Merge branch 'dev'

We lookup freed extents for testing in post_reset.  Take background_thread lock
so that the extents are not modified at the same time.

.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
+  - file lib/jemalloc.dll
+  - mingw32-make tests
+  - mingw32-make -k check

.autom4te.cfg

@@ -0,0 +1,3 @@
+begin-language: "Autoconf-without-aclocal-m4"
+args: --no-cache
+end-language: "Autoconf-without-aclocal-m4"

.gitattributes

@@ -0,0 +1 @@
+* text=auto eol=lf

.gitignore

@@ -1,16 +1,91 @@
-/jemalloc/autom4te.cache/
-/jemalloc/config.stamp
-/jemalloc/config.log
-/jemalloc/config.status
-/jemalloc/configure
-/jemalloc/doc/jemalloc.3
-/jemalloc/lib/
-/jemalloc/Makefile
-/jemalloc/include/jemalloc/internal/jemalloc_internal\.h
-/jemalloc/include/jemalloc/jemalloc\.h
-/jemalloc/include/jemalloc/jemalloc_defs\.h
-/jemalloc/test/jemalloc_test\.h
-/jemalloc/src/*.[od]
-/jemalloc/test/*.[od]
-/jemalloc/test/*.out
-/jemalloc/VERSION
+/bin/jemalloc-config
+/bin/jemalloc.sh
+/bin/jeprof
+
+/config.stamp
+/config.log
+/config.status
+/configure
+
+/doc/html.xsl
+/doc/manpages.xsl
+/doc/jemalloc.xml
+/doc/jemalloc.html
+/doc/jemalloc.3
+
+/jemalloc.pc
+
+/lib/
+
+/Makefile
+
+/include/jemalloc/internal/jemalloc_preamble.h
+/include/jemalloc/internal/jemalloc_internal_defs.h
+/include/jemalloc/internal/private_namespace.gen.h
+/include/jemalloc/internal/private_namespace.h
+/include/jemalloc/internal/private_namespace_jet.gen.h
+/include/jemalloc/internal/private_namespace_jet.h
+/include/jemalloc/internal/private_symbols.awk
+/include/jemalloc/internal/private_symbols_jet.awk
+/include/jemalloc/internal/public_namespace.h
+/include/jemalloc/internal/public_symbols.txt
+/include/jemalloc/internal/public_unnamespace.h
+/include/jemalloc/internal/size_classes.h
+/include/jemalloc/jemalloc.h
+/include/jemalloc/jemalloc_defs.h
+/include/jemalloc/jemalloc_macros.h
+/include/jemalloc/jemalloc_mangle.h
+/include/jemalloc/jemalloc_mangle_jet.h
+/include/jemalloc/jemalloc_protos.h
+/include/jemalloc/jemalloc_protos_jet.h
+/include/jemalloc/jemalloc_rename.h
+/include/jemalloc/jemalloc_typedefs.h
+
+/src/*.[od]
+/src/*.sym
+
+/run_tests.out/
+
+/test/test.sh
+test/include/test/jemalloc_test.h
+test/include/test/jemalloc_test_defs.h
+
+/test/integration/[A-Za-z]*
+!/test/integration/[A-Za-z]*.*
+/test/integration/*.[od]
+/test/integration/*.out
+
+/test/integration/cpp/[A-Za-z]*
+!/test/integration/cpp/[A-Za-z]*.*
+/test/integration/cpp/*.[od]
+/test/integration/cpp/*.out
+
+/test/src/*.[od]
+
+/test/stress/[A-Za-z]*
+!/test/stress/[A-Za-z]*.*
+/test/stress/*.[od]
+/test/stress/*.out
+
+/test/unit/[A-Za-z]*
+!/test/unit/[A-Za-z]*.*
+/test/unit/*.[od]
+/test/unit/*.out
+
+/VERSION
+
+*.pdb
+*.sdf
+*.opendb
+*.opensdf
+*.cachefile
+*.suo
+*.user
+*.sln.docstates
+*.tmp
+/msvc/Win32/
+/msvc/x64/
+/msvc/projects/*/*/Debug*/
+/msvc/projects/*/*/Release*/
+/msvc/projects/*/*/Win32/
+/msvc/projects/*/*/x64/

.travis.yml

@@ -0,0 +1,155 @@
+language: generic
+
+matrix:
+  include:
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: osx
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=clang CXX=clang++ COMPILER_FLAGS="" CONFIGURE_FLAGS="" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="-m32" CONFIGURE_FLAGS="" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+      addons:
+        apt:
+          packages:
+            - gcc-multilib
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-debug" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-prof" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--disable-stats" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--with-malloc-conf=tcache:false" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--with-malloc-conf=dss:primary" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--with-malloc-conf=percpu_arena:percpu" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--with-malloc-conf=background_thread:true" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: osx
+      env: CC=clang CXX=clang++ COMPILER_FLAGS="" CONFIGURE_FLAGS="" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: osx
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="-m32" CONFIGURE_FLAGS="" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: osx
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-debug" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: osx
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--disable-stats" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: osx
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--with-malloc-conf=tcache:false" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=clang CXX=clang++ COMPILER_FLAGS="-m32" CONFIGURE_FLAGS="" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+      addons:
+        apt:
+          packages:
+            - gcc-multilib
+    - os: linux
+      env: CC=clang CXX=clang++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-debug" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=clang CXX=clang++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-prof" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=clang CXX=clang++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--disable-stats" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=clang CXX=clang++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--with-malloc-conf=tcache:false" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=clang CXX=clang++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--with-malloc-conf=dss:primary" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=clang CXX=clang++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--with-malloc-conf=percpu_arena:percpu" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=clang CXX=clang++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--with-malloc-conf=background_thread:true" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="-m32" CONFIGURE_FLAGS="--enable-debug" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+      addons:
+        apt:
+          packages:
+            - gcc-multilib
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="-m32" CONFIGURE_FLAGS="--enable-prof" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+      addons:
+        apt:
+          packages:
+            - gcc-multilib
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="-m32" CONFIGURE_FLAGS="--disable-stats" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+      addons:
+        apt:
+          packages:
+            - gcc-multilib
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="-m32" CONFIGURE_FLAGS="--with-malloc-conf=tcache:false" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+      addons:
+        apt:
+          packages:
+            - gcc-multilib
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="-m32" CONFIGURE_FLAGS="--with-malloc-conf=dss:primary" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+      addons:
+        apt:
+          packages:
+            - gcc-multilib
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="-m32" CONFIGURE_FLAGS="--with-malloc-conf=percpu_arena:percpu" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+      addons:
+        apt:
+          packages:
+            - gcc-multilib
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="-m32" CONFIGURE_FLAGS="--with-malloc-conf=background_thread:true" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+      addons:
+        apt:
+          packages:
+            - gcc-multilib
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-debug --enable-prof" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-debug --disable-stats" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-debug --with-malloc-conf=tcache:false" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-debug --with-malloc-conf=dss:primary" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-debug --with-malloc-conf=percpu_arena:percpu" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-debug --with-malloc-conf=background_thread:true" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-prof --disable-stats" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-prof --with-malloc-conf=tcache:false" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-prof --with-malloc-conf=dss:primary" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-prof --with-malloc-conf=percpu_arena:percpu" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--enable-prof --with-malloc-conf=background_thread:true" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--disable-stats --with-malloc-conf=tcache:false" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--disable-stats --with-malloc-conf=dss:primary" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--disable-stats --with-malloc-conf=percpu_arena:percpu" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--disable-stats --with-malloc-conf=background_thread:true" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--with-malloc-conf=tcache:false,dss:primary" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--with-malloc-conf=tcache:false,percpu_arena:percpu" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--with-malloc-conf=tcache:false,background_thread:true" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--with-malloc-conf=dss:primary,percpu_arena:percpu" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--with-malloc-conf=dss:primary,background_thread:true" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+    - os: linux
+      env: CC=gcc CXX=g++ COMPILER_FLAGS="" CONFIGURE_FLAGS="--with-malloc-conf=percpu_arena:percpu,background_thread:true" EXTRA_CFLAGS="-Werror -Wno-array-bounds"
+
+
+before_script:
+  - autoconf
+  - ./configure ${COMPILER_FLAGS:+       CC="$CC $COMPILER_FLAGS"       CXX="$CXX $COMPILER_FLAGS" }       $CONFIGURE_FLAGS
+  - make -j3
+  - make -j3 tests
+
+script:
+  - make check
+

COPYING

@@ -0,0 +1,27 @@
+Unless otherwise specified, files in the jemalloc source distribution are
+subject to the following license:
+--------------------------------------------------------------------------------
+Copyright (C) 2002-2017 Jason Evans <jasone@canonware.com>.
+All rights reserved.
+Copyright (C) 2007-2012 Mozilla Foundation.  All rights reserved.
+Copyright (C) 2009-2017 Facebook, Inc.  All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are met:
+1. Redistributions of source code must retain the above copyright notice(s),
+   this list of conditions and the following disclaimer.
+2. Redistributions in binary form must reproduce the above copyright notice(s),
+   this list of conditions and the following disclaimer in the documentation
+   and/or other materials provided with the distribution.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDER(S) ``AS IS'' AND ANY EXPRESS
+OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO
+EVENT SHALL THE COPYRIGHT HOLDER(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE
+OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
+ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+--------------------------------------------------------------------------------

INSTALL.md

@@ -0,0 +1,411 @@
+Building and installing a packaged release of jemalloc can be as simple as
+typing the following while in the root directory of the source tree:
+
+    ./configure
+    make
+    make install
+
+If building from unpackaged developer sources, the simplest command sequence
+that might work is:
+
+    ./autogen.sh
+    make dist
+    make
+    make install
+
+Note that documentation is not built by the default target because doing so
+would create a dependency on xsltproc in packaged releases, hence the
+requirement to either run 'make dist' or avoid installing docs via the various
+install_* targets documented below.
+
+
+## Advanced configuration
+
+The 'configure' script supports numerous options that allow control of which
+functionality is enabled, where jemalloc is installed, etc.  Optionally, pass
+any of the following arguments (not a definitive list) to 'configure':
+
+* `--help`
+
+    Print a definitive list of options.
+
+* `--prefix=<install-root-dir>`
+
+    Set the base directory in which to install.  For example:
+
+        ./configure --prefix=/usr/local
+
+    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>|VERSION)`
+
+    The VERSION file is mandatory for successful configuration, and the
+    following steps are taken to assure its presence:
+    1) If --with-version=<major>.<minor>.<bugfix>-<nrev>-g<gid> is specified,
+       generate VERSION using the specified value.
+    2) If --with-version is not specified in either form and the source
+       directory is inside a git repository, try to generate VERSION via 'git
+       describe' invocations that pattern-match release tags.
+    3) If VERSION is missing, generate it with a bogus version:
+       0.0.0-0-g0000000000000000000000000000000000000000
+
+    Note that --with-version=VERSION bypasses (1) and (2), which simplifies
+    VERSION configuration when embedding a jemalloc release into another
+    project's git repository.
+
+* `--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.
+
+* `--with-mangling=<map>`
+
+    Mangle public symbols specified in <map> which is a comma-separated list of
+    name:mangled pairs.
+
+    For example, to use ld's --wrap option as an alternative method for
+    overriding libc's malloc implementation, specify something like:
+
+      --with-mangling=malloc:__wrap_malloc,free:__wrap_free[...]
+
+    Note that mangling happens prior to application of the prefix specified by
+    --with-jemalloc-prefix, and mangled symbols are then ignored when applying
+    the prefix.
+
+* `--with-jemalloc-prefix=<prefix>`
+
+    Prefix all public APIs with <prefix>.  For example, if <prefix> is
+    "prefix_", API changes like the following occur:
+
+      malloc()         --> prefix_malloc()
+      malloc_conf      --> prefix_malloc_conf
+      /etc/malloc.conf --> /etc/prefix_malloc.conf
+      MALLOC_CONF      --> PREFIX_MALLOC_CONF
+
+    This makes it possible to use jemalloc at the same time as the system
+    allocator, or even to use multiple copies of jemalloc simultaneously.
+
+    By default, the prefix is "", except on OS X, where it is "je_".  On OS X,
+    jemalloc overlays the default malloc zone, but makes no attempt to actually
+    replace the "malloc", "calloc", etc. symbols.
+
+* `--without-export`
+
+    Don't export public APIs.  This can be useful when building jemalloc as a
+    static library, or to avoid exporting public APIs when using the zone
+    allocator on OSX.
+
+* `--with-private-namespace=<prefix>`
+
+    Prefix all library-private APIs with <prefix>je_.  For shared libraries,
+    symbol visibility mechanisms prevent these symbols from being exported, but
+    for static libraries, naming collisions are a real possibility.  By
+    default, <prefix> is empty, which results in a symbol prefix of je_ .
+
+* `--with-install-suffix=<suffix>`
+
+    Append <suffix> to the base name of all installed files, such that multiple
+    versions of jemalloc can coexist in the same installation directory.  For
+    example, libjemalloc.so.0 becomes libjemalloc<suffix>.so.0.
+
+* `--with-malloc-conf=<malloc_conf>`
+
+    Embed `<malloc_conf>` as a run-time options string that is processed prior to
+    the malloc_conf global variable, the /etc/malloc.conf symlink, and the
+    MALLOC_CONF environment variable.  For example, to change the default decay
+    time to 30 seconds:
+
+      --with-malloc-conf=decay_ms:30000
+
+* `--enable-debug`
+
+    Enable assertions and validation code.  This incurs a substantial
+    performance hit, but is very useful during application development.
+
+* `--disable-stats`
+
+    Disable statistics gathering functionality.  See the "opt.stats_print"
+    option documentation for usage details.
+
+* `--enable-prof`
+
+    Enable heap profiling and leak detection functionality.  See the "opt.prof"
+    option documentation for usage details.  When enabled, there are several
+    approaches to backtracing, and the configure script chooses the first one
+    in the following list that appears to function correctly:
+
+    + libunwind      (requires --enable-prof-libunwind)
+    + libgcc         (unless --disable-prof-libgcc)
+    + gcc intrinsics (unless --disable-prof-gcc)
+
+* `--enable-prof-libunwind`
+
+    Use the libunwind library (http://www.nongnu.org/libunwind/) for stack
+    backtracing.
+
+* `--disable-prof-libgcc`
+
+    Disable the use of libgcc's backtracing functionality.
+
+* `--disable-prof-gcc`
+
+    Disable the use of gcc intrinsics for backtracing.
+
+* `--with-static-libunwind=<libunwind.a>`
+
+    Statically link against the specified libunwind.a rather than dynamically
+    linking with -lunwind.
+
+* `--disable-thp`
+
+    Disable transparent huge page (THP) integration.  This option can be useful
+    when cross compiling.
+
+* `--disable-fill`
+
+    Disable support for junk/zero filling of memory.  See the "opt.junk" and
+    "opt.zero" option documentation for usage details.
+
+* `--disable-zone-allocator`
+
+    Disable zone allocator for Darwin.  This means jemalloc won't be hooked as
+    the default allocator on OSX/iOS.
+
+* `--enable-utrace`
+
+    Enable utrace(2)-based allocation tracing.  This feature is not broadly
+    portable (FreeBSD has it, but Linux and OS X do not).
+
+* `--enable-xmalloc`
+
+    Enable support for optional immediate termination due to out-of-memory
+    errors, as is commonly implemented by "xmalloc" wrapper function for malloc.
+    See the "opt.xmalloc" option documentation for usage details.
+
+* `--enable-lazy-lock`
+
+    Enable code that wraps pthread_create() to detect when an application
+    switches from single-threaded to multi-threaded mode, so that it can avoid
+    mutex locking/unlocking operations while in single-threaded mode.  In
+    practice, this feature usually has little impact on performance unless
+    thread-specific caching is disabled.
+
+* `--disable-cache-oblivious`
+
+    Disable cache-oblivious large allocation alignment for large allocation
+    requests with no alignment constraints.  If this feature is disabled, all
+    large allocations are page-aligned as an implementation artifact, which can
+    severely harm CPU cache utilization.  However, the cache-oblivious layout
+    comes at the cost of one extra page per large allocation, which in the
+    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).
+
+* `--disable-cxx`
+
+    Disable C++ integration.  This will cause new and delete operator
+    implementations to be omitted.
+
+* `--with-xslroot=<path>`
+
+    Specify where to find DocBook XSL stylesheets when building the
+    documentation.
+
+* `--with-lg-page=<lg-page>`
+
+    Specify the base 2 log of the allocator page size, which must in turn be at
+    least as large as the system page size.  By default the configure script
+    determines the host's page size and sets the allocator page size equal to
+    the system page size, so this option need not be specified unless the
+    system page size may change between configuration and execution, e.g. when
+    cross compiling.
+
+* `--with-lg-page-sizes=<lg-page-sizes>`
+
+    Specify the comma-separated base 2 logs of the page sizes to support.  This
+    option may be useful when cross compiling in combination with
+    `--with-lg-page`, but its primary use case is for integration with FreeBSD's
+    libc, wherein jemalloc is embedded.
+
+* `--with-lg-hugepage=<lg-hugepage>`
+
+    Specify the base 2 log of the system huge page size.  This option is useful
+    when cross compiling, or when overriding the default for systems that do
+    not explicitly support huge pages.
+
+* `--with-lg-quantum=<lg-quantum>`
+
+    Specify the base 2 log of the minimum allocation alignment.  jemalloc needs
+    to know the minimum alignment that meets the following C standard
+    requirement (quoted from the April 12, 2011 draft of the C11 standard):
+
+    >  The pointer returned if the allocation succeeds is suitably aligned so
+      that it may be assigned to a pointer to any type of object with a
+      fundamental alignment requirement and then used to access such an object
+      or an array of such objects in the space allocated [...]
+
+    This setting is architecture-specific, and although jemalloc includes known
+    safe values for the most commonly used modern architectures, there is a
+    wrinkle related to GNU libc (glibc) that may impact your choice of
+    <lg-quantum>.  On most modern architectures, this mandates 16-byte
+    alignment (<lg-quantum>=4), but the glibc developers chose not to meet this
+    requirement for performance reasons.  An old discussion can be found at
+    <https://sourceware.org/bugzilla/show_bug.cgi?id=206> .  Unlike glibc,
+    jemalloc does follow the C standard by default (caveat: jemalloc
+    technically cheats for size classes smaller than the quantum), but the fact
+    that Linux systems already work around this allocator noncompliance means
+    that it is generally safe in practice to let jemalloc's minimum alignment
+    follow glibc's lead.  If you specify `--with-lg-quantum=3` during
+    configuration, jemalloc will provide additional size classes that are not
+    16-byte-aligned (24, 40, and 56).
+
+The following environment variables (not a definitive list) impact configure's
+behavior:
+
+* `CFLAGS="?"`
+* `CXXFLAGS="?"`
+
+    Pass these flags to the C/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="?"`
+* `EXTRA_CXXFLAGS="?"`
+
+    Append these flags to CFLAGS/CXXFLAGS, without passing them to the
+    compiler(s) 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
+    'cpp' when 'configure' is looking for include files, so you must use
+    CPPFLAGS instead if you need to help 'configure' find header files.
+
+* `LD_LIBRARY_PATH="?"`
+
+    'ld' uses this colon-separated list to find libraries.
+
+* `LDFLAGS="?"`
+
+    Pass these flags when linking.
+
+* `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:
+
+    build_lib_shared
+    build_lib_static
+    build_lib
+    build_doc_html
+    build_doc_man
+    build_doc
+
+To install only parts of jemalloc, use the following targets:
+
+    install_bin
+    install_include
+    install_lib_shared
+    install_lib_static
+    install_lib
+    install_doc_html
+    install_doc_man
+    install_doc
+
+To clean up build results to varying degrees, use the following make targets:
+
+    clean
+    distclean
+    relclean
+
+
+## Advanced installation
+
+Optionally, define make variables when invoking make, including (not
+exclusively):
+
+* `INCLUDEDIR="?"`
+
+    Use this as the installation prefix for header files.
+
+* `LIBDIR="?"`
+
+    Use this as the installation prefix for libraries.
+
+* `MANDIR="?"`
+
+    Use this as the installation prefix for man pages.
+
+* `DESTDIR="?"`
+
+    Prepend DESTDIR to INCLUDEDIR, LIBDIR, DATADIR, and MANDIR.  This is useful
+    when installing to a different path than was specified via --prefix.
+
+* `CC="?"`
+
+    Use this to invoke the C compiler.
+
+* `CFLAGS="?"`
+
+    Pass these flags to the compiler.
+
+* `CPPFLAGS="?"`
+
+    Pass these flags to the C preprocessor.
+
+* `LDFLAGS="?"`
+
+    Pass these flags when linking.
+
+* `PATH="?"`
+
+    Use this to search for programs used during configuration and building.
+
+
+## Development
+
+If you intend to make non-trivial changes to jemalloc, use the 'autogen.sh'
+script rather than 'configure'.  This re-generates 'configure', enables
+configuration dependency rules, and enables re-generation of automatically
+generated source files.
+
+The build system supports using an object directory separate from the source
+tree.  For example, you can create an 'obj' directory, and from within that
+directory, issue configuration and build commands:
+
+    autoconf
+    mkdir obj
+    cd obj
+    ../configure --enable-autogen
+    make
+
+
+## Documentation
+
+The manual page is generated in both html and roff formats.  Any web browser
+can be used to view the html manual.  The roff manual page can be formatted
+prior to installation via the following command:
+
+    nroff -man -t doc/jemalloc.3

Makefile.in

@@ -0,0 +1,563 @@
+# Clear out all vpaths, then set just one (default vpath) for the main build
+# directory.
+vpath
+vpath % .
+
+# Clear the default suffixes, so that built-in rules are not used.
+.SUFFIXES :
+
+SHELL := /bin/sh
+
+CC := @CC@
+CXX := @CXX@
+
+# Configuration parameters.
+DESTDIR =
+BINDIR := $(DESTDIR)@BINDIR@
+INCLUDEDIR := $(DESTDIR)@INCLUDEDIR@
+LIBDIR := $(DESTDIR)@LIBDIR@
+DATADIR := $(DESTDIR)@DATADIR@
+MANDIR := $(DESTDIR)@MANDIR@
+srcroot := @srcroot@
+objroot := @objroot@
+abs_srcroot := @abs_srcroot@
+abs_objroot := @abs_objroot@
+
+# Build parameters.
+CPPFLAGS := @CPPFLAGS@ -I$(srcroot)include -I$(objroot)include
+CONFIGURE_CFLAGS := @CONFIGURE_CFLAGS@
+SPECIFIED_CFLAGS := @SPECIFIED_CFLAGS@
+EXTRA_CFLAGS := @EXTRA_CFLAGS@
+CFLAGS := $(strip $(CONFIGURE_CFLAGS) $(SPECIFIED_CFLAGS) $(EXTRA_CFLAGS))
+CONFIGURE_CXXFLAGS := @CONFIGURE_CXXFLAGS@
+SPECIFIED_CXXFLAGS := @SPECIFIED_CXXFLAGS@
+EXTRA_CXXFLAGS := @EXTRA_CXXFLAGS@
+CXXFLAGS := $(strip $(CONFIGURE_CXXFLAGS) $(SPECIFIED_CXXFLAGS) $(EXTRA_CXXFLAGS))
+LDFLAGS := @LDFLAGS@
+EXTRA_LDFLAGS := @EXTRA_LDFLAGS@
+LIBS := @LIBS@
+RPATH_EXTRA := @RPATH_EXTRA@
+SO := @so@
+IMPORTLIB := @importlib@
+O := @o@
+A := @a@
+EXE := @exe@
+LIBPREFIX := @libprefix@
+REV := @rev@
+install_suffix := @install_suffix@
+ABI := @abi@
+XSLTPROC := @XSLTPROC@
+AUTOCONF := @AUTOCONF@
+_RPATH = @RPATH@
+RPATH = $(if $(1),$(call _RPATH,$(1)))
+cfghdrs_in := $(addprefix $(srcroot),@cfghdrs_in@)
+cfghdrs_out := @cfghdrs_out@
+cfgoutputs_in := $(addprefix $(srcroot),@cfgoutputs_in@)
+cfgoutputs_out := @cfgoutputs_out@
+enable_autogen := @enable_autogen@
+enable_prof := @enable_prof@
+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@
+DUMP_SYMS = @DUMP_SYMS@
+AWK := @AWK@
+CC_MM = @CC_MM@
+LM := @LM@
+INSTALL = @INSTALL@
+
+ifeq (macho, $(ABI))
+TEST_LIBRARY_PATH := DYLD_FALLBACK_LIBRARY_PATH="$(objroot)lib"
+else
+ifeq (pecoff, $(ABI))
+TEST_LIBRARY_PATH := PATH="$(PATH):$(objroot)lib"
+else
+TEST_LIBRARY_PATH :=
+endif
+endif
+
+LIBJEMALLOC := $(LIBPREFIX)jemalloc$(install_suffix)
+
+# Lists of files.
+BINS := $(objroot)bin/jemalloc-config $(objroot)bin/jemalloc.sh $(objroot)bin/jeprof
+C_HDRS := $(objroot)include/jemalloc/jemalloc$(install_suffix).h
+C_SRCS := $(srcroot)src/jemalloc.c \
+	$(srcroot)src/arena.c \
+	$(srcroot)src/background_thread.c \
+	$(srcroot)src/base.c \
+	$(srcroot)src/bitmap.c \
+	$(srcroot)src/ckh.c \
+	$(srcroot)src/ctl.c \
+	$(srcroot)src/extent.c \
+	$(srcroot)src/extent_dss.c \
+	$(srcroot)src/extent_mmap.c \
+	$(srcroot)src/hash.c \
+	$(srcroot)src/hooks.c \
+	$(srcroot)src/large.c \
+	$(srcroot)src/malloc_io.c \
+	$(srcroot)src/mutex.c \
+	$(srcroot)src/mutex_pool.c \
+	$(srcroot)src/nstime.c \
+	$(srcroot)src/pages.c \
+	$(srcroot)src/prng.c \
+	$(srcroot)src/prof.c \
+	$(srcroot)src/rtree.c \
+	$(srcroot)src/stats.c \
+	$(srcroot)src/spin.c \
+	$(srcroot)src/sz.c \
+	$(srcroot)src/tcache.c \
+	$(srcroot)src/ticker.c \
+	$(srcroot)src/tsd.c \
+	$(srcroot)src/witness.c
+ifeq ($(enable_zone_allocator), 1)
+C_SRCS += $(srcroot)src/zone.c
+endif
+ifeq ($(IMPORTLIB),$(SO))
+STATIC_LIBS := $(objroot)lib/$(LIBJEMALLOC).$(A)
+endif
+ifdef PIC_CFLAGS
+STATIC_LIBS += $(objroot)lib/$(LIBJEMALLOC)_pic.$(A)
+else
+STATIC_LIBS += $(objroot)lib/$(LIBJEMALLOC)_s.$(A)
+endif
+DSOS := $(objroot)lib/$(LIBJEMALLOC).$(SOREV)
+ifneq ($(SOREV),$(SO))
+DSOS += $(objroot)lib/$(LIBJEMALLOC).$(SO)
+endif
+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
+DOCS_HTML := $(DOCS_XML:$(objroot)%.xml=$(objroot)%.html)
+DOCS_MAN3 := $(DOCS_XML:$(objroot)%.xml=$(objroot)%.3)
+DOCS := $(DOCS_HTML) $(DOCS_MAN3)
+C_TESTLIB_SRCS := $(srcroot)test/src/btalloc.c $(srcroot)test/src/btalloc_0.c \
+	$(srcroot)test/src/btalloc_1.c $(srcroot)test/src/math.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 :=
+C_UTIL_CPP_SRCS :=
+else
+C_UTIL_INTEGRATION_SRCS := $(srcroot)src/nstime.c $(srcroot)src/malloc_io.c
+C_UTIL_CPP_SRCS := $(srcroot)src/nstime.c $(srcroot)src/malloc_io.c
+endif
+TESTS_UNIT := \
+	$(srcroot)test/unit/a0.c \
+	$(srcroot)test/unit/arena_reset.c \
+	$(srcroot)test/unit/atomic.c \
+	$(srcroot)test/unit/background_thread.c \
+	$(srcroot)test/unit/base.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/hooks.c \
+	$(srcroot)test/unit/junk.c \
+	$(srcroot)test/unit/junk_alloc.c \
+	$(srcroot)test/unit/junk_free.c \
+	$(srcroot)test/unit/mallctl.c \
+	$(srcroot)test/unit/malloc_io.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 \
+	$(srcroot)test/unit/prof_gdump.c \
+	$(srcroot)test/unit/prof_idump.c \
+	$(srcroot)test/unit/prof_reset.c \
+	$(srcroot)test/unit/prof_tctx.c \
+	$(srcroot)test/unit/prof_thread_name.c \
+	$(srcroot)test/unit/ql.c \
+	$(srcroot)test/unit/qr.c \
+	$(srcroot)test/unit/rb.c \
+	$(srcroot)test/unit/retained.c \
+	$(srcroot)test/unit/rtree.c \
+	$(srcroot)test/unit/SFMT.c \
+	$(srcroot)test/unit/size_classes.c \
+	$(srcroot)test/unit/slab.c \
+	$(srcroot)test/unit/smoothstep.c \
+	$(srcroot)test/unit/spin.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/witness.c \
+	$(srcroot)test/unit/zero.c
+ifeq (@enable_prof@, 1)
+TESTS_UNIT += \
+	$(srcroot)test/unit/arena_reset_prof.c
+endif
+TESTS_INTEGRATION := $(srcroot)test/integration/aligned_alloc.c \
+	$(srcroot)test/integration/allocated.c \
+	$(srcroot)test/integration/extent.c \
+	$(srcroot)test/integration/mallocx.c \
+	$(srcroot)test/integration/MALLOCX_ARENA.c \
+	$(srcroot)test/integration/overflow.c \
+	$(srcroot)test/integration/posix_memalign.c \
+	$(srcroot)test/integration/rallocx.c \
+	$(srcroot)test/integration/sdallocx.c \
+	$(srcroot)test/integration/thread_arena.c \
+	$(srcroot)test/integration/thread_tcache_enabled.c \
+	$(srcroot)test/integration/xallocx.c
+ifeq (@enable_cxx@, 1)
+CPP_SRCS := $(srcroot)src/jemalloc_cpp.cpp
+TESTS_INTEGRATION_CPP := $(srcroot)test/integration/cpp/basic.cpp
+else
+CPP_SRCS :=
+TESTS_INTEGRATION_CPP :=
+endif
+TESTS_STRESS := $(srcroot)test/stress/microbench.c
+
+TESTS := $(TESTS_UNIT) $(TESTS_INTEGRATION) $(TESTS_INTEGRATION_CPP) $(TESTS_STRESS)
+
+PRIVATE_NAMESPACE_HDRS := $(objroot)include/jemalloc/internal/private_namespace.h $(objroot)include/jemalloc/internal/private_namespace_jet.h
+PRIVATE_NAMESPACE_GEN_HDRS := $(PRIVATE_NAMESPACE_HDRS:%.h=%.gen.h)
+C_SYM_OBJS := $(C_SRCS:$(srcroot)%.c=$(objroot)%.sym.$(O))
+C_SYMS := $(C_SRCS:$(srcroot)%.c=$(objroot)%.sym)
+C_OBJS := $(C_SRCS:$(srcroot)%.c=$(objroot)%.$(O))
+CPP_OBJS := $(CPP_SRCS:$(srcroot)%.cpp=$(objroot)%.$(O))
+C_PIC_OBJS := $(C_SRCS:$(srcroot)%.c=$(objroot)%.pic.$(O))
+CPP_PIC_OBJS := $(CPP_SRCS:$(srcroot)%.cpp=$(objroot)%.pic.$(O))
+C_JET_SYM_OBJS := $(C_SRCS:$(srcroot)%.c=$(objroot)%.jet.sym.$(O))
+C_JET_SYMS := $(C_SRCS:$(srcroot)%.c=$(objroot)%.jet.sym)
+C_JET_OBJS := $(C_SRCS:$(srcroot)%.c=$(objroot)%.jet.$(O))
+C_TESTLIB_UNIT_OBJS := $(C_TESTLIB_SRCS:$(srcroot)%.c=$(objroot)%.unit.$(O))
+C_TESTLIB_INTEGRATION_OBJS := $(C_TESTLIB_SRCS:$(srcroot)%.c=$(objroot)%.integration.$(O))
+C_UTIL_INTEGRATION_OBJS := $(C_UTIL_INTEGRATION_SRCS:$(srcroot)%.c=$(objroot)%.integration.$(O))
+C_TESTLIB_STRESS_OBJS := $(C_TESTLIB_SRCS:$(srcroot)%.c=$(objroot)%.stress.$(O))
+C_TESTLIB_OBJS := $(C_TESTLIB_UNIT_OBJS) $(C_TESTLIB_INTEGRATION_OBJS) $(C_UTIL_INTEGRATION_OBJS) $(C_TESTLIB_STRESS_OBJS)
+
+TESTS_UNIT_OBJS := $(TESTS_UNIT:$(srcroot)%.c=$(objroot)%.$(O))
+TESTS_INTEGRATION_OBJS := $(TESTS_INTEGRATION:$(srcroot)%.c=$(objroot)%.$(O))
+TESTS_INTEGRATION_CPP_OBJS := $(TESTS_INTEGRATION_CPP:$(srcroot)%.cpp=$(objroot)%.$(O))
+TESTS_STRESS_OBJS := $(TESTS_STRESS:$(srcroot)%.c=$(objroot)%.$(O))
+TESTS_OBJS := $(TESTS_UNIT_OBJS) $(TESTS_INTEGRATION_OBJS) $(TESTS_STRESS_OBJS)
+TESTS_CPP_OBJS := $(TESTS_INTEGRATION_CPP_OBJS)
+
+.PHONY: all dist build_doc_html build_doc_man build_doc
+.PHONY: install_bin install_include install_lib
+.PHONY: install_doc_html install_doc_man install_doc install
+.PHONY: tests check clean distclean relclean
+
+.SECONDARY : $(PRIVATE_NAMESPACE_GEN_HDRS) $(TESTS_OBJS) $(TESTS_CPP_OBJS)
+
+# Default target.
+all: build_lib
+
+dist: build_doc
+
+$(objroot)doc/%.html : $(objroot)doc/%.xml $(srcroot)doc/stylesheet.xsl $(objroot)doc/html.xsl
+	$(XSLTPROC) -o $@ $(objroot)doc/html.xsl $<
+
+$(objroot)doc/%.3 : $(objroot)doc/%.xml $(srcroot)doc/stylesheet.xsl $(objroot)doc/manpages.xsl
+	$(XSLTPROC) -o $@ $(objroot)doc/manpages.xsl $<
+
+build_doc_html: $(DOCS_HTML)
+build_doc_man: $(DOCS_MAN3)
+build_doc: $(DOCS)
+
+#
+# Include generated dependency files.
+#
+ifdef CC_MM
+-include $(C_SYM_OBJS:%.$(O)=%.d)
+-include $(C_OBJS:%.$(O)=%.d)
+-include $(CPP_OBJS:%.$(O)=%.d)
+-include $(C_PIC_OBJS:%.$(O)=%.d)
+-include $(CPP_PIC_OBJS:%.$(O)=%.d)
+-include $(C_JET_SYM_OBJS:%.$(O)=%.d)
+-include $(C_JET_OBJS:%.$(O)=%.d)
+-include $(C_TESTLIB_OBJS:%.$(O)=%.d)
+-include $(TESTS_OBJS:%.$(O)=%.d)
+-include $(TESTS_CPP_OBJS:%.$(O)=%.d)
+endif
+
+$(C_SYM_OBJS): $(objroot)src/%.sym.$(O): $(srcroot)src/%.c
+$(C_SYM_OBJS): CPPFLAGS += -DJEMALLOC_NO_PRIVATE_NAMESPACE
+$(C_SYMS): $(objroot)src/%.sym: $(objroot)src/%.sym.$(O)
+$(C_OBJS): $(objroot)src/%.$(O): $(srcroot)src/%.c
+$(CPP_OBJS): $(objroot)src/%.$(O): $(srcroot)src/%.cpp
+$(C_PIC_OBJS): $(objroot)src/%.pic.$(O): $(srcroot)src/%.c
+$(C_PIC_OBJS): CFLAGS += $(PIC_CFLAGS)
+$(CPP_PIC_OBJS): $(objroot)src/%.pic.$(O): $(srcroot)src/%.cpp
+$(CPP_PIC_OBJS): CXXFLAGS += $(PIC_CFLAGS)
+$(C_JET_SYM_OBJS): $(objroot)src/%.jet.sym.$(O): $(srcroot)src/%.c
+$(C_JET_SYM_OBJS): CPPFLAGS += -DJEMALLOC_JET -DJEMALLOC_NO_PRIVATE_NAMESPACE
+$(C_JET_SYMS): $(objroot)src/%.jet.sym: $(objroot)src/%.jet.sym.$(O)
+$(C_JET_OBJS): $(objroot)src/%.jet.$(O): $(srcroot)src/%.c
+$(C_JET_OBJS): CPPFLAGS += -DJEMALLOC_JET
+$(C_TESTLIB_UNIT_OBJS): $(objroot)test/src/%.unit.$(O): $(srcroot)test/src/%.c
+$(C_TESTLIB_UNIT_OBJS): CPPFLAGS += -DJEMALLOC_UNIT_TEST
+$(C_TESTLIB_INTEGRATION_OBJS): $(objroot)test/src/%.integration.$(O): $(srcroot)test/src/%.c
+$(C_TESTLIB_INTEGRATION_OBJS): CPPFLAGS += -DJEMALLOC_INTEGRATION_TEST
+$(C_UTIL_INTEGRATION_OBJS): $(objroot)src/%.integration.$(O): $(srcroot)src/%.c
+$(C_TESTLIB_STRESS_OBJS): $(objroot)test/src/%.stress.$(O): $(srcroot)test/src/%.c
+$(C_TESTLIB_STRESS_OBJS): CPPFLAGS += -DJEMALLOC_STRESS_TEST -DJEMALLOC_STRESS_TESTLIB
+$(C_TESTLIB_OBJS): CPPFLAGS += -I$(srcroot)test/include -I$(objroot)test/include
+$(TESTS_UNIT_OBJS): CPPFLAGS += -DJEMALLOC_UNIT_TEST
+$(TESTS_INTEGRATION_OBJS): CPPFLAGS += -DJEMALLOC_INTEGRATION_TEST
+$(TESTS_INTEGRATION_CPP_OBJS): CPPFLAGS += -DJEMALLOC_INTEGRATION_CPP_TEST
+$(TESTS_STRESS_OBJS): CPPFLAGS += -DJEMALLOC_STRESS_TEST
+$(TESTS_OBJS): $(objroot)test/%.$(O): $(srcroot)test/%.c
+$(TESTS_CPP_OBJS): $(objroot)test/%.$(O): $(srcroot)test/%.cpp
+$(TESTS_OBJS): CPPFLAGS += -I$(srcroot)test/include -I$(objroot)test/include
+$(TESTS_CPP_OBJS): CPPFLAGS += -I$(srcroot)test/include -I$(objroot)test/include
+ifneq ($(IMPORTLIB),$(SO))
+$(CPP_OBJS) $(C_SYM_OBJS) $(C_OBJS) $(C_JET_SYM_OBJS) $(C_JET_OBJS): CPPFLAGS += -DDLLEXPORT
+endif
+
+# Dependencies.
+ifndef CC_MM
+HEADER_DIRS = $(srcroot)include/jemalloc/internal \
+	$(objroot)include/jemalloc $(objroot)include/jemalloc/internal
+HEADERS = $(filter-out $(PRIVATE_NAMESPACE_HDRS),$(wildcard $(foreach dir,$(HEADER_DIRS),$(dir)/*.h)))
+$(C_SYM_OBJS) $(C_OBJS) $(CPP_OBJS) $(C_PIC_OBJS) $(CPP_PIC_OBJS) $(C_JET_SYM_OBJS) $(C_JET_OBJS) $(C_TESTLIB_OBJS) $(TESTS_OBJS) $(TESTS_CPP_OBJS): $(HEADERS)
+$(TESTS_OBJS) $(TESTS_CPP_OBJS): $(objroot)test/include/test/jemalloc_test.h
+endif
+
+$(C_OBJS) $(CPP_OBJS) $(C_PIC_OBJS) $(CPP_PIC_OBJS) $(C_TESTLIB_INTEGRATION_OBJS) $(C_UTIL_INTEGRATION_OBJS) $(TESTS_INTEGRATION_OBJS) $(TESTS_INTEGRATION_CPP_OBJS): $(objroot)include/jemalloc/internal/private_namespace.h
+$(C_JET_OBJS) $(C_TESTLIB_UNIT_OBJS) $(C_TESTLIB_STRESS_OBJS) $(TESTS_UNIT_OBJS) $(TESTS_STRESS_OBJS): $(objroot)include/jemalloc/internal/private_namespace_jet.h
+
+$(C_SYM_OBJS) $(C_OBJS) $(C_PIC_OBJS) $(C_JET_SYM_OBJS) $(C_JET_OBJS) $(C_TESTLIB_OBJS) $(TESTS_OBJS): %.$(O):
+	@mkdir -p $(@D)
+	$(CC) $(CFLAGS) -c $(CPPFLAGS) $(CTARGET) $<
+ifdef CC_MM
+	@$(CC) -MM $(CPPFLAGS) -MT $@ -o $(@:%.$(O)=%.d) $<
+endif
+
+$(C_SYMS): %.sym:
+	@mkdir -p $(@D)
+	$(DUMP_SYMS) $< | $(AWK) -f $(objroot)include/jemalloc/internal/private_symbols.awk > $@
+
+$(C_JET_SYMS): %.sym:
+	@mkdir -p $(@D)
+	$(DUMP_SYMS) $< | $(AWK) -f $(objroot)include/jemalloc/internal/private_symbols_jet.awk > $@
+
+$(objroot)include/jemalloc/internal/private_namespace.gen.h: $(C_SYMS)
+	$(SHELL) $(srcroot)include/jemalloc/internal/private_namespace.sh $^ > $@
+
+$(objroot)include/jemalloc/internal/private_namespace_jet.gen.h: $(C_JET_SYMS)
+	$(SHELL) $(srcroot)include/jemalloc/internal/private_namespace.sh $^ > $@
+
+%.h: %.gen.h
+	@if ! `cmp -s $< $@` ; then echo "cp $< $<"; cp $< $@ ; fi
+
+$(CPP_OBJS) $(CPP_PIC_OBJS) $(TESTS_CPP_OBJS): %.$(O):
+	@mkdir -p $(@D)
+	$(CXX) $(CXXFLAGS) -c $(CPPFLAGS) $(CTARGET) $<
+ifdef CC_MM
+	@$(CXX) -MM $(CPPFLAGS) -MT $@ -o $(@:%.$(O)=%.d) $<
+endif
+
+ifneq ($(SOREV),$(SO))
+%.$(SO) : %.$(SOREV)
+	@mkdir -p $(@D)
+	ln -sf $(<F) $@
+endif
+
+$(objroot)lib/$(LIBJEMALLOC).$(SOREV) : $(if $(PIC_CFLAGS),$(C_PIC_OBJS),$(C_OBJS)) $(if $(PIC_CFLAGS),$(CPP_PIC_OBJS),$(CPP_OBJS))
+	@mkdir -p $(@D)
+	$(CC) $(DSO_LDFLAGS) $(call RPATH,$(RPATH_EXTRA)) $(LDTARGET) $+ $(LDFLAGS) $(LIBS) $(EXTRA_LDFLAGS)
+
+$(objroot)lib/$(LIBJEMALLOC)_pic.$(A) : $(C_PIC_OBJS) $(CPP_PIC_OBJS)
+$(objroot)lib/$(LIBJEMALLOC).$(A) : $(C_OBJS) $(CPP_OBJS)
+$(objroot)lib/$(LIBJEMALLOC)_s.$(A) : $(C_OBJS) $(CPP_OBJS)
+
+$(STATIC_LIBS):
+	@mkdir -p $(@D)
+	$(AR) $(ARFLAGS)@AROUT@ $+
+
+$(objroot)test/unit/%$(EXE): $(objroot)test/unit/%.$(O) $(C_JET_OBJS) $(C_TESTLIB_UNIT_OBJS)
+	@mkdir -p $(@D)
+	$(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) $(TEST_LD_MODE) $(LDTARGET) $(filter %.$(O),$^) $(call RPATH,$(objroot)lib) $(LJEMALLOC) $(LDFLAGS) $(filter-out -lm,$(filter -lrt -lpthread -lstdc++,$(LIBS))) $(LM) $(EXTRA_LDFLAGS)
+
+$(objroot)test/integration/cpp/%$(EXE): $(objroot)test/integration/cpp/%.$(O) $(C_TESTLIB_INTEGRATION_OBJS) $(C_UTIL_INTEGRATION_OBJS) $(objroot)lib/$(LIBJEMALLOC).$(IMPORTLIB)
+	@mkdir -p $(@D)
+	$(CXX) $(LDTARGET) $(filter %.$(O),$^) $(call RPATH,$(objroot)lib) $(objroot)lib/$(LIBJEMALLOC).$(IMPORTLIB) $(LDFLAGS) $(filter-out -lm,$(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) $(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)
+	@for b in $(BINS); do \
+	echo "$(INSTALL) -m 755 $$b $(BINDIR)"; \
+	$(INSTALL) -m 755 $$b $(BINDIR); \
+done
+
+install_include:
+	$(INSTALL) -d $(INCLUDEDIR)/jemalloc
+	@for h in $(C_HDRS); do \
+	echo "$(INSTALL) -m 644 $$h $(INCLUDEDIR)/jemalloc"; \
+	$(INSTALL) -m 644 $$h $(INCLUDEDIR)/jemalloc; \
+done
+
+install_lib_shared: $(DSOS)
+	$(INSTALL) -d $(LIBDIR)
+	$(INSTALL) -m 755 $(objroot)lib/$(LIBJEMALLOC).$(SOREV) $(LIBDIR)
+ifneq ($(SOREV),$(SO))
+	ln -sf $(LIBJEMALLOC).$(SOREV) $(LIBDIR)/$(LIBJEMALLOC).$(SO)
+endif
+
+install_lib_static: $(STATIC_LIBS)
+	$(INSTALL) -d $(LIBDIR)
+	@for l in $(STATIC_LIBS); do \
+	echo "$(INSTALL) -m 755 $$l $(LIBDIR)"; \
+	$(INSTALL) -m 755 $$l $(LIBDIR); \
+done
+
+install_lib_pc: $(PC)
+	$(INSTALL) -d $(LIBDIR)/pkgconfig
+	@for l in $(PC); do \
+	echo "$(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)
+	@for d in $(DOCS_HTML); do \
+	echo "$(INSTALL) -m 644 $$d $(DATADIR)/doc/jemalloc$(install_suffix)"; \
+	$(INSTALL) -m 644 $$d $(DATADIR)/doc/jemalloc$(install_suffix); \
+done
+
+install_doc_man:
+	$(INSTALL) -d $(MANDIR)/man3
+	@for d in $(DOCS_MAN3); do \
+	echo "$(INSTALL) -m 644 $$d $(MANDIR)/man3"; \
+	$(INSTALL) -m 644 $$d $(MANDIR)/man3; \
+done
+
+install_doc: install_doc_html install_doc_man
+
+install: install_bin install_include install_lib install_doc
+
+tests_unit: $(TESTS_UNIT:$(srcroot)%.c=$(objroot)%$(EXE))
+tests_integration: $(TESTS_INTEGRATION:$(srcroot)%.c=$(objroot)%$(EXE)) $(TESTS_INTEGRATION_CPP:$(srcroot)%.cpp=$(objroot)%$(EXE))
+tests_stress: $(TESTS_STRESS:$(srcroot)%.c=$(objroot)%$(EXE))
+tests: tests_unit tests_integration tests_stress
+
+check_unit_dir:
+	@mkdir -p $(objroot)test/unit
+check_integration_dir:
+	@mkdir -p $(objroot)test/integration
+stress_dir:
+	@mkdir -p $(objroot)test/stress
+check_dir: check_unit_dir check_integration_dir
+
+check_unit: tests_unit check_unit_dir
+	$(SHELL) $(objroot)test/test.sh $(TESTS_UNIT:$(srcroot)%.c=$(objroot)%)
+check_integration_prof: tests_integration check_integration_dir
+ifeq ($(enable_prof), 1)
+	$(MALLOC_CONF)="prof:true" $(SHELL) $(objroot)test/test.sh $(TESTS_INTEGRATION:$(srcroot)%.c=$(objroot)%) $(TESTS_INTEGRATION_CPP:$(srcroot)%.cpp=$(objroot)%)
+	$(MALLOC_CONF)="prof:true,prof_active:false" $(SHELL) $(objroot)test/test.sh $(TESTS_INTEGRATION:$(srcroot)%.c=$(objroot)%) $(TESTS_INTEGRATION_CPP:$(srcroot)%.cpp=$(objroot)%)
+endif
+check_integration_decay: tests_integration check_integration_dir
+	$(MALLOC_CONF)="dirty_decay_ms:-1,muzzy_decay_ms:-1" $(SHELL) $(objroot)test/test.sh $(TESTS_INTEGRATION:$(srcroot)%.c=$(objroot)%) $(TESTS_INTEGRATION_CPP:$(srcroot)%.cpp=$(objroot)%)
+	$(MALLOC_CONF)="dirty_decay_ms:0,muzzy_decay_ms:0" $(SHELL) $(objroot)test/test.sh $(TESTS_INTEGRATION:$(srcroot)%.c=$(objroot)%) $(TESTS_INTEGRATION_CPP:$(srcroot)%.cpp=$(objroot)%)
+check_integration: tests_integration check_integration_dir
+	$(SHELL) $(objroot)test/test.sh $(TESTS_INTEGRATION:$(srcroot)%.c=$(objroot)%) $(TESTS_INTEGRATION_CPP:$(srcroot)%.cpp=$(objroot)%)
+stress: tests_stress stress_dir
+	$(SHELL) $(objroot)test/test.sh $(TESTS_STRESS:$(srcroot)%.c=$(objroot)%)
+check: check_unit check_integration check_integration_decay check_integration_prof
+
+clean:
+	rm -f $(PRIVATE_NAMESPACE_HDRS)
+	rm -f $(PRIVATE_NAMESPACE_GEN_HDRS)
+	rm -f $(C_SYM_OBJS)
+	rm -f $(C_SYMS)
+	rm -f $(C_OBJS)
+	rm -f $(CPP_OBJS)
+	rm -f $(C_PIC_OBJS)
+	rm -f $(CPP_PIC_OBJS)
+	rm -f $(C_JET_SYM_OBJS)
+	rm -f $(C_JET_SYMS)
+	rm -f $(C_JET_OBJS)
+	rm -f $(C_TESTLIB_OBJS)
+	rm -f $(C_SYM_OBJS:%.$(O)=%.d)
+	rm -f $(C_OBJS:%.$(O)=%.d)
+	rm -f $(CPP_OBJS:%.$(O)=%.d)
+	rm -f $(C_PIC_OBJS:%.$(O)=%.d)
+	rm -f $(CPP_PIC_OBJS:%.$(O)=%.d)
+	rm -f $(C_JET_SYM_OBJS:%.$(O)=%.d)
+	rm -f $(C_JET_OBJS:%.$(O)=%.d)
+	rm -f $(C_TESTLIB_OBJS:%.$(O)=%.d)
+	rm -f $(TESTS_OBJS:%.$(O)=%$(EXE))
+	rm -f $(TESTS_OBJS)
+	rm -f $(TESTS_OBJS:%.$(O)=%.d)
+	rm -f $(TESTS_OBJS:%.$(O)=%.out)
+	rm -f $(TESTS_CPP_OBJS:%.$(O)=%$(EXE))
+	rm -f $(TESTS_CPP_OBJS)
+	rm -f $(TESTS_CPP_OBJS:%.$(O)=%.d)
+	rm -f $(TESTS_CPP_OBJS:%.$(O)=%.out)
+	rm -f $(DSOS) $(STATIC_LIBS)
+
+distclean: clean
+	rm -f $(objroot)bin/jemalloc-config
+	rm -f $(objroot)bin/jemalloc.sh
+	rm -f $(objroot)bin/jeprof
+	rm -f $(objroot)config.log
+	rm -f $(objroot)config.status
+	rm -f $(objroot)config.stamp
+	rm -f $(cfghdrs_out)
+	rm -f $(cfgoutputs_out)
+
+relclean: distclean
+	rm -f $(objroot)configure
+	rm -f $(objroot)VERSION
+	rm -f $(DOCS_HTML)
+	rm -f $(DOCS_MAN3)
+
+#===============================================================================
+# Re-configuration rules.
+
+ifeq ($(enable_autogen), 1)
+$(srcroot)configure : $(srcroot)configure.ac
+	cd ./$(srcroot) && $(AUTOCONF)
+
+$(objroot)config.status : $(srcroot)configure
+	./$(objroot)config.status --recheck
+
+$(srcroot)config.stamp.in : $(srcroot)configure.ac
+	echo stamp > $(srcroot)config.stamp.in
+
+$(objroot)config.stamp : $(cfgoutputs_in) $(cfghdrs_in) $(srcroot)configure
+	./$(objroot)config.status
+	@touch $@
+
+# There must be some action in order for make to re-read Makefile when it is
+# out of date.
+$(cfgoutputs_out) $(cfghdrs_out) : $(objroot)config.stamp
+	@true
+endif

README

@@ -0,0 +1,20 @@
+jemalloc is a general purpose malloc(3) implementation that emphasizes
+fragmentation avoidance and scalable concurrency support.  jemalloc first came
+into use as the FreeBSD libc allocator in 2005, and since then it has found its
+way into numerous applications that rely on its predictable behavior.  In 2010
+jemalloc development efforts broadened to include developer support features
+such as heap profiling and extensive monitoring/tuning hooks.  Modern jemalloc
+releases continue to be integrated back into FreeBSD, and therefore versatility
+remains critical.  Ongoing development efforts trend toward making jemalloc
+among the best allocators for a broad range of demanding applications, and
+eliminating/mitigating weaknesses that have practical repercussions for real
+world applications.
+
+The COPYING file contains copyright and licensing information.
+
+The INSTALL file contains information on how to configure, build, and install
+jemalloc.
+
+The ChangeLog file contains a brief summary of changes for each release.
+
+URL: http://jemalloc.net/

bin/jemalloc-config.in

@@ -0,0 +1,83 @@
+#!/bin/sh
+
+usage() {
+	cat <<EOF
+Usage:
+  @BINDIR@/jemalloc-config <option>
+Options:
+  --help | -h  : Print usage.
+  --version    : Print jemalloc version.
+  --revision   : Print shared library revision number.
+  --config     : Print configure options used to build jemalloc.
+  --prefix     : Print installation directory prefix.
+  --bindir     : Print binary installation directory.
+  --datadir    : Print data installation directory.
+  --includedir : Print include installation directory.
+  --libdir     : Print library installation directory.
+  --mandir     : Print manual page installation directory.
+  --cc         : Print compiler used to build jemalloc.
+  --cflags     : Print compiler flags used to build jemalloc.
+  --cppflags   : Print preprocessor flags used to build jemalloc.
+  --cxxflags   : Print C++ compiler flags used to build jemalloc.
+  --ldflags    : Print library flags used to build jemalloc.
+  --libs       : Print libraries jemalloc was linked against.
+EOF
+}
+
+prefix="@prefix@"
+exec_prefix="@exec_prefix@"
+
+case "$1" in
+--help | -h)
+	usage
+	exit 0
+	;;
+--version)
+	echo "@jemalloc_version@"
+	;;
+--revision)
+	echo "@rev@"
+	;;
+--config)
+	echo "@CONFIG@"
+	;;
+--prefix)
+	echo "@PREFIX@"
+	;;
+--bindir)
+	echo "@BINDIR@"
+	;;
+--datadir)
+	echo "@DATADIR@"
+	;;
+--includedir)
+	echo "@INCLUDEDIR@"
+	;;
+--libdir)
+	echo "@LIBDIR@"
+	;;
+--mandir)
+	echo "@MANDIR@"
+	;;
+--cc)
+	echo "@CC@"
+	;;
+--cflags)
+	echo "@CFLAGS@"
+	;;
+--cppflags)
+	echo "@CPPFLAGS@"
+	;;
+--cxxflags)
+	echo "@CXXFLAGS@"
+	;;
+--ldflags)
+	echo "@LDFLAGS@ @EXTRA_LDFLAGS@"
+	;;
+--libs)
+	echo "@LIBS@"
+	;;
+*)
+	usage
+	exit 1
+esac

bin/jemalloc.sh.in

@@ -0,0 +1,9 @@
+#!/bin/sh
+
+prefix=@prefix@
+exec_prefix=@exec_prefix@
+libdir=@libdir@
+
+@LD_PRELOAD_VAR@=${libdir}/libjemalloc.@SOREV@
+export @LD_PRELOAD_VAR@
+exec "$@"

build-aux/config.sub

@@ -1,42 +1,40 @@
 #! /bin/sh
 # Configuration validation subroutine script.
-#   Copyright (C) 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
-#   2000, 2001, 2002, 2003 Free Software Foundation, Inc.
+#   Copyright 1992-2016 Free Software Foundation, Inc.
 
-timestamp='2004-02-23'
+timestamp='2016-11-04'
 
-# This file is (in principle) common to ALL GNU software.
-# The presence of a machine in this file suggests that SOME GNU software
-# can handle that machine.  It does not imply ALL GNU software can.
-#
-# 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
-# the Free Software Foundation; either version 2 of the License, or
+# 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
+# the Free Software Foundation; either version 3 of the License, or
 # (at your option) any later version.
 #
-# This program is distributed in the hope that it will be useful,
-# but WITHOUT ANY WARRANTY; without even the implied warranty of
-# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-# GNU General Public License for more details.
+# This program is distributed in the hope that it will be useful, but
+# WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+# General Public License for more details.
 #
 # You should have received a copy of the GNU General Public License
-# along with this program; if not, write to the Free Software
-# Foundation, Inc., 59 Temple Place - Suite 330,
-# Boston, MA 02111-1307, USA.
-
+# along with this program; if not, see <http://www.gnu.org/licenses/>.
+#
 # As a special exception to the GNU General Public License, if you
 # distribute this file as part of a program that contains a
 # configuration script generated by Autoconf, you may include it under
-# the same distribution terms that you use for the rest of that program.
+# the same distribution terms that you use for the rest of that
+# program.  This Exception is an additional permission under section 7
+# of the GNU General Public License, version 3 ("GPLv3").
 
-# Please send patches to <config-patches@gnu.org>.  Submit a context
-# diff and a properly formatted ChangeLog entry.
+
+# 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.
 # If it is invalid, we print an error message on stderr and exit with code 1.
 # Otherwise, we print the canonical config type on stdout and succeed.
 
+# You can get the latest version of this script from:
+# http://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.sub
+
 # This file is supposed to be the same for all GNU packages
 # and recognize all the CPU types, system types and aliases
 # that are meaningful with *any* GNU software.
@@ -55,8 +53,7 @@ timestamp='2004-02-23'
 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.
 
@@ -70,8 +67,7 @@ Report bugs and patches to <config-patches@gnu.org>."
 version="\
 GNU config.sub ($timestamp)
 
-Copyright (C) 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001
-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."
@@ -83,11 +79,11 @@ Try \`$me --help' for more information."
 while test $# -gt 0 ; do
   case $1 in
     --time-stamp | --time* | -t )
-       echo "$timestamp" ; exit 0 ;;
+       echo "$timestamp" ; exit ;;
     --version | -v )
-       echo "$version" ; exit 0 ;;
+       echo "$version" ; exit ;;
     --help | --h* | -h )
-       echo "$usage"; exit 0 ;;
+       echo "$usage"; exit ;;
     -- )     # Stop option processing
        shift; break ;;
     - )	# Use stdin as input.
@@ -99,7 +95,7 @@ while test $# -gt 0 ; do
     *local*)
        # First pass through any local machine types.
        echo $1
-       exit 0;;
+       exit ;;
 
     * )
        break ;;
@@ -118,11 +114,18 @@ esac
 # Here we must recognize all the valid KERNEL-OS combinations.
 maybe_os=`echo $1 | sed 's/^\(.*\)-\([^-]*-[^-]*\)$/\2/'`
 case $maybe_os in
-  nto-qnx* | linux-gnu* | linux-dietlibc | linux-uclibc* | uclinux-uclibc* | uclinux-gnu* | \
-  kfreebsd*-gnu* | knetbsd*-gnu* | netbsd*-gnu* | storm-chaos* | os2-emx* | rtmk-nova*)
+  nto-qnx* | linux-gnu* | linux-android* | linux-dietlibc | linux-newlib* | \
+  linux-musl* | linux-uclibc* | uclinux-uclibc* | uclinux-gnu* | kfreebsd*-gnu* | \
+  knetbsd*-gnu* | netbsd*-gnu* | netbsd*-eabi* | \
+  kopensolaris*-gnu* | cloudabi*-eabi* | \
+  storm-chaos* | os2-emx* | rtmk-nova*)
     os=-$maybe_os
     basic_machine=`echo $1 | sed 's/^\(.*\)-\([^-]*-[^-]*\)$/\1/'`
     ;;
+  android-linux)
+    os=-linux-android
+    basic_machine=`echo $1 | sed 's/^\(.*\)-\([^-]*-[^-]*\)$/\1/'`-unknown
+    ;;
   *)
     basic_machine=`echo $1 | sed 's/-[^-]*$//'`
     if [ $basic_machine != $1 ]
@@ -145,10 +148,13 @@ case $os in
 	-convergent* | -ncr* | -news | -32* | -3600* | -3100* | -hitachi* |\
 	-c[123]* | -convex* | -sun | -crds | -omron* | -dg | -ultra | -tti* | \
 	-harris | -dolphin | -highlevel | -gould | -cbm | -ns | -masscomp | \
-	-apple | -axis)
+	-apple | -axis | -knuth | -cray | -microblaze*)
 		os=
 		basic_machine=$1
 		;;
+	-bluegene*)
+		os=-cnk
+		;;
 	-sim | -cisco | -oki | -wec | -winbond)
 		os=
 		basic_machine=$1
@@ -163,13 +169,17 @@ case $os in
 		os=-chorusos
 		basic_machine=$1
 		;;
- 	-chorusrdb)
- 		os=-chorusrdb
+	-chorusrdb)
+		os=-chorusrdb
 		basic_machine=$1
- 		;;
+		;;
 	-hiux*)
 		os=-hiuxwe2
 		;;
+	-sco6)
+		os=-sco5v6
+		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
+		;;
 	-sco5)
 		os=-sco3.2v5
 		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
@@ -186,6 +196,10 @@ case $os in
 		# Don't forget version if it is 3.2v4 or newer.
 		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
 		;;
+	-sco5v6*)
+		# Don't forget version if it is 3.2v4 or newer.
+		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
+		;;
 	-sco*)
 		os=-sco3.2v2
 		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
@@ -203,6 +217,12 @@ case $os in
 	-isc*)
 		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
 		;;
+	-lynx*178)
+		os=-lynxos178
+		;;
+	-lynx*5)
+		os=-lynxos5
+		;;
 	-lynx*)
 		os=-lynxos
 		;;
@@ -227,57 +247,115 @@ case $basic_machine in
 	# Some are omitted here because they have special meanings below.
 	1750a | 580 \
 	| a29k \
+	| aarch64 | aarch64_be \
 	| alpha | alphaev[4-8] | alphaev56 | alphaev6[78] | alphapca5[67] \
 	| alpha64 | alpha64ev[4-8] | alpha64ev56 | alpha64ev6[78] | alpha64pca5[67] \
 	| am33_2.0 \
-	| arc | arm | arm[bl]e | arme[lb] | armv[2345] | armv[345][lb] | avr \
-	| c4x | clipper \
+	| 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 \
-	| fr30 | frv \
+	| e2k | epiphany \
+	| fido | fr30 | frv | ft32 \
 	| h8300 | h8500 | hppa | hppa1.[01] | hppa2.0 | hppa2.0[nw] | hppa64 \
+	| hexagon \
 	| i370 | i860 | i960 | ia64 \
 	| ip2k | iq2000 \
-	| m32r | m68000 | m68k | m88k | mcore \
+	| k1om \
+	| le32 | le64 \
+	| lm32 \
+	| m32c | m32r | m32rle | m68000 | m68k | m88k \
+	| maxq | mb | microblaze | microblazeel | mcore | mep | metag \
 	| mips | mipsbe | mipseb | mipsel | mipsle \
 	| mips16 \
 	| mips64 | mips64el \
-	| mips64vr | mips64vrel \
+	| mips64octeon | mips64octeonel \
 	| mips64orion | mips64orionel \
+	| mips64r5900 | mips64r5900el \
+	| mips64vr | mips64vrel \
 	| mips64vr4100 | mips64vr4100el \
 	| mips64vr4300 | mips64vr4300el \
 	| mips64vr5000 | mips64vr5000el \
+	| mips64vr5900 | mips64vr5900el \
 	| mipsisa32 | mipsisa32el \
 	| mipsisa32r2 | mipsisa32r2el \
+	| mipsisa32r6 | mipsisa32r6el \
 	| mipsisa64 | mipsisa64el \
 	| mipsisa64r2 | mipsisa64r2el \
+	| mipsisa64r6 | mipsisa64r6el \
 	| mipsisa64sb1 | mipsisa64sb1el \
 	| mipsisa64sr71k | mipsisa64sr71kel \
+	| mipsr5900 | mipsr5900el \
 	| mipstx39 | mipstx39el \
 	| mn10200 | mn10300 \
+	| moxie \
+	| mt \
 	| msp430 \
+	| nds32 | nds32le | nds32be \
+	| nios | nios2 | nios2eb | nios2el \
 	| ns16k | ns32k \
-	| openrisc | or32 \
+	| open8 | or1k | or1knd | or32 \
 	| pdp10 | pdp11 | pj | pjl \
-	| powerpc | powerpc64 | powerpc64le | powerpcle | ppcbe \
+	| powerpc | powerpc64 | powerpc64le | powerpcle \
+	| pru \
 	| pyramid \
-	| sh | sh[1234] | sh[23]e | sh[34]eb | shbe | shle | sh[1234]le | sh3ele \
+	| riscv32 | riscv64 \
+	| rl78 | rx \
+	| score \
+	| 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 | sparc86x | sparclet | sparclite | sparcv9 | sparcv9b \
-	| strongarm \
-	| tahoe | thumb | tic4x | tic80 | tron \
-	| v850 | v850e \
+	| sparc | sparc64 | sparc64b | sparc64v | sparc86x | sparclet | sparclite \
+	| sparcv8 | sparcv9 | sparcv9b | sparcv9v \
+	| spu \
+	| tahoe | tic4x | tic54x | tic55x | tic6x | tic80 | tron \
+	| ubicom32 \
+	| v850 | v850e | v850e1 | v850e2 | v850es | v850e2v3 \
+	| visium \
 	| we32k \
-	| x86 | xscale | xstormy16 | xtensa \
-	| z8k)
+	| x86 | xc16x | xstormy16 | xtensa \
+	| z8k | z80)
 		basic_machine=$basic_machine-unknown
 		;;
-	m6811 | m68hc11 | m6812 | m68hc12)
-		# Motorola 68HC11/12.
+	c54x)
+		basic_machine=tic54x-unknown
+		;;
+	c55x)
+		basic_machine=tic55x-unknown
+		;;
+	c6x)
+		basic_machine=tic6x-unknown
+		;;
+	leon|leon[3-9])
+		basic_machine=sparc-$basic_machine
+		;;
+	m6811 | m68hc11 | m6812 | m68hc12 | m68hcs12x | nvptx | picochip)
 		basic_machine=$basic_machine-unknown
 		os=-none
 		;;
 	m88110 | m680[12346]0 | m683?2 | m68360 | m5200 | v70 | w65 | z8k)
 		;;
+	ms1)
+		basic_machine=mt-unknown
+		;;
+
+	strongarm | thumb | xscale)
+		basic_machine=arm-unknown
+		;;
+	xgate)
+		basic_machine=$basic_machine-unknown
+		os=-none
+		;;
+	xscaleeb)
+		basic_machine=armeb-unknown
+		;;
+
+	xscaleel)
+		basic_machine=armel-unknown
+		;;
 
 	# We use `pc' rather than `unknown'
 	# because (1) that's what they normally are, and
@@ -293,59 +371,90 @@ case $basic_machine in
 	# Recognize the basic CPU types with company name.
 	580-* \
 	| a29k-* \
+	| aarch64-* | aarch64_be-* \
 	| alpha-* | alphaev[4-8]-* | alphaev56-* | alphaev6[78]-* \
 	| alpha64-* | alpha64ev[4-8]-* | alpha64ev56-* | alpha64ev6[78]-* \
-	| alphapca5[67]-* | alpha64pca5[67]-* | arc-* \
+	| alphapca5[67]-* | alpha64pca5[67]-* | arc-* | arceb-* \
 	| arm-*  | armbe-* | armle-* | armeb-* | armv*-* \
-	| avr-* \
-	| bs2000-* \
-	| c[123]* | c30-* | [cjt]90-* | c4x-* | c54x-* | c55x-* | c6x-* \
-	| clipper-* | cydra-* \
+	| avr-* | avr32-* \
+	| ba-* \
+	| be32-* | be64-* \
+	| bfin-* | bs2000-* \
+	| c[123]* | c30-* | [cjt]90-* | c4x-* \
+	| c8051-* | clipper-* | craynv-* | cydra-* \
 	| d10v-* | d30v-* | dlx-* \
-	| elxsi-* \
-	| f30[01]-* | f700-* | fr30-* | frv-* | fx80-* \
+	| e2k-* | elxsi-* \
+	| f30[01]-* | f700-* | fido-* | fr30-* | frv-* | fx80-* \
 	| h8300-* | h8500-* \
 	| hppa-* | hppa1.[01]-* | hppa2.0-* | hppa2.0[nw]-* | hppa64-* \
+	| hexagon-* \
 	| i*86-* | i860-* | i960-* | ia64-* \
 	| ip2k-* | iq2000-* \
-	| m32r-* \
+	| k1om-* \
+	| le32-* | le64-* \
+	| lm32-* \
+	| m32c-* | m32r-* | m32rle-* \
 	| m68000-* | m680[012346]0-* | m68360-* | m683?2-* | m68k-* \
-	| m88110-* | m88k-* | mcore-* \
+	| m88110-* | m88k-* | maxq-* | mcore-* | metag-* \
+	| microblaze-* | microblazeel-* \
 	| mips-* | mipsbe-* | mipseb-* | mipsel-* | mipsle-* \
 	| mips16-* \
 	| mips64-* | mips64el-* \
-	| mips64vr-* | mips64vrel-* \
+	| mips64octeon-* | mips64octeonel-* \
 	| mips64orion-* | mips64orionel-* \
+	| mips64r5900-* | mips64r5900el-* \
+	| mips64vr-* | mips64vrel-* \
 	| mips64vr4100-* | mips64vr4100el-* \
 	| mips64vr4300-* | mips64vr4300el-* \
 	| mips64vr5000-* | mips64vr5000el-* \
+	| mips64vr5900-* | mips64vr5900el-* \
 	| mipsisa32-* | mipsisa32el-* \
 	| mipsisa32r2-* | mipsisa32r2el-* \
+	| mipsisa32r6-* | mipsisa32r6el-* \
 	| mipsisa64-* | mipsisa64el-* \
 	| mipsisa64r2-* | mipsisa64r2el-* \
+	| mipsisa64r6-* | mipsisa64r6el-* \
 	| mipsisa64sb1-* | mipsisa64sb1el-* \
 	| mipsisa64sr71k-* | mipsisa64sr71kel-* \
+	| mipsr5900-* | mipsr5900el-* \
 	| mipstx39-* | mipstx39el-* \
+	| mmix-* \
+	| mt-* \
 	| msp430-* \
-	| none-* | np1-* | nv1-* | ns16k-* | ns32k-* \
+	| nds32-* | nds32le-* | nds32be-* \
+	| nios-* | nios2-* | nios2eb-* | nios2el-* \
+	| none-* | np1-* | ns16k-* | ns32k-* \
+	| open8-* \
+	| or1k*-* \
 	| orion-* \
 	| pdp10-* | pdp11-* | pj-* | pjl-* | pn-* | power-* \
-	| powerpc-* | powerpc64-* | powerpc64le-* | powerpcle-* | ppcbe-* \
+	| powerpc-* | powerpc64-* | powerpc64le-* | powerpcle-* \
+	| pru-* \
 	| pyramid-* \
-	| romp-* | rs6000-* \
-	| sh-* | sh[1234]-* | sh[23]e-* | sh[34]eb-* | shbe-* \
+	| 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-* | sparc86x-* | sparclet-* | sparclite-* \
-	| sparcv9-* | sparcv9b-* | strongarm-* | sv1-* | sx?-* \
-	| tahoe-* | thumb-* \
+	| sparc-* | sparc64-* | sparc64b-* | sparc64v-* | sparc86x-* | sparclet-* \
+	| sparclite-* \
+	| sparcv8-* | sparcv9-* | sparcv9b-* | sparcv9v-* | sv1-* | sx*-* \
+	| tahoe-* \
 	| tic30-* | tic4x-* | tic54x-* | tic55x-* | tic6x-* | tic80-* \
+	| tile*-* \
 	| tron-* \
-	| v850-* | v850e-* | vax-* \
+	| ubicom32-* \
+	| v850-* | v850e-* | v850e1-* | v850es-* | v850e2-* | v850e2v3-* \
+	| vax-* \
+	| visium-* \
 	| we32k-* \
-	| x86-* | x86_64-* | xps100-* | xscale-* | xstormy16-* \
-	| xtensa-* \
+	| x86-* | x86_64-* | xc16x-* | xps100-* \
+	| xstormy16-* | xtensa*-* \
 	| ymp-* \
-	| z8k-*)
+	| z8k-* | z80-*)
+		;;
+	# Recognize the basic CPU types without company name, with glob match.
+	xtensa*)
+		basic_machine=$basic_machine-unknown
 		;;
 	# Recognize the various machine names and aliases which stand
 	# for a CPU type and a company and sometimes even an OS.
@@ -363,7 +472,7 @@ case $basic_machine in
 		basic_machine=a29k-amd
 		os=-udi
 		;;
-    	abacus)
+	abacus)
 		basic_machine=abacus-unknown
 		;;
 	adobe68k)
@@ -409,6 +518,13 @@ case $basic_machine in
 		basic_machine=m68k-apollo
 		os=-bsd
 		;;
+	aros)
+		basic_machine=i386-pc
+		os=-aros
+		;;
+	asmjs)
+		basic_machine=asmjs-unknown
+		;;
 	aux)
 		basic_machine=m68k-apple
 		os=-aux
@@ -417,10 +533,35 @@ case $basic_machine in
 		basic_machine=ns32k-sequent
 		os=-dynix
 		;;
+	blackfin)
+		basic_machine=bfin-unknown
+		os=-linux
+		;;
+	blackfin-*)
+		basic_machine=bfin-`echo $basic_machine | sed 's/^[^-]*-//'`
+		os=-linux
+		;;
+	bluegene*)
+		basic_machine=powerpc-ibm
+		os=-cnk
+		;;
+	c54x-*)
+		basic_machine=tic54x-`echo $basic_machine | sed 's/^[^-]*-//'`
+		;;
+	c55x-*)
+		basic_machine=tic55x-`echo $basic_machine | sed 's/^[^-]*-//'`
+		;;
+	c6x-*)
+		basic_machine=tic6x-`echo $basic_machine | sed 's/^[^-]*-//'`
+		;;
 	c90)
 		basic_machine=c90-cray
 		os=-unicos
 		;;
+	cegcc)
+		basic_machine=arm-unknown
+		os=-cegcc
+		;;
 	convex-c1)
 		basic_machine=c1-convex
 		os=-bsd
@@ -445,13 +586,20 @@ case $basic_machine in
 		basic_machine=j90-cray
 		os=-unicos
 		;;
-	cr16c)
-		basic_machine=cr16c-unknown
+	craynv)
+		basic_machine=craynv-cray
+		os=-unicosmp
+		;;
+	cr16 | cr16-*)
+		basic_machine=cr16-unknown
 		os=-elf
 		;;
 	crds | unos)
 		basic_machine=m68k-crds
 		;;
+	crisv32 | crisv32-* | etraxfs*)
+		basic_machine=crisv32-axis
+		;;
 	cris | cris-* | etrax*)
 		basic_machine=cris-axis
 		;;
@@ -481,6 +629,14 @@ case $basic_machine in
 		basic_machine=m88k-motorola
 		os=-sysv3
 		;;
+	dicos)
+		basic_machine=i686-pc
+		os=-dicos
+		;;
+	djgpp)
+		basic_machine=i586-pc
+		os=-msdosdjgpp
+		;;
 	dpx20 | dpx20-*)
 		basic_machine=rs6000-bull
 		os=-bosx
@@ -489,6 +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
@@ -592,7 +756,6 @@ case $basic_machine in
 	i370-ibm* | ibm*)
 		basic_machine=i370-ibm
 		;;
-# I'm not sure what "Sysv32" means.  Should this be sysv3.2?
 	i*86v32)
 		basic_machine=`echo $1 | sed -e 's/86.*/86-pc/'`
 		os=-sysv32
@@ -631,6 +794,17 @@ 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
+		;;
+	m68knommu-*)
+		basic_machine=m68k-`echo $basic_machine | sed 's/^[^-]*-//'`
+		os=-linux
+		;;
 	m88k-omron*)
 		basic_machine=m88k-omron
 		;;
@@ -642,10 +816,21 @@ case $basic_machine in
 		basic_machine=ns32k-utek
 		os=-sysv
 		;;
+	microblaze*)
+		basic_machine=microblaze-xilinx
+		;;
+	mingw64)
+		basic_machine=x86_64-pc
+		os=-mingw64
+		;;
 	mingw32)
-		basic_machine=i386-pc
+		basic_machine=i686-pc
 		os=-mingw32
 		;;
+	mingw32ce)
+		basic_machine=arm-unknown
+		os=-mingw32ce
+		;;
 	miniframe)
 		basic_machine=m68000-convergent
 		;;
@@ -659,10 +844,6 @@ case $basic_machine in
 	mips3*)
 		basic_machine=`echo $basic_machine | sed -e 's/mips3/mips64/'`-unknown
 		;;
-	mmix*)
-		basic_machine=mmix-knuth
-		os=-mmixware
-		;;
 	monitor)
 		basic_machine=m68k-rom68k
 		os=-coff
@@ -671,14 +852,29 @@ case $basic_machine in
 		basic_machine=powerpc-unknown
 		os=-morphos
 		;;
+	moxiebox)
+		basic_machine=moxie-unknown
+		os=-moxiebox
+		;;
 	msdos)
 		basic_machine=i386-pc
 		os=-msdos
 		;;
+	ms1-*)
+		basic_machine=`echo $basic_machine | sed -e 's/ms1-/mt-/'`
+		;;
+	msys)
+		basic_machine=i686-pc
+		os=-msys
+		;;
 	mvs)
 		basic_machine=i370-ibm
 		os=-mvs
 		;;
+	nacl)
+		basic_machine=le32-unknown
+		os=-nacl
+		;;
 	ncr3000)
 		basic_machine=i486-ncr
 		os=-sysv4
@@ -743,9 +939,11 @@ case $basic_machine in
 	np1)
 		basic_machine=np1-gould
 		;;
-	nv1)
-		basic_machine=nv1-cray
-		os=-unicosmp
+	neo-tandem)
+		basic_machine=neo-tandem
+		;;
+	nse-tandem)
+		basic_machine=nse-tandem
 		;;
 	nsr-tandem)
 		basic_machine=nsr-tandem
@@ -754,9 +952,8 @@ case $basic_machine in
 		basic_machine=hppa1.1-oki
 		os=-proelf
 		;;
-	or32 | or32-*)
+	openrisc | openrisc-*)
 		basic_machine=or32-unknown
-		os=-coff
 		;;
 	os400)
 		basic_machine=powerpc-ibm
@@ -778,6 +975,14 @@ case $basic_machine in
 		basic_machine=i860-intel
 		os=-osf
 		;;
+	parisc)
+		basic_machine=hppa-unknown
+		os=-linux
+		;;
+	parisc-*)
+		basic_machine=hppa-`echo $basic_machine | sed 's/^[^-]*-//'`
+		os=-linux
+		;;
 	pbd)
 		basic_machine=sparc-tti
 		;;
@@ -787,6 +992,12 @@ case $basic_machine in
 	pc532 | pc532-*)
 		basic_machine=ns32k-pc532
 		;;
+	pc98)
+		basic_machine=i386-pc
+		;;
+	pc98-*)
+		basic_machine=i386-`echo $basic_machine | sed 's/^[^-]*-//'`
+		;;
 	pentium | p5 | k5 | k6 | nexgen | viac3)
 		basic_machine=i586-pc
 		;;
@@ -816,11 +1027,12 @@ case $basic_machine in
 		;;
 	power)	basic_machine=power-ibm
 		;;
-	ppc)	basic_machine=powerpc-unknown
+	ppc | ppcbe)	basic_machine=powerpc-unknown
 		;;
-	ppc-*)	basic_machine=powerpc-`echo $basic_machine | sed 's/^[^-]*-//'`
+	ppc-* | ppcbe-*)
+		basic_machine=powerpc-`echo $basic_machine | sed 's/^[^-]*-//'`
 		;;
-	ppcle | powerpclittle | ppc-le | powerpc-little)
+	ppcle | powerpclittle)
 		basic_machine=powerpcle-unknown
 		;;
 	ppcle-* | powerpclittle-*)
@@ -830,7 +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-*)
@@ -843,6 +1055,14 @@ case $basic_machine in
 		basic_machine=i586-unknown
 		os=-pw32
 		;;
+	rdos | rdos64)
+		basic_machine=x86_64-pc
+		os=-rdos
+		;;
+	rdos32)
+		basic_machine=i386-pc
+		os=-rdos
+		;;
 	rom68k)
 		basic_machine=m68k-rom68k
 		os=-coff
@@ -869,6 +1089,10 @@ case $basic_machine in
 	sb1el)
 		basic_machine=mipsisa64sb1el-unknown
 		;;
+	sde)
+		basic_machine=mipsisa32-sde
+		os=-elf
+		;;
 	sei)
 		basic_machine=mips-sei
 		os=-seiux
@@ -880,6 +1104,9 @@ case $basic_machine in
 		basic_machine=sh-hitachi
 		os=-hms
 		;;
+	sh5el)
+		basic_machine=sh5le-unknown
+		;;
 	sh64)
 		basic_machine=sh64-unknown
 		;;
@@ -901,6 +1128,9 @@ case $basic_machine in
 		basic_machine=i860-stratus
 		os=-sysv4
 		;;
+	strongarm-* | thumb-*)
+		basic_machine=arm-`echo $basic_machine | sed 's/^[^-]*-//'`
+		;;
 	sun2)
 		basic_machine=m68000-sun
 		;;
@@ -957,17 +1187,9 @@ case $basic_machine in
 		basic_machine=t90-cray
 		os=-unicos
 		;;
-	tic54x | c54x*)
-		basic_machine=tic54x-unknown
-		os=-coff
-		;;
-	tic55x | c55x*)
-		basic_machine=tic55x-unknown
-		os=-coff
-		;;
-	tic6x | c6x*)
-		basic_machine=tic6x-unknown
-		os=-coff
+	tile*)
+		basic_machine=$basic_machine-unknown
+		os=-linux-gnu
 		;;
 	tx39)
 		basic_machine=mipstx39-unknown
@@ -1029,9 +1251,16 @@ case $basic_machine in
 		basic_machine=hppa1.1-winbond
 		os=-proelf
 		;;
+	xbox)
+		basic_machine=i686-pc
+		os=-mingw32
+		;;
 	xps | xps100)
 		basic_machine=xps100-honeywell
 		;;
+	xscale-* | xscalee[bl]-*)
+		basic_machine=`echo $basic_machine | sed 's/^xscale/arm/'`
+		;;
 	ymp)
 		basic_machine=ymp-cray
 		os=-unicos
@@ -1040,6 +1269,10 @@ case $basic_machine in
 		basic_machine=z8k-unknown
 		os=-sim
 		;;
+	z80-*-coff)
+		basic_machine=z80-unknown
+		os=-sim
+		;;
 	none)
 		basic_machine=none-none
 		os=-none
@@ -1059,6 +1292,9 @@ case $basic_machine in
 	romp)
 		basic_machine=romp-ibm
 		;;
+	mmix)
+		basic_machine=mmix-knuth
+		;;
 	rs6000)
 		basic_machine=rs6000-ibm
 		;;
@@ -1075,13 +1311,10 @@ case $basic_machine in
 	we32k)
 		basic_machine=we32k-att
 		;;
-	sh3 | sh4 | sh[34]eb | sh[1234]le | sh[23]ele)
+	sh[1234] | sh[24]a | sh[24]aeb | sh[34]eb | sh[1234]le | sh[23]ele)
 		basic_machine=sh-unknown
 		;;
-	sh64)
-		basic_machine=sh64-unknown
-		;;
-	sparc | sparcv9 | sparcv9b)
+	sparc | sparcv8 | sparcv9 | sparcv9b | sparcv9v)
 		basic_machine=sparc-sun
 		;;
 	cydra)
@@ -1125,9 +1358,12 @@ esac
 if [ x"$os" != x"" ]
 then
 case $os in
-        # First match some system type aliases
-        # that might get confused with valid system types.
+	# First match some system type aliases
+	# that might get confused with valid system types.
 	# -solaris* is a basic system type, with this one exception.
+	-auroraux)
+		os=-auroraux
+		;;
 	-solaris1 | -solaris1.*)
 		os=`echo $os | sed -e 's|solaris1|sunos4|'`
 		;;
@@ -1148,26 +1384,32 @@ case $os in
 	# Each alternative MUST END IN A *, to match a version number.
 	# -sysv* is not here because it comes later, after sysvr4.
 	-gnu* | -bsd* | -mach* | -minix* | -genix* | -ultrix* | -irix* \
-	      | -*vms* | -sco* | -esix* | -isc* | -aix* | -sunos | -sunos[34]*\
-	      | -hpux* | -unos* | -osf* | -luna* | -dgux* | -solaris* | -sym* \
+	      | -*vms* | -sco* | -esix* | -isc* | -aix* | -cnk* | -sunos | -sunos[34]*\
+	      | -hpux* | -unos* | -osf* | -luna* | -dgux* | -auroraux* | -solaris* \
+	      | -sym* | -kopensolaris* | -plan9* \
 	      | -amigaos* | -amigados* | -msdos* | -newsos* | -unicos* | -aof* \
-	      | -aos* \
+	      | -aos* | -aros* | -cloudabi* | -sortix* \
 	      | -nindy* | -vxsim* | -vxworks* | -ebmon* | -hms* | -mvs* \
 	      | -clix* | -riscos* | -uniplus* | -iris* | -rtu* | -xenix* \
-	      | -hiux* | -386bsd* | -knetbsd* | -mirbsd* | -netbsd* | -openbsd* \
+	      | -hiux* | -386bsd* | -knetbsd* | -mirbsd* | -netbsd* \
+	      | -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* \
-	      | -cygwin* | -pe* | -psos* | -moss* | -proelf* | -rtems* \
-	      | -mingw32* | -linux-gnu* | -linux-uclibc* | -uxpv* | -beos* | -mpeix* | -udk* \
+	      | -chorusos* | -chorusrdb* | -cegcc* \
+	      | -cygwin* | -msys* | -pe* | -psos* | -moss* | -proelf* | -rtems* \
+	      | -midipix* | -mingw32* | -mingw64* | -linux-gnu* | -linux-android* \
+	      | -linux-newlib* | -linux-musl* | -linux-uclibc* \
+	      | -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*)
+	      | -powermax* | -dnix* | -nx6 | -nx7 | -sei* | -dragonfly* \
+	      | -skyos* | -haiku* | -rdos* | -toppers* | -drops* | -es* \
+	      | -onefs* | -tirtos* | -phoenix* | -fuchsia*)
 	# Remember, each alternative MUST END IN *, to match a version number.
 		;;
 	-qnx*)
@@ -1185,7 +1427,7 @@ case $os in
 		os=`echo $os | sed -e 's|nto|nto-qnx|'`
 		;;
 	-sim | -es1800* | -hms* | -xray | -os68k* | -none* | -v88r* \
-	      | -windows* | -osx | -abug | -netware* | -os9* | -beos* \
+	      | -windows* | -osx | -abug | -netware* | -os9* | -beos* | -haiku* \
 	      | -macos* | -mpw* | -magic* | -mmixware* | -mon960* | -lnews*)
 		;;
 	-mac*)
@@ -1206,7 +1448,7 @@ case $os in
 	-opened*)
 		os=-openedition
 		;;
-        -os400*)
+	-os400*)
 		os=-os400
 		;;
 	-wince*)
@@ -1255,7 +1497,7 @@ case $os in
 	-sinix*)
 		os=-sysv4
 		;;
-        -tpf*)
+	-tpf*)
 		os=-tpf
 		;;
 	-triton*)
@@ -1291,8 +1533,15 @@ case $os in
 	-aros*)
 		os=-aros
 		;;
-	-kaos*)
-		os=-kaos
+	-zvmoe)
+		os=-zvmoe
+		;;
+	-dicos*)
+		os=-dicos
+		;;
+	-nacl*)
+		;;
+	-ios)
 		;;
 	-none)
 		;;
@@ -1316,6 +1565,12 @@ else
 # system, and we'll never get to this point.
 
 case $basic_machine in
+	score-*)
+		os=-elf
+		;;
+	spu-*)
+		os=-elf
+		;;
 	*-acorn)
 		os=-riscix1.2
 		;;
@@ -1325,9 +1580,24 @@ case $basic_machine in
 	arm*-semi)
 		os=-aout
 		;;
-    c4x-* | tic4x-*)
-        os=-coff
-        ;;
+	c4x-* | tic4x-*)
+		os=-coff
+		;;
+	c8051-*)
+		os=-elf
+		;;
+	hexagon-*)
+		os=-elf
+		;;
+	tic54x-*)
+		os=-coff
+		;;
+	tic55x-*)
+		os=-coff
+		;;
+	tic6x-*)
+		os=-coff
+		;;
 	# This must come before the *-dec entry.
 	pdp10-*)
 		os=-tops20
@@ -1346,13 +1616,13 @@ case $basic_machine in
 		;;
 	m68000-sun)
 		os=-sunos3
-		# This also exists in the configure program, but was not the
-		# default.
-		# os=-sunos4
 		;;
 	m68*-cisco)
 		os=-aout
 		;;
+	mep-*)
+		os=-elf
+		;;
 	mips*-cisco)
 		os=-elf
 		;;
@@ -1371,9 +1641,15 @@ case $basic_machine in
 	*-be)
 		os=-beos
 		;;
+	*-haiku)
+		os=-haiku
+		;;
 	*-ibm)
 		os=-aix
 		;;
+	*-knuth)
+		os=-mmixware
+		;;
 	*-wec)
 		os=-proelf
 		;;
@@ -1476,7 +1752,7 @@ case $basic_machine in
 			-sunos*)
 				vendor=sun
 				;;
-			-aix*)
+			-cnk*|-aix*)
 				vendor=ibm
 				;;
 			-beos*)
@@ -1539,7 +1815,7 @@ case $basic_machine in
 esac
 
 echo $basic_machine$os
-exit 0
+exit
 
 # Local variables:
 # eval: (add-hook 'write-file-hooks 'time-stamp)

doc/html.xsl.in

@@ -0,0 +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>

doc/manpages.xsl.in

@@ -0,0 +1,4 @@
+<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">
+  <xsl:import href="@XSLROOT@/manpages/docbook.xsl"/>
+  <xsl:import href="@abs_srcroot@doc/stylesheet.xsl"/>
+</xsl:stylesheet>

doc/stylesheet.xsl

@@ -0,0 +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="0"/>
+  <xsl:template match="function">
+    <xsl:call-template name="inline.monoseq"/>
+  </xsl:template>
+  <xsl:template match="mallctl">
+    <quote><xsl:call-template name="inline.monoseq"/></quote>
+  </xsl:template>
+</xsl:stylesheet>

include/jemalloc/internal/arena_externs.h

@@ -0,0 +1,96 @@
+#ifndef JEMALLOC_INTERNAL_ARENA_EXTERNS_H
+#define JEMALLOC_INTERNAL_ARENA_EXTERNS_H
+
+#include "jemalloc/internal/extent_dss.h"
+#include "jemalloc/internal/pages.h"
+#include "jemalloc/internal/size_classes.h"
+#include "jemalloc/internal/stats.h"
+
+extern ssize_t opt_dirty_decay_ms;
+extern ssize_t opt_muzzy_decay_ms;
+
+extern const arena_bin_info_t arena_bin_info[NBINS];
+
+extern percpu_arena_mode_t opt_percpu_arena;
+extern const char *percpu_arena_mode_names[];
+
+extern const uint64_t h_steps[SMOOTHSTEP_NSTEPS];
+extern malloc_mutex_t arenas_lock;
+
+void arena_stats_large_nrequests_add(tsdn_t *tsdn, arena_stats_t *arena_stats,
+    szind_t szind, uint64_t nrequests);
+void arena_stats_mapped_add(tsdn_t *tsdn, arena_stats_t *arena_stats,
+    size_t size);
+void arena_basic_stats_merge(tsdn_t *tsdn, arena_t *arena,
+    unsigned *nthreads, const char **dss, ssize_t *dirty_decay_ms,
+    ssize_t *muzzy_decay_ms, size_t *nactive, size_t *ndirty, size_t *nmuzzy);
+void arena_stats_merge(tsdn_t *tsdn, arena_t *arena, unsigned *nthreads,
+    const char **dss, ssize_t *dirty_decay_ms, ssize_t *muzzy_decay_ms,
+    size_t *nactive, size_t *ndirty, size_t *nmuzzy, arena_stats_t *astats,
+    malloc_bin_stats_t *bstats, malloc_large_stats_t *lstats);
+void arena_extents_dirty_dalloc(tsdn_t *tsdn, arena_t *arena,
+    extent_hooks_t **r_extent_hooks, extent_t *extent);
+#ifdef JEMALLOC_JET
+size_t arena_slab_regind(extent_t *slab, szind_t binind, const void *ptr);
+#endif
+extent_t *arena_extent_alloc_large(tsdn_t *tsdn, arena_t *arena,
+    size_t usize, size_t alignment, bool *zero);
+void arena_extent_dalloc_large_prep(tsdn_t *tsdn, arena_t *arena,
+    extent_t *extent);
+void arena_extent_ralloc_large_shrink(tsdn_t *tsdn, arena_t *arena,
+    extent_t *extent, size_t oldsize);
+void arena_extent_ralloc_large_expand(tsdn_t *tsdn, arena_t *arena,
+    extent_t *extent, size_t oldsize);
+ssize_t arena_dirty_decay_ms_get(arena_t *arena);
+bool arena_dirty_decay_ms_set(tsdn_t *tsdn, arena_t *arena, ssize_t decay_ms);
+ssize_t arena_muzzy_decay_ms_get(arena_t *arena);
+bool arena_muzzy_decay_ms_set(tsdn_t *tsdn, arena_t *arena, ssize_t decay_ms);
+void arena_decay(tsdn_t *tsdn, arena_t *arena, bool is_background_thread,
+    bool all);
+void arena_reset(tsd_t *tsd, arena_t *arena);
+void arena_destroy(tsd_t *tsd, arena_t *arena);
+void arena_tcache_fill_small(tsdn_t *tsdn, arena_t *arena, tcache_t *tcache,
+    tcache_bin_t *tbin, szind_t binind, uint64_t prof_accumbytes);
+void arena_alloc_junk_small(void *ptr, const arena_bin_info_t *bin_info,
+    bool zero);
+
+typedef void (arena_dalloc_junk_small_t)(void *, const arena_bin_info_t *);
+extern arena_dalloc_junk_small_t *JET_MUTABLE arena_dalloc_junk_small;
+
+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_promote(tsdn_t *tsdn, const void *ptr, size_t usize);
+void arena_dalloc_promoted(tsdn_t *tsdn, void *ptr, tcache_t *tcache,
+    bool slow_path);
+void arena_dalloc_bin_junked_locked(tsdn_t *tsdn, arena_t *arena,
+    extent_t *extent, void *ptr);
+void arena_dalloc_small(tsdn_t *tsdn, void *ptr);
+bool arena_ralloc_no_move(tsdn_t *tsdn, void *ptr, size_t oldsize, size_t size,
+    size_t extra, bool zero);
+void *arena_ralloc(tsdn_t *tsdn, 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);
+ssize_t arena_dirty_decay_ms_default_get(void);
+bool arena_dirty_decay_ms_default_set(ssize_t decay_ms);
+ssize_t arena_muzzy_decay_ms_default_get(void);
+bool arena_muzzy_decay_ms_default_set(ssize_t decay_ms);
+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, extent_hooks_t *extent_hooks);
+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_prefork4(tsdn_t *tsdn, arena_t *arena);
+void arena_prefork5(tsdn_t *tsdn, arena_t *arena);
+void arena_prefork6(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_INTERNAL_ARENA_EXTERNS_H */

include/jemalloc/internal/arena_inlines_a.h

@@ -0,0 +1,57 @@
+#ifndef JEMALLOC_INTERNAL_ARENA_INLINES_A_H
+#define JEMALLOC_INTERNAL_ARENA_INLINES_A_H
+
+static inline unsigned
+arena_ind_get(const arena_t *arena) {
+	return base_ind_get(arena->base);
+}
+
+static inline void
+arena_internal_add(arena_t *arena, size_t size) {
+	atomic_fetch_add_zu(&arena->stats.internal, size, ATOMIC_RELAXED);
+}
+
+static inline void
+arena_internal_sub(arena_t *arena, size_t size) {
+	atomic_fetch_sub_zu(&arena->stats.internal, size, ATOMIC_RELAXED);
+}
+
+static inline size_t
+arena_internal_get(arena_t *arena) {
+	return atomic_load_zu(&arena->stats.internal, ATOMIC_RELAXED);
+}
+
+static inline bool
+arena_prof_accum(tsdn_t *tsdn, arena_t *arena, uint64_t accumbytes) {
+	cassert(config_prof);
+
+	if (likely(prof_interval == 0)) {
+		return false;
+	}
+
+	return prof_accum_add(tsdn, &arena->prof_accum, accumbytes);
+}
+
+static inline void
+percpu_arena_update(tsd_t *tsd, unsigned cpu) {
+	assert(have_percpu_arena);
+	arena_t *oldarena = tsd_arena_get(tsd);
+	assert(oldarena != NULL);
+	unsigned oldind = arena_ind_get(oldarena);
+
+	if (oldind != cpu) {
+		unsigned newind = cpu;
+		arena_t *newarena = arena_get(tsd_tsdn(tsd), newind, true);
+		assert(newarena != NULL);
+
+		/* Set new arena/tcache associations. */
+		arena_migrate(tsd, oldind, newind);
+		tcache_t *tcache = tcache_get(tsd);
+		if (tcache != NULL) {
+			tcache_arena_reassociate(tsd_tsdn(tsd), tcache,
+			    newarena);
+		}
+	}
+}
+
+#endif /* JEMALLOC_INTERNAL_ARENA_INLINES_A_H */

include/jemalloc/internal/arena_inlines_b.h

@@ -0,0 +1,361 @@
+#ifndef JEMALLOC_INTERNAL_ARENA_INLINES_B_H
+#define JEMALLOC_INTERNAL_ARENA_INLINES_B_H
+
+#include "jemalloc/internal/jemalloc_internal_types.h"
+#include "jemalloc/internal/mutex.h"
+#include "jemalloc/internal/rtree.h"
+#include "jemalloc/internal/size_classes.h"
+#include "jemalloc/internal/sz.h"
+#include "jemalloc/internal/ticker.h"
+
+static inline szind_t
+arena_bin_index(arena_t *arena, arena_bin_t *bin) {
+	szind_t binind = (szind_t)(bin - arena->bins);
+	assert(binind < NBINS);
+	return binind;
+}
+
+JEMALLOC_ALWAYS_INLINE prof_tctx_t *
+arena_prof_tctx_get(tsdn_t *tsdn, const void *ptr, alloc_ctx_t *alloc_ctx) {
+	cassert(config_prof);
+	assert(ptr != NULL);
+
+	/* Static check. */
+	if (alloc_ctx == NULL) {
+		const extent_t *extent = iealloc(tsdn, ptr);
+		if (unlikely(!extent_slab_get(extent))) {
+			return large_prof_tctx_get(tsdn, extent);
+		}
+	} else {
+		if (unlikely(!alloc_ctx->slab)) {
+			return large_prof_tctx_get(tsdn, iealloc(tsdn, ptr));
+		}
+	}
+	return (prof_tctx_t *)(uintptr_t)1U;
+}
+
+JEMALLOC_ALWAYS_INLINE void
+arena_prof_tctx_set(tsdn_t *tsdn, const void *ptr, size_t usize,
+    alloc_ctx_t *alloc_ctx, prof_tctx_t *tctx) {
+	cassert(config_prof);
+	assert(ptr != NULL);
+
+	/* Static check. */
+	if (alloc_ctx == NULL) {
+		extent_t *extent = iealloc(tsdn, ptr);
+		if (unlikely(!extent_slab_get(extent))) {
+			large_prof_tctx_set(tsdn, extent, tctx);
+		}
+	} else {
+		if (unlikely(!alloc_ctx->slab)) {
+			large_prof_tctx_set(tsdn, iealloc(tsdn, ptr), tctx);
+		}
+	}
+}
+
+static inline void
+arena_prof_tctx_reset(tsdn_t *tsdn, const void *ptr, prof_tctx_t *tctx) {
+	cassert(config_prof);
+	assert(ptr != NULL);
+
+	extent_t *extent = iealloc(tsdn, ptr);
+	assert(!extent_slab_get(extent));
+
+	large_prof_tctx_reset(tsdn, extent);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+arena_decay_ticks(tsdn_t *tsdn, arena_t *arena, unsigned nticks) {
+	tsd_t *tsd;
+	ticker_t *decay_ticker;
+
+	if (unlikely(tsdn_null(tsdn))) {
+		return;
+	}
+	tsd = tsdn_tsd(tsdn);
+	decay_ticker = decay_ticker_get(tsd, arena_ind_get(arena));
+	if (unlikely(decay_ticker == NULL)) {
+		return;
+	}
+	if (unlikely(ticker_ticks(decay_ticker, nticks))) {
+		arena_decay(tsdn, arena, false, false);
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE void
+arena_decay_tick(tsdn_t *tsdn, arena_t *arena) {
+	malloc_mutex_assert_not_owner(tsdn, &arena->decay_dirty.mtx);
+	malloc_mutex_assert_not_owner(tsdn, &arena->decay_muzzy.mtx);
+
+	arena_decay_ticks(tsdn, arena, 1);
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+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(tsdn_tsd(tsdn), arena,
+			    tcache, size, ind, zero, slow_path);
+		}
+		if (likely(size <= tcache_maxclass)) {
+			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(tsdn, arena, size, ind, zero);
+}
+
+JEMALLOC_ALWAYS_INLINE arena_t *
+arena_aalloc(tsdn_t *tsdn, const void *ptr) {
+	return extent_arena_get(iealloc(tsdn, ptr));
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+arena_salloc(tsdn_t *tsdn, const void *ptr) {
+	assert(ptr != NULL);
+
+	rtree_ctx_t rtree_ctx_fallback;
+	rtree_ctx_t *rtree_ctx = tsdn_rtree_ctx(tsdn, &rtree_ctx_fallback);
+
+	szind_t szind = rtree_szind_read(tsdn, &extents_rtree, rtree_ctx,
+	    (uintptr_t)ptr, true);
+	assert(szind != NSIZES);
+
+	return sz_index2size(szind);
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+arena_vsalloc(tsdn_t *tsdn, const void *ptr) {
+	/*
+	 * Return 0 if ptr is not within an extent managed by jemalloc.  This
+	 * function has two extra costs relative to isalloc():
+	 * - The rtree calls cannot claim to be dependent lookups, which induces
+	 *   rtree lookup load dependencies.
+	 * - The lookup may fail, so there is an extra branch to check for
+	 *   failure.
+	 */
+
+	rtree_ctx_t rtree_ctx_fallback;
+	rtree_ctx_t *rtree_ctx = tsdn_rtree_ctx(tsdn, &rtree_ctx_fallback);
+
+	extent_t *extent;
+	szind_t szind;
+	if (rtree_extent_szind_read(tsdn, &extents_rtree, rtree_ctx,
+	    (uintptr_t)ptr, false, &extent, &szind)) {
+		return 0;
+	}
+
+	if (extent == NULL) {
+		return 0;
+	}
+	assert(extent_state_get(extent) == extent_state_active);
+	/* Only slab members should be looked up via interior pointers. */
+	assert(extent_addr_get(extent) == ptr || extent_slab_get(extent));
+
+	assert(szind != NSIZES);
+
+	return sz_index2size(szind);
+}
+
+static inline void
+arena_dalloc_no_tcache(tsdn_t *tsdn, void *ptr) {
+	assert(ptr != NULL);
+
+	rtree_ctx_t rtree_ctx_fallback;
+	rtree_ctx_t *rtree_ctx = tsdn_rtree_ctx(tsdn, &rtree_ctx_fallback);
+
+	szind_t szind;
+	bool slab;
+	rtree_szind_slab_read(tsdn, &extents_rtree, rtree_ctx, (uintptr_t)ptr,
+	    true, &szind, &slab);
+
+	if (config_debug) {
+		extent_t *extent = rtree_extent_read(tsdn, &extents_rtree,
+		    rtree_ctx, (uintptr_t)ptr, true);
+		assert(szind == extent_szind_get(extent));
+		assert(szind < NSIZES);
+		assert(slab == extent_slab_get(extent));
+	}
+
+	if (likely(slab)) {
+		/* Small allocation. */
+		arena_dalloc_small(tsdn, ptr);
+	} else {
+		extent_t *extent = iealloc(tsdn, ptr);
+		large_dalloc(tsdn, extent);
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE void
+arena_dalloc(tsdn_t *tsdn, void *ptr, tcache_t *tcache,
+    alloc_ctx_t *alloc_ctx, bool slow_path) {
+	assert(!tsdn_null(tsdn) || tcache == NULL);
+	assert(ptr != NULL);
+
+	if (unlikely(tcache == NULL)) {
+		arena_dalloc_no_tcache(tsdn, ptr);
+		return;
+	}
+
+	szind_t szind;
+	bool slab;
+	rtree_ctx_t *rtree_ctx;
+	if (alloc_ctx != NULL) {
+		szind = alloc_ctx->szind;
+		slab = alloc_ctx->slab;
+		assert(szind != NSIZES);
+	} else {
+		rtree_ctx = tsd_rtree_ctx(tsdn_tsd(tsdn));
+		rtree_szind_slab_read(tsdn, &extents_rtree, rtree_ctx,
+		    (uintptr_t)ptr, true, &szind, &slab);
+	}
+
+	if (config_debug) {
+		rtree_ctx = tsd_rtree_ctx(tsdn_tsd(tsdn));
+		extent_t *extent = rtree_extent_read(tsdn, &extents_rtree,
+		    rtree_ctx, (uintptr_t)ptr, true);
+		assert(szind == extent_szind_get(extent));
+		assert(szind < NSIZES);
+		assert(slab == extent_slab_get(extent));
+	}
+
+	if (likely(slab)) {
+		/* Small allocation. */
+		tcache_dalloc_small(tsdn_tsd(tsdn), tcache, ptr, szind,
+		    slow_path);
+	} else {
+		if (szind < nhbins) {
+			if (config_prof && unlikely(szind < NBINS)) {
+				arena_dalloc_promoted(tsdn, ptr, tcache,
+				    slow_path);
+			} else {
+				tcache_dalloc_large(tsdn_tsd(tsdn), tcache, ptr,
+				    szind, slow_path);
+			}
+		} else {
+			extent_t *extent = iealloc(tsdn, ptr);
+			large_dalloc(tsdn, extent);
+		}
+	}
+}
+
+static inline void
+arena_sdalloc_no_tcache(tsdn_t *tsdn, void *ptr, size_t size) {
+	assert(ptr != NULL);
+	assert(size <= LARGE_MAXCLASS);
+
+	szind_t szind;
+	bool slab;
+	if (!config_prof || !opt_prof) {
+		/*
+		 * There is no risk of being confused by a promoted sampled
+		 * object, so base szind and slab on the given size.
+		 */
+		szind = sz_size2index(size);
+		slab = (szind < NBINS);
+	}
+
+	if ((config_prof && opt_prof) || config_debug) {
+		rtree_ctx_t rtree_ctx_fallback;
+		rtree_ctx_t *rtree_ctx = tsdn_rtree_ctx(tsdn,
+		    &rtree_ctx_fallback);
+
+		rtree_szind_slab_read(tsdn, &extents_rtree, rtree_ctx,
+		    (uintptr_t)ptr, true, &szind, &slab);
+
+		assert(szind == sz_size2index(size));
+		assert((config_prof && opt_prof) || slab == (szind < NBINS));
+
+		if (config_debug) {
+			extent_t *extent = rtree_extent_read(tsdn,
+			    &extents_rtree, rtree_ctx, (uintptr_t)ptr, true);
+			assert(szind == extent_szind_get(extent));
+			assert(slab == extent_slab_get(extent));
+		}
+	}
+
+	if (likely(slab)) {
+		/* Small allocation. */
+		arena_dalloc_small(tsdn, ptr);
+	} else {
+		extent_t *extent = iealloc(tsdn, ptr);
+		large_dalloc(tsdn, extent);
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE void
+arena_sdalloc(tsdn_t *tsdn, void *ptr, size_t size, tcache_t *tcache,
+    alloc_ctx_t *alloc_ctx, bool slow_path) {
+	assert(!tsdn_null(tsdn) || tcache == NULL);
+	assert(ptr != NULL);
+	assert(size <= LARGE_MAXCLASS);
+
+	if (unlikely(tcache == NULL)) {
+		arena_sdalloc_no_tcache(tsdn, ptr, size);
+		return;
+	}
+
+	szind_t szind;
+	bool slab;
+	UNUSED alloc_ctx_t local_ctx;
+	if (config_prof && opt_prof) {
+		if (alloc_ctx == NULL) {
+			/* Uncommon case and should be a static check. */
+			rtree_ctx_t rtree_ctx_fallback;
+			rtree_ctx_t *rtree_ctx = tsdn_rtree_ctx(tsdn,
+			    &rtree_ctx_fallback);
+			rtree_szind_slab_read(tsdn, &extents_rtree, rtree_ctx,
+			    (uintptr_t)ptr, true, &local_ctx.szind,
+			    &local_ctx.slab);
+			assert(local_ctx.szind == sz_size2index(size));
+			alloc_ctx = &local_ctx;
+		}
+		slab = alloc_ctx->slab;
+		szind = alloc_ctx->szind;
+	} else {
+		/*
+		 * There is no risk of being confused by a promoted sampled
+		 * object, so base szind and slab on the given size.
+		 */
+		szind = sz_size2index(size);
+		slab = (szind < NBINS);
+	}
+
+	if (config_debug) {
+		rtree_ctx_t *rtree_ctx = tsd_rtree_ctx(tsdn_tsd(tsdn));
+		rtree_szind_slab_read(tsdn, &extents_rtree, rtree_ctx,
+		    (uintptr_t)ptr, true, &szind, &slab);
+		extent_t *extent = rtree_extent_read(tsdn,
+		    &extents_rtree, rtree_ctx, (uintptr_t)ptr, true);
+		assert(szind == extent_szind_get(extent));
+		assert(slab == extent_slab_get(extent));
+	}
+
+	if (likely(slab)) {
+		/* Small allocation. */
+		tcache_dalloc_small(tsdn_tsd(tsdn), tcache, ptr, szind,
+		    slow_path);
+	} else {
+		if (szind < nhbins) {
+			if (config_prof && unlikely(szind < NBINS)) {
+				arena_dalloc_promoted(tsdn, ptr, tcache,
+				    slow_path);
+			} else {
+				tcache_dalloc_large(tsdn_tsd(tsdn),
+				    tcache, ptr, szind, slow_path);
+			}
+		} else {
+			extent_t *extent = iealloc(tsdn, ptr);
+			large_dalloc(tsdn, extent);
+		}
+	}
+}
+
+#endif /* JEMALLOC_INTERNAL_ARENA_INLINES_B_H */

include/jemalloc/internal/arena_structs_a.h

@@ -0,0 +1,11 @@
+#ifndef JEMALLOC_INTERNAL_ARENA_STRUCTS_A_H
+#define JEMALLOC_INTERNAL_ARENA_STRUCTS_A_H
+
+#include "jemalloc/internal/bitmap.h"
+
+struct arena_slab_data_s {
+	/* Per region allocated/deallocated bitmap. */
+	bitmap_t	bitmap[BITMAP_GROUPS_MAX];
+};
+
+#endif /* JEMALLOC_INTERNAL_ARENA_STRUCTS_A_H */

include/jemalloc/internal/arena_structs_b.h

@@ -0,0 +1,284 @@
+#ifndef JEMALLOC_INTERNAL_ARENA_STRUCTS_B_H
+#define JEMALLOC_INTERNAL_ARENA_STRUCTS_B_H
+
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/bitmap.h"
+#include "jemalloc/internal/extent_dss.h"
+#include "jemalloc/internal/jemalloc_internal_types.h"
+#include "jemalloc/internal/mutex.h"
+#include "jemalloc/internal/nstime.h"
+#include "jemalloc/internal/ql.h"
+#include "jemalloc/internal/size_classes.h"
+#include "jemalloc/internal/smoothstep.h"
+#include "jemalloc/internal/stats.h"
+#include "jemalloc/internal/ticker.h"
+
+/*
+ * Read-only information associated with each element of arena_t's bins array
+ * is stored separately, partly to reduce memory usage (only one copy, rather
+ * than one per arena), but mainly to avoid false cacheline sharing.
+ *
+ * Each slab has the following layout:
+ *
+ *   /--------------------\
+ *   | region 0           |
+ *   |--------------------|
+ *   | region 1           |
+ *   |--------------------|
+ *   | ...                |
+ *   | ...                |
+ *   | ...                |
+ *   |--------------------|
+ *   | region nregs-1     |
+ *   \--------------------/
+ */
+struct arena_bin_info_s {
+	/* Size of regions in a slab for this bin's size class. */
+	size_t			reg_size;
+
+	/* Total size of a slab for this bin's size class. */
+	size_t			slab_size;
+
+	/* Total number of regions in a slab for this bin's size class. */
+	uint32_t		nregs;
+
+	/*
+	 * Metadata used to manipulate bitmaps for slabs associated with this
+	 * bin.
+	 */
+	bitmap_info_t		bitmap_info;
+};
+
+struct arena_decay_s {
+	/* Synchronizes all non-atomic fields. */
+	malloc_mutex_t		mtx;
+	/*
+	 * True if a thread is currently purging the extents associated with
+	 * this decay structure.
+	 */
+	bool			purging;
+	/*
+	 * Approximate time in milliseconds from the creation of a set of unused
+	 * dirty pages until an equivalent set of unused dirty pages is purged
+	 * and/or reused.
+	 */
+	atomic_zd_t		time_ms;
+	/* 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 unpurged pages at beginning of current epoch.  During epoch
+	 * advancement we use the delta between arena->decay_*.nunpurged and
+	 * extents_npages_get(&arena->extents_*) to determine how many dirty
+	 * pages, if any, were generated.
+	 */
+	size_t			nunpurged;
+	/*
+	 * 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];
+
+	/*
+	 * Pointer to associated stats.  These stats are embedded directly in
+	 * the arena's stats due to how stats structures are shared between the
+	 * arena and ctl code.
+	 *
+	 * Synchronization: Same as associated arena's stats field. */
+	decay_stats_t		*stats;
+	/* Peak number of pages in associated extents.  Used for debug only. */
+	uint64_t		ceil_npages;
+};
+
+struct arena_bin_s {
+	/* All operations on arena_bin_t fields require lock ownership. */
+	malloc_mutex_t		lock;
+
+	/*
+	 * Current slab being used to service allocations of this bin's size
+	 * class.  slabcur is independent of slabs_{nonfull,full}; whenever
+	 * slabcur is reassigned, the previous slab must be deallocated or
+	 * inserted into slabs_{nonfull,full}.
+	 */
+	extent_t		*slabcur;
+
+	/*
+	 * Heap of non-full slabs.  This heap is used to assure that new
+	 * allocations come from the non-full slab that is oldest/lowest in
+	 * memory.
+	 */
+	extent_heap_t		slabs_nonfull;
+
+	/* List used to track full slabs. */
+	extent_list_t		slabs_full;
+
+	/* Bin statistics. */
+	malloc_bin_stats_t	stats;
+};
+
+struct arena_s {
+	/*
+	 * Number of threads currently assigned to this arena.  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 explicitly created via the arenas.create
+	 * mallctl, because the arena.<i>.reset mallctl indiscriminately
+	 * discards all allocations for the affected arena.
+	 *
+	 *   0: Application allocation.
+	 *   1: Internal metadata allocation.
+	 *
+	 * Synchronization: atomic.
+	 */
+	atomic_u_t		nthreads[2];
+
+	/*
+	 * When percpu_arena is enabled, to amortize the cost of reading /
+	 * updating the current CPU id, track the most recent thread accessing
+	 * this arena, and only read CPU if there is a mismatch.
+	 */
+	tsdn_t		*last_thd;
+
+	/* Synchronization: internal. */
+	arena_stats_t		stats;
+
+	/*
+	 * List of tcaches for extant threads associated with this arena.
+	 * Stats from these are merged incrementally, and at exit if
+	 * opt_stats_print is enabled.
+	 *
+	 * Synchronization: tcache_ql_mtx.
+	 */
+	ql_head(tcache_t)	tcache_ql;
+	malloc_mutex_t		tcache_ql_mtx;
+
+	/* Synchronization: internal. */
+	prof_accum_t		prof_accum;
+	uint64_t		prof_accumbytes;
+
+	/*
+	 * PRNG state for cache index randomization of large allocation base
+	 * pointers.
+	 *
+	 * Synchronization: atomic.
+	 */
+	atomic_zu_t		offset_state;
+
+	/*
+	 * Extent serial number generator state.
+	 *
+	 * Synchronization: atomic.
+	 */
+	atomic_zu_t		extent_sn_next;
+
+	/*
+	 * Represents a dss_prec_t, but atomically.
+	 *
+	 * Synchronization: atomic.
+	 */
+	atomic_u_t		dss_prec;
+
+	/*
+	 * Number of pages in active extents.
+	 *
+	 * Synchronization: atomic.
+	 */
+	atomic_zu_t		nactive;
+
+	/*
+	 * Extant large allocations.
+	 *
+	 * Synchronization: large_mtx.
+	 */
+	extent_list_t		large;
+	/* Synchronizes all large allocation/update/deallocation. */
+	malloc_mutex_t		large_mtx;
+
+	/*
+	 * Collections of extents that were previously allocated.  These are
+	 * used when allocating extents, in an attempt to re-use address space.
+	 *
+	 * Synchronization: internal.
+	 */
+	extents_t		extents_dirty;
+	extents_t		extents_muzzy;
+	extents_t		extents_retained;
+
+	/*
+	 * Decay-based purging state, responsible for scheduling extent state
+	 * transitions.
+	 *
+	 * Synchronization: internal.
+	 */
+	arena_decay_t		decay_dirty; /* dirty --> muzzy */
+	arena_decay_t		decay_muzzy; /* muzzy --> retained */
+
+	/*
+	 * Next extent size class in a growing series to use when satisfying a
+	 * request via the extent hooks (only if opt_retain).  This limits the
+	 * number of disjoint virtual memory ranges so that extent merging can
+	 * be effective even if multiple arenas' extent allocation requests are
+	 * highly interleaved.
+	 *
+	 * Synchronization: extent_grow_mtx
+	 */
+	pszind_t		extent_grow_next;
+	malloc_mutex_t		extent_grow_mtx;
+
+	/*
+	 * Available extent structures that were allocated via
+	 * base_alloc_extent().
+	 *
+	 * Synchronization: extent_avail_mtx.
+	 */
+	extent_tree_t		extent_avail;
+	malloc_mutex_t		extent_avail_mtx;
+
+	/*
+	 * bins is used to store heaps of free regions.
+	 *
+	 * Synchronization: internal.
+	 */
+	arena_bin_t		bins[NBINS];
+
+	/*
+	 * Base allocator, from which arena metadata are allocated.
+	 *
+	 * Synchronization: internal.
+	 */
+	base_t			*base;
+	/* Used to determine uptime.  Read-only after initialization. */
+	nstime_t		create_time;
+};
+
+/* Used in conjunction with tsd for fast arena-related context lookup. */
+struct arena_tdata_s {
+	ticker_t		decay_ticker;
+};
+
+/* Used to pass rtree lookup context down the path. */
+struct alloc_ctx_s {
+	szind_t szind;
+	bool slab;
+};
+
+#endif /* JEMALLOC_INTERNAL_ARENA_STRUCTS_B_H */

include/jemalloc/internal/arena_types.h

@@ -0,0 +1,45 @@
+#ifndef JEMALLOC_INTERNAL_ARENA_TYPES_H
+#define JEMALLOC_INTERNAL_ARENA_TYPES_H
+
+/* Maximum number of regions in one slab. */
+#define LG_SLAB_MAXREGS		(LG_PAGE - LG_TINY_MIN)
+#define SLAB_MAXREGS		(1U << LG_SLAB_MAXREGS)
+
+/* Default decay times in milliseconds. */
+#define DIRTY_DECAY_MS_DEFAULT	ZD(10 * 1000)
+#define MUZZY_DECAY_MS_DEFAULT	ZD(10 * 1000)
+/* Number of event ticks between time checks. */
+#define DECAY_NTICKS_PER_UPDATE	1000
+
+typedef struct arena_slab_data_s arena_slab_data_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;
+typedef struct alloc_ctx_s alloc_ctx_t;
+
+typedef enum {
+	percpu_arena_mode_names_base   = 0, /* Used for options processing. */
+
+	/*
+	 * *_uninit are used only during bootstrapping, and must correspond
+	 * to initialized variant plus percpu_arena_mode_enabled_base.
+	 */
+	percpu_arena_uninit            = 0,
+	per_phycpu_arena_uninit        = 1,
+
+	/* All non-disabled modes must come after percpu_arena_disabled. */
+	percpu_arena_disabled          = 2,
+
+	percpu_arena_mode_names_limit  = 3, /* Used for options processing. */
+	percpu_arena_mode_enabled_base = 3,
+
+	percpu_arena                   = 3,
+	per_phycpu_arena               = 4  /* Hyper threads share arena. */
+} percpu_arena_mode_t;
+
+#define PERCPU_ARENA_ENABLED(m)	((m) >= percpu_arena_mode_enabled_base)
+#define PERCPU_ARENA_DEFAULT	percpu_arena_disabled
+
+#endif /* JEMALLOC_INTERNAL_ARENA_TYPES_H */

include/jemalloc/internal/assert.h

@@ -0,0 +1,56 @@
+#include "jemalloc/internal/malloc_io.h"
+#include "jemalloc/internal/util.h"
+
+/*
+ * Define a custom assert() in order to reduce the chances of deadlock during
+ * assertion failure.
+ */
+#ifndef assert
+#define assert(e) do {							\
+	if (unlikely(config_debug && !(e))) {				\
+		malloc_printf(						\
+		    "<jemalloc>: %s:%d: Failed assertion: \"%s\"\n",	\
+		    __FILE__, __LINE__, #e);				\
+		abort();						\
+	}								\
+} while (0)
+#endif
+
+#ifndef not_reached
+#define not_reached() do {						\
+	if (config_debug) {						\
+		malloc_printf(						\
+		    "<jemalloc>: %s:%d: Unreachable code reached\n",	\
+		    __FILE__, __LINE__);				\
+		abort();						\
+	}								\
+	unreachable();							\
+} while (0)
+#endif
+
+#ifndef not_implemented
+#define not_implemented() do {						\
+	if (config_debug) {						\
+		malloc_printf("<jemalloc>: %s:%d: Not implemented\n",	\
+		    __FILE__, __LINE__);				\
+		abort();						\
+	}								\
+} while (0)
+#endif
+
+#ifndef assert_not_implemented
+#define assert_not_implemented(e) do {					\
+	if (unlikely(config_debug && !(e))) {				\
+		not_implemented();					\
+	}								\
+} while (0)
+#endif
+
+/* Use to assert a particular configuration, e.g., cassert(config_debug). */
+#ifndef cassert
+#define cassert(c) do {							\
+	if (unlikely(!(c))) {						\
+		not_reached();						\
+	}								\
+} while (0)
+#endif

include/jemalloc/internal/atomic.h

@@ -0,0 +1,77 @@
+#ifndef JEMALLOC_INTERNAL_ATOMIC_H
+#define JEMALLOC_INTERNAL_ATOMIC_H
+
+#define ATOMIC_INLINE static inline
+
+#if defined(JEMALLOC_GCC_ATOMIC_ATOMICS)
+#  include "jemalloc/internal/atomic_gcc_atomic.h"
+#elif defined(JEMALLOC_GCC_SYNC_ATOMICS)
+#  include "jemalloc/internal/atomic_gcc_sync.h"
+#elif defined(_MSC_VER)
+#  include "jemalloc/internal/atomic_msvc.h"
+#elif defined(JEMALLOC_C11_ATOMICS)
+#  include "jemalloc/internal/atomic_c11.h"
+#else
+#  error "Don't have atomics implemented on this platform."
+#endif
+
+/*
+ * This header gives more or less a backport of C11 atomics. The user can write
+ * JEMALLOC_GENERATE_ATOMICS(type, short_type, lg_sizeof_type); to generate
+ * counterparts of the C11 atomic functions for type, as so:
+ *   JEMALLOC_GENERATE_ATOMICS(int *, pi, 3);
+ * and then write things like:
+ *   int *some_ptr;
+ *   atomic_pi_t atomic_ptr_to_int;
+ *   atomic_store_pi(&atomic_ptr_to_int, some_ptr, ATOMIC_RELAXED);
+ *   int *prev_value = atomic_exchange_pi(&ptr_to_int, NULL, ATOMIC_ACQ_REL);
+ *   assert(some_ptr == prev_value);
+ * and expect things to work in the obvious way.
+ *
+ * Also included (with naming differences to avoid conflicts with the standard
+ * library):
+ *   atomic_fence(atomic_memory_order_t) (mimics C11's atomic_thread_fence).
+ *   ATOMIC_INIT (mimics C11's ATOMIC_VAR_INIT).
+ */
+
+/*
+ * Pure convenience, so that we don't have to type "atomic_memory_order_"
+ * quite so often.
+ */
+#define ATOMIC_RELAXED atomic_memory_order_relaxed
+#define ATOMIC_ACQUIRE atomic_memory_order_acquire
+#define ATOMIC_RELEASE atomic_memory_order_release
+#define ATOMIC_ACQ_REL atomic_memory_order_acq_rel
+#define ATOMIC_SEQ_CST atomic_memory_order_seq_cst
+
+/*
+ * Not all platforms have 64-bit atomics.  If we do, this #define exposes that
+ * fact.
+ */
+#if (LG_SIZEOF_PTR == 3 || LG_SIZEOF_INT == 3)
+#  define JEMALLOC_ATOMIC_U64
+#endif
+
+JEMALLOC_GENERATE_ATOMICS(void *, p, LG_SIZEOF_PTR)
+
+/*
+ * There's no actual guarantee that sizeof(bool) == 1, but it's true on the only
+ * platform that actually needs to know the size, MSVC.
+ */
+JEMALLOC_GENERATE_ATOMICS(bool, b, 0)
+
+JEMALLOC_GENERATE_INT_ATOMICS(unsigned, u, LG_SIZEOF_INT)
+
+JEMALLOC_GENERATE_INT_ATOMICS(size_t, zu, LG_SIZEOF_PTR)
+
+JEMALLOC_GENERATE_INT_ATOMICS(ssize_t, zd, LG_SIZEOF_PTR)
+
+JEMALLOC_GENERATE_INT_ATOMICS(uint32_t, u32, 2)
+
+#ifdef JEMALLOC_ATOMIC_U64
+JEMALLOC_GENERATE_INT_ATOMICS(uint64_t, u64, 3)
+#endif
+
+#undef ATOMIC_INLINE
+
+#endif /* JEMALLOC_INTERNAL_ATOMIC_H */

include/jemalloc/internal/atomic_c11.h

@@ -0,0 +1,97 @@
+#ifndef JEMALLOC_INTERNAL_ATOMIC_C11_H
+#define JEMALLOC_INTERNAL_ATOMIC_C11_H
+
+#include <stdatomic.h>
+
+#define ATOMIC_INIT(...) ATOMIC_VAR_INIT(__VA_ARGS__)
+
+#define atomic_memory_order_t memory_order
+#define atomic_memory_order_relaxed memory_order_relaxed
+#define atomic_memory_order_acquire memory_order_acquire
+#define atomic_memory_order_release memory_order_release
+#define atomic_memory_order_acq_rel memory_order_acq_rel
+#define atomic_memory_order_seq_cst memory_order_seq_cst
+
+#define atomic_fence atomic_thread_fence
+
+#define JEMALLOC_GENERATE_ATOMICS(type, short_type,			\
+    /* unused */ lg_size)						\
+typedef _Atomic(type) atomic_##short_type##_t;				\
+									\
+ATOMIC_INLINE type							\
+atomic_load_##short_type(const atomic_##short_type##_t *a,		\
+    atomic_memory_order_t mo) {						\
+	/*								\
+	 * A strict interpretation of the C standard prevents		\
+	 * atomic_load from taking a const argument, but it's		\
+	 * convenient for our purposes. This cast is a workaround.	\
+	 */								\
+	atomic_##short_type##_t* a_nonconst =				\
+	    (atomic_##short_type##_t*)a;				\
+	return atomic_load_explicit(a_nonconst, mo);			\
+}									\
+									\
+ATOMIC_INLINE void							\
+atomic_store_##short_type(atomic_##short_type##_t *a,			\
+    type val, atomic_memory_order_t mo) {				\
+	atomic_store_explicit(a, val, mo);				\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_exchange_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return atomic_exchange_explicit(a, val, mo);			\
+}									\
+									\
+ATOMIC_INLINE bool							\
+atomic_compare_exchange_weak_##short_type(atomic_##short_type##_t *a,	\
+    type *expected, type desired, atomic_memory_order_t success_mo,	\
+    atomic_memory_order_t failure_mo) {					\
+	return atomic_compare_exchange_weak_explicit(a, expected,	\
+	    desired, success_mo, failure_mo);				\
+}									\
+									\
+ATOMIC_INLINE bool							\
+atomic_compare_exchange_strong_##short_type(atomic_##short_type##_t *a,	\
+    type *expected, type desired, atomic_memory_order_t success_mo,	\
+    atomic_memory_order_t failure_mo) {					\
+	return atomic_compare_exchange_strong_explicit(a, expected,	\
+	    desired, success_mo, failure_mo);				\
+}
+
+/*
+ * Integral types have some special operations available that non-integral ones
+ * lack.
+ */
+#define JEMALLOC_GENERATE_INT_ATOMICS(type, short_type, 		\
+    /* unused */ lg_size)						\
+JEMALLOC_GENERATE_ATOMICS(type, short_type, /* unused */ lg_size)	\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_add_##short_type(atomic_##short_type##_t *a,		\
+    type val, atomic_memory_order_t mo) {				\
+	return atomic_fetch_add_explicit(a, val, mo);			\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_sub_##short_type(atomic_##short_type##_t *a,		\
+    type val, atomic_memory_order_t mo) {				\
+	return atomic_fetch_sub_explicit(a, val, mo);			\
+}									\
+ATOMIC_INLINE type							\
+atomic_fetch_and_##short_type(atomic_##short_type##_t *a,		\
+    type val, atomic_memory_order_t mo) {				\
+	return atomic_fetch_and_explicit(a, val, mo);			\
+}									\
+ATOMIC_INLINE type							\
+atomic_fetch_or_##short_type(atomic_##short_type##_t *a,		\
+    type val, atomic_memory_order_t mo) {				\
+	return atomic_fetch_or_explicit(a, val, mo);			\
+}									\
+ATOMIC_INLINE type							\
+atomic_fetch_xor_##short_type(atomic_##short_type##_t *a,		\
+    type val, atomic_memory_order_t mo) {				\
+	return atomic_fetch_xor_explicit(a, val, mo);			\
+}
+
+#endif /* JEMALLOC_INTERNAL_ATOMIC_C11_H */

include/jemalloc/internal/atomic_gcc_atomic.h

@@ -0,0 +1,127 @@
+#ifndef JEMALLOC_INTERNAL_ATOMIC_GCC_ATOMIC_H
+#define JEMALLOC_INTERNAL_ATOMIC_GCC_ATOMIC_H
+
+#include "jemalloc/internal/assert.h"
+
+#define ATOMIC_INIT(...) {__VA_ARGS__}
+
+typedef enum {
+	atomic_memory_order_relaxed,
+	atomic_memory_order_acquire,
+	atomic_memory_order_release,
+	atomic_memory_order_acq_rel,
+	atomic_memory_order_seq_cst
+} atomic_memory_order_t;
+
+ATOMIC_INLINE int
+atomic_enum_to_builtin(atomic_memory_order_t mo) {
+	switch (mo) {
+	case atomic_memory_order_relaxed:
+		return __ATOMIC_RELAXED;
+	case atomic_memory_order_acquire:
+		return __ATOMIC_ACQUIRE;
+	case atomic_memory_order_release:
+		return __ATOMIC_RELEASE;
+	case atomic_memory_order_acq_rel:
+		return __ATOMIC_ACQ_REL;
+	case atomic_memory_order_seq_cst:
+		return __ATOMIC_SEQ_CST;
+	}
+	/* Can't happen; the switch is exhaustive. */
+	not_reached();
+}
+
+ATOMIC_INLINE void
+atomic_fence(atomic_memory_order_t mo) {
+	__atomic_thread_fence(atomic_enum_to_builtin(mo));
+}
+
+#define JEMALLOC_GENERATE_ATOMICS(type, short_type,			\
+    /* unused */ lg_size)						\
+typedef struct {							\
+	type repr;							\
+} atomic_##short_type##_t;						\
+									\
+ATOMIC_INLINE type							\
+atomic_load_##short_type(const atomic_##short_type##_t *a,		\
+    atomic_memory_order_t mo) {						\
+	type result;							\
+	__atomic_load(&a->repr, &result, atomic_enum_to_builtin(mo));	\
+	return result;							\
+}									\
+									\
+ATOMIC_INLINE void							\
+atomic_store_##short_type(atomic_##short_type##_t *a, type val,		\
+    atomic_memory_order_t mo) {						\
+	__atomic_store(&a->repr, &val, atomic_enum_to_builtin(mo));	\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_exchange_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	type result;							\
+	__atomic_exchange(&a->repr, &val, &result,			\
+	    atomic_enum_to_builtin(mo));				\
+	return result;							\
+}									\
+									\
+ATOMIC_INLINE bool							\
+atomic_compare_exchange_weak_##short_type(atomic_##short_type##_t *a,	\
+    type *expected, type desired, atomic_memory_order_t success_mo,	\
+    atomic_memory_order_t failure_mo) {					\
+	return __atomic_compare_exchange(&a->repr, expected, &desired,	\
+	    true, atomic_enum_to_builtin(success_mo),			\
+	    atomic_enum_to_builtin(failure_mo));			\
+}									\
+									\
+ATOMIC_INLINE bool							\
+atomic_compare_exchange_strong_##short_type(atomic_##short_type##_t *a,	\
+    type *expected, type desired, atomic_memory_order_t success_mo,	\
+    atomic_memory_order_t failure_mo) {					\
+	return __atomic_compare_exchange(&a->repr, expected, &desired,	\
+	    false,							\
+	    atomic_enum_to_builtin(success_mo),				\
+	    atomic_enum_to_builtin(failure_mo));			\
+}
+
+
+#define JEMALLOC_GENERATE_INT_ATOMICS(type, short_type,			\
+    /* unused */ lg_size)						\
+JEMALLOC_GENERATE_ATOMICS(type, short_type, /* unused */ lg_size)	\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_add_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return __atomic_fetch_add(&a->repr, val,			\
+	    atomic_enum_to_builtin(mo));				\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_sub_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return __atomic_fetch_sub(&a->repr, val,			\
+	    atomic_enum_to_builtin(mo));				\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_and_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return __atomic_fetch_and(&a->repr, val,			\
+	    atomic_enum_to_builtin(mo));				\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_or_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return __atomic_fetch_or(&a->repr, val,				\
+	    atomic_enum_to_builtin(mo));				\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_xor_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return __atomic_fetch_xor(&a->repr, val,			\
+	    atomic_enum_to_builtin(mo));				\
+}
+
+#endif /* JEMALLOC_INTERNAL_ATOMIC_GCC_ATOMIC_H */

include/jemalloc/internal/atomic_gcc_sync.h

@@ -0,0 +1,191 @@
+#ifndef JEMALLOC_INTERNAL_ATOMIC_GCC_SYNC_H
+#define JEMALLOC_INTERNAL_ATOMIC_GCC_SYNC_H
+
+#define ATOMIC_INIT(...) {__VA_ARGS__}
+
+typedef enum {
+	atomic_memory_order_relaxed,
+	atomic_memory_order_acquire,
+	atomic_memory_order_release,
+	atomic_memory_order_acq_rel,
+	atomic_memory_order_seq_cst
+} atomic_memory_order_t;
+
+ATOMIC_INLINE void
+atomic_fence(atomic_memory_order_t mo) {
+	/* Easy cases first: no barrier, and full barrier. */
+	if (mo == atomic_memory_order_relaxed) {
+		asm volatile("" ::: "memory");
+		return;
+	}
+	if (mo == atomic_memory_order_seq_cst) {
+		asm volatile("" ::: "memory");
+		__sync_synchronize();
+		asm volatile("" ::: "memory");
+		return;
+	}
+	asm volatile("" ::: "memory");
+#  if defined(__i386__) || defined(__x86_64__)
+	/* This is implicit on x86. */
+#  elif defined(__ppc__)
+	asm volatile("lwsync");
+#  elif defined(__sparc__) && defined(__arch64__)
+	if (mo == atomic_memory_order_acquire) {
+		asm volatile("membar #LoadLoad | #LoadStore");
+	} else if (mo == atomic_memory_order_release) {
+		asm volatile("membar #LoadStore | #StoreStore");
+	} else {
+		asm volatile("membar #LoadLoad | #LoadStore | #StoreStore");
+	}
+#  else
+	__sync_synchronize();
+#  endif
+	asm volatile("" ::: "memory");
+}
+
+/*
+ * A correct implementation of seq_cst loads and stores on weakly ordered
+ * architectures could do either of the following:
+ *   1. store() is weak-fence -> store -> strong fence, load() is load ->
+ *      strong-fence.
+ *   2. store() is strong-fence -> store, load() is strong-fence -> load ->
+ *      weak-fence.
+ * The tricky thing is, load() and store() above can be the load or store
+ * portions of a gcc __sync builtin, so we have to follow GCC's lead, which
+ * means going with strategy 2.
+ * On strongly ordered architectures, the natural strategy is to stick a strong
+ * fence after seq_cst stores, and have naked loads.  So we want the strong
+ * fences in different places on different architectures.
+ * atomic_pre_sc_load_fence and atomic_post_sc_store_fence allow us to
+ * accomplish this.
+ */
+
+ATOMIC_INLINE void
+atomic_pre_sc_load_fence() {
+#  if defined(__i386__) || defined(__x86_64__) ||			\
+    (defined(__sparc__) && defined(__arch64__))
+	atomic_fence(atomic_memory_order_relaxed);
+#  else
+	atomic_fence(atomic_memory_order_seq_cst);
+#  endif
+}
+
+ATOMIC_INLINE void
+atomic_post_sc_store_fence() {
+#  if defined(__i386__) || defined(__x86_64__) ||			\
+    (defined(__sparc__) && defined(__arch64__))
+	atomic_fence(atomic_memory_order_seq_cst);
+#  else
+	atomic_fence(atomic_memory_order_relaxed);
+#  endif
+
+}
+
+#define JEMALLOC_GENERATE_ATOMICS(type, short_type,			\
+    /* unused */ lg_size)						\
+typedef struct {							\
+	type volatile repr;						\
+} atomic_##short_type##_t;						\
+									\
+ATOMIC_INLINE type							\
+atomic_load_##short_type(const atomic_##short_type##_t *a,		\
+    atomic_memory_order_t mo) {						\
+	if (mo == atomic_memory_order_seq_cst) {			\
+		atomic_pre_sc_load_fence();				\
+	}								\
+	type result = a->repr;						\
+	if (mo != atomic_memory_order_relaxed) {			\
+		atomic_fence(atomic_memory_order_acquire);		\
+	}								\
+	return result;							\
+}									\
+									\
+ATOMIC_INLINE void							\
+atomic_store_##short_type(atomic_##short_type##_t *a,			\
+    type val, atomic_memory_order_t mo) {				\
+	if (mo != atomic_memory_order_relaxed) {			\
+		atomic_fence(atomic_memory_order_release);		\
+	}								\
+	a->repr = val;							\
+	if (mo == atomic_memory_order_seq_cst) {			\
+		atomic_post_sc_store_fence();				\
+	}								\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_exchange_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	/*								\
+	 * Because of FreeBSD, we care about gcc 4.2, which doesn't have\
+	 * an atomic exchange builtin.  We fake it with a CAS loop.	\
+	 */								\
+	while (true) {							\
+		type old = a->repr;					\
+		if (__sync_bool_compare_and_swap(&a->repr, old, val)) {	\
+			return old;					\
+		}							\
+	}								\
+}									\
+									\
+ATOMIC_INLINE bool							\
+atomic_compare_exchange_weak_##short_type(atomic_##short_type##_t *a,	\
+    type *expected, type desired, atomic_memory_order_t success_mo,	\
+    atomic_memory_order_t failure_mo) {					\
+	type prev = __sync_val_compare_and_swap(&a->repr, *expected,	\
+	    desired);							\
+	if (prev == *expected) {					\
+		return true;						\
+	} else {							\
+		*expected = prev;					\
+		return false;						\
+	}								\
+}									\
+ATOMIC_INLINE bool							\
+atomic_compare_exchange_strong_##short_type(atomic_##short_type##_t *a,	\
+    type *expected, type desired, atomic_memory_order_t success_mo,	\
+    atomic_memory_order_t failure_mo) {					\
+	type prev = __sync_val_compare_and_swap(&a->repr, *expected,	\
+	    desired);							\
+	if (prev == *expected) {					\
+		return true;						\
+	} else {							\
+		*expected = prev;					\
+		return false;						\
+	}								\
+}
+
+#define JEMALLOC_GENERATE_INT_ATOMICS(type, short_type,			\
+    /* unused */ lg_size)						\
+JEMALLOC_GENERATE_ATOMICS(type, short_type, /* unused */ lg_size)	\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_add_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return __sync_fetch_and_add(&a->repr, val);			\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_sub_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return __sync_fetch_and_sub(&a->repr, val);			\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_and_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return __sync_fetch_and_and(&a->repr, val);			\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_or_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return __sync_fetch_and_or(&a->repr, val);			\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_xor_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return __sync_fetch_and_xor(&a->repr, val);			\
+}
+
+#endif /* JEMALLOC_INTERNAL_ATOMIC_GCC_SYNC_H */

include/jemalloc/internal/atomic_msvc.h

@@ -0,0 +1,158 @@
+#ifndef JEMALLOC_INTERNAL_ATOMIC_MSVC_H
+#define JEMALLOC_INTERNAL_ATOMIC_MSVC_H
+
+#define ATOMIC_INIT(...) {__VA_ARGS__}
+
+typedef enum {
+	atomic_memory_order_relaxed,
+	atomic_memory_order_acquire,
+	atomic_memory_order_release,
+	atomic_memory_order_acq_rel,
+	atomic_memory_order_seq_cst
+} atomic_memory_order_t;
+
+typedef char atomic_repr_0_t;
+typedef short atomic_repr_1_t;
+typedef long atomic_repr_2_t;
+typedef __int64 atomic_repr_3_t;
+
+ATOMIC_INLINE void
+atomic_fence(atomic_memory_order_t mo) {
+	_ReadWriteBarrier();
+#  if defined(_M_ARM) || defined(_M_ARM64)
+	/* ARM needs a barrier for everything but relaxed. */
+	if (mo != atomic_memory_order_relaxed) {
+		MemoryBarrier();
+	}
+#  elif defined(_M_IX86) || defined (_M_X64)
+	/* x86 needs a barrier only for seq_cst. */
+	if (mo == atomic_memory_order_seq_cst) {
+		MemoryBarrier();
+	}
+#  else
+#  error "Don't know how to create atomics for this platform for MSVC."
+#  endif
+	_ReadWriteBarrier();
+}
+
+#define ATOMIC_INTERLOCKED_REPR(lg_size) atomic_repr_ ## lg_size ## _t
+
+#define ATOMIC_CONCAT(a, b) ATOMIC_RAW_CONCAT(a, b)
+#define ATOMIC_RAW_CONCAT(a, b) a ## b
+
+#define ATOMIC_INTERLOCKED_NAME(base_name, lg_size) ATOMIC_CONCAT(	\
+    base_name, ATOMIC_INTERLOCKED_SUFFIX(lg_size))
+
+#define ATOMIC_INTERLOCKED_SUFFIX(lg_size)				\
+    ATOMIC_CONCAT(ATOMIC_INTERLOCKED_SUFFIX_, lg_size)
+
+#define ATOMIC_INTERLOCKED_SUFFIX_0 8
+#define ATOMIC_INTERLOCKED_SUFFIX_1 16
+#define ATOMIC_INTERLOCKED_SUFFIX_2
+#define ATOMIC_INTERLOCKED_SUFFIX_3 64
+
+#define JEMALLOC_GENERATE_ATOMICS(type, short_type, lg_size)		\
+typedef struct {							\
+	ATOMIC_INTERLOCKED_REPR(lg_size) repr;				\
+} atomic_##short_type##_t;						\
+									\
+ATOMIC_INLINE type							\
+atomic_load_##short_type(const atomic_##short_type##_t *a,		\
+    atomic_memory_order_t mo) {						\
+	ATOMIC_INTERLOCKED_REPR(lg_size) ret = a->repr;			\
+	if (mo != atomic_memory_order_relaxed) {			\
+		atomic_fence(atomic_memory_order_acquire);		\
+	}								\
+	return (type) ret;						\
+}									\
+									\
+ATOMIC_INLINE void							\
+atomic_store_##short_type(atomic_##short_type##_t *a,			\
+    type val, atomic_memory_order_t mo) {				\
+	if (mo != atomic_memory_order_relaxed) {			\
+		atomic_fence(atomic_memory_order_release);		\
+	}								\
+	a->repr = (ATOMIC_INTERLOCKED_REPR(lg_size)) val;		\
+	if (mo == atomic_memory_order_seq_cst) {			\
+		atomic_fence(atomic_memory_order_seq_cst);		\
+	}								\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_exchange_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return (type)ATOMIC_INTERLOCKED_NAME(_InterlockedExchange,	\
+	    lg_size)(&a->repr, (ATOMIC_INTERLOCKED_REPR(lg_size))val);	\
+}									\
+									\
+ATOMIC_INLINE bool							\
+atomic_compare_exchange_weak_##short_type(atomic_##short_type##_t *a,	\
+    type *expected, type desired, atomic_memory_order_t success_mo,	\
+    atomic_memory_order_t failure_mo) {					\
+	ATOMIC_INTERLOCKED_REPR(lg_size) e =				\
+	    (ATOMIC_INTERLOCKED_REPR(lg_size))*expected;		\
+	ATOMIC_INTERLOCKED_REPR(lg_size) d =				\
+	    (ATOMIC_INTERLOCKED_REPR(lg_size))desired;			\
+	ATOMIC_INTERLOCKED_REPR(lg_size) old =				\
+	    ATOMIC_INTERLOCKED_NAME(_InterlockedCompareExchange, 	\
+		lg_size)(&a->repr, d, e);				\
+	if (old == e) {							\
+		return true;						\
+	} else {							\
+		*expected = (type)old;					\
+		return false;						\
+	}								\
+}									\
+									\
+ATOMIC_INLINE bool							\
+atomic_compare_exchange_strong_##short_type(atomic_##short_type##_t *a,	\
+    type *expected, type desired, atomic_memory_order_t success_mo,	\
+    atomic_memory_order_t failure_mo) {					\
+	/* We implement the weak version with strong semantics. */	\
+	return atomic_compare_exchange_weak_##short_type(a, expected,	\
+	    desired, success_mo, failure_mo);				\
+}
+
+
+#define JEMALLOC_GENERATE_INT_ATOMICS(type, short_type, lg_size)	\
+JEMALLOC_GENERATE_ATOMICS(type, short_type, lg_size)			\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_add_##short_type(atomic_##short_type##_t *a,		\
+    type val, atomic_memory_order_t mo) {				\
+	return (type)ATOMIC_INTERLOCKED_NAME(_InterlockedExchangeAdd,	\
+	    lg_size)(&a->repr, (ATOMIC_INTERLOCKED_REPR(lg_size))val);	\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_sub_##short_type(atomic_##short_type##_t *a,		\
+    type val, atomic_memory_order_t mo) {				\
+	/*								\
+	 * MSVC warns on negation of unsigned operands, but for us it	\
+	 * gives exactly the right semantics (MAX_TYPE + 1 - operand).	\
+	 */								\
+	__pragma(warning(push))						\
+	__pragma(warning(disable: 4146))				\
+	return atomic_fetch_add_##short_type(a, -val, mo);		\
+	__pragma(warning(pop))						\
+}									\
+ATOMIC_INLINE type							\
+atomic_fetch_and_##short_type(atomic_##short_type##_t *a,		\
+    type val, atomic_memory_order_t mo) {				\
+	return (type)ATOMIC_INTERLOCKED_NAME(_InterlockedAnd, lg_size)(	\
+	    &a->repr, (ATOMIC_INTERLOCKED_REPR(lg_size))val);		\
+}									\
+ATOMIC_INLINE type							\
+atomic_fetch_or_##short_type(atomic_##short_type##_t *a,		\
+    type val, atomic_memory_order_t mo) {				\
+	return (type)ATOMIC_INTERLOCKED_NAME(_InterlockedOr, lg_size)(	\
+	    &a->repr, (ATOMIC_INTERLOCKED_REPR(lg_size))val);		\
+}									\
+ATOMIC_INLINE type							\
+atomic_fetch_xor_##short_type(atomic_##short_type##_t *a,		\
+    type val, atomic_memory_order_t mo) {				\
+	return (type)ATOMIC_INTERLOCKED_NAME(_InterlockedXor, lg_size)(	\
+	    &a->repr, (ATOMIC_INTERLOCKED_REPR(lg_size))val);		\
+}
+
+#endif /* JEMALLOC_INTERNAL_ATOMIC_MSVC_H */

include/jemalloc/internal/background_thread_externs.h

@@ -0,0 +1,30 @@
+#ifndef JEMALLOC_INTERNAL_BACKGROUND_THREAD_EXTERNS_H
+#define JEMALLOC_INTERNAL_BACKGROUND_THREAD_EXTERNS_H
+
+extern bool opt_background_thread;
+extern malloc_mutex_t background_thread_lock;
+extern atomic_b_t background_thread_enabled_state;
+extern size_t n_background_threads;
+extern background_thread_info_t *background_thread_info;
+
+bool background_thread_create(tsd_t *tsd, unsigned arena_ind);
+bool background_threads_enable(tsd_t *tsd);
+bool background_threads_disable(tsd_t *tsd);
+void background_thread_interval_check(tsdn_t *tsdn, arena_t *arena,
+    arena_decay_t *decay, size_t npages_new);
+void background_thread_prefork0(tsdn_t *tsdn);
+void background_thread_prefork1(tsdn_t *tsdn);
+void background_thread_postfork_parent(tsdn_t *tsdn);
+void background_thread_postfork_child(tsdn_t *tsdn);
+bool background_thread_stats_read(tsdn_t *tsdn,
+    background_thread_stats_t *stats);
+void background_thread_ctl_init(tsdn_t *tsdn);
+
+#ifdef JEMALLOC_PTHREAD_CREATE_WRAPPER
+extern int pthread_create_wrapper(pthread_t *__restrict, const pthread_attr_t *,
+    void *(*)(void *), void *__restrict);
+#endif
+bool background_thread_boot0(void);
+bool background_thread_boot1(tsdn_t *tsdn);
+
+#endif /* JEMALLOC_INTERNAL_BACKGROUND_THREAD_EXTERNS_H */

include/jemalloc/internal/background_thread_inlines.h

@@ -0,0 +1,56 @@
+#ifndef JEMALLOC_INTERNAL_BACKGROUND_THREAD_INLINES_H
+#define JEMALLOC_INTERNAL_BACKGROUND_THREAD_INLINES_H
+
+JEMALLOC_ALWAYS_INLINE bool
+background_thread_enabled(void) {
+	return atomic_load_b(&background_thread_enabled_state, ATOMIC_RELAXED);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+background_thread_enabled_set(tsdn_t *tsdn, bool state) {
+	malloc_mutex_assert_owner(tsdn, &background_thread_lock);
+	atomic_store_b(&background_thread_enabled_state, state, ATOMIC_RELAXED);
+}
+
+JEMALLOC_ALWAYS_INLINE background_thread_info_t *
+arena_background_thread_info_get(arena_t *arena) {
+	unsigned arena_ind = arena_ind_get(arena);
+	return &background_thread_info[arena_ind % ncpus];
+}
+
+JEMALLOC_ALWAYS_INLINE uint64_t
+background_thread_wakeup_time_get(background_thread_info_t *info) {
+	uint64_t next_wakeup = nstime_ns(&info->next_wakeup);
+	assert(atomic_load_b(&info->indefinite_sleep, ATOMIC_ACQUIRE) ==
+	    (next_wakeup == BACKGROUND_THREAD_INDEFINITE_SLEEP));
+	return next_wakeup;
+}
+
+JEMALLOC_ALWAYS_INLINE void
+background_thread_wakeup_time_set(tsdn_t *tsdn, background_thread_info_t *info,
+    uint64_t wakeup_time) {
+	malloc_mutex_assert_owner(tsdn, &info->mtx);
+	atomic_store_b(&info->indefinite_sleep,
+	    wakeup_time == BACKGROUND_THREAD_INDEFINITE_SLEEP, ATOMIC_RELEASE);
+	nstime_init(&info->next_wakeup, wakeup_time);
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+background_thread_indefinite_sleep(background_thread_info_t *info) {
+	return atomic_load_b(&info->indefinite_sleep, ATOMIC_ACQUIRE);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+arena_background_thread_inactivity_check(tsdn_t *tsdn, arena_t *arena) {
+	if (!background_thread_enabled()) {
+		return;
+	}
+	background_thread_info_t *info =
+	    arena_background_thread_info_get(arena);
+	if (background_thread_indefinite_sleep(info)) {
+		background_thread_interval_check(tsdn, arena,
+		    &arena->decay_dirty, 0);
+	}
+}
+
+#endif /* JEMALLOC_INTERNAL_BACKGROUND_THREAD_INLINES_H */

include/jemalloc/internal/background_thread_structs.h

@@ -0,0 +1,52 @@
+#ifndef JEMALLOC_INTERNAL_BACKGROUND_THREAD_STRUCTS_H
+#define JEMALLOC_INTERNAL_BACKGROUND_THREAD_STRUCTS_H
+
+/* This file really combines "structs" and "types", but only transitionally. */
+
+#if defined(JEMALLOC_BACKGROUND_THREAD) || defined(JEMALLOC_LAZY_LOCK)
+#  define JEMALLOC_PTHREAD_CREATE_WRAPPER
+#endif
+
+#define BACKGROUND_THREAD_INDEFINITE_SLEEP UINT64_MAX
+
+typedef enum {
+	background_thread_stopped,
+	background_thread_started,
+	/* Thread waits on the global lock when paused (for arena_reset). */
+	background_thread_paused,
+} background_thread_state_t;
+
+struct background_thread_info_s {
+#ifdef JEMALLOC_BACKGROUND_THREAD
+	/* Background thread is pthread specific. */
+	pthread_t		thread;
+	pthread_cond_t		cond;
+#endif
+	malloc_mutex_t		mtx;
+	background_thread_state_t	state;
+	/* When true, it means no wakeup scheduled. */
+	atomic_b_t		indefinite_sleep;
+	/* Next scheduled wakeup time (absolute time in ns). */
+	nstime_t		next_wakeup;
+	/*
+	 *  Since the last background thread run, newly added number of pages
+	 *  that need to be purged by the next wakeup.  This is adjusted on
+	 *  epoch advance, and is used to determine whether we should signal the
+	 *  background thread to wake up earlier.
+	 */
+	size_t			npages_to_purge_new;
+	/* Stats: total number of runs since started. */
+	uint64_t		tot_n_runs;
+	/* Stats: total sleep time since started. */
+	nstime_t		tot_sleep_time;
+};
+typedef struct background_thread_info_s background_thread_info_t;
+
+struct background_thread_stats_s {
+	size_t num_threads;
+	uint64_t num_runs;
+	nstime_t run_interval;
+};
+typedef struct background_thread_stats_s background_thread_stats_t;
+
+#endif /* JEMALLOC_INTERNAL_BACKGROUND_THREAD_STRUCTS_H */

include/jemalloc/internal/base_externs.h

@@ -0,0 +1,19 @@
+#ifndef JEMALLOC_INTERNAL_BASE_EXTERNS_H
+#define JEMALLOC_INTERNAL_BASE_EXTERNS_H
+
+base_t *b0get(void);
+base_t *base_new(tsdn_t *tsdn, unsigned ind, extent_hooks_t *extent_hooks);
+void base_delete(base_t *base);
+extent_hooks_t *base_extent_hooks_get(base_t *base);
+extent_hooks_t *base_extent_hooks_set(base_t *base,
+    extent_hooks_t *extent_hooks);
+void *base_alloc(tsdn_t *tsdn, base_t *base, size_t size, size_t alignment);
+extent_t *base_alloc_extent(tsdn_t *tsdn, base_t *base);
+void base_stats_get(tsdn_t *tsdn, base_t *base, size_t *allocated,
+    size_t *resident, size_t *mapped);
+void base_prefork(tsdn_t *tsdn, base_t *base);
+void base_postfork_parent(tsdn_t *tsdn, base_t *base);
+void base_postfork_child(tsdn_t *tsdn, base_t *base);
+bool base_boot(tsdn_t *tsdn);
+
+#endif /* JEMALLOC_INTERNAL_BASE_EXTERNS_H */

include/jemalloc/internal/base_inlines.h

@@ -0,0 +1,9 @@
+#ifndef JEMALLOC_INTERNAL_BASE_INLINES_H
+#define JEMALLOC_INTERNAL_BASE_INLINES_H
+
+static inline unsigned
+base_ind_get(const base_t *base) {
+	return base->ind;
+}
+
+#endif /* JEMALLOC_INTERNAL_BASE_INLINES_H */

include/jemalloc/internal/base_structs.h

@@ -0,0 +1,55 @@
+#ifndef JEMALLOC_INTERNAL_BASE_STRUCTS_H
+#define JEMALLOC_INTERNAL_BASE_STRUCTS_H
+
+#include "jemalloc/internal/jemalloc_internal_types.h"
+#include "jemalloc/internal/mutex.h"
+#include "jemalloc/internal/size_classes.h"
+
+/* Embedded at the beginning of every block of base-managed virtual memory. */
+struct base_block_s {
+	/* Total size of block's virtual memory mapping. */
+	size_t		size;
+
+	/* Next block in list of base's blocks. */
+	base_block_t	*next;
+
+	/* Tracks unused trailing space. */
+	extent_t	extent;
+};
+
+struct base_s {
+	/* Associated arena's index within the arenas array. */
+	unsigned	ind;
+
+	/*
+	 * User-configurable extent hook functions.  Points to an
+	 * extent_hooks_t.
+	 */
+	atomic_p_t	extent_hooks;
+
+	/* Protects base_alloc() and base_stats_get() operations. */
+	malloc_mutex_t	mtx;
+
+	/*
+	 * Most recent size class in the series of increasingly large base
+	 * extents.  Logarithmic spacing between subsequent allocations ensures
+	 * that the total number of distinct mappings remains small.
+	 */
+	pszind_t	pind_last;
+
+	/* Serial number generation state. */
+	size_t		extent_sn_next;
+
+	/* Chain of all blocks associated with base. */
+	base_block_t	*blocks;
+
+	/* Heap of extents that track unused trailing space within blocks. */
+	extent_heap_t	avail[NSIZES];
+
+	/* Stats, only maintained if config_stats. */
+	size_t		allocated;
+	size_t		resident;
+	size_t		mapped;
+};
+
+#endif /* JEMALLOC_INTERNAL_BASE_STRUCTS_H */

include/jemalloc/internal/base_types.h

@@ -0,0 +1,7 @@
+#ifndef JEMALLOC_INTERNAL_BASE_TYPES_H
+#define JEMALLOC_INTERNAL_BASE_TYPES_H
+
+typedef struct base_block_s base_block_t;
+typedef struct base_s base_t;
+
+#endif /* JEMALLOC_INTERNAL_BASE_TYPES_H */

include/jemalloc/internal/bit_util.h

@@ -0,0 +1,165 @@
+#ifndef JEMALLOC_INTERNAL_BIT_UTIL_H
+#define JEMALLOC_INTERNAL_BIT_UTIL_H
+
+#include "jemalloc/internal/assert.h"
+
+#define BIT_UTIL_INLINE static inline
+
+/* Sanity check. */
+#if !defined(JEMALLOC_INTERNAL_FFSLL) || !defined(JEMALLOC_INTERNAL_FFSL) \
+    || !defined(JEMALLOC_INTERNAL_FFS)
+#  error JEMALLOC_INTERNAL_FFS{,L,LL} should have been defined by configure
+#endif
+
+
+BIT_UTIL_INLINE unsigned
+ffs_llu(unsigned long long bitmap) {
+	return JEMALLOC_INTERNAL_FFSLL(bitmap);
+}
+
+BIT_UTIL_INLINE unsigned
+ffs_lu(unsigned long bitmap) {
+	return JEMALLOC_INTERNAL_FFSL(bitmap);
+}
+
+BIT_UTIL_INLINE unsigned
+ffs_u(unsigned bitmap) {
+	return JEMALLOC_INTERNAL_FFS(bitmap);
+}
+
+BIT_UTIL_INLINE unsigned
+ffs_zu(size_t bitmap) {
+#if LG_SIZEOF_PTR == LG_SIZEOF_INT
+	return ffs_u(bitmap);
+#elif LG_SIZEOF_PTR == LG_SIZEOF_LONG
+	return ffs_lu(bitmap);
+#elif LG_SIZEOF_PTR == LG_SIZEOF_LONG_LONG
+	return ffs_llu(bitmap);
+#else
+#error No implementation for size_t ffs()
+#endif
+}
+
+BIT_UTIL_INLINE unsigned
+ffs_u64(uint64_t bitmap) {
+#if LG_SIZEOF_LONG == 3
+	return ffs_lu(bitmap);
+#elif LG_SIZEOF_LONG_LONG == 3
+	return ffs_llu(bitmap);
+#else
+#error No implementation for 64-bit ffs()
+#endif
+}
+
+BIT_UTIL_INLINE unsigned
+ffs_u32(uint32_t bitmap) {
+#if LG_SIZEOF_INT == 2
+	return ffs_u(bitmap);
+#else
+#error No implementation for 32-bit ffs()
+#endif
+	return ffs_u(bitmap);
+}
+
+BIT_UTIL_INLINE uint64_t
+pow2_ceil_u64(uint64_t x) {
+	x--;
+	x |= x >> 1;
+	x |= x >> 2;
+	x |= x >> 4;
+	x |= x >> 8;
+	x |= x >> 16;
+	x |= x >> 32;
+	x++;
+	return x;
+}
+
+BIT_UTIL_INLINE uint32_t
+pow2_ceil_u32(uint32_t x) {
+	x--;
+	x |= x >> 1;
+	x |= x >> 2;
+	x |= x >> 4;
+	x |= x >> 8;
+	x |= x >> 16;
+	x++;
+	return x;
+}
+
+/* Compute the smallest power of 2 that is >= x. */
+BIT_UTIL_INLINE size_t
+pow2_ceil_zu(size_t x) {
+#if (LG_SIZEOF_PTR == 3)
+	return pow2_ceil_u64(x);
+#else
+	return pow2_ceil_u32(x);
+#endif
+}
+
+#if (defined(__i386__) || defined(__amd64__) || defined(__x86_64__))
+BIT_UTIL_INLINE unsigned
+lg_floor(size_t x) {
+	size_t ret;
+	assert(x != 0);
+
+	asm ("bsr %1, %0"
+	    : "=r"(ret) // Outputs.
+	    : "r"(x)    // Inputs.
+	    );
+	assert(ret < UINT_MAX);
+	return (unsigned)ret;
+}
+#elif (defined(_MSC_VER))
+BIT_UTIL_INLINE unsigned
+lg_floor(size_t x) {
+	unsigned long ret;
+
+	assert(x != 0);
+
+#if (LG_SIZEOF_PTR == 3)
+	_BitScanReverse64(&ret, x);
+#elif (LG_SIZEOF_PTR == 2)
+	_BitScanReverse(&ret, x);
+#else
+#  error "Unsupported type size for lg_floor()"
+#endif
+	assert(ret < UINT_MAX);
+	return (unsigned)ret;
+}
+#elif (defined(JEMALLOC_HAVE_BUILTIN_CLZ))
+BIT_UTIL_INLINE unsigned
+lg_floor(size_t x) {
+	assert(x != 0);
+
+#if (LG_SIZEOF_PTR == LG_SIZEOF_INT)
+	return ((8 << LG_SIZEOF_PTR) - 1) - __builtin_clz(x);
+#elif (LG_SIZEOF_PTR == LG_SIZEOF_LONG)
+	return ((8 << LG_SIZEOF_PTR) - 1) - __builtin_clzl(x);
+#else
+#  error "Unsupported type size for lg_floor()"
+#endif
+}
+#else
+BIT_UTIL_INLINE unsigned
+lg_floor(size_t x) {
+	assert(x != 0);
+
+	x |= (x >> 1);
+	x |= (x >> 2);
+	x |= (x >> 4);
+	x |= (x >> 8);
+	x |= (x >> 16);
+#if (LG_SIZEOF_PTR == 3)
+	x |= (x >> 32);
+#endif
+	if (x == SIZE_T_MAX) {
+		return (8 << LG_SIZEOF_PTR) - 1;
+	}
+	x++;
+	return ffs_zu(x) - 2;
+}
+#endif
+
+#undef BIT_UTIL_INLINE
+
+#endif /* JEMALLOC_INTERNAL_BIT_UTIL_H */

include/jemalloc/internal/bitmap.h

@@ -0,0 +1,369 @@
+#ifndef JEMALLOC_INTERNAL_BITMAP_H
+#define JEMALLOC_INTERNAL_BITMAP_H
+
+#include "jemalloc/internal/arena_types.h"
+#include "jemalloc/internal/bit_util.h"
+#include "jemalloc/internal/size_classes.h"
+
+typedef unsigned long bitmap_t;
+#define LG_SIZEOF_BITMAP	LG_SIZEOF_LONG
+
+/* Maximum bitmap bit count is 2^LG_BITMAP_MAXBITS. */
+#if LG_SLAB_MAXREGS > LG_CEIL_NSIZES
+/* Maximum bitmap bit count is determined by maximum regions per slab. */
+#  define LG_BITMAP_MAXBITS	LG_SLAB_MAXREGS
+#else
+/* Maximum bitmap bit count is determined by number of extent size classes. */
+#  define LG_BITMAP_MAXBITS	LG_CEIL_NSIZES
+#endif
+#define BITMAP_MAXBITS		(ZU(1) << LG_BITMAP_MAXBITS)
+
+/* Number of bits per group. */
+#define LG_BITMAP_GROUP_NBITS		(LG_SIZEOF_BITMAP + 3)
+#define BITMAP_GROUP_NBITS		(1U << LG_BITMAP_GROUP_NBITS)
+#define BITMAP_GROUP_NBITS_MASK		(BITMAP_GROUP_NBITS-1)
+
+/*
+ * 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 ffs_lu() more than 2^3 times,
+ * use a tree instead.
+ */
+#if LG_BITMAP_MAXBITS - LG_BITMAP_GROUP_NBITS > 3
+#  define BITMAP_USE_TREE
+#endif
+
+/* Number of groups required to store a given number of bits. */
+#define BITMAP_BITS2GROUPS(nbits)					\
+    (((nbits) + BITMAP_GROUP_NBITS_MASK) >> LG_BITMAP_GROUP_NBITS)
+
+/*
+ * Number of groups required at a particular level for a given number of bits.
+ */
+#define BITMAP_GROUPS_L0(nbits)						\
+    BITMAP_BITS2GROUPS(nbits)
+#define BITMAP_GROUPS_L1(nbits)						\
+    BITMAP_BITS2GROUPS(BITMAP_BITS2GROUPS(nbits))
+#define BITMAP_GROUPS_L2(nbits)						\
+    BITMAP_BITS2GROUPS(BITMAP_BITS2GROUPS(BITMAP_BITS2GROUPS((nbits))))
+#define BITMAP_GROUPS_L3(nbits)						\
+    BITMAP_BITS2GROUPS(BITMAP_BITS2GROUPS(BITMAP_BITS2GROUPS(		\
+	BITMAP_BITS2GROUPS((nbits)))))
+#define BITMAP_GROUPS_L4(nbits)						\
+    BITMAP_BITS2GROUPS(BITMAP_BITS2GROUPS(BITMAP_BITS2GROUPS(		\
+	BITMAP_BITS2GROUPS(BITMAP_BITS2GROUPS((nbits))))))
+
+/*
+ * Assuming the number of levels, number of groups required for a given number
+ * of bits.
+ */
+#define BITMAP_GROUPS_1_LEVEL(nbits)					\
+    BITMAP_GROUPS_L0(nbits)
+#define BITMAP_GROUPS_2_LEVEL(nbits)					\
+    (BITMAP_GROUPS_1_LEVEL(nbits) + BITMAP_GROUPS_L1(nbits))
+#define BITMAP_GROUPS_3_LEVEL(nbits)					\
+    (BITMAP_GROUPS_2_LEVEL(nbits) + BITMAP_GROUPS_L2(nbits))
+#define BITMAP_GROUPS_4_LEVEL(nbits)					\
+    (BITMAP_GROUPS_3_LEVEL(nbits) + BITMAP_GROUPS_L3(nbits))
+#define BITMAP_GROUPS_5_LEVEL(nbits)					\
+    (BITMAP_GROUPS_4_LEVEL(nbits) + BITMAP_GROUPS_L4(nbits))
+
+/*
+ * Maximum number of groups required to support LG_BITMAP_MAXBITS.
+ */
+#ifdef BITMAP_USE_TREE
+
+#if LG_BITMAP_MAXBITS <= LG_BITMAP_GROUP_NBITS
+#  define BITMAP_GROUPS(nbits)	BITMAP_GROUPS_1_LEVEL(nbits)
+#  define BITMAP_GROUPS_MAX	BITMAP_GROUPS_1_LEVEL(BITMAP_MAXBITS)
+#elif LG_BITMAP_MAXBITS <= LG_BITMAP_GROUP_NBITS * 2
+#  define BITMAP_GROUPS(nbits)	BITMAP_GROUPS_2_LEVEL(nbits)
+#  define BITMAP_GROUPS_MAX	BITMAP_GROUPS_2_LEVEL(BITMAP_MAXBITS)
+#elif LG_BITMAP_MAXBITS <= LG_BITMAP_GROUP_NBITS * 3
+#  define BITMAP_GROUPS(nbits)	BITMAP_GROUPS_3_LEVEL(nbits)
+#  define BITMAP_GROUPS_MAX	BITMAP_GROUPS_3_LEVEL(BITMAP_MAXBITS)
+#elif LG_BITMAP_MAXBITS <= LG_BITMAP_GROUP_NBITS * 4
+#  define BITMAP_GROUPS(nbits)	BITMAP_GROUPS_4_LEVEL(nbits)
+#  define BITMAP_GROUPS_MAX	BITMAP_GROUPS_4_LEVEL(BITMAP_MAXBITS)
+#elif LG_BITMAP_MAXBITS <= LG_BITMAP_GROUP_NBITS * 5
+#  define BITMAP_GROUPS(nbits)	BITMAP_GROUPS_5_LEVEL(nbits)
+#  define BITMAP_GROUPS_MAX	BITMAP_GROUPS_5_LEVEL(BITMAP_MAXBITS)
+#else
+#  error "Unsupported bitmap size"
+#endif
+
+/*
+ * Maximum number of levels possible.  This could be statically computed based
+ * on LG_BITMAP_MAXBITS:
+ *
+ * #define BITMAP_MAX_LEVELS \
+ *     (LG_BITMAP_MAXBITS / LG_SIZEOF_BITMAP) \
+ *     + !!(LG_BITMAP_MAXBITS % LG_SIZEOF_BITMAP)
+ *
+ * However, that would not allow the generic BITMAP_INFO_INITIALIZER() macro, so
+ * instead hardcode BITMAP_MAX_LEVELS to the largest number supported by the
+ * various cascading macros.  The only additional cost this incurs is some
+ * unused trailing entries in bitmap_info_t structures; the bitmaps themselves
+ * are not impacted.
+ */
+#define BITMAP_MAX_LEVELS	5
+
+#define BITMAP_INFO_INITIALIZER(nbits) {				\
+	/* nbits. */							\
+	nbits,								\
+	/* nlevels. */							\
+	(BITMAP_GROUPS_L0(nbits) > BITMAP_GROUPS_L1(nbits)) +		\
+	    (BITMAP_GROUPS_L1(nbits) > BITMAP_GROUPS_L2(nbits)) +	\
+	    (BITMAP_GROUPS_L2(nbits) > BITMAP_GROUPS_L3(nbits)) +	\
+	    (BITMAP_GROUPS_L3(nbits) > BITMAP_GROUPS_L4(nbits)) + 1,	\
+	/* levels. */							\
+	{								\
+		{0},							\
+		{BITMAP_GROUPS_L0(nbits)},				\
+		{BITMAP_GROUPS_L1(nbits) + BITMAP_GROUPS_L0(nbits)},	\
+		{BITMAP_GROUPS_L2(nbits) + BITMAP_GROUPS_L1(nbits) +	\
+		    BITMAP_GROUPS_L0(nbits)},				\
+		{BITMAP_GROUPS_L3(nbits) + BITMAP_GROUPS_L2(nbits) +	\
+		    BITMAP_GROUPS_L1(nbits) + BITMAP_GROUPS_L0(nbits)},	\
+		{BITMAP_GROUPS_L4(nbits) + BITMAP_GROUPS_L3(nbits) +	\
+		     BITMAP_GROUPS_L2(nbits) + BITMAP_GROUPS_L1(nbits)	\
+		     + BITMAP_GROUPS_L0(nbits)}				\
+	}								\
+}
+
+#else /* BITMAP_USE_TREE */
+
+#define BITMAP_GROUPS(nbits)	BITMAP_BITS2GROUPS(nbits)
+#define BITMAP_GROUPS_MAX	BITMAP_BITS2GROUPS(BITMAP_MAXBITS)
+
+#define BITMAP_INFO_INITIALIZER(nbits) {				\
+	/* nbits. */							\
+	nbits,								\
+	/* ngroups. */							\
+	BITMAP_BITS2GROUPS(nbits)					\
+}
+
+#endif /* BITMAP_USE_TREE */
+
+typedef struct bitmap_level_s {
+	/* Offset of this level's groups within the array of groups. */
+	size_t group_offset;
+} bitmap_level_t;
+
+typedef struct bitmap_info_s {
+	/* Logical number of bits in bitmap (stored at bottom level). */
+	size_t nbits;
+
+#ifdef BITMAP_USE_TREE
+	/* Number of levels necessary for nbits. */
+	unsigned nlevels;
+
+	/*
+	 * Only the first (nlevels+1) elements are used, and levels are ordered
+	 * bottom to top (e.g. the bottom level is stored in levels[0]).
+	 */
+	bitmap_level_t levels[BITMAP_MAX_LEVELS+1];
+#else /* BITMAP_USE_TREE */
+	/* Number of groups necessary for nbits. */
+	size_t ngroups;
+#endif /* BITMAP_USE_TREE */
+} bitmap_info_t;
+
+void bitmap_info_init(bitmap_info_t *binfo, size_t nbits);
+void bitmap_init(bitmap_t *bitmap, const bitmap_info_t *binfo, bool fill);
+size_t bitmap_size(const bitmap_info_t *binfo);
+
+static inline bool
+bitmap_full(bitmap_t *bitmap, const bitmap_info_t *binfo) {
+#ifdef BITMAP_USE_TREE
+	size_t rgoff = binfo->levels[binfo->nlevels].group_offset - 1;
+	bitmap_t rg = bitmap[rgoff];
+	/* The bitmap is full iff the root group is 0. */
+	return (rg == 0);
+#else
+	size_t i;
+
+	for (i = 0; i < binfo->ngroups; i++) {
+		if (bitmap[i] != 0) {
+			return false;
+		}
+	}
+	return true;
+#endif
+}
+
+static inline bool
+bitmap_get(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit) {
+	size_t goff;
+	bitmap_t g;
+
+	assert(bit < binfo->nbits);
+	goff = bit >> LG_BITMAP_GROUP_NBITS;
+	g = bitmap[goff];
+	return !(g & (ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK)));
+}
+
+static inline void
+bitmap_set(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit) {
+	size_t goff;
+	bitmap_t *gp;
+	bitmap_t g;
+
+	assert(bit < binfo->nbits);
+	assert(!bitmap_get(bitmap, binfo, bit));
+	goff = bit >> LG_BITMAP_GROUP_NBITS;
+	gp = &bitmap[goff];
+	g = *gp;
+	assert(g & (ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK)));
+	g ^= ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK);
+	*gp = g;
+	assert(bitmap_get(bitmap, binfo, bit));
+#ifdef BITMAP_USE_TREE
+	/* Propagate group state transitions up the tree. */
+	if (g == 0) {
+		unsigned i;
+		for (i = 1; i < binfo->nlevels; i++) {
+			bit = goff;
+			goff = bit >> LG_BITMAP_GROUP_NBITS;
+			gp = &bitmap[binfo->levels[i].group_offset + goff];
+			g = *gp;
+			assert(g & (ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK)));
+			g ^= ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK);
+			*gp = g;
+			if (g != 0) {
+				break;
+			}
+		}
+	}
+#endif
+}
+
+/* ffu: find first unset >= bit. */
+static inline size_t
+bitmap_ffu(const bitmap_t *bitmap, const bitmap_info_t *binfo, size_t min_bit) {
+	assert(min_bit < binfo->nbits);
+
+#ifdef BITMAP_USE_TREE
+	size_t bit = 0;
+	for (unsigned level = binfo->nlevels; level--;) {
+		size_t lg_bits_per_group = (LG_BITMAP_GROUP_NBITS * (level +
+		    1));
+		bitmap_t group = bitmap[binfo->levels[level].group_offset + (bit
+		    >> lg_bits_per_group)];
+		unsigned group_nmask = (unsigned)(((min_bit > bit) ? (min_bit -
+		    bit) : 0) >> (lg_bits_per_group - LG_BITMAP_GROUP_NBITS));
+		assert(group_nmask <= BITMAP_GROUP_NBITS);
+		bitmap_t group_mask = ~((1LU << group_nmask) - 1);
+		bitmap_t group_masked = group & group_mask;
+		if (group_masked == 0LU) {
+			if (group == 0LU) {
+				return binfo->nbits;
+			}
+			/*
+			 * min_bit was preceded by one or more unset bits in
+			 * this group, but there are no other unset bits in this
+			 * group.  Try again starting at the first bit of the
+			 * next sibling.  This will recurse at most once per
+			 * non-root level.
+			 */
+			size_t sib_base = bit + (ZU(1) << lg_bits_per_group);
+			assert(sib_base > min_bit);
+			assert(sib_base > bit);
+			if (sib_base >= binfo->nbits) {
+				return binfo->nbits;
+			}
+			return bitmap_ffu(bitmap, binfo, sib_base);
+		}
+		bit += ((size_t)(ffs_lu(group_masked) - 1)) <<
+		    (lg_bits_per_group - LG_BITMAP_GROUP_NBITS);
+	}
+	assert(bit >= min_bit);
+	assert(bit < binfo->nbits);
+	return bit;
+#else
+	size_t i = min_bit >> LG_BITMAP_GROUP_NBITS;
+	bitmap_t g = bitmap[i] & ~((1LU << (min_bit & BITMAP_GROUP_NBITS_MASK))
+	    - 1);
+	size_t bit;
+	do {
+		bit = ffs_lu(g);
+		if (bit != 0) {
+			return (i << LG_BITMAP_GROUP_NBITS) + (bit - 1);
+		}
+		i++;
+		g = bitmap[i];
+	} while (i < binfo->ngroups);
+	return binfo->nbits;
+#endif
+}
+
+/* sfu: set first unset. */
+static inline size_t
+bitmap_sfu(bitmap_t *bitmap, const bitmap_info_t *binfo) {
+	size_t bit;
+	bitmap_t g;
+	unsigned i;
+
+	assert(!bitmap_full(bitmap, binfo));
+
+#ifdef BITMAP_USE_TREE
+	i = binfo->nlevels - 1;
+	g = bitmap[binfo->levels[i].group_offset];
+	bit = ffs_lu(g) - 1;
+	while (i > 0) {
+		i--;
+		g = bitmap[binfo->levels[i].group_offset + bit];
+		bit = (bit << LG_BITMAP_GROUP_NBITS) + (ffs_lu(g) - 1);
+	}
+#else
+	i = 0;
+	g = bitmap[0];
+	while ((bit = ffs_lu(g)) == 0) {
+		i++;
+		g = bitmap[i];
+	}
+	bit = (i << LG_BITMAP_GROUP_NBITS) + (bit - 1);
+#endif
+	bitmap_set(bitmap, binfo, bit);
+	return bit;
+}
+
+static inline void
+bitmap_unset(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit) {
+	size_t goff;
+	bitmap_t *gp;
+	bitmap_t g;
+	UNUSED bool propagate;
+
+	assert(bit < binfo->nbits);
+	assert(bitmap_get(bitmap, binfo, bit));
+	goff = bit >> LG_BITMAP_GROUP_NBITS;
+	gp = &bitmap[goff];
+	g = *gp;
+	propagate = (g == 0);
+	assert((g & (ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK))) == 0);
+	g ^= ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK);
+	*gp = g;
+	assert(!bitmap_get(bitmap, binfo, bit));
+#ifdef BITMAP_USE_TREE
+	/* Propagate group state transitions up the tree. */
+	if (propagate) {
+		unsigned i;
+		for (i = 1; i < binfo->nlevels; i++) {
+			bit = goff;
+			goff = bit >> LG_BITMAP_GROUP_NBITS;
+			gp = &bitmap[binfo->levels[i].group_offset + goff];
+			g = *gp;
+			propagate = (g == 0);
+			assert((g & (ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK)))
+			    == 0);
+			g ^= ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK);
+			*gp = g;
+			if (!propagate) {
+				break;
+			}
+		}
+	}
+#endif /* BITMAP_USE_TREE */
+}
+
+#endif /* JEMALLOC_INTERNAL_BITMAP_H */

include/jemalloc/internal/ckh.h

@@ -0,0 +1,101 @@
+#ifndef JEMALLOC_INTERNAL_CKH_H
+#define JEMALLOC_INTERNAL_CKH_H
+
+#include "jemalloc/internal/tsd.h"
+
+/* Cuckoo hashing implementation.  Skip to the end for the interface. */
+
+/******************************************************************************/
+/* INTERNAL DEFINITIONS -- IGNORE */
+/******************************************************************************/
+
+/* Maintain counters used to get an idea of performance. */
+/* #define CKH_COUNT */
+/* Print counter values in ckh_delete() (requires CKH_COUNT). */
+/* #define CKH_VERBOSE */
+
+/*
+ * There are 2^LG_CKH_BUCKET_CELLS cells in each hash table bucket.  Try to fit
+ * one bucket per L1 cache line.
+ */
+#define LG_CKH_BUCKET_CELLS (LG_CACHELINE - LG_SIZEOF_PTR - 1)
+
+/* Typedefs to allow easy function pointer passing. */
+typedef void ckh_hash_t (const void *, size_t[2]);
+typedef bool ckh_keycomp_t (const void *, const void *);
+
+/* Hash table cell. */
+typedef struct {
+	const void *key;
+	const void *data;
+} ckhc_t;
+
+/* The hash table itself. */
+typedef struct {
+#ifdef CKH_COUNT
+	/* Counters used to get an idea of performance. */
+	uint64_t ngrows;
+	uint64_t nshrinks;
+	uint64_t nshrinkfails;
+	uint64_t ninserts;
+	uint64_t nrelocs;
+#endif
+
+	/* Used for pseudo-random number generation. */
+	uint64_t prng_state;
+
+	/* Total number of items. */
+	size_t count;
+
+	/*
+	 * Minimum and current number of hash table buckets.  There are
+	 * 2^LG_CKH_BUCKET_CELLS cells per bucket.
+	 */
+	unsigned lg_minbuckets;
+	unsigned lg_curbuckets;
+
+	/* Hash and comparison functions. */
+	ckh_hash_t *hash;
+	ckh_keycomp_t *keycomp;
+
+	/* Hash table with 2^lg_curbuckets buckets. */
+	ckhc_t *tab;
+} ckh_t;
+
+/******************************************************************************/
+/* BEGIN PUBLIC API */
+/******************************************************************************/
+
+/* Lifetime management.  Minitems is the initial capacity. */
+bool ckh_new(tsd_t *tsd, ckh_t *ckh, size_t minitems, ckh_hash_t *hash,
+    ckh_keycomp_t *keycomp);
+void ckh_delete(tsd_t *tsd, ckh_t *ckh);
+
+/* Get the number of elements in the set. */
+size_t ckh_count(ckh_t *ckh);
+
+/*
+ * To iterate over the elements in the table, initialize *tabind to 0 and call
+ * this function until it returns true.  Each call that returns false will
+ * update *key and *data to the next element in the table, assuming the pointers
+ * are non-NULL.
+ */
+bool ckh_iter(ckh_t *ckh, size_t *tabind, void **key, void **data);
+
+/*
+ * Basic hash table operations -- insert, removal, lookup.  For ckh_remove and
+ * ckh_search, key or data can be NULL.  The hash-table only stores pointers to
+ * the key and value, and doesn't do any lifetime management.
+ */
+bool ckh_insert(tsd_t *tsd, ckh_t *ckh, const void *key, const void *data);
+bool ckh_remove(tsd_t *tsd, ckh_t *ckh, const void *searchkey, void **key,
+    void **data);
+bool ckh_search(ckh_t *ckh, const void *searchkey, void **key, void **data);
+
+/* Some useful hash and comparison functions for strings and pointers. */
+void ckh_string_hash(const void *key, size_t r_hash[2]);
+bool ckh_string_keycomp(const void *k1, const void *k2);
+void ckh_pointer_hash(const void *key, size_t r_hash[2]);
+bool ckh_pointer_keycomp(const void *k1, const void *k2);
+
+#endif /* JEMALLOC_INTERNAL_CKH_H */

include/jemalloc/internal/ctl.h

@@ -0,0 +1,131 @@
+#ifndef JEMALLOC_INTERNAL_CTL_H
+#define JEMALLOC_INTERNAL_CTL_H
+
+#include "jemalloc/internal/jemalloc_internal_types.h"
+#include "jemalloc/internal/malloc_io.h"
+#include "jemalloc/internal/mutex_prof.h"
+#include "jemalloc/internal/ql.h"
+#include "jemalloc/internal/size_classes.h"
+#include "jemalloc/internal/stats.h"
+
+/* Maximum ctl tree depth. */
+#define CTL_MAX_DEPTH	7
+
+typedef struct ctl_node_s {
+	bool named;
+} ctl_node_t;
+
+typedef struct ctl_named_node_s {
+	ctl_node_t node;
+	const char *name;
+	/* If (nchildren == 0), this is a terminal node. */
+	size_t nchildren;
+	const ctl_node_t *children;
+	int (*ctl)(tsd_t *, const size_t *, size_t, void *, size_t *, void *,
+	    size_t);
+} ctl_named_node_t;
+
+typedef struct ctl_indexed_node_s {
+	struct ctl_node_s node;
+	const ctl_named_node_t *(*index)(tsdn_t *, const size_t *, size_t,
+	    size_t);
+} ctl_indexed_node_t;
+
+typedef struct ctl_arena_stats_s {
+	arena_stats_t astats;
+
+	/* Aggregate stats for small size classes, based on bin stats. */
+	size_t allocated_small;
+	uint64_t nmalloc_small;
+	uint64_t ndalloc_small;
+	uint64_t nrequests_small;
+
+	malloc_bin_stats_t bstats[NBINS];
+	malloc_large_stats_t lstats[NSIZES - NBINS];
+} ctl_arena_stats_t;
+
+typedef struct ctl_stats_s {
+	size_t allocated;
+	size_t active;
+	size_t metadata;
+	size_t resident;
+	size_t mapped;
+	size_t retained;
+
+	background_thread_stats_t background_thread;
+	mutex_prof_data_t mutex_prof_data[mutex_prof_num_global_mutexes];
+} ctl_stats_t;
+
+typedef struct ctl_arena_s ctl_arena_t;
+struct ctl_arena_s {
+	unsigned arena_ind;
+	bool initialized;
+	ql_elm(ctl_arena_t) destroyed_link;
+
+	/* Basic stats, supported even if !config_stats. */
+	unsigned nthreads;
+	const char *dss;
+	ssize_t dirty_decay_ms;
+	ssize_t muzzy_decay_ms;
+	size_t pactive;
+	size_t pdirty;
+	size_t pmuzzy;
+
+	/* NULL if !config_stats. */
+	ctl_arena_stats_t *astats;
+};
+
+typedef struct ctl_arenas_s {
+	uint64_t epoch;
+	unsigned narenas;
+	ql_head(ctl_arena_t) destroyed;
+
+	/*
+	 * Element 0 corresponds to merged stats for extant arenas (accessed via
+	 * MALLCTL_ARENAS_ALL), element 1 corresponds to merged stats for
+	 * destroyed arenas (accessed via MALLCTL_ARENAS_DESTROYED), and the
+	 * remaining MALLOCX_ARENA_LIMIT elements correspond to arenas.
+	 */
+	ctl_arena_t *arenas[2 + MALLOCX_ARENA_LIMIT];
+} ctl_arenas_t;
+
+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(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)		\
+	    != 0) {							\
+		malloc_printf(						\
+		    "<jemalloc>: Failure in xmallctl(\"%s\", ...)\n",	\
+		    name);						\
+		abort();						\
+	}								\
+} while (0)
+
+#define xmallctlnametomib(name, mibp, miblenp) do {			\
+	if (je_mallctlnametomib(name, mibp, miblenp) != 0) {		\
+		malloc_printf("<jemalloc>: Failure in "			\
+		    "xmallctlnametomib(\"%s\", ...)\n", name);		\
+		abort();						\
+	}								\
+} while (0)
+
+#define xmallctlbymib(mib, miblen, oldp, oldlenp, newp, newlen) do {	\
+	if (je_mallctlbymib(mib, miblen, oldp, oldlenp, newp,		\
+	    newlen) != 0) {						\
+		malloc_write(						\
+		    "<jemalloc>: Failure in xmallctlbymib()\n");	\
+		abort();						\
+	}								\
+} while (0)
+
+#endif /* JEMALLOC_INTERNAL_CTL_H */

include/jemalloc/internal/extent_dss.h

@@ -0,0 +1,26 @@
+#ifndef JEMALLOC_INTERNAL_EXTENT_DSS_H
+#define JEMALLOC_INTERNAL_EXTENT_DSS_H
+
+typedef enum {
+	dss_prec_disabled  = 0,
+	dss_prec_primary   = 1,
+	dss_prec_secondary = 2,
+
+	dss_prec_limit     = 3
+} dss_prec_t;
+#define DSS_PREC_DEFAULT dss_prec_secondary
+#define DSS_DEFAULT "secondary"
+
+extern const char *dss_prec_names[];
+
+extern const char *opt_dss;
+
+dss_prec_t extent_dss_prec_get(void);
+bool extent_dss_prec_set(dss_prec_t dss_prec);
+void *extent_alloc_dss(tsdn_t *tsdn, arena_t *arena, void *new_addr,
+    size_t size, size_t alignment, bool *zero, bool *commit);
+bool extent_in_dss(void *addr);
+bool extent_dss_mergeable(void *addr_a, void *addr_b);
+void extent_dss_boot(void);
+
+#endif /* JEMALLOC_INTERNAL_EXTENT_DSS_H */

include/jemalloc/internal/extent_externs.h

@@ -0,0 +1,72 @@
+#ifndef JEMALLOC_INTERNAL_EXTENT_EXTERNS_H
+#define JEMALLOC_INTERNAL_EXTENT_EXTERNS_H
+
+#include "jemalloc/internal/mutex.h"
+#include "jemalloc/internal/mutex_pool.h"
+#include "jemalloc/internal/ph.h"
+#include "jemalloc/internal/rb.h"
+#include "jemalloc/internal/rtree.h"
+
+extern rtree_t			extents_rtree;
+extern const extent_hooks_t	extent_hooks_default;
+extern mutex_pool_t		extent_mutex_pool;
+
+extent_t *extent_alloc(tsdn_t *tsdn, arena_t *arena);
+void extent_dalloc(tsdn_t *tsdn, arena_t *arena, extent_t *extent);
+
+extent_hooks_t *extent_hooks_get(arena_t *arena);
+extent_hooks_t *extent_hooks_set(tsd_t *tsd, arena_t *arena,
+    extent_hooks_t *extent_hooks);
+
+#ifdef JEMALLOC_JET
+size_t extent_size_quantize_floor(size_t size);
+size_t extent_size_quantize_ceil(size_t size);
+#endif
+
+rb_proto(, extent_avail_, extent_tree_t, extent_t)
+ph_proto(, extent_heap_, extent_heap_t, extent_t)
+
+bool extents_init(tsdn_t *tsdn, extents_t *extents, extent_state_t state,
+    bool delay_coalesce);
+extent_state_t extents_state_get(const extents_t *extents);
+size_t extents_npages_get(extents_t *extents);
+extent_t *extents_alloc(tsdn_t *tsdn, arena_t *arena,
+    extent_hooks_t **r_extent_hooks, extents_t *extents, void *new_addr,
+    size_t size, size_t pad, size_t alignment, bool slab, szind_t szind,
+    bool *zero, bool *commit);
+void extents_dalloc(tsdn_t *tsdn, arena_t *arena,
+    extent_hooks_t **r_extent_hooks, extents_t *extents, extent_t *extent);
+extent_t *extents_evict(tsdn_t *tsdn, arena_t *arena,
+    extent_hooks_t **r_extent_hooks, extents_t *extents, size_t npages_min);
+void extents_prefork(tsdn_t *tsdn, extents_t *extents);
+void extents_postfork_parent(tsdn_t *tsdn, extents_t *extents);
+void extents_postfork_child(tsdn_t *tsdn, extents_t *extents);
+extent_t *extent_alloc_wrapper(tsdn_t *tsdn, arena_t *arena,
+    extent_hooks_t **r_extent_hooks, void *new_addr, size_t size, size_t pad,
+    size_t alignment, bool slab, szind_t szind, bool *zero, bool *commit);
+void extent_dalloc_gap(tsdn_t *tsdn, arena_t *arena, extent_t *extent);
+void extent_dalloc_wrapper(tsdn_t *tsdn, arena_t *arena,
+    extent_hooks_t **r_extent_hooks, extent_t *extent);
+void extent_destroy_wrapper(tsdn_t *tsdn, arena_t *arena,
+    extent_hooks_t **r_extent_hooks, extent_t *extent);
+bool extent_commit_wrapper(tsdn_t *tsdn, arena_t *arena,
+    extent_hooks_t **r_extent_hooks, extent_t *extent, size_t offset,
+    size_t length);
+bool extent_decommit_wrapper(tsdn_t *tsdn, arena_t *arena,
+    extent_hooks_t **r_extent_hooks, extent_t *extent, size_t offset,
+    size_t length);
+bool extent_purge_lazy_wrapper(tsdn_t *tsdn, arena_t *arena,
+    extent_hooks_t **r_extent_hooks, extent_t *extent, size_t offset,
+    size_t length);
+bool extent_purge_forced_wrapper(tsdn_t *tsdn, arena_t *arena,
+    extent_hooks_t **r_extent_hooks, extent_t *extent, size_t offset,
+    size_t length);
+extent_t *extent_split_wrapper(tsdn_t *tsdn, arena_t *arena,
+    extent_hooks_t **r_extent_hooks, extent_t *extent, size_t size_a,
+    szind_t szind_a, bool slab_a, size_t size_b, szind_t szind_b, bool slab_b);
+bool extent_merge_wrapper(tsdn_t *tsdn, arena_t *arena,
+    extent_hooks_t **r_extent_hooks, extent_t *a, extent_t *b);
+
+bool extent_boot(void);
+
+#endif /* JEMALLOC_INTERNAL_EXTENT_EXTERNS_H */

include/jemalloc/internal/extent_inlines.h

@@ -0,0 +1,407 @@
+#ifndef JEMALLOC_INTERNAL_EXTENT_INLINES_H
+#define JEMALLOC_INTERNAL_EXTENT_INLINES_H
+
+#include "jemalloc/internal/mutex.h"
+#include "jemalloc/internal/mutex_pool.h"
+#include "jemalloc/internal/pages.h"
+#include "jemalloc/internal/prng.h"
+#include "jemalloc/internal/ql.h"
+#include "jemalloc/internal/sz.h"
+
+static inline void
+extent_lock(tsdn_t *tsdn, extent_t *extent) {
+	assert(extent != NULL);
+	mutex_pool_lock(tsdn, &extent_mutex_pool, (uintptr_t)extent);
+}
+
+static inline void
+extent_unlock(tsdn_t *tsdn, extent_t *extent) {
+	assert(extent != NULL);
+	mutex_pool_unlock(tsdn, &extent_mutex_pool, (uintptr_t)extent);
+}
+
+static inline void
+extent_lock2(tsdn_t *tsdn, extent_t *extent1, extent_t *extent2) {
+	assert(extent1 != NULL && extent2 != NULL);
+	mutex_pool_lock2(tsdn, &extent_mutex_pool, (uintptr_t)extent1,
+	    (uintptr_t)extent2);
+}
+
+static inline void
+extent_unlock2(tsdn_t *tsdn, extent_t *extent1, extent_t *extent2) {
+	assert(extent1 != NULL && extent2 != NULL);
+	mutex_pool_unlock2(tsdn, &extent_mutex_pool, (uintptr_t)extent1,
+	    (uintptr_t)extent2);
+}
+
+static inline arena_t *
+extent_arena_get(const extent_t *extent) {
+	unsigned arena_ind = (unsigned)((extent->e_bits &
+	    EXTENT_BITS_ARENA_MASK) >> EXTENT_BITS_ARENA_SHIFT);
+	/*
+	 * The following check is omitted because we should never actually read
+	 * a NULL arena pointer.
+	 */
+	if (false && arena_ind >= MALLOCX_ARENA_LIMIT) {
+		return NULL;
+	}
+	assert(arena_ind < MALLOCX_ARENA_LIMIT);
+	return (arena_t *)atomic_load_p(&arenas[arena_ind], ATOMIC_ACQUIRE);
+}
+
+static inline szind_t
+extent_szind_get_maybe_invalid(const extent_t *extent) {
+	szind_t szind = (szind_t)((extent->e_bits & EXTENT_BITS_SZIND_MASK) >>
+	    EXTENT_BITS_SZIND_SHIFT);
+	assert(szind <= NSIZES);
+	return szind;
+}
+
+static inline szind_t
+extent_szind_get(const extent_t *extent) {
+	szind_t szind = extent_szind_get_maybe_invalid(extent);
+	assert(szind < NSIZES); /* Never call when "invalid". */
+	return szind;
+}
+
+static inline size_t
+extent_usize_get(const extent_t *extent) {
+	return sz_index2size(extent_szind_get(extent));
+}
+
+static inline size_t
+extent_sn_get(const extent_t *extent) {
+	return (size_t)((extent->e_bits & EXTENT_BITS_SN_MASK) >>
+	    EXTENT_BITS_SN_SHIFT);
+}
+
+static inline extent_state_t
+extent_state_get(const extent_t *extent) {
+	return (extent_state_t)((extent->e_bits & EXTENT_BITS_STATE_MASK) >>
+	    EXTENT_BITS_STATE_SHIFT);
+}
+
+static inline bool
+extent_zeroed_get(const extent_t *extent) {
+	return (bool)((extent->e_bits & EXTENT_BITS_ZEROED_MASK) >>
+	    EXTENT_BITS_ZEROED_SHIFT);
+}
+
+static inline bool
+extent_committed_get(const extent_t *extent) {
+	return (bool)((extent->e_bits & EXTENT_BITS_COMMITTED_MASK) >>
+	    EXTENT_BITS_COMMITTED_SHIFT);
+}
+
+static inline bool
+extent_slab_get(const extent_t *extent) {
+	return (bool)((extent->e_bits & EXTENT_BITS_SLAB_MASK) >>
+	    EXTENT_BITS_SLAB_SHIFT);
+}
+
+static inline unsigned
+extent_nfree_get(const extent_t *extent) {
+	assert(extent_slab_get(extent));
+	return (unsigned)((extent->e_bits & EXTENT_BITS_NFREE_MASK) >>
+	    EXTENT_BITS_NFREE_SHIFT);
+}
+
+static inline void *
+extent_base_get(const extent_t *extent) {
+	assert(extent->e_addr == PAGE_ADDR2BASE(extent->e_addr) ||
+	    !extent_slab_get(extent));
+	return PAGE_ADDR2BASE(extent->e_addr);
+}
+
+static inline void *
+extent_addr_get(const extent_t *extent) {
+	assert(extent->e_addr == PAGE_ADDR2BASE(extent->e_addr) ||
+	    !extent_slab_get(extent));
+	return extent->e_addr;
+}
+
+static inline size_t
+extent_size_get(const extent_t *extent) {
+	return (extent->e_size_esn & EXTENT_SIZE_MASK);
+}
+
+static inline size_t
+extent_esn_get(const extent_t *extent) {
+	return (extent->e_size_esn & EXTENT_ESN_MASK);
+}
+
+static inline size_t
+extent_bsize_get(const extent_t *extent) {
+	return extent->e_bsize;
+}
+
+static inline void *
+extent_before_get(const extent_t *extent) {
+	return (void *)((uintptr_t)extent_base_get(extent) - PAGE);
+}
+
+static inline void *
+extent_last_get(const extent_t *extent) {
+	return (void *)((uintptr_t)extent_base_get(extent) +
+	    extent_size_get(extent) - PAGE);
+}
+
+static inline void *
+extent_past_get(const extent_t *extent) {
+	return (void *)((uintptr_t)extent_base_get(extent) +
+	    extent_size_get(extent));
+}
+
+static inline arena_slab_data_t *
+extent_slab_data_get(extent_t *extent) {
+	assert(extent_slab_get(extent));
+	return &extent->e_slab_data;
+}
+
+static inline const arena_slab_data_t *
+extent_slab_data_get_const(const extent_t *extent) {
+	assert(extent_slab_get(extent));
+	return &extent->e_slab_data;
+}
+
+static inline prof_tctx_t *
+extent_prof_tctx_get(const extent_t *extent) {
+	return (prof_tctx_t *)atomic_load_p(&extent->e_prof_tctx,
+	    ATOMIC_ACQUIRE);
+}
+
+static inline void
+extent_arena_set(extent_t *extent, arena_t *arena) {
+	unsigned arena_ind = (arena != NULL) ? arena_ind_get(arena) : ((1U <<
+	    MALLOCX_ARENA_BITS) - 1);
+	extent->e_bits = (extent->e_bits & ~EXTENT_BITS_ARENA_MASK) |
+	    ((uint64_t)arena_ind << EXTENT_BITS_ARENA_SHIFT);
+}
+
+static inline void
+extent_addr_set(extent_t *extent, void *addr) {
+	extent->e_addr = addr;
+}
+
+static inline void
+extent_addr_randomize(tsdn_t *tsdn, extent_t *extent, size_t alignment) {
+	assert(extent_base_get(extent) == extent_addr_get(extent));
+
+	if (alignment < PAGE) {
+		unsigned lg_range = LG_PAGE -
+		    lg_floor(CACHELINE_CEILING(alignment));
+		size_t r =
+		    prng_lg_range_zu(&extent_arena_get(extent)->offset_state,
+		    lg_range, true);
+		uintptr_t random_offset = ((uintptr_t)r) << (LG_PAGE -
+		    lg_range);
+		extent->e_addr = (void *)((uintptr_t)extent->e_addr +
+		    random_offset);
+		assert(ALIGNMENT_ADDR2BASE(extent->e_addr, alignment) ==
+		    extent->e_addr);
+	}
+}
+
+static inline void
+extent_size_set(extent_t *extent, size_t size) {
+	assert((size & ~EXTENT_SIZE_MASK) == 0);
+	extent->e_size_esn = size | (extent->e_size_esn & ~EXTENT_SIZE_MASK);
+}
+
+static inline void
+extent_esn_set(extent_t *extent, size_t esn) {
+	extent->e_size_esn = (extent->e_size_esn & ~EXTENT_ESN_MASK) | (esn &
+	    EXTENT_ESN_MASK);
+}
+
+static inline void
+extent_bsize_set(extent_t *extent, size_t bsize) {
+	extent->e_bsize = bsize;
+}
+
+static inline void
+extent_szind_set(extent_t *extent, szind_t szind) {
+	assert(szind <= NSIZES); /* NSIZES means "invalid". */
+	extent->e_bits = (extent->e_bits & ~EXTENT_BITS_SZIND_MASK) |
+	    ((uint64_t)szind << EXTENT_BITS_SZIND_SHIFT);
+}
+
+static inline void
+extent_nfree_set(extent_t *extent, unsigned nfree) {
+	assert(extent_slab_get(extent));
+	extent->e_bits = (extent->e_bits & ~EXTENT_BITS_NFREE_MASK) |
+	    ((uint64_t)nfree << EXTENT_BITS_NFREE_SHIFT);
+}
+
+static inline void
+extent_nfree_inc(extent_t *extent) {
+	assert(extent_slab_get(extent));
+	extent->e_bits += ((uint64_t)1U << EXTENT_BITS_NFREE_SHIFT);
+}
+
+static inline void
+extent_nfree_dec(extent_t *extent) {
+	assert(extent_slab_get(extent));
+	extent->e_bits -= ((uint64_t)1U << EXTENT_BITS_NFREE_SHIFT);
+}
+
+static inline void
+extent_sn_set(extent_t *extent, size_t sn) {
+	extent->e_bits = (extent->e_bits & ~EXTENT_BITS_SN_MASK) |
+	    ((uint64_t)sn << EXTENT_BITS_SN_SHIFT);
+}
+
+static inline void
+extent_state_set(extent_t *extent, extent_state_t state) {
+	extent->e_bits = (extent->e_bits & ~EXTENT_BITS_STATE_MASK) |
+	    ((uint64_t)state << EXTENT_BITS_STATE_SHIFT);
+}
+
+static inline void
+extent_zeroed_set(extent_t *extent, bool zeroed) {
+	extent->e_bits = (extent->e_bits & ~EXTENT_BITS_ZEROED_MASK) |
+	    ((uint64_t)zeroed << EXTENT_BITS_ZEROED_SHIFT);
+}
+
+static inline void
+extent_committed_set(extent_t *extent, bool committed) {
+	extent->e_bits = (extent->e_bits & ~EXTENT_BITS_COMMITTED_MASK) |
+	    ((uint64_t)committed << EXTENT_BITS_COMMITTED_SHIFT);
+}
+
+static inline void
+extent_slab_set(extent_t *extent, bool slab) {
+	extent->e_bits = (extent->e_bits & ~EXTENT_BITS_SLAB_MASK) |
+	    ((uint64_t)slab << EXTENT_BITS_SLAB_SHIFT);
+}
+
+static inline void
+extent_prof_tctx_set(extent_t *extent, prof_tctx_t *tctx) {
+	atomic_store_p(&extent->e_prof_tctx, tctx, ATOMIC_RELEASE);
+}
+
+static inline void
+extent_init(extent_t *extent, arena_t *arena, void *addr, size_t size,
+    bool slab, szind_t szind, size_t sn, extent_state_t state, bool zeroed,
+    bool committed) {
+	assert(addr == PAGE_ADDR2BASE(addr) || !slab);
+
+	extent_arena_set(extent, arena);
+	extent_addr_set(extent, addr);
+	extent_size_set(extent, size);
+	extent_slab_set(extent, slab);
+	extent_szind_set(extent, szind);
+	extent_sn_set(extent, sn);
+	extent_state_set(extent, state);
+	extent_zeroed_set(extent, zeroed);
+	extent_committed_set(extent, committed);
+	ql_elm_new(extent, ql_link);
+	if (config_prof) {
+		extent_prof_tctx_set(extent, NULL);
+	}
+}
+
+static inline void
+extent_binit(extent_t *extent, void *addr, size_t bsize, size_t sn) {
+	extent_arena_set(extent, NULL);
+	extent_addr_set(extent, addr);
+	extent_bsize_set(extent, bsize);
+	extent_slab_set(extent, false);
+	extent_szind_set(extent, NSIZES);
+	extent_sn_set(extent, sn);
+	extent_state_set(extent, extent_state_active);
+	extent_zeroed_set(extent, true);
+	extent_committed_set(extent, true);
+}
+
+static inline void
+extent_list_init(extent_list_t *list) {
+	ql_new(list);
+}
+
+static inline extent_t *
+extent_list_first(const extent_list_t *list) {
+	return ql_first(list);
+}
+
+static inline extent_t *
+extent_list_last(const extent_list_t *list) {
+	return ql_last(list, ql_link);
+}
+
+static inline void
+extent_list_append(extent_list_t *list, extent_t *extent) {
+	ql_tail_insert(list, extent, ql_link);
+}
+
+static inline void
+extent_list_replace(extent_list_t *list, extent_t *to_remove,
+    extent_t *to_insert) {
+	ql_after_insert(to_remove, to_insert, ql_link);
+	ql_remove(list, to_remove, ql_link);
+}
+
+static inline void
+extent_list_remove(extent_list_t *list, extent_t *extent) {
+	ql_remove(list, extent, ql_link);
+}
+
+static inline int
+extent_sn_comp(const extent_t *a, const extent_t *b) {
+	size_t a_sn = extent_sn_get(a);
+	size_t b_sn = extent_sn_get(b);
+
+	return (a_sn > b_sn) - (a_sn < b_sn);
+}
+
+static inline int
+extent_esn_comp(const extent_t *a, const extent_t *b) {
+	size_t a_esn = extent_esn_get(a);
+	size_t b_esn = extent_esn_get(b);
+
+	return (a_esn > b_esn) - (a_esn < b_esn);
+}
+
+static inline int
+extent_ad_comp(const extent_t *a, const extent_t *b) {
+	uintptr_t a_addr = (uintptr_t)extent_addr_get(a);
+	uintptr_t b_addr = (uintptr_t)extent_addr_get(b);
+
+	return (a_addr > b_addr) - (a_addr < b_addr);
+}
+
+static inline int
+extent_ead_comp(const extent_t *a, const extent_t *b) {
+	uintptr_t a_eaddr = (uintptr_t)a;
+	uintptr_t b_eaddr = (uintptr_t)b;
+
+	return (a_eaddr > b_eaddr) - (a_eaddr < b_eaddr);
+}
+
+static inline int
+extent_snad_comp(const extent_t *a, const extent_t *b) {
+	int ret;
+
+	ret = extent_sn_comp(a, b);
+	if (ret != 0) {
+		return ret;
+	}
+
+	ret = extent_ad_comp(a, b);
+	return ret;
+}
+
+static inline int
+extent_esnead_comp(const extent_t *a, const extent_t *b) {
+	int ret;
+
+	ret = extent_esn_comp(a, b);
+	if (ret != 0) {
+		return ret;
+	}
+
+	ret = extent_ead_comp(a, b);
+	return ret;
+}
+
+#endif /* JEMALLOC_INTERNAL_EXTENT_INLINES_H */

include/jemalloc/internal/extent_mmap.h

@@ -0,0 +1,10 @@
+#ifndef JEMALLOC_INTERNAL_EXTENT_MMAP_EXTERNS_H
+#define JEMALLOC_INTERNAL_EXTENT_MMAP_EXTERNS_H
+
+extern bool opt_retain;
+
+void *extent_alloc_mmap(void *new_addr, size_t size, size_t alignment,
+    bool *zero, bool *commit);
+bool extent_dalloc_mmap(void *addr, size_t size);
+
+#endif /* JEMALLOC_INTERNAL_EXTENT_MMAP_EXTERNS_H */

include/jemalloc/internal/extent_structs.h

@@ -0,0 +1,199 @@
+#ifndef JEMALLOC_INTERNAL_EXTENT_STRUCTS_H
+#define JEMALLOC_INTERNAL_EXTENT_STRUCTS_H
+
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/bitmap.h"
+#include "jemalloc/internal/mutex.h"
+#include "jemalloc/internal/ql.h"
+#include "jemalloc/internal/rb.h"
+#include "jemalloc/internal/ph.h"
+#include "jemalloc/internal/size_classes.h"
+
+typedef enum {
+	extent_state_active   = 0,
+	extent_state_dirty    = 1,
+	extent_state_muzzy    = 2,
+	extent_state_retained = 3
+} extent_state_t;
+
+/* Extent (span of pages).  Use accessor functions for e_* fields. */
+struct extent_s {
+	/*
+	 * Bitfield containing several fields:
+	 *
+	 * a: arena_ind
+	 * b: slab
+	 * c: committed
+	 * z: zeroed
+	 * t: state
+	 * i: szind
+	 * f: nfree
+	 * n: sn
+	 *
+	 * nnnnnnnn ... nnnnnfff fffffffi iiiiiiit tzcbaaaa aaaaaaaa
+	 *
+	 * arena_ind: Arena from which this extent came, or all 1 bits if
+	 *            unassociated.
+	 *
+	 * slab: The slab flag indicates whether the extent is used for a slab
+	 *       of small regions.  This helps differentiate small size classes,
+	 *       and it indicates whether interior pointers can be looked up via
+	 *       iealloc().
+	 *
+	 * committed: The committed flag indicates whether physical memory is
+	 *            committed to the extent, whether explicitly or implicitly
+	 *            as on a system that overcommits and satisfies physical
+	 *            memory needs on demand via soft page faults.
+	 *
+	 * zeroed: The zeroed flag is used by extent recycling code to track
+	 *         whether memory is zero-filled.
+	 *
+	 * state: The state flag is an extent_state_t.
+	 *
+	 * szind: The szind flag indicates usable size class index for
+	 *        allocations residing in this extent, regardless of whether the
+	 *        extent is a slab.  Extent size and usable size often differ
+	 *        even for non-slabs, either due to sz_large_pad or promotion of
+	 *        sampled small regions.
+	 *
+	 * nfree: Number of free regions in slab.
+	 *
+	 * sn: Serial number (potentially non-unique).
+	 *
+	 *     Serial numbers may wrap around if !opt_retain, 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.
+	 */
+	uint64_t		e_bits;
+#define EXTENT_BITS_ARENA_SHIFT		0
+#define EXTENT_BITS_ARENA_MASK \
+    (((uint64_t)(1U << MALLOCX_ARENA_BITS) - 1) << EXTENT_BITS_ARENA_SHIFT)
+
+#define EXTENT_BITS_SLAB_SHIFT		MALLOCX_ARENA_BITS
+#define EXTENT_BITS_SLAB_MASK \
+    ((uint64_t)0x1U << EXTENT_BITS_SLAB_SHIFT)
+
+#define EXTENT_BITS_COMMITTED_SHIFT	(MALLOCX_ARENA_BITS + 1)
+#define EXTENT_BITS_COMMITTED_MASK \
+    ((uint64_t)0x1U << EXTENT_BITS_COMMITTED_SHIFT)
+
+#define EXTENT_BITS_ZEROED_SHIFT	(MALLOCX_ARENA_BITS + 2)
+#define EXTENT_BITS_ZEROED_MASK \
+    ((uint64_t)0x1U << EXTENT_BITS_ZEROED_SHIFT)
+
+#define EXTENT_BITS_STATE_SHIFT		(MALLOCX_ARENA_BITS + 3)
+#define EXTENT_BITS_STATE_MASK \
+    ((uint64_t)0x3U << EXTENT_BITS_STATE_SHIFT)
+
+#define EXTENT_BITS_SZIND_SHIFT		(MALLOCX_ARENA_BITS + 5)
+#define EXTENT_BITS_SZIND_MASK \
+    (((uint64_t)(1U << LG_CEIL_NSIZES) - 1) << EXTENT_BITS_SZIND_SHIFT)
+
+#define EXTENT_BITS_NFREE_SHIFT \
+    (MALLOCX_ARENA_BITS + 5 + LG_CEIL_NSIZES)
+#define EXTENT_BITS_NFREE_MASK \
+    ((uint64_t)((1U << (LG_SLAB_MAXREGS + 1)) - 1) << EXTENT_BITS_NFREE_SHIFT)
+
+#define EXTENT_BITS_SN_SHIFT \
+    (MALLOCX_ARENA_BITS + 5 + LG_CEIL_NSIZES + (LG_SLAB_MAXREGS + 1))
+#define EXTENT_BITS_SN_MASK		(UINT64_MAX << EXTENT_BITS_SN_SHIFT)
+
+	/* Pointer to the extent that this structure is responsible for. */
+	void			*e_addr;
+
+	union {
+		/*
+		 * Extent size and serial number associated with the extent
+		 * structure (different than the serial number for the extent at
+		 * e_addr).
+		 *
+		 * ssssssss [...] ssssssss ssssnnnn nnnnnnnn
+		 */
+		size_t			e_size_esn;
+	#define EXTENT_SIZE_MASK	((size_t)~(PAGE-1))
+	#define EXTENT_ESN_MASK		((size_t)PAGE-1)
+		/* Base extent size, which may not be a multiple of PAGE. */
+		size_t			e_bsize;
+	};
+
+	union {
+		/*
+		 * List linkage, used by a variety of lists:
+		 * - arena_bin_t's slabs_full
+		 * - extents_t's LRU
+		 * - stashed dirty extents
+		 * - arena's large allocations
+		 */
+		ql_elm(extent_t)	ql_link;
+		/* Red-black tree linkage, used by arena's extent_avail. */
+		rb_node(extent_t)	rb_link;
+	};
+
+	/* Linkage for per size class sn/address-ordered heaps. */
+	phn(extent_t)		ph_link;
+
+	union {
+		/* Small region slab metadata. */
+		arena_slab_data_t	e_slab_data;
+
+		/*
+		 * Profile counters, used for large objects.  Points to a
+		 * prof_tctx_t.
+		 */
+		atomic_p_t		e_prof_tctx;
+	};
+};
+typedef ql_head(extent_t) extent_list_t;
+typedef rb_tree(extent_t) extent_tree_t;
+typedef ph(extent_t) extent_heap_t;
+
+/* Quantized collection of extents, with built-in LRU queue. */
+struct extents_s {
+	malloc_mutex_t		mtx;
+
+	/*
+	 * Quantized per size class heaps of extents.
+	 *
+	 * Synchronization: mtx.
+	 */
+	extent_heap_t		heaps[NPSIZES+1];
+
+	/*
+	 * Bitmap for which set bits correspond to non-empty heaps.
+	 *
+	 * Synchronization: mtx.
+	 */
+	bitmap_t		bitmap[BITMAP_GROUPS(NPSIZES+1)];
+
+	/*
+	 * LRU of all extents in heaps.
+	 *
+	 * Synchronization: mtx.
+	 */
+	extent_list_t		lru;
+
+	/*
+	 * Page sum for all extents in heaps.
+	 *
+	 * The synchronization here is a little tricky.  Modifications to npages
+	 * must hold mtx, but reads need not (though, a reader who sees npages
+	 * without holding the mutex can't assume anything about the rest of the
+	 * state of the extents_t).
+	 */
+	atomic_zu_t		npages;
+
+	/* All stored extents must be in the same state. */
+	extent_state_t		state;
+
+	/*
+	 * If true, delay coalescing until eviction; otherwise coalesce during
+	 * deallocation.
+	 */
+	bool			delay_coalesce;
+};
+
+#endif /* JEMALLOC_INTERNAL_EXTENT_STRUCTS_H */

include/jemalloc/internal/extent_types.h

@@ -0,0 +1,9 @@
+#ifndef JEMALLOC_INTERNAL_EXTENT_TYPES_H
+#define JEMALLOC_INTERNAL_EXTENT_TYPES_H
+
+typedef struct extent_s extent_t;
+typedef struct extents_s extents_t;
+
+#define EXTENT_HOOKS_INITIALIZER	NULL
+
+#endif /* JEMALLOC_INTERNAL_EXTENT_TYPES_H */

include/jemalloc/internal/hash.h

@@ -0,0 +1,318 @@
+#ifndef JEMALLOC_INTERNAL_HASH_H
+#define JEMALLOC_INTERNAL_HASH_H
+
+#include "jemalloc/internal/assert.h"
+
+/*
+ * The following hash function is based on MurmurHash3, placed into the public
+ * domain by Austin Appleby.  See https://github.com/aappleby/smhasher for
+ * details.
+ */
+
+/******************************************************************************/
+/* Internal implementation. */
+static inline uint32_t
+hash_rotl_32(uint32_t x, int8_t r) {
+	return ((x << r) | (x >> (32 - r)));
+}
+
+static inline uint64_t
+hash_rotl_64(uint64_t x, int8_t r) {
+	return ((x << r) | (x >> (64 - r)));
+}
+
+static inline uint32_t
+hash_get_block_32(const uint32_t *p, int i) {
+	/* Handle unaligned read. */
+	if (unlikely((uintptr_t)p & (sizeof(uint32_t)-1)) != 0) {
+		uint32_t ret;
+
+		memcpy(&ret, (uint8_t *)(p + i), sizeof(uint32_t));
+		return ret;
+	}
+
+	return p[i];
+}
+
+static inline uint64_t
+hash_get_block_64(const uint64_t *p, int i) {
+	/* Handle unaligned read. */
+	if (unlikely((uintptr_t)p & (sizeof(uint64_t)-1)) != 0) {
+		uint64_t ret;
+
+		memcpy(&ret, (uint8_t *)(p + i), sizeof(uint64_t));
+		return ret;
+	}
+
+	return p[i];
+}
+
+static inline uint32_t
+hash_fmix_32(uint32_t h) {
+	h ^= h >> 16;
+	h *= 0x85ebca6b;
+	h ^= h >> 13;
+	h *= 0xc2b2ae35;
+	h ^= h >> 16;
+
+	return h;
+}
+
+static inline uint64_t
+hash_fmix_64(uint64_t k) {
+	k ^= k >> 33;
+	k *= KQU(0xff51afd7ed558ccd);
+	k ^= k >> 33;
+	k *= KQU(0xc4ceb9fe1a85ec53);
+	k ^= k >> 33;
+
+	return k;
+}
+
+static inline uint32_t
+hash_x86_32(const void *key, int len, uint32_t seed) {
+	const uint8_t *data = (const uint8_t *) key;
+	const int nblocks = len / 4;
+
+	uint32_t h1 = seed;
+
+	const uint32_t c1 = 0xcc9e2d51;
+	const uint32_t c2 = 0x1b873593;
+
+	/* body */
+	{
+		const uint32_t *blocks = (const uint32_t *) (data + nblocks*4);
+		int i;
+
+		for (i = -nblocks; i; i++) {
+			uint32_t k1 = hash_get_block_32(blocks, i);
+
+			k1 *= c1;
+			k1 = hash_rotl_32(k1, 15);
+			k1 *= c2;
+
+			h1 ^= k1;
+			h1 = hash_rotl_32(h1, 13);
+			h1 = h1*5 + 0xe6546b64;
+		}
+	}
+
+	/* tail */
+	{
+		const uint8_t *tail = (const uint8_t *) (data + nblocks*4);
+
+		uint32_t k1 = 0;
+
+		switch (len & 3) {
+		case 3: k1 ^= tail[2] << 16;
+		case 2: k1 ^= tail[1] << 8;
+		case 1: k1 ^= tail[0]; k1 *= c1; k1 = hash_rotl_32(k1, 15);
+			k1 *= c2; h1 ^= k1;
+		}
+	}
+
+	/* finalization */
+	h1 ^= len;
+
+	h1 = hash_fmix_32(h1);
+
+	return h1;
+}
+
+UNUSED static inline void
+hash_x86_128(const void *key, const int len, uint32_t seed,
+    uint64_t r_out[2]) {
+	const uint8_t * data = (const uint8_t *) key;
+	const int nblocks = len / 16;
+
+	uint32_t h1 = seed;
+	uint32_t h2 = seed;
+	uint32_t h3 = seed;
+	uint32_t h4 = seed;
+
+	const uint32_t c1 = 0x239b961b;
+	const uint32_t c2 = 0xab0e9789;
+	const uint32_t c3 = 0x38b34ae5;
+	const uint32_t c4 = 0xa1e38b93;
+
+	/* body */
+	{
+		const uint32_t *blocks = (const uint32_t *) (data + nblocks*16);
+		int i;
+
+		for (i = -nblocks; i; i++) {
+			uint32_t k1 = hash_get_block_32(blocks, i*4 + 0);
+			uint32_t k2 = hash_get_block_32(blocks, i*4 + 1);
+			uint32_t k3 = hash_get_block_32(blocks, i*4 + 2);
+			uint32_t k4 = hash_get_block_32(blocks, i*4 + 3);
+
+			k1 *= c1; k1 = hash_rotl_32(k1, 15); k1 *= c2; h1 ^= k1;
+
+			h1 = hash_rotl_32(h1, 19); h1 += h2;
+			h1 = h1*5 + 0x561ccd1b;
+
+			k2 *= c2; k2 = hash_rotl_32(k2, 16); k2 *= c3; h2 ^= k2;
+
+			h2 = hash_rotl_32(h2, 17); h2 += h3;
+			h2 = h2*5 + 0x0bcaa747;
+
+			k3 *= c3; k3 = hash_rotl_32(k3, 17); k3 *= c4; h3 ^= k3;
+
+			h3 = hash_rotl_32(h3, 15); h3 += h4;
+			h3 = h3*5 + 0x96cd1c35;
+
+			k4 *= c4; k4 = hash_rotl_32(k4, 18); k4 *= c1; h4 ^= k4;
+
+			h4 = hash_rotl_32(h4, 13); h4 += h1;
+			h4 = h4*5 + 0x32ac3b17;
+		}
+	}
+
+	/* tail */
+	{
+		const uint8_t *tail = (const uint8_t *) (data + nblocks*16);
+		uint32_t k1 = 0;
+		uint32_t k2 = 0;
+		uint32_t k3 = 0;
+		uint32_t k4 = 0;
+
+		switch (len & 15) {
+		case 15: k4 ^= tail[14] << 16;
+		case 14: k4 ^= tail[13] << 8;
+		case 13: k4 ^= tail[12] << 0;
+			k4 *= c4; k4 = hash_rotl_32(k4, 18); k4 *= c1; h4 ^= k4;
+
+		case 12: k3 ^= tail[11] << 24;
+		case 11: k3 ^= tail[10] << 16;
+		case 10: k3 ^= tail[ 9] << 8;
+		case  9: k3 ^= tail[ 8] << 0;
+		     k3 *= c3; k3 = hash_rotl_32(k3, 17); k3 *= c4; h3 ^= k3;
+
+		case  8: k2 ^= tail[ 7] << 24;
+		case  7: k2 ^= tail[ 6] << 16;
+		case  6: k2 ^= tail[ 5] << 8;
+		case  5: k2 ^= tail[ 4] << 0;
+			k2 *= c2; k2 = hash_rotl_32(k2, 16); k2 *= c3; h2 ^= k2;
+
+		case  4: k1 ^= tail[ 3] << 24;
+		case  3: k1 ^= tail[ 2] << 16;
+		case  2: k1 ^= tail[ 1] << 8;
+		case  1: k1 ^= tail[ 0] << 0;
+			k1 *= c1; k1 = hash_rotl_32(k1, 15); k1 *= c2; h1 ^= k1;
+		}
+	}
+
+	/* finalization */
+	h1 ^= len; h2 ^= len; h3 ^= len; h4 ^= len;
+
+	h1 += h2; h1 += h3; h1 += h4;
+	h2 += h1; h3 += h1; h4 += h1;
+
+	h1 = hash_fmix_32(h1);
+	h2 = hash_fmix_32(h2);
+	h3 = hash_fmix_32(h3);
+	h4 = hash_fmix_32(h4);
+
+	h1 += h2; h1 += h3; h1 += h4;
+	h2 += h1; h3 += h1; h4 += h1;
+
+	r_out[0] = (((uint64_t) h2) << 32) | h1;
+	r_out[1] = (((uint64_t) h4) << 32) | h3;
+}
+
+UNUSED static inline void
+hash_x64_128(const void *key, const int len, const uint32_t seed,
+    uint64_t r_out[2]) {
+	const uint8_t *data = (const uint8_t *) key;
+	const int nblocks = len / 16;
+
+	uint64_t h1 = seed;
+	uint64_t h2 = seed;
+
+	const uint64_t c1 = KQU(0x87c37b91114253d5);
+	const uint64_t c2 = KQU(0x4cf5ad432745937f);
+
+	/* body */
+	{
+		const uint64_t *blocks = (const uint64_t *) (data);
+		int i;
+
+		for (i = 0; i < nblocks; i++) {
+			uint64_t k1 = hash_get_block_64(blocks, i*2 + 0);
+			uint64_t k2 = hash_get_block_64(blocks, i*2 + 1);
+
+			k1 *= c1; k1 = hash_rotl_64(k1, 31); k1 *= c2; h1 ^= k1;
+
+			h1 = hash_rotl_64(h1, 27); h1 += h2;
+			h1 = h1*5 + 0x52dce729;
+
+			k2 *= c2; k2 = hash_rotl_64(k2, 33); k2 *= c1; h2 ^= k2;
+
+			h2 = hash_rotl_64(h2, 31); h2 += h1;
+			h2 = h2*5 + 0x38495ab5;
+		}
+	}
+
+	/* tail */
+	{
+		const uint8_t *tail = (const uint8_t*)(data + nblocks*16);
+		uint64_t k1 = 0;
+		uint64_t k2 = 0;
+
+		switch (len & 15) {
+		case 15: k2 ^= ((uint64_t)(tail[14])) << 48;
+		case 14: k2 ^= ((uint64_t)(tail[13])) << 40;
+		case 13: k2 ^= ((uint64_t)(tail[12])) << 32;
+		case 12: k2 ^= ((uint64_t)(tail[11])) << 24;
+		case 11: k2 ^= ((uint64_t)(tail[10])) << 16;
+		case 10: k2 ^= ((uint64_t)(tail[ 9])) << 8;
+		case  9: k2 ^= ((uint64_t)(tail[ 8])) << 0;
+			k2 *= c2; k2 = hash_rotl_64(k2, 33); k2 *= c1; h2 ^= k2;
+
+		case  8: k1 ^= ((uint64_t)(tail[ 7])) << 56;
+		case  7: k1 ^= ((uint64_t)(tail[ 6])) << 48;
+		case  6: k1 ^= ((uint64_t)(tail[ 5])) << 40;
+		case  5: k1 ^= ((uint64_t)(tail[ 4])) << 32;
+		case  4: k1 ^= ((uint64_t)(tail[ 3])) << 24;
+		case  3: k1 ^= ((uint64_t)(tail[ 2])) << 16;
+		case  2: k1 ^= ((uint64_t)(tail[ 1])) << 8;
+		case  1: k1 ^= ((uint64_t)(tail[ 0])) << 0;
+			k1 *= c1; k1 = hash_rotl_64(k1, 31); k1 *= c2; h1 ^= k1;
+		}
+	}
+
+	/* finalization */
+	h1 ^= len; h2 ^= len;
+
+	h1 += h2;
+	h2 += h1;
+
+	h1 = hash_fmix_64(h1);
+	h2 = hash_fmix_64(h2);
+
+	h1 += h2;
+	h2 += h1;
+
+	r_out[0] = h1;
+	r_out[1] = h2;
+}
+
+/******************************************************************************/
+/* API. */
+static inline void
+hash(const void *key, size_t len, const uint32_t seed, size_t r_hash[2]) {
+	assert(len <= INT_MAX); /* Unfortunate implementation limitation. */
+
+#if (LG_SIZEOF_PTR == 3 && !defined(JEMALLOC_BIG_ENDIAN))
+	hash_x64_128(key, (int)len, seed, (uint64_t *)r_hash);
+#else
+	{
+		uint64_t hashes[2];
+		hash_x86_128(key, (int)len, seed, hashes);
+		r_hash[0] = (size_t)hashes[0];
+		r_hash[1] = (size_t)hashes[1];
+	}
+#endif
+}
+
+#endif /* JEMALLOC_INTERNAL_HASH_H */

include/jemalloc/internal/hooks.h

@@ -0,0 +1,19 @@
+#ifndef JEMALLOC_INTERNAL_HOOKS_H
+#define JEMALLOC_INTERNAL_HOOKS_H
+
+extern JEMALLOC_EXPORT void (*hooks_arena_new_hook)();
+extern JEMALLOC_EXPORT void (*hooks_libc_hook)();
+
+#define JEMALLOC_HOOK(fn, hook) ((void)(hook != NULL && (hook(), 0)), fn)
+
+#define open JEMALLOC_HOOK(open, hooks_libc_hook)
+#define read JEMALLOC_HOOK(read, hooks_libc_hook)
+#define write JEMALLOC_HOOK(write, hooks_libc_hook)
+#define readlink JEMALLOC_HOOK(readlink, hooks_libc_hook)
+#define close JEMALLOC_HOOK(close, hooks_libc_hook)
+#define creat JEMALLOC_HOOK(creat, hooks_libc_hook)
+#define secure_getenv JEMALLOC_HOOK(secure_getenv, hooks_libc_hook)
+/* Note that this is undef'd and re-define'd in src/prof.c. */
+#define _Unwind_Backtrace JEMALLOC_HOOK(_Unwind_Backtrace, hooks_libc_hook)
+
+#endif /* JEMALLOC_INTERNAL_HOOKS_H */

include/jemalloc/internal/jemalloc_internal_decls.h

@@ -0,0 +1,81 @@
+#ifndef JEMALLOC_INTERNAL_DECLS_H
+#define JEMALLOC_INTERNAL_DECLS_H
+
+#include <math.h>
+#ifdef _WIN32
+#  include <windows.h>
+#  include "msvc_compat/windows_extra.h"
+
+#else
+#  include <sys/param.h>
+#  include <sys/mman.h>
+#  if !defined(__pnacl__) && !defined(__native_client__)
+#    include <sys/syscall.h>
+#    if !defined(SYS_write) && defined(__NR_write)
+#      define SYS_write __NR_write
+#    endif
+#    if defined(SYS_open) && defined(__aarch64__)
+       /* Android headers may define SYS_open to __NR_open even though
+        * __NR_open may not exist on AArch64 (superseded by __NR_openat). */
+#      undef SYS_open
+#    endif
+#    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>
+
+#include <limits.h>
+#ifndef SIZE_T_MAX
+#  define SIZE_T_MAX	SIZE_MAX
+#endif
+#ifndef SSIZE_MAX
+#  define SSIZE_MAX	((ssize_t)(SIZE_T_MAX >> 1))
+#endif
+#include <stdarg.h>
+#include <stdbool.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdint.h>
+#include <stddef.h>
+#ifndef offsetof
+#  define offsetof(type, member)	((size_t)&(((type *)NULL)->member))
+#endif
+#include <string.h>
+#include <strings.h>
+#include <ctype.h>
+#ifdef _MSC_VER
+#  include <io.h>
+typedef intptr_t ssize_t;
+#  define PATH_MAX 1024
+#  define STDERR_FILENO 2
+#  define __func__ __FUNCTION__
+#  ifdef JEMALLOC_HAS_RESTRICT
+#    define restrict __restrict
+#  endif
+/* Disable warnings about deprecated system functions. */
+#  pragma warning(disable: 4996)
+#if _MSC_VER < 1800
+static int
+isblank(int c) {
+	return (c == '\t' || c == ' ');
+}
+#endif
+#else
+#  include <unistd.h>
+#endif
+#include <fcntl.h>
+
+#endif /* JEMALLOC_INTERNAL_H */

include/jemalloc/internal/jemalloc_internal_defs.h.in

@@ -0,0 +1,336 @@
+#ifndef JEMALLOC_INTERNAL_DEFS_H_
+#define JEMALLOC_INTERNAL_DEFS_H_
+/*
+ * If JEMALLOC_PREFIX is defined via --with-jemalloc-prefix, it will cause all
+ * public APIs to be prefixed.  This makes it possible, with some care, to use
+ * multiple allocators simultaneously.
+ */
+#undef JEMALLOC_PREFIX
+#undef JEMALLOC_CPREFIX
+
+/*
+ * Define overrides for non-standard allocator-related functions if they are
+ * present on the system.
+ */
+#undef JEMALLOC_OVERRIDE___LIBC_CALLOC
+#undef JEMALLOC_OVERRIDE___LIBC_FREE
+#undef JEMALLOC_OVERRIDE___LIBC_MALLOC
+#undef JEMALLOC_OVERRIDE___LIBC_MEMALIGN
+#undef JEMALLOC_OVERRIDE___LIBC_REALLOC
+#undef JEMALLOC_OVERRIDE___LIBC_VALLOC
+#undef JEMALLOC_OVERRIDE___POSIX_MEMALIGN
+
+/*
+ * JEMALLOC_PRIVATE_NAMESPACE is used as a prefix for all library-private APIs.
+ * For shared libraries, symbol visibility mechanisms prevent these symbols
+ * from being exported, but for static libraries, naming collisions are a real
+ * possibility.
+ */
+#undef JEMALLOC_PRIVATE_NAMESPACE
+
+/*
+ * Hyper-threaded CPUs may need a special instruction inside spin loops in
+ * order to yield to another virtual CPU.
+ */
+#undef CPU_SPINWAIT
+
+/*
+ * Number of significant bits in virtual addresses.  This may be less than the
+ * total number of bits in a pointer, e.g. on x64, for which the uppermost 16
+ * bits are the same as bit 47.
+ */
+#undef LG_VADDR
+
+/* Defined if C11 atomics are available. */
+#undef JEMALLOC_C11_ATOMICS
+
+/* Defined if GCC __atomic atomics are available. */
+#undef JEMALLOC_GCC_ATOMIC_ATOMICS
+
+/* Defined if GCC __sync atomics are available. */
+#undef JEMALLOC_GCC_SYNC_ATOMICS
+
+/*
+ * Defined if __sync_add_and_fetch(uint32_t *, uint32_t) and
+ * __sync_sub_and_fetch(uint32_t *, uint32_t) are available, despite
+ * __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4 not being defined (which means the
+ * functions are defined in libgcc instead of being inlines).
+ */
+#undef JE_FORCE_SYNC_COMPARE_AND_SWAP_4
+
+/*
+ * Defined if __sync_add_and_fetch(uint64_t *, uint64_t) and
+ * __sync_sub_and_fetch(uint64_t *, uint64_t) are available, despite
+ * __GCC_HAVE_SYNC_COMPARE_AND_SWAP_8 not being defined (which means the
+ * functions are defined in libgcc instead of being inlines).
+ */
+#undef JE_FORCE_SYNC_COMPARE_AND_SWAP_8
+
+/*
+ * Defined if __builtin_clz() and __builtin_clzl() are available.
+ */
+#undef JEMALLOC_HAVE_BUILTIN_CLZ
+
+/*
+ * Defined if os_unfair_lock_*() functions are available, as provided by Darwin.
+ */
+#undef JEMALLOC_OS_UNFAIR_LOCK
+
+/*
+ * Defined if OSSpin*() functions are available, as provided by Darwin, and
+ * documented in the spinlock(3) manual page.
+ */
+#undef JEMALLOC_OSSPIN
+
+/* Defined if syscall(2) is usable. */
+#undef JEMALLOC_USE_SYSCALL
+
+/*
+ * Defined if secure_getenv(3) is available.
+ */
+#undef JEMALLOC_HAVE_SECURE_GETENV
+
+/*
+ * Defined if issetugid(2) is available.
+ */
+#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
+ * bootstrapping will cause recursion into the pthreads library.  Therefore, if
+ * _malloc_thread_cleanup() exists, use it as the basis for thread cleanup in
+ * malloc_tsd.
+ */
+#undef JEMALLOC_MALLOC_THREAD_CLEANUP
+
+/*
+ * Defined if threaded initialization is known to be safe on this platform.
+ * Among other things, it must be possible to initialize a mutex without
+ * triggering allocation in order for threaded allocation to be safe.
+ */
+#undef JEMALLOC_THREADED_INIT
+
+/*
+ * Defined if the pthreads implementation defines
+ * _pthread_mutex_init_calloc_cb(), in which case the function is used in order
+ * to avoid recursive allocation during mutex initialization.
+ */
+#undef JEMALLOC_MUTEX_INIT_CB
+
+/* Non-empty if the tls_model attribute is supported. */
+#undef JEMALLOC_TLS_MODEL
+
+/*
+ * JEMALLOC_DEBUG enables assertions and other sanity checks, and disables
+ * inline functions.
+ */
+#undef JEMALLOC_DEBUG
+
+/* JEMALLOC_STATS enables statistics calculation. */
+#undef JEMALLOC_STATS
+
+/* JEMALLOC_PROF enables allocation profiling. */
+#undef JEMALLOC_PROF
+
+/* Use libunwind for profile backtracing if defined. */
+#undef JEMALLOC_PROF_LIBUNWIND
+
+/* Use libgcc for profile backtracing if defined. */
+#undef JEMALLOC_PROF_LIBGCC
+
+/* Use gcc intrinsics for profile backtracing if defined. */
+#undef JEMALLOC_PROF_GCC
+
+/*
+ * JEMALLOC_DSS enables use of sbrk(2) to allocate extents from the data storage
+ * segment (DSS).
+ */
+#undef JEMALLOC_DSS
+
+/* Support memory filling (junk/zero). */
+#undef JEMALLOC_FILL
+
+/* Support utrace(2)-based tracing. */
+#undef JEMALLOC_UTRACE
+
+/* Support optional abort() on OOM. */
+#undef JEMALLOC_XMALLOC
+
+/* Support lazy locking (avoid locking unless a second thread is launched). */
+#undef JEMALLOC_LAZY_LOCK
+
+/*
+ * Minimum allocation alignment is 2^LG_QUANTUM bytes (ignoring tiny size
+ * classes).
+ */
+#undef LG_QUANTUM
+
+/* One page is 2^LG_PAGE bytes. */
+#undef LG_PAGE
+
+/*
+ * One huge page is 2^LG_HUGEPAGE bytes.  Note that this is defined even if the
+ * system does not explicitly support huge pages; system calls that require
+ * explicit huge page support are separately configured.
+ */
+#undef LG_HUGEPAGE
+
+/*
+ * If defined, adjacent virtual memory mappings with identical attributes
+ * automatically coalesce, and they fragment when changes are made to subranges.
+ * This is the normal order of things for mmap()/munmap(), but on Windows
+ * VirtualAlloc()/VirtualFree() operations must be precisely matched, i.e.
+ * mappings do *not* coalesce/fragment.
+ */
+#undef JEMALLOC_MAPS_COALESCE
+
+/*
+ * If defined, retain memory for later reuse by default rather than using e.g.
+ * munmap() to unmap freed extents.  This is enabled on 64-bit Linux because
+ * common sequences of mmap()/munmap() calls will cause virtual memory map
+ * holes.
+ */
+#undef JEMALLOC_RETAIN
+
+/* 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.
+ */
+#undef JEMALLOC_INTERNAL_FFSLL
+#undef JEMALLOC_INTERNAL_FFSL
+#undef JEMALLOC_INTERNAL_FFS
+
+/*
+ * If defined, explicitly attempt to more uniformly distribute large allocation
+ * pointer alignments across all cache indices.
+ */
+#undef JEMALLOC_CACHE_OBLIVIOUS
+
+/*
+ * Darwin (OS X) uses zones to work around Mach-O symbol override shortcomings.
+ */
+#undef JEMALLOC_ZONE
+
+/*
+ * 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
+
+/*
+ * Methods for purging unused pages differ between operating systems.
+ *
+ *   madvise(..., MADV_FREE) : This marks pages as being unused, such that they
+ *                             will be discarded rather than swapped out.
+ *   madvise(..., MADV_DONTNEED) : If JEMALLOC_PURGE_MADVISE_DONTNEED_ZEROS is
+ *                                 defined, this immediately discards pages,
+ *                                 such that new pages will be demand-zeroed if
+ *                                 the address region is later touched;
+ *                                 otherwise this behaves similarly to
+ *                                 MADV_FREE, though typically with higher
+ *                                 system overhead.
+ */
+#undef JEMALLOC_PURGE_MADVISE_FREE
+#undef JEMALLOC_PURGE_MADVISE_DONTNEED
+#undef JEMALLOC_PURGE_MADVISE_DONTNEED_ZEROS
+
+/*
+ * Defined if transparent huge pages (THPs) are supported via the
+ * MADV_[NO]HUGEPAGE arguments to madvise(2), and THP support is enabled.
+ */
+#undef JEMALLOC_THP
+
+/* Define if operating system has alloca.h header. */
+#undef JEMALLOC_HAS_ALLOCA_H
+
+/* C99 restrict keyword supported. */
+#undef JEMALLOC_HAS_RESTRICT
+
+/* For use by hash code. */
+#undef JEMALLOC_BIG_ENDIAN
+
+/* sizeof(int) == 2^LG_SIZEOF_INT. */
+#undef LG_SIZEOF_INT
+
+/* sizeof(long) == 2^LG_SIZEOF_LONG. */
+#undef LG_SIZEOF_LONG
+
+/* sizeof(long long) == 2^LG_SIZEOF_LONG_LONG. */
+#undef LG_SIZEOF_LONG_LONG
+
+/* sizeof(intmax_t) == 2^LG_SIZEOF_INTMAX_T. */
+#undef LG_SIZEOF_INTMAX_T
+
+/* glibc malloc hooks (__malloc_hook, __realloc_hook, __free_hook). */
+#undef JEMALLOC_GLIBC_MALLOC_HOOK
+
+/* glibc memalign hook. */
+#undef JEMALLOC_GLIBC_MEMALIGN_HOOK
+
+/* pthread support */
+#undef JEMALLOC_HAVE_PTHREAD
+
+/* dlsym() support */
+#undef JEMALLOC_HAVE_DLSYM
+
+/* Adaptive mutex support in pthreads. */
+#undef JEMALLOC_HAVE_PTHREAD_MUTEX_ADAPTIVE_NP
+
+/* GNU specific sched_getcpu support */
+#undef JEMALLOC_HAVE_SCHED_GETCPU
+
+/* GNU specific sched_setaffinity support */
+#undef JEMALLOC_HAVE_SCHED_SETAFFINITY
+
+/*
+ * If defined, all the features necessary for background threads are present.
+ */
+#undef JEMALLOC_BACKGROUND_THREAD
+
+/*
+ * If defined, jemalloc symbols are not exported (doesn't work when
+ * JEMALLOC_PREFIX is not defined).
+ */
+#undef JEMALLOC_EXPORT
+
+/* config.malloc_conf options string. */
+#undef JEMALLOC_CONFIG_MALLOC_CONF
+
+/* If defined, jemalloc takes the malloc/free/etc. symbol names. */
+#undef JEMALLOC_IS_MALLOC
+
+#endif /* JEMALLOC_INTERNAL_DEFS_H_ */

include/jemalloc/internal/jemalloc_internal_externs.h

@@ -0,0 +1,53 @@
+#ifndef JEMALLOC_INTERNAL_EXTERNS_H
+#define JEMALLOC_INTERNAL_EXTERNS_H
+
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/size_classes.h"
+#include "jemalloc/internal/tsd_types.h"
+
+/* TSD checks this to set thread local slow state accordingly. */
+extern bool malloc_slow;
+
+/* Run-time options. */
+extern bool opt_abort;
+extern bool opt_abort_conf;
+extern const char *opt_junk;
+extern bool opt_junk_alloc;
+extern bool opt_junk_free;
+extern bool opt_utrace;
+extern bool opt_xmalloc;
+extern bool opt_zero;
+extern unsigned opt_narenas;
+
+/* 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 atomic_p_t arenas[];
+
+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);
+void arena_set(unsigned ind, arena_t *arena);
+unsigned narenas_total_get(void);
+arena_t *arena_init(tsdn_t *tsdn, unsigned ind, extent_hooks_t *extent_hooks);
+arena_tdata_t *arena_tdata_get_hard(tsd_t *tsd, unsigned ind);
+arena_t *arena_choose_hard(tsd_t *tsd, bool internal);
+void arena_migrate(tsd_t *tsd, unsigned oldind, unsigned newind);
+void iarena_cleanup(tsd_t *tsd);
+void arena_cleanup(tsd_t *tsd);
+void arenas_tdata_cleanup(tsd_t *tsd);
+void jemalloc_prefork(void);
+void jemalloc_postfork_parent(void);
+void jemalloc_postfork_child(void);
+bool malloc_initialized(void);
+
+#endif /* JEMALLOC_INTERNAL_EXTERNS_H */

include/jemalloc/internal/jemalloc_internal_includes.h

@@ -0,0 +1,94 @@
+#ifndef JEMALLOC_INTERNAL_INCLUDES_H
+#define JEMALLOC_INTERNAL_INCLUDES_H
+
+/*
+ * jemalloc can conceptually be broken into components (arena, tcache, etc.),
+ * but there are circular dependencies that cannot be broken without
+ * substantial performance degradation.
+ *
+ * Historically, we dealt with this by each header into four sections (types,
+ * structs, externs, and inlines), and included each header file multiple times
+ * in this file, picking out the portion we want on each pass using the
+ * following #defines:
+ *   JEMALLOC_H_TYPES   : Preprocessor-defined constants and psuedo-opaque data
+ *                        types.
+ *   JEMALLOC_H_STRUCTS : Data structures.
+ *   JEMALLOC_H_EXTERNS : Extern data declarations and function prototypes.
+ *   JEMALLOC_H_INLINES : Inline functions.
+ *
+ * We're moving toward a world in which the dependencies are explicit; each file
+ * will #include the headers it depends on (rather than relying on them being
+ * implicitly available via this file including every header file in the
+ * project).
+ *
+ * We're now in an intermediate state: we've broken up the header files to avoid
+ * having to include each one multiple times, but have not yet moved the
+ * dependency information into the header files (i.e. we still rely on the
+ * ordering in this file to ensure all a header's dependencies are available in
+ * its translation unit).  Each component is now broken up into multiple header
+ * files, corresponding to the sections above (e.g. instead of "foo.h", we now
+ * have "foo_types.h", "foo_structs.h", "foo_externs.h", "foo_inlines.h").
+ *
+ * Those files which have been converted to explicitly include their
+ * inter-component dependencies are now in the initial HERMETIC HEADERS
+ * section.  All headers may still rely on jemalloc_preamble.h (which, by fiat,
+ * must be included first in every translation unit) for system headers and
+ * global jemalloc definitions, however.
+ */
+
+/******************************************************************************/
+/* TYPES */
+/******************************************************************************/
+
+#include "jemalloc/internal/extent_types.h"
+#include "jemalloc/internal/base_types.h"
+#include "jemalloc/internal/arena_types.h"
+#include "jemalloc/internal/tcache_types.h"
+#include "jemalloc/internal/prof_types.h"
+
+/******************************************************************************/
+/* STRUCTS */
+/******************************************************************************/
+
+#include "jemalloc/internal/arena_structs_a.h"
+#include "jemalloc/internal/extent_structs.h"
+#include "jemalloc/internal/base_structs.h"
+#include "jemalloc/internal/prof_structs.h"
+#include "jemalloc/internal/arena_structs_b.h"
+#include "jemalloc/internal/tcache_structs.h"
+#include "jemalloc/internal/background_thread_structs.h"
+
+/******************************************************************************/
+/* EXTERNS */
+/******************************************************************************/
+
+#include "jemalloc/internal/jemalloc_internal_externs.h"
+#include "jemalloc/internal/extent_externs.h"
+#include "jemalloc/internal/base_externs.h"
+#include "jemalloc/internal/arena_externs.h"
+#include "jemalloc/internal/large_externs.h"
+#include "jemalloc/internal/tcache_externs.h"
+#include "jemalloc/internal/prof_externs.h"
+#include "jemalloc/internal/background_thread_externs.h"
+
+/******************************************************************************/
+/* INLINES */
+/******************************************************************************/
+
+#include "jemalloc/internal/jemalloc_internal_inlines_a.h"
+#include "jemalloc/internal/base_inlines.h"
+/*
+ * Include portions of arena code interleaved with tcache code in order to
+ * resolve circular dependencies.
+ */
+#include "jemalloc/internal/prof_inlines_a.h"
+#include "jemalloc/internal/arena_inlines_a.h"
+#include "jemalloc/internal/extent_inlines.h"
+#include "jemalloc/internal/jemalloc_internal_inlines_b.h"
+#include "jemalloc/internal/tcache_inlines.h"
+#include "jemalloc/internal/arena_inlines_b.h"
+#include "jemalloc/internal/jemalloc_internal_inlines_c.h"
+#include "jemalloc/internal/prof_inlines_b.h"
+#include "jemalloc/internal/background_thread_inlines.h"
+
+#endif /* JEMALLOC_INTERNAL_INCLUDES_H */

include/jemalloc/internal/jemalloc_internal_inlines_a.h

@@ -0,0 +1,168 @@
+#ifndef JEMALLOC_INTERNAL_INLINES_A_H
+#define JEMALLOC_INTERNAL_INLINES_A_H
+
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/bit_util.h"
+#include "jemalloc/internal/jemalloc_internal_types.h"
+#include "jemalloc/internal/size_classes.h"
+#include "jemalloc/internal/ticker.h"
+
+JEMALLOC_ALWAYS_INLINE malloc_cpuid_t
+malloc_getcpu(void) {
+	assert(have_percpu_arena);
+#if defined(JEMALLOC_HAVE_SCHED_GETCPU)
+	return (malloc_cpuid_t)sched_getcpu();
+#else
+	not_reached();
+	return -1;
+#endif
+}
+
+/* Return the chosen arena index based on current cpu. */
+JEMALLOC_ALWAYS_INLINE unsigned
+percpu_arena_choose(void) {
+	assert(have_percpu_arena && PERCPU_ARENA_ENABLED(opt_percpu_arena));
+
+	malloc_cpuid_t cpuid = malloc_getcpu();
+	assert(cpuid >= 0);
+
+	unsigned arena_ind;
+	if ((opt_percpu_arena == percpu_arena) || ((unsigned)cpuid < ncpus /
+	    2)) {
+		arena_ind = cpuid;
+	} else {
+		assert(opt_percpu_arena == per_phycpu_arena);
+		/* Hyper threads on the same physical CPU share arena. */
+		arena_ind = cpuid - ncpus / 2;
+	}
+
+	return arena_ind;
+}
+
+/* Return the limit of percpu auto arena range, i.e. arenas[0...ind_limit). */
+JEMALLOC_ALWAYS_INLINE unsigned
+percpu_arena_ind_limit(percpu_arena_mode_t mode) {
+	assert(have_percpu_arena && PERCPU_ARENA_ENABLED(mode));
+	if (mode == per_phycpu_arena && ncpus > 1) {
+		if (ncpus % 2) {
+			/* This likely means a misconfig. */
+			return ncpus / 2 + 1;
+		}
+		return ncpus / 2;
+	} else {
+		return ncpus;
+	}
+}
+
+static inline arena_tdata_t *
+arena_tdata_get(tsd_t *tsd, unsigned ind, bool refresh_if_missing) {
+	arena_tdata_t *tdata;
+	arena_tdata_t *arenas_tdata = tsd_arenas_tdata_get(tsd);
+
+	if (unlikely(arenas_tdata == NULL)) {
+		/* arenas_tdata hasn't been initialized yet. */
+		return arena_tdata_get_hard(tsd, ind);
+	}
+	if (unlikely(ind >= tsd_narenas_tdata_get(tsd))) {
+		/*
+		 * ind is invalid, cache is old (too small), or tdata to be
+		 * initialized.
+		 */
+		return (refresh_if_missing ? arena_tdata_get_hard(tsd, ind) :
+		    NULL);
+	}
+
+	tdata = &arenas_tdata[ind];
+	if (likely(tdata != NULL) || !refresh_if_missing) {
+		return tdata;
+	}
+	return arena_tdata_get_hard(tsd, ind);
+}
+
+static inline arena_t *
+arena_get(tsdn_t *tsdn, unsigned ind, bool init_if_missing) {
+	arena_t *ret;
+
+	assert(ind < MALLOCX_ARENA_LIMIT);
+
+	ret = (arena_t *)atomic_load_p(&arenas[ind], ATOMIC_ACQUIRE);
+	if (unlikely(ret == NULL)) {
+		if (init_if_missing) {
+			ret = arena_init(tsdn, ind,
+			    (extent_hooks_t *)&extent_hooks_default);
+		}
+	}
+	return ret;
+}
+
+static inline ticker_t *
+decay_ticker_get(tsd_t *tsd, unsigned ind) {
+	arena_tdata_t *tdata;
+
+	tdata = arena_tdata_get(tsd, ind, true);
+	if (unlikely(tdata == NULL)) {
+		return NULL;
+	}
+	return &tdata->decay_ticker;
+}
+
+JEMALLOC_ALWAYS_INLINE tcache_bin_t *
+tcache_small_bin_get(tcache_t *tcache, szind_t binind) {
+	assert(binind < NBINS);
+	return &tcache->tbins_small[binind];
+}
+
+JEMALLOC_ALWAYS_INLINE tcache_bin_t *
+tcache_large_bin_get(tcache_t *tcache, szind_t binind) {
+	assert(binind >= NBINS &&binind < nhbins);
+	return &tcache->tbins_large[binind - NBINS];
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+tcache_available(tsd_t *tsd) {
+	/*
+	 * Thread specific auto tcache might be unavailable if: 1) during tcache
+	 * initialization, or 2) disabled through thread.tcache.enabled mallctl
+	 * or config options.  This check covers all cases.
+	 */
+	if (likely(tsd_tcache_enabled_get(tsd))) {
+		/* Associated arena == NULL implies tcache init in progress. */
+		assert(tsd_tcachep_get(tsd)->arena == NULL ||
+		    tcache_small_bin_get(tsd_tcachep_get(tsd), 0)->avail !=
+		    NULL);
+		return true;
+	}
+
+	return false;
+}
+
+JEMALLOC_ALWAYS_INLINE tcache_t *
+tcache_get(tsd_t *tsd) {
+	if (!tcache_available(tsd)) {
+		return NULL;
+	}
+
+	return tsd_tcachep_get(tsd);
+}
+
+static inline void
+pre_reentrancy(tsd_t *tsd) {
+	bool fast = tsd_fast(tsd);
+	++*tsd_reentrancy_levelp_get(tsd);
+	if (fast) {
+		/* Prepare slow path for reentrancy. */
+		tsd_slow_update(tsd);
+		assert(tsd->state == tsd_state_nominal_slow);
+	}
+}
+
+static inline void
+post_reentrancy(tsd_t *tsd) {
+	int8_t *reentrancy_level = tsd_reentrancy_levelp_get(tsd);
+	assert(*reentrancy_level > 0);
+	if (--*reentrancy_level == 0) {
+		tsd_slow_update(tsd);
+	}
+}
+
+#endif /* JEMALLOC_INTERNAL_INLINES_A_H */

include/jemalloc/internal/jemalloc_internal_inlines_b.h

@@ -0,0 +1,86 @@
+#ifndef JEMALLOC_INTERNAL_INLINES_B_H
+#define JEMALLOC_INTERNAL_INLINES_B_H
+
+#include "jemalloc/internal/rtree.h"
+
+/* Choose an arena based on a per-thread value. */
+static inline arena_t *
+arena_choose_impl(tsd_t *tsd, arena_t *arena, bool internal) {
+	arena_t *ret;
+
+	if (arena != NULL) {
+		return arena;
+	}
+
+	/* During reentrancy, arena 0 is the safest bet. */
+	if (unlikely(tsd_reentrancy_level_get(tsd) > 0)) {
+		return arena_get(tsd_tsdn(tsd), 0, true);
+	}
+
+	ret = internal ? tsd_iarena_get(tsd) : tsd_arena_get(tsd);
+	if (unlikely(ret == NULL)) {
+		ret = arena_choose_hard(tsd, internal);
+		assert(ret);
+		if (tcache_available(tsd)) {
+			tcache_t *tcache = tcache_get(tsd);
+			if (tcache->arena != NULL) {
+				/* See comments in tcache_data_init().*/
+				assert(tcache->arena ==
+				    arena_get(tsd_tsdn(tsd), 0, false));
+				if (tcache->arena != ret) {
+					tcache_arena_reassociate(tsd_tsdn(tsd),
+					    tcache, ret);
+				}
+			} else {
+				tcache_arena_associate(tsd_tsdn(tsd), tcache,
+				    ret);
+			}
+		}
+	}
+
+	/*
+	 * Note that for percpu arena, if the current arena is outside of the
+	 * auto percpu arena range, (i.e. thread is assigned to a manually
+	 * managed arena), then percpu arena is skipped.
+	 */
+	if (have_percpu_arena && PERCPU_ARENA_ENABLED(opt_percpu_arena) &&
+	    !internal && (arena_ind_get(ret) <
+	    percpu_arena_ind_limit(opt_percpu_arena)) && (ret->last_thd !=
+	    tsd_tsdn(tsd))) {
+		unsigned ind = percpu_arena_choose();
+		if (arena_ind_get(ret) != ind) {
+			percpu_arena_update(tsd, ind);
+			ret = tsd_arena_get(tsd);
+		}
+		ret->last_thd = tsd_tsdn(tsd);
+	}
+
+	return ret;
+}
+
+static inline arena_t *
+arena_choose(tsd_t *tsd, arena_t *arena) {
+	return arena_choose_impl(tsd, arena, false);
+}
+
+static inline arena_t *
+arena_ichoose(tsd_t *tsd, arena_t *arena) {
+	return arena_choose_impl(tsd, arena, true);
+}
+
+static inline bool
+arena_is_auto(arena_t *arena) {
+	assert(narenas_auto > 0);
+	return (arena_ind_get(arena) < narenas_auto);
+}
+
+JEMALLOC_ALWAYS_INLINE extent_t *
+iealloc(tsdn_t *tsdn, const void *ptr) {
+	rtree_ctx_t rtree_ctx_fallback;
+	rtree_ctx_t *rtree_ctx = tsdn_rtree_ctx(tsdn, &rtree_ctx_fallback);
+
+	return rtree_extent_read(tsdn, &extents_rtree, rtree_ctx,
+	    (uintptr_t)ptr, true);
+}
+
+#endif /* JEMALLOC_INTERNAL_INLINES_B_H */

include/jemalloc/internal/jemalloc_internal_inlines_c.h

@@ -0,0 +1,197 @@
+#ifndef JEMALLOC_INTERNAL_INLINES_C_H
+#define JEMALLOC_INTERNAL_INLINES_C_H
+
+#include "jemalloc/internal/jemalloc_internal_types.h"
+#include "jemalloc/internal/sz.h"
+#include "jemalloc/internal/witness.h"
+
+JEMALLOC_ALWAYS_INLINE arena_t *
+iaalloc(tsdn_t *tsdn, const void *ptr) {
+	assert(ptr != NULL);
+
+	return arena_aalloc(tsdn, ptr);
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+isalloc(tsdn_t *tsdn, const void *ptr) {
+	assert(ptr != NULL);
+
+	return arena_salloc(tsdn, ptr);
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+iallocztm(tsdn_t *tsdn, size_t size, szind_t ind, bool zero, tcache_t *tcache,
+    bool is_internal, arena_t *arena, bool slow_path) {
+	void *ret;
+
+	assert(size != 0);
+	assert(!is_internal || tcache == NULL);
+	assert(!is_internal || arena == NULL || arena_is_auto(arena));
+	witness_assert_depth_to_rank(tsdn_witness_tsdp_get(tsdn),
+	    WITNESS_RANK_CORE, 0);
+
+	ret = arena_malloc(tsdn, arena, size, ind, zero, tcache, slow_path);
+	if (config_stats && is_internal && likely(ret != NULL)) {
+		arena_internal_add(iaalloc(tsdn, ret), isalloc(tsdn, ret));
+	}
+	return ret;
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+ialloc(tsd_t *tsd, size_t size, szind_t ind, bool zero, bool slow_path) {
+	return iallocztm(tsd_tsdn(tsd), size, ind, zero, tcache_get(tsd), false,
+	    NULL, slow_path);
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+ipallocztm(tsdn_t *tsdn, size_t usize, size_t alignment, bool zero,
+    tcache_t *tcache, bool is_internal, arena_t *arena) {
+	void *ret;
+
+	assert(usize != 0);
+	assert(usize == sz_sa2u(usize, alignment));
+	assert(!is_internal || tcache == NULL);
+	assert(!is_internal || arena == NULL || arena_is_auto(arena));
+	witness_assert_depth_to_rank(tsdn_witness_tsdp_get(tsdn),
+	    WITNESS_RANK_CORE, 0);
+
+	ret = arena_palloc(tsdn, arena, usize, alignment, zero, tcache);
+	assert(ALIGNMENT_ADDR2BASE(ret, alignment) == ret);
+	if (config_stats && is_internal && likely(ret != NULL)) {
+		arena_internal_add(iaalloc(tsdn, ret), isalloc(tsdn, ret));
+	}
+	return ret;
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+ipalloct(tsdn_t *tsdn, size_t usize, size_t alignment, bool zero,
+    tcache_t *tcache, arena_t *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_tsdn(tsd), usize, alignment, zero,
+	    tcache_get(tsd), false, NULL);
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+ivsalloc(tsdn_t *tsdn, const void *ptr) {
+	return arena_vsalloc(tsdn, ptr);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+idalloctm(tsdn_t *tsdn, void *ptr, tcache_t *tcache, alloc_ctx_t *alloc_ctx,
+    bool is_internal, bool slow_path) {
+	assert(ptr != NULL);
+	assert(!is_internal || tcache == NULL);
+	assert(!is_internal || arena_is_auto(iaalloc(tsdn, ptr)));
+	witness_assert_depth_to_rank(tsdn_witness_tsdp_get(tsdn),
+	    WITNESS_RANK_CORE, 0);
+	if (config_stats && is_internal) {
+		arena_internal_sub(iaalloc(tsdn, ptr), isalloc(tsdn, ptr));
+	}
+	if (!is_internal && tsd_reentrancy_level_get(tsdn_tsd(tsdn)) != 0) {
+		assert(tcache == NULL);
+	}
+	arena_dalloc(tsdn, ptr, tcache, alloc_ctx, slow_path);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+idalloc(tsd_t *tsd, void *ptr) {
+	idalloctm(tsd_tsdn(tsd), ptr, tcache_get(tsd), NULL, false, true);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+isdalloct(tsdn_t *tsdn, void *ptr, size_t size, tcache_t *tcache,
+    alloc_ctx_t *alloc_ctx, bool slow_path) {
+	witness_assert_depth_to_rank(tsdn_witness_tsdp_get(tsdn),
+	    WITNESS_RANK_CORE, 0);
+	arena_sdalloc(tsdn, ptr, size, tcache, alloc_ctx, slow_path);
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+iralloct_realign(tsdn_t *tsdn, void *ptr, size_t oldsize, size_t size,
+    size_t extra, size_t alignment, bool zero, tcache_t *tcache,
+    arena_t *arena) {
+	witness_assert_depth_to_rank(tsdn_witness_tsdp_get(tsdn),
+	    WITNESS_RANK_CORE, 0);
+	void *p;
+	size_t usize, copysize;
+
+	usize = sz_sa2u(size + extra, alignment);
+	if (unlikely(usize == 0 || usize > LARGE_MAXCLASS)) {
+		return NULL;
+	}
+	p = ipalloct(tsdn, usize, alignment, zero, tcache, arena);
+	if (p == NULL) {
+		if (extra == 0) {
+			return NULL;
+		}
+		/* Try again, without extra this time. */
+		usize = sz_sa2u(size, alignment);
+		if (unlikely(usize == 0 || usize > LARGE_MAXCLASS)) {
+			return NULL;
+		}
+		p = ipalloct(tsdn, usize, alignment, zero, tcache, arena);
+		if (p == NULL) {
+			return NULL;
+		}
+	}
+	/*
+	 * Copy at most size bytes (not size+extra), since the caller has no
+	 * expectation that the extra bytes will be reliably preserved.
+	 */
+	copysize = (size < oldsize) ? size : oldsize;
+	memcpy(p, ptr, copysize);
+	isdalloct(tsdn, ptr, oldsize, tcache, NULL, true);
+	return p;
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+iralloct(tsdn_t *tsdn, void *ptr, size_t oldsize, size_t size, size_t alignment,
+    bool zero, tcache_t *tcache, arena_t *arena) {
+	assert(ptr != NULL);
+	assert(size != 0);
+	witness_assert_depth_to_rank(tsdn_witness_tsdp_get(tsdn),
+	    WITNESS_RANK_CORE, 0);
+
+	if (alignment != 0 && ((uintptr_t)ptr & ((uintptr_t)alignment-1))
+	    != 0) {
+		/*
+		 * Existing object alignment is inadequate; allocate new space
+		 * and copy.
+		 */
+		return iralloct_realign(tsdn, ptr, oldsize, size, 0, alignment,
+		    zero, tcache, arena);
+	}
+
+	return arena_ralloc(tsdn, arena, ptr, oldsize, size, alignment, zero,
+	    tcache);
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+iralloc(tsd_t *tsd, void *ptr, size_t oldsize, size_t size, size_t alignment,
+    bool zero) {
+	return iralloct(tsd_tsdn(tsd), ptr, oldsize, size, alignment, zero,
+	    tcache_get(tsd), NULL);
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+ixalloc(tsdn_t *tsdn, void *ptr, size_t oldsize, size_t size, size_t extra,
+    size_t alignment, bool zero) {
+	assert(ptr != NULL);
+	assert(size != 0);
+	witness_assert_depth_to_rank(tsdn_witness_tsdp_get(tsdn),
+	    WITNESS_RANK_CORE, 0);
+
+	if (alignment != 0 && ((uintptr_t)ptr & ((uintptr_t)alignment-1))
+	    != 0) {
+		/* Existing object alignment is inadequate. */
+		return true;
+	}
+
+	return arena_ralloc_no_move(tsdn, ptr, oldsize, size, extra, zero);
+}
+
+#endif /* JEMALLOC_INTERNAL_INLINES_C_H */

include/jemalloc/internal/jemalloc_internal_macros.h

@@ -0,0 +1,40 @@
+#ifndef JEMALLOC_INTERNAL_MACROS_H
+#define JEMALLOC_INTERNAL_MACROS_H
+
+#ifdef JEMALLOC_DEBUG
+#  define JEMALLOC_ALWAYS_INLINE static inline
+#else
+#  define JEMALLOC_ALWAYS_INLINE JEMALLOC_ATTR(always_inline) static inline
+#endif
+#ifdef _MSC_VER
+#  define inline _inline
+#endif
+
+#define UNUSED JEMALLOC_ATTR(unused)
+
+#define ZU(z)	((size_t)z)
+#define ZD(z)	((ssize_t)z)
+#define QU(q)	((uint64_t)q)
+#define QD(q)	((int64_t)q)
+
+#define KZU(z)	ZU(z##ULL)
+#define KZD(z)	ZD(z##LL)
+#define KQU(q)	QU(q##ULL)
+#define KQD(q)	QI(q##LL)
+
+#ifndef __DECONST
+#  define	__DECONST(type, var)	((type)(uintptr_t)(const void *)(var))
+#endif
+
+#if !defined(JEMALLOC_HAS_RESTRICT) || defined(__cplusplus)
+#  define restrict
+#endif
+
+/* Various function pointers are statick and immutable except during testing. */
+#ifdef JEMALLOC_JET
+#  define JET_MUTABLE
+#else
+#  define JET_MUTABLE const
+#endif
+
+#endif /* JEMALLOC_INTERNAL_MACROS_H */

include/jemalloc/internal/jemalloc_internal_types.h

@@ -0,0 +1,178 @@
+#ifndef JEMALLOC_INTERNAL_TYPES_H
+#define JEMALLOC_INTERNAL_TYPES_H
+
+/* Page size index type. */
+typedef unsigned pszind_t;
+
+/* Size class index type. */
+typedef unsigned szind_t;
+
+/* Processor / core id type. */
+typedef int malloc_cpuid_t;
+
+/*
+ * Flags bits:
+ *
+ * a: arena
+ * t: tcache
+ * 0: unused
+ * z: zero
+ * n: alignment
+ *
+ * aaaaaaaa aaaatttt tttttttt 0znnnnnn
+ */
+#define MALLOCX_ARENA_BITS	12
+#define MALLOCX_TCACHE_BITS	12
+#define MALLOCX_LG_ALIGN_BITS	6
+#define MALLOCX_ARENA_SHIFT	20
+#define MALLOCX_TCACHE_SHIFT	8
+#define MALLOCX_ARENA_MASK \
+    (((1 << MALLOCX_ARENA_BITS) - 1) << MALLOCX_ARENA_SHIFT)
+/* NB: Arena index bias decreases the maximum number of arenas by 1. */
+#define MALLOCX_ARENA_LIMIT	((1 << MALLOCX_ARENA_BITS) - 1)
+#define MALLOCX_TCACHE_MASK \
+    (((1 << MALLOCX_TCACHE_BITS) - 1) << MALLOCX_TCACHE_SHIFT)
+#define MALLOCX_TCACHE_MAX	((1 << MALLOCX_TCACHE_BITS) - 3)
+#define MALLOCX_LG_ALIGN_MASK	((1 << MALLOCX_LG_ALIGN_BITS) - 1)
+/* Use MALLOCX_ALIGN_GET() if alignment may not be specified in flags. */
+#define MALLOCX_ALIGN_GET_SPECIFIED(flags)				\
+    (ZU(1) << (flags & MALLOCX_LG_ALIGN_MASK))
+#define MALLOCX_ALIGN_GET(flags)					\
+    (MALLOCX_ALIGN_GET_SPECIFIED(flags) & (SIZE_T_MAX-1))
+#define MALLOCX_ZERO_GET(flags)						\
+    ((bool)(flags & MALLOCX_ZERO))
+
+#define MALLOCX_TCACHE_GET(flags)					\
+    (((unsigned)((flags & MALLOCX_TCACHE_MASK) >> MALLOCX_TCACHE_SHIFT)) - 2)
+#define MALLOCX_ARENA_GET(flags)					\
+    (((unsigned)(((unsigned)flags) >> MALLOCX_ARENA_SHIFT)) - 1)
+
+/* Smallest size class to support. */
+#define TINY_MIN		(1U << LG_TINY_MIN)
+
+/*
+ * Minimum allocation alignment is 2^LG_QUANTUM bytes (ignoring tiny size
+ * classes).
+ */
+#ifndef LG_QUANTUM
+#  if (defined(__i386__) || defined(_M_IX86))
+#    define LG_QUANTUM		4
+#  endif
+#  ifdef __ia64__
+#    define LG_QUANTUM		4
+#  endif
+#  ifdef __alpha__
+#    define LG_QUANTUM		4
+#  endif
+#  if (defined(__sparc64__) || defined(__sparcv9) || defined(__sparc_v9__))
+#    define LG_QUANTUM		4
+#  endif
+#  if (defined(__amd64__) || defined(__x86_64__) || defined(_M_X64))
+#    define LG_QUANTUM		4
+#  endif
+#  ifdef __arm__
+#    define LG_QUANTUM		3
+#  endif
+#  ifdef __aarch64__
+#    define LG_QUANTUM		4
+#  endif
+#  ifdef __hppa__
+#    define LG_QUANTUM		4
+#  endif
+#  ifdef __mips__
+#    define LG_QUANTUM		3
+#  endif
+#  ifdef __or1k__
+#    define LG_QUANTUM		3
+#  endif
+#  ifdef __powerpc__
+#    define LG_QUANTUM		4
+#  endif
+#  ifdef __riscv__
+#    define LG_QUANTUM		4
+#  endif
+#  ifdef __s390__
+#    define LG_QUANTUM		4
+#  endif
+#  ifdef __SH4__
+#    define LG_QUANTUM		4
+#  endif
+#  ifdef __tile__
+#    define LG_QUANTUM		4
+#  endif
+#  ifdef __le32__
+#    define LG_QUANTUM		4
+#  endif
+#  ifndef LG_QUANTUM
+#    error "Unknown minimum alignment for architecture; specify via "
+	 "--with-lg-quantum"
+#  endif
+#endif
+
+#define QUANTUM			((size_t)(1U << LG_QUANTUM))
+#define QUANTUM_MASK		(QUANTUM - 1)
+
+/* Return the smallest quantum multiple that is >= a. */
+#define QUANTUM_CEILING(a)						\
+	(((a) + QUANTUM_MASK) & ~QUANTUM_MASK)
+
+#define LONG			((size_t)(1U << LG_SIZEOF_LONG))
+#define LONG_MASK		(LONG - 1)
+
+/* Return the smallest long multiple that is >= a. */
+#define LONG_CEILING(a)							\
+	(((a) + LONG_MASK) & ~LONG_MASK)
+
+#define SIZEOF_PTR		(1U << LG_SIZEOF_PTR)
+#define PTR_MASK		(SIZEOF_PTR - 1)
+
+/* Return the smallest (void *) multiple that is >= a. */
+#define PTR_CEILING(a)							\
+	(((a) + PTR_MASK) & ~PTR_MASK)
+
+/*
+ * Maximum size of L1 cache line.  This is used to avoid cache line aliasing.
+ * In addition, this controls the spacing of cacheline-spaced size classes.
+ *
+ * CACHELINE cannot be based on LG_CACHELINE because __declspec(align()) can
+ * only handle raw constants.
+ */
+#define LG_CACHELINE		6
+#define CACHELINE		64
+#define CACHELINE_MASK		(CACHELINE - 1)
+
+/* Return the smallest cacheline multiple that is >= s. */
+#define CACHELINE_CEILING(s)						\
+	(((s) + CACHELINE_MASK) & ~CACHELINE_MASK)
+
+/* Return the nearest aligned address at or below a. */
+#define ALIGNMENT_ADDR2BASE(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)				\
+	((size_t)((uintptr_t)(a) & (alignment - 1)))
+
+/* Return the smallest alignment multiple that is >= s. */
+#define ALIGNMENT_CEILING(s, alignment)					\
+	(((s) + (alignment - 1)) & ((~(alignment)) + 1))
+
+/* Declare a variable-length array. */
+#if __STDC_VERSION__ < 199901L
+#  ifdef _MSC_VER
+#    include <malloc.h>
+#    define alloca _alloca
+#  else
+#    ifdef JEMALLOC_HAS_ALLOCA_H
+#      include <alloca.h>
+#    else
+#      include <stdlib.h>
+#    endif
+#  endif
+#  define VARIABLE_ARRAY(type, name, count) \
+	type *name = alloca(sizeof(type) * (count))
+#else
+#  define VARIABLE_ARRAY(type, name, count) type name[(count)]
+#endif
+
+#endif /* JEMALLOC_INTERNAL_TYPES_H */

include/jemalloc/internal/jemalloc_preamble.h.in

@@ -0,0 +1,179 @@
+#ifndef JEMALLOC_PREAMBLE_H
+#define JEMALLOC_PREAMBLE_H
+
+#include "jemalloc_internal_defs.h"
+#include "jemalloc/internal/jemalloc_internal_decls.h"
+
+#ifdef JEMALLOC_UTRACE
+#include <sys/ktrace.h>
+#endif
+
+#define JEMALLOC_NO_DEMANGLE
+#ifdef JEMALLOC_JET
+#  undef JEMALLOC_IS_MALLOC
+#  define JEMALLOC_N(n) jet_##n
+#  include "jemalloc/internal/public_namespace.h"
+#  define JEMALLOC_NO_RENAME
+#  include "../jemalloc@install_suffix@.h"
+#  undef JEMALLOC_NO_RENAME
+#else
+#  define JEMALLOC_N(n) @private_namespace@##n
+#  include "../jemalloc@install_suffix@.h"
+#endif
+
+#if (defined(JEMALLOC_OSATOMIC) || defined(JEMALLOC_OSSPIN))
+#include <libkern/OSAtomic.h>
+#endif
+
+#ifdef JEMALLOC_ZONE
+#include <mach/mach_error.h>
+#include <mach/mach_init.h>
+#include <mach/vm_map.h>
+#endif
+
+#include "jemalloc/internal/jemalloc_internal_macros.h"
+
+/*
+ * Note that the ordering matters here; the hook itself is name-mangled.  We
+ * want the inclusion of hooks to happen early, so that we hook as much as
+ * possible.
+ */
+#ifndef JEMALLOC_NO_PRIVATE_NAMESPACE
+#  ifndef JEMALLOC_JET
+#    include "jemalloc/internal/private_namespace.h"
+#  else
+#    include "jemalloc/internal/private_namespace_jet.h"
+#  endif
+#endif
+#include "jemalloc/internal/hooks.h"
+
+static const bool config_debug =
+#ifdef JEMALLOC_DEBUG
+    true
+#else
+    false
+#endif
+    ;
+static const bool have_dss =
+#ifdef JEMALLOC_DSS
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_fill =
+#ifdef JEMALLOC_FILL
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_lazy_lock =
+#ifdef JEMALLOC_LAZY_LOCK
+    true
+#else
+    false
+#endif
+    ;
+static const char * const config_malloc_conf = JEMALLOC_CONFIG_MALLOC_CONF;
+static const bool config_prof =
+#ifdef JEMALLOC_PROF
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_prof_libgcc =
+#ifdef JEMALLOC_PROF_LIBGCC
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_prof_libunwind =
+#ifdef JEMALLOC_PROF_LIBUNWIND
+    true
+#else
+    false
+#endif
+    ;
+static const bool maps_coalesce =
+#ifdef JEMALLOC_MAPS_COALESCE
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_stats =
+#ifdef JEMALLOC_STATS
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_thp =
+#ifdef JEMALLOC_THP
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_tls =
+#ifdef JEMALLOC_TLS
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_utrace =
+#ifdef JEMALLOC_UTRACE
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_xmalloc =
+#ifdef JEMALLOC_XMALLOC
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_cache_oblivious =
+#ifdef JEMALLOC_CACHE_OBLIVIOUS
+    true
+#else
+    false
+#endif
+    ;
+#ifdef JEMALLOC_HAVE_SCHED_GETCPU
+/* Currently percpu_arena depends on sched_getcpu. */
+#define JEMALLOC_PERCPU_ARENA
+#endif
+static const bool have_percpu_arena =
+#ifdef JEMALLOC_PERCPU_ARENA
+    true
+#else
+    false
+#endif
+    ;
+/*
+ * Undocumented, and not recommended; the application should take full
+ * responsibility for tracking provenance.
+ */
+static const bool force_ivsalloc =
+#ifdef JEMALLOC_FORCE_IVSALLOC
+    true
+#else
+    false
+#endif
+    ;
+static const bool have_background_thread =
+#ifdef JEMALLOC_BACKGROUND_THREAD
+    true
+#else
+    false
+#endif
+    ;
+
+#endif /* JEMALLOC_PREAMBLE_H */

include/jemalloc/internal/large_externs.h

@@ -0,0 +1,26 @@
+#ifndef JEMALLOC_INTERNAL_LARGE_EXTERNS_H
+#define JEMALLOC_INTERNAL_LARGE_EXTERNS_H
+
+void *large_malloc(tsdn_t *tsdn, arena_t *arena, size_t usize, bool zero);
+void *large_palloc(tsdn_t *tsdn, arena_t *arena, size_t usize, size_t alignment,
+    bool zero);
+bool large_ralloc_no_move(tsdn_t *tsdn, extent_t *extent, size_t usize_min,
+    size_t usize_max, bool zero);
+void *large_ralloc(tsdn_t *tsdn, arena_t *arena, extent_t *extent, size_t usize,
+    size_t alignment, bool zero, tcache_t *tcache);
+
+typedef void (large_dalloc_junk_t)(void *, size_t);
+extern large_dalloc_junk_t *JET_MUTABLE large_dalloc_junk;
+
+typedef void (large_dalloc_maybe_junk_t)(void *, size_t);
+extern large_dalloc_maybe_junk_t *JET_MUTABLE large_dalloc_maybe_junk;
+
+void large_dalloc_prep_junked_locked(tsdn_t *tsdn, extent_t *extent);
+void large_dalloc_finish(tsdn_t *tsdn, extent_t *extent);
+void large_dalloc(tsdn_t *tsdn, extent_t *extent);
+size_t large_salloc(tsdn_t *tsdn, const extent_t *extent);
+prof_tctx_t *large_prof_tctx_get(tsdn_t *tsdn, const extent_t *extent);
+void large_prof_tctx_set(tsdn_t *tsdn, extent_t *extent, prof_tctx_t *tctx);
+void large_prof_tctx_reset(tsdn_t *tsdn, extent_t *extent);
+
+#endif /* JEMALLOC_INTERNAL_LARGE_EXTERNS_H */

include/jemalloc/internal/malloc_io.h

@@ -0,0 +1,62 @@
+#ifndef JEMALLOC_INTERNAL_MALLOC_IO_H
+#define JEMALLOC_INTERNAL_MALLOC_IO_H
+
+#ifdef _WIN32
+#  ifdef _WIN64
+#    define FMT64_PREFIX "ll"
+#    define FMTPTR_PREFIX "ll"
+#  else
+#    define FMT64_PREFIX "ll"
+#    define FMTPTR_PREFIX ""
+#  endif
+#  define FMTd32 "d"
+#  define FMTu32 "u"
+#  define FMTx32 "x"
+#  define FMTd64 FMT64_PREFIX "d"
+#  define FMTu64 FMT64_PREFIX "u"
+#  define FMTx64 FMT64_PREFIX "x"
+#  define FMTdPTR FMTPTR_PREFIX "d"
+#  define FMTuPTR FMTPTR_PREFIX "u"
+#  define FMTxPTR FMTPTR_PREFIX "x"
+#else
+#  include <inttypes.h>
+#  define FMTd32 PRId32
+#  define FMTu32 PRIu32
+#  define FMTx32 PRIx32
+#  define FMTd64 PRId64
+#  define FMTu64 PRIu64
+#  define FMTx64 PRIx64
+#  define FMTdPTR PRIdPTR
+#  define FMTuPTR PRIuPTR
+#  define FMTxPTR PRIxPTR
+#endif
+
+/* Size of stack-allocated buffer passed to buferror(). */
+#define BUFERROR_BUF		64
+
+/*
+ * Size of stack-allocated buffer used by malloc_{,v,vc}printf().  This must be
+ * large enough for all possible uses within jemalloc.
+ */
+#define MALLOC_PRINTF_BUFSIZE	4096
+
+int buferror(int err, char *buf, size_t buflen);
+uintmax_t malloc_strtoumax(const char *restrict nptr, char **restrict endptr,
+    int base);
+void malloc_write(const char *s);
+
+/*
+ * malloc_vsnprintf() supports a subset of snprintf(3) that avoids floating
+ * point math.
+ */
+size_t malloc_vsnprintf(char *str, size_t size, const char *format,
+    va_list ap);
+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);
+void malloc_cprintf(void (*write_cb)(void *, const char *), void *cbopaque,
+    const char *format, ...) JEMALLOC_FORMAT_PRINTF(3, 4);
+void malloc_printf(const char *format, ...) JEMALLOC_FORMAT_PRINTF(1, 2);
+
+#endif /* JEMALLOC_INTERNAL_MALLOC_IO_H */

include/jemalloc/internal/mutex.h

@@ -0,0 +1,248 @@
+#ifndef JEMALLOC_INTERNAL_MUTEX_H
+#define JEMALLOC_INTERNAL_MUTEX_H
+
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/mutex_prof.h"
+#include "jemalloc/internal/tsd.h"
+#include "jemalloc/internal/witness.h"
+
+typedef enum {
+	/* Can only acquire one mutex of a given witness rank at a time. */
+	malloc_mutex_rank_exclusive,
+	/*
+	 * Can acquire multiple mutexes of the same witness rank, but in
+	 * address-ascending order only.
+	 */
+	malloc_mutex_address_ordered
+} malloc_mutex_lock_order_t;
+
+typedef struct malloc_mutex_s malloc_mutex_t;
+struct malloc_mutex_s {
+	union {
+		struct {
+			/*
+			 * prof_data is defined first to reduce cacheline
+			 * bouncing: the data is not touched by the mutex holder
+			 * during unlocking, while might be modified by
+			 * contenders.  Having it before the mutex itself could
+			 * avoid prefetching a modified cacheline (for the
+			 * unlocking thread).
+			 */
+			mutex_prof_data_t	prof_data;
+#ifdef _WIN32
+#  if _WIN32_WINNT >= 0x0600
+			SRWLOCK         	lock;
+#  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))
+			pthread_mutex_t		lock;
+			malloc_mutex_t		*postponed_next;
+#else
+			pthread_mutex_t		lock;
+#endif
+		};
+		/*
+		 * We only touch witness when configured w/ debug.  However we
+		 * keep the field in a union when !debug so that we don't have
+		 * to pollute the code base with #ifdefs, while avoid paying the
+		 * memory cost.
+		 */
+#if !defined(JEMALLOC_DEBUG)
+		witness_t			witness;
+		malloc_mutex_lock_order_t	lock_order;
+#endif
+	};
+
+#if defined(JEMALLOC_DEBUG)
+	witness_t			witness;
+	malloc_mutex_lock_order_t	lock_order;
+#endif
+};
+
+/*
+ * Based on benchmark results, a fixed spin with this amount of retries works
+ * well for our critical sections.
+ */
+#define MALLOC_MUTEX_MAX_SPIN 250
+
+#ifdef _WIN32
+#  if _WIN32_WINNT >= 0x0600
+#    define MALLOC_MUTEX_LOCK(m)    AcquireSRWLockExclusive(&(m)->lock)
+#    define MALLOC_MUTEX_UNLOCK(m)  ReleaseSRWLockExclusive(&(m)->lock)
+#    define MALLOC_MUTEX_TRYLOCK(m) (!TryAcquireSRWLockExclusive(&(m)->lock))
+#  else
+#    define MALLOC_MUTEX_LOCK(m)    EnterCriticalSection(&(m)->lock)
+#    define MALLOC_MUTEX_UNLOCK(m)  LeaveCriticalSection(&(m)->lock)
+#    define MALLOC_MUTEX_TRYLOCK(m) (!TryEnterCriticalSection(&(m)->lock))
+#  endif
+#elif (defined(JEMALLOC_OS_UNFAIR_LOCK))
+#    define MALLOC_MUTEX_LOCK(m)    os_unfair_lock_lock(&(m)->lock)
+#    define MALLOC_MUTEX_UNLOCK(m)  os_unfair_lock_unlock(&(m)->lock)
+#    define MALLOC_MUTEX_TRYLOCK(m) (!os_unfair_lock_trylock(&(m)->lock))
+#elif (defined(JEMALLOC_OSSPIN))
+#    define MALLOC_MUTEX_LOCK(m)    OSSpinLockLock(&(m)->lock)
+#    define MALLOC_MUTEX_UNLOCK(m)  OSSpinLockUnlock(&(m)->lock)
+#    define MALLOC_MUTEX_TRYLOCK(m) (!OSSpinLockTry(&(m)->lock))
+#else
+#    define MALLOC_MUTEX_LOCK(m)    pthread_mutex_lock(&(m)->lock)
+#    define MALLOC_MUTEX_UNLOCK(m)  pthread_mutex_unlock(&(m)->lock)
+#    define MALLOC_MUTEX_TRYLOCK(m) (pthread_mutex_trylock(&(m)->lock) != 0)
+#endif
+
+#define LOCK_PROF_DATA_INITIALIZER					\
+    {NSTIME_ZERO_INITIALIZER, NSTIME_ZERO_INITIALIZER, 0, 0, 0,		\
+	    ATOMIC_INIT(0), 0, NULL, 0}
+
+#ifdef _WIN32
+#  define MALLOC_MUTEX_INITIALIZER
+#elif (defined(JEMALLOC_OS_UNFAIR_LOCK))
+#  define MALLOC_MUTEX_INITIALIZER					\
+     {{{LOCK_PROF_DATA_INITIALIZER, OS_UNFAIR_LOCK_INIT}},		\
+      WITNESS_INITIALIZER("mutex", WITNESS_RANK_OMIT)}
+#elif (defined(JEMALLOC_OSSPIN))
+#  define MALLOC_MUTEX_INITIALIZER					\
+     {{{LOCK_PROF_DATA_INITIALIZER, 0}},				\
+      WITNESS_INITIALIZER("mutex", WITNESS_RANK_OMIT)}
+#elif (defined(JEMALLOC_MUTEX_INIT_CB))
+#  define MALLOC_MUTEX_INITIALIZER					\
+     {{{LOCK_PROF_DATA_INITIALIZER, PTHREAD_MUTEX_INITIALIZER, NULL}},	\
+      WITNESS_INITIALIZER("mutex", WITNESS_RANK_OMIT)}
+#else
+#    define MALLOC_MUTEX_TYPE PTHREAD_MUTEX_DEFAULT
+#    define MALLOC_MUTEX_INITIALIZER					\
+       {{{LOCK_PROF_DATA_INITIALIZER, PTHREAD_MUTEX_INITIALIZER}},	\
+        WITNESS_INITIALIZER("mutex", WITNESS_RANK_OMIT)}
+#endif
+
+#ifdef JEMALLOC_LAZY_LOCK
+extern bool isthreaded;
+#else
+#  undef isthreaded /* Undo private_namespace.h definition. */
+#  define isthreaded true
+#endif
+
+bool malloc_mutex_init(malloc_mutex_t *mutex, const char *name,
+    witness_rank_t rank, malloc_mutex_lock_order_t lock_order);
+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);
+void malloc_mutex_prof_data_reset(tsdn_t *tsdn, malloc_mutex_t *mutex);
+
+void malloc_mutex_lock_slow(malloc_mutex_t *mutex);
+
+static inline void
+malloc_mutex_lock_final(malloc_mutex_t *mutex) {
+	MALLOC_MUTEX_LOCK(mutex);
+}
+
+static inline bool
+malloc_mutex_trylock_final(malloc_mutex_t *mutex) {
+	return MALLOC_MUTEX_TRYLOCK(mutex);
+}
+
+static inline void
+mutex_owner_stats_update(tsdn_t *tsdn, malloc_mutex_t *mutex) {
+	if (config_stats) {
+		mutex_prof_data_t *data = &mutex->prof_data;
+		data->n_lock_ops++;
+		if (data->prev_owner != tsdn) {
+			data->prev_owner = tsdn;
+			data->n_owner_switches++;
+		}
+	}
+}
+
+/* Trylock: return false if the lock is successfully acquired. */
+static inline bool
+malloc_mutex_trylock(tsdn_t *tsdn, malloc_mutex_t *mutex) {
+	witness_assert_not_owner(tsdn_witness_tsdp_get(tsdn), &mutex->witness);
+	if (isthreaded) {
+		if (malloc_mutex_trylock_final(mutex)) {
+			return true;
+		}
+		mutex_owner_stats_update(tsdn, mutex);
+	}
+	witness_lock(tsdn_witness_tsdp_get(tsdn), &mutex->witness);
+
+	return false;
+}
+
+/* Aggregate lock prof data. */
+static inline void
+malloc_mutex_prof_merge(mutex_prof_data_t *sum, mutex_prof_data_t *data) {
+	nstime_add(&sum->tot_wait_time, &data->tot_wait_time);
+	if (nstime_compare(&sum->max_wait_time, &data->max_wait_time) < 0) {
+		nstime_copy(&sum->max_wait_time, &data->max_wait_time);
+	}
+
+	sum->n_wait_times += data->n_wait_times;
+	sum->n_spin_acquired += data->n_spin_acquired;
+
+	if (sum->max_n_thds < data->max_n_thds) {
+		sum->max_n_thds = data->max_n_thds;
+	}
+	uint32_t cur_n_waiting_thds = atomic_load_u32(&sum->n_waiting_thds,
+	    ATOMIC_RELAXED);
+	uint32_t new_n_waiting_thds = cur_n_waiting_thds + atomic_load_u32(
+	    &data->n_waiting_thds, ATOMIC_RELAXED);
+	atomic_store_u32(&sum->n_waiting_thds, new_n_waiting_thds,
+	    ATOMIC_RELAXED);
+	sum->n_owner_switches += data->n_owner_switches;
+	sum->n_lock_ops += data->n_lock_ops;
+}
+
+static inline void
+malloc_mutex_lock(tsdn_t *tsdn, malloc_mutex_t *mutex) {
+	witness_assert_not_owner(tsdn_witness_tsdp_get(tsdn), &mutex->witness);
+	if (isthreaded) {
+		if (malloc_mutex_trylock_final(mutex)) {
+			malloc_mutex_lock_slow(mutex);
+		}
+		mutex_owner_stats_update(tsdn, mutex);
+	}
+	witness_lock(tsdn_witness_tsdp_get(tsdn), &mutex->witness);
+}
+
+static inline void
+malloc_mutex_unlock(tsdn_t *tsdn, malloc_mutex_t *mutex) {
+	witness_unlock(tsdn_witness_tsdp_get(tsdn), &mutex->witness);
+	if (isthreaded) {
+		MALLOC_MUTEX_UNLOCK(mutex);
+	}
+}
+
+static inline void
+malloc_mutex_assert_owner(tsdn_t *tsdn, malloc_mutex_t *mutex) {
+	witness_assert_owner(tsdn_witness_tsdp_get(tsdn), &mutex->witness);
+}
+
+static inline void
+malloc_mutex_assert_not_owner(tsdn_t *tsdn, malloc_mutex_t *mutex) {
+	witness_assert_not_owner(tsdn_witness_tsdp_get(tsdn), &mutex->witness);
+}
+
+/* Copy the prof data from mutex for processing. */
+static inline void
+malloc_mutex_prof_read(tsdn_t *tsdn, mutex_prof_data_t *data,
+    malloc_mutex_t *mutex) {
+	mutex_prof_data_t *source = &mutex->prof_data;
+	/* Can only read holding the mutex. */
+	malloc_mutex_assert_owner(tsdn, mutex);
+
+	/*
+	 * Not *really* allowed (we shouldn't be doing non-atomic loads of
+	 * atomic data), but the mutex protection makes this safe, and writing
+	 * a member-for-member copy is tedious for this situation.
+	 */
+	*data = *source;
+	/* n_wait_thds is not reported (modified w/o locking). */
+	atomic_store_u32(&data->n_waiting_thds, 0, ATOMIC_RELAXED);
+}
+
+#endif /* JEMALLOC_INTERNAL_MUTEX_H */

include/jemalloc/internal/mutex_pool.h

@@ -0,0 +1,94 @@
+#ifndef JEMALLOC_INTERNAL_MUTEX_POOL_H
+#define JEMALLOC_INTERNAL_MUTEX_POOL_H
+
+#include "jemalloc/internal/hash.h"
+#include "jemalloc/internal/mutex.h"
+#include "jemalloc/internal/witness.h"
+
+/* We do mod reductions by this value, so it should be kept a power of 2. */
+#define MUTEX_POOL_SIZE 256
+
+typedef struct mutex_pool_s mutex_pool_t;
+struct mutex_pool_s {
+	malloc_mutex_t mutexes[MUTEX_POOL_SIZE];
+};
+
+bool mutex_pool_init(mutex_pool_t *pool, const char *name, witness_rank_t rank);
+
+/* Internal helper - not meant to be called outside this module. */
+static inline malloc_mutex_t *
+mutex_pool_mutex(mutex_pool_t *pool, uintptr_t key) {
+	size_t hash_result[2];
+	hash(&key, sizeof(key), 0xd50dcc1b, hash_result);
+	return &pool->mutexes[hash_result[0] % MUTEX_POOL_SIZE];
+}
+
+static inline void
+mutex_pool_assert_not_held(tsdn_t *tsdn, mutex_pool_t *pool) {
+	for (int i = 0; i < MUTEX_POOL_SIZE; i++) {
+		malloc_mutex_assert_not_owner(tsdn, &pool->mutexes[i]);
+	}
+}
+
+/*
+ * Note that a mutex pool doesn't work exactly the way an embdedded mutex would.
+ * You're not allowed to acquire mutexes in the pool one at a time.  You have to
+ * acquire all the mutexes you'll need in a single function call, and then
+ * release them all in a single function call.
+ */
+
+static inline void
+mutex_pool_lock(tsdn_t *tsdn, mutex_pool_t *pool, uintptr_t key) {
+	mutex_pool_assert_not_held(tsdn, pool);
+
+	malloc_mutex_t *mutex = mutex_pool_mutex(pool, key);
+	malloc_mutex_lock(tsdn, mutex);
+}
+
+static inline void
+mutex_pool_unlock(tsdn_t *tsdn, mutex_pool_t *pool, uintptr_t key) {
+	malloc_mutex_t *mutex = mutex_pool_mutex(pool, key);
+	malloc_mutex_unlock(tsdn, mutex);
+
+	mutex_pool_assert_not_held(tsdn, pool);
+}
+
+static inline void
+mutex_pool_lock2(tsdn_t *tsdn, mutex_pool_t *pool, uintptr_t key1,
+    uintptr_t key2) {
+	mutex_pool_assert_not_held(tsdn, pool);
+
+	malloc_mutex_t *mutex1 = mutex_pool_mutex(pool, key1);
+	malloc_mutex_t *mutex2 = mutex_pool_mutex(pool, key2);
+	if ((uintptr_t)mutex1 < (uintptr_t)mutex2) {
+		malloc_mutex_lock(tsdn, mutex1);
+		malloc_mutex_lock(tsdn, mutex2);
+	} else if ((uintptr_t)mutex1 == (uintptr_t)mutex2) {
+		malloc_mutex_lock(tsdn, mutex1);
+	} else {
+		malloc_mutex_lock(tsdn, mutex2);
+		malloc_mutex_lock(tsdn, mutex1);
+	}
+}
+
+static inline void
+mutex_pool_unlock2(tsdn_t *tsdn, mutex_pool_t *pool, uintptr_t key1,
+    uintptr_t key2) {
+	malloc_mutex_t *mutex1 = mutex_pool_mutex(pool, key1);
+	malloc_mutex_t *mutex2 = mutex_pool_mutex(pool, key2);
+	if (mutex1 == mutex2) {
+		malloc_mutex_unlock(tsdn, mutex1);
+	} else {
+		malloc_mutex_unlock(tsdn, mutex1);
+		malloc_mutex_unlock(tsdn, mutex2);
+	}
+
+	mutex_pool_assert_not_held(tsdn, pool);
+}
+
+static inline void
+mutex_pool_assert_owner(tsdn_t *tsdn, mutex_pool_t *pool, uintptr_t key) {
+	malloc_mutex_assert_owner(tsdn, mutex_pool_mutex(pool, key));
+}
+
+#endif /* JEMALLOC_INTERNAL_MUTEX_POOL_H */

include/jemalloc/internal/mutex_prof.h

@@ -0,0 +1,86 @@
+#ifndef JEMALLOC_INTERNAL_MUTEX_PROF_H
+#define JEMALLOC_INTERNAL_MUTEX_PROF_H
+
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/nstime.h"
+#include "jemalloc/internal/tsd_types.h"
+
+#define MUTEX_PROF_GLOBAL_MUTEXES					\
+    OP(background_thread)						\
+    OP(ctl)								\
+    OP(prof)
+
+typedef enum {
+#define OP(mtx) global_prof_mutex_##mtx,
+	MUTEX_PROF_GLOBAL_MUTEXES
+#undef OP
+	mutex_prof_num_global_mutexes
+} mutex_prof_global_ind_t;
+
+#define MUTEX_PROF_ARENA_MUTEXES					\
+    OP(large)								\
+    OP(extent_avail)							\
+    OP(extents_dirty)							\
+    OP(extents_muzzy)							\
+    OP(extents_retained)						\
+    OP(decay_dirty)							\
+    OP(decay_muzzy)							\
+    OP(base)								\
+    OP(tcache_list)
+
+typedef enum {
+#define OP(mtx) arena_prof_mutex_##mtx,
+	MUTEX_PROF_ARENA_MUTEXES
+#undef OP
+	mutex_prof_num_arena_mutexes
+} mutex_prof_arena_ind_t;
+
+#define MUTEX_PROF_COUNTERS						\
+    OP(num_ops, uint64_t)						\
+    OP(num_wait, uint64_t)						\
+    OP(num_spin_acq, uint64_t)						\
+    OP(num_owner_switch, uint64_t)					\
+    OP(total_wait_time, uint64_t)					\
+    OP(max_wait_time, uint64_t)						\
+    OP(max_num_thds, uint32_t)
+
+typedef enum {
+#define OP(counter, type) mutex_counter_##counter,
+	MUTEX_PROF_COUNTERS
+#undef OP
+	mutex_prof_num_counters
+} mutex_prof_counter_ind_t;
+
+typedef struct {
+	/*
+	 * Counters touched on the slow path, i.e. when there is lock
+	 * contention.  We update them once we have the lock.
+	 */
+	/* Total time (in nano seconds) spent waiting on this mutex. */
+	nstime_t		tot_wait_time;
+	/* Max time (in nano seconds) spent on a single lock operation. */
+	nstime_t		max_wait_time;
+	/* # of times have to wait for this mutex (after spinning). */
+	uint64_t		n_wait_times;
+	/* # of times acquired the mutex through local spinning. */
+	uint64_t		n_spin_acquired;
+	/* Max # of threads waiting for the mutex at the same time. */
+	uint32_t		max_n_thds;
+	/* Current # of threads waiting on the lock.  Atomic synced. */
+	atomic_u32_t		n_waiting_thds;
+
+	/*
+	 * Data touched on the fast path.  These are modified right after we
+	 * grab the lock, so it's placed closest to the end (i.e. right before
+	 * the lock) so that we have a higher chance of them being on the same
+	 * cacheline.
+	 */
+	/* # of times the mutex holder is different than the previous one. */
+	uint64_t		n_owner_switches;
+	/* Previous mutex holder, to facilitate n_owner_switches. */
+	tsdn_t			*prev_owner;
+	/* # of lock() operations in total. */
+	uint64_t		n_lock_ops;
+} mutex_prof_data_t;
+
+#endif /* JEMALLOC_INTERNAL_MUTEX_PROF_H */

include/jemalloc/internal/nstime.h

@@ -0,0 +1,34 @@
+#ifndef JEMALLOC_INTERNAL_NSTIME_H
+#define JEMALLOC_INTERNAL_NSTIME_H
+
+/* Maximum supported number of seconds (~584 years). */
+#define NSTIME_SEC_MAX KQU(18446744072)
+#define NSTIME_ZERO_INITIALIZER {0}
+
+typedef struct {
+	uint64_t ns;
+} nstime_t;
+
+void nstime_init(nstime_t *time, uint64_t ns);
+void nstime_init2(nstime_t *time, uint64_t sec, uint64_t nsec);
+uint64_t nstime_ns(const nstime_t *time);
+uint64_t nstime_sec(const nstime_t *time);
+uint64_t nstime_msec(const nstime_t *time);
+uint64_t nstime_nsec(const nstime_t *time);
+void nstime_copy(nstime_t *time, const nstime_t *source);
+int nstime_compare(const nstime_t *a, const nstime_t *b);
+void nstime_add(nstime_t *time, const nstime_t *addend);
+void nstime_iadd(nstime_t *time, uint64_t addend);
+void nstime_subtract(nstime_t *time, const nstime_t *subtrahend);
+void nstime_isubtract(nstime_t *time, uint64_t subtrahend);
+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);
+
+typedef bool (nstime_monotonic_t)(void);
+extern nstime_monotonic_t *JET_MUTABLE nstime_monotonic;
+
+typedef bool (nstime_update_t)(nstime_t *);
+extern nstime_update_t *JET_MUTABLE nstime_update;
+
+#endif /* JEMALLOC_INTERNAL_NSTIME_H */

include/jemalloc/internal/pages.h

@@ -0,0 +1,71 @@
+#ifndef JEMALLOC_INTERNAL_PAGES_EXTERNS_H
+#define JEMALLOC_INTERNAL_PAGES_EXTERNS_H
+
+/* Page size.  LG_PAGE is determined by the configure script. */
+#ifdef PAGE_MASK
+#  undef PAGE_MASK
+#endif
+#define PAGE		((size_t)(1U << LG_PAGE))
+#define PAGE_MASK	((size_t)(PAGE - 1))
+/* Return the page base address for the page containing address a. */
+#define PAGE_ADDR2BASE(a)						\
+	((void *)((uintptr_t)(a) & ~PAGE_MASK))
+/* Return the smallest pagesize multiple that is >= s. */
+#define PAGE_CEILING(s)							\
+	(((s) + PAGE_MASK) & ~PAGE_MASK)
+
+/* Huge page size.  LG_HUGEPAGE is determined by the configure script. */
+#define HUGEPAGE	((size_t)(1U << LG_HUGEPAGE))
+#define HUGEPAGE_MASK	((size_t)(HUGEPAGE - 1))
+/* Return the huge page base address for the huge page containing address a. */
+#define HUGEPAGE_ADDR2BASE(a)						\
+	((void *)((uintptr_t)(a) & ~HUGEPAGE_MASK))
+/* Return the smallest pagesize multiple that is >= s. */
+#define HUGEPAGE_CEILING(s)						\
+	(((s) + HUGEPAGE_MASK) & ~HUGEPAGE_MASK)
+
+/* PAGES_CAN_PURGE_LAZY is defined if lazy purging is supported. */
+#if defined(_WIN32) || defined(JEMALLOC_PURGE_MADVISE_FREE)
+#  define PAGES_CAN_PURGE_LAZY
+#endif
+/*
+ * PAGES_CAN_PURGE_FORCED is defined if forced purging is supported.
+ *
+ * The only supported way to hard-purge on Windows is to decommit and then
+ * re-commit, but doing so is racy, and if re-commit fails it's a pain to
+ * propagate the "poisoned" memory state.  Since we typically decommit as the
+ * next step after purging on Windows anyway, there's no point in adding such
+ * complexity.
+ */
+#if !defined(_WIN32) && ((defined(JEMALLOC_PURGE_MADVISE_DONTNEED) && \
+    defined(JEMALLOC_PURGE_MADVISE_DONTNEED_ZEROS)) || \
+    defined(JEMALLOC_MAPS_COALESCE))
+#  define PAGES_CAN_PURGE_FORCED
+#endif
+
+static const bool pages_can_purge_lazy =
+#ifdef PAGES_CAN_PURGE_LAZY
+    true
+#else
+    false
+#endif
+    ;
+static const bool pages_can_purge_forced =
+#ifdef PAGES_CAN_PURGE_FORCED
+    true
+#else
+    false
+#endif
+    ;
+
+void *pages_map(void *addr, size_t size, size_t alignment, bool *commit);
+void pages_unmap(void *addr, size_t size);
+bool pages_commit(void *addr, size_t size);
+bool pages_decommit(void *addr, size_t size);
+bool pages_purge_lazy(void *addr, size_t size);
+bool pages_purge_forced(void *addr, size_t size);
+bool pages_huge(void *addr, size_t size);
+bool pages_nohuge(void *addr, size_t size);
+bool pages_boot(void);
+
+#endif /* JEMALLOC_INTERNAL_PAGES_EXTERNS_H */

include/jemalloc/internal/ph.h

@@ -0,0 +1,391 @@
+/*
+ * 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 a_type	*a_prefix##any(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 a_type	*a_prefix##remove_any(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 a_type *								\
+a_prefix##any(a_ph_type *ph) {						\
+	if (ph->ph_root == NULL) {					\
+		return NULL;						\
+	}								\
+	a_type *aux = phn_next_get(a_type, a_field, ph->ph_root);	\
+	if (aux != NULL) {						\
+		return aux;						\
+	}								\
+	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 a_type *								\
+a_prefix##remove_any(a_ph_type *ph) {					\
+	/*								\
+	 * Remove the most recently inserted aux list element, or the	\
+	 * root if the aux list is empty.  This has the effect of	\
+	 * behaving as a LIFO (and insertion/removal is therefore	\
+	 * constant-time) if a_prefix##[remove_]first() are never	\
+	 * called.							\
+	 */								\
+	if (ph->ph_root == NULL) {					\
+		return NULL;						\
+	}								\
+	a_type *ret = phn_next_get(a_type, a_field, ph->ph_root);	\
+	if (ret != NULL) {						\
+		a_type *aux = phn_next_get(a_type, a_field, ret);	\
+		phn_next_set(a_type, a_field, ph->ph_root, aux);	\
+		if (aux != NULL) {					\
+			phn_prev_set(a_type, a_field, aux,		\
+			    ph->ph_root);				\
+		}							\
+		return ret;						\
+	}								\
+	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;					\
+									\
+	if (ph->ph_root == phn) {					\
+		/*							\
+		 * We can delete from aux list without merging it, but	\
+		 * we need to merge if we are dealing with the root	\
+		 * node and it has children.				\
+		 */							\
+		if (phn_lchild_get(a_type, a_field, phn) == NULL) {	\
+			ph->ph_root = phn_next_get(a_type, a_field,	\
+			    phn);					\
+			if (ph->ph_root != NULL) {			\
+				phn_prev_set(a_type, a_field,		\
+				    ph->ph_root, NULL);			\
+			}						\
+			return;						\
+		}							\
+		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_ */

include/jemalloc/internal/private_namespace.sh

@@ -0,0 +1,5 @@
+#!/bin/sh
+
+for symbol in `cat "$@"` ; do
+  echo "#define ${symbol} JEMALLOC_N(${symbol})"
+done

include/jemalloc/internal/private_symbols.sh

@@ -0,0 +1,51 @@
+#!/bin/sh
+#
+# Generate private_symbols[_jet].awk.
+#
+# Usage: private_symbols.sh <sym_prefix> <sym>*
+#
+# <sym_prefix> is typically "" or "_".
+
+sym_prefix=$1
+shift
+
+cat <<EOF
+#!/usr/bin/env awk -f
+
+BEGIN {
+  sym_prefix = "${sym_prefix}"
+  split("\\
+EOF
+
+for public_sym in "$@" ; do
+  cat <<EOF
+        ${sym_prefix}${public_sym} \\
+EOF
+done
+
+cat <<"EOF"
+        ", exported_symbol_names)
+  # Store exported symbol names as keys in exported_symbols.
+  for (i in exported_symbol_names) {
+    exported_symbols[exported_symbol_names[i]] = 1
+  }
+}
+
+# Process 'nm -a <c_source.o>' output.
+#
+# Handle lines like:
+#   0000000000000008 D opt_junk
+#   0000000000007574 T malloc_initialized
+(NF == 3 && $2 ~ /^[ABCDGRSTVW]$/ && !($3 in exported_symbols) && $3 ~ /^[A-Za-z0-9_]+$/) {
+  print substr($3, 1+length(sym_prefix), length($3)-length(sym_prefix))
+}
+
+# Process 'dumpbin /SYMBOLS <c_source.obj>' output.
+#
+# Handle lines like:
+#   353 00008098 SECT4  notype       External     | opt_junk
+#   3F1 00000000 SECT7  notype ()    External     | malloc_initialized
+($3 ~ /^SECT[0-9]+/ && $(NF-2) == "External" && !($NF in exported_symbols)) {
+  print $NF
+}
+EOF

include/jemalloc/internal/prng.h

@@ -0,0 +1,185 @@
+#ifndef JEMALLOC_INTERNAL_PRNG_H
+#define JEMALLOC_INTERNAL_PRNG_H
+
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/bit_util.h"
+
+/*
+ * Simple linear congruential pseudo-random number generator:
+ *
+ *   prng(y) = (a*x + c) % m
+ *
+ * where the following constants ensure maximal period:
+ *
+ *   a == Odd number (relatively prime to 2^n), and (a-1) is a multiple of 4.
+ *   c == Odd number (relatively prime to 2^n).
+ *   m == 2^32
+ *
+ * See Knuth's TAOCP 3rd Ed., Vol. 2, pg. 17 for details on these constraints.
+ *
+ * This choice of m has the disadvantage that the quality of the bits is
+ * proportional to bit position.  For example, the lowest bit has a cycle of 2,
+ * the next has a cycle of 4, etc.  For this reason, we prefer to use the upper
+ * bits.
+ */
+
+/******************************************************************************/
+/* INTERNAL DEFINITIONS -- IGNORE */
+/******************************************************************************/
+#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)
+
+JEMALLOC_ALWAYS_INLINE uint32_t
+prng_state_next_u32(uint32_t state) {
+	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
+}
+
+/******************************************************************************/
+/* BEGIN PUBLIC API */
+/******************************************************************************/
+
+/*
+ * The prng_lg_range functions give a uniform int in the half-open range [0,
+ * 2**lg_range).  If atomic is true, they do so safely from multiple threads.
+ * Multithreaded 64-bit prngs aren't supported.
+ */
+
+JEMALLOC_ALWAYS_INLINE uint32_t
+prng_lg_range_u32(atomic_u32_t *state, unsigned lg_range, bool atomic) {
+	uint32_t ret, state0, state1;
+
+	assert(lg_range > 0);
+	assert(lg_range <= 32);
+
+	state0 = atomic_load_u32(state, ATOMIC_RELAXED);
+
+	if (atomic) {
+		do {
+			state1 = prng_state_next_u32(state0);
+		} while (!atomic_compare_exchange_weak_u32(state, &state0,
+		    state1, ATOMIC_RELAXED, ATOMIC_RELAXED));
+	} else {
+		state1 = prng_state_next_u32(state0);
+		atomic_store_u32(state, state1, ATOMIC_RELAXED);
+	}
+	ret = state1 >> (32 - lg_range);
+
+	return ret;
+}
+
+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);
+
+	state1 = prng_state_next_u64(*state);
+	*state = state1;
+	ret = state1 >> (64 - lg_range);
+
+	return ret;
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+prng_lg_range_zu(atomic_zu_t *state, unsigned lg_range, bool atomic) {
+	size_t ret, state0, state1;
+
+	assert(lg_range > 0);
+	assert(lg_range <= ZU(1) << (3 + LG_SIZEOF_PTR));
+
+	state0 = atomic_load_zu(state, ATOMIC_RELAXED);
+
+	if (atomic) {
+		do {
+			state1 = prng_state_next_zu(state0);
+		} while (atomic_compare_exchange_weak_zu(state, &state0,
+		    state1, ATOMIC_RELAXED, ATOMIC_RELAXED));
+	} else {
+		state1 = prng_state_next_zu(state0);
+		atomic_store_zu(state, state1, ATOMIC_RELAXED);
+	}
+	ret = state1 >> ((ZU(1) << (3 + LG_SIZEOF_PTR)) - lg_range);
+
+	return ret;
+}
+
+/*
+ * The prng_range functions behave like the prng_lg_range, but return a result
+ * in [0, range) instead of [0, 2**lg_range).
+ */
+
+JEMALLOC_ALWAYS_INLINE uint32_t
+prng_range_u32(atomic_u32_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_u64(uint64_t *state, uint64_t range) {
+	uint64_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_u64(state, lg_range);
+	} while (ret >= range);
+
+	return ret;
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+prng_range_zu(atomic_zu_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;
+}
+
+#endif /* JEMALLOC_INTERNAL_PRNG_H */

include/jemalloc/internal/prof_externs.h

@@ -0,0 +1,92 @@
+#ifndef JEMALLOC_INTERNAL_PROF_EXTERNS_H
+#define JEMALLOC_INTERNAL_PROF_EXTERNS_H
+
+#include "jemalloc/internal/mutex.h"
+
+extern malloc_mutex_t	bt2gctx_mtx;
+
+extern bool	opt_prof;
+extern bool	opt_prof_active;
+extern bool	opt_prof_thread_active_init;
+extern size_t	opt_lg_prof_sample;   /* Mean bytes between samples. */
+extern ssize_t	opt_lg_prof_interval; /* lg(prof_interval). */
+extern bool	opt_prof_gdump;       /* High-water memory dumping. */
+extern bool	opt_prof_final;       /* Final profile dumping. */
+extern bool	opt_prof_leak;        /* Dump leak summary at exit. */
+extern bool	opt_prof_accum;       /* Report cumulative bytes. */
+extern char	opt_prof_prefix[
+    /* Minimize memory bloat for non-prof builds. */
+#ifdef JEMALLOC_PROF
+    PATH_MAX +
+#endif
+    1];
+
+/* Accessed via prof_active_[gs]et{_unlocked,}(). */
+extern bool	prof_active;
+
+/* Accessed via prof_gdump_[gs]et{_unlocked,}(). */
+extern bool	prof_gdump_val;
+
+/*
+ * Profile dump interval, measured in bytes allocated.  Each arena triggers a
+ * profile dump when it reaches this threshold.  The effect is that the
+ * interval between profile dumps averages prof_interval, though the actual
+ * interval between dumps will tend to be sporadic, and the interval will be a
+ * maximum of approximately (prof_interval * narenas).
+ */
+extern uint64_t	prof_interval;
+
+/*
+ * Initialized as opt_lg_prof_sample, and potentially modified during profiling
+ * resets.
+ */
+extern size_t	lg_prof_sample;
+
+void prof_alloc_rollback(tsd_t *tsd, prof_tctx_t *tctx, bool updated);
+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);
+void prof_backtrace(prof_bt_t *bt);
+prof_tctx_t *prof_lookup(tsd_t *tsd, prof_bt_t *bt);
+#ifdef JEMALLOC_JET
+size_t prof_tdata_count(void);
+size_t prof_bt_count(void);
+#endif
+typedef int (prof_dump_open_t)(bool, const char *);
+extern prof_dump_open_t *JET_MUTABLE prof_dump_open;
+
+typedef bool (prof_dump_header_t)(tsdn_t *, bool, const prof_cnt_t *);
+extern prof_dump_header_t *JET_MUTABLE prof_dump_header;
+#ifdef JEMALLOC_JET
+void prof_cnt_all(uint64_t *curobjs, uint64_t *curbytes, uint64_t *accumobjs,
+    uint64_t *accumbytes);
+#endif
+bool prof_accum_init(tsdn_t *tsdn, prof_accum_t *prof_accum);
+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);
+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(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(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_INTERNAL_PROF_EXTERNS_H */

include/jemalloc/internal/prof_inlines_a.h

@@ -0,0 +1,72 @@
+#ifndef JEMALLOC_INTERNAL_PROF_INLINES_A_H
+#define JEMALLOC_INTERNAL_PROF_INLINES_A_H
+
+#include "jemalloc/internal/mutex.h"
+
+static inline bool
+prof_accum_add(tsdn_t *tsdn, prof_accum_t *prof_accum, uint64_t accumbytes) {
+	cassert(config_prof);
+
+	bool overflow;
+	uint64_t a0, a1;
+
+	/*
+	 * If the application allocates fast enough (and/or if idump is slow
+	 * enough), extreme overflow here (a1 >= prof_interval * 2) can cause
+	 * idump trigger coalescing.  This is an intentional mechanism that
+	 * avoids rate-limiting allocation.
+	 */
+#ifdef JEMALLOC_ATOMIC_U64
+	a0 = atomic_load_u64(&prof_accum->accumbytes, ATOMIC_RELAXED);
+	do {
+		a1 = a0 + accumbytes;
+		assert(a1 >= a0);
+		overflow = (a1 >= prof_interval);
+		if (overflow) {
+			a1 %= prof_interval;
+		}
+	} while (!atomic_compare_exchange_weak_u64(&prof_accum->accumbytes, &a0,
+	    a1, ATOMIC_RELAXED, ATOMIC_RELAXED));
+#else
+	malloc_mutex_lock(tsdn, &prof_accum->mtx);
+	a0 = prof_accum->accumbytes;
+	a1 = a0 + accumbytes;
+	overflow = (a1 >= prof_interval);
+	if (overflow) {
+		a1 %= prof_interval;
+	}
+	prof_accum->accumbytes = a1;
+	malloc_mutex_unlock(tsdn, &prof_accum->mtx);
+#endif
+	return overflow;
+}
+
+static inline void
+prof_accum_cancel(tsdn_t *tsdn, prof_accum_t *prof_accum, size_t usize) {
+	cassert(config_prof);
+
+	/*
+	 * Cancel out as much of the excessive prof_accumbytes increase as
+	 * possible without underflowing.  Interval-triggered dumps occur
+	 * slightly more often than intended as a result of incomplete
+	 * canceling.
+	 */
+	uint64_t a0, a1;
+#ifdef JEMALLOC_ATOMIC_U64
+	a0 = atomic_load_u64(&prof_accum->accumbytes, ATOMIC_RELAXED);
+	do {
+		a1 = (a0 >= LARGE_MINCLASS - usize) ?  a0 - (LARGE_MINCLASS -
+		    usize) : 0;
+	} while (!atomic_compare_exchange_weak_u64(&prof_accum->accumbytes, &a0,
+	    a1, ATOMIC_RELAXED, ATOMIC_RELAXED));
+#else
+	malloc_mutex_lock(tsdn, &prof_accum->mtx);
+	a0 = prof_accum->accumbytes;
+	a1 = (a0 >= LARGE_MINCLASS - usize) ?  a0 - (LARGE_MINCLASS - usize) :
+	    0;
+	prof_accum->accumbytes = a1;
+	malloc_mutex_unlock(tsdn, &prof_accum->mtx);
+#endif
+}
+
+#endif /* JEMALLOC_INTERNAL_PROF_INLINES_A_H */

include/jemalloc/internal/prof_inlines_b.h

@@ -0,0 +1,217 @@
+#ifndef JEMALLOC_INTERNAL_PROF_INLINES_B_H
+#define JEMALLOC_INTERNAL_PROF_INLINES_B_H
+
+#include "jemalloc/internal/sz.h"
+
+JEMALLOC_ALWAYS_INLINE bool
+prof_active_get_unlocked(void) {
+	/*
+	 * Even if opt_prof is true, sampling can be temporarily disabled by
+	 * setting prof_active to false.  No locking is used when reading
+	 * prof_active in the fast path, so there are no guarantees regarding
+	 * how long it will take for all threads to notice state changes.
+	 */
+	return prof_active;
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+prof_gdump_get_unlocked(void) {
+	/*
+	 * No locking is used when reading prof_gdump_val in the fast path, so
+	 * there are no guarantees regarding how long it will take for all
+	 * threads to notice state changes.
+	 */
+	return prof_gdump_val;
+}
+
+JEMALLOC_ALWAYS_INLINE prof_tdata_t *
+prof_tdata_get(tsd_t *tsd, bool create) {
+	prof_tdata_t *tdata;
+
+	cassert(config_prof);
+
+	tdata = tsd_prof_tdata_get(tsd);
+	if (create) {
+		if (unlikely(tdata == NULL)) {
+			if (tsd_nominal(tsd)) {
+				tdata = prof_tdata_init(tsd);
+				tsd_prof_tdata_set(tsd, tdata);
+			}
+		} else if (unlikely(tdata->expired)) {
+			tdata = prof_tdata_reinit(tsd, tdata);
+			tsd_prof_tdata_set(tsd, tdata);
+		}
+		assert(tdata == NULL || tdata->attached);
+	}
+
+	return tdata;
+}
+
+JEMALLOC_ALWAYS_INLINE prof_tctx_t *
+prof_tctx_get(tsdn_t *tsdn, const void *ptr, alloc_ctx_t *alloc_ctx) {
+	cassert(config_prof);
+	assert(ptr != NULL);
+
+	return arena_prof_tctx_get(tsdn, ptr, alloc_ctx);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+prof_tctx_set(tsdn_t *tsdn, const void *ptr, size_t usize,
+    alloc_ctx_t *alloc_ctx, prof_tctx_t *tctx) {
+	cassert(config_prof);
+	assert(ptr != NULL);
+
+	arena_prof_tctx_set(tsdn, ptr, usize, alloc_ctx, tctx);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+prof_tctx_reset(tsdn_t *tsdn, const void *ptr, prof_tctx_t *tctx) {
+	cassert(config_prof);
+	assert(ptr != NULL);
+
+	arena_prof_tctx_reset(tsdn, ptr, tctx);
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+prof_sample_accum_update(tsd_t *tsd, size_t usize, bool update,
+    prof_tdata_t **tdata_out) {
+	prof_tdata_t *tdata;
+
+	cassert(config_prof);
+
+	tdata = prof_tdata_get(tsd, true);
+	if (unlikely((uintptr_t)tdata <= (uintptr_t)PROF_TDATA_STATE_MAX)) {
+		tdata = NULL;
+	}
+
+	if (tdata_out != NULL) {
+		*tdata_out = tdata;
+	}
+
+	if (unlikely(tdata == NULL)) {
+		return true;
+	}
+
+	if (likely(tdata->bytes_until_sample >= usize)) {
+		if (update) {
+			tdata->bytes_until_sample -= usize;
+		}
+		return true;
+	} else {
+		if (tsd_reentrancy_level_get(tsd) > 0) {
+			return true;
+		}
+		/* Compute new sample threshold. */
+		if (update) {
+			prof_sample_threshold_update(tdata);
+		}
+		return !tdata->active;
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE prof_tctx_t *
+prof_alloc_prep(tsd_t *tsd, size_t usize, bool prof_active, bool update) {
+	prof_tctx_t *ret;
+	prof_tdata_t *tdata;
+	prof_bt_t bt;
+
+	assert(usize == sz_s2u(usize));
+
+	if (!prof_active || likely(prof_sample_accum_update(tsd, usize, update,
+	    &tdata))) {
+		ret = (prof_tctx_t *)(uintptr_t)1U;
+	} else {
+		bt_init(&bt, tdata->vec);
+		prof_backtrace(&bt);
+		ret = prof_lookup(tsd, &bt);
+	}
+
+	return ret;
+}
+
+JEMALLOC_ALWAYS_INLINE void
+prof_malloc(tsdn_t *tsdn, const void *ptr, size_t usize, alloc_ctx_t *alloc_ctx,
+    prof_tctx_t *tctx) {
+	cassert(config_prof);
+	assert(ptr != NULL);
+	assert(usize == isalloc(tsdn, ptr));
+
+	if (unlikely((uintptr_t)tctx > (uintptr_t)1U)) {
+		prof_malloc_sample_object(tsdn, ptr, usize, tctx);
+	} else {
+		prof_tctx_set(tsdn, ptr, usize, alloc_ctx,
+		    (prof_tctx_t *)(uintptr_t)1U);
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE 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) {
+	bool sampled, old_sampled, moved;
+
+	cassert(config_prof);
+	assert(ptr != NULL || (uintptr_t)tctx <= (uintptr_t)1U);
+
+	if (prof_active && !updated && ptr != NULL) {
+		assert(usize == isalloc(tsd_tsdn(tsd), ptr));
+		if (prof_sample_accum_update(tsd, usize, true, NULL)) {
+			/*
+			 * Don't sample.  The usize passed to prof_alloc_prep()
+			 * was larger than what actually got allocated, so a
+			 * backtrace was captured for this allocation, even
+			 * though its actual usize was insufficient to cross the
+			 * sample threshold.
+			 */
+			prof_alloc_rollback(tsd, tctx, true);
+			tctx = (prof_tctx_t *)(uintptr_t)1U;
+		}
+	}
+
+	sampled = ((uintptr_t)tctx > (uintptr_t)1U);
+	old_sampled = ((uintptr_t)old_tctx > (uintptr_t)1U);
+	moved = (ptr != old_ptr);
+
+	if (unlikely(sampled)) {
+		prof_malloc_sample_object(tsd_tsdn(tsd), ptr, usize, tctx);
+	} else if (moved) {
+		prof_tctx_set(tsd_tsdn(tsd), ptr, usize, NULL,
+		    (prof_tctx_t *)(uintptr_t)1U);
+	} else if (unlikely(old_sampled)) {
+		/*
+		 * prof_tctx_set() would work for the !moved case as well, but
+		 * prof_tctx_reset() is slightly cheaper, and the proper thing
+		 * to do here in the presence of explicit knowledge re: moved
+		 * state.
+		 */
+		prof_tctx_reset(tsd_tsdn(tsd), ptr, tctx);
+	} else {
+		assert((uintptr_t)prof_tctx_get(tsd_tsdn(tsd), ptr, NULL) ==
+		    (uintptr_t)1U);
+	}
+
+	/*
+	 * The prof_free_sampled_object() call must come after the
+	 * prof_malloc_sample_object() call, because tctx and old_tctx may be
+	 * the same, in which case reversing the call order could cause the tctx
+	 * to be prematurely destroyed as a side effect of momentarily zeroed
+	 * counters.
+	 */
+	if (unlikely(old_sampled)) {
+		prof_free_sampled_object(tsd, old_usize, old_tctx);
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE void
+prof_free(tsd_t *tsd, const void *ptr, size_t usize, alloc_ctx_t *alloc_ctx) {
+	prof_tctx_t *tctx = prof_tctx_get(tsd_tsdn(tsd), ptr, alloc_ctx);
+
+	cassert(config_prof);
+	assert(usize == isalloc(tsd_tsdn(tsd), ptr));
+
+	if (unlikely((uintptr_t)tctx > (uintptr_t)1U)) {
+		prof_free_sampled_object(tsd, usize, tctx);
+	}
+}
+
+#endif /* JEMALLOC_INTERNAL_PROF_INLINES_B_H */

include/jemalloc/internal/prof_structs.h

@@ -0,0 +1,201 @@
+#ifndef JEMALLOC_INTERNAL_PROF_STRUCTS_H
+#define JEMALLOC_INTERNAL_PROF_STRUCTS_H
+
+#include "jemalloc/internal/ckh.h"
+#include "jemalloc/internal/mutex.h"
+#include "jemalloc/internal/prng.h"
+#include "jemalloc/internal/rb.h"
+
+struct prof_bt_s {
+	/* Backtrace, stored as len program counters. */
+	void		**vec;
+	unsigned	len;
+};
+
+#ifdef JEMALLOC_PROF_LIBGCC
+/* Data structure passed to libgcc _Unwind_Backtrace() callback functions. */
+typedef struct {
+	prof_bt_t	*bt;
+	unsigned	max;
+} prof_unwind_data_t;
+#endif
+
+struct prof_accum_s {
+#ifndef JEMALLOC_ATOMIC_U64
+	malloc_mutex_t	mtx;
+	uint64_t	accumbytes;
+#else
+	atomic_u64_t	accumbytes;
+#endif
+};
+
+struct prof_cnt_s {
+	/* Profiling counters. */
+	uint64_t	curobjs;
+	uint64_t	curbytes;
+	uint64_t	accumobjs;
+	uint64_t	accumbytes;
+};
+
+typedef enum {
+	prof_tctx_state_initializing,
+	prof_tctx_state_nominal,
+	prof_tctx_state_dumping,
+	prof_tctx_state_purgatory /* Dumper must finish destroying. */
+} prof_tctx_state_t;
+
+struct prof_tctx_s {
+	/* Thread data for thread that performed the allocation. */
+	prof_tdata_t		*tdata;
+
+	/*
+	 * Copy of tdata->thr_{uid,discrim}, necessary because tdata may be
+	 * defunct during teardown.
+	 */
+	uint64_t		thr_uid;
+	uint64_t		thr_discrim;
+
+	/* Profiling counters, protected by tdata->lock. */
+	prof_cnt_t		cnts;
+
+	/* Associated global context. */
+	prof_gctx_t		*gctx;
+
+	/*
+	 * UID that distinguishes multiple tctx's created by the same thread,
+	 * but coexisting in gctx->tctxs.  There are two ways that such
+	 * coexistence can occur:
+	 * - A dumper thread can cause a tctx to be retained in the purgatory
+	 *   state.
+	 * - Although a single "producer" thread must create all tctx's which
+	 *   share the same thr_uid, multiple "consumers" can each concurrently
+	 *   execute portions of prof_tctx_destroy().  prof_tctx_destroy() only
+	 *   gets called once each time cnts.cur{objs,bytes} drop to 0, but this
+	 *   threshold can be hit again before the first consumer finishes
+	 *   executing prof_tctx_destroy().
+	 */
+	uint64_t		tctx_uid;
+
+	/* Linkage into gctx's tctxs. */
+	rb_node(prof_tctx_t)	tctx_link;
+
+	/*
+	 * True during prof_alloc_prep()..prof_malloc_sample_object(), prevents
+	 * sample vs destroy race.
+	 */
+	bool			prepared;
+
+	/* Current dump-related state, protected by gctx->lock. */
+	prof_tctx_state_t	state;
+
+	/*
+	 * Copy of cnts snapshotted during early dump phase, protected by
+	 * dump_mtx.
+	 */
+	prof_cnt_t		dump_cnts;
+};
+typedef rb_tree(prof_tctx_t) prof_tctx_tree_t;
+
+struct prof_gctx_s {
+	/* Protects nlimbo, cnt_summed, and tctxs. */
+	malloc_mutex_t		*lock;
+
+	/*
+	 * Number of threads that currently cause this gctx to be in a state of
+	 * limbo due to one of:
+	 *   - Initializing this gctx.
+	 *   - Initializing per thread counters associated with this gctx.
+	 *   - Preparing to destroy this gctx.
+	 *   - Dumping a heap profile that includes this gctx.
+	 * nlimbo must be 1 (single destroyer) in order to safely destroy the
+	 * gctx.
+	 */
+	unsigned		nlimbo;
+
+	/*
+	 * Tree of profile counters, one for each thread that has allocated in
+	 * this context.
+	 */
+	prof_tctx_tree_t	tctxs;
+
+	/* Linkage for tree of contexts to be dumped. */
+	rb_node(prof_gctx_t)	dump_link;
+
+	/* Temporary storage for summation during dump. */
+	prof_cnt_t		cnt_summed;
+
+	/* Associated backtrace. */
+	prof_bt_t		bt;
+
+	/* Backtrace vector, variable size, referred to by bt. */
+	void			*vec[1];
+};
+typedef rb_tree(prof_gctx_t) prof_gctx_tree_t;
+
+struct prof_tdata_s {
+	malloc_mutex_t		*lock;
+
+	/* Monotonically increasing unique thread identifier. */
+	uint64_t		thr_uid;
+
+	/*
+	 * Monotonically increasing discriminator among tdata structures
+	 * associated with the same thr_uid.
+	 */
+	uint64_t		thr_discrim;
+
+	/* Included in heap profile dumps if non-NULL. */
+	char			*thread_name;
+
+	bool			attached;
+	bool			expired;
+
+	rb_node(prof_tdata_t)	tdata_link;
+
+	/*
+	 * Counter used to initialize prof_tctx_t's tctx_uid.  No locking is
+	 * necessary when incrementing this field, because only one thread ever
+	 * does so.
+	 */
+	uint64_t		tctx_uid_next;
+
+	/*
+	 * Hash of (prof_bt_t *)-->(prof_tctx_t *).  Each thread tracks
+	 * backtraces for which it has non-zero allocation/deallocation counters
+	 * associated with thread-specific prof_tctx_t objects.  Other threads
+	 * may write to prof_tctx_t contents when freeing associated objects.
+	 */
+	ckh_t			bt2tctx;
+
+	/* Sampling state. */
+	uint64_t		prng_state;
+	uint64_t		bytes_until_sample;
+
+	/* State used to avoid dumping while operating on prof internals. */
+	bool			enq;
+	bool			enq_idump;
+	bool			enq_gdump;
+
+	/*
+	 * Set to true during an early dump phase for tdata's which are
+	 * currently being dumped.  New threads' tdata's have this initialized
+	 * to false so that they aren't accidentally included in later dump
+	 * phases.
+	 */
+	bool			dumping;
+
+	/*
+	 * True if profiling is active for this tdata's thread
+	 * (thread.prof.active mallctl).
+	 */
+	bool			active;
+
+	/* Temporary storage for summation during dump. */
+	prof_cnt_t		cnt_summed;
+
+	/* Backtrace vector, used for calls to prof_backtrace(). */
+	void			*vec[PROF_BT_MAX];
+};
+typedef rb_tree(prof_tdata_t) prof_tdata_tree_t;
+
+#endif /* JEMALLOC_INTERNAL_PROF_STRUCTS_H */

include/jemalloc/internal/prof_types.h

@@ -0,0 +1,56 @@
+#ifndef JEMALLOC_INTERNAL_PROF_TYPES_H
+#define JEMALLOC_INTERNAL_PROF_TYPES_H
+
+typedef struct prof_bt_s prof_bt_t;
+typedef struct prof_accum_s prof_accum_t;
+typedef struct prof_cnt_s prof_cnt_t;
+typedef struct prof_tctx_s prof_tctx_t;
+typedef struct prof_gctx_s prof_gctx_t;
+typedef struct prof_tdata_s prof_tdata_t;
+
+/* Option defaults. */
+#ifdef JEMALLOC_PROF
+#  define PROF_PREFIX_DEFAULT		"jeprof"
+#else
+#  define PROF_PREFIX_DEFAULT		""
+#endif
+#define LG_PROF_SAMPLE_DEFAULT		19
+#define LG_PROF_INTERVAL_DEFAULT	-1
+
+/*
+ * Hard limit on stack backtrace depth.  The version of prof_backtrace() that
+ * is based on __builtin_return_address() necessarily has a hard-coded number
+ * of backtrace frame handlers, and should be kept in sync with this setting.
+ */
+#define PROF_BT_MAX			128
+
+/* Initial hash table size. */
+#define PROF_CKH_MINITEMS		64
+
+/* Size of memory buffer to use when writing dump files. */
+#define PROF_DUMP_BUFSIZE		65536
+
+/* Size of stack-allocated buffer used by prof_printf(). */
+#define PROF_PRINTF_BUFSIZE		128
+
+/*
+ * Number of mutexes shared among all gctx's.  No space is allocated for these
+ * unless profiling is enabled, so it's okay to over-provision.
+ */
+#define PROF_NCTX_LOCKS			1024
+
+/*
+ * Number of mutexes shared among all tdata's.  No space is allocated for these
+ * unless profiling is enabled, so it's okay to over-provision.
+ */
+#define PROF_NTDATA_LOCKS		256
+
+/*
+ * prof_tdata pointers close to NULL are used to encode state information that
+ * is used for cleaning up during thread shutdown.
+ */
+#define PROF_TDATA_STATE_REINCARNATED	((prof_tdata_t *)(uintptr_t)1)
+#define PROF_TDATA_STATE_PURGATORY	((prof_tdata_t *)(uintptr_t)2)
+#define PROF_TDATA_STATE_MAX		PROF_TDATA_STATE_PURGATORY
+
+#endif /* JEMALLOC_INTERNAL_PROF_TYPES_H */

include/jemalloc/internal/public_namespace.sh

@@ -0,0 +1,6 @@
+#!/bin/sh
+
+for nm in `cat $1` ; do
+  n=`echo ${nm} |tr ':' ' ' |awk '{print $1}'`
+  echo "#define je_${n} JEMALLOC_N(${n})"
+done

include/jemalloc/internal/public_unnamespace.sh

@@ -0,0 +1,6 @@
+#!/bin/sh
+
+for nm in `cat $1` ; do
+  n=`echo ${nm} |tr ':' ' ' |awk '{print $1}'`
+  echo "#undef je_${n}"
+done

include/jemalloc/internal/ql.h

@@ -1,6 +1,9 @@
-/*
- * List definitions.
- */
+#ifndef JEMALLOC_INTERNAL_QL_H
+#define JEMALLOC_INTERNAL_QL_H
+
+#include "jemalloc/internal/qr.h"
+
+/* List definitions. */
 #define ql_head(a_type)							\
 struct {								\
 	a_type *qlh_first;						\
@@ -81,3 +84,5 @@ struct {								\
 
 #define ql_reverse_foreach(a_var, a_head, a_field)			\
 	qr_reverse_foreach((a_var), ql_first(a_head), a_field)
+
+#endif /* JEMALLOC_INTERNAL_QL_H */

include/jemalloc/internal/qr.h

@@ -1,3 +1,6 @@
+#ifndef JEMALLOC_INTERNAL_QR_H
+#define JEMALLOC_INTERNAL_QR_H
+
 /* Ring definitions. */
 #define qr(a_type)							\
 struct {								\
@@ -22,17 +25,15 @@ struct {								\
 	(a_qrelm)->a_field.qre_prev = (a_qr);				\
 } while (0)
 
-#define qr_after_insert(a_qrelm, a_qr, a_field)				\
-    do									\
-    {									\
+#define qr_after_insert(a_qrelm, a_qr, a_field) do {			\
 	(a_qr)->a_field.qre_next = (a_qrelm)->a_field.qre_next;		\
 	(a_qr)->a_field.qre_prev = (a_qrelm);				\
 	(a_qr)->a_field.qre_next->a_field.qre_prev = (a_qr);		\
 	(a_qrelm)->a_field.qre_next = (a_qr);				\
-    } while (0)
+} while (0)
 
-#define qr_meld(a_qr_a, a_qr_b, a_field) do {				\
-	void *t;							\
+#define qr_meld(a_qr_a, a_qr_b, a_type, a_field) do {			\
+	a_type *t;							\
 	(a_qr_a)->a_field.qre_prev->a_field.qre_next = (a_qr_b);	\
 	(a_qr_b)->a_field.qre_prev->a_field.qre_next = (a_qr_a);	\
 	t = (a_qr_a)->a_field.qre_prev;					\
@@ -40,10 +41,12 @@ struct {								\
 	(a_qr_b)->a_field.qre_prev = t;					\
 } while (0)
 
-/* qr_meld() and qr_split() are functionally equivalent, so there's no need to
- * have two copies of the code. */
-#define qr_split(a_qr_a, a_qr_b, a_field)				\
-	qr_meld((a_qr_a), (a_qr_b), a_field)
+/*
+ * qr_meld() and qr_split() are functionally equivalent, so there's no need to
+ * have two copies of the code.
+ */
+#define qr_split(a_qr_a, a_qr_b, a_type, a_field)			\
+	qr_meld((a_qr_a), (a_qr_b), a_type, a_field)
 
 #define qr_remove(a_qr, a_field) do {					\
 	(a_qr)->a_field.qre_prev->a_field.qre_next			\
@@ -65,3 +68,5 @@ struct {								\
 	    (var) != NULL;						\
 	    (var) = (((var) != (a_qr))					\
 	    ? (var)->a_field.qre_prev : NULL))
+
+#endif /* JEMALLOC_INTERNAL_QR_H */

include/jemalloc/internal/rb.h

@@ -20,21 +20,21 @@
  */
 
 #ifndef RB_H_
-#define	RB_H_
+#define RB_H_
 
-#if 0
-__FBSDID("$FreeBSD: head/lib/libc/stdlib/rb.h 204493 2010-02-28 22:57:13Z jasone $");
+#ifndef __PGI
+#define RB_COMPACT
 #endif
 
 #ifdef RB_COMPACT
 /* Node structure. */
-#define	rb_node(a_type)							\
+#define rb_node(a_type)							\
 struct {								\
     a_type *rbn_left;							\
     a_type *rbn_right_red;						\
 }
 #else
-#define	rb_node(a_type)							\
+#define rb_node(a_type)							\
 struct {								\
     a_type *rbn_left;							\
     a_type *rbn_right;							\
@@ -43,111 +43,116 @@ struct {								\
 #endif
 
 /* Root structure. */
-#define	rb_tree(a_type)							\
+#define rb_tree(a_type)							\
 struct {								\
     a_type *rbt_root;							\
-    a_type rbt_nil;							\
 }
 
 /* Left accessors. */
-#define	rbtn_left_get(a_type, a_field, a_node)				\
+#define rbtn_left_get(a_type, a_field, a_node)				\
     ((a_node)->a_field.rbn_left)
-#define	rbtn_left_set(a_type, a_field, a_node, a_left) do {		\
+#define rbtn_left_set(a_type, a_field, a_node, a_left) do {		\
     (a_node)->a_field.rbn_left = a_left;				\
 } while (0)
 
 #ifdef RB_COMPACT
 /* Right accessors. */
-#define	rbtn_right_get(a_type, a_field, a_node)				\
+#define rbtn_right_get(a_type, a_field, a_node)				\
     ((a_type *) (((intptr_t) (a_node)->a_field.rbn_right_red)		\
       & ((ssize_t)-2)))
-#define	rbtn_right_set(a_type, a_field, a_node, a_right) do {		\
+#define rbtn_right_set(a_type, a_field, a_node, a_right) do {		\
     (a_node)->a_field.rbn_right_red = (a_type *) (((uintptr_t) a_right)	\
       | (((uintptr_t) (a_node)->a_field.rbn_right_red) & ((size_t)1)));	\
 } while (0)
 
 /* Color accessors. */
-#define	rbtn_red_get(a_type, a_field, a_node)				\
+#define rbtn_red_get(a_type, a_field, a_node)				\
     ((bool) (((uintptr_t) (a_node)->a_field.rbn_right_red)		\
       & ((size_t)1)))
-#define	rbtn_color_set(a_type, a_field, a_node, a_red) do {		\
+#define rbtn_color_set(a_type, a_field, a_node, a_red) do {		\
     (a_node)->a_field.rbn_right_red = (a_type *) ((((intptr_t)		\
       (a_node)->a_field.rbn_right_red) & ((ssize_t)-2))			\
       | ((ssize_t)a_red));						\
 } while (0)
-#define	rbtn_red_set(a_type, a_field, a_node) do {			\
+#define rbtn_red_set(a_type, a_field, a_node) do {			\
     (a_node)->a_field.rbn_right_red = (a_type *) (((uintptr_t)		\
       (a_node)->a_field.rbn_right_red) | ((size_t)1));			\
 } while (0)
-#define	rbtn_black_set(a_type, a_field, a_node) do {			\
+#define rbtn_black_set(a_type, a_field, a_node) do {			\
     (a_node)->a_field.rbn_right_red = (a_type *) (((intptr_t)		\
       (a_node)->a_field.rbn_right_red) & ((ssize_t)-2));		\
 } while (0)
+
+/* Node initializer. */
+#define rbt_node_new(a_type, a_field, a_rbt, a_node) do {		\
+    /* Bookkeeping bit cannot be used by node pointer. */		\
+    assert(((uintptr_t)(a_node) & 0x1) == 0);				\
+    rbtn_left_set(a_type, a_field, (a_node), NULL);	\
+    rbtn_right_set(a_type, a_field, (a_node), NULL);	\
+    rbtn_red_set(a_type, a_field, (a_node));				\
+} while (0)
 #else
 /* Right accessors. */
-#define	rbtn_right_get(a_type, a_field, a_node)				\
+#define rbtn_right_get(a_type, a_field, a_node)				\
     ((a_node)->a_field.rbn_right)
-#define	rbtn_right_set(a_type, a_field, a_node, a_right) do {		\
+#define rbtn_right_set(a_type, a_field, a_node, a_right) do {		\
     (a_node)->a_field.rbn_right = a_right;				\
 } while (0)
 
 /* Color accessors. */
-#define	rbtn_red_get(a_type, a_field, a_node)				\
+#define rbtn_red_get(a_type, a_field, a_node)				\
     ((a_node)->a_field.rbn_red)
-#define	rbtn_color_set(a_type, a_field, a_node, a_red) do {		\
+#define rbtn_color_set(a_type, a_field, a_node, a_red) do {		\
     (a_node)->a_field.rbn_red = (a_red);				\
 } while (0)
-#define	rbtn_red_set(a_type, a_field, a_node) do {			\
+#define rbtn_red_set(a_type, a_field, a_node) do {			\
     (a_node)->a_field.rbn_red = true;					\
 } while (0)
-#define	rbtn_black_set(a_type, a_field, a_node) do {			\
+#define rbtn_black_set(a_type, a_field, a_node) do {			\
     (a_node)->a_field.rbn_red = false;					\
 } while (0)
+
+/* Node initializer. */
+#define rbt_node_new(a_type, a_field, a_rbt, a_node) do {		\
+    rbtn_left_set(a_type, a_field, (a_node), NULL);	\
+    rbtn_right_set(a_type, a_field, (a_node), NULL);	\
+    rbtn_red_set(a_type, a_field, (a_node));				\
+} while (0)
 #endif
 
-/* Node initializer. */
-#define	rbt_node_new(a_type, a_field, a_rbt, a_node) do {		\
-    rbtn_left_set(a_type, a_field, (a_node), &(a_rbt)->rbt_nil);	\
-    rbtn_right_set(a_type, a_field, (a_node), &(a_rbt)->rbt_nil);	\
-    rbtn_red_set(a_type, a_field, (a_node));				\
-} while (0)
-
 /* Tree initializer. */
-#define	rb_new(a_type, a_field, a_rbt) do {				\
-    (a_rbt)->rbt_root = &(a_rbt)->rbt_nil;				\
-    rbt_node_new(a_type, a_field, a_rbt, &(a_rbt)->rbt_nil);		\
-    rbtn_black_set(a_type, a_field, &(a_rbt)->rbt_nil);			\
+#define rb_new(a_type, a_field, a_rbt) do {				\
+    (a_rbt)->rbt_root = NULL;						\
 } while (0)
 
 /* Internal utility macros. */
-#define	rbtn_first(a_type, a_field, a_rbt, a_root, r_node) do {		\
+#define rbtn_first(a_type, a_field, a_rbt, a_root, r_node) do {		\
     (r_node) = (a_root);						\
-    if ((r_node) != &(a_rbt)->rbt_nil) {				\
+    if ((r_node) != NULL) {						\
 	for (;								\
-	  rbtn_left_get(a_type, a_field, (r_node)) != &(a_rbt)->rbt_nil;\
+	  rbtn_left_get(a_type, a_field, (r_node)) != NULL;		\
 	  (r_node) = rbtn_left_get(a_type, a_field, (r_node))) {	\
 	}								\
     }									\
 } while (0)
 
-#define	rbtn_last(a_type, a_field, a_rbt, a_root, r_node) do {		\
+#define rbtn_last(a_type, a_field, a_rbt, a_root, r_node) do {		\
     (r_node) = (a_root);						\
-    if ((r_node) != &(a_rbt)->rbt_nil) {				\
-	for (; rbtn_right_get(a_type, a_field, (r_node)) !=		\
-	  &(a_rbt)->rbt_nil; (r_node) = rbtn_right_get(a_type, a_field,	\
-	  (r_node))) {							\
+    if ((r_node) != NULL) {						\
+	for (; rbtn_right_get(a_type, a_field, (r_node)) != NULL;	\
+	  (r_node) = rbtn_right_get(a_type, a_field, (r_node))) {	\
 	}								\
     }									\
 } while (0)
 
-#define	rbtn_rotate_left(a_type, a_field, a_node, r_node) do {		\
+#define rbtn_rotate_left(a_type, a_field, a_node, r_node) do {		\
     (r_node) = rbtn_right_get(a_type, a_field, (a_node));		\
     rbtn_right_set(a_type, a_field, (a_node),				\
       rbtn_left_get(a_type, a_field, (r_node)));			\
     rbtn_left_set(a_type, a_field, (r_node), (a_node));			\
 } while (0)
 
-#define	rbtn_rotate_right(a_type, a_field, a_node, r_node) do {		\
+#define rbtn_rotate_right(a_type, a_field, a_node, r_node) do {		\
     (r_node) = rbtn_left_get(a_type, a_field, (a_node));		\
     rbtn_left_set(a_type, a_field, (a_node),				\
       rbtn_right_get(a_type, a_field, (r_node)));			\
@@ -159,9 +164,11 @@ struct {								\
  * functions generated by an equivalently parameterized call to rb_gen().
  */
 
-#define	rb_proto(a_attr, a_prefix, a_rbt_type, a_type)			\
+#define rb_proto(a_attr, a_prefix, a_rbt_type, a_type)			\
 a_attr void								\
 a_prefix##new(a_rbt_type *rbtree);					\
+a_attr bool								\
+a_prefix##empty(a_rbt_type *rbtree);					\
 a_attr a_type *								\
 a_prefix##first(a_rbt_type *rbtree);					\
 a_attr a_type *								\
@@ -171,11 +178,11 @@ a_prefix##next(a_rbt_type *rbtree, a_type *node);			\
 a_attr a_type *								\
 a_prefix##prev(a_rbt_type *rbtree, a_type *node);			\
 a_attr a_type *								\
-a_prefix##search(a_rbt_type *rbtree, a_type *key);			\
+a_prefix##search(a_rbt_type *rbtree, const a_type *key);		\
 a_attr a_type *								\
-a_prefix##nsearch(a_rbt_type *rbtree, a_type *key);			\
+a_prefix##nsearch(a_rbt_type *rbtree, const a_type *key);		\
 a_attr a_type *								\
-a_prefix##psearch(a_rbt_type *rbtree, a_type *key);			\
+a_prefix##psearch(a_rbt_type *rbtree, const a_type *key);		\
 a_attr void								\
 a_prefix##insert(a_rbt_type *rbtree, a_type *node);			\
 a_attr void								\
@@ -185,7 +192,10 @@ a_prefix##iter(a_rbt_type *rbtree, a_type *start, a_type *(*cb)(	\
   a_rbt_type *, a_type *, void *), void *arg);				\
 a_attr a_type *								\
 a_prefix##reverse_iter(a_rbt_type *rbtree, a_type *start,		\
-  a_type *(*cb)(a_rbt_type *, a_type *, void *), void *arg);
+  a_type *(*cb)(a_rbt_type *, a_type *, void *), void *arg);		\
+a_attr void								\
+a_prefix##destroy(a_rbt_type *rbtree, void (*cb)(a_type *, void *),	\
+  void *arg);
 
 /*
  * The rb_gen() macro generates a type-specific red-black tree implementation,
@@ -202,7 +212,7 @@ a_prefix##reverse_iter(a_rbt_type *rbtree, a_type *start,		\
  *                 int (a_cmp *)(a_type *a_node, a_type *a_other);
  *                                       ^^^^^^
  *                                    or a_key
- *               Interpretation of comparision function return values:
+ *               Interpretation of comparison function return values:
  *                 -1 : a_node <  a_other
  *                  0 : a_node == a_other
  *                  1 : a_node >  a_other
@@ -223,124 +233,143 @@ a_prefix##reverse_iter(a_rbt_type *rbtree, a_type *start,		\
  * The following API is generated:
  *
  *   static void
- *   ex_new(ex_t *extree);
+ *   ex_new(ex_t *tree);
  *       Description: Initialize a red-black tree structure.
  *       Args:
- *         extree: Pointer to an uninitialized red-black tree object.
+ *         tree: Pointer to an uninitialized red-black tree object.
  *
- *   static ex_node_t *
- *   ex_first(ex_t *extree);
- *   static ex_node_t *
- *   ex_last(ex_t *extree);
- *       Description: Get the first/last node in extree.
+ *   static bool
+ *   ex_empty(ex_t *tree);
+ *       Description: Determine whether tree is empty.
  *       Args:
- *         extree: Pointer to an initialized red-black tree object.
- *       Ret: First/last node in extree, or NULL if extree is empty.
+ *         tree: Pointer to an initialized red-black tree object.
+ *       Ret: True if tree is empty, false otherwise.
  *
  *   static ex_node_t *
- *   ex_next(ex_t *extree, ex_node_t *node);
+ *   ex_first(ex_t *tree);
  *   static ex_node_t *
- *   ex_prev(ex_t *extree, ex_node_t *node);
+ *   ex_last(ex_t *tree);
+ *       Description: Get the first/last node in tree.
+ *       Args:
+ *         tree: Pointer to an initialized red-black tree object.
+ *       Ret: First/last node in tree, or NULL if tree is empty.
+ *
+ *   static ex_node_t *
+ *   ex_next(ex_t *tree, ex_node_t *node);
+ *   static ex_node_t *
+ *   ex_prev(ex_t *tree, ex_node_t *node);
  *       Description: Get node's successor/predecessor.
  *       Args:
- *         extree: Pointer to an initialized red-black tree object.
- *         node : A node in extree.
- *       Ret: node's successor/predecessor in extree, or NULL if node is
+ *         tree: Pointer to an initialized red-black tree object.
+ *         node: A node in tree.
+ *       Ret: node's successor/predecessor in tree, or NULL if node is
  *            last/first.
  *
  *   static ex_node_t *
- *   ex_search(ex_t *extree, ex_node_t *key);
+ *   ex_search(ex_t *tree, const ex_node_t *key);
  *       Description: Search for node that matches key.
  *       Args:
- *         extree: Pointer to an initialized red-black tree object.
- *         key  : Search key.
- *       Ret: Node in extree that matches key, or NULL if no match.
+ *         tree: Pointer to an initialized red-black tree object.
+ *         key : Search key.
+ *       Ret: Node in tree that matches key, or NULL if no match.
  *
  *   static ex_node_t *
- *   ex_nsearch(ex_t *extree, ex_node_t *key);
+ *   ex_nsearch(ex_t *tree, const ex_node_t *key);
  *   static ex_node_t *
- *   ex_psearch(ex_t *extree, ex_node_t *key);
+ *   ex_psearch(ex_t *tree, const ex_node_t *key);
  *       Description: Search for node that matches key.  If no match is found,
  *                    return what would be key's successor/predecessor, were
- *                    key in extree.
+ *                    key in tree.
  *       Args:
- *         extree: Pointer to an initialized red-black tree object.
- *         key   : Search key.
- *       Ret: Node in extree that matches key, or if no match, hypothetical
- *            node's successor/predecessor (NULL if no successor/predecessor).
+ *         tree: Pointer to an initialized red-black tree object.
+ *         key : Search key.
+ *       Ret: Node in tree that matches key, or if no match, hypothetical node's
+ *            successor/predecessor (NULL if no successor/predecessor).
  *
  *   static void
- *   ex_insert(ex_t *extree, ex_node_t *node);
- *       Description: Insert node into extree.
+ *   ex_insert(ex_t *tree, ex_node_t *node);
+ *       Description: Insert node into tree.
  *       Args:
- *         extree: Pointer to an initialized red-black tree object.
- *         node  : Node to be inserted into extree.
+ *         tree: Pointer to an initialized red-black tree object.
+ *         node: Node to be inserted into tree.
  *
  *   static void
- *   ex_remove(ex_t *extree, ex_node_t *node);
- *       Description: Remove node from extree.
+ *   ex_remove(ex_t *tree, ex_node_t *node);
+ *       Description: Remove node from tree.
  *       Args:
- *         extree: Pointer to an initialized red-black tree object.
- *         node  : Node in extree to be removed.
+ *         tree: Pointer to an initialized red-black tree object.
+ *         node: Node in tree to be removed.
  *
  *   static ex_node_t *
- *   ex_iter(ex_t *extree, ex_node_t *start, ex_node_t *(*cb)(ex_t *,
+ *   ex_iter(ex_t *tree, ex_node_t *start, ex_node_t *(*cb)(ex_t *,
  *     ex_node_t *, void *), void *arg);
  *   static ex_node_t *
- *   ex_reverse_iter(ex_t *extree, ex_node_t *start, ex_node *(*cb)(ex_t *,
+ *   ex_reverse_iter(ex_t *tree, ex_node_t *start, ex_node *(*cb)(ex_t *,
  *     ex_node_t *, void *), void *arg);
- *       Description: Iterate forward/backward over extree, starting at node.
- *                    If extree is modified, iteration must be immediately
+ *       Description: Iterate forward/backward over tree, starting at node.  If
+ *                    tree is modified, iteration must be immediately
  *                    terminated by the callback function that causes the
  *                    modification.
  *       Args:
- *         extree: Pointer to an initialized red-black tree object.
- *         start : Node at which to start iteration, or NULL to start at
- *                 first/last node.
- *         cb    : Callback function, which is called for each node during
- *                 iteration.  Under normal circumstances the callback function
- *                 should return NULL, which causes iteration to continue.  If a
- *                 callback function returns non-NULL, iteration is immediately
- *                 terminated and the non-NULL return value is returned by the
- *                 iterator.  This is useful for re-starting iteration after
- *                 modifying extree.
- *         arg   : Opaque pointer passed to cb().
+ *         tree : Pointer to an initialized red-black tree object.
+ *         start: Node at which to start iteration, or NULL to start at
+ *                first/last node.
+ *         cb   : Callback function, which is called for each node during
+ *                iteration.  Under normal circumstances the callback function
+ *                should return NULL, which causes iteration to continue.  If a
+ *                callback function returns non-NULL, iteration is immediately
+ *                terminated and the non-NULL return value is returned by the
+ *                iterator.  This is useful for re-starting iteration after
+ *                modifying tree.
+ *         arg  : Opaque pointer passed to cb().
  *       Ret: NULL if iteration completed, or the non-NULL callback return value
  *            that caused termination of the iteration.
+ *
+ *   static void
+ *   ex_destroy(ex_t *tree, void (*cb)(ex_node_t *, void *), void *arg);
+ *       Description: Iterate over the tree with post-order traversal, remove
+ *                    each node, and run the callback if non-null.  This is
+ *                    used for destroying a tree without paying the cost to
+ *                    rebalance it.  The tree must not be otherwise altered
+ *                    during traversal.
+ *       Args:
+ *         tree: Pointer to an initialized red-black tree object.
+ *         cb  : Callback function, which, if non-null, is called for each node
+ *               during iteration.  There is no way to stop iteration once it
+ *               has begun.
+ *         arg : Opaque pointer passed to cb().
  */
-#define	rb_gen(a_attr, a_prefix, a_rbt_type, a_type, a_field, a_cmp)	\
+#define rb_gen(a_attr, a_prefix, a_rbt_type, a_type, a_field, a_cmp)	\
 a_attr void								\
 a_prefix##new(a_rbt_type *rbtree) {					\
     rb_new(a_type, a_field, rbtree);					\
 }									\
+a_attr bool								\
+a_prefix##empty(a_rbt_type *rbtree) {					\
+    return (rbtree->rbt_root == NULL);					\
+}									\
 a_attr a_type *								\
 a_prefix##first(a_rbt_type *rbtree) {					\
     a_type *ret;							\
     rbtn_first(a_type, a_field, rbtree, rbtree->rbt_root, ret);		\
-    if (ret == &rbtree->rbt_nil) {					\
-	ret = NULL;							\
-    }									\
-    return (ret);							\
+    return ret;								\
 }									\
 a_attr a_type *								\
 a_prefix##last(a_rbt_type *rbtree) {					\
     a_type *ret;							\
     rbtn_last(a_type, a_field, rbtree, rbtree->rbt_root, ret);		\
-    if (ret == &rbtree->rbt_nil) {					\
-	ret = NULL;							\
-    }									\
-    return (ret);							\
+    return ret;								\
 }									\
 a_attr a_type *								\
 a_prefix##next(a_rbt_type *rbtree, a_type *node) {			\
     a_type *ret;							\
-    if (rbtn_right_get(a_type, a_field, node) != &rbtree->rbt_nil) {	\
+    if (rbtn_right_get(a_type, a_field, node) != NULL) {		\
 	rbtn_first(a_type, a_field, rbtree, rbtn_right_get(a_type,	\
 	  a_field, node), ret);						\
     } else {								\
 	a_type *tnode = rbtree->rbt_root;				\
-	assert(tnode != &rbtree->rbt_nil);				\
-	ret = &rbtree->rbt_nil;						\
+	assert(tnode != NULL);						\
+	ret = NULL;							\
 	while (true) {							\
 	    int cmp = (a_cmp)(node, tnode);				\
 	    if (cmp < 0) {						\
@@ -351,24 +380,21 @@ a_prefix##next(a_rbt_type *rbtree, a_type *node) {			\
 	    } else {							\
 		break;							\
 	    }								\
-	    assert(tnode != &rbtree->rbt_nil);				\
+	    assert(tnode != NULL);					\
 	}								\
     }									\
-    if (ret == &rbtree->rbt_nil) {					\
-	ret = (NULL);							\
-    }									\
-    return (ret);							\
+    return ret;								\
 }									\
 a_attr a_type *								\
 a_prefix##prev(a_rbt_type *rbtree, a_type *node) {			\
     a_type *ret;							\
-    if (rbtn_left_get(a_type, a_field, node) != &rbtree->rbt_nil) {	\
+    if (rbtn_left_get(a_type, a_field, node) != NULL) {			\
 	rbtn_last(a_type, a_field, rbtree, rbtn_left_get(a_type,	\
 	  a_field, node), ret);						\
     } else {								\
 	a_type *tnode = rbtree->rbt_root;				\
-	assert(tnode != &rbtree->rbt_nil);				\
-	ret = &rbtree->rbt_nil;						\
+	assert(tnode != NULL);						\
+	ret = NULL;							\
 	while (true) {							\
 	    int cmp = (a_cmp)(node, tnode);				\
 	    if (cmp < 0) {						\
@@ -379,20 +405,17 @@ a_prefix##prev(a_rbt_type *rbtree, a_type *node) {			\
 	    } else {							\
 		break;							\
 	    }								\
-	    assert(tnode != &rbtree->rbt_nil);				\
+	    assert(tnode != NULL);					\
 	}								\
     }									\
-    if (ret == &rbtree->rbt_nil) {					\
-	ret = (NULL);							\
-    }									\
-    return (ret);							\
+    return ret;								\
 }									\
 a_attr a_type *								\
-a_prefix##search(a_rbt_type *rbtree, a_type *key) {			\
+a_prefix##search(a_rbt_type *rbtree, const a_type *key) {		\
     a_type *ret;							\
     int cmp;								\
     ret = rbtree->rbt_root;						\
-    while (ret != &rbtree->rbt_nil					\
+    while (ret != NULL							\
       && (cmp = (a_cmp)(key, ret)) != 0) {				\
 	if (cmp < 0) {							\
 	    ret = rbtn_left_get(a_type, a_field, ret);			\
@@ -400,17 +423,14 @@ a_prefix##search(a_rbt_type *rbtree, a_type *key) {			\
 	    ret = rbtn_right_get(a_type, a_field, ret);			\
 	}								\
     }									\
-    if (ret == &rbtree->rbt_nil) {					\
-	ret = (NULL);							\
-    }									\
-    return (ret);							\
+    return ret;								\
 }									\
 a_attr a_type *								\
-a_prefix##nsearch(a_rbt_type *rbtree, a_type *key) {			\
+a_prefix##nsearch(a_rbt_type *rbtree, const a_type *key) {		\
     a_type *ret;							\
     a_type *tnode = rbtree->rbt_root;					\
-    ret = &rbtree->rbt_nil;						\
-    while (tnode != &rbtree->rbt_nil) {					\
+    ret = NULL;								\
+    while (tnode != NULL) {						\
 	int cmp = (a_cmp)(key, tnode);					\
 	if (cmp < 0) {							\
 	    ret = tnode;						\
@@ -422,17 +442,14 @@ a_prefix##nsearch(a_rbt_type *rbtree, a_type *key) {			\
 	    break;							\
 	}								\
     }									\
-    if (ret == &rbtree->rbt_nil) {					\
-	ret = (NULL);							\
-    }									\
-    return (ret);							\
+    return ret;								\
 }									\
 a_attr a_type *								\
-a_prefix##psearch(a_rbt_type *rbtree, a_type *key) {			\
+a_prefix##psearch(a_rbt_type *rbtree, const a_type *key) {		\
     a_type *ret;							\
     a_type *tnode = rbtree->rbt_root;					\
-    ret = &rbtree->rbt_nil;						\
-    while (tnode != &rbtree->rbt_nil) {					\
+    ret = NULL;								\
+    while (tnode != NULL) {						\
 	int cmp = (a_cmp)(key, tnode);					\
 	if (cmp < 0) {							\
 	    tnode = rbtn_left_get(a_type, a_field, tnode);		\
@@ -444,10 +461,7 @@ a_prefix##psearch(a_rbt_type *rbtree, a_type *key) {			\
 	    break;							\
 	}								\
     }									\
-    if (ret == &rbtree->rbt_nil) {					\
-	ret = (NULL);							\
-    }									\
-    return (ret);							\
+    return ret;								\
 }									\
 a_attr void								\
 a_prefix##insert(a_rbt_type *rbtree, a_type *node) {			\
@@ -458,7 +472,7 @@ a_prefix##insert(a_rbt_type *rbtree, a_type *node) {			\
     rbt_node_new(a_type, a_field, rbtree, node);			\
     /* Wind. */								\
     path->node = rbtree->rbt_root;					\
-    for (pathp = path; pathp->node != &rbtree->rbt_nil; pathp++) {	\
+    for (pathp = path; pathp->node != NULL; pathp++) {			\
 	int cmp = pathp->cmp = a_cmp(node, pathp->node);		\
 	assert(cmp != 0);						\
 	if (cmp < 0) {							\
@@ -478,7 +492,8 @@ a_prefix##insert(a_rbt_type *rbtree, a_type *node) {			\
 	    rbtn_left_set(a_type, a_field, cnode, left);		\
 	    if (rbtn_red_get(a_type, a_field, left)) {			\
 		a_type *leftleft = rbtn_left_get(a_type, a_field, left);\
-		if (rbtn_red_get(a_type, a_field, leftleft)) {		\
+		if (leftleft != NULL && rbtn_red_get(a_type, a_field,	\
+		  leftleft)) {						\
 		    /* Fix up 4-node. */				\
 		    a_type *tnode;					\
 		    rbtn_black_set(a_type, a_field, leftleft);		\
@@ -493,7 +508,8 @@ a_prefix##insert(a_rbt_type *rbtree, a_type *node) {			\
 	    rbtn_right_set(a_type, a_field, cnode, right);		\
 	    if (rbtn_red_get(a_type, a_field, right)) {			\
 		a_type *left = rbtn_left_get(a_type, a_field, cnode);	\
-		if (rbtn_red_get(a_type, a_field, left)) {		\
+		if (left != NULL && rbtn_red_get(a_type, a_field,	\
+		  left)) {						\
 		    /* Split 4-node. */					\
 		    rbtn_black_set(a_type, a_field, left);		\
 		    rbtn_black_set(a_type, a_field, right);		\
@@ -526,7 +542,7 @@ a_prefix##remove(a_rbt_type *rbtree, a_type *node) {			\
     /* Wind. */								\
     nodep = NULL; /* Silence compiler warning. */			\
     path->node = rbtree->rbt_root;					\
-    for (pathp = path; pathp->node != &rbtree->rbt_nil; pathp++) {	\
+    for (pathp = path; pathp->node != NULL; pathp++) {			\
 	int cmp = pathp->cmp = a_cmp(node, pathp->node);		\
 	if (cmp < 0) {							\
 	    pathp[1].node = rbtn_left_get(a_type, a_field,		\
@@ -538,8 +554,7 @@ a_prefix##remove(a_rbt_type *rbtree, a_type *node) {			\
 	        /* Find node's successor, in preparation for swap. */	\
 		pathp->cmp = 1;						\
 		nodep = pathp;						\
-		for (pathp++; pathp->node != &rbtree->rbt_nil;		\
-		  pathp++) {						\
+		for (pathp++; pathp->node != NULL; pathp++) {		\
 		    pathp->cmp = -1;					\
 		    pathp[1].node = rbtn_left_get(a_type, a_field,	\
 		      pathp->node);					\
@@ -581,10 +596,10 @@ a_prefix##remove(a_rbt_type *rbtree, a_type *node) {			\
 	}								\
     } else {								\
 	a_type *left = rbtn_left_get(a_type, a_field, node);		\
-	if (left != &rbtree->rbt_nil) {					\
+	if (left != NULL) {						\
 	    /* node has no successor, but it has a left child.        */\
 	    /* Splice node out, without losing the left child.        */\
-	    assert(rbtn_red_get(a_type, a_field, node) == false);	\
+	    assert(!rbtn_red_get(a_type, a_field, node));		\
 	    assert(rbtn_red_get(a_type, a_field, left));		\
 	    rbtn_black_set(a_type, a_field, left);			\
 	    if (pathp == path) {					\
@@ -601,34 +616,32 @@ a_prefix##remove(a_rbt_type *rbtree, a_type *node) {			\
 	    return;							\
 	} else if (pathp == path) {					\
 	    /* The tree only contained one node. */			\
-	    rbtree->rbt_root = &rbtree->rbt_nil;			\
+	    rbtree->rbt_root = NULL;					\
 	    return;							\
 	}								\
     }									\
     if (rbtn_red_get(a_type, a_field, pathp->node)) {			\
 	/* Prune red node, which requires no fixup. */			\
 	assert(pathp[-1].cmp < 0);					\
-	rbtn_left_set(a_type, a_field, pathp[-1].node,			\
-	  &rbtree->rbt_nil);						\
+	rbtn_left_set(a_type, a_field, pathp[-1].node, NULL);		\
 	return;								\
     }									\
     /* The node to be pruned is black, so unwind until balance is     */\
     /* restored.                                                      */\
-    pathp->node = &rbtree->rbt_nil;					\
+    pathp->node = NULL;							\
     for (pathp--; (uintptr_t)pathp >= (uintptr_t)path; pathp--) {	\
 	assert(pathp->cmp != 0);					\
 	if (pathp->cmp < 0) {						\
 	    rbtn_left_set(a_type, a_field, pathp->node,			\
 	      pathp[1].node);						\
-	    assert(rbtn_red_get(a_type, a_field, pathp[1].node)		\
-	      == false);						\
 	    if (rbtn_red_get(a_type, a_field, pathp->node)) {		\
 		a_type *right = rbtn_right_get(a_type, a_field,		\
 		  pathp->node);						\
 		a_type *rightleft = rbtn_left_get(a_type, a_field,	\
 		  right);						\
 		a_type *tnode;						\
-		if (rbtn_red_get(a_type, a_field, rightleft)) {		\
+		if (rightleft != NULL && rbtn_red_get(a_type, a_field,	\
+		  rightleft)) {						\
 		    /* In the following diagrams, ||, //, and \\      */\
 		    /* indicate the path to the removed node.         */\
 		    /*                                                */\
@@ -671,7 +684,8 @@ a_prefix##remove(a_rbt_type *rbtree, a_type *node) {			\
 		  pathp->node);						\
 		a_type *rightleft = rbtn_left_get(a_type, a_field,	\
 		  right);						\
-		if (rbtn_red_get(a_type, a_field, rightleft)) {		\
+		if (rightleft != NULL && rbtn_red_get(a_type, a_field,	\
+		  rightleft)) {						\
 		    /*      ||                                        */\
 		    /*    pathp(b)                                    */\
 		    /*  //        \                                   */\
@@ -685,7 +699,7 @@ a_prefix##remove(a_rbt_type *rbtree, a_type *node) {			\
 		    rbtn_rotate_left(a_type, a_field, pathp->node,	\
 		      tnode);						\
 		    /* Balance restored, but rotation modified        */\
-		    /* subree root, which may actually be the tree    */\
+		    /* subtree root, which may actually be the tree   */\
 		    /* root.                                          */\
 		    if (pathp == path) {				\
 			/* Set root. */					\
@@ -725,7 +739,8 @@ a_prefix##remove(a_rbt_type *rbtree, a_type *node) {			\
 		  left);						\
 		a_type *leftrightleft = rbtn_left_get(a_type, a_field,	\
 		  leftright);						\
-		if (rbtn_red_get(a_type, a_field, leftrightleft)) {	\
+		if (leftrightleft != NULL && rbtn_red_get(a_type,	\
+		  a_field, leftrightleft)) {				\
 		    /*      ||                                        */\
 		    /*    pathp(b)                                    */\
 		    /*   /        \\                                  */\
@@ -751,7 +766,7 @@ a_prefix##remove(a_rbt_type *rbtree, a_type *node) {			\
 		    /*   (b)                                          */\
 		    /*   /                                            */\
 		    /* (b)                                            */\
-		    assert(leftright != &rbtree->rbt_nil);		\
+		    assert(leftright != NULL);				\
 		    rbtn_red_set(a_type, a_field, leftright);		\
 		    rbtn_rotate_right(a_type, a_field, pathp->node,	\
 		      tnode);						\
@@ -774,7 +789,8 @@ a_prefix##remove(a_rbt_type *rbtree, a_type *node) {			\
 		return;							\
 	    } else if (rbtn_red_get(a_type, a_field, pathp->node)) {	\
 		a_type *leftleft = rbtn_left_get(a_type, a_field, left);\
-		if (rbtn_red_get(a_type, a_field, leftleft)) {		\
+		if (leftleft != NULL && rbtn_red_get(a_type, a_field,	\
+		  leftleft)) {						\
 		    /*        ||                                      */\
 		    /*      pathp(r)                                  */\
 		    /*     /        \\                                */\
@@ -812,7 +828,8 @@ a_prefix##remove(a_rbt_type *rbtree, a_type *node) {			\
 		}							\
 	    } else {							\
 		a_type *leftleft = rbtn_left_get(a_type, a_field, left);\
-		if (rbtn_red_get(a_type, a_field, leftleft)) {		\
+		if (leftleft != NULL && rbtn_red_get(a_type, a_field,	\
+		  leftleft)) {						\
 		    /*               ||                               */\
 		    /*             pathp(b)                           */\
 		    /*            /        \\                         */\
@@ -853,22 +870,22 @@ a_prefix##remove(a_rbt_type *rbtree, a_type *node) {			\
     }									\
     /* Set root. */							\
     rbtree->rbt_root = path->node;					\
-    assert(rbtn_red_get(a_type, a_field, rbtree->rbt_root) == false);	\
+    assert(!rbtn_red_get(a_type, a_field, rbtree->rbt_root));		\
 }									\
 a_attr a_type *								\
 a_prefix##iter_recurse(a_rbt_type *rbtree, a_type *node,		\
   a_type *(*cb)(a_rbt_type *, a_type *, void *), void *arg) {		\
-    if (node == &rbtree->rbt_nil) {					\
-	return (&rbtree->rbt_nil);					\
+    if (node == NULL) {							\
+	return NULL;							\
     } else {								\
 	a_type *ret;							\
 	if ((ret = a_prefix##iter_recurse(rbtree, rbtn_left_get(a_type,	\
-	  a_field, node), cb, arg)) != &rbtree->rbt_nil			\
-	  || (ret = cb(rbtree, node, arg)) != NULL) {			\
-	    return (ret);						\
+	  a_field, node), cb, arg)) != NULL || (ret = cb(rbtree, node,	\
+	  arg)) != NULL) {						\
+	    return ret;							\
 	}								\
-	return (a_prefix##iter_recurse(rbtree, rbtn_right_get(a_type,	\
-	  a_field, node), cb, arg));					\
+	return a_prefix##iter_recurse(rbtree, rbtn_right_get(a_type,	\
+	  a_field, node), cb, arg);					\
     }									\
 }									\
 a_attr a_type *								\
@@ -878,22 +895,22 @@ a_prefix##iter_start(a_rbt_type *rbtree, a_type *start, a_type *node,	\
     if (cmp < 0) {							\
 	a_type *ret;							\
 	if ((ret = a_prefix##iter_start(rbtree, start,			\
-	  rbtn_left_get(a_type, a_field, node), cb, arg)) !=		\
-	  &rbtree->rbt_nil || (ret = cb(rbtree, node, arg)) != NULL) {	\
-	    return (ret);						\
+	  rbtn_left_get(a_type, a_field, node), cb, arg)) != NULL ||	\
+	  (ret = cb(rbtree, node, arg)) != NULL) {			\
+	    return ret;							\
 	}								\
-	return (a_prefix##iter_recurse(rbtree, rbtn_right_get(a_type,	\
-	  a_field, node), cb, arg));					\
+	return a_prefix##iter_recurse(rbtree, rbtn_right_get(a_type,	\
+	  a_field, node), cb, arg);					\
     } else if (cmp > 0) {						\
-	return (a_prefix##iter_start(rbtree, start,			\
-	  rbtn_right_get(a_type, a_field, node), cb, arg));		\
+	return a_prefix##iter_start(rbtree, start,			\
+	  rbtn_right_get(a_type, a_field, node), cb, arg);		\
     } else {								\
 	a_type *ret;							\
 	if ((ret = cb(rbtree, node, arg)) != NULL) {			\
-	    return (ret);						\
+	    return ret;							\
 	}								\
-	return (a_prefix##iter_recurse(rbtree, rbtn_right_get(a_type,	\
-	  a_field, node), cb, arg));					\
+	return a_prefix##iter_recurse(rbtree, rbtn_right_get(a_type,	\
+	  a_field, node), cb, arg);					\
     }									\
 }									\
 a_attr a_type *								\
@@ -906,25 +923,22 @@ a_prefix##iter(a_rbt_type *rbtree, a_type *start, a_type *(*cb)(	\
     } else {								\
 	ret = a_prefix##iter_recurse(rbtree, rbtree->rbt_root, cb, arg);\
     }									\
-    if (ret == &rbtree->rbt_nil) {					\
-	ret = NULL;							\
-    }									\
-    return (ret);							\
+    return ret;								\
 }									\
 a_attr a_type *								\
 a_prefix##reverse_iter_recurse(a_rbt_type *rbtree, a_type *node,	\
   a_type *(*cb)(a_rbt_type *, a_type *, void *), void *arg) {		\
-    if (node == &rbtree->rbt_nil) {					\
-	return (&rbtree->rbt_nil);					\
+    if (node == NULL) {							\
+	return NULL;							\
     } else {								\
 	a_type *ret;							\
 	if ((ret = a_prefix##reverse_iter_recurse(rbtree,		\
-	  rbtn_right_get(a_type, a_field, node), cb, arg)) !=		\
-	  &rbtree->rbt_nil || (ret = cb(rbtree, node, arg)) != NULL) {	\
-	    return (ret);						\
+	  rbtn_right_get(a_type, a_field, node), cb, arg)) != NULL ||	\
+	  (ret = cb(rbtree, node, arg)) != NULL) {			\
+	    return ret;							\
 	}								\
-	return (a_prefix##reverse_iter_recurse(rbtree,			\
-	  rbtn_left_get(a_type, a_field, node), cb, arg));		\
+	return a_prefix##reverse_iter_recurse(rbtree,			\
+	  rbtn_left_get(a_type, a_field, node), cb, arg);		\
     }									\
 }									\
 a_attr a_type *								\
@@ -935,22 +949,22 @@ a_prefix##reverse_iter_start(a_rbt_type *rbtree, a_type *start,		\
     if (cmp > 0) {							\
 	a_type *ret;							\
 	if ((ret = a_prefix##reverse_iter_start(rbtree, start,		\
-	  rbtn_right_get(a_type, a_field, node), cb, arg)) !=		\
-	  &rbtree->rbt_nil || (ret = cb(rbtree, node, arg)) != NULL) {	\
-	    return (ret);						\
+	  rbtn_right_get(a_type, a_field, node), cb, arg)) != NULL ||	\
+	  (ret = cb(rbtree, node, arg)) != NULL) {			\
+	    return ret;							\
 	}								\
-	return (a_prefix##reverse_iter_recurse(rbtree,			\
-	  rbtn_left_get(a_type, a_field, node), cb, arg));		\
+	return a_prefix##reverse_iter_recurse(rbtree,			\
+	  rbtn_left_get(a_type, a_field, node), cb, arg);		\
     } else if (cmp < 0) {						\
-	return (a_prefix##reverse_iter_start(rbtree, start,		\
-	  rbtn_left_get(a_type, a_field, node), cb, arg));		\
+	return a_prefix##reverse_iter_start(rbtree, start,		\
+	  rbtn_left_get(a_type, a_field, node), cb, arg);		\
     } else {								\
 	a_type *ret;							\
 	if ((ret = cb(rbtree, node, arg)) != NULL) {			\
-	    return (ret);						\
+	    return ret;							\
 	}								\
-	return (a_prefix##reverse_iter_recurse(rbtree,			\
-	  rbtn_left_get(a_type, a_field, node), cb, arg));		\
+	return a_prefix##reverse_iter_recurse(rbtree,			\
+	  rbtn_left_get(a_type, a_field, node), cb, arg);		\
     }									\
 }									\
 a_attr a_type *								\
@@ -964,10 +978,29 @@ a_prefix##reverse_iter(a_rbt_type *rbtree, a_type *start,		\
 	ret = a_prefix##reverse_iter_recurse(rbtree, rbtree->rbt_root,	\
 	  cb, arg);							\
     }									\
-    if (ret == &rbtree->rbt_nil) {					\
-	ret = NULL;							\
+    return ret;								\
+}									\
+a_attr void								\
+a_prefix##destroy_recurse(a_rbt_type *rbtree, a_type *node, void (*cb)(	\
+  a_type *, void *), void *arg) {					\
+    if (node == NULL) {							\
+	return;								\
     }									\
-    return (ret);							\
+    a_prefix##destroy_recurse(rbtree, rbtn_left_get(a_type, a_field,	\
+      node), cb, arg);							\
+    rbtn_left_set(a_type, a_field, (node), NULL);			\
+    a_prefix##destroy_recurse(rbtree, rbtn_right_get(a_type, a_field,	\
+      node), cb, arg);							\
+    rbtn_right_set(a_type, a_field, (node), NULL);			\
+    if (cb) {								\
+	cb(node, arg);							\
+    }									\
+}									\
+a_attr void								\
+a_prefix##destroy(a_rbt_type *rbtree, void (*cb)(a_type *, void *),	\
+  void *arg) {								\
+    a_prefix##destroy_recurse(rbtree, rbtree->rbt_root, cb, arg);	\
+    rbtree->rbt_root = NULL;						\
 }
 
 #endif /* RB_H_ */

include/jemalloc/internal/rtree.h

@@ -0,0 +1,474 @@
+#ifndef JEMALLOC_INTERNAL_RTREE_H
+#define JEMALLOC_INTERNAL_RTREE_H
+
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/mutex.h"
+#include "jemalloc/internal/rtree_tsd.h"
+#include "jemalloc/internal/size_classes.h"
+#include "jemalloc/internal/tsd.h"
+
+/*
+ * This radix tree implementation is tailored to the singular purpose of
+ * associating metadata with extents that are currently owned by jemalloc.
+ *
+ *******************************************************************************
+ */
+
+/* Number of high insignificant bits. */
+#define RTREE_NHIB ((1U << (LG_SIZEOF_PTR+3)) - LG_VADDR)
+/* Number of low insigificant bits. */
+#define RTREE_NLIB LG_PAGE
+/* Number of significant bits. */
+#define RTREE_NSB (LG_VADDR - RTREE_NLIB)
+/* Number of levels in radix tree. */
+#if RTREE_NSB <= 10
+#  define RTREE_HEIGHT 1
+#elif RTREE_NSB <= 36
+#  define RTREE_HEIGHT 2
+#elif RTREE_NSB <= 52
+#  define RTREE_HEIGHT 3
+#else
+#  error Unsupported number of significant virtual address bits
+#endif
+/* Use compact leaf representation if virtual address encoding allows. */
+#if RTREE_NHIB >= LG_CEIL_NSIZES
+#  define RTREE_LEAF_COMPACT
+#endif
+
+/* Needed for initialization only. */
+#define RTREE_LEAFKEY_INVALID ((uintptr_t)1)
+
+typedef struct rtree_node_elm_s rtree_node_elm_t;
+struct rtree_node_elm_s {
+	atomic_p_t	child; /* (rtree_{node,leaf}_elm_t *) */
+};
+
+struct rtree_leaf_elm_s {
+#ifdef RTREE_LEAF_COMPACT
+	/*
+	 * Single pointer-width field containing all three leaf element fields.
+	 * For example, on a 64-bit x64 system with 48 significant virtual
+	 * memory address bits, the index, extent, and slab fields are packed as
+	 * such:
+	 *
+	 * x: index
+	 * e: extent
+	 * b: slab
+	 *
+	 *   00000000 xxxxxxxx eeeeeeee [...] eeeeeeee eeee000b
+	 */
+	atomic_p_t	le_bits;
+#else
+	atomic_p_t	le_extent; /* (extent_t *) */
+	atomic_u_t	le_szind; /* (szind_t) */
+	atomic_b_t	le_slab; /* (bool) */
+#endif
+};
+
+typedef struct rtree_level_s rtree_level_t;
+struct rtree_level_s {
+	/* Number of key bits distinguished by this level. */
+	unsigned		bits;
+	/*
+	 * Cumulative number of key bits distinguished by traversing to
+	 * corresponding tree level.
+	 */
+	unsigned		cumbits;
+};
+
+typedef struct rtree_s rtree_t;
+struct rtree_s {
+	malloc_mutex_t		init_lock;
+	/* Number of elements based on rtree_levels[0].bits. */
+#if RTREE_HEIGHT > 1
+	rtree_node_elm_t	root[1U << (RTREE_NSB/RTREE_HEIGHT)];
+#else
+	rtree_leaf_elm_t	root[1U << (RTREE_NSB/RTREE_HEIGHT)];
+#endif
+};
+
+/*
+ * Split the bits into one to three partitions depending on number of
+ * significant bits.  It the number of bits does not divide evenly into the
+ * number of levels, place one remainder bit per level starting at the leaf
+ * level.
+ */
+static const rtree_level_t rtree_levels[] = {
+#if RTREE_HEIGHT == 1
+	{RTREE_NSB, RTREE_NHIB + RTREE_NSB}
+#elif RTREE_HEIGHT == 2
+	{RTREE_NSB/2, RTREE_NHIB + RTREE_NSB/2},
+	{RTREE_NSB/2 + RTREE_NSB%2, RTREE_NHIB + RTREE_NSB}
+#elif RTREE_HEIGHT == 3
+	{RTREE_NSB/3, RTREE_NHIB + RTREE_NSB/3},
+	{RTREE_NSB/3 + RTREE_NSB%3/2,
+	    RTREE_NHIB + RTREE_NSB/3*2 + RTREE_NSB%3/2},
+	{RTREE_NSB/3 + RTREE_NSB%3 - RTREE_NSB%3/2, RTREE_NHIB + RTREE_NSB}
+#else
+#  error Unsupported rtree height
+#endif
+};
+
+bool rtree_new(rtree_t *rtree, bool zeroed);
+
+typedef rtree_node_elm_t *(rtree_node_alloc_t)(tsdn_t *, rtree_t *, size_t);
+extern rtree_node_alloc_t *JET_MUTABLE rtree_node_alloc;
+
+typedef rtree_leaf_elm_t *(rtree_leaf_alloc_t)(tsdn_t *, rtree_t *, size_t);
+extern rtree_leaf_alloc_t *JET_MUTABLE rtree_leaf_alloc;
+
+typedef void (rtree_node_dalloc_t)(tsdn_t *, rtree_t *, rtree_node_elm_t *);
+extern rtree_node_dalloc_t *JET_MUTABLE rtree_node_dalloc;
+
+typedef void (rtree_leaf_dalloc_t)(tsdn_t *, rtree_t *, rtree_leaf_elm_t *);
+extern rtree_leaf_dalloc_t *JET_MUTABLE rtree_leaf_dalloc;
+#ifdef JEMALLOC_JET
+void rtree_delete(tsdn_t *tsdn, rtree_t *rtree);
+#endif
+rtree_leaf_elm_t *rtree_leaf_elm_lookup_hard(tsdn_t *tsdn, rtree_t *rtree,
+    rtree_ctx_t *rtree_ctx, uintptr_t key, bool dependent, bool init_missing);
+
+JEMALLOC_ALWAYS_INLINE uintptr_t
+rtree_leafkey(uintptr_t key) {
+	unsigned ptrbits = ZU(1) << (LG_SIZEOF_PTR+3);
+	unsigned cumbits = (rtree_levels[RTREE_HEIGHT-1].cumbits -
+	    rtree_levels[RTREE_HEIGHT-1].bits);
+	unsigned maskbits = ptrbits - cumbits;
+	uintptr_t mask = ~((ZU(1) << maskbits) - 1);
+	return (key & mask);
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+rtree_cache_direct_map(uintptr_t key) {
+	unsigned ptrbits = ZU(1) << (LG_SIZEOF_PTR+3);
+	unsigned cumbits = (rtree_levels[RTREE_HEIGHT-1].cumbits -
+	    rtree_levels[RTREE_HEIGHT-1].bits);
+	unsigned maskbits = ptrbits - cumbits;
+	return (size_t)((key >> maskbits) & (RTREE_CTX_NCACHE - 1));
+}
+
+JEMALLOC_ALWAYS_INLINE uintptr_t
+rtree_subkey(uintptr_t key, unsigned level) {
+	unsigned ptrbits = ZU(1) << (LG_SIZEOF_PTR+3);
+	unsigned cumbits = rtree_levels[level].cumbits;
+	unsigned shiftbits = ptrbits - cumbits;
+	unsigned maskbits = rtree_levels[level].bits;
+	uintptr_t mask = (ZU(1) << maskbits) - 1;
+	return ((key >> shiftbits) & mask);
+}
+
+/*
+ * Atomic getters.
+ *
+ * dependent: Reading a value on behalf of a pointer to a valid allocation
+ *            is guaranteed to be a clean read even without synchronization,
+ *            because the rtree update became visible in memory before the
+ *            pointer came into existence.
+ * !dependent: An arbitrary read, e.g. on behalf of ivsalloc(), may not be
+ *             dependent on a previous rtree write, which means a stale read
+ *             could result if synchronization were omitted here.
+ */
+#  ifdef RTREE_LEAF_COMPACT
+JEMALLOC_ALWAYS_INLINE uintptr_t
+rtree_leaf_elm_bits_read(tsdn_t *tsdn, rtree_t *rtree, rtree_leaf_elm_t *elm,
+    bool dependent) {
+	return (uintptr_t)atomic_load_p(&elm->le_bits, dependent
+	    ? ATOMIC_RELAXED : ATOMIC_ACQUIRE);
+}
+
+JEMALLOC_ALWAYS_INLINE extent_t *
+rtree_leaf_elm_bits_extent_get(uintptr_t bits) {
+	/* Restore sign-extended high bits, mask slab bit. */
+	return (extent_t *)((uintptr_t)((intptr_t)(bits << RTREE_NHIB) >>
+	    RTREE_NHIB) & ~((uintptr_t)0x1));
+}
+
+JEMALLOC_ALWAYS_INLINE szind_t
+rtree_leaf_elm_bits_szind_get(uintptr_t bits) {
+	return (szind_t)(bits >> LG_VADDR);
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+rtree_leaf_elm_bits_slab_get(uintptr_t bits) {
+	return (bool)(bits & (uintptr_t)0x1);
+}
+
+#  endif
+
+JEMALLOC_ALWAYS_INLINE extent_t *
+rtree_leaf_elm_extent_read(tsdn_t *tsdn, rtree_t *rtree, rtree_leaf_elm_t *elm,
+    bool dependent) {
+#ifdef RTREE_LEAF_COMPACT
+	uintptr_t bits = rtree_leaf_elm_bits_read(tsdn, rtree, elm, dependent);
+	return rtree_leaf_elm_bits_extent_get(bits);
+#else
+	extent_t *extent = (extent_t *)atomic_load_p(&elm->le_extent, dependent
+	    ? ATOMIC_RELAXED : ATOMIC_ACQUIRE);
+	return extent;
+#endif
+}
+
+JEMALLOC_ALWAYS_INLINE szind_t
+rtree_leaf_elm_szind_read(tsdn_t *tsdn, rtree_t *rtree, rtree_leaf_elm_t *elm,
+    bool dependent) {
+#ifdef RTREE_LEAF_COMPACT
+	uintptr_t bits = rtree_leaf_elm_bits_read(tsdn, rtree, elm, dependent);
+	return rtree_leaf_elm_bits_szind_get(bits);
+#else
+	return (szind_t)atomic_load_u(&elm->le_szind, dependent ? ATOMIC_RELAXED
+	    : ATOMIC_ACQUIRE);
+#endif
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+rtree_leaf_elm_slab_read(tsdn_t *tsdn, rtree_t *rtree, rtree_leaf_elm_t *elm,
+    bool dependent) {
+#ifdef RTREE_LEAF_COMPACT
+	uintptr_t bits = rtree_leaf_elm_bits_read(tsdn, rtree, elm, dependent);
+	return rtree_leaf_elm_bits_slab_get(bits);
+#else
+	return atomic_load_b(&elm->le_slab, dependent ? ATOMIC_RELAXED :
+	    ATOMIC_ACQUIRE);
+#endif
+}
+
+static inline void
+rtree_leaf_elm_extent_write(tsdn_t *tsdn, rtree_t *rtree, rtree_leaf_elm_t *elm,
+    extent_t *extent) {
+#ifdef RTREE_LEAF_COMPACT
+	uintptr_t old_bits = rtree_leaf_elm_bits_read(tsdn, rtree, elm, true);
+	uintptr_t bits = ((uintptr_t)rtree_leaf_elm_bits_szind_get(old_bits) <<
+	    LG_VADDR) | ((uintptr_t)extent & (((uintptr_t)0x1 << LG_VADDR) - 1))
+	    | ((uintptr_t)rtree_leaf_elm_bits_slab_get(old_bits));
+	atomic_store_p(&elm->le_bits, (void *)bits, ATOMIC_RELEASE);
+#else
+	atomic_store_p(&elm->le_extent, extent, ATOMIC_RELEASE);
+#endif
+}
+
+static inline void
+rtree_leaf_elm_szind_write(tsdn_t *tsdn, rtree_t *rtree, rtree_leaf_elm_t *elm,
+    szind_t szind) {
+	assert(szind <= NSIZES);
+
+#ifdef RTREE_LEAF_COMPACT
+	uintptr_t old_bits = rtree_leaf_elm_bits_read(tsdn, rtree, elm,
+	    true);
+	uintptr_t bits = ((uintptr_t)szind << LG_VADDR) |
+	    ((uintptr_t)rtree_leaf_elm_bits_extent_get(old_bits) &
+	    (((uintptr_t)0x1 << LG_VADDR) - 1)) |
+	    ((uintptr_t)rtree_leaf_elm_bits_slab_get(old_bits));
+	atomic_store_p(&elm->le_bits, (void *)bits, ATOMIC_RELEASE);
+#else
+	atomic_store_u(&elm->le_szind, szind, ATOMIC_RELEASE);
+#endif
+}
+
+static inline void
+rtree_leaf_elm_slab_write(tsdn_t *tsdn, rtree_t *rtree, rtree_leaf_elm_t *elm,
+     bool slab) {
+#ifdef RTREE_LEAF_COMPACT
+	uintptr_t old_bits = rtree_leaf_elm_bits_read(tsdn, rtree, elm,
+	    true);
+	uintptr_t bits = ((uintptr_t)rtree_leaf_elm_bits_szind_get(old_bits) <<
+	    LG_VADDR) | ((uintptr_t)rtree_leaf_elm_bits_extent_get(old_bits) &
+	    (((uintptr_t)0x1 << LG_VADDR) - 1)) | ((uintptr_t)slab);
+	atomic_store_p(&elm->le_bits, (void *)bits, ATOMIC_RELEASE);
+#else
+	atomic_store_b(&elm->le_slab, slab, ATOMIC_RELEASE);
+#endif
+}
+
+static inline void
+rtree_leaf_elm_write(tsdn_t *tsdn, rtree_t *rtree, rtree_leaf_elm_t *elm,
+    extent_t *extent, szind_t szind, bool slab) {
+#ifdef RTREE_LEAF_COMPACT
+	uintptr_t bits = ((uintptr_t)szind << LG_VADDR) |
+	    ((uintptr_t)extent & (((uintptr_t)0x1 << LG_VADDR) - 1)) |
+	    ((uintptr_t)slab);
+	atomic_store_p(&elm->le_bits, (void *)bits, ATOMIC_RELEASE);
+#else
+	rtree_leaf_elm_slab_write(tsdn, rtree, elm, slab);
+	rtree_leaf_elm_szind_write(tsdn, rtree, elm, szind);
+	/*
+	 * Write extent last, since the element is atomically considered valid
+	 * as soon as the extent field is non-NULL.
+	 */
+	rtree_leaf_elm_extent_write(tsdn, rtree, elm, extent);
+#endif
+}
+
+static inline void
+rtree_leaf_elm_szind_slab_update(tsdn_t *tsdn, rtree_t *rtree,
+    rtree_leaf_elm_t *elm, szind_t szind, bool slab) {
+	assert(!slab || szind < NBINS);
+
+	/*
+	 * The caller implicitly assures that it is the only writer to the szind
+	 * and slab fields, and that the extent field cannot currently change.
+	 */
+	rtree_leaf_elm_slab_write(tsdn, rtree, elm, slab);
+	rtree_leaf_elm_szind_write(tsdn, rtree, elm, szind);
+}
+
+JEMALLOC_ALWAYS_INLINE rtree_leaf_elm_t *
+rtree_leaf_elm_lookup(tsdn_t *tsdn, rtree_t *rtree, rtree_ctx_t *rtree_ctx,
+    uintptr_t key, bool dependent, bool init_missing) {
+	assert(key != 0);
+	assert(!dependent || !init_missing);
+
+	size_t slot = rtree_cache_direct_map(key);
+	uintptr_t leafkey = rtree_leafkey(key);
+	assert(leafkey != RTREE_LEAFKEY_INVALID);
+
+	/* Fast path: L1 direct mapped cache. */
+	if (likely(rtree_ctx->cache[slot].leafkey == leafkey)) {
+		rtree_leaf_elm_t *leaf = rtree_ctx->cache[slot].leaf;
+		assert(leaf != NULL);
+		uintptr_t subkey = rtree_subkey(key, RTREE_HEIGHT-1);
+		return &leaf[subkey];
+	}
+	/*
+	 * Search the L2 LRU cache.  On hit, swap the matching element into the
+	 * slot in L1 cache, and move the position in L2 up by 1.
+	 */
+#define RTREE_CACHE_CHECK_L2(i) do {					\
+	if (likely(rtree_ctx->l2_cache[i].leafkey == leafkey)) {	\
+		rtree_leaf_elm_t *leaf = rtree_ctx->l2_cache[i].leaf;	\
+		assert(leaf != NULL);					\
+		if (i > 0) {						\
+			/* Bubble up by one. */				\
+			rtree_ctx->l2_cache[i].leafkey =		\
+				rtree_ctx->l2_cache[i - 1].leafkey;	\
+			rtree_ctx->l2_cache[i].leaf =			\
+				rtree_ctx->l2_cache[i - 1].leaf;	\
+			rtree_ctx->l2_cache[i - 1].leafkey =		\
+			    rtree_ctx->cache[slot].leafkey;		\
+			rtree_ctx->l2_cache[i - 1].leaf =		\
+			    rtree_ctx->cache[slot].leaf;		\
+		} else {						\
+			rtree_ctx->l2_cache[0].leafkey =		\
+			    rtree_ctx->cache[slot].leafkey;		\
+			rtree_ctx->l2_cache[0].leaf =			\
+			    rtree_ctx->cache[slot].leaf;		\
+		}							\
+		rtree_ctx->cache[slot].leafkey = leafkey;		\
+		rtree_ctx->cache[slot].leaf = leaf;			\
+		uintptr_t subkey = rtree_subkey(key, RTREE_HEIGHT-1);	\
+		return &leaf[subkey];					\
+	}								\
+} while (0)
+	/* Check the first cache entry. */
+	RTREE_CACHE_CHECK_L2(0);
+	/* Search the remaining cache elements. */
+	for (unsigned i = 1; i < RTREE_CTX_NCACHE_L2; i++) {
+		RTREE_CACHE_CHECK_L2(i);
+	}
+#undef RTREE_CACHE_CHECK_L2
+
+	return rtree_leaf_elm_lookup_hard(tsdn, rtree, rtree_ctx, key,
+	    dependent, init_missing);
+}
+
+static inline bool
+rtree_write(tsdn_t *tsdn, rtree_t *rtree, rtree_ctx_t *rtree_ctx, uintptr_t key,
+    extent_t *extent, szind_t szind, bool slab) {
+	/* Use rtree_clear() to set the extent to NULL. */
+	assert(extent != NULL);
+
+	rtree_leaf_elm_t *elm = rtree_leaf_elm_lookup(tsdn, rtree, rtree_ctx,
+	    key, false, true);
+	if (elm == NULL) {
+		return true;
+	}
+
+	assert(rtree_leaf_elm_extent_read(tsdn, rtree, elm, false) == NULL);
+	rtree_leaf_elm_write(tsdn, rtree, elm, extent, szind, slab);
+
+	return false;
+}
+
+JEMALLOC_ALWAYS_INLINE rtree_leaf_elm_t *
+rtree_read(tsdn_t *tsdn, rtree_t *rtree, rtree_ctx_t *rtree_ctx, uintptr_t key,
+    bool dependent) {
+	rtree_leaf_elm_t *elm = rtree_leaf_elm_lookup(tsdn, rtree, rtree_ctx,
+	    key, dependent, false);
+	if (!dependent && elm == NULL) {
+		return NULL;
+	}
+	assert(elm != NULL);
+	return elm;
+}
+
+JEMALLOC_ALWAYS_INLINE extent_t *
+rtree_extent_read(tsdn_t *tsdn, rtree_t *rtree, rtree_ctx_t *rtree_ctx,
+    uintptr_t key, bool dependent) {
+	rtree_leaf_elm_t *elm = rtree_read(tsdn, rtree, rtree_ctx, key,
+	    dependent);
+	if (!dependent && elm == NULL) {
+		return NULL;
+	}
+	return rtree_leaf_elm_extent_read(tsdn, rtree, elm, dependent);
+}
+
+JEMALLOC_ALWAYS_INLINE szind_t
+rtree_szind_read(tsdn_t *tsdn, rtree_t *rtree, rtree_ctx_t *rtree_ctx,
+    uintptr_t key, bool dependent) {
+	rtree_leaf_elm_t *elm = rtree_read(tsdn, rtree, rtree_ctx, key,
+	    dependent);
+	if (!dependent && elm == NULL) {
+		return NSIZES;
+	}
+	return rtree_leaf_elm_szind_read(tsdn, rtree, elm, dependent);
+}
+
+/*
+ * rtree_slab_read() is intentionally omitted because slab is always read in
+ * conjunction with szind, which makes rtree_szind_slab_read() a better choice.
+ */
+
+JEMALLOC_ALWAYS_INLINE bool
+rtree_extent_szind_read(tsdn_t *tsdn, rtree_t *rtree, rtree_ctx_t *rtree_ctx,
+    uintptr_t key, bool dependent, extent_t **r_extent, szind_t *r_szind) {
+	rtree_leaf_elm_t *elm = rtree_read(tsdn, rtree, rtree_ctx, key,
+	    dependent);
+	if (!dependent && elm == NULL) {
+		return true;
+	}
+	*r_extent = rtree_leaf_elm_extent_read(tsdn, rtree, elm, dependent);
+	*r_szind = rtree_leaf_elm_szind_read(tsdn, rtree, elm, dependent);
+	return false;
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+rtree_szind_slab_read(tsdn_t *tsdn, rtree_t *rtree, rtree_ctx_t *rtree_ctx,
+    uintptr_t key, bool dependent, szind_t *r_szind, bool *r_slab) {
+	rtree_leaf_elm_t *elm = rtree_read(tsdn, rtree, rtree_ctx, key,
+	    dependent);
+	if (!dependent && elm == NULL) {
+		return true;
+	}
+	*r_szind = rtree_leaf_elm_szind_read(tsdn, rtree, elm, dependent);
+	*r_slab = rtree_leaf_elm_slab_read(tsdn, rtree, elm, dependent);
+	return false;
+}
+
+static inline void
+rtree_szind_slab_update(tsdn_t *tsdn, rtree_t *rtree, rtree_ctx_t *rtree_ctx,
+    uintptr_t key, szind_t szind, bool slab) {
+	assert(!slab || szind < NBINS);
+
+	rtree_leaf_elm_t *elm = rtree_read(tsdn, rtree, rtree_ctx, key, true);
+	rtree_leaf_elm_szind_slab_update(tsdn, rtree, elm, szind, slab);
+}
+
+static inline void
+rtree_clear(tsdn_t *tsdn, rtree_t *rtree, rtree_ctx_t *rtree_ctx,
+    uintptr_t key) {
+	rtree_leaf_elm_t *elm = rtree_read(tsdn, rtree, rtree_ctx, key, true);
+	assert(rtree_leaf_elm_extent_read(tsdn, rtree, elm, false) !=
+	    NULL);
+	rtree_leaf_elm_write(tsdn, rtree, elm, NULL, NSIZES, false);
+}
+
+#endif /* JEMALLOC_INTERNAL_RTREE_H */

include/jemalloc/internal/rtree_tsd.h

@@ -0,0 +1,50 @@
+#ifndef JEMALLOC_INTERNAL_RTREE_CTX_H
+#define JEMALLOC_INTERNAL_RTREE_CTX_H
+
+/*
+ * Number of leafkey/leaf pairs to cache in L1 and L2 level respectively.  Each
+ * entry supports an entire leaf, so the cache hit rate is typically high even
+ * with a small number of entries.  In rare cases extent activity will straddle
+ * the boundary between two leaf nodes.  Furthermore, an arena may use a
+ * combination of dss and mmap.  Note that as memory usage grows past the amount
+ * that this cache can directly cover, the cache will become less effective if
+ * locality of reference is low, but the consequence is merely cache misses
+ * while traversing the tree nodes.
+ *
+ * The L1 direct mapped cache offers consistent and low cost on cache hit.
+ * However collision could affect hit rate negatively.  This is resolved by
+ * combining with a L2 LRU cache, which requires linear search and re-ordering
+ * on access but suffers no collision.  Note that, the cache will itself suffer
+ * cache misses if made overly large, plus the cost of linear search in the LRU
+ * cache.
+ */
+#define RTREE_CTX_LG_NCACHE 4
+#define RTREE_CTX_NCACHE (1 << RTREE_CTX_LG_NCACHE)
+#define RTREE_CTX_NCACHE_L2 8
+
+/*
+ * Zero initializer required for tsd initialization only.  Proper initialization
+ * done via rtree_ctx_data_init().
+ */
+#define RTREE_CTX_ZERO_INITIALIZER {{{0}}}
+
+
+typedef struct rtree_leaf_elm_s rtree_leaf_elm_t;
+
+typedef struct rtree_ctx_cache_elm_s rtree_ctx_cache_elm_t;
+struct rtree_ctx_cache_elm_s {
+	uintptr_t		leafkey;
+	rtree_leaf_elm_t	*leaf;
+};
+
+typedef struct rtree_ctx_s rtree_ctx_t;
+struct rtree_ctx_s {
+	/* Direct mapped cache. */
+	rtree_ctx_cache_elm_t	cache[RTREE_CTX_NCACHE];
+	/* L2 LRU cache. */
+	rtree_ctx_cache_elm_t	l2_cache[RTREE_CTX_NCACHE_L2];
+};
+
+void rtree_ctx_data_init(rtree_ctx_t *ctx);
+
+#endif /* JEMALLOC_INTERNAL_RTREE_CTX_H */

include/jemalloc/internal/size_classes.sh

@@ -0,0 +1,361 @@
+#!/bin/sh
+#
+# Usage: size_classes.sh <lg_qarr> <lg_tmin> <lg_parr> <lg_g>
+
+# The following limits are chosen such that they cover all supported platforms.
+
+# Pointer sizes.
+lg_zarr="2 3"
+
+# Quanta.
+lg_qarr=$1
+
+# The range of tiny size classes is [2^lg_tmin..2^(lg_q-1)].
+lg_tmin=$2
+
+# Maximum lookup size.
+lg_kmax=12
+
+# Page sizes.
+lg_parr=`echo $3 | tr ',' ' '`
+
+# Size class group size (number of size classes for each size doubling).
+lg_g=$4
+
+pow2() {
+  e=$1
+  pow2_result=1
+  while [ ${e} -gt 0 ] ; do
+    pow2_result=$((${pow2_result} + ${pow2_result}))
+    e=$((${e} - 1))
+  done
+}
+
+lg() {
+  x=$1
+  lg_result=0
+  while [ ${x} -gt 1 ] ; do
+    lg_result=$((${lg_result} + 1))
+    x=$((${x} / 2))
+  done
+}
+
+lg_ceil() {
+  y=$1
+  lg ${y}; lg_floor=${lg_result}
+  pow2 ${lg_floor}; pow2_floor=${pow2_result}
+  if [ ${pow2_floor} -lt ${y} ] ; then
+    lg_ceil_result=$((${lg_floor} + 1))
+  else
+    lg_ceil_result=${lg_floor}
+  fi
+}
+
+reg_size_compute() {
+  lg_grp=$1
+  lg_delta=$2
+  ndelta=$3
+
+  pow2 ${lg_grp}; grp=${pow2_result}
+  pow2 ${lg_delta}; delta=${pow2_result}
+  reg_size=$((${grp} + ${delta}*${ndelta}))
+}
+
+slab_size() {
+  lg_p=$1
+  lg_grp=$2
+  lg_delta=$3
+  ndelta=$4
+
+  pow2 ${lg_p}; p=${pow2_result}
+  reg_size_compute ${lg_grp} ${lg_delta} ${ndelta}
+
+  # Compute smallest slab size that is an integer multiple of reg_size.
+  try_slab_size=${p}
+  try_nregs=$((${try_slab_size} / ${reg_size}))
+  perfect=0
+  while [ ${perfect} -eq 0 ] ; do
+    perfect_slab_size=${try_slab_size}
+    perfect_nregs=${try_nregs}
+
+    try_slab_size=$((${try_slab_size} + ${p}))
+    try_nregs=$((${try_slab_size} / ${reg_size}))
+    if [ ${perfect_slab_size} -eq $((${perfect_nregs} * ${reg_size})) ] ; then
+      perfect=1
+    fi
+  done
+
+  slab_size_pgs=$((${perfect_slab_size} / ${p}))
+}
+
+size_class() {
+  index=$1
+  lg_grp=$2
+  lg_delta=$3
+  ndelta=$4
+  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 ${ndelta} ] ; then
+    rem="yes"
+  else
+    rem="no"
+  fi
+
+  lg_size=${lg_grp}
+  if [ $((${lg_delta} + ${lg_ndelta})) -eq ${lg_grp} ] ; then
+    lg_size=$((${lg_grp} + 1))
+  else
+    lg_size=${lg_grp}
+    rem="yes"
+  fi
+
+  if [ ${lg_size} -lt $((${lg_p} + ${lg_g})) ] ; then
+    bin="yes"
+    slab_size ${lg_p} ${lg_grp} ${lg_delta} ${ndelta}; pgs=${slab_size_pgs}
+  else
+    bin="no"
+    pgs=0
+  fi
+  if [ ${lg_size} -lt ${lg_kmax} \
+      -o ${lg_size} -eq ${lg_kmax} -a ${rem} = "no" ] ; then
+    lg_delta_lookup=${lg_delta}
+  else
+    lg_delta_lookup="no"
+  fi
+  printf '    SC(%3d, %6d, %8d, %6d, %3s, %3s, %3d, %2s) \\\n' ${index} ${lg_grp} ${lg_delta} ${ndelta} ${psz} ${bin} ${pgs} ${lg_delta_lookup}
+  # Defined upon return:
+  # - psz ("yes" or "no")
+  # - bin ("yes" or "no")
+  # - pgs
+  # - lg_delta_lookup (${lg_delta} or "no")
+}
+
+sep_line() {
+  echo "                                                         \\"
+}
+
+size_classes() {
+  lg_z=$1
+  lg_q=$2
+  lg_t=$3
+  lg_p=$4
+  lg_g=$5
+
+  pow2 $((${lg_z} + 3)); ptr_bits=${pow2_result}
+  pow2 ${lg_g}; g=${pow2_result}
+
+  echo "#define SIZE_CLASSES \\"
+  echo "  /* index, lg_grp, lg_delta, ndelta, psz, bin, pgs, lg_delta_lookup */ \\"
+
+  ntbins=0
+  nlbins=0
+  lg_tiny_maxclass='"NA"'
+  nbins=0
+  npsizes=0
+
+  # Tiny size classes.
+  ndelta=0
+  index=0
+  lg_grp=${lg_t}
+  lg_delta=${lg_grp}
+  while [ ${lg_grp} -lt ${lg_q} ] ; do
+    size_class ${index} ${lg_grp} ${lg_delta} ${ndelta} ${lg_p} ${lg_kmax}
+    if [ ${lg_delta_lookup} != "no" ] ; then
+      nlbins=$((${index} + 1))
+    fi
+    if [ ${psz} = "yes" ] ; then
+      npsizes=$((${npsizes} + 1))
+    fi
+    if [ ${bin} != "no" ] ; then
+      nbins=$((${index} + 1))
+    fi
+    ntbins=$((${ntbins} + 1))
+    lg_tiny_maxclass=${lg_grp} # Final written value is correct.
+    index=$((${index} + 1))
+    lg_delta=${lg_grp}
+    lg_grp=$((${lg_grp} + 1))
+  done
+
+  # First non-tiny group.
+  if [ ${ntbins} -gt 0 ] ; then
+    sep_line
+    # The first size class has an unusual encoding, because the size has to be
+    # split between grp and delta*ndelta.
+    lg_grp=$((${lg_grp} - 1))
+    ndelta=1
+    size_class ${index} ${lg_grp} ${lg_delta} ${ndelta} ${lg_p} ${lg_kmax}
+    index=$((${index} + 1))
+    lg_grp=$((${lg_grp} + 1))
+    lg_delta=$((${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 [ ${psz} = "yes" ] ; then
+      npsizes=$((${npsizes} + 1))
+    fi
+  done
+
+  # All remaining groups.
+  lg_grp=$((${lg_grp} + ${lg_g}))
+  while [ ${lg_grp} -lt $((${ptr_bits} - 1)) ] ; do
+    sep_line
+    ndelta=1
+    if [ ${lg_grp} -eq $((${ptr_bits} - 2)) ] ; then
+      ndelta_limit=$((${g} - 1))
+    else
+      ndelta_limit=${g}
+    fi
+    while [ ${ndelta} -le ${ndelta_limit} ] ; do
+      size_class ${index} ${lg_grp} ${lg_delta} ${ndelta} ${lg_p} ${lg_kmax}
+      if [ ${lg_delta_lookup} != "no" ] ; then
+        nlbins=$((${index} + 1))
+        # 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:
+        small_maxclass="((((size_t)1) << ${lg_grp}) + (((size_t)${ndelta}) << ${lg_delta}))"
+        if [ ${lg_g} -gt 0 ] ; then
+          lg_large_minclass=$((${lg_grp} + 1))
+        else
+          lg_large_minclass=$((${lg_grp} + 2))
+        fi
+      fi
+      # Final written value is correct:
+      large_maxclass="((((size_t)1) << ${lg_grp}) + (((size_t)${ndelta}) << ${lg_delta}))"
+      index=$((${index} + 1))
+      ndelta=$((${ndelta} + 1))
+    done
+    lg_grp=$((${lg_grp} + 1))
+    lg_delta=$((${lg_delta} + 1))
+  done
+  echo
+  nsizes=${index}
+  lg_ceil ${nsizes}; lg_ceil_nsizes=${lg_ceil_result}
+
+  # Defined upon completion:
+  # - ntbins
+  # - nlbins
+  # - nbins
+  # - nsizes
+  # - lg_ceil_nsizes
+  # - npsizes
+  # - lg_tiny_maxclass
+  # - lookup_maxclass
+  # - small_maxclass
+  # - lg_large_minclass
+  # - large_maxclass
+}
+
+cat <<EOF
+#ifndef JEMALLOC_INTERNAL_SIZE_CLASSES_H
+#define JEMALLOC_INTERNAL_SIZE_CLASSES_H
+
+/* This file was automatically generated by size_classes.sh. */
+
+#include "jemalloc/internal/jemalloc_internal_types.h"
+
+/*
+ * This header file defines:
+ *
+ *   LG_SIZE_CLASS_GROUP: Lg of size class count for each size doubling.
+ *   LG_TINY_MIN: Lg of minimum size class to support.
+ *   SIZE_CLASSES: Complete table of SC(index, lg_grp, lg_delta, ndelta, psz,
+ *                 bin, pgs, 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.
+ *     pgs: Slab page count if a small bin size class, 0 otherwise.
+ *     lg_delta_lookup: Same as lg_delta if a lookup table size class, 'no'
+ *                      otherwise.
+ *   NTBINS: Number of tiny bins.
+ *   NLBINS: Number of bins supported by the lookup table.
+ *   NBINS: Number of small size class bins.
+ *   NSIZES: Number of size classes.
+ *   LG_CEIL_NSIZES: Number of bits required to store NSIZES.
+ *   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.
+ *   LG_LARGE_MINCLASS: Lg of minimum large size class.
+ *   LARGE_MAXCLASS: Maximum (large) size class.
+ */
+
+#define LG_SIZE_CLASS_GROUP	${lg_g}
+#define LG_TINY_MIN		${lg_tmin}
+
+EOF
+
+for lg_z in ${lg_zarr} ; do
+  for lg_q in ${lg_qarr} ; do
+    lg_t=${lg_tmin}
+    while [ ${lg_t} -le ${lg_q} ] ; do
+      # Iterate through page sizes and compute how many bins there are.
+      for lg_p in ${lg_parr} ; do
+        echo "#if (LG_SIZEOF_PTR == ${lg_z} && LG_TINY_MIN == ${lg_t} && LG_QUANTUM == ${lg_q} && LG_PAGE == ${lg_p})"
+        size_classes ${lg_z} ${lg_q} ${lg_t} ${lg_p} ${lg_g}
+        echo "#define SIZE_CLASSES_DEFINED"
+        echo "#define NTBINS			${ntbins}"
+        echo "#define NLBINS			${nlbins}"
+        echo "#define NBINS			${nbins}"
+        echo "#define NSIZES			${nsizes}"
+        echo "#define LG_CEIL_NSIZES		${lg_ceil_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}"
+        echo "#define LG_LARGE_MINCLASS	${lg_large_minclass}"
+        echo "#define LARGE_MINCLASS		(ZU(1) << LG_LARGE_MINCLASS)"
+        echo "#define LARGE_MAXCLASS		${large_maxclass}"
+        echo "#endif"
+        echo
+      done
+      lg_t=$((${lg_t} + 1))
+    done
+  done
+done
+
+cat <<EOF
+#ifndef SIZE_CLASSES_DEFINED
+#  error "No size class definitions match configuration"
+#endif
+#undef SIZE_CLASSES_DEFINED
+/*
+ * The size2index_tab lookup table uses uint8_t to encode each bin index, so we
+ * cannot support more than 256 small size classes.
+ */
+#if (NBINS > 256)
+#  error "Too many small size classes"
+#endif
+
+#endif /* JEMALLOC_INTERNAL_SIZE_CLASSES_H */
+EOF

include/jemalloc/internal/smoothstep.h

@@ -0,0 +1,232 @@
+#ifndef JEMALLOC_INTERNAL_SMOOTHSTEP_H
+#define JEMALLOC_INTERNAL_SMOOTHSTEP_H
+
+/*
+ * This file was generated by the following command:
+ *   sh smoothstep.sh smoother 200 24 3 15
+ */
+/******************************************************************************/
+
+/*
+ * This header defines a precomputed table based on the smoothstep family of
+ * sigmoidal curves (https://en.wikipedia.org/wiki/Smoothstep) that grow from 0
+ * to 1 in 0 <= x <= 1.  The table is stored as integer fixed point values so
+ * that floating point math can be avoided.
+ *
+ *                      3     2
+ *   smoothstep(x) = -2x  + 3x
+ *
+ *                       5      4      3
+ *   smootherstep(x) = 6x  - 15x  + 10x
+ *
+ *                          7      6      5      4
+ *   smootheststep(x) = -20x  + 70x  - 84x  + 35x
+ */
+
+#define SMOOTHSTEP_VARIANT	"smoother"
+#define SMOOTHSTEP_NSTEPS	200
+#define SMOOTHSTEP_BFP		24
+#define SMOOTHSTEP \
+ /* STEP(step, h,                            x,     y) */ \
+    STEP(   1, UINT64_C(0x0000000000000014), 0.005, 0.000001240643750) \
+    STEP(   2, UINT64_C(0x00000000000000a5), 0.010, 0.000009850600000) \
+    STEP(   3, UINT64_C(0x0000000000000229), 0.015, 0.000032995181250) \
+    STEP(   4, UINT64_C(0x0000000000000516), 0.020, 0.000077619200000) \
+    STEP(   5, UINT64_C(0x00000000000009dc), 0.025, 0.000150449218750) \
+    STEP(   6, UINT64_C(0x00000000000010e8), 0.030, 0.000257995800000) \
+    STEP(   7, UINT64_C(0x0000000000001aa4), 0.035, 0.000406555756250) \
+    STEP(   8, UINT64_C(0x0000000000002777), 0.040, 0.000602214400000) \
+    STEP(   9, UINT64_C(0x00000000000037c2), 0.045, 0.000850847793750) \
+    STEP(  10, UINT64_C(0x0000000000004be6), 0.050, 0.001158125000000) \
+    STEP(  11, UINT64_C(0x000000000000643c), 0.055, 0.001529510331250) \
+    STEP(  12, UINT64_C(0x000000000000811f), 0.060, 0.001970265600000) \
+    STEP(  13, UINT64_C(0x000000000000a2e2), 0.065, 0.002485452368750) \
+    STEP(  14, UINT64_C(0x000000000000c9d8), 0.070, 0.003079934200000) \
+    STEP(  15, UINT64_C(0x000000000000f64f), 0.075, 0.003758378906250) \
+    STEP(  16, UINT64_C(0x0000000000012891), 0.080, 0.004525260800000) \
+    STEP(  17, UINT64_C(0x00000000000160e7), 0.085, 0.005384862943750) \
+    STEP(  18, UINT64_C(0x0000000000019f95), 0.090, 0.006341279400000) \
+    STEP(  19, UINT64_C(0x000000000001e4dc), 0.095, 0.007398417481250) \
+    STEP(  20, UINT64_C(0x00000000000230fc), 0.100, 0.008560000000000) \
+    STEP(  21, UINT64_C(0x0000000000028430), 0.105, 0.009829567518750) \
+    STEP(  22, UINT64_C(0x000000000002deb0), 0.110, 0.011210480600000) \
+    STEP(  23, UINT64_C(0x00000000000340b1), 0.115, 0.012705922056250) \
+    STEP(  24, UINT64_C(0x000000000003aa67), 0.120, 0.014318899200000) \
+    STEP(  25, UINT64_C(0x0000000000041c00), 0.125, 0.016052246093750) \
+    STEP(  26, UINT64_C(0x00000000000495a8), 0.130, 0.017908625800000) \
+    STEP(  27, UINT64_C(0x000000000005178b), 0.135, 0.019890532631250) \
+    STEP(  28, UINT64_C(0x000000000005a1cf), 0.140, 0.022000294400000) \
+    STEP(  29, UINT64_C(0x0000000000063498), 0.145, 0.024240074668750) \
+    STEP(  30, UINT64_C(0x000000000006d009), 0.150, 0.026611875000000) \
+    STEP(  31, UINT64_C(0x000000000007743f), 0.155, 0.029117537206250) \
+    STEP(  32, UINT64_C(0x0000000000082157), 0.160, 0.031758745600000) \
+    STEP(  33, UINT64_C(0x000000000008d76b), 0.165, 0.034537029243750) \
+    STEP(  34, UINT64_C(0x0000000000099691), 0.170, 0.037453764200000) \
+    STEP(  35, UINT64_C(0x00000000000a5edf), 0.175, 0.040510175781250) \
+    STEP(  36, UINT64_C(0x00000000000b3067), 0.180, 0.043707340800000) \
+    STEP(  37, UINT64_C(0x00000000000c0b38), 0.185, 0.047046189818750) \
+    STEP(  38, UINT64_C(0x00000000000cef5e), 0.190, 0.050527509400000) \
+    STEP(  39, UINT64_C(0x00000000000ddce6), 0.195, 0.054151944356250) \
+    STEP(  40, UINT64_C(0x00000000000ed3d8), 0.200, 0.057920000000000) \
+    STEP(  41, UINT64_C(0x00000000000fd439), 0.205, 0.061832044393750) \
+    STEP(  42, UINT64_C(0x000000000010de0e), 0.210, 0.065888310600000) \
+    STEP(  43, UINT64_C(0x000000000011f158), 0.215, 0.070088898931250) \
+    STEP(  44, UINT64_C(0x0000000000130e17), 0.220, 0.074433779200000) \
+    STEP(  45, UINT64_C(0x0000000000143448), 0.225, 0.078922792968750) \
+    STEP(  46, UINT64_C(0x00000000001563e7), 0.230, 0.083555655800000) \
+    STEP(  47, UINT64_C(0x0000000000169cec), 0.235, 0.088331959506250) \
+    STEP(  48, UINT64_C(0x000000000017df4f), 0.240, 0.093251174400000) \
+    STEP(  49, UINT64_C(0x0000000000192b04), 0.245, 0.098312651543750) \
+    STEP(  50, UINT64_C(0x00000000001a8000), 0.250, 0.103515625000000) \
+    STEP(  51, UINT64_C(0x00000000001bde32), 0.255, 0.108859214081250) \
+    STEP(  52, UINT64_C(0x00000000001d458b), 0.260, 0.114342425600000) \
+    STEP(  53, UINT64_C(0x00000000001eb5f8), 0.265, 0.119964156118750) \
+    STEP(  54, UINT64_C(0x0000000000202f65), 0.270, 0.125723194200000) \
+    STEP(  55, UINT64_C(0x000000000021b1bb), 0.275, 0.131618222656250) \
+    STEP(  56, UINT64_C(0x0000000000233ce3), 0.280, 0.137647820800000) \
+    STEP(  57, UINT64_C(0x000000000024d0c3), 0.285, 0.143810466693750) \
+    STEP(  58, UINT64_C(0x0000000000266d40), 0.290, 0.150104539400000) \
+    STEP(  59, UINT64_C(0x000000000028123d), 0.295, 0.156528321231250) \
+    STEP(  60, UINT64_C(0x000000000029bf9c), 0.300, 0.163080000000000) \
+    STEP(  61, UINT64_C(0x00000000002b753d), 0.305, 0.169757671268750) \
+    STEP(  62, UINT64_C(0x00000000002d32fe), 0.310, 0.176559340600000) \
+    STEP(  63, UINT64_C(0x00000000002ef8bc), 0.315, 0.183482925806250) \
+    STEP(  64, UINT64_C(0x000000000030c654), 0.320, 0.190526259200000) \
+    STEP(  65, UINT64_C(0x0000000000329b9f), 0.325, 0.197687089843750) \
+    STEP(  66, UINT64_C(0x0000000000347875), 0.330, 0.204963085800000) \
+    STEP(  67, UINT64_C(0x0000000000365cb0), 0.335, 0.212351836381250) \
+    STEP(  68, UINT64_C(0x0000000000384825), 0.340, 0.219850854400000) \
+    STEP(  69, UINT64_C(0x00000000003a3aa8), 0.345, 0.227457578418750) \
+    STEP(  70, UINT64_C(0x00000000003c340f), 0.350, 0.235169375000000) \
+    STEP(  71, UINT64_C(0x00000000003e342b), 0.355, 0.242983540956250) \
+    STEP(  72, UINT64_C(0x0000000000403ace), 0.360, 0.250897305600000) \
+    STEP(  73, UINT64_C(0x00000000004247c8), 0.365, 0.258907832993750) \
+    STEP(  74, UINT64_C(0x0000000000445ae9), 0.370, 0.267012224200000) \
+    STEP(  75, UINT64_C(0x0000000000467400), 0.375, 0.275207519531250) \
+    STEP(  76, UINT64_C(0x00000000004892d8), 0.380, 0.283490700800000) \
+    STEP(  77, UINT64_C(0x00000000004ab740), 0.385, 0.291858693568750) \
+    STEP(  78, UINT64_C(0x00000000004ce102), 0.390, 0.300308369400000) \
+    STEP(  79, UINT64_C(0x00000000004f0fe9), 0.395, 0.308836548106250) \
+    STEP(  80, UINT64_C(0x00000000005143bf), 0.400, 0.317440000000000) \
+    STEP(  81, UINT64_C(0x0000000000537c4d), 0.405, 0.326115448143750) \
+    STEP(  82, UINT64_C(0x000000000055b95b), 0.410, 0.334859570600000) \
+    STEP(  83, UINT64_C(0x000000000057fab1), 0.415, 0.343669002681250) \
+    STEP(  84, UINT64_C(0x00000000005a4015), 0.420, 0.352540339200000) \
+    STEP(  85, UINT64_C(0x00000000005c894e), 0.425, 0.361470136718750) \
+    STEP(  86, UINT64_C(0x00000000005ed622), 0.430, 0.370454915800000) \
+    STEP(  87, UINT64_C(0x0000000000612655), 0.435, 0.379491163256250) \
+    STEP(  88, UINT64_C(0x00000000006379ac), 0.440, 0.388575334400000) \
+    STEP(  89, UINT64_C(0x000000000065cfeb), 0.445, 0.397703855293750) \
+    STEP(  90, UINT64_C(0x00000000006828d6), 0.450, 0.406873125000000) \
+    STEP(  91, UINT64_C(0x00000000006a842f), 0.455, 0.416079517831250) \
+    STEP(  92, UINT64_C(0x00000000006ce1bb), 0.460, 0.425319385600000) \
+    STEP(  93, UINT64_C(0x00000000006f413a), 0.465, 0.434589059868750) \
+    STEP(  94, UINT64_C(0x000000000071a270), 0.470, 0.443884854200000) \
+    STEP(  95, UINT64_C(0x000000000074051d), 0.475, 0.453203066406250) \
+    STEP(  96, UINT64_C(0x0000000000766905), 0.480, 0.462539980800000) \
+    STEP(  97, UINT64_C(0x000000000078cde7), 0.485, 0.471891870443750) \
+    STEP(  98, UINT64_C(0x00000000007b3387), 0.490, 0.481254999400000) \
+    STEP(  99, UINT64_C(0x00000000007d99a4), 0.495, 0.490625624981250) \
+    STEP( 100, UINT64_C(0x0000000000800000), 0.500, 0.500000000000000) \
+    STEP( 101, UINT64_C(0x000000000082665b), 0.505, 0.509374375018750) \
+    STEP( 102, UINT64_C(0x000000000084cc78), 0.510, 0.518745000600000) \
+    STEP( 103, UINT64_C(0x0000000000873218), 0.515, 0.528108129556250) \
+    STEP( 104, UINT64_C(0x00000000008996fa), 0.520, 0.537460019200000) \
+    STEP( 105, UINT64_C(0x00000000008bfae2), 0.525, 0.546796933593750) \
+    STEP( 106, UINT64_C(0x00000000008e5d8f), 0.530, 0.556115145800000) \
+    STEP( 107, UINT64_C(0x000000000090bec5), 0.535, 0.565410940131250) \
+    STEP( 108, UINT64_C(0x0000000000931e44), 0.540, 0.574680614400000) \
+    STEP( 109, UINT64_C(0x0000000000957bd0), 0.545, 0.583920482168750) \
+    STEP( 110, UINT64_C(0x000000000097d729), 0.550, 0.593126875000000) \
+    STEP( 111, UINT64_C(0x00000000009a3014), 0.555, 0.602296144706250) \
+    STEP( 112, UINT64_C(0x00000000009c8653), 0.560, 0.611424665600000) \
+    STEP( 113, UINT64_C(0x00000000009ed9aa), 0.565, 0.620508836743750) \
+    STEP( 114, UINT64_C(0x0000000000a129dd), 0.570, 0.629545084200000) \
+    STEP( 115, UINT64_C(0x0000000000a376b1), 0.575, 0.638529863281250) \
+    STEP( 116, UINT64_C(0x0000000000a5bfea), 0.580, 0.647459660800000) \
+    STEP( 117, UINT64_C(0x0000000000a8054e), 0.585, 0.656330997318750) \
+    STEP( 118, UINT64_C(0x0000000000aa46a4), 0.590, 0.665140429400000) \
+    STEP( 119, UINT64_C(0x0000000000ac83b2), 0.595, 0.673884551856250) \
+    STEP( 120, UINT64_C(0x0000000000aebc40), 0.600, 0.682560000000000) \
+    STEP( 121, UINT64_C(0x0000000000b0f016), 0.605, 0.691163451893750) \
+    STEP( 122, UINT64_C(0x0000000000b31efd), 0.610, 0.699691630600000) \
+    STEP( 123, UINT64_C(0x0000000000b548bf), 0.615, 0.708141306431250) \
+    STEP( 124, UINT64_C(0x0000000000b76d27), 0.620, 0.716509299200000) \
+    STEP( 125, UINT64_C(0x0000000000b98c00), 0.625, 0.724792480468750) \
+    STEP( 126, UINT64_C(0x0000000000bba516), 0.630, 0.732987775800000) \
+    STEP( 127, UINT64_C(0x0000000000bdb837), 0.635, 0.741092167006250) \
+    STEP( 128, UINT64_C(0x0000000000bfc531), 0.640, 0.749102694400000) \
+    STEP( 129, UINT64_C(0x0000000000c1cbd4), 0.645, 0.757016459043750) \
+    STEP( 130, UINT64_C(0x0000000000c3cbf0), 0.650, 0.764830625000000) \
+    STEP( 131, UINT64_C(0x0000000000c5c557), 0.655, 0.772542421581250) \
+    STEP( 132, UINT64_C(0x0000000000c7b7da), 0.660, 0.780149145600000) \
+    STEP( 133, UINT64_C(0x0000000000c9a34f), 0.665, 0.787648163618750) \
+    STEP( 134, UINT64_C(0x0000000000cb878a), 0.670, 0.795036914200000) \
+    STEP( 135, UINT64_C(0x0000000000cd6460), 0.675, 0.802312910156250) \
+    STEP( 136, UINT64_C(0x0000000000cf39ab), 0.680, 0.809473740800000) \
+    STEP( 137, UINT64_C(0x0000000000d10743), 0.685, 0.816517074193750) \
+    STEP( 138, UINT64_C(0x0000000000d2cd01), 0.690, 0.823440659400000) \
+    STEP( 139, UINT64_C(0x0000000000d48ac2), 0.695, 0.830242328731250) \
+    STEP( 140, UINT64_C(0x0000000000d64063), 0.700, 0.836920000000000) \
+    STEP( 141, UINT64_C(0x0000000000d7edc2), 0.705, 0.843471678768750) \
+    STEP( 142, UINT64_C(0x0000000000d992bf), 0.710, 0.849895460600000) \
+    STEP( 143, UINT64_C(0x0000000000db2f3c), 0.715, 0.856189533306250) \
+    STEP( 144, UINT64_C(0x0000000000dcc31c), 0.720, 0.862352179200000) \
+    STEP( 145, UINT64_C(0x0000000000de4e44), 0.725, 0.868381777343750) \
+    STEP( 146, UINT64_C(0x0000000000dfd09a), 0.730, 0.874276805800000) \
+    STEP( 147, UINT64_C(0x0000000000e14a07), 0.735, 0.880035843881250) \
+    STEP( 148, UINT64_C(0x0000000000e2ba74), 0.740, 0.885657574400000) \
+    STEP( 149, UINT64_C(0x0000000000e421cd), 0.745, 0.891140785918750) \
+    STEP( 150, UINT64_C(0x0000000000e58000), 0.750, 0.896484375000000) \
+    STEP( 151, UINT64_C(0x0000000000e6d4fb), 0.755, 0.901687348456250) \
+    STEP( 152, UINT64_C(0x0000000000e820b0), 0.760, 0.906748825600000) \
+    STEP( 153, UINT64_C(0x0000000000e96313), 0.765, 0.911668040493750) \
+    STEP( 154, UINT64_C(0x0000000000ea9c18), 0.770, 0.916444344200000) \
+    STEP( 155, UINT64_C(0x0000000000ebcbb7), 0.775, 0.921077207031250) \
+    STEP( 156, UINT64_C(0x0000000000ecf1e8), 0.780, 0.925566220800000) \
+    STEP( 157, UINT64_C(0x0000000000ee0ea7), 0.785, 0.929911101068750) \
+    STEP( 158, UINT64_C(0x0000000000ef21f1), 0.790, 0.934111689400000) \
+    STEP( 159, UINT64_C(0x0000000000f02bc6), 0.795, 0.938167955606250) \
+    STEP( 160, UINT64_C(0x0000000000f12c27), 0.800, 0.942080000000000) \
+    STEP( 161, UINT64_C(0x0000000000f22319), 0.805, 0.945848055643750) \
+    STEP( 162, UINT64_C(0x0000000000f310a1), 0.810, 0.949472490600000) \
+    STEP( 163, UINT64_C(0x0000000000f3f4c7), 0.815, 0.952953810181250) \
+    STEP( 164, UINT64_C(0x0000000000f4cf98), 0.820, 0.956292659200000) \
+    STEP( 165, UINT64_C(0x0000000000f5a120), 0.825, 0.959489824218750) \
+    STEP( 166, UINT64_C(0x0000000000f6696e), 0.830, 0.962546235800000) \
+    STEP( 167, UINT64_C(0x0000000000f72894), 0.835, 0.965462970756250) \
+    STEP( 168, UINT64_C(0x0000000000f7dea8), 0.840, 0.968241254400000) \
+    STEP( 169, UINT64_C(0x0000000000f88bc0), 0.845, 0.970882462793750) \
+    STEP( 170, UINT64_C(0x0000000000f92ff6), 0.850, 0.973388125000000) \
+    STEP( 171, UINT64_C(0x0000000000f9cb67), 0.855, 0.975759925331250) \
+    STEP( 172, UINT64_C(0x0000000000fa5e30), 0.860, 0.977999705600000) \
+    STEP( 173, UINT64_C(0x0000000000fae874), 0.865, 0.980109467368750) \
+    STEP( 174, UINT64_C(0x0000000000fb6a57), 0.870, 0.982091374200000) \
+    STEP( 175, UINT64_C(0x0000000000fbe400), 0.875, 0.983947753906250) \
+    STEP( 176, UINT64_C(0x0000000000fc5598), 0.880, 0.985681100800000) \
+    STEP( 177, UINT64_C(0x0000000000fcbf4e), 0.885, 0.987294077943750) \
+    STEP( 178, UINT64_C(0x0000000000fd214f), 0.890, 0.988789519400000) \
+    STEP( 179, UINT64_C(0x0000000000fd7bcf), 0.895, 0.990170432481250) \
+    STEP( 180, UINT64_C(0x0000000000fdcf03), 0.900, 0.991440000000000) \
+    STEP( 181, UINT64_C(0x0000000000fe1b23), 0.905, 0.992601582518750) \
+    STEP( 182, UINT64_C(0x0000000000fe606a), 0.910, 0.993658720600000) \
+    STEP( 183, UINT64_C(0x0000000000fe9f18), 0.915, 0.994615137056250) \
+    STEP( 184, UINT64_C(0x0000000000fed76e), 0.920, 0.995474739200000) \
+    STEP( 185, UINT64_C(0x0000000000ff09b0), 0.925, 0.996241621093750) \
+    STEP( 186, UINT64_C(0x0000000000ff3627), 0.930, 0.996920065800000) \
+    STEP( 187, UINT64_C(0x0000000000ff5d1d), 0.935, 0.997514547631250) \
+    STEP( 188, UINT64_C(0x0000000000ff7ee0), 0.940, 0.998029734400000) \
+    STEP( 189, UINT64_C(0x0000000000ff9bc3), 0.945, 0.998470489668750) \
+    STEP( 190, UINT64_C(0x0000000000ffb419), 0.950, 0.998841875000000) \
+    STEP( 191, UINT64_C(0x0000000000ffc83d), 0.955, 0.999149152206250) \
+    STEP( 192, UINT64_C(0x0000000000ffd888), 0.960, 0.999397785600000) \
+    STEP( 193, UINT64_C(0x0000000000ffe55b), 0.965, 0.999593444243750) \
+    STEP( 194, UINT64_C(0x0000000000ffef17), 0.970, 0.999742004200000) \
+    STEP( 195, UINT64_C(0x0000000000fff623), 0.975, 0.999849550781250) \
+    STEP( 196, UINT64_C(0x0000000000fffae9), 0.980, 0.999922380800000) \
+    STEP( 197, UINT64_C(0x0000000000fffdd6), 0.985, 0.999967004818750) \
+    STEP( 198, UINT64_C(0x0000000000ffff5a), 0.990, 0.999990149400000) \
+    STEP( 199, UINT64_C(0x0000000000ffffeb), 0.995, 0.999998759356250) \
+    STEP( 200, UINT64_C(0x0000000001000000), 1.000, 1.000000000000000) \
+
+#endif /* JEMALLOC_INTERNAL_SMOOTHSTEP_H */

include/jemalloc/internal/smoothstep.sh

@@ -0,0 +1,101 @@
+#!/bin/sh
+#
+# Generate a discrete lookup table for a sigmoid function in the smoothstep
+# family (https://en.wikipedia.org/wiki/Smoothstep), where the lookup table
+# entries correspond to x in [1/nsteps, 2/nsteps, ..., nsteps/nsteps].  Encode
+# the entries using a binary fixed point representation.
+#
+# Usage: smoothstep.sh <variant> <nsteps> <bfp> <xprec> <yprec>
+#
+#        <variant> is in {smooth, smoother, smoothest}.
+#        <nsteps> must be greater than zero.
+#        <bfp> must be in [0..62]; reasonable values are roughly [10..30].
+#        <xprec> is x decimal precision.
+#        <yprec> is y decimal precision.
+
+#set -x
+
+cmd="sh smoothstep.sh $*"
+variant=$1
+nsteps=$2
+bfp=$3
+xprec=$4
+yprec=$5
+
+case "${variant}" in
+  smooth)
+    ;;
+  smoother)
+    ;;
+  smoothest)
+    ;;
+  *)
+    echo "Unsupported variant"
+    exit 1
+    ;;
+esac
+
+smooth() {
+  step=$1
+  y=`echo ${yprec} k ${step} ${nsteps} / sx _2 lx 3 ^ '*' 3 lx 2 ^ '*' + p | dc | tr -d '\\\\\n' | sed -e 's#^\.#0.#g'`
+  h=`echo ${yprec} k 2 ${bfp} ^ ${y} '*' p | dc | tr -d '\\\\\n' | sed -e 's#^\.#0.#g' | tr '.' ' ' | awk '{print $1}' `
+}
+
+smoother() {
+  step=$1
+  y=`echo ${yprec} k ${step} ${nsteps} / sx 6 lx 5 ^ '*' _15 lx 4 ^ '*' + 10 lx 3 ^ '*' + p | dc | tr -d '\\\\\n' | sed -e 's#^\.#0.#g'`
+  h=`echo ${yprec} k 2 ${bfp} ^ ${y} '*' p | dc | tr -d '\\\\\n' | sed -e 's#^\.#0.#g' | tr '.' ' ' | awk '{print $1}' `
+}
+
+smoothest() {
+  step=$1
+  y=`echo ${yprec} k ${step} ${nsteps} / sx _20 lx 7 ^ '*' 70 lx 6 ^ '*' + _84 lx 5 ^ '*' + 35 lx 4 ^ '*' + p | dc | tr -d '\\\\\n' | sed -e 's#^\.#0.#g'`
+  h=`echo ${yprec} k 2 ${bfp} ^ ${y} '*' p | dc | tr -d '\\\\\n' | sed -e 's#^\.#0.#g' | tr '.' ' ' | awk '{print $1}' `
+}
+
+cat <<EOF
+#ifndef JEMALLOC_INTERNAL_SMOOTHSTEP_H
+#define JEMALLOC_INTERNAL_SMOOTHSTEP_H
+
+/*
+ * This file was generated by the following command:
+ *   $cmd
+ */
+/******************************************************************************/
+
+/*
+ * This header defines a precomputed table based on the smoothstep family of
+ * sigmoidal curves (https://en.wikipedia.org/wiki/Smoothstep) that grow from 0
+ * to 1 in 0 <= x <= 1.  The table is stored as integer fixed point values so
+ * that floating point math can be avoided.
+ *
+ *                      3     2
+ *   smoothstep(x) = -2x  + 3x
+ *
+ *                       5      4      3
+ *   smootherstep(x) = 6x  - 15x  + 10x
+ *
+ *                          7      6      5      4
+ *   smootheststep(x) = -20x  + 70x  - 84x  + 35x
+ */
+
+#define SMOOTHSTEP_VARIANT	"${variant}"
+#define SMOOTHSTEP_NSTEPS	${nsteps}
+#define SMOOTHSTEP_BFP		${bfp}
+#define SMOOTHSTEP \\
+ /* STEP(step, h,                            x,     y) */ \\
+EOF
+
+s=1
+while [ $s -le $nsteps ] ; do
+  $variant ${s}
+  x=`echo ${xprec} k ${s} ${nsteps} / p | dc | tr -d '\\\\\n' | sed -e 's#^\.#0.#g'`
+  printf '    STEP(%4d, UINT64_C(0x%016x), %s, %s) \\\n' ${s} ${h} ${x} ${y}
+
+  s=$((s+1))
+done
+echo
+
+cat <<EOF
+#endif /* JEMALLOC_INTERNAL_SMOOTHSTEP_H */
+EOF

include/jemalloc/internal/spin.h

@@ -0,0 +1,36 @@
+#ifndef JEMALLOC_INTERNAL_SPIN_H
+#define JEMALLOC_INTERNAL_SPIN_H
+
+#ifdef JEMALLOC_SPIN_C_
+#  define SPIN_INLINE extern inline
+#else
+#  define SPIN_INLINE inline
+#endif
+
+#define SPIN_INITIALIZER {0U}
+
+typedef struct {
+	unsigned iteration;
+} spin_t;
+
+SPIN_INLINE void
+spin_adaptive(spin_t *spin) {
+	volatile uint32_t i;
+
+	if (spin->iteration < 5) {
+		for (i = 0; i < (1U << spin->iteration); i++) {
+			CPU_SPINWAIT;
+		}
+		spin->iteration++;
+	} else {
+#ifdef _WIN32
+		SwitchToThread();
+#else
+		sched_yield();
+#endif
+	}
+}
+
+#undef SPIN_INLINE
+
+#endif /* JEMALLOC_INTERNAL_SPIN_H */

include/jemalloc/internal/stats.h

@@ -0,0 +1,164 @@
+#ifndef JEMALLOC_INTERNAL_STATS_H
+#define JEMALLOC_INTERNAL_STATS_H
+
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/mutex_prof.h"
+#include "jemalloc/internal/mutex.h"
+#include "jemalloc/internal/size_classes.h"
+#include "jemalloc/internal/stats_tsd.h"
+
+/*  OPTION(opt,		var_name,	default,	set_value_to) */
+#define STATS_PRINT_OPTIONS						\
+    OPTION('J',		json,		false,		true)		\
+    OPTION('g',		general,	true,		false)		\
+    OPTION('m',		merged,		config_stats,	false)		\
+    OPTION('d',		destroyed,	config_stats,	false)		\
+    OPTION('a',		unmerged,	config_stats,	false)		\
+    OPTION('b',		bins,		true,		false)		\
+    OPTION('l',		large,		true,		false)		\
+    OPTION('x',		mutex,		true,		false)
+
+enum {
+#define OPTION(o, v, d, s) stats_print_option_num_##v,
+    STATS_PRINT_OPTIONS
+#undef OPTION
+    stats_print_tot_num_options
+};
+
+/* Options for stats_print. */
+extern bool opt_stats_print;
+extern char opt_stats_print_opts[stats_print_tot_num_options+1];
+
+/* Implements je_malloc_stats_print. */
+void stats_print(void (*write_cb)(void *, const char *), void *cbopaque,
+    const char *opts);
+
+/*
+ * In those architectures that support 64-bit atomics, we use atomic updates for
+ * our 64-bit values.  Otherwise, we use a plain uint64_t and synchronize
+ * externally.
+ */
+#ifdef JEMALLOC_ATOMIC_U64
+typedef atomic_u64_t arena_stats_u64_t;
+#else
+/* Must hold the arena stats mutex while reading atomically. */
+typedef uint64_t arena_stats_u64_t;
+#endif
+
+typedef struct malloc_bin_stats_s {
+	/*
+	 * Total number of allocation/deallocation requests served directly by
+	 * the bin.  Note that tcache may allocate an object, then recycle it
+	 * many times, resulting many increments to nrequests, but only one
+	 * each to nmalloc and ndalloc.
+	 */
+	uint64_t	nmalloc;
+	uint64_t	ndalloc;
+
+	/*
+	 * Number of allocation requests that correspond to the size of this
+	 * bin.  This includes requests served by tcache, though tcache only
+	 * periodically merges into this counter.
+	 */
+	uint64_t	nrequests;
+
+	/*
+	 * Current number of regions of this size class, including regions
+	 * currently cached by tcache.
+	 */
+	size_t		curregs;
+
+	/* Number of tcache fills from this bin. */
+	uint64_t	nfills;
+
+	/* Number of tcache flushes to this bin. */
+	uint64_t	nflushes;
+
+	/* Total number of slabs created for this bin's size class. */
+	uint64_t	nslabs;
+
+	/*
+	 * Total number of slabs reused by extracting them from the slabs heap
+	 * for this bin's size class.
+	 */
+	uint64_t	reslabs;
+
+	/* Current number of slabs in this bin. */
+	size_t		curslabs;
+
+	mutex_prof_data_t mutex_data;
+} malloc_bin_stats_t;
+
+typedef struct malloc_large_stats_s {
+	/*
+	 * Total number of allocation/deallocation requests served directly by
+	 * the arena.
+	 */
+	arena_stats_u64_t	nmalloc;
+	arena_stats_u64_t	ndalloc;
+
+	/*
+	 * Number of allocation requests that correspond to this size class.
+	 * This includes requests served by tcache, though tcache only
+	 * periodically merges into this counter.
+	 */
+	arena_stats_u64_t	nrequests; /* Partially derived. */
+
+	/* Current number of allocations of this size class. */
+	size_t		curlextents; /* Derived. */
+} malloc_large_stats_t;
+
+typedef struct decay_stats_s {
+	/* Total number of purge sweeps. */
+	arena_stats_u64_t	npurge;
+	/* Total number of madvise calls made. */
+	arena_stats_u64_t	nmadvise;
+	/* Total number of pages purged. */
+	arena_stats_u64_t	purged;
+} decay_stats_t;
+
+/*
+ * Arena stats.  Note that fields marked "derived" are not directly maintained
+ * within the arena code; rather their values are derived during stats merge
+ * requests.
+ */
+typedef struct arena_stats_s {
+#ifndef JEMALLOC_ATOMIC_U64
+	malloc_mutex_t		mtx;
+#endif
+
+	/* Number of bytes currently mapped, excluding retained memory. */
+	atomic_zu_t		mapped; /* Partially derived. */
+
+	/*
+	 * Number of unused virtual memory bytes currently retained.  Retained
+	 * bytes are technically mapped (though always decommitted or purged),
+	 * but they are excluded from the mapped statistic (above).
+	 */
+	atomic_zu_t		retained; /* Derived. */
+
+	decay_stats_t		decay_dirty;
+	decay_stats_t		decay_muzzy;
+
+	atomic_zu_t		base; /* Derived. */
+	atomic_zu_t		internal;
+	atomic_zu_t		resident; /* Derived. */
+
+	atomic_zu_t		allocated_large; /* Derived. */
+	arena_stats_u64_t	nmalloc_large; /* Derived. */
+	arena_stats_u64_t	ndalloc_large; /* Derived. */
+	arena_stats_u64_t	nrequests_large; /* Derived. */
+
+	/* Number of bytes cached in tcache associated with this arena. */
+	atomic_zu_t		tcache_bytes; /* Derived. */
+
+	mutex_prof_data_t mutex_prof_data[mutex_prof_num_arena_mutexes];
+
+	/* One element for each large size class. */
+	malloc_large_stats_t	lstats[NSIZES - NBINS];
+
+	/* Arena uptime. */
+	nstime_t		uptime;
+} arena_stats_t;
+
+#endif /* JEMALLOC_INTERNAL_STATS_H */

include/jemalloc/internal/stats_tsd.h

@@ -0,0 +1,12 @@
+#ifndef JEMALLOC_INTERNAL_STATS_TSD_H
+#define JEMALLOC_INTERNAL_STATS_TSD_H
+
+typedef struct tcache_bin_stats_s {
+	/*
+	 * Number of allocation requests that corresponded to the size of this
+	 * bin.
+	 */
+	uint64_t	nrequests;
+} tcache_bin_stats_t;
+
+#endif /* JEMALLOC_INTERNAL_STATS_TSD_H */

include/jemalloc/internal/sz.h

@@ -0,0 +1,317 @@
+#ifndef JEMALLOC_INTERNAL_SIZE_H
+#define JEMALLOC_INTERNAL_SIZE_H
+
+#include "jemalloc/internal/bit_util.h"
+#include "jemalloc/internal/pages.h"
+#include "jemalloc/internal/size_classes.h"
+#include "jemalloc/internal/util.h"
+
+/*
+ * sz module: Size computations.
+ *
+ * Some abbreviations used here:
+ *   p: Page
+ *   ind: Index
+ *   s, sz: Size
+ *   u: Usable size
+ *   a: Aligned
+ *
+ * These are not always used completely consistently, but should be enough to
+ * interpret function names.  E.g. sz_psz2ind converts page size to page size
+ * index; sz_sa2u converts a (size, alignment) allocation request to the usable
+ * size that would result from such an allocation.
+ */
+
+/*
+ * sz_pind2sz_tab encodes the same information as could be computed by
+ * sz_pind2sz_compute().
+ */
+extern size_t const sz_pind2sz_tab[NPSIZES+1];
+/*
+ * sz_index2size_tab encodes the same information as could be computed (at
+ * unacceptable cost in some code paths) by sz_index2size_compute().
+ */
+extern size_t const sz_index2size_tab[NSIZES];
+/*
+ * sz_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,
+ * and all accesses are via sz_size2index().
+ */
+extern uint8_t const sz_size2index_tab[];
+
+static const size_t sz_large_pad =
+#ifdef JEMALLOC_CACHE_OBLIVIOUS
+    PAGE
+#else
+    0
+#endif
+    ;
+
+JEMALLOC_ALWAYS_INLINE pszind_t
+sz_psz2ind(size_t psz) {
+	if (unlikely(psz > LARGE_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 = ZD(-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;
+	}
+}
+
+static inline size_t
+sz_pind2sz_compute(pszind_t pind) {
+	if (unlikely(pind == NPSIZES)) {
+		return LARGE_MAXCLASS + PAGE;
+	}
+	{
+		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;
+	}
+}
+
+static inline size_t
+sz_pind2sz_lookup(pszind_t pind) {
+	size_t ret = (size_t)sz_pind2sz_tab[pind];
+	assert(ret == sz_pind2sz_compute(pind));
+	return ret;
+}
+
+static inline size_t
+sz_pind2sz(pszind_t pind) {
+	assert(pind < NPSIZES+1);
+	return sz_pind2sz_lookup(pind);
+}
+
+static inline size_t
+sz_psz2u(size_t psz) {
+	if (unlikely(psz > LARGE_MAXCLASS)) {
+		return LARGE_MAXCLASS + PAGE;
+	}
+	{
+		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;
+	}
+}
+
+static inline szind_t
+sz_size2index_compute(size_t size) {
+	if (unlikely(size > LARGE_MAXCLASS)) {
+		return NSIZES;
+	}
+#if (NTBINS != 0)
+	if (size <= (ZU(1) << LG_TINY_MAXCLASS)) {
+		szind_t lg_tmin = LG_TINY_MAXCLASS - NTBINS + 1;
+		szind_t lg_ceil = lg_floor(pow2_ceil_zu(size));
+		return (lg_ceil < lg_tmin ? 0 : lg_ceil - lg_tmin);
+	}
+#endif
+	{
+		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;
+
+		szind_t lg_delta = (x < LG_SIZE_CLASS_GROUP + LG_QUANTUM + 1)
+		    ? LG_QUANTUM : x - LG_SIZE_CLASS_GROUP - 1;
+
+		size_t delta_inverse_mask = ZD(-1) << lg_delta;
+		szind_t mod = ((((size-1) & delta_inverse_mask) >> lg_delta)) &
+		    ((ZU(1) << LG_SIZE_CLASS_GROUP) - 1);
+
+		szind_t index = NTBINS + grp + mod;
+		return index;
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE szind_t
+sz_size2index_lookup(size_t size) {
+	assert(size <= LOOKUP_MAXCLASS);
+	{
+		szind_t ret = (sz_size2index_tab[(size-1) >> LG_TINY_MIN]);
+		assert(ret == sz_size2index_compute(size));
+		return ret;
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE szind_t
+sz_size2index(size_t size) {
+	assert(size > 0);
+	if (likely(size <= LOOKUP_MAXCLASS)) {
+		return sz_size2index_lookup(size);
+	}
+	return sz_size2index_compute(size);
+}
+
+static inline size_t
+sz_index2size_compute(szind_t index) {
+#if (NTBINS > 0)
+	if (index < NTBINS) {
+		return (ZU(1) << (LG_TINY_MAXCLASS - NTBINS + 1 + index));
+	}
+#endif
+	{
+		size_t reduced_index = index - NTBINS;
+		size_t grp = reduced_index >> LG_SIZE_CLASS_GROUP;
+		size_t mod = reduced_index & ((ZU(1) << LG_SIZE_CLASS_GROUP) -
+		    1);
+
+		size_t grp_size_mask = ~((!!grp)-1);
+		size_t grp_size = ((ZU(1) << (LG_QUANTUM +
+		    (LG_SIZE_CLASS_GROUP-1))) << grp) & grp_size_mask;
+
+		size_t shift = (grp == 0) ? 1 : grp;
+		size_t lg_delta = shift + (LG_QUANTUM-1);
+		size_t mod_size = (mod+1) << lg_delta;
+
+		size_t usize = grp_size + mod_size;
+		return usize;
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+sz_index2size_lookup(szind_t index) {
+	size_t ret = (size_t)sz_index2size_tab[index];
+	assert(ret == sz_index2size_compute(index));
+	return ret;
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+sz_index2size(szind_t index) {
+	assert(index < NSIZES);
+	return sz_index2size_lookup(index);
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+sz_s2u_compute(size_t size) {
+	if (unlikely(size > LARGE_MAXCLASS)) {
+		return 0;
+	}
+#if (NTBINS > 0)
+	if (size <= (ZU(1) << LG_TINY_MAXCLASS)) {
+		size_t lg_tmin = LG_TINY_MAXCLASS - NTBINS + 1;
+		size_t lg_ceil = lg_floor(pow2_ceil_zu(size));
+		return (lg_ceil < lg_tmin ? (ZU(1) << lg_tmin) :
+		    (ZU(1) << lg_ceil));
+	}
+#endif
+	{
+		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;
+		size_t delta_mask = delta - 1;
+		size_t usize = (size + delta_mask) & ~delta_mask;
+		return usize;
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+sz_s2u_lookup(size_t size) {
+	size_t ret = sz_index2size_lookup(sz_size2index_lookup(size));
+
+	assert(ret == sz_s2u_compute(size));
+	return ret;
+}
+
+/*
+ * Compute usable size that would result from allocating an object with the
+ * specified size.
+ */
+JEMALLOC_ALWAYS_INLINE size_t
+sz_s2u(size_t size) {
+	assert(size > 0);
+	if (likely(size <= LOOKUP_MAXCLASS)) {
+		return sz_s2u_lookup(size);
+	}
+	return sz_s2u_compute(size);
+}
+
+/*
+ * Compute usable size that would result from allocating an object with the
+ * specified size and alignment.
+ */
+JEMALLOC_ALWAYS_INLINE size_t
+sz_sa2u(size_t size, size_t alignment) {
+	size_t usize;
+
+	assert(alignment != 0 && ((alignment - 1) & alignment) == 0);
+
+	/* Try for a small size class. */
+	if (size <= SMALL_MAXCLASS && alignment < PAGE) {
+		/*
+		 * Round size up to the nearest multiple of alignment.
+		 *
+		 * This done, we can take advantage of the fact that for each
+		 * small size class, every object is aligned at the smallest
+		 * power of two that is non-zero in the base two representation
+		 * of the size.  For example:
+		 *
+		 *   Size |   Base 2 | Minimum alignment
+		 *   -----+----------+------------------
+		 *     96 |  1100000 |  32
+		 *    144 | 10100000 |  32
+		 *    192 | 11000000 |  64
+		 */
+		usize = sz_s2u(ALIGNMENT_CEILING(size, alignment));
+		if (usize < LARGE_MINCLASS) {
+			return usize;
+		}
+	}
+
+	/* Large size class.  Beware of overflow. */
+
+	if (unlikely(alignment > LARGE_MAXCLASS)) {
+		return 0;
+	}
+
+	/* Make sure result is a large size class. */
+	if (size <= LARGE_MINCLASS) {
+		usize = LARGE_MINCLASS;
+	} else {
+		usize = sz_s2u(size);
+		if (usize < size) {
+			/* size_t overflow. */
+			return 0;
+		}
+	}
+
+	/*
+	 * Calculate the multi-page mapping that large_palloc() would need in
+	 * order to guarantee the alignment.
+	 */
+	if (usize + sz_large_pad + PAGE_CEILING(alignment) - PAGE < usize) {
+		/* size_t overflow. */
+		return 0;
+	}
+	return usize;
+}
+
+#endif /* JEMALLOC_INTERNAL_SIZE_H */

include/jemalloc/internal/tcache_externs.h

@@ -0,0 +1,55 @@
+#ifndef JEMALLOC_INTERNAL_TCACHE_EXTERNS_H
+#define JEMALLOC_INTERNAL_TCACHE_EXTERNS_H
+
+#include "jemalloc/internal/size_classes.h"
+
+extern bool	opt_tcache;
+extern ssize_t	opt_lg_tcache_max;
+
+extern tcache_bin_info_t	*tcache_bin_info;
+
+/*
+ * Number of tcache bins.  There are NBINS small-object bins, plus 0 or more
+ * large-object bins.
+ */
+extern unsigned	nhbins;
+
+/* Maximum cached size class. */
+extern size_t	tcache_maxclass;
+
+/*
+ * Explicit tcaches, managed via the tcache.{create,flush,destroy} mallctls and
+ * usable via the MALLOCX_TCACHE() flag.  The automatic per thread tcaches are
+ * completely disjoint from this data structure.  tcaches starts off as a sparse
+ * array, so it has no physical memory footprint until individual pages are
+ * touched.  This allows the entire array to be allocated the first time an
+ * explicit tcache is created without a disproportionate impact on memory usage.
+ */
+extern tcaches_t	*tcaches;
+
+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(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_reassociate(tsdn_t *tsdn, tcache_t *tcache,
+    arena_t *arena);
+tcache_t *tcache_create_explicit(tsd_t *tsd);
+void	tcache_cleanup(tsd_t *tsd);
+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(tsdn_t *tsdn);
+void tcache_arena_associate(tsdn_t *tsdn, tcache_t *tcache, arena_t *arena);
+void tcache_prefork(tsdn_t *tsdn);
+void tcache_postfork_parent(tsdn_t *tsdn);
+void tcache_postfork_child(tsdn_t *tsdn);
+void tcache_flush(void);
+bool tsd_tcache_data_init(tsd_t *tsd);
+bool tsd_tcache_enabled_data_init(tsd_t *tsd);
+
+#endif /* JEMALLOC_INTERNAL_TCACHE_EXTERNS_H */

include/jemalloc/internal/tcache_inlines.h

@@ -0,0 +1,250 @@
+#ifndef JEMALLOC_INTERNAL_TCACHE_INLINES_H
+#define JEMALLOC_INTERNAL_TCACHE_INLINES_H
+
+#include "jemalloc/internal/jemalloc_internal_types.h"
+#include "jemalloc/internal/size_classes.h"
+#include "jemalloc/internal/sz.h"
+#include "jemalloc/internal/ticker.h"
+#include "jemalloc/internal/util.h"
+
+static inline bool
+tcache_enabled_get(tsd_t *tsd) {
+	return tsd_tcache_enabled_get(tsd);
+}
+
+static inline void
+tcache_enabled_set(tsd_t *tsd, bool enabled) {
+	bool was_enabled = tsd_tcache_enabled_get(tsd);
+
+	if (!was_enabled && enabled) {
+		tsd_tcache_data_init(tsd);
+	} else if (was_enabled && !enabled) {
+		tcache_cleanup(tsd);
+	}
+	/* Commit the state last.  Above calls check current state. */
+	tsd_tcache_enabled_set(tsd, enabled);
+	tsd_slow_update(tsd);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+tcache_event(tsd_t *tsd, tcache_t *tcache) {
+	if (TCACHE_GC_INCR == 0) {
+		return;
+	}
+
+	if (unlikely(ticker_tick(&tcache->gc_ticker))) {
+		tcache_event_hard(tsd, tcache);
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+tcache_alloc_easy(tcache_bin_t *tbin, bool *tcache_success) {
+	void *ret;
+
+	if (unlikely(tbin->ncached == 0)) {
+		tbin->low_water = -1;
+		*tcache_success = false;
+		return NULL;
+	}
+	/*
+	 * tcache_success (instead of ret) should be checked upon the return of
+	 * this function.  We avoid checking (ret == NULL) because there is
+	 * never a null stored on the avail stack (which is unknown to the
+	 * compiler), and eagerly checking ret would cause pipeline stall
+	 * (waiting for the cacheline).
+	 */
+	*tcache_success = true;
+	ret = *(tbin->avail - tbin->ncached);
+	tbin->ncached--;
+
+	if (unlikely((low_water_t)tbin->ncached < tbin->low_water)) {
+		tbin->low_water = tbin->ncached;
+	}
+
+	return ret;
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+tcache_alloc_small(tsd_t *tsd, arena_t *arena, tcache_t *tcache, size_t size,
+    szind_t binind, bool zero, bool slow_path) {
+	void *ret;
+	tcache_bin_t *tbin;
+	bool tcache_success;
+	size_t usize JEMALLOC_CC_SILENCE_INIT(0);
+
+	assert(binind < NBINS);
+	tbin = tcache_small_bin_get(tcache, binind);
+	ret = tcache_alloc_easy(tbin, &tcache_success);
+	assert(tcache_success == (ret != NULL));
+	if (unlikely(!tcache_success)) {
+		bool tcache_hard_success;
+		arena = arena_choose(tsd, arena);
+		if (unlikely(arena == NULL)) {
+			return NULL;
+		}
+
+		ret = tcache_alloc_small_hard(tsd_tsdn(tsd), arena, tcache,
+		    tbin, binind, &tcache_hard_success);
+		if (tcache_hard_success == false) {
+			return NULL;
+		}
+	}
+
+	assert(ret);
+	/*
+	 * Only compute usize if required.  The checks in the following if
+	 * statement are all static.
+	 */
+	if (config_prof || (slow_path && config_fill) || unlikely(zero)) {
+		usize = sz_index2size(binind);
+		assert(tcache_salloc(tsd_tsdn(tsd), ret) == usize);
+	}
+
+	if (likely(!zero)) {
+		if (slow_path && config_fill) {
+			if (unlikely(opt_junk_alloc)) {
+				arena_alloc_junk_small(ret,
+				    &arena_bin_info[binind], false);
+			} else if (unlikely(opt_zero)) {
+				memset(ret, 0, usize);
+			}
+		}
+	} else {
+		if (slow_path && config_fill && unlikely(opt_junk_alloc)) {
+			arena_alloc_junk_small(ret, &arena_bin_info[binind],
+			    true);
+		}
+		memset(ret, 0, usize);
+	}
+
+	if (config_stats) {
+		tbin->tstats.nrequests++;
+	}
+	if (config_prof) {
+		tcache->prof_accumbytes += usize;
+	}
+	tcache_event(tsd, tcache);
+	return ret;
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+tcache_alloc_large(tsd_t *tsd, arena_t *arena, tcache_t *tcache, size_t size,
+    szind_t binind, bool zero, bool slow_path) {
+	void *ret;
+	tcache_bin_t *tbin;
+	bool tcache_success;
+
+	assert(binind >= NBINS &&binind < nhbins);
+	tbin = tcache_large_bin_get(tcache, binind);
+	ret = tcache_alloc_easy(tbin, &tcache_success);
+	assert(tcache_success == (ret != NULL));
+	if (unlikely(!tcache_success)) {
+		/*
+		 * Only allocate one large object at a time, because it's quite
+		 * expensive to create one and not use it.
+		 */
+		arena = arena_choose(tsd, arena);
+		if (unlikely(arena == NULL)) {
+			return NULL;
+		}
+
+		ret = large_malloc(tsd_tsdn(tsd), arena, sz_s2u(size), zero);
+		if (ret == NULL) {
+			return NULL;
+		}
+	} else {
+		size_t usize JEMALLOC_CC_SILENCE_INIT(0);
+
+		/* Only compute usize on demand */
+		if (config_prof || (slow_path && config_fill) ||
+		    unlikely(zero)) {
+			usize = sz_index2size(binind);
+			assert(usize <= tcache_maxclass);
+		}
+
+		if (likely(!zero)) {
+			if (slow_path && config_fill) {
+				if (unlikely(opt_junk_alloc)) {
+					memset(ret, JEMALLOC_ALLOC_JUNK,
+					    usize);
+				} else if (unlikely(opt_zero)) {
+					memset(ret, 0, usize);
+				}
+			}
+		} else {
+			memset(ret, 0, usize);
+		}
+
+		if (config_stats) {
+			tbin->tstats.nrequests++;
+		}
+		if (config_prof) {
+			tcache->prof_accumbytes += usize;
+		}
+	}
+
+	tcache_event(tsd, tcache);
+	return ret;
+}
+
+JEMALLOC_ALWAYS_INLINE void
+tcache_dalloc_small(tsd_t *tsd, tcache_t *tcache, void *ptr, szind_t binind,
+    bool slow_path) {
+	tcache_bin_t *tbin;
+	tcache_bin_info_t *tbin_info;
+
+	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]);
+	}
+
+	tbin = tcache_small_bin_get(tcache, binind);
+	tbin_info = &tcache_bin_info[binind];
+	if (unlikely(tbin->ncached == tbin_info->ncached_max)) {
+		tcache_bin_flush_small(tsd, tcache, tbin, binind,
+		    (tbin_info->ncached_max >> 1));
+	}
+	assert(tbin->ncached < tbin_info->ncached_max);
+	tbin->ncached++;
+	*(tbin->avail - tbin->ncached) = ptr;
+
+	tcache_event(tsd, tcache);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+tcache_dalloc_large(tsd_t *tsd, tcache_t *tcache, void *ptr, szind_t binind,
+    bool slow_path) {
+	tcache_bin_t *tbin;
+	tcache_bin_info_t *tbin_info;
+
+	assert(tcache_salloc(tsd_tsdn(tsd), ptr) > SMALL_MAXCLASS);
+	assert(tcache_salloc(tsd_tsdn(tsd), ptr) <= tcache_maxclass);
+
+	if (slow_path && config_fill && unlikely(opt_junk_free)) {
+		large_dalloc_junk(ptr, sz_index2size(binind));
+	}
+
+	tbin = tcache_large_bin_get(tcache, binind);
+	tbin_info = &tcache_bin_info[binind];
+	if (unlikely(tbin->ncached == tbin_info->ncached_max)) {
+		tcache_bin_flush_large(tsd, tbin, binind,
+		    (tbin_info->ncached_max >> 1), tcache);
+	}
+	assert(tbin->ncached < tbin_info->ncached_max);
+	tbin->ncached++;
+	*(tbin->avail - tbin->ncached) = ptr;
+
+	tcache_event(tsd, tcache);
+}
+
+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_explicit(tsd);
+	}
+	return elm->tcache;
+}
+
+#endif /* JEMALLOC_INTERNAL_TCACHE_INLINES_H */

include/jemalloc/internal/tcache_structs.h

@@ -0,0 +1,64 @@
+#ifndef JEMALLOC_INTERNAL_TCACHE_STRUCTS_H
+#define JEMALLOC_INTERNAL_TCACHE_STRUCTS_H
+
+#include "jemalloc/internal/ql.h"
+#include "jemalloc/internal/size_classes.h"
+#include "jemalloc/internal/stats_tsd.h"
+#include "jemalloc/internal/ticker.h"
+
+/*
+ * Read-only information associated with each element of tcache_t's tbins array
+ * is stored separately, mainly to reduce memory usage.
+ */
+struct tcache_bin_info_s {
+	unsigned	ncached_max;	/* Upper limit on ncached. */
+};
+
+struct tcache_bin_s {
+	low_water_t	low_water;	/* Min # cached since last GC. */
+	uint32_t	ncached;	/* # of cached objects. */
+	/*
+	 * ncached and stats are both modified frequently.  Let's keep them
+	 * close so that they have a higher chance of being on the same
+	 * cacheline, thus less write-backs.
+	 */
+	tcache_bin_stats_t tstats;
+	/*
+	 * To make use of adjacent cacheline prefetch, the items in the avail
+	 * stack goes to higher address for newer allocations.  avail points
+	 * just above the available space, which means that
+	 * avail[-ncached, ... -1] are available items and the lowest item will
+	 * be allocated first.
+	 */
+	void		**avail;	/* Stack of available objects. */
+};
+
+struct tcache_s {
+	/* Data accessed frequently first: prof, ticker and small bins. */
+	uint64_t	prof_accumbytes;/* Cleared after arena_prof_accum(). */
+	ticker_t	gc_ticker;	/* Drives incremental GC. */
+	/*
+	 * The pointer stacks associated with tbins follow as a contiguous
+	 * array.  During tcache initialization, the avail pointer in each
+	 * element of tbins is initialized to point to the proper offset within
+	 * this array.
+	 */
+	tcache_bin_t	tbins_small[NBINS];
+	/* Data accessed less often below. */
+	ql_elm(tcache_t) link;		/* Used for aggregating stats. */
+	arena_t		*arena;		/* Associated arena. */
+	szind_t		next_gc_bin;	/* Next bin to GC. */
+	/* For small bins, fill (ncached_max >> lg_fill_div). */
+	uint8_t		lg_fill_div[NBINS];
+	tcache_bin_t	tbins_large[NSIZES-NBINS];
+};
+
+/* Linkage for list of available (previously used) explicit tcache IDs. */
+struct tcaches_s {
+	union {
+		tcache_t	*tcache;
+		tcaches_t	*next;
+	};
+};
+
+#endif /* JEMALLOC_INTERNAL_TCACHE_STRUCTS_H */

include/jemalloc/internal/tcache_types.h

@@ -0,0 +1,61 @@
+#ifndef JEMALLOC_INTERNAL_TCACHE_TYPES_H
+#define JEMALLOC_INTERNAL_TCACHE_TYPES_H
+
+#include "jemalloc/internal/size_classes.h"
+
+typedef struct tcache_bin_info_s tcache_bin_info_t;
+typedef struct tcache_bin_s tcache_bin_t;
+typedef struct tcache_s tcache_t;
+typedef struct tcaches_s tcaches_t;
+
+/* ncached is cast to this type for comparison. */
+typedef int32_t low_water_t;
+
+/*
+ * tcache pointers close to NULL are used to encode state information that is
+ * used for two purposes: preventing thread caching on a per thread basis and
+ * cleaning up during thread shutdown.
+ */
+#define TCACHE_STATE_DISABLED		((tcache_t *)(uintptr_t)1)
+#define TCACHE_STATE_REINCARNATED	((tcache_t *)(uintptr_t)2)
+#define TCACHE_STATE_PURGATORY		((tcache_t *)(uintptr_t)3)
+#define TCACHE_STATE_MAX		TCACHE_STATE_PURGATORY
+
+/*
+ * Absolute minimum number of cache slots for each small bin.
+ */
+#define TCACHE_NSLOTS_SMALL_MIN		20
+
+/*
+ * Absolute maximum number of cache slots for each small bin in the thread
+ * cache.  This is an additional constraint beyond that imposed as: twice the
+ * number of regions per slab for this size class.
+ *
+ * This constant must be an even number.
+ */
+#define TCACHE_NSLOTS_SMALL_MAX		200
+
+/* Number of cache slots for large size classes. */
+#define TCACHE_NSLOTS_LARGE		20
+
+/* (1U << opt_lg_tcache_max) is used to compute tcache_maxclass. */
+#define LG_TCACHE_MAXCLASS_DEFAULT	15
+
+/*
+ * TCACHE_GC_SWEEP is the approximate number of allocation events between
+ * full GC sweeps.  Integer rounding may cause the actual number to be
+ * slightly higher, since GC is performed incrementally.
+ */
+#define TCACHE_GC_SWEEP			8192
+
+/* Number of tcache allocation/deallocation events between incremental GCs. */
+#define TCACHE_GC_INCR							\
+    ((TCACHE_GC_SWEEP / NBINS) + ((TCACHE_GC_SWEEP / NBINS == 0) ? 0 : 1))
+
+/* Used in TSD static initializer only. Real init in tcache_data_init(). */
+#define TCACHE_ZERO_INITIALIZER {0}
+
+/* Used in TSD static initializer only. Will be initialized to opt_tcache. */
+#define TCACHE_ENABLED_ZERO_INITIALIZER false
+
+#endif /* JEMALLOC_INTERNAL_TCACHE_TYPES_H */

include/jemalloc/internal/ticker.h

@@ -0,0 +1,50 @@
+#ifndef JEMALLOC_INTERNAL_TICKER_H
+#define JEMALLOC_INTERNAL_TICKER_H
+
+#include "jemalloc/internal/util.h"
+
+/**
+ * A ticker makes it easy to count-down events until some limit.  You
+ * ticker_init the ticker to trigger every nticks events.  You then notify it
+ * that an event has occurred with calls to ticker_tick (or that nticks events
+ * have occurred with a call to ticker_ticks), which will return true (and reset
+ * the counter) if the countdown hit zero.
+ */
+
+typedef struct {
+	int32_t tick;
+	int32_t nticks;
+} ticker_t;
+
+static inline void
+ticker_init(ticker_t *ticker, int32_t nticks) {
+	ticker->tick = nticks;
+	ticker->nticks = nticks;
+}
+
+static inline void
+ticker_copy(ticker_t *ticker, const ticker_t *other) {
+	*ticker = *other;
+}
+
+static inline int32_t
+ticker_read(const ticker_t *ticker) {
+	return ticker->tick;
+}
+
+static inline bool
+ticker_ticks(ticker_t *ticker, int32_t nticks) {
+	if (unlikely(ticker->tick < nticks)) {
+		ticker->tick = ticker->nticks;
+		return true;
+	}
+	ticker->tick -= nticks;
+	return(false);
+}
+
+static inline bool
+ticker_tick(ticker_t *ticker) {
+	return ticker_ticks(ticker, 1);
+}
+
+#endif /* JEMALLOC_INTERNAL_TICKER_H */

include/jemalloc/internal/tsd.h

@@ -0,0 +1,310 @@
+#ifndef JEMALLOC_INTERNAL_TSD_H
+#define JEMALLOC_INTERNAL_TSD_H
+
+#include "jemalloc/internal/arena_types.h"
+#include "jemalloc/internal/assert.h"
+#include "jemalloc/internal/jemalloc_internal_externs.h"
+#include "jemalloc/internal/prof_types.h"
+#include "jemalloc/internal/ql.h"
+#include "jemalloc/internal/rtree_tsd.h"
+#include "jemalloc/internal/tcache_types.h"
+#include "jemalloc/internal/tcache_structs.h"
+#include "jemalloc/internal/util.h"
+#include "jemalloc/internal/witness.h"
+
+/*
+ * Thread-Specific-Data layout
+ * --- data accessed on tcache fast path: state, rtree_ctx, stats, prof ---
+ * s: state
+ * e: tcache_enabled
+ * m: thread_allocated (config_stats)
+ * f: thread_deallocated (config_stats)
+ * p: prof_tdata (config_prof)
+ * c: rtree_ctx (rtree cache accessed on deallocation)
+ * t: tcache
+ * --- data not accessed on tcache fast path: arena-related fields ---
+ * d: arenas_tdata_bypass
+ * r: reentrancy_level
+ * x: narenas_tdata
+ * i: iarena
+ * a: arena
+ * o: arenas_tdata
+ * Loading TSD data is on the critical path of basically all malloc operations.
+ * In particular, tcache and rtree_ctx rely on hot CPU cache to be effective.
+ * Use a compact layout to reduce cache footprint.
+ * +--- 64-bit and 64B cacheline; 1B each letter; First byte on the left. ---+
+ * |----------------------------  1st cacheline  ----------------------------|
+ * | sedrxxxx mmmmmmmm ffffffff pppppppp [c * 32  ........ ........ .......] |
+ * |----------------------------  2nd cacheline  ----------------------------|
+ * | [c * 64  ........ ........ ........ ........ ........ ........ .......] |
+ * |----------------------------  3nd cacheline  ----------------------------|
+ * | [c * 32  ........ ........ .......] iiiiiiii aaaaaaaa oooooooo [t...... |
+ * +-------------------------------------------------------------------------+
+ * Note: the entire tcache is embedded into TSD and spans multiple cachelines.
+ *
+ * The last 3 members (i, a and o) before tcache isn't really needed on tcache
+ * fast path.  However we have a number of unused tcache bins and witnesses
+ * (never touched unless config_debug) at the end of tcache, so we place them
+ * there to avoid breaking the cachelines and possibly paging in an extra page.
+ */
+#ifdef JEMALLOC_JET
+typedef void (*test_callback_t)(int *);
+#  define MALLOC_TSD_TEST_DATA_INIT 0x72b65c10
+#  define MALLOC_TEST_TSD \
+    O(test_data,		int,			int)		\
+    O(test_callback,		test_callback_t,	int)
+#  define MALLOC_TEST_TSD_INITIALIZER , MALLOC_TSD_TEST_DATA_INIT, NULL
+#else
+#  define MALLOC_TEST_TSD
+#  define MALLOC_TEST_TSD_INITIALIZER
+#endif
+
+/*  O(name,			type,			nullable type */
+#define MALLOC_TSD							\
+    O(tcache_enabled,		bool,			bool)		\
+    O(arenas_tdata_bypass,	bool,			bool)		\
+    O(reentrancy_level,		int8_t,			int8_t)		\
+    O(narenas_tdata,		uint32_t,		uint32_t)	\
+    O(thread_allocated,		uint64_t,		uint64_t)	\
+    O(thread_deallocated,	uint64_t,		uint64_t)	\
+    O(prof_tdata,		prof_tdata_t *,		prof_tdata_t *)	\
+    O(rtree_ctx,		rtree_ctx_t,		rtree_ctx_t)	\
+    O(iarena,			arena_t *,		arena_t *)	\
+    O(arena,			arena_t *,		arena_t *)	\
+    O(arenas_tdata,		arena_tdata_t *,	arena_tdata_t *)\
+    O(tcache,			tcache_t,		tcache_t)	\
+    O(witness_tsd,              witness_tsd_t,		witness_tsdn_t)	\
+    MALLOC_TEST_TSD
+
+#define TSD_INITIALIZER {						\
+    tsd_state_uninitialized,						\
+    TCACHE_ENABLED_ZERO_INITIALIZER,					\
+    false,								\
+    0,									\
+    0,									\
+    0,									\
+    0,									\
+    NULL,								\
+    RTREE_CTX_ZERO_INITIALIZER,						\
+    NULL,								\
+    NULL,								\
+    NULL,								\
+    TCACHE_ZERO_INITIALIZER,						\
+    WITNESS_TSD_INITIALIZER						\
+    MALLOC_TEST_TSD_INITIALIZER						\
+}
+
+enum {
+	tsd_state_nominal = 0, /* Common case --> jnz. */
+	tsd_state_nominal_slow = 1, /* Initialized but on slow path. */
+	/* the above 2 nominal states should be lower values. */
+	tsd_state_nominal_max = 1, /* used for comparison only. */
+	tsd_state_purgatory = 2,
+	tsd_state_reincarnated = 3,
+	tsd_state_uninitialized = 4
+};
+
+/* Manually limit tsd_state_t to a single byte. */
+typedef uint8_t tsd_state_t;
+
+/* The actual tsd. */
+struct tsd_s {
+	/*
+	 * The contents should be treated as totally opaque outside the tsd
+	 * module.  Access any thread-local state through the getters and
+	 * setters below.
+	 */
+	tsd_state_t	state;
+#define O(n, t, nt)							\
+	t use_a_getter_or_setter_instead_##n;
+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;
+};
+#define TSDN_NULL ((tsdn_t *)0)
+JEMALLOC_ALWAYS_INLINE tsdn_t *
+tsd_tsdn(tsd_t *tsd) {
+	return (tsdn_t *)tsd;
+}
+
+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;
+}
+
+void *malloc_tsd_malloc(size_t size);
+void malloc_tsd_dalloc(void *wrapper);
+void malloc_tsd_cleanup_register(bool (*f)(void));
+tsd_t *malloc_tsd_boot0(void);
+void malloc_tsd_boot1(void);
+void tsd_cleanup(void *arg);
+tsd_t *tsd_fetch_slow(tsd_t *tsd, bool internal);
+void tsd_slow_update(tsd_t *tsd);
+
+/*
+ * We put the platform-specific data declarations and inlines into their own
+ * header files to avoid cluttering this file.  They define tsd_boot0,
+ * tsd_boot1, tsd_boot, tsd_booted_get, tsd_get_allocates, tsd_get, and tsd_set.
+ */
+#ifdef JEMALLOC_MALLOC_THREAD_CLEANUP
+#include "jemalloc/internal/tsd_malloc_thread_cleanup.h"
+#elif (defined(JEMALLOC_TLS))
+#include "jemalloc/internal/tsd_tls.h"
+#elif (defined(_WIN32))
+#include "jemalloc/internal/tsd_win.h"
+#else
+#include "jemalloc/internal/tsd_generic.h"
+#endif
+
+/*
+ * tsd_foop_get_unsafe(tsd) returns a pointer to the thread-local instance of
+ * foo.  This omits some safety checks, and so can be used during tsd
+ * initialization and cleanup.
+ */
+#define O(n, t, nt)							\
+JEMALLOC_ALWAYS_INLINE t *						\
+tsd_##n##p_get_unsafe(tsd_t *tsd) {					\
+	return &tsd->use_a_getter_or_setter_instead_##n;		\
+}
+MALLOC_TSD
+#undef O
+
+/* tsd_foop_get(tsd) returns a pointer to the thread-local instance of foo. */
+#define O(n, t, nt)							\
+JEMALLOC_ALWAYS_INLINE t *						\
+tsd_##n##p_get(tsd_t *tsd) {						\
+	assert(tsd->state == tsd_state_nominal ||			\
+	    tsd->state == tsd_state_nominal_slow ||			\
+	    tsd->state == tsd_state_reincarnated);			\
+	return tsd_##n##p_get_unsafe(tsd);				\
+}
+MALLOC_TSD
+#undef O
+
+/*
+ * tsdn_foop_get(tsdn) returns either the thread-local instance of foo (if tsdn
+ * isn't NULL), or NULL (if tsdn is NULL), cast to the nullable pointer type.
+ */
+#define O(n, t, nt)							\
+JEMALLOC_ALWAYS_INLINE nt *						\
+tsdn_##n##p_get(tsdn_t *tsdn) {						\
+	if (tsdn_null(tsdn)) {						\
+		return NULL;						\
+	}								\
+	tsd_t *tsd = tsdn_tsd(tsdn);					\
+	return (nt *)tsd_##n##p_get(tsd);				\
+}
+MALLOC_TSD
+#undef O
+
+/* tsd_foo_get(tsd) returns the value of the thread-local instance of foo. */
+#define O(n, t, nt)							\
+JEMALLOC_ALWAYS_INLINE t						\
+tsd_##n##_get(tsd_t *tsd) {						\
+	return *tsd_##n##p_get(tsd);					\
+}
+MALLOC_TSD
+#undef O
+
+/* tsd_foo_set(tsd, val) updates the thread-local instance of foo to be val. */
+#define O(n, t, nt)							\
+JEMALLOC_ALWAYS_INLINE void						\
+tsd_##n##_set(tsd_t *tsd, t val) {					\
+	assert(tsd->state != tsd_state_reincarnated);			\
+	*tsd_##n##p_get(tsd) = val;					\
+}
+MALLOC_TSD
+#undef O
+
+JEMALLOC_ALWAYS_INLINE void
+tsd_assert_fast(tsd_t *tsd) {
+	assert(!malloc_slow && tsd_tcache_enabled_get(tsd) &&
+	    tsd_reentrancy_level_get(tsd) == 0);
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+tsd_fast(tsd_t *tsd) {
+	bool fast = (tsd->state == tsd_state_nominal);
+	if (fast) {
+		tsd_assert_fast(tsd);
+	}
+
+	return fast;
+}
+
+JEMALLOC_ALWAYS_INLINE tsd_t *
+tsd_fetch_impl(bool init, bool internal) {
+	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)) {
+		return tsd_fetch_slow(tsd, internal);
+	}
+	assert(tsd_fast(tsd));
+	tsd_assert_fast(tsd);
+
+	return tsd;
+}
+
+JEMALLOC_ALWAYS_INLINE tsd_t *
+tsd_internal_fetch(void) {
+	return tsd_fetch_impl(true, true);
+}
+
+JEMALLOC_ALWAYS_INLINE tsd_t *
+tsd_fetch(void) {
+	return tsd_fetch_impl(true, false);
+}
+
+static inline bool
+tsd_nominal(tsd_t *tsd) {
+	return (tsd->state <= tsd_state_nominal_max);
+}
+
+JEMALLOC_ALWAYS_INLINE tsdn_t *
+tsdn_fetch(void) {
+	if (!tsd_booted_get()) {
+		return NULL;
+	}
+
+	return tsd_tsdn(tsd_fetch_impl(false, false));
+}
+
+JEMALLOC_ALWAYS_INLINE rtree_ctx_t *
+tsd_rtree_ctx(tsd_t *tsd) {
+	return tsd_rtree_ctxp_get(tsd);
+}
+
+JEMALLOC_ALWAYS_INLINE rtree_ctx_t *
+tsdn_rtree_ctx(tsdn_t *tsdn, rtree_ctx_t *fallback) {
+	/*
+	 * If tsd cannot be accessed, initialize the fallback rtree_ctx and
+	 * return a pointer to it.
+	 */
+	if (unlikely(tsdn_null(tsdn))) {
+		rtree_ctx_data_init(fallback);
+		return fallback;
+	}
+	return tsd_rtree_ctx(tsdn_tsd(tsdn));
+}
+
+#endif /* JEMALLOC_INTERNAL_TSD_H */