &sz, NULL,
+	    0))) {
 		if (err == ENOENT)
 			goto label_ENOENT;
 		test_fail("%s(): Error in mallctl(): %s", __func__,
@@ -57,9 +58,9 @@ thd_start(void *arg)
 	assert_ptr_not_null(p, "Unexpected malloc() error");

 	sz = sizeof(a1);
-	mallctl("thread.allocated", &a1, &sz, NULL, 0);
+	mallctl("thread.allocated", (void *)&a1, &sz, NULL, 0);
 	sz = sizeof(ap1);
-	mallctl("thread.allocatedp", &ap1, &sz, NULL, 0);
+	mallctl("thread.allocatedp", (void *)&ap1, &sz, NULL, 0);
 	assert_u64_eq(*ap1, a1,
 	    "Dereferenced \"thread.allocatedp\" value should equal "
 	    "\"thread.allocated\" value");
@@ -74,9 +75,9 @@ thd_start(void *arg)
 	free(p);

 	sz = sizeof(d1);
-	mallctl("thread.deallocated", &d1, &sz, NULL, 0);
+	mallctl("thread.deallocated", (void *)&d1, &sz, NULL, 0);
 	sz = sizeof(dp1);
-	mallctl("thread.deallocatedp", &dp1, &sz, NULL, 0);
+	mallctl("thread.deallocatedp", (void *)&dp1, &sz, NULL, 0);
 	assert_u64_eq(*dp1, d1,
 	    "Dereferenced \"thread.deallocatedp\" value should equal "
 	    "\"thread.deallocated\" value");
--- a/test/integration/chunk.c
+++ b/test/integration/chunk.c
@@ -1,9 +1,5 @@
 #include "test/jemalloc_test.h"

-#ifdef JEMALLOC_FILL
-const char *malloc_conf = "junk:false";
-#endif
-
 static chunk_hooks_t orig_hooks;
 static chunk_hooks_t old_hooks;

@@ -137,8 +133,8 @@ TEST_BEGIN(test_chunk)
 	bool xallocx_success_a, xallocx_success_b, xallocx_success_c;

 	sz = sizeof(unsigned);
-	assert_d_eq(mallctl("arenas.extend", &arena_ind, &sz, NULL, 0), 0,
-	    "Unexpected mallctl() failure");
+	assert_d_eq(mallctl("arenas.extend", (void *)&arena_ind, &sz, NULL, 0),
+	    0, "Unexpected mallctl() failure");
 	flags = MALLOCX_ARENA(arena_ind) | MALLOCX_TCACHE_NONE;

 	/* Install custom chunk hooks. */
@@ -148,8 +144,9 @@ TEST_BEGIN(test_chunk)
 	hooks_mib[1] = (size_t)arena_ind;
 	old_size = sizeof(chunk_hooks_t);
 	new_size = sizeof(chunk_hooks_t);
-	assert_d_eq(mallctlbymib(hooks_mib, hooks_miblen, &old_hooks, &old_size,
-	    &new_hooks, new_size), 0, "Unexpected chunk_hooks error");
+	assert_d_eq(mallctlbymib(hooks_mib, hooks_miblen, (void *)&old_hooks,
+	    &old_size, (void *)&new_hooks, new_size), 0,
+	    "Unexpected chunk_hooks error");
 	orig_hooks = old_hooks;
 	assert_ptr_ne(old_hooks.alloc, chunk_alloc, "Unexpected alloc error");
 	assert_ptr_ne(old_hooks.dalloc, chunk_dalloc,
@@ -164,18 +161,18 @@ TEST_BEGIN(test_chunk)

 	/* Get large size classes. */
 	sz = sizeof(size_t);
-	assert_d_eq(mallctl("arenas.lrun.0.size", &large0, &sz, NULL, 0), 0,
-	    "Unexpected arenas.lrun.0.size failure");
-	assert_d_eq(mallctl("arenas.lrun.1.size", &large1, &sz, NULL, 0), 0,
-	    "Unexpected arenas.lrun.1.size failure");
+	assert_d_eq(mallctl("arenas.lrun.0.size", (void *)&large0, &sz, NULL,
+	    0), 0, "Unexpected arenas.lrun.0.size failure");
+	assert_d_eq(mallctl("arenas.lrun.1.size", (void *)&large1, &sz, NULL,
+	    0), 0, "Unexpected arenas.lrun.1.size failure");

 	/* Get huge size classes. */
-	assert_d_eq(mallctl("arenas.hchunk.0.size", &huge0, &sz, NULL, 0), 0,
-	    "Unexpected arenas.hchunk.0.size failure");
-	assert_d_eq(mallctl("arenas.hchunk.1.size", &huge1, &sz, NULL, 0), 0,
-	    "Unexpected arenas.hchunk.1.size failure");
-	assert_d_eq(mallctl("arenas.hchunk.2.size", &huge2, &sz, NULL, 0), 0,
-	    "Unexpected arenas.hchunk.2.size failure");
+	assert_d_eq(mallctl("arenas.hchunk.0.size", (void *)&huge0, &sz, NULL,
+	    0), 0, "Unexpected arenas.hchunk.0.size failure");
+	assert_d_eq(mallctl("arenas.hchunk.1.size", (void *)&huge1, &sz, NULL,
+	    0), 0, "Unexpected arenas.hchunk.1.size failure");
+	assert_d_eq(mallctl("arenas.hchunk.2.size", (void *)&huge2, &sz, NULL,
+	    0), 0, "Unexpected arenas.hchunk.2.size failure");

 	/* Test dalloc/decommit/purge cascade. */
 	purge_miblen = sizeof(purge_mib)/sizeof(size_t);
@@ -265,9 +262,9 @@ TEST_BEGIN(test_chunk)

 	/* Restore chunk hooks. */
 	assert_d_eq(mallctlbymib(hooks_mib, hooks_miblen, NULL, NULL,
-	    &old_hooks, new_size), 0, "Unexpected chunk_hooks error");
-	assert_d_eq(mallctlbymib(hooks_mib, hooks_miblen, &old_hooks, &old_size,
-	    NULL, 0), 0, "Unexpected chunk_hooks error");
+	    (void *)&old_hooks, new_size), 0, "Unexpected chunk_hooks error");
+	assert_d_eq(mallctlbymib(hooks_mib, hooks_miblen, (void *)&old_hooks,
+	    &old_size, NULL, 0), 0, "Unexpected chunk_hooks error");
 	assert_ptr_eq(old_hooks.alloc, orig_hooks.alloc,
 	    "Unexpected alloc error");
 	assert_ptr_eq(old_hooks.dalloc, orig_hooks.dalloc,
--- a/test/integration/chunk.sh
+++ b/test/integration/chunk.sh
@@ -0,0 +1,5 @@
+#!/bin/sh
+
+if [ "x${enable_fill}" = "x1" ] ; then
+  export MALLOC_CONF="junk:false"
+fi
--- a/test/integration/mallocx.c
+++ b/test/integration/mallocx.c
@@ -7,7 +7,7 @@ get_nsizes_impl(const char *cmd)
 	size_t z;

 	z = sizeof(unsigned);
-	assert_d_eq(mallctl(cmd, &ret, &z, NULL, 0), 0,
+	assert_d_eq(mallctl(cmd, (void *)&ret, &z, NULL, 0), 0,
 	    "Unexpected mallctl(\"%s\", ...) failure", cmd);

 	return (ret);
@@ -33,7 +33,7 @@ get_size_impl(const char *cmd, size_t ind)
 	    0, "Unexpected mallctlnametomib(\"%s\", ...) failure", cmd);
 	mib[2] = ind;
 	z = sizeof(size_t);
-	assert_d_eq(mallctlbymib(mib, miblen, &ret, &z, NULL, 0),
+	assert_d_eq(mallctlbymib(mib, miblen, (void *)&ret, &z, NULL, 0),
 	    0, "Unexpected mallctlbymib([\"%s\", %zu], ...) failure", cmd, ind);

 	return (ret);
@@ -46,6 +46,19 @@ get_huge_size(size_t ind)
 	return (get_size_impl("arenas.hchunk.0.size", ind));
 }

+/*
+ * On systems which can't merge extents, tests that call this function generate
+ * a lot of dirty memory very quickly.  Purging between cycles mitigates
+ * potential OOM on e.g. 32-bit Windows.
+ */
+static void
+purge(void)
+{
+
+	assert_d_eq(mallctl("arena.0.purge", NULL, NULL, NULL, 0), 0,
+	    "Unexpected mallctl error");
+}
+
 TEST_BEGIN(test_overflow)
 {
 	size_t hugemax;
@@ -69,7 +82,7 @@ TEST_END

 TEST_BEGIN(test_oom)
 {
-	size_t hugemax, size, alignment;
+	size_t hugemax;
 	bool oom;
 	void *ptrs[3];
 	unsigned i;
@@ -92,23 +105,25 @@ TEST_BEGIN(test_oom)
 		if (ptrs[i] != NULL)
 			dallocx(ptrs[i], 0);
 	}
+	purge();

 #if LG_SIZEOF_PTR == 3
-	size      = ZU(0x8000000000000000);
-	alignment = ZU(0x8000000000000000);
+	assert_ptr_null(mallocx(0x8000000000000000ULL,
+	    MALLOCX_ALIGN(0x8000000000000000ULL)),
+	    "Expected OOM for mallocx()");
+	assert_ptr_null(mallocx(0x8000000000000000ULL,
+	    MALLOCX_ALIGN(0x80000000)),
+	    "Expected OOM for mallocx()");
 #else
-	size      = ZU(0x80000000);
-	alignment = ZU(0x80000000);
+	assert_ptr_null(mallocx(0x80000000UL, MALLOCX_ALIGN(0x80000000UL)),
+	    "Expected OOM for mallocx()");
 #endif
-	assert_ptr_null(mallocx(size, MALLOCX_ALIGN(alignment)),
-	    "Expected OOM for mallocx(size=%#zx, MALLOCX_ALIGN(%#zx)", size,
-	    alignment);
 }
 TEST_END

 TEST_BEGIN(test_basic)
 {
-#define	MAXSZ (((size_t)1) << 26)
+#define	MAXSZ (((size_t)1) << 23)
 	size_t sz;

 	for (sz = 1; sz < MAXSZ; sz = nallocx(sz, 0) + 1) {
@@ -117,23 +132,28 @@ TEST_BEGIN(test_basic)
 		nsz = nallocx(sz, 0);
 		assert_zu_ne(nsz, 0, "Unexpected nallocx() error");
 		p = mallocx(sz, 0);
-		assert_ptr_not_null(p, "Unexpected mallocx() error");
+		assert_ptr_not_null(p,
+		    "Unexpected mallocx(size=%zx, flags=0) error", sz);
 		rsz = sallocx(p, 0);
 		assert_zu_ge(rsz, sz, "Real size smaller than expected");
 		assert_zu_eq(nsz, rsz, "nallocx()/sallocx() size mismatch");
 		dallocx(p, 0);

 		p = mallocx(sz, 0);
-		assert_ptr_not_null(p, "Unexpected mallocx() error");
+		assert_ptr_not_null(p,
+		    "Unexpected mallocx(size=%zx, flags=0) error", sz);
 		dallocx(p, 0);

 		nsz = nallocx(sz, MALLOCX_ZERO);
 		assert_zu_ne(nsz, 0, "Unexpected nallocx() error");
 		p = mallocx(sz, MALLOCX_ZERO);
-		assert_ptr_not_null(p, "Unexpected mallocx() error");
+		assert_ptr_not_null(p,
+		    "Unexpected mallocx(size=%zx, flags=MALLOCX_ZERO) error",
+		    nsz);
 		rsz = sallocx(p, 0);
 		assert_zu_eq(nsz, rsz, "nallocx()/sallocx() rsize mismatch");
 		dallocx(p, 0);
+		purge();
 	}
 #undef MAXSZ
 }
@@ -141,7 +161,7 @@ TEST_END

 TEST_BEGIN(test_alignment_and_size)
 {
-#define	MAXALIGN (((size_t)1) << 25)
+#define	MAXALIGN (((size_t)1) << 23)
 #define	NITER 4
 	size_t nsz, rsz, sz, alignment, total;
 	unsigned i;
@@ -191,6 +211,7 @@ TEST_BEGIN(test_alignment_and_size)
 				}
 			}
 		}
+		purge();
 	}
 #undef MAXALIGN
 #undef NITER
--- a/test/integration/mallocx.sh
+++ b/test/integration/mallocx.sh
@@ -0,0 +1,5 @@
+#!/bin/sh
+
+if [ "x${enable_fill}" = "x1" ] ; then
+  export MALLOC_CONF="junk:false"
+fi
--- a/test/integration/overflow.c
+++ b/test/integration/overflow.c
@@ -8,8 +8,8 @@ TEST_BEGIN(test_overflow)
 	void *p;

 	sz = sizeof(unsigned);
-	assert_d_eq(mallctl("arenas.nhchunks", &nhchunks, &sz, NULL, 0), 0,
-	    "Unexpected mallctl() error");
+	assert_d_eq(mallctl("arenas.nhchunks", (void *)&nhchunks, &sz, NULL, 0),
+	    0, "Unexpected mallctl() error");

 	miblen = sizeof(mib) / sizeof(size_t);
 	assert_d_eq(mallctlnametomib("arenas.hchunk.0.size", mib, &miblen), 0,
@@ -17,8 +17,8 @@ TEST_BEGIN(test_overflow)
 	mib[2] = nhchunks - 1;

 	sz = sizeof(size_t);
-	assert_d_eq(mallctlbymib(mib, miblen, &max_size_class, &sz, NULL, 0), 0,
-	    "Unexpected mallctlbymib() error");
+	assert_d_eq(mallctlbymib(mib, miblen, (void *)&max_size_class, &sz,
+	    NULL, 0), 0, "Unexpected mallctlbymib() error");

 	assert_ptr_null(malloc(max_size_class + 1),
 	    "Expected OOM due to over-sized allocation request");
--- a/test/integration/posix_memalign.c
+++ b/test/integration/posix_memalign.c
@@ -1,9 +1,20 @@
 #include "test/jemalloc_test.h"

 #define	CHUNK 0x400000
-/* #define MAXALIGN ((size_t)UINT64_C(0x80000000000)) */
-#define	MAXALIGN ((size_t)0x2000000LU)
-#define	NITER 4
+#define	MAXALIGN (((size_t)1) << 23)
+
+/*
+ * On systems which can't merge extents, tests that call this function generate
+ * a lot of dirty memory very quickly.  Purging between cycles mitigates
+ * potential OOM on e.g. 32-bit Windows.
+ */
+static void
+purge(void)
+{
+
+	assert_d_eq(mallctl("arena.0.purge", NULL, NULL, NULL, 0), 0,
+	    "Unexpected mallctl error");
+}

 TEST_BEGIN(test_alignment_errors)
 {
@@ -66,6 +77,7 @@ TEST_END

 TEST_BEGIN(test_alignment_and_size)
 {
+#define	NITER 4
 	size_t alignment, size, total;
 	unsigned i;
 	int err;
@@ -104,7 +116,9 @@ TEST_BEGIN(test_alignment_and_size)
 				}
 			}
 		}
+		purge();
 	}
+#undef NITER
 }
 TEST_END

--- a/test/integration/rallocx.c
+++ b/test/integration/rallocx.c
@@ -7,7 +7,7 @@ get_nsizes_impl(const char *cmd)
 	size_t z;

 	z = sizeof(unsigned);
-	assert_d_eq(mallctl(cmd, &ret, &z, NULL, 0), 0,
+	assert_d_eq(mallctl(cmd, (void *)&ret, &z, NULL, 0), 0,
 	    "Unexpected mallctl(\"%s\", ...) failure", cmd);

 	return (ret);
@@ -33,7 +33,7 @@ get_size_impl(const char *cmd, size_t ind)
 	    0, "Unexpected mallctlnametomib(\"%s\", ...) failure", cmd);
 	mib[2] = ind;
 	z = sizeof(size_t);
-	assert_d_eq(mallctlbymib(mib, miblen, &ret, &z, NULL, 0),
+	assert_d_eq(mallctlbymib(mib, miblen, (void *)&ret, &z, NULL, 0),
 	    0, "Unexpected mallctlbymib([\"%s\", %zu], ...) failure", cmd, ind);

 	return (ret);
--- a/test/integration/sdallocx.c
+++ b/test/integration/sdallocx.c
@@ -1,7 +1,7 @@
 #include "test/jemalloc_test.h"

-#define	MAXALIGN (((size_t)1) << 25)
-#define	NITER 4
+#define	MAXALIGN (((size_t)1) << 22)
+#define	NITER 3

 TEST_BEGIN(test_basic)
 {
--- a/test/integration/thread_arena.c
+++ b/test/integration/thread_arena.c
@@ -16,8 +16,8 @@ thd_start(void *arg)
 	free(p);

 	size = sizeof(arena_ind);
-	if ((err = mallctl("thread.arena", &arena_ind, &size, &main_arena_ind,
-	    sizeof(main_arena_ind)))) {
+	if ((err = mallctl("thread.arena", (void *)&arena_ind, &size,
+	    (void *)&main_arena_ind, sizeof(main_arena_ind)))) {
 		char buf[BUFERROR_BUF];

 		buferror(err, buf, sizeof(buf));
@@ -25,7 +25,8 @@ thd_start(void *arg)
 	}

 	size = sizeof(arena_ind);
-	if ((err = mallctl("thread.arena", &arena_ind, &size, NULL, 0))) {
+	if ((err = mallctl("thread.arena", (void *)&arena_ind, &size, NULL,
+	    0))) {
 		char buf[BUFERROR_BUF];

 		buferror(err, buf, sizeof(buf));
@@ -50,7 +51,8 @@ TEST_BEGIN(test_thread_arena)
 	assert_ptr_not_null(p, "Error in malloc()");

 	size = sizeof(arena_ind);
-	if ((err = mallctl("thread.arena", &arena_ind, &size, NULL, 0))) {
+	if ((err = mallctl("thread.arena", (void *)&arena_ind, &size, NULL,
+	    0))) {
 		char buf[BUFERROR_BUF];

 		buferror(err, buf, sizeof(buf));
--- a/test/integration/thread_tcache_enabled.c
+++ b/test/integration/thread_tcache_enabled.c
@@ -16,7 +16,8 @@ thd_start(void *arg)
 	bool e0, e1;

 	sz = sizeof(bool);
-	if ((err = mallctl("thread.tcache.enabled", &e0, &sz, NULL, 0))) {
+	if ((err = mallctl("thread.tcache.enabled", (void *)&e0, &sz, NULL,
+	    0))) {
 		if (err == ENOENT) {
 			assert_false(config_tcache,
 			    "ENOENT should only be returned if tcache is "
@@ -27,53 +28,53 @@ thd_start(void *arg)

 	if (e0) {
 		e1 = false;
-		assert_d_eq(mallctl("thread.tcache.enabled", &e0, &sz, &e1, sz),
-		    0, "Unexpected mallctl() error");
+		assert_d_eq(mallctl("thread.tcache.enabled", (void *)&e0, &sz,
+		    (void *)&e1, sz), 0, "Unexpected mallctl() error");
 		assert_true(e0, "tcache should be enabled");
 	}

 	e1 = true;
-	assert_d_eq(mallctl("thread.tcache.enabled", &e0, &sz, &e1, sz), 0,
-	    "Unexpected mallctl() error");
+	assert_d_eq(mallctl("thread.tcache.enabled", (void *)&e0, &sz,
+	    (void *)&e1, sz), 0, "Unexpected mallctl() error");
 	assert_false(e0, "tcache should be disabled");

 	e1 = true;
-	assert_d_eq(mallctl("thread.tcache.enabled", &e0, &sz, &e1, sz), 0,
-	    "Unexpected mallctl() error");
+	assert_d_eq(mallctl("thread.tcache.enabled", (void *)&e0, &sz,
+	    (void *)&e1, sz), 0, "Unexpected mallctl() error");
 	assert_true(e0, "tcache should be enabled");

 	e1 = false;
-	assert_d_eq(mallctl("thread.tcache.enabled", &e0, &sz, &e1, sz), 0,
-	    "Unexpected mallctl() error");
+	assert_d_eq(mallctl("thread.tcache.enabled", (void *)&e0, &sz,
+	    (void *)&e1, sz), 0, "Unexpected mallctl() error");
 	assert_true(e0, "tcache should be enabled");

 	e1 = false;
-	assert_d_eq(mallctl("thread.tcache.enabled", &e0, &sz, &e1, sz), 0,
-	    "Unexpected mallctl() error");
+	assert_d_eq(mallctl("thread.tcache.enabled", (void *)&e0, &sz,
+	    (void *)&e1, sz), 0, "Unexpected mallctl() error");
 	assert_false(e0, "tcache should be disabled");

 	free(malloc(1));
 	e1 = true;
-	assert_d_eq(mallctl("thread.tcache.enabled", &e0, &sz, &e1, sz), 0,
-	    "Unexpected mallctl() error");
+	assert_d_eq(mallctl("thread.tcache.enabled", (void *)&e0, &sz,
+	    (void *)&e1, sz), 0, "Unexpected mallctl() error");
 	assert_false(e0, "tcache should be disabled");

 	free(malloc(1));
 	e1 = true;
-	assert_d_eq(mallctl("thread.tcache.enabled", &e0, &sz, &e1, sz), 0,
-	    "Unexpected mallctl() error");
+	assert_d_eq(mallctl("thread.tcache.enabled", (void *)&e0, &sz,
+	    (void *)&e1, sz), 0, "Unexpected mallctl() error");
 	assert_true(e0, "tcache should be enabled");

 	free(malloc(1));
 	e1 = false;
-	assert_d_eq(mallctl("thread.tcache.enabled", &e0, &sz, &e1, sz), 0,
-	    "Unexpected mallctl() error");
+	assert_d_eq(mallctl("thread.tcache.enabled", (void *)&e0, &sz,
+	    (void *)&e1, sz), 0, "Unexpected mallctl() error");
 	assert_true(e0, "tcache should be enabled");

 	free(malloc(1));
 	e1 = false;
-	assert_d_eq(mallctl("thread.tcache.enabled", &e0, &sz, &e1, sz), 0,
-	    "Unexpected mallctl() error");
+	assert_d_eq(mallctl("thread.tcache.enabled", (void *)&e0, &sz,
+	    (void *)&e1, sz), 0, "Unexpected mallctl() error");
 	assert_false(e0, "tcache should be disabled");

 	free(malloc(1));
--- a/test/integration/xallocx.c
+++ b/test/integration/xallocx.c
@@ -12,8 +12,8 @@ arena_ind(void)

 	if (ind == 0) {
 		size_t sz = sizeof(ind);
-		assert_d_eq(mallctl("arenas.extend", &ind, &sz, NULL, 0), 0,
-		    "Unexpected mallctl failure creating arena");
+		assert_d_eq(mallctl("arenas.extend", (void *)&ind, &sz, NULL,
+		    0), 0, "Unexpected mallctl failure creating arena");
 	}

 	return (ind);
@@ -74,7 +74,7 @@ get_nsizes_impl(const char *cmd)
 	size_t z;

 	z = sizeof(unsigned);
-	assert_d_eq(mallctl(cmd, &ret, &z, NULL, 0), 0,
+	assert_d_eq(mallctl(cmd, (void *)&ret, &z, NULL, 0), 0,
 	    "Unexpected mallctl(\"%s\", ...) failure", cmd);

 	return (ret);
@@ -114,7 +114,7 @@ get_size_impl(const char *cmd, size_t ind)
 	    0, "Unexpected mallctlnametomib(\"%s\", ...) failure", cmd);
 	mib[2] = ind;
 	z = sizeof(size_t);
-	assert_d_eq(mallctlbymib(mib, miblen, &ret, &z, NULL, 0),
+	assert_d_eq(mallctlbymib(mib, miblen, (void *)&ret, &z, NULL, 0),
 	    0, "Unexpected mallctlbymib([\"%s\", %zu], ...) failure", cmd, ind);

 	return (ret);
--- a/test/integration/xallocx.sh
+++ b/test/integration/xallocx.sh
@@ -0,0 +1,5 @@
+#!/bin/sh
+
+if [ "x${enable_fill}" = "x1" ] ; then
+  export MALLOC_CONF="junk:false"
+fi
--- a/test/src/mtx.c
+++ b/test/src/mtx.c
@@ -11,6 +11,8 @@ mtx_init(mtx_t *mtx)
 #ifdef _WIN32
 	if (!InitializeCriticalSectionAndSpinCount(&mtx->lock, _CRT_SPINCOUNT))
 		return (true);
+#elif (defined(JEMALLOC_OS_UNFAIR_LOCK))
+	mtx->lock = OS_UNFAIR_LOCK_INIT;
 #elif (defined(JEMALLOC_OSSPIN))
 	mtx->lock = 0;
 #else
@@ -33,6 +35,7 @@ mtx_fini(mtx_t *mtx)
 {

 #ifdef _WIN32
+#elif (defined(JEMALLOC_OS_UNFAIR_LOCK))
 #elif (defined(JEMALLOC_OSSPIN))
 #else
 	pthread_mutex_destroy(&mtx->lock);
@@ -45,6 +48,8 @@ mtx_lock(mtx_t *mtx)

 #ifdef _WIN32
 	EnterCriticalSection(&mtx->lock);
+#elif (defined(JEMALLOC_OS_UNFAIR_LOCK))
+	os_unfair_lock_lock(&mtx->lock);
 #elif (defined(JEMALLOC_OSSPIN))
 	OSSpinLockLock(&mtx->lock);
 #else
@@ -58,6 +63,8 @@ mtx_unlock(mtx_t *mtx)

 #ifdef _WIN32
 	LeaveCriticalSection(&mtx->lock);
+#elif (defined(JEMALLOC_OS_UNFAIR_LOCK))
+	os_unfair_lock_unlock(&mtx->lock);
 #elif (defined(JEMALLOC_OSSPIN))
 	OSSpinLockUnlock(&mtx->lock);
 #else
--- a/test/src/test.c
+++ b/test/src/test.c
@@ -60,32 +60,30 @@ p_test_fini(void)
 	malloc_printf("%s: %s\n", test_name, test_status_string(test_status));
 }

-test_status_t
-p_test(test_t *t, ...)
+static test_status_t
+p_test_impl(bool do_malloc_init, test_t *t, va_list ap)
 {
 	test_status_t ret;
-	va_list ap;

-	/*
-	 * Make sure initialization occurs prior to running tests.  Tests are
-	 * special because they may use internal facilities prior to triggering
-	 * initialization as a side effect of calling into the public API.  This
-	 * is a final safety that works even if jemalloc_constructor() doesn't
-	 * run, as for MSVC builds.
-	 */
-	if (nallocx(1, 0) == 0) {
-		malloc_printf("Initialization error");
-		return (test_status_fail);
+	if (do_malloc_init) {
+		/*
+		 * Make sure initialization occurs prior to running tests.
+		 * Tests are special because they may use internal facilities
+		 * prior to triggering initialization as a side effect of
+		 * calling into the public API.
+		 */
+		if (nallocx(1, 0) == 0) {
+			malloc_printf("Initialization error");
+			return (test_status_fail);
+		}
 	}

 	ret = test_status_pass;
-	va_start(ap, t);
 	for (; t != NULL; t = va_arg(ap, test_t *)) {
 		t();
 		if (test_status > ret)
 			ret = test_status;
 	}
-	va_end(ap);

 	malloc_printf("--- %s: %u/%u, %s: %u/%u, %s: %u/%u ---\n",
 	    test_status_string(test_status_pass),
@@ -98,6 +96,34 @@ p_test(test_t *t, ...)
 	return (ret);
 }

+test_status_t
+p_test(test_t *t, ...)
+{
+	test_status_t ret;
+	va_list ap;
+
+	ret = test_status_pass;
+	va_start(ap, t);
+	ret = p_test_impl(true, t, ap);
+	va_end(ap);
+
+	return (ret);
+}
+
+test_status_t
+p_test_no_malloc_init(test_t *t, ...)
+{
+	test_status_t ret;
+	va_list ap;
+
+	ret = test_status_pass;
+	va_start(ap, t);
+	ret = p_test_impl(false, t, ap);
+	va_end(ap);
+
+	return (ret);
+}
+
 void
 p_test_fail(const char *prefix, const char *message)
 {
--- a/test/src/timer.c
+++ b/test/src/timer.c
@@ -32,9 +32,8 @@ timer_ratio(timedelta_t *a, timedelta_t *b, char *buf, size_t buflen)
 	uint64_t t0 = timer_usec(a);
 	uint64_t t1 = timer_usec(b);
 	uint64_t mult;
-	unsigned i = 0;
-	unsigned j;
-	int n;
+	size_t i = 0;
+	size_t j, n;

 	/* Whole. */
 	n = malloc_snprintf(&buf[i], buflen-i, "%"FMTu64, t0 / t1);
--- a/test/stress/microbench.c
+++ b/test/stress/microbench.c
@@ -1,7 +1,8 @@
 #include "test/jemalloc_test.h"

 JEMALLOC_INLINE_C void
-time_func(timedelta_t *timer, uint64_t nwarmup, uint64_t niter, void (*func)(void))
+time_func(timedelta_t *timer, uint64_t nwarmup, uint64_t niter,
+    void (*func)(void))
 {
 	uint64_t i;

--- a/test/test.sh.in
+++ b/test/test.sh.in
@@ -11,6 +11,18 @@ case @abi@ in
    ;;
 esac

+# Make a copy of the @JEMALLOC_CPREFIX@MALLOC_CONF passed in to this script, so
+# it can be repeatedly concatenated with per test settings.
+export MALLOC_CONF_ALL=${@JEMALLOC_CPREFIX@MALLOC_CONF}
+# Concatenate the individual test's MALLOC_CONF and MALLOC_CONF_ALL.
+export_malloc_conf() {
+  if [ "x${MALLOC_CONF}" != "x" -a "x${MALLOC_CONF_ALL}" != "x" ] ; then
+    export @JEMALLOC_CPREFIX@MALLOC_CONF="${MALLOC_CONF},${MALLOC_CONF_ALL}"
+  else
+    export @JEMALLOC_CPREFIX@MALLOC_CONF="${MALLOC_CONF}${MALLOC_CONF_ALL}"
+  fi
+}
+
 # Corresponds to test_status_t.
 pass_code=0
 skip_code=1
@@ -24,7 +36,22 @@ for t in $@; do
    echo
  fi
  echo "=== ${t} ==="
-  ${t}@exe@ @abs_srcroot@ @abs_objroot@
+  if [ -e "@srcroot@${t}.sh" ] ; then
+    # Source the shell script corresponding to the test in a subshell and
+    # execute the test.  This allows the shell script to set MALLOC_CONF, which
+    # is then used to set @JEMALLOC_CPREFIX@MALLOC_CONF (thus allowing the
+    # per test shell script to ignore the @JEMALLOC_CPREFIX@ detail).
+    $(enable_fill=@enable_fill@ \
+      enable_prof=@enable_prof@ \
+      enable_tcache=@enable_tcache@ \
+      . @srcroot@${t}.sh && \
+      export_malloc_conf && \
+      ${t}@exe@ @abs_srcroot@ @abs_objroot@)
+  else
+    $(export MALLOC_CONF= && \
+      export_malloc_conf &&
+      ${t}@exe@ @abs_srcroot@ @abs_objroot@)
+  fi
  result_code=$?
  case ${result_code} in
    ${pass_code})
--- a/test/unit/a0.c
+++ b/test/unit/a0.c
@@ -0,0 +1,19 @@
+#include "test/jemalloc_test.h"
+
+TEST_BEGIN(test_a0)
+{
+	void *p;
+
+	p = a0malloc(1);
+	assert_ptr_not_null(p, "Unexpected a0malloc() error");
+	a0dalloc(p);
+}
+TEST_END
+
+int
+main(void)
+{
+
+	return (test_no_malloc_init(
+	    test_a0));
+}
--- a/test/unit/arena_reset.c
+++ b/test/unit/arena_reset.c
@@ -0,0 +1,155 @@
+#include "test/jemalloc_test.h"
+
+static unsigned
+get_nsizes_impl(const char *cmd)
+{
+	unsigned ret;
+	size_t z;
+
+	z = sizeof(unsigned);
+	assert_d_eq(mallctl(cmd, (void *)&ret, &z, NULL, 0), 0,
+	    "Unexpected mallctl(\"%s\", ...) failure", cmd);
+
+	return (ret);
+}
+
+static unsigned
+get_nsmall(void)
+{
+
+	return (get_nsizes_impl("arenas.nbins"));
+}
+
+static unsigned
+get_nlarge(void)
+{
+
+	return (get_nsizes_impl("arenas.nlruns"));
+}
+
+static unsigned
+get_nhuge(void)
+{
+
+	return (get_nsizes_impl("arenas.nhchunks"));
+}
+
+static size_t
+get_size_impl(const char *cmd, size_t ind)
+{
+	size_t ret;
+	size_t z;
+	size_t mib[4];
+	size_t miblen = 4;
+
+	z = sizeof(size_t);
+	assert_d_eq(mallctlnametomib(cmd, mib, &miblen),
+	    0, "Unexpected mallctlnametomib(\"%s\", ...) failure", cmd);
+	mib[2] = ind;
+	z = sizeof(size_t);
+	assert_d_eq(mallctlbymib(mib, miblen, (void *)&ret, &z, NULL, 0),
+	    0, "Unexpected mallctlbymib([\"%s\", %zu], ...) failure", cmd, ind);
+
+	return (ret);
+}
+
+static size_t
+get_small_size(size_t ind)
+{
+
+	return (get_size_impl("arenas.bin.0.size", ind));
+}
+
+static size_t
+get_large_size(size_t ind)
+{
+
+	return (get_size_impl("arenas.lrun.0.size", ind));
+}
+
+static size_t
+get_huge_size(size_t ind)
+{
+
+	return (get_size_impl("arenas.hchunk.0.size", ind));
+}
+
+TEST_BEGIN(test_arena_reset)
+{
+#define	NHUGE	4
+	unsigned arena_ind, nsmall, nlarge, nhuge, nptrs, i;
+	size_t sz, miblen;
+	void **ptrs;
+	int flags;
+	size_t mib[3];
+	tsdn_t *tsdn;
+
+	test_skip_if((config_valgrind && unlikely(in_valgrind)) || (config_fill
+	    && unlikely(opt_quarantine)));
+
+	sz = sizeof(unsigned);
+	assert_d_eq(mallctl("arenas.extend", (void *)&arena_ind, &sz, NULL, 0),
+	    0, "Unexpected mallctl() failure");
+
+	flags = MALLOCX_ARENA(arena_ind) | MALLOCX_TCACHE_NONE;
+
+	nsmall = get_nsmall();
+	nlarge = get_nlarge();
+	nhuge = get_nhuge() > NHUGE ? NHUGE : get_nhuge();
+	nptrs = nsmall + nlarge + nhuge;
+	ptrs = (void **)malloc(nptrs * sizeof(void *));
+	assert_ptr_not_null(ptrs, "Unexpected malloc() failure");
+
+	/* Allocate objects with a wide range of sizes. */
+	for (i = 0; i < nsmall; i++) {
+		sz = get_small_size(i);
+		ptrs[i] = mallocx(sz, flags);
+		assert_ptr_not_null(ptrs[i],
+		    "Unexpected mallocx(%zu, %#x) failure", sz, flags);
+	}
+	for (i = 0; i < nlarge; i++) {
+		sz = get_large_size(i);
+		ptrs[nsmall + i] = mallocx(sz, flags);
+		assert_ptr_not_null(ptrs[i],
+		    "Unexpected mallocx(%zu, %#x) failure", sz, flags);
+	}
+	for (i = 0; i < nhuge; i++) {
+		sz = get_huge_size(i);
+		ptrs[nsmall + nlarge + i] = mallocx(sz, flags);
+		assert_ptr_not_null(ptrs[i],
+		    "Unexpected mallocx(%zu, %#x) failure", sz, flags);
+	}
+
+	tsdn = tsdn_fetch();
+
+	/* Verify allocations. */
+	for (i = 0; i < nptrs; i++) {
+		assert_zu_gt(ivsalloc(tsdn, ptrs[i], false), 0,
+		    "Allocation should have queryable size");
+	}
+
+	/* Reset. */
+	miblen = sizeof(mib)/sizeof(size_t);
+	assert_d_eq(mallctlnametomib("arena.0.reset", mib, &miblen), 0,
+	    "Unexpected mallctlnametomib() failure");
+	mib[1] = (size_t)arena_ind;
+	assert_d_eq(mallctlbymib(mib, miblen, NULL, NULL, NULL, 0), 0,
+	    "Unexpected mallctlbymib() failure");
+
+	/* Verify allocations no longer exist. */
+	for (i = 0; i < nptrs; i++) {
+		assert_zu_eq(ivsalloc(tsdn, ptrs[i], false), 0,
+		    "Allocation should no longer exist");
+	}
+
+	free(ptrs);
+}
+TEST_END
+
+int
+main(void)
+{
+
+	return (test(
+	    test_arena_reset));
+}
--- a/test/unit/arena_reset.sh
+++ b/test/unit/arena_reset.sh
@@ -0,0 +1,5 @@
+#!/bin/sh
+
+if [ "x${enable_prof}" = "x1" ] ; then
+  export MALLOC_CONF="prof:true,lg_prof_sample:0"
+fi
--- a/test/unit/bitmap.c
+++ b/test/unit/bitmap.c
@@ -101,7 +101,7 @@ TEST_BEGIN(test_bitmap_sfu)
 		bitmap_info_t binfo;
 		bitmap_info_init(&binfo, i);
 		{
-			ssize_t j;
+			size_t j;
 			bitmap_t *bitmap = (bitmap_t *)malloc(
 			    bitmap_size(&binfo));
 			bitmap_init(bitmap, &binfo);
@@ -119,7 +119,7 @@ TEST_BEGIN(test_bitmap_sfu)
 			 * Iteratively unset bits starting at the end, and
 			 * verify that bitmap_sfu() reaches the unset bits.
 			 */
-			for (j = i - 1; j >= 0; j--) {
+			for (j = i - 1; j < i; j--) { /* (i..0] */
 				bitmap_unset(bitmap, &binfo, j);
 				assert_zd_eq(bitmap_sfu(bitmap, &binfo), j,
 				    "First unset bit should the bit previously "
--- a/test/unit/ckh.c
+++ b/test/unit/ckh.c
@@ -7,8 +7,8 @@ TEST_BEGIN(test_new_delete)

 	tsd = tsd_fetch();

-	assert_false(ckh_new(tsd, &ckh, 2, ckh_string_hash, ckh_string_keycomp),
-	    "Unexpected ckh_new() error");
+	assert_false(ckh_new(tsd, &ckh, 2, ckh_string_hash,
+	    ckh_string_keycomp), "Unexpected ckh_new() error");
 	ckh_delete(tsd, &ckh);

 	assert_false(ckh_new(tsd, &ckh, 3, ckh_pointer_hash,
@@ -32,8 +32,8 @@ TEST_BEGIN(test_count_insert_search_remove)

 	tsd = tsd_fetch();

-	assert_false(ckh_new(tsd, &ckh, 2, ckh_string_hash, ckh_string_keycomp),
-	    "Unexpected ckh_new() error");
+	assert_false(ckh_new(tsd, &ckh, 2, ckh_string_hash,
+	    ckh_string_keycomp), "Unexpected ckh_new() error");
 	assert_zu_eq(ckh_count(&ckh), 0,
 	    "ckh_count() should return %zu, but it returned %zu", ZU(0),
 	    ckh_count(&ckh));
--- a/Show More
+++ b/Show More