2.4.0 Release notes.

Ref: issue #103

.gitignore

@@ -1,4 +1,3 @@
-00-RELEASENOTES
 *.o
 *.rdb
 *.log
@@ -6,8 +5,17 @@ redis-cli
 redis-server
 redis-benchmark
 redis-check-dump
+redis-check-aof
 doc-tools
 mkrelease.sh
 release
 myredis.conf
 misc/*
+src/release.h
+appendonly.aof
+SHORT_TERM_TODO
+redis.conf.*
+release.h
+src/transfer.sh
+src/configs
+src/redis-server.dSYM

00-RELEASENOTES

@@ -0,0 +1,102 @@
+Redis 2.4 release notes
+
+Migrating from 2.2 to 2.4
+=========================
+
+Redis 2.2 is mostly a strict subset of 2.4.
+The only thing you should be aware is that you can't use .rdb and AOF files
+generated with 2.4 into a 2.2 instance.
+
+2.4 slaves can be attached to 2.2 masters, but not the contrary, and only for
+the time needed to perform the version upgrade.
+
+From the point of view of the API Redis 2.4 only adds new commands
+(other commands now accepts a variable number of arguments) so you don't need
+to modify your program in order to use Redis 2.4.
+
+---------
+CHANGELOG
+---------
+
+What's new in Redis 2.4.0
+=========================
+
+* [BUGFIX] redis-cli segfault with single numerical argument fixed.
+* [BUGFIX] OpenBSD compilation problem fixed.
+* [BUGFIX] More robust Redis test, with better random port selection.
+* [BUGFIX] Fix for bug #128 about the RENAME command.
+* [BUGFIX] Fixed Issue #131. stime/utime reported in INFO was inverted.
+* [BUGFIX] Unlink Unix socket file on shutdown.
+* [BUGFIX] AUTH now returns error if no password is set on the server.
+* [BUGFIX] Exit with Fatal error at startup on RDB loading errors.
+* redis-check-dump: RDB version 2 now supported.
+* More informative error when DEBUG RELOAD fails.
+* Added a config directive for a Unix socket mask.
+* CONFIG SET/GET for loglevel.
+
+What's new in Redis 2.3.11 (2.4 Release Candidate 8)
+====================================================
+
+* [BUGFIX] Fixed a rare but possible AOF race condition that could result into
+           duplicated commands inside the AOF.
+* [BUGFIX] Fixed issue 620, don't segfault on corrupted (by hand) AOF.
+* [BUGFIX] Fixed compilation on Mac/PPC.
+* [BUGFIX] Don't replicate SAVE.
+* LRANGE optimization may drastically improve performances when querying the
+  final part of a long list.
+* redis-cli now implements a --latency mode to monitory Redis delay.
+* Hash type settings removed from INFO (same info is available via config GET)
+* Include port number on error when can't bind.
+* AOF fsync is now performed in background when fsync policy is 'everysec'.
+* AOF performances improved moving in background a possibly slow close(2) call.
+* AOF protocol synthesis speedup.
+
+What's new in Redis 2.3.10 (2.4 Release Candidate 7)
+====================================================
+
+* [BUGFIX] Fixed issue 593 (BRPOPLPUSH related crash).
+* [BUGFIX] Fixed an issue with the networking layer that may prevent Redis from sending the whole reply back to client under extreme conditions.
+
+What's new in Redis 2.3.9 (2.4 Release Candidate 6)
+===================================================
+
+* [BUGFIX] Fixed a bug with the automatic AOF rewrite causing continuous
+  rewrites for AOF files bigger than 4 GB.
+* New maxmemory tests.
+
+What's new in Redis 2.3.8 (2.4 Release Candidate 5)
+===================================================
+
+Compared to Redis 2.3.7 (RC4) there are the following changes:
+
+* [BUGFIX] HDEL: Abort deleting fields when hash is removed.
+* [BUGFIX] Fix adding bulk reply when getcwd fails.
+
+WHAT'S NEW IN REDIS 2.4 compared to the 2.2 version?
+====================================================
+
+* Specially encoded sorted sets, now small sorted sets will use little memory.
+* Native persistence of specially encoded data types (ziplists, zipmaps,
+  intsets). Many data sets will be saved and loaded an order of magnitude
+  faster.
+* Variadic versions of commands: SADD, HDEL, SREM, ZREM, ZADD, L/RPUSH.
+* Jemalloc support (enabled by default for Linux build) in order to
+  avoid fragmentation issues.
+* Reduced memory usage while saving.
+* More info fields (peak memory, fork time, ...)
+* OBJECT command for objects introspection.
+* CLIENT command for clients introspection.
+* Non blocking slave -> master connection.
+* Better redis-cli connection handling. New redis-cli features.
+* Better redis-benchmark, now able to benchmark user provided commands.
+* Colorized Make.
+* VM deprecated. Still supported but with a big warning... don't use it.
+* Many speed optimizations and bug fixes.
+
+Credits: Where not specified the implementation and design are done by
+Salvatore Sanfilippo and Pieter Noordhuis. Thanks to VMware for making all
+this possible. Also many thanks to all the other contributors and the amazing
+community we have.
+
+Cheers,
+Salvatore

BETATESTING.txt

@@ -1,9 +0,0 @@
-This is a stable release, for beta testing make sure to download the latest source code from Git:
-
-    git clone git://github.com/antirez/redis.git
-
-It's also possibe to download the latest source code as a tarball:
-
-    http://github.com/antirez/redis/tree/master
-
-(use the download button)

CONTRIBUTING

@@ -0,0 +1,13 @@
+1. Enter irc.freenode.org #redis and start talking with 'antirez' and/or 'pietern' to check if there is interest for such a feature and to understand the probability of it being merged. We'll try hard to keep Redis simple... so you'll likely encounter an high resistence.
+
+2. Drop a message to the Redis Google Group with a proposal of semantics/API.
+
+3. If steps 1 and 2 are ok, use the following procedure to submit a patch:
+
+    a. Fork Redis on github
+    b. Create a topic branch (git checkout -b my_branch)
+    c. Push to your branch (git push origin my_branch)
+    d. Create an issue in the Redis google code site with a link to your patch
+    e. Done :)
+
+Thanks!

INSTALL

@@ -0,0 +1 @@
+Please check the README file.

Makefile

@@ -1,103 +1,23 @@
-# Redis Makefile
-# Copyright (C) 2009 Salvatore Sanfilippo <antirez at gmail dot com>
-# This file is released under the BSD license, see the COPYING file
+# Top level makefile, the real shit is at src/Makefile
 
-uname_S := $(shell sh -c 'uname -s 2>/dev/null || echo not')
-OPTIMIZATION?=-O2
-ifeq ($(uname_S),SunOS)
-  CFLAGS?= -std=c99 -pedantic $(OPTIMIZATION) -Wall -W -D__EXTENSIONS__ -D_XPG6
-  CCLINK?= -ldl -lnsl -lsocket -lm -lpthread
-else
-  CFLAGS?= -std=c99 -pedantic $(OPTIMIZATION) -Wall -W $(ARCH) $(PROF)
-  CCLINK?= -lm -pthread
-endif
-CCOPT= $(CFLAGS) $(CCLINK) $(ARCH) $(PROF)
-DEBUG?= -g -rdynamic -ggdb 
+TARGETS=32bit noopt test
 
-OBJ = adlist.o ae.o anet.o dict.o redis.o sds.o zmalloc.o lzf_c.o lzf_d.o pqsort.o zipmap.o
-BENCHOBJ = ae.o anet.o redis-benchmark.o sds.o adlist.o zmalloc.o
-CLIOBJ = anet.o sds.o adlist.o redis-cli.o zmalloc.o
-CHECKDUMPOBJ = redis-check-dump.o lzf_c.o lzf_d.o
+all:
+	cd src && $(MAKE) $@
 
-PRGNAME = redis-server
-BENCHPRGNAME = redis-benchmark
-CLIPRGNAME = redis-cli
-CHECKDUMPPRGNAME = redis-check-dump
-
-all: redis-server redis-benchmark redis-cli redis-check-dump
-
-# Deps (use make dep to generate this)
-adlist.o: adlist.c adlist.h zmalloc.h
-ae.o: ae.c ae.h zmalloc.h config.h ae_kqueue.c
-ae_epoll.o: ae_epoll.c
-ae_kqueue.o: ae_kqueue.c
-ae_select.o: ae_select.c
-anet.o: anet.c fmacros.h anet.h
-dict.o: dict.c fmacros.h dict.h zmalloc.h
-lzf_c.o: lzf_c.c lzfP.h
-lzf_d.o: lzf_d.c lzfP.h
-pqsort.o: pqsort.c
-redis-benchmark.o: redis-benchmark.c fmacros.h ae.h anet.h sds.h adlist.h \
-  zmalloc.h
-redis-cli.o: redis-cli.c fmacros.h anet.h sds.h adlist.h zmalloc.h
-redis.o: redis.c fmacros.h config.h redis.h ae.h sds.h anet.h dict.h \
-  adlist.h zmalloc.h lzf.h pqsort.h zipmap.h staticsymbols.h
-sds.o: sds.c sds.h zmalloc.h
-zipmap.o: zipmap.c zmalloc.h
-zmalloc.o: zmalloc.c config.h
-
-redis-server: $(OBJ)
-	$(CC) -o $(PRGNAME) $(CCOPT) $(DEBUG) $(OBJ)
-	@echo ""
-	@echo "Hint: To run the test-redis.tcl script is a good idea."
-	@echo "Launch the redis server with ./redis-server, then in another"
-	@echo "terminal window enter this directory and run 'make test'."
-	@echo ""
-
-redis-benchmark: $(BENCHOBJ)
-	$(CC) -o $(BENCHPRGNAME) $(CCOPT) $(DEBUG) $(BENCHOBJ)
-
-redis-cli: $(CLIOBJ)
-	$(CC) -o $(CLIPRGNAME) $(CCOPT) $(DEBUG) $(CLIOBJ)
-
-redis-check-dump: $(CHECKDUMPOBJ)
-	$(CC) -o $(CHECKDUMPPRGNAME) $(CCOPT) $(DEBUG) $(CHECKDUMPOBJ)
-
-.c.o:
-	$(CC) -c $(CFLAGS) $(DEBUG) $(COMPILE_TIME) $<
+install: dummy
+	cd src && $(MAKE) $@
 
 clean:
-	rm -rf $(PRGNAME) $(BENCHPRGNAME) $(CLIPRGNAME) $(CHECKDUMPPRGNAME) *.o *.gcda *.gcno *.gcov
+	cd src && $(MAKE) $@
+	cd deps/hiredis && $(MAKE) $@
+	cd deps/linenoise && $(MAKE) $@
+	-(cd deps/jemalloc && $(MAKE) distclean)
 
-dep:
-	$(CC) -MM *.c
+$(TARGETS):
+	cd src && $(MAKE) $@
 
-staticsymbols:
-	tclsh utils/build-static-symbols.tcl > staticsymbols.h
+src/help.h:
+	@./utils/generate-command-help.rb > $@
 
-test:
-	tclsh test-redis.tcl
-
-bench:
-	./redis-benchmark
-
-log:
-	git log '--pretty=format:%ad %s' --date=short > Changelog
-
-32bit:
-	@echo ""
-	@echo "WARNING: if it fails under Linux you probably need to install libc6-dev-i386"
-	@echo ""
-	make ARCH="-m32"
-
-gprof:
-	make PROF="-pg"
-
-gcov:
-	make PROF="-fprofile-arcs -ftest-coverage"
-
-noopt:
-	make OPTIMIZATION=""
-
-32bitgprof:
-	make PROF="-pg" ARCH="-arch i386"
+dummy:

README

@@ -1 +1,94 @@
-Check the 'doc' directory. doc/README.html is a good starting point :)
+Where to find complete Redis documentation?
+-------------------------------------------
+
+This README is just a fast "quick start" document. You can find more detailed
+documentation at http://redis.io
+
+Building Redis
+--------------
+
+It is as simple as:
+
+    % make
+
+Redis is just a single binary, but if you want to install it you can use
+the "make install" target that will copy the binary in /usr/local/bin
+for default. You can also use "make PREFIX=/some/other/directory install"
+if you wish to use a different destination.
+
+You can run a 32 bit Redis binary using:
+
+    % make 32bit
+
+After building Redis is a good idea to test it, using:
+
+    % make test
+
+NOTE: if after building Redis with a 32 bit target you need to rebuild it
+      with a 64 bit target you need to perform a "make clean" in the root
+      directory of the Redis distribution.
+
+Allocator
+---------
+
+By default Redis compiles and links against jemalloc under Linux, since
+glibc malloc() has memory fragmentation problems.
+
+To force a libc malloc() build use:
+
+    make FORCE_LIBC_MALLOC=yes
+
+In all the other non Linux systems the libc malloc() is used by default.
+
+On Mac OS X you can force a jemalloc based build using the following:
+
+    make USE_JEMALLOC=yes
+
+Verbose build
+-------------
+
+Redis will build with a user friendly colorized output by default.
+If you want to see a more verbose output use the following:
+
+    make V=1
+
+Running Redis
+-------------
+
+To run Redis with the default configuration just type:
+
+    % cd src
+    % ./redis-server
+    
+If you want to provide your redis.conf, you have to run it using an additional
+parameter (the path of the configuration file):
+
+    % cd src
+    % ./redis-server /path/to/redis.conf
+
+Playing with Redis
+------------------
+
+You can use redis-cli to play with Redis. Start a redis-server instance,
+then in another terminal try the following:
+
+    % cd src
+    % ./redis-cli
+    redis> ping
+    PONG
+    redis> set foo bar
+    OK
+    redis> get foo
+    "bar"
+    redis> incr mycounter
+    (integer) 1
+    redis> incr mycounter
+    (integer) 2
+    redis> 
+
+You can find the list of all the available commands here:
+
+    http://redis.io/commands
+
+Enjoy!
+

TODO

@@ -1,90 +1,4 @@
-Redis TODO and Roadmap
+This is a stable release! No TODO file here.
+Please check the TODO file in the master branch on github.
 
-VERSION 2.0 TODO
-================
-
-* BRPOPLPUSH
-* List ops like L/RPUSH L/RPOP should return the new list length.
-* Save dataset / fsync() on SIGTERM
-* MULTI/EXEC should support the "EXEC FSYNC" form?
-* BLPOP & C. tests (write a non blocking Tcl client as first step)
-* ZCOUNT sortedset min max
-* ZRANK: http://docs.google.com/viewer?a=v&q=cache:tCQaP3ZeN4YJ:courses.csail.mit.edu/6.046/spring04/handouts/ps5-sol.pdf+skip+list+rank+operation+augmented&hl=en&pid=bl&srcid=ADGEEShXuNjTcZyXw_1cq9OaWpSXy3PprjXqVzmM-LE0ETFznLyrDXJKQ_mBPNT10R8ErkoiXD9JbMw_FaoHmOA4yoGVrA7tZWiy393JwfCwuewuP93sjbkzZ_gnEp83jYhPYjThaIzw&sig=AHIEtbRF0GkYCdYRFtTJBE69senXZwFY0w
-* Once ZRANK is implemented, change the implementation of ZCOUNT to use the augmented skiplist in order to be much faster.
-* Write doc for ZCOUNT, and for open / closed intervals of sorted sets range operations.
-
-Virtual Memory sub-TODO:
-* Check if the page selection algorithm is working well
-* Divide swappability of objects by refcount
-* Use multiple open FDs against the VM file, one for thread.
-* it should be possible to give the vm-max-memory option in megabyte, gigabyte, ..., just using 2GB, 100MB, and so forth.
-* Try to understand what can be moved into I/O threads that currently is instead handled by the main thread. For instance swapping file table scannig to find contiguous page could be a potential candidate (but I'm not convinced it's a good idea, better to improve the algorithm, for instance double the fast forward at every step?).
-* Possibly decrRefCount() against swapped objects can be moved into I/O threads, as it's a slow operation against million elements list, and in general consumes CPU time that can be consumed by other threads (and cores).
-* EXISTS should avoid loading the object if possible without too make the code too specialized.
-* vm-min-age <seconds> option
-* Make sure objects loaded from the VM are specially encoded when possible.
-* Check what happens performance-wise if instead to create threads again and again the same threads are reused forever. Note: this requires a way to disable this clients in the child, but waiting for empty new jobs queue can be enough.
-* Sets of integers are slow to load, for a number of reasons. Fix it. (use slow_sets.rdb file for debugging). (p.s. this was now partially fixed).
-* On EXEC try to block the client until relevant keys are loaded.
-
-* Hashes (GET/SET/DEL/INCRBY/EXISTS/FIELDS/LEN/MSET/MGET). Special encoding for hashes with less than N elements.
-* Write documentation for APPEND
-* Implement LEN, SUBSTR, PEEK, POKE, SETBIT, GETBIT
-
-VERSION 2.2 TODO (Fault tolerant sharding)
-===========================================
-
-* Redis-cluster, a fast intermediate layer (proxy) that implements consistent hashing and fault tollerant nodes handling.
-
-Interesting readings about this:
-
-    - http://ayende.com/Blog/archive/2009/04/06/designing-rhino-dht-a-fault-tolerant-dynamically-distributed-hash.aspx
-
-VERSION 2.4 TODO (Optimizations and latency)
-============================================
-
-* Lower the CPU usage.
-* Lower the RAM usage everywhere possible.
-* Use epool and alike to rewrite ae.c for Linux and other platforms suppporting fater-than-select() mutiplexing APIs.
-* Implement an UDP interface for low-latency GET/SET operations.
-
-OTHER IMPORTANT THINGS THAT WILL BE ADDED BUT I'M NOT SURE WHEN
-===============================================================
-
-BIG ONES:
-
-* Specially encoded memory-saving integer sets.
-* A command to export a JSON dump (there should be mostly working patch needing major reworking).
-* Specially encoded sets of integers (this includes a big refactoring providing an higher level layer for Sets manipulation)
-
-SMALL ONES:
-
-* If sizeof(double) == sizeof(void*) we could store the double value of sorted sets directly in place of the pointer instead of allocating it in the heap.
-* Delete on writes against expire policy should only happen after argument parsing for commands doing their own arg parsing stuff.
-* Give errors when incrementing a key that does not look like an integer, when providing as a sorted set score something can't be parsed as a double, and so forth.
-* MSADD (n keys) (n values). See this thread in the Redis google group: http://groups.google.com/group/redis-db/browse_thread/thread/e766d84eb375cd41
-* Don't save empty lists / sets / zsets on disk with snapshotting.
-* Remove keys when a list / set / zset reaches length of 0.
-* An option to exec a command slave-side if the master connection is lost: even cooler: if the script returns "0" the slave elects itself as master, otherwise continue trying to reconnect.
-
-THE "MAYBE" TODO LIST: things that may or may not get implemented
-=================================================================
-
-Most of this can be seen just as proposals, the fact they are in this list
-it's not a guarantee they'll ever get implemented ;)
-
-* Move dict.c from hash table to skip list, in order to avoid the blocking resize operation needed for the hash table.
-* FORK command (fork()s executing the commands received by the current
-  client in the new process). Hint: large SORTs can use more cores,
-  copy-on-write will avoid memory problems.
-* DUP command? DUP srckey dstkey, creates an exact clone of srckey value in dstkey.
-* SORT: Don't copy the list into a vector when BY argument is constant.
-* Write the hash table size of every db in the dump, so that Redis can resize the hash table just one time when loading a big DB.
-* LOCK / TRYLOCK / UNLOCK as described many times in the google group
-* Replication automated tests
-* Byte Array type (BA prefixed commands): BASETBIT BAGETBIT BASETU8 U16 U32 U64 S8 S16 S32 S64, ability to atomically INCRBY all the base types. BARANGE to get a range of bytes as a bulk value, BASETRANGE to set a range of bytes.
-* zmalloc() should avoid to add a private header for archs where there is some other kind of libc-specific way to get the size of a malloced block. Already done for Mac OS X.
-* Read-only mode.
-* Pattern-matching replication.
-* Add an option to relax the delete-expiring-keys-on-write semantic *denying* replication and AOF when this is on? Can be handy sometimes, when using Redis for non persistent state, but can create problems. For instance should rename and move also "move" the timeouts? How does this affect other commands?
-* Multiple BY in SORT.
+    https://github.com/antirez/redis/raw/master/TODO

client-libraries/README

@@ -1,11 +0,0 @@
-An updated list of client libraries for Redis can be found here:
-
-    http://code.google.com/p/redis
-
-All the links are in the front page.
-
-Most mainstream (and less mainstream) languages are supported.
-
-If you plan to write a new client library for a missing language, please
-get in touch with the other developers writing a message to the Redis google
-group that is located here: http://groups.google.com/group/redis-db

deps/hiredis/.gitignore

@@ -0,0 +1,6 @@
+/hiredis-test
+/hiredis-example*
+/*.o
+/*.so
+/*.dylib
+/*.a

deps/hiredis/COPYING

@@ -0,0 +1,10 @@
+Copyright (c) 2006-2009, Salvatore Sanfilippo
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
+
+    * Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
+    * Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
+    * Neither the name of Redis nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

deps/hiredis/Makefile

@@ -0,0 +1,115 @@
+# Hiredis Makefile
+# Copyright (C) 2010 Salvatore Sanfilippo <antirez at gmail dot com>
+# This file is released under the BSD license, see the COPYING file
+
+OBJ = net.o hiredis.o sds.o async.o
+BINS = hiredis-example hiredis-test
+
+uname_S := $(shell sh -c 'uname -s 2>/dev/null || echo not')
+OPTIMIZATION?=-O3
+ifeq ($(uname_S),SunOS)
+  CFLAGS?=-std=c99 -pedantic $(OPTIMIZATION) -fPIC -Wall -W -D__EXTENSIONS__ -D_XPG6 $(ARCH) $(PROF)
+  CCLINK?=-ldl -lnsl -lsocket -lm -lpthread
+  LDFLAGS?=-L. -Wl,-R,.
+  DYLIBNAME?=libhiredis.so
+  DYLIB_MAKE_CMD?=$(CC) -G -o ${DYLIBNAME} ${OBJ}
+  STLIBNAME?=libhiredis.a
+  STLIB_MAKE_CMD?=ar rcs ${STLIBNAME} ${OBJ}
+else
+ifeq ($(uname_S),Darwin)
+  CFLAGS?=-std=c99 -pedantic $(OPTIMIZATION) -fPIC -Wall -W -Wstrict-prototypes -Wwrite-strings $(ARCH) $(PROF)
+  CCLINK?=-lm -pthread
+  LDFLAGS?=-L. -Wl,-rpath,.
+  OBJARCH?=-arch i386 -arch x86_64
+  DYLIBNAME?=libhiredis.dylib
+  DYLIB_MAKE_CMD?=libtool -dynamic -o ${DYLIBNAME} -lm ${DEBUG} - ${OBJ}
+  STLIBNAME?=libhiredis.a
+  STLIB_MAKE_CMD?=libtool -static -o ${STLIBNAME} - ${OBJ}
+else
+  CFLAGS?=-std=c99 -pedantic $(OPTIMIZATION) -fPIC -Wall -W -Wstrict-prototypes -Wwrite-strings $(ARCH) $(PROF)
+  CCLINK?=-lm -pthread
+  LDFLAGS?=-L. -Wl,-rpath,.
+  DYLIBNAME?=libhiredis.so
+  DYLIB_MAKE_CMD?=gcc -shared -Wl,-soname,${DYLIBNAME} -o ${DYLIBNAME} ${OBJ}
+  STLIBNAME?=libhiredis.a
+  STLIB_MAKE_CMD?=ar rcs ${STLIBNAME} ${OBJ}
+endif
+endif
+
+CCOPT= $(CFLAGS) $(CCLINK)
+DEBUG?= -g -ggdb
+
+PREFIX?= /usr/local
+INSTALL_INC= $(PREFIX)/include/hiredis
+INSTALL_LIB= $(PREFIX)/lib
+INSTALL= cp -a
+
+all: ${DYLIBNAME} ${BINS}
+
+# Deps (use make dep to generate this)
+net.o: net.c fmacros.h net.h
+async.o: async.c async.h hiredis.h sds.h util.h dict.c dict.h
+example.o: example.c hiredis.h
+hiredis.o: hiredis.c hiredis.h net.h sds.h util.h
+sds.o: sds.c sds.h
+test.o: test.c hiredis.h
+
+${DYLIBNAME}: ${OBJ}
+	${DYLIB_MAKE_CMD}
+
+${STLIBNAME}: ${OBJ}
+	${STLIB_MAKE_CMD}
+
+dynamic: ${DYLIBNAME}
+static: ${STLIBNAME}
+
+# Binaries:
+hiredis-example-libevent: example-libevent.c adapters/libevent.h ${DYLIBNAME}
+	$(CC) -o $@ $(CCOPT) $(DEBUG) $(LDFLAGS) -lhiredis -levent example-libevent.c
+
+hiredis-example-libev: example-libev.c adapters/libev.h ${DYLIBNAME}
+	$(CC) -o $@ $(CCOPT) $(DEBUG) $(LDFLAGS) -lhiredis -lev example-libev.c
+
+ifndef AE_DIR
+hiredis-example-ae:
+	@echo "Please specify AE_DIR (e.g. <redis repository>/src)"
+	@false
+else
+hiredis-example-ae: example-ae.c adapters/ae.h ${DYLIBNAME}
+	$(CC) -o $@ $(CCOPT) $(DEBUG) -I$(AE_DIR) $(LDFLAGS) -lhiredis example-ae.c $(AE_DIR)/ae.o $(AE_DIR)/zmalloc.o
+endif
+
+hiredis-%: %.o ${DYLIBNAME}
+	$(CC) -o $@ $(CCOPT) $(DEBUG) $(LDFLAGS) -lhiredis $<
+
+test: hiredis-test
+	./hiredis-test
+
+.c.o:
+	$(CC) -c $(CFLAGS) $(OBJARCH) $(DEBUG) $(COMPILE_TIME) $<
+
+clean:
+	rm -rf ${DYLIBNAME} ${STLIBNAME} $(BINS) hiredis-example* *.o *.gcda *.gcno *.gcov
+
+dep:
+	$(CC) -MM *.c
+
+install: ${DYLIBNAME} ${STLIBNAME}
+	mkdir -p $(INSTALL_INC) $(INSTALL_LIB)
+	$(INSTALL) hiredis.h async.h adapters $(INSTALL_INC)
+	$(INSTALL) ${DYLIBNAME} ${STLIBNAME} $(INSTALL_LIB)
+
+32bit:
+	@echo ""
+	@echo "WARNING: if it fails under Linux you probably need to install libc6-dev-i386"
+	@echo ""
+	$(MAKE) ARCH="-m32"
+
+gprof:
+	$(MAKE) PROF="-pg"
+
+gcov:
+	$(MAKE) PROF="-fprofile-arcs -ftest-coverage"
+
+noopt:
+	$(MAKE) OPTIMIZATION=""

deps/hiredis/README.md

@@ -0,0 +1,311 @@
+# HIREDIS
+
+Hiredis is a minimalistic C client library for the [Redis](http://redis.io/) database.
+
+It is minimalistic because it just adds minimal support for the protocol, but
+at the same time it uses an high level printf-alike API in order to make it
+much higher level than otherwise suggested by its minimal code base and the
+lack of explicit bindings for every Redis command.
+
+Apart from supporting sending commands and receiving replies, it comes with
+a reply parser that is decoupled from the I/O layer. It
+is a stream parser designed for easy reusability, which can for instance be used
+in higher level language bindings for efficient reply parsing.
+
+Hiredis only supports the binary-safe Redis protocol, so you can use it with any
+Redis version >= 1.2.0.
+
+The library comes with multiple APIs. There is the
+*synchronous API*, the *asynchronous API* and the *reply parsing API*.
+
+## UPGRADING
+
+Version 0.9.0 is a major overhaul of hiredis in every aspect. However, upgrading existing
+code using hiredis should not be a big pain. The key thing to keep in mind when
+upgrading is that hiredis >= 0.9.0 uses a `redisContext*` to keep state, in contrast to
+the stateless 0.0.1 that only has a file descriptor to work with.
+
+## Synchronous API
+
+To consume the synchronous API, there are only a few function calls that need to be introduced:
+
+    redisContext *redisConnect(const char *ip, int port);
+    void *redisCommand(redisContext *c, const char *format, ...);
+    void freeReplyObject(void *reply);
+
+### Connecting
+
+The function `redisConnect` is used to create a so-called `redisContext`. The
+context is where Hiredis holds state for a connection. The `redisContext`
+struct has an integer `err` field that is non-zero when an the connection is in
+an error state. The field `errstr` will contain a string with a description of
+the error. More information on errors can be found in the **Errors** section.
+After trying to connect to Redis using `redisConnect` you should
+check the `err` field to see if establishing the connection was successful:
+
+    redisContext *c = redisConnect("127.0.0.1", 6379);
+    if (c->err) {
+        printf("Error: %s\n", c->errstr);
+        // handle error
+    }
+
+### Sending commands
+
+There are several ways to issue commands to Redis. The first that will be introduced is
+`redisCommand`. This function takes a format similar to printf. In the simplest form,
+it is used like this:
+
+    reply = redisCommand(context, "SET foo bar");
+
+The specifier `%s` interpolates a string in the command, and uses `strlen` to
+determine the length of the string:
+
+    reply = redisCommand(context, "SET foo %s", value);
+
+When you need to pass binary safe strings in a command, the `%b` specifier can be
+used. Together with a pointer to the string, it requires a `size_t` length argument
+of the string:
+
+    reply = redisCommand(context, "SET foo %b", value, valuelen);
+
+Internally, Hiredis splits the command in different arguments and will
+convert it to the protocol used to communicate with Redis.
+One or more spaces separates arguments, so you can use the specifiers
+anywhere in an argument:
+
+    reply = redisCommand("SET key:%s %s", myid, value);
+
+### Using replies
+
+The return value of `redisCommand` holds a reply when the command was
+successfully executed. When an error occurs, the return value is `NULL` and
+the `err` field in the context will be set (see section on **Errors**).
+Once an error is returned the context cannot be reused and you should set up
+a new connection.
+
+The standard replies that `redisCommand` are of the type `redisReply`. The
+`type` field in the `redisReply` should be used to test what kind of reply
+was received:
+
+* **`REDIS_REPLY_STATUS`**:
+    * The command replied with a status reply. The status string can be accessed using `reply->str`.
+      The length of this string can be accessed using `reply->len`.
+
+* **`REDIS_REPLY_ERROR`**:
+    *  The command replied with an error. The error string can be accessed identical to `REDIS_REPLY_STATUS`.
+
+* **`REDIS_REPLY_INTEGER`**:
+    * The command replied with an integer. The integer value can be accessed using the
+      `reply->integer` field of type `long long`.
+
+* **`REDIS_REPLY_NIL`**:
+    * The command replied with a **nil** object. There is no data to access.
+
+* **`REDIS_REPLY_STRING`**:
+    * A bulk (string) reply. The value of the reply can be accessed using `reply->str`.
+      The length of this string can be accessed using `reply->len`.
+
+* **`REDIS_REPLY_ARRAY`**:
+    * A multi bulk reply. The number of elements in the multi bulk reply is stored in
+      `reply->elements`. Every element in the multi bulk reply is a `redisReply` object as well
+      and can be accessed via `reply->element[..index..]`.
+      Redis may reply with nested arrays but this is fully supported.
+
+Replies should be freed using the `freeReplyObject()` function.
+Note that this function will take care of freeing sub-replies objects
+contained in arrays and nested arrays, so there is no need for the user to
+free the sub replies (it is actually harmful and will corrupt the memory).
+
+### Cleaning up
+
+To disconnect and free the context the following function can be used:
+
+    void redisFree(redisContext *c);
+
+This function immediately closes the socket and then free's the allocations done in
+creating the context.
+
+### Sending commands (cont'd)
+
+Together with `redisCommand`, the function `redisCommandArgv` can be used to issue commands.
+It has the following prototype:
+
+    void *redisCommandArgv(redisContext *c, int argc, const char **argv, const size_t *argvlen);
+
+It takes the number of arguments `argc`, an array of strings `argv` and the lengths of the
+arguments `argvlen`. For convenience, `argvlen` may be set to `NULL` and the function will
+use `strlen(3)` on every argument to determine its length. Obviously, when any of the arguments
+need to be binary safe, the entire array of lengths `argvlen` should be provided.
+
+The return value has the same semantic as `redisCommand`.
+
+### Pipelining
+
+To explain how Hiredis supports pipelining in a blocking connection, there needs to be
+understanding of the internal execution flow.
+
+When any of the functions in the `redisCommand` family is called, Hiredis first formats the
+command according to the Redis protocol. The formatted command is then put in the output buffer
+of the context. This output buffer is dynamic, so it can hold any number of commands.
+After the command is put in the output buffer, `redisGetReply` is called. This function has the
+following two execution paths:
+
+1. The input buffer is non-empty:
+    * Try to parse a single reply from the input buffer and return it
+    * If no reply could be parsed, continue at *2*
+2. The input buffer is empty:
+    * Write the **entire** output buffer to the socket
+    * Read from the socket until a single reply could be parsed
+
+The function `redisGetReply` is exported as part of the Hiredis API and can be used when a reply
+is expected on the socket. To pipeline commands, the only things that needs to be done is
+filling up the output buffer. For this cause, two commands can be used that are identical
+to the `redisCommand` family, apart from not returning a reply:
+
+    void redisAppendCommand(redisContext *c, const char *format, ...);
+    void redisAppendCommandArgv(redisContext *c, int argc, const char **argv, const size_t *argvlen);
+
+After calling either function one or more times, `redisGetReply` can be used to receive the
+subsequent replies. The return value for this function is either `REDIS_OK` or `REDIS_ERR`, where
+the latter means an error occurred while reading a reply. Just as with the other commands,
+the `err` field in the context can be used to find out what the cause of this error is.
+
+The following examples shows a simple pipeline (resulting in only a single call to `write(2)` and
+a single call to `read(2)`):
+
+    redisReply *reply;
+    redisAppendCommand(context,"SET foo bar");
+    redisAppendCommand(context,"GET foo");
+    redisGetReply(context,&reply); // reply for SET
+    freeReplyObject(reply);
+    redisGetReply(context,&reply); // reply for GET
+    freeReplyObject(reply);
+
+This API can also be used to implement a blocking subscriber:
+
+    reply = redisCommand(context,"SUBSCRIBE foo");
+    freeReplyObject(reply);
+    while(redisGetReply(context,&reply) == REDIS_OK) {
+        // consume message
+        freeReplyObject(reply);
+    }
+
+### Errors
+
+When a function call is not successful, depending on the function either `NULL` or `REDIS_ERR` is
+returned. The `err` field inside the context will be non-zero and set to one of the
+following constants:
+
+* **`REDIS_ERR_IO`**:
+    There was an I/O error while creating the connection, trying to write
+    to the socket or read from the socket. If you included `errno.h` in your
+    application, you can use the global `errno` variable to find out what is
+    wrong.
+
+* **`REDIS_ERR_EOF`**:
+    The server closed the connection which resulted in an empty read.
+
+* **`REDIS_ERR_PROTOCOL`**:
+    There was an error while parsing the protocol.
+
+* **`REDIS_ERR_OTHER`**:
+    Any other error. Currently, it is only used when a specified hostname to connect
+    to cannot be resolved.
+
+In every case, the `errstr` field in the context will be set to hold a string representation
+of the error.
+
+## Asynchronous API
+
+Hiredis comes with an asynchronous API that works easily with any event library.
+Examples are bundled that show using Hiredis with [libev](http://software.schmorp.de/pkg/libev.html)
+and [libevent](http://monkey.org/~provos/libevent/).
+
+### Connecting
+
+The function `redisAsyncConnect` can be used to establish a non-blocking connection to
+Redis. It returns a pointer to the newly created `redisAsyncContext` struct. The `err` field
+should be checked after creation to see if there were errors creating the connection.
+Because the connection that will be created is non-blocking, the kernel is not able to
+instantly return if the specified host and port is able to accept a connection.
+
+    redisAsyncContext *c = redisAsyncConnect("127.0.0.1", 6379);
+    if (c->err) {
+        printf("Error: %s\n", c->errstr);
+        // handle error
+    }
+
+The asynchronous context can hold a disconnect callback function that is called when the
+connection is disconnected (either because of an error or per user request). This function should
+have the following prototype:
+
+    void(const redisAsyncContext *c, int status);
+
+On a disconnect, the `status` argument is set to `REDIS_OK` when disconnection was initiated by the
+user, or `REDIS_ERR` when the disconnection was caused by an error. When it is `REDIS_ERR`, the `err`
+field in the context can be accessed to find out the cause of the error.
+
+The context object is always free'd after the disconnect callback fired. When a reconnect is needed,
+the disconnect callback is a good point to do so.
+
+Setting the disconnect callback can only be done once per context. For subsequent calls it will
+return `REDIS_ERR`. The function to set the disconnect callback has the following prototype:
+
+    int redisAsyncSetDisconnectCallback(redisAsyncContext *ac, redisDisconnectCallback *fn);
+
+### Sending commands and their callbacks
+
+In an asynchronous context, commands are automatically pipelined due to the nature of an event loop.
+Therefore, unlike the synchronous API, there is only a single way to send commands.
+Because commands are sent to Redis asynchronously, issuing a command requires a callback function
+that is called when the reply is received. Reply callbacks should have the following prototype:
+
+    void(redisAsyncContext *c, void *reply, void *privdata);
+
+The `privdata` argument can be used to curry arbitrary data to the callback from the point where
+the command is initially queued for execution.
+
+The functions that can be used to issue commands in an asynchronous context are:
+
+    int redisAsyncCommand(
+      redisAsyncContext *ac, redisCallbackFn *fn, void *privdata,
+      const char *format, ...);
+    int redisAsyncCommandArgv(
+      redisAsyncContext *ac, redisCallbackFn *fn, void *privdata,
+      int argc, const char **argv, const size_t *argvlen);
+
+Both functions work like their blocking counterparts. The return value is `REDIS_OK` when the command
+was successfully added to the output buffer and `REDIS_ERR` otherwise. Example: when the connection
+is being disconnected per user-request, no new commands may be added to the output buffer and `REDIS_ERR` is
+returned on calls to the `redisAsyncCommand` family.
+
+If the reply for a command with a `NULL` callback is read, it is immediately free'd. When the callback
+for a command is non-`NULL`, it is responsible for cleaning up the reply.
+
+All pending callbacks are called with a `NULL` reply when the context encountered an error.
+
+### Disconnecting
+
+An asynchronous connection can be terminated using:
+
+    void redisAsyncDisconnect(redisAsyncContext *ac);
+
+When this function is called, the connection is **not** immediately terminated. Instead, new
+commands are no longer accepted and the connection is only terminated when all pending commands
+have been written to the socket, their respective replies have been read and their respective
+callbacks have been executed. After this, the disconnection callback is executed with the
+`REDIS_OK` status and the context object is free'd.
+
+### Hooking it up to event library *X*
+
+There are a few hooks that need to be set on the context object after it is created.
+See the `adapters/` directory for bindings to *libev* and *libevent*.
+
+## Reply parsing API
+
+To be done.
+
+## AUTHORS
+
+Hiredis was written by Salvatore Sanfilippo (antirez at gmail) and
+Pieter Noordhuis (pcnoordhuis at gmail) and is released under the BSD license.

deps/hiredis/TODO

@@ -0,0 +1,2 @@
+- add redisCommandVector()
+- add support for pipelining

deps/hiredis/adapters/ae.h

@@ -0,0 +1,95 @@
+#include <sys/types.h>
+#include <ae.h>
+#include "../hiredis.h"
+#include "../async.h"
+
+typedef struct redisAeEvents {
+    redisAsyncContext *context;
+    aeEventLoop *loop;
+    int fd;
+    int reading, writing;
+} redisAeEvents;
+
+void redisAeReadEvent(aeEventLoop *el, int fd, void *privdata, int mask) {
+    ((void)el); ((void)fd); ((void)mask);
+
+    redisAeEvents *e = (redisAeEvents*)privdata;
+    redisAsyncHandleRead(e->context);
+}
+
+void redisAeWriteEvent(aeEventLoop *el, int fd, void *privdata, int mask) {
+    ((void)el); ((void)fd); ((void)mask);
+
+    redisAeEvents *e = (redisAeEvents*)privdata;
+    redisAsyncHandleWrite(e->context);
+}
+
+void redisAeAddRead(void *privdata) {
+    redisAeEvents *e = (redisAeEvents*)privdata;
+    aeEventLoop *loop = e->loop;
+    if (!e->reading) {
+        e->reading = 1;
+        aeCreateFileEvent(loop,e->fd,AE_READABLE,redisAeReadEvent,e);
+    }
+}
+
+void redisAeDelRead(void *privdata) {
+    redisAeEvents *e = (redisAeEvents*)privdata;
+    aeEventLoop *loop = e->loop;
+    if (e->reading) {
+        e->reading = 0;
+        aeDeleteFileEvent(loop,e->fd,AE_READABLE);
+    }
+}
+
+void redisAeAddWrite(void *privdata) {
+    redisAeEvents *e = (redisAeEvents*)privdata;
+    aeEventLoop *loop = e->loop;
+    if (!e->writing) {
+        e->writing = 1;
+        aeCreateFileEvent(loop,e->fd,AE_WRITABLE,redisAeWriteEvent,e);
+    }
+}
+
+void redisAeDelWrite(void *privdata) {
+    redisAeEvents *e = (redisAeEvents*)privdata;
+    aeEventLoop *loop = e->loop;
+    if (e->writing) {
+        e->writing = 0;
+        aeDeleteFileEvent(loop,e->fd,AE_WRITABLE);
+    }
+}
+
+void redisAeCleanup(void *privdata) {
+    redisAeEvents *e = (redisAeEvents*)privdata;
+    redisAeDelRead(privdata);
+    redisAeDelWrite(privdata);
+    free(e);
+}
+
+int redisAeAttach(aeEventLoop *loop, redisAsyncContext *ac) {
+    redisContext *c = &(ac->c);
+    redisAeEvents *e;
+
+    /* Nothing should be attached when something is already attached */
+    if (ac->_adapter_data != NULL)
+        return REDIS_ERR;
+
+    /* Create container for context and r/w events */
+    e = (redisAeEvents*)malloc(sizeof(*e));
+    e->context = ac;
+    e->loop = loop;
+    e->fd = c->fd;
+    e->reading = e->writing = 0;
+
+    /* Register functions to start/stop listening for events */
+    ac->evAddRead = redisAeAddRead;
+    ac->evDelRead = redisAeDelRead;
+    ac->evAddWrite = redisAeAddWrite;
+    ac->evDelWrite = redisAeDelWrite;
+    ac->evCleanup = redisAeCleanup;
+    ac->_adapter_data = e;
+
+    return REDIS_OK;
+}
+

deps/hiredis/adapters/libev.h

@@ -0,0 +1,113 @@
+#include <sys/types.h>
+#include <ev.h>
+#include "../hiredis.h"
+#include "../async.h"
+
+typedef struct redisLibevEvents {
+    redisAsyncContext *context;
+    struct ev_loop *loop;
+    int reading, writing;
+    ev_io rev, wev;
+} redisLibevEvents;
+
+void redisLibevReadEvent(EV_P_ ev_io *watcher, int revents) {
+#if EV_MULTIPLICITY
+    ((void)loop);
+#endif
+    ((void)revents);
+
+    redisLibevEvents *e = (redisLibevEvents*)watcher->data;
+    redisAsyncHandleRead(e->context);
+}
+
+void redisLibevWriteEvent(EV_P_ ev_io *watcher, int revents) {
+#if EV_MULTIPLICITY
+    ((void)loop);
+#endif
+    ((void)revents);
+
+    redisLibevEvents *e = (redisLibevEvents*)watcher->data;
+    redisAsyncHandleWrite(e->context);
+}
+
+void redisLibevAddRead(void *privdata) {
+    redisLibevEvents *e = (redisLibevEvents*)privdata;
+    struct ev_loop *loop = e->loop;
+    ((void)loop);
+    if (!e->reading) {
+        e->reading = 1;
+        ev_io_start(EV_A_ &e->rev);
+    }
+}
+
+void redisLibevDelRead(void *privdata) {
+    redisLibevEvents *e = (redisLibevEvents*)privdata;
+    struct ev_loop *loop = e->loop;
+    ((void)loop);
+    if (e->reading) {
+        e->reading = 0;
+        ev_io_stop(EV_A_ &e->rev);
+    }
+}
+
+void redisLibevAddWrite(void *privdata) {
+    redisLibevEvents *e = (redisLibevEvents*)privdata;
+    struct ev_loop *loop = e->loop;
+    ((void)loop);
+    if (!e->writing) {
+        e->writing = 1;
+        ev_io_start(EV_A_ &e->wev);
+    }
+}
+
+void redisLibevDelWrite(void *privdata) {
+    redisLibevEvents *e = (redisLibevEvents*)privdata;
+    struct ev_loop *loop = e->loop;
+    ((void)loop);
+    if (e->writing) {
+        e->writing = 0;
+        ev_io_stop(EV_A_ &e->wev);
+    }
+}
+
+void redisLibevCleanup(void *privdata) {
+    redisLibevEvents *e = (redisLibevEvents*)privdata;
+    redisLibevDelRead(privdata);
+    redisLibevDelWrite(privdata);
+    free(e);
+}
+
+int redisLibevAttach(EV_P_ redisAsyncContext *ac) {
+    redisContext *c = &(ac->c);
+    redisLibevEvents *e;
+
+    /* Nothing should be attached when something is already attached */
+    if (ac->_adapter_data != NULL)
+        return REDIS_ERR;
+
+    /* Create container for context and r/w events */
+    e = (redisLibevEvents*)malloc(sizeof(*e));
+    e->context = ac;
+#if EV_MULTIPLICITY
+    e->loop = loop;
+#else
+    e->loop = NULL;
+#endif
+    e->reading = e->writing = 0;
+    e->rev.data = e;
+    e->wev.data = e;
+
+    /* Register functions to start/stop listening for events */
+    ac->evAddRead = redisLibevAddRead;
+    ac->evDelRead = redisLibevDelRead;
+    ac->evAddWrite = redisLibevAddWrite;
+    ac->evDelWrite = redisLibevDelWrite;
+    ac->evCleanup = redisLibevCleanup;
+    ac->_adapter_data = e;
+
+    /* Initialize read/write events */
+    ev_io_init(&e->rev,redisLibevReadEvent,c->fd,EV_READ);
+    ev_io_init(&e->wev,redisLibevWriteEvent,c->fd,EV_WRITE);
+    return REDIS_OK;
+}
+

deps/hiredis/adapters/libevent.h

@@ -0,0 +1,76 @@
+#include <sys/types.h>
+#include <event.h>
+#include "../hiredis.h"
+#include "../async.h"
+
+typedef struct redisLibeventEvents {
+    redisAsyncContext *context;
+    struct event rev, wev;
+} redisLibeventEvents;
+
+void redisLibeventReadEvent(int fd, short event, void *arg) {
+    ((void)fd); ((void)event);
+    redisLibeventEvents *e = (redisLibeventEvents*)arg;
+    redisAsyncHandleRead(e->context);
+}
+
+void redisLibeventWriteEvent(int fd, short event, void *arg) {
+    ((void)fd); ((void)event);
+    redisLibeventEvents *e = (redisLibeventEvents*)arg;
+    redisAsyncHandleWrite(e->context);
+}
+
+void redisLibeventAddRead(void *privdata) {
+    redisLibeventEvents *e = (redisLibeventEvents*)privdata;
+    event_add(&e->rev,NULL);
+}
+
+void redisLibeventDelRead(void *privdata) {
+    redisLibeventEvents *e = (redisLibeventEvents*)privdata;
+    event_del(&e->rev);
+}
+
+void redisLibeventAddWrite(void *privdata) {
+    redisLibeventEvents *e = (redisLibeventEvents*)privdata;
+    event_add(&e->wev,NULL);
+}
+
+void redisLibeventDelWrite(void *privdata) {
+    redisLibeventEvents *e = (redisLibeventEvents*)privdata;
+    event_del(&e->wev);
+}
+
+void redisLibeventCleanup(void *privdata) {
+    redisLibeventEvents *e = (redisLibeventEvents*)privdata;
+    event_del(&e->rev);
+    event_del(&e->wev);
+    free(e);
+}
+
+int redisLibeventAttach(redisAsyncContext *ac, struct event_base *base) {
+    redisContext *c = &(ac->c);
+    redisLibeventEvents *e;
+
+    /* Nothing should be attached when something is already attached */
+    if (ac->_adapter_data != NULL)
+        return REDIS_ERR;
+
+    /* Create container for context and r/w events */
+    e = (redisLibeventEvents*)malloc(sizeof(*e));
+    e->context = ac;
+
+    /* Register functions to start/stop listening for events */
+    ac->evAddRead = redisLibeventAddRead;
+    ac->evDelRead = redisLibeventDelRead;
+    ac->evAddWrite = redisLibeventAddWrite;
+    ac->evDelWrite = redisLibeventDelWrite;
+    ac->evCleanup = redisLibeventCleanup;
+    ac->_adapter_data = e;
+
+    /* Initialize and install read/write events */
+    event_set(&e->rev,c->fd,EV_READ,redisLibeventReadEvent,e);
+    event_set(&e->wev,c->fd,EV_WRITE,redisLibeventWriteEvent,e);
+    event_base_set(base,&e->rev);
+    event_base_set(base,&e->wev);
+    return REDIS_OK;
+}

deps/hiredis/async.c

@@ -0,0 +1,545 @@
+/*
+ * Copyright (c) 2009-2010, Salvatore Sanfilippo <antirez at gmail dot com>
+ * Copyright (c) 2010, Pieter Noordhuis <pcnoordhuis at gmail dot com>
+ *
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ *   * Redistributions of source code must retain the above copyright notice,
+ *     this list of conditions and the following disclaimer.
+ *   * Redistributions in binary form must reproduce the above copyright
+ *     notice, this list of conditions and the following disclaimer in the
+ *     documentation and/or other materials provided with the distribution.
+ *   * Neither the name of Redis nor the names of its contributors may be used
+ *     to endorse or promote products derived from this software without
+ *     specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <string.h>
+#include <strings.h>
+#include <assert.h>
+#include <ctype.h>
+#include "async.h"
+#include "dict.c"
+#include "sds.h"
+#include "util.h"
+
+/* Forward declaration of function in hiredis.c */
+void __redisAppendCommand(redisContext *c, char *cmd, size_t len);
+
+/* Functions managing dictionary of callbacks for pub/sub. */
+static unsigned int callbackHash(const void *key) {
+    return dictGenHashFunction((unsigned char*)key,sdslen((char*)key));
+}
+
+static void *callbackValDup(void *privdata, const void *src) {
+    ((void) privdata);
+    redisCallback *dup = malloc(sizeof(*dup));
+    memcpy(dup,src,sizeof(*dup));
+    return dup;
+}
+
+static int callbackKeyCompare(void *privdata, const void *key1, const void *key2) {
+    int l1, l2;
+    ((void) privdata);
+
+    l1 = sdslen((sds)key1);
+    l2 = sdslen((sds)key2);
+    if (l1 != l2) return 0;
+    return memcmp(key1,key2,l1) == 0;
+}
+
+static void callbackKeyDestructor(void *privdata, void *key) {
+    ((void) privdata);
+    sdsfree((sds)key);
+}
+
+static void callbackValDestructor(void *privdata, void *val) {
+    ((void) privdata);
+    free(val);
+}
+
+static dictType callbackDict = {
+    callbackHash,
+    NULL,
+    callbackValDup,
+    callbackKeyCompare,
+    callbackKeyDestructor,
+    callbackValDestructor
+};
+
+static redisAsyncContext *redisAsyncInitialize(redisContext *c) {
+    redisAsyncContext *ac = realloc(c,sizeof(redisAsyncContext));
+    c = &(ac->c);
+
+    /* The regular connect functions will always set the flag REDIS_CONNECTED.
+     * For the async API, we want to wait until the first write event is
+     * received up before setting this flag, so reset it here. */
+    c->flags &= ~REDIS_CONNECTED;
+
+    ac->err = 0;
+    ac->errstr = NULL;
+    ac->data = NULL;
+
+    ac->ev.data = NULL;
+    ac->ev.addRead = NULL;
+    ac->ev.delRead = NULL;
+    ac->ev.addWrite = NULL;
+    ac->ev.delWrite = NULL;
+    ac->ev.cleanup = NULL;
+
+    ac->onConnect = NULL;
+    ac->onDisconnect = NULL;
+
+    ac->replies.head = NULL;
+    ac->replies.tail = NULL;
+    ac->sub.invalid.head = NULL;
+    ac->sub.invalid.tail = NULL;
+    ac->sub.channels = dictCreate(&callbackDict,NULL);
+    ac->sub.patterns = dictCreate(&callbackDict,NULL);
+    return ac;
+}
+
+/* We want the error field to be accessible directly instead of requiring
+ * an indirection to the redisContext struct. */
+static void __redisAsyncCopyError(redisAsyncContext *ac) {
+    redisContext *c = &(ac->c);
+    ac->err = c->err;
+    ac->errstr = c->errstr;
+}
+
+redisAsyncContext *redisAsyncConnect(const char *ip, int port) {
+    redisContext *c = redisConnectNonBlock(ip,port);
+    redisAsyncContext *ac = redisAsyncInitialize(c);
+    __redisAsyncCopyError(ac);
+    return ac;
+}
+
+redisAsyncContext *redisAsyncConnectUnix(const char *path) {
+    redisContext *c = redisConnectUnixNonBlock(path);
+    redisAsyncContext *ac = redisAsyncInitialize(c);
+    __redisAsyncCopyError(ac);
+    return ac;
+}
+
+int redisAsyncSetReplyObjectFunctions(redisAsyncContext *ac, redisReplyObjectFunctions *fn) {
+    redisContext *c = &(ac->c);
+    return redisSetReplyObjectFunctions(c,fn);
+}
+
+int redisAsyncSetConnectCallback(redisAsyncContext *ac, redisConnectCallback *fn) {
+    if (ac->onConnect == NULL) {
+        ac->onConnect = fn;
+
+        /* The common way to detect an established connection is to wait for
+         * the first write event to be fired. This assumes the related event
+         * library functions are already set. */
+        if (ac->ev.addWrite) ac->ev.addWrite(ac->ev.data);
+        return REDIS_OK;
+    }
+    return REDIS_ERR;
+}
+
+int redisAsyncSetDisconnectCallback(redisAsyncContext *ac, redisDisconnectCallback *fn) {
+    if (ac->onDisconnect == NULL) {
+        ac->onDisconnect = fn;
+        return REDIS_OK;
+    }
+    return REDIS_ERR;
+}
+
+/* Helper functions to push/shift callbacks */
+static int __redisPushCallback(redisCallbackList *list, redisCallback *source) {
+    redisCallback *cb;
+
+    /* Copy callback from stack to heap */
+    cb = malloc(sizeof(*cb));
+    if (!cb) redisOOM();
+    if (source != NULL) {
+        memcpy(cb,source,sizeof(*cb));
+        cb->next = NULL;
+    }
+
+    /* Store callback in list */
+    if (list->head == NULL)
+        list->head = cb;
+    if (list->tail != NULL)
+        list->tail->next = cb;
+    list->tail = cb;
+    return REDIS_OK;
+}
+
+static int __redisShiftCallback(redisCallbackList *list, redisCallback *target) {
+    redisCallback *cb = list->head;
+    if (cb != NULL) {
+        list->head = cb->next;
+        if (cb == list->tail)
+            list->tail = NULL;
+
+        /* Copy callback from heap to stack */
+        if (target != NULL)
+            memcpy(target,cb,sizeof(*cb));
+        free(cb);
+        return REDIS_OK;
+    }
+    return REDIS_ERR;
+}
+
+static void __redisRunCallback(redisAsyncContext *ac, redisCallback *cb, redisReply *reply) {
+    redisContext *c = &(ac->c);
+    if (cb->fn != NULL) {
+        c->flags |= REDIS_IN_CALLBACK;
+        cb->fn(ac,reply,cb->privdata);
+        c->flags &= ~REDIS_IN_CALLBACK;
+    }
+}
+
+/* Helper function to free the context. */
+static void __redisAsyncFree(redisAsyncContext *ac) {
+    redisContext *c = &(ac->c);
+    redisCallback cb;
+    dictIterator *it;
+    dictEntry *de;
+
+    /* Execute pending callbacks with NULL reply. */
+    while (__redisShiftCallback(&ac->replies,&cb) == REDIS_OK)
+        __redisRunCallback(ac,&cb,NULL);
+
+    /* Execute callbacks for invalid commands */
+    while (__redisShiftCallback(&ac->sub.invalid,&cb) == REDIS_OK)
+        __redisRunCallback(ac,&cb,NULL);
+
+    /* Run subscription callbacks callbacks with NULL reply */
+    it = dictGetIterator(ac->sub.channels);
+    while ((de = dictNext(it)) != NULL)
+        __redisRunCallback(ac,dictGetEntryVal(de),NULL);
+    dictReleaseIterator(it);
+    dictRelease(ac->sub.channels);
+
+    it = dictGetIterator(ac->sub.patterns);
+    while ((de = dictNext(it)) != NULL)
+        __redisRunCallback(ac,dictGetEntryVal(de),NULL);
+    dictReleaseIterator(it);
+    dictRelease(ac->sub.patterns);
+
+    /* Signal event lib to clean up */
+    if (ac->ev.cleanup) ac->ev.cleanup(ac->ev.data);
+
+    /* Execute disconnect callback. When redisAsyncFree() initiated destroying
+     * this context, the status will always be REDIS_OK. */
+    if (ac->onDisconnect && (c->flags & REDIS_CONNECTED)) {
+        if (c->flags & REDIS_FREEING) {
+            ac->onDisconnect(ac,REDIS_OK);
+        } else {
+            ac->onDisconnect(ac,(ac->err == 0) ? REDIS_OK : REDIS_ERR);
+        }
+    }
+
+    /* Cleanup self */
+    redisFree(c);
+}
+
+/* Free the async context. When this function is called from a callback,
+ * control needs to be returned to redisProcessCallbacks() before actual
+ * free'ing. To do so, a flag is set on the context which is picked up by
+ * redisProcessCallbacks(). Otherwise, the context is immediately free'd. */
+void redisAsyncFree(redisAsyncContext *ac) {
+    redisContext *c = &(ac->c);
+    c->flags |= REDIS_FREEING;
+    if (!(c->flags & REDIS_IN_CALLBACK))
+        __redisAsyncFree(ac);
+}
+
+/* Helper function to make the disconnect happen and clean up. */
+static void __redisAsyncDisconnect(redisAsyncContext *ac) {
+    redisContext *c = &(ac->c);
+
+    /* Make sure error is accessible if there is any */
+    __redisAsyncCopyError(ac);
+
+    if (ac->err == 0) {
+        /* For clean disconnects, there should be no pending callbacks. */
+        assert(__redisShiftCallback(&ac->replies,NULL) == REDIS_ERR);
+    } else {
+        /* Disconnection is caused by an error, make sure that pending
+         * callbacks cannot call new commands. */
+        c->flags |= REDIS_DISCONNECTING;
+    }
+
+    /* For non-clean disconnects, __redisAsyncFree() will execute pending
+     * callbacks with a NULL-reply. */
+    __redisAsyncFree(ac);
+}
+
+/* Tries to do a clean disconnect from Redis, meaning it stops new commands
+ * from being issued, but tries to flush the output buffer and execute
+ * callbacks for all remaining replies. When this function is called from a
+ * callback, there might be more replies and we can safely defer disconnecting
+ * to redisProcessCallbacks(). Otherwise, we can only disconnect immediately
+ * when there are no pending callbacks. */
+void redisAsyncDisconnect(redisAsyncContext *ac) {
+    redisContext *c = &(ac->c);
+    c->flags |= REDIS_DISCONNECTING;
+    if (!(c->flags & REDIS_IN_CALLBACK) && ac->replies.head == NULL)
+        __redisAsyncDisconnect(ac);
+}
+
+static int __redisGetSubscribeCallback(redisAsyncContext *ac, redisReply *reply, redisCallback *dstcb) {
+    redisContext *c = &(ac->c);
+    dict *callbacks;
+    dictEntry *de;
+    int pvariant;
+    char *stype;
+    sds sname;
+
+    /* Custom reply functions are not supported for pub/sub. This will fail
+     * very hard when they are used... */
+    if (reply->type == REDIS_REPLY_ARRAY) {
+        assert(reply->elements >= 2);
+        assert(reply->element[0]->type == REDIS_REPLY_STRING);
+        stype = reply->element[0]->str;
+        pvariant = (tolower(stype[0]) == 'p') ? 1 : 0;
+
+        if (pvariant)
+            callbacks = ac->sub.patterns;
+        else
+            callbacks = ac->sub.channels;
+
+        /* Locate the right callback */
+        assert(reply->element[1]->type == REDIS_REPLY_STRING);
+        sname = sdsnewlen(reply->element[1]->str,reply->element[1]->len);
+        de = dictFind(callbacks,sname);
+        if (de != NULL) {
+            memcpy(dstcb,dictGetEntryVal(de),sizeof(*dstcb));
+
+            /* If this is an unsubscribe message, remove it. */
+            if (strcasecmp(stype+pvariant,"unsubscribe") == 0) {
+                dictDelete(callbacks,sname);
+
+                /* If this was the last unsubscribe message, revert to
+                 * non-subscribe mode. */
+                assert(reply->element[2]->type == REDIS_REPLY_INTEGER);
+                if (reply->element[2]->integer == 0)
+                    c->flags &= ~REDIS_SUBSCRIBED;
+            }
+        }
+        sdsfree(sname);
+    } else {
+        /* Shift callback for invalid commands. */
+        __redisShiftCallback(&ac->sub.invalid,dstcb);
+    }
+    return REDIS_OK;
+}
+
+void redisProcessCallbacks(redisAsyncContext *ac) {
+    redisContext *c = &(ac->c);
+    redisCallback cb;
+    void *reply = NULL;
+    int status;
+
+    while((status = redisGetReply(c,&reply)) == REDIS_OK) {
+        if (reply == NULL) {
+            /* When the connection is being disconnected and there are
+             * no more replies, this is the cue to really disconnect. */
+            if (c->flags & REDIS_DISCONNECTING && sdslen(c->obuf) == 0) {
+                __redisAsyncDisconnect(ac);
+                return;
+            }
+
+            /* When the connection is not being disconnected, simply stop
+             * trying to get replies and wait for the next loop tick. */
+            break;
+        }
+
+        /* Even if the context is subscribed, pending regular callbacks will
+         * get a reply before pub/sub messages arrive. */
+        if (__redisShiftCallback(&ac->replies,&cb) != REDIS_OK) {
+            /* No more regular callbacks, the context *must* be subscribed. */
+            assert(c->flags & REDIS_SUBSCRIBED);
+            __redisGetSubscribeCallback(ac,reply,&cb);
+        }
+
+        if (cb.fn != NULL) {
+            __redisRunCallback(ac,&cb,reply);
+            c->fn->freeObject(reply);
+
+            /* Proceed with free'ing when redisAsyncFree() was called. */
+            if (c->flags & REDIS_FREEING) {
+                __redisAsyncFree(ac);
+                return;
+            }
+        } else {
+            /* No callback for this reply. This can either be a NULL callback,
+             * or there were no callbacks to begin with. Either way, don't
+             * abort with an error, but simply ignore it because the client
+             * doesn't know what the server will spit out over the wire. */
+            c->fn->freeObject(reply);
+        }
+    }
+
+    /* Disconnect when there was an error reading the reply */
+    if (status != REDIS_OK)
+        __redisAsyncDisconnect(ac);
+}
+
+/* This function should be called when the socket is readable.
+ * It processes all replies that can be read and executes their callbacks.
+ */
+void redisAsyncHandleRead(redisAsyncContext *ac) {
+    redisContext *c = &(ac->c);
+
+    if (redisBufferRead(c) == REDIS_ERR) {
+        __redisAsyncDisconnect(ac);
+    } else {
+        /* Always re-schedule reads */
+        if (ac->ev.addRead) ac->ev.addRead(ac->ev.data);
+        redisProcessCallbacks(ac);
+    }
+}
+
+void redisAsyncHandleWrite(redisAsyncContext *ac) {
+    redisContext *c = &(ac->c);
+    int done = 0;
+
+    if (redisBufferWrite(c,&done) == REDIS_ERR) {
+        __redisAsyncDisconnect(ac);
+    } else {
+        /* Continue writing when not done, stop writing otherwise */
+        if (!done) {
+            if (ac->ev.addWrite) ac->ev.addWrite(ac->ev.data);
+        } else {
+            if (ac->ev.delWrite) ac->ev.delWrite(ac->ev.data);
+        }
+
+        /* Always schedule reads after writes */
+        if (ac->ev.addRead) ac->ev.addRead(ac->ev.data);
+
+        /* Fire onConnect when this is the first write event. */
+        if (!(c->flags & REDIS_CONNECTED)) {
+            c->flags |= REDIS_CONNECTED;
+            if (ac->onConnect) ac->onConnect(ac);
+        }
+    }
+}
+
+/* Sets a pointer to the first argument and its length starting at p. Returns
+ * the number of bytes to skip to get to the following argument. */
+static char *nextArgument(char *start, char **str, size_t *len) {
+    char *p = start;
+    if (p[0] != '$') {
+        p = strchr(p,'$');
+        if (p == NULL) return NULL;
+    }
+
+    *len = (int)strtol(p+1,NULL,10);
+    p = strchr(p,'\r');
+    assert(p);
+    *str = p+2;
+    return p+2+(*len)+2;
+}
+
+/* Helper function for the redisAsyncCommand* family of functions. Writes a
+ * formatted command to the output buffer and registers the provided callback
+ * function with the context. */
+static int __redisAsyncCommand(redisAsyncContext *ac, redisCallbackFn *fn, void *privdata, char *cmd, size_t len) {
+    redisContext *c = &(ac->c);
+    redisCallback cb;
+    int pvariant, hasnext;
+    char *cstr, *astr;
+    size_t clen, alen;
+    char *p;
+    sds sname;
+
+    /* Don't accept new commands when the connection is about to be closed. */
+    if (c->flags & (REDIS_DISCONNECTING | REDIS_FREEING)) return REDIS_ERR;
+
+    /* Setup callback */
+    cb.fn = fn;
+    cb.privdata = privdata;
+
+    /* Find out which command will be appended. */
+    p = nextArgument(cmd,&cstr,&clen);
+    assert(p != NULL);
+    hasnext = (p[0] == '$');
+    pvariant = (tolower(cstr[0]) == 'p') ? 1 : 0;
+    cstr += pvariant;
+    clen -= pvariant;
+
+    if (hasnext && strncasecmp(cstr,"subscribe\r\n",11) == 0) {
+        c->flags |= REDIS_SUBSCRIBED;
+
+        /* Add every channel/pattern to the list of subscription callbacks. */
+        while ((p = nextArgument(p,&astr,&alen)) != NULL) {
+            sname = sdsnewlen(astr,alen);
+            if (pvariant)
+                dictReplace(ac->sub.patterns,sname,&cb);
+            else
+                dictReplace(ac->sub.channels,sname,&cb);
+        }
+    } else if (strncasecmp(cstr,"unsubscribe\r\n",13) == 0) {
+        /* It is only useful to call (P)UNSUBSCRIBE when the context is
+         * subscribed to one or more channels or patterns. */
+        if (!(c->flags & REDIS_SUBSCRIBED)) return REDIS_ERR;
+
+        /* (P)UNSUBSCRIBE does not have its own response: every channel or
+         * pattern that is unsubscribed will receive a message. This means we
+         * should not append a callback function for this command. */
+    } else {
+        if (c->flags & REDIS_SUBSCRIBED)
+            /* This will likely result in an error reply, but it needs to be
+             * received and passed to the callback. */
+            __redisPushCallback(&ac->sub.invalid,&cb);
+        else
+            __redisPushCallback(&ac->replies,&cb);
+    }
+
+    __redisAppendCommand(c,cmd,len);
+
+    /* Always schedule a write when the write buffer is non-empty */
+    if (ac->ev.addWrite) ac->ev.addWrite(ac->ev.data);
+
+    return REDIS_OK;
+}
+
+int redisvAsyncCommand(redisAsyncContext *ac, redisCallbackFn *fn, void *privdata, const char *format, va_list ap) {
+    char *cmd;
+    int len;
+    int status;
+    len = redisvFormatCommand(&cmd,format,ap);
+    status = __redisAsyncCommand(ac,fn,privdata,cmd,len);
+    free(cmd);
+    return status;
+}
+
+int redisAsyncCommand(redisAsyncContext *ac, redisCallbackFn *fn, void *privdata, const char *format, ...) {
+    va_list ap;
+    int status;
+    va_start(ap,format);
+    status = redisvAsyncCommand(ac,fn,privdata,format,ap);
+    va_end(ap);
+    return status;
+}
+
+int redisAsyncCommandArgv(redisAsyncContext *ac, redisCallbackFn *fn, void *privdata, int argc, const char **argv, const size_t *argvlen) {
+    char *cmd;
+    int len;
+    int status;
+    len = redisFormatCommandArgv(&cmd,argc,argv,argvlen);
+    status = __redisAsyncCommand(ac,fn,privdata,cmd,len);
+    free(cmd);
+    return status;
+}

deps/hiredis/async.h

@@ -0,0 +1,126 @@
+/*
+ * Copyright (c) 2009-2010, Salvatore Sanfilippo <antirez at gmail dot com>
+ * Copyright (c) 2010, Pieter Noordhuis <pcnoordhuis at gmail dot com>
+ *
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ *   * Redistributions of source code must retain the above copyright notice,
+ *     this list of conditions and the following disclaimer.
+ *   * Redistributions in binary form must reproduce the above copyright
+ *     notice, this list of conditions and the following disclaimer in the
+ *     documentation and/or other materials provided with the distribution.
+ *   * Neither the name of Redis nor the names of its contributors may be used
+ *     to endorse or promote products derived from this software without
+ *     specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#ifndef __HIREDIS_ASYNC_H
+#define __HIREDIS_ASYNC_H
+#include "hiredis.h"
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+struct redisAsyncContext; /* need forward declaration of redisAsyncContext */
+struct dict; /* dictionary header is included in async.c */
+
+/* Reply callback prototype and container */
+typedef void (redisCallbackFn)(struct redisAsyncContext*, void*, void*);
+typedef struct redisCallback {
+    struct redisCallback *next; /* simple singly linked list */
+    redisCallbackFn *fn;
+    void *privdata;
+} redisCallback;
+
+/* List of callbacks for either regular replies or pub/sub */
+typedef struct redisCallbackList {
+    redisCallback *head, *tail;
+} redisCallbackList;
+
+/* Connection callback prototypes */
+typedef void (redisDisconnectCallback)(const struct redisAsyncContext*, int status);
+typedef void (redisConnectCallback)(const struct redisAsyncContext*);
+
+/* Context for an async connection to Redis */
+typedef struct redisAsyncContext {
+    /* Hold the regular context, so it can be realloc'ed. */
+    redisContext c;
+
+    /* Setup error flags so they can be used directly. */
+    int err;
+    char *errstr;
+
+    /* Not used by hiredis */
+    void *data;
+
+    /* Event library data and hooks */
+    struct {
+        void *data;
+
+        /* Hooks that are called when the library expects to start
+         * reading/writing. These functions should be idempotent. */
+        void (*addRead)(void *privdata);
+        void (*delRead)(void *privdata);
+        void (*addWrite)(void *privdata);
+        void (*delWrite)(void *privdata);
+        void (*cleanup)(void *privdata);
+    } ev;
+
+    /* Called when either the connection is terminated due to an error or per
+     * user request. The status is set accordingly (REDIS_OK, REDIS_ERR). */
+    redisDisconnectCallback *onDisconnect;
+
+    /* Called when the first write event was received. */
+    redisConnectCallback *onConnect;
+
+    /* Regular command callbacks */
+    redisCallbackList replies;
+
+    /* Subscription callbacks */
+    struct {
+        redisCallbackList invalid;
+        struct dict *channels;
+        struct dict *patterns;
+    } sub;
+} redisAsyncContext;
+
+/* Functions that proxy to hiredis */
+redisAsyncContext *redisAsyncConnect(const char *ip, int port);
+redisAsyncContext *redisAsyncConnectUnix(const char *path);
+int redisAsyncSetReplyObjectFunctions(redisAsyncContext *ac, redisReplyObjectFunctions *fn);
+int redisAsyncSetConnectCallback(redisAsyncContext *ac, redisConnectCallback *fn);
+int redisAsyncSetDisconnectCallback(redisAsyncContext *ac, redisDisconnectCallback *fn);
+void redisAsyncDisconnect(redisAsyncContext *ac);
+void redisAsyncFree(redisAsyncContext *ac);
+
+/* Handle read/write events */
+void redisAsyncHandleRead(redisAsyncContext *ac);
+void redisAsyncHandleWrite(redisAsyncContext *ac);
+
+/* Command functions for an async context. Write the command to the
+ * output buffer and register the provided callback. */
+int redisvAsyncCommand(redisAsyncContext *ac, redisCallbackFn *fn, void *privdata, const char *format, va_list ap);
+int redisAsyncCommand(redisAsyncContext *ac, redisCallbackFn *fn, void *privdata, const char *format, ...);
+int redisAsyncCommandArgv(redisAsyncContext *ac, redisCallbackFn *fn, void *privdata, int argc, const char **argv, const size_t *argvlen);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif

deps/hiredis/dict.c

@@ -1,4 +1,4 @@
-/* Hash Tables Implementation.
+/* Hash table implementation.
  *
  * This file implements in memory hash tables with insert/del/replace/find/
  * get-random-element operations. Hash tables will auto resize if needed
@@ -34,43 +34,10 @@
  */
 
 #include "fmacros.h"
-
-#include <stdio.h>
 #include <stdlib.h>
-#include <string.h>
-#include <stdarg.h>
 #include <assert.h>
 #include <limits.h>
-
 #include "dict.h"
-#include "zmalloc.h"
-
-/* ---------------------------- Utility funcitons --------------------------- */
-
-static void _dictPanic(const char *fmt, ...)
-{
-    va_list ap;
-
-    va_start(ap, fmt);
-    fprintf(stderr, "\nDICT LIBRARY PANIC: ");
-    vfprintf(stderr, fmt, ap);
-    fprintf(stderr, "\n\n");
-    va_end(ap);
-}
-
-/* ------------------------- Heap Management Wrappers------------------------ */
-
-static void *_dictAlloc(size_t size)
-{
-    void *p = zmalloc(size);
-    if (p == NULL)
-        _dictPanic("Out of memory");
-    return p;
-}
-
-static void _dictFree(void *ptr) {
-    zfree(ptr);
-}
 
 /* -------------------------- private prototypes ---------------------------- */
 
@@ -81,27 +48,9 @@ static int _dictInit(dict *ht, dictType *type, void *privDataPtr);
 
 /* -------------------------- hash functions -------------------------------- */
 
-/* Thomas Wang's 32 bit Mix Function */
-unsigned int dictIntHashFunction(unsigned int key)
-{
-    key += ~(key << 15);
-    key ^=  (key >> 10);
-    key +=  (key << 3);
-    key ^=  (key >> 6);
-    key += ~(key << 11);
-    key ^=  (key >> 16);
-    return key;
-}
-
-/* Identity hash function for integer keys */
-unsigned int dictIdentityHashFunction(unsigned int key)
-{
-    return key;
-}
-
 /* Generic hash function (a popular one from Bernstein).
  * I tested a few and this was the best. */
-unsigned int dictGenHashFunction(const unsigned char *buf, int len) {
+static unsigned int dictGenHashFunction(const unsigned char *buf, int len) {
     unsigned int hash = 5381;
 
     while (len--)
@@ -113,8 +62,7 @@ unsigned int dictGenHashFunction(const unsigned char *buf, int len) {
 
 /* Reset an hashtable already initialized with ht_init().
  * NOTE: This function should only called by ht_destroy(). */
-static void _dictReset(dict *ht)
-{
+static void _dictReset(dict *ht) {
     ht->table = NULL;
     ht->size = 0;
     ht->sizemask = 0;
@@ -122,39 +70,22 @@ static void _dictReset(dict *ht)
 }
 
 /* Create a new hash table */
-dict *dictCreate(dictType *type,
-        void *privDataPtr)
-{
-    dict *ht = _dictAlloc(sizeof(*ht));
-
+static dict *dictCreate(dictType *type, void *privDataPtr) {
+    dict *ht = malloc(sizeof(*ht));
     _dictInit(ht,type,privDataPtr);
     return ht;
 }
 
 /* Initialize the hash table */
-int _dictInit(dict *ht, dictType *type,
-        void *privDataPtr)
-{
+static int _dictInit(dict *ht, dictType *type, void *privDataPtr) {
     _dictReset(ht);
     ht->type = type;
     ht->privdata = privDataPtr;
     return DICT_OK;
 }
 
-/* Resize the table to the minimal size that contains all the elements,
- * but with the invariant of a USER/BUCKETS ration near to <= 1 */
-int dictResize(dict *ht)
-{
-    int minimal = ht->used;
-
-    if (minimal < DICT_HT_INITIAL_SIZE)
-        minimal = DICT_HT_INITIAL_SIZE;
-    return dictExpand(ht, minimal);
-}
-
 /* Expand or create the hashtable */
-int dictExpand(dict *ht, unsigned long size)
-{
+static int dictExpand(dict *ht, unsigned long size) {
     dict n; /* the new hashtable */
     unsigned long realsize = _dictNextPower(size), i;
 
@@ -166,10 +97,7 @@ int dictExpand(dict *ht, unsigned long size)
     _dictInit(&n, ht->type, ht->privdata);
     n.size = realsize;
     n.sizemask = realsize-1;
-    n.table = _dictAlloc(realsize*sizeof(dictEntry*));
-
-    /* Initialize all the pointers to NULL */
-    memset(n.table, 0, realsize*sizeof(dictEntry*));
+    n.table = calloc(realsize,sizeof(dictEntry*));
 
     /* Copy all the elements from the old to the new table:
      * note that if the old hash table is empty ht->size is zero,
@@ -179,7 +107,7 @@ int dictExpand(dict *ht, unsigned long size)
         dictEntry *he, *nextHe;
 
         if (ht->table[i] == NULL) continue;
-        
+
         /* For each hash entry on this slot... */
         he = ht->table[i];
         while(he) {
@@ -196,7 +124,7 @@ int dictExpand(dict *ht, unsigned long size)
         }
     }
     assert(ht->used == 0);
-    _dictFree(ht->table);
+    free(ht->table);
 
     /* Remap the new hashtable in the old */
     *ht = n;
@@ -204,8 +132,7 @@ int dictExpand(dict *ht, unsigned long size)
 }
 
 /* Add an element to the target hash table */
-int dictAdd(dict *ht, void *key, void *val)
-{
+static int dictAdd(dict *ht, void *key, void *val) {
     int index;
     dictEntry *entry;
 
@@ -215,7 +142,7 @@ int dictAdd(dict *ht, void *key, void *val)
         return DICT_ERR;
 
     /* Allocates the memory and stores key */
-    entry = _dictAlloc(sizeof(*entry));
+    entry = malloc(sizeof(*entry));
     entry->next = ht->table[index];
     ht->table[index] = entry;
 
@@ -230,8 +157,7 @@ int dictAdd(dict *ht, void *key, void *val)
  * Return 1 if the key was added from scratch, 0 if there was already an
  * element with such key and dictReplace() just performed a value update
  * operation. */
-int dictReplace(dict *ht, void *key, void *val)
-{
+static int dictReplace(dict *ht, void *key, void *val) {
     dictEntry *entry, auxentry;
 
     /* Try to add the element. If the key
@@ -253,49 +179,38 @@ int dictReplace(dict *ht, void *key, void *val)
 }
 
 /* Search and remove an element */
-static int dictGenericDelete(dict *ht, const void *key, int nofree)
-{
+static int dictDelete(dict *ht, const void *key) {
     unsigned int h;
-    dictEntry *he, *prevHe;
+    dictEntry *de, *prevde;
 
     if (ht->size == 0)
         return DICT_ERR;
     h = dictHashKey(ht, key) & ht->sizemask;
-    he = ht->table[h];
+    de = ht->table[h];
 
-    prevHe = NULL;
-    while(he) {
-        if (dictCompareHashKeys(ht, key, he->key)) {
+    prevde = NULL;
+    while(de) {
+        if (dictCompareHashKeys(ht,key,de->key)) {
             /* Unlink the element from the list */
-            if (prevHe)
-                prevHe->next = he->next;
+            if (prevde)
+                prevde->next = de->next;
             else
-                ht->table[h] = he->next;
-            if (!nofree) {
-                dictFreeEntryKey(ht, he);
-                dictFreeEntryVal(ht, he);
-            }
-            _dictFree(he);
+                ht->table[h] = de->next;
+
+            dictFreeEntryKey(ht,de);
+            dictFreeEntryVal(ht,de);
+            free(de);
             ht->used--;
             return DICT_OK;
         }
-        prevHe = he;
-        he = he->next;
+        prevde = de;
+        de = de->next;
     }
     return DICT_ERR; /* not found */
 }
 
-int dictDelete(dict *ht, const void *key) {
-    return dictGenericDelete(ht,key,0);
-}
-
-int dictDeleteNoFree(dict *ht, const void *key) {
-    return dictGenericDelete(ht,key,1);
-}
-
 /* Destroy an entire hash table */
-int _dictClear(dict *ht)
-{
+static int _dictClear(dict *ht) {
     unsigned long i;
 
     /* Free all the elements */
@@ -307,27 +222,25 @@ int _dictClear(dict *ht)
             nextHe = he->next;
             dictFreeEntryKey(ht, he);
             dictFreeEntryVal(ht, he);
-            _dictFree(he);
+            free(he);
             ht->used--;
             he = nextHe;
         }
     }
     /* Free the table and the allocated cache structure */
-    _dictFree(ht->table);
+    free(ht->table);
     /* Re-initialize the table */
     _dictReset(ht);
     return DICT_OK; /* never fails */
 }
 
 /* Clear & Release the hash table */
-void dictRelease(dict *ht)
-{
+static void dictRelease(dict *ht) {
     _dictClear(ht);
-    _dictFree(ht);
+    free(ht);
 }
 
-dictEntry *dictFind(dict *ht, const void *key)
-{
+static dictEntry *dictFind(dict *ht, const void *key) {
     dictEntry *he;
     unsigned int h;
 
@@ -342,9 +255,8 @@ dictEntry *dictFind(dict *ht, const void *key)
     return NULL;
 }
 
-dictIterator *dictGetIterator(dict *ht)
-{
-    dictIterator *iter = _dictAlloc(sizeof(*iter));
+static dictIterator *dictGetIterator(dict *ht) {
+    dictIterator *iter = malloc(sizeof(*iter));
 
     iter->ht = ht;
     iter->index = -1;
@@ -353,8 +265,7 @@ dictIterator *dictGetIterator(dict *ht)
     return iter;
 }
 
-dictEntry *dictNext(dictIterator *iter)
-{
+static dictEntry *dictNext(dictIterator *iter) {
     while (1) {
         if (iter->entry == NULL) {
             iter->index++;
@@ -374,45 +285,14 @@ dictEntry *dictNext(dictIterator *iter)
     return NULL;
 }
 
-void dictReleaseIterator(dictIterator *iter)
-{
-    _dictFree(iter);
-}
-
-/* Return a random entry from the hash table. Useful to
- * implement randomized algorithms */
-dictEntry *dictGetRandomKey(dict *ht)
-{
-    dictEntry *he;
-    unsigned int h;
-    int listlen, listele;
-
-    if (ht->used == 0) return NULL;
-    do {
-        h = random() & ht->sizemask;
-        he = ht->table[h];
-    } while(he == NULL);
-
-    /* Now we found a non empty bucket, but it is a linked
-     * list and we need to get a random element from the list.
-     * The only sane way to do so is to count the element and
-     * select a random index. */
-    listlen = 0;
-    while(he) {
-        he = he->next;
-        listlen++;
-    }
-    listele = random() % listlen;
-    he = ht->table[h];
-    while(listele--) he = he->next;
-    return he;
+static void dictReleaseIterator(dictIterator *iter) {
+    free(iter);
 }
 
 /* ------------------------- private functions ------------------------------ */
 
 /* Expand the hash table if needed */
-static int _dictExpandIfNeeded(dict *ht)
-{
+static int _dictExpandIfNeeded(dict *ht) {
     /* If the hash table is empty expand it to the intial size,
      * if the table is "full" dobule its size. */
     if (ht->size == 0)
@@ -423,8 +303,7 @@ static int _dictExpandIfNeeded(dict *ht)
 }
 
 /* Our hash table capability is a power of two */
-static unsigned long _dictNextPower(unsigned long size)
-{
+static unsigned long _dictNextPower(unsigned long size) {
     unsigned long i = DICT_HT_INITIAL_SIZE;
 
     if (size >= LONG_MAX) return LONG_MAX;
@@ -438,8 +317,7 @@ static unsigned long _dictNextPower(unsigned long size)
 /* Returns the index of a free slot that can be populated with
  * an hash entry for the given 'key'.
  * If the key already exists, -1 is returned. */
-static int _dictKeyIndex(dict *ht, const void *key)
-{
+static int _dictKeyIndex(dict *ht, const void *key) {
     unsigned int h;
     dictEntry *he;
 
@@ -458,133 +336,3 @@ static int _dictKeyIndex(dict *ht, const void *key)
     return h;
 }
 
-void dictEmpty(dict *ht) {
-    _dictClear(ht);
-}
-
-#define DICT_STATS_VECTLEN 50
-void dictPrintStats(dict *ht) {
-    unsigned long i, slots = 0, chainlen, maxchainlen = 0;
-    unsigned long totchainlen = 0;
-    unsigned long clvector[DICT_STATS_VECTLEN];
-
-    if (ht->used == 0) {
-        printf("No stats available for empty dictionaries\n");
-        return;
-    }
-
-    for (i = 0; i < DICT_STATS_VECTLEN; i++) clvector[i] = 0;
-    for (i = 0; i < ht->size; i++) {
-        dictEntry *he;
-
-        if (ht->table[i] == NULL) {
-            clvector[0]++;
-            continue;
-        }
-        slots++;
-        /* For each hash entry on this slot... */
-        chainlen = 0;
-        he = ht->table[i];
-        while(he) {
-            chainlen++;
-            he = he->next;
-        }
-        clvector[(chainlen < DICT_STATS_VECTLEN) ? chainlen : (DICT_STATS_VECTLEN-1)]++;
-        if (chainlen > maxchainlen) maxchainlen = chainlen;
-        totchainlen += chainlen;
-    }
-    printf("Hash table stats:\n");
-    printf(" table size: %ld\n", ht->size);
-    printf(" number of elements: %ld\n", ht->used);
-    printf(" different slots: %ld\n", slots);
-    printf(" max chain length: %ld\n", maxchainlen);
-    printf(" avg chain length (counted): %.02f\n", (float)totchainlen/slots);
-    printf(" avg chain length (computed): %.02f\n", (float)ht->used/slots);
-    printf(" Chain length distribution:\n");
-    for (i = 0; i < DICT_STATS_VECTLEN-1; i++) {
-        if (clvector[i] == 0) continue;
-        printf("   %s%ld: %ld (%.02f%%)\n",(i == DICT_STATS_VECTLEN-1)?">= ":"", i, clvector[i], ((float)clvector[i]/ht->size)*100);
-    }
-}
-
-/* ----------------------- StringCopy Hash Table Type ------------------------*/
-
-static unsigned int _dictStringCopyHTHashFunction(const void *key)
-{
-    return dictGenHashFunction(key, strlen(key));
-}
-
-static void *_dictStringCopyHTKeyDup(void *privdata, const void *key)
-{
-    int len = strlen(key);
-    char *copy = _dictAlloc(len+1);
-    DICT_NOTUSED(privdata);
-
-    memcpy(copy, key, len);
-    copy[len] = '\0';
-    return copy;
-}
-
-static void *_dictStringKeyValCopyHTValDup(void *privdata, const void *val)
-{
-    int len = strlen(val);
-    char *copy = _dictAlloc(len+1);
-    DICT_NOTUSED(privdata);
-
-    memcpy(copy, val, len);
-    copy[len] = '\0';
-    return copy;
-}
-
-static int _dictStringCopyHTKeyCompare(void *privdata, const void *key1,
-        const void *key2)
-{
-    DICT_NOTUSED(privdata);
-
-    return strcmp(key1, key2) == 0;
-}
-
-static void _dictStringCopyHTKeyDestructor(void *privdata, void *key)
-{
-    DICT_NOTUSED(privdata);
-
-    _dictFree((void*)key); /* ATTENTION: const cast */
-}
-
-static void _dictStringKeyValCopyHTValDestructor(void *privdata, void *val)
-{
-    DICT_NOTUSED(privdata);
-
-    _dictFree((void*)val); /* ATTENTION: const cast */
-}
-
-dictType dictTypeHeapStringCopyKey = {
-    _dictStringCopyHTHashFunction,        /* hash function */
-    _dictStringCopyHTKeyDup,              /* key dup */
-    NULL,                               /* val dup */
-    _dictStringCopyHTKeyCompare,          /* key compare */
-    _dictStringCopyHTKeyDestructor,       /* key destructor */
-    NULL                                /* val destructor */
-};
-
-/* This is like StringCopy but does not auto-duplicate the key.
- * It's used for intepreter's shared strings. */
-dictType dictTypeHeapStrings = {
-    _dictStringCopyHTHashFunction,        /* hash function */
-    NULL,                               /* key dup */
-    NULL,                               /* val dup */
-    _dictStringCopyHTKeyCompare,          /* key compare */
-    _dictStringCopyHTKeyDestructor,       /* key destructor */
-    NULL                                /* val destructor */
-};
-
-/* This is like StringCopy but also automatically handle dynamic
- * allocated C strings as values. */
-dictType dictTypeHeapStringCopyKeyValue = {
-    _dictStringCopyHTHashFunction,        /* hash function */
-    _dictStringCopyHTKeyDup,              /* key dup */
-    _dictStringKeyValCopyHTValDup,        /* val dup */
-    _dictStringCopyHTKeyCompare,          /* key compare */
-    _dictStringCopyHTKeyDestructor,       /* key destructor */
-    _dictStringKeyValCopyHTValDestructor, /* val destructor */
-};

deps/hiredis/dict.h

@@ -1,4 +1,4 @@
-/* Hash Tables Implementation.
+/* Hash table implementation.
  *
  * This file implements in memory hash tables with insert/del/replace/find/
  * get-random-element operations. Hash tables will auto resize if needed
@@ -111,26 +111,16 @@ typedef struct dictIterator {
 #define dictSize(ht) ((ht)->used)
 
 /* API */
-dict *dictCreate(dictType *type, void *privDataPtr);
-int dictExpand(dict *ht, unsigned long size);
-int dictAdd(dict *ht, void *key, void *val);
-int dictReplace(dict *ht, void *key, void *val);
-int dictDelete(dict *ht, const void *key);
-int dictDeleteNoFree(dict *ht, const void *key);
-void dictRelease(dict *ht);
-dictEntry * dictFind(dict *ht, const void *key);
-int dictResize(dict *ht);
-dictIterator *dictGetIterator(dict *ht);
-dictEntry *dictNext(dictIterator *iter);
-void dictReleaseIterator(dictIterator *iter);
-dictEntry *dictGetRandomKey(dict *ht);
-void dictPrintStats(dict *ht);
-unsigned int dictGenHashFunction(const unsigned char *buf, int len);
-void dictEmpty(dict *ht);
-
-/* Hash table types */
-extern dictType dictTypeHeapStringCopyKey;
-extern dictType dictTypeHeapStrings;
-extern dictType dictTypeHeapStringCopyKeyValue;
+static unsigned int dictGenHashFunction(const unsigned char *buf, int len);
+static dict *dictCreate(dictType *type, void *privDataPtr);
+static int dictExpand(dict *ht, unsigned long size);
+static int dictAdd(dict *ht, void *key, void *val);
+static int dictReplace(dict *ht, void *key, void *val);
+static int dictDelete(dict *ht, const void *key);
+static void dictRelease(dict *ht);
+static dictEntry * dictFind(dict *ht, const void *key);
+static dictIterator *dictGetIterator(dict *ht);
+static dictEntry *dictNext(dictIterator *iter);
+static void dictReleaseIterator(dictIterator *iter);
 
 #endif /* __DICT_H */

deps/hiredis/example-ae.c

@@ -0,0 +1,53 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <signal.h>
+#include "hiredis.h"
+#include "async.h"
+#include "adapters/ae.h"
+
+/* Put event loop in the global scope, so it can be explicitly stopped */
+static aeEventLoop *loop;
+
+void getCallback(redisAsyncContext *c, void *r, void *privdata) {
+    redisReply *reply = r;
+    if (reply == NULL) return;
+    printf("argv[%s]: %s\n", (char*)privdata, reply->str);
+
+    /* Disconnect after receiving the reply to GET */
+    redisAsyncDisconnect(c);
+}
+
+void connectCallback(const redisAsyncContext *c) {
+    ((void)c);
+    printf("connected...\n");
+}
+
+void disconnectCallback(const redisAsyncContext *c, int status) {
+    if (status != REDIS_OK) {
+        printf("Error: %s\n", c->errstr);
+    }
+    printf("disconnected...\n");
+    aeStop(loop);
+}
+
+int main (int argc, char **argv) {
+    signal(SIGPIPE, SIG_IGN);
+
+    redisAsyncContext *c = redisAsyncConnect("127.0.0.1", 6379);
+    if (c->err) {
+        /* Let *c leak for now... */
+        printf("Error: %s\n", c->errstr);
+        return 1;
+    }
+
+    loop = aeCreateEventLoop();
+    redisAeAttach(loop, c);
+    redisAsyncSetConnectCallback(c,connectCallback);
+    redisAsyncSetDisconnectCallback(c,disconnectCallback);
+    redisAsyncCommand(c, NULL, NULL, "SET key %b", argv[argc-1], strlen(argv[argc-1]));
+    redisAsyncCommand(c, getCallback, (char*)"end-1", "GET key");
+    aeMain(loop);
+    return 0;
+}
+

deps/hiredis/example-libev.c

@@ -0,0 +1,47 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <signal.h>
+#include "hiredis.h"
+#include "async.h"
+#include "adapters/libev.h"
+
+void getCallback(redisAsyncContext *c, void *r, void *privdata) {
+    redisReply *reply = r;
+    if (reply == NULL) return;
+    printf("argv[%s]: %s\n", (char*)privdata, reply->str);
+
+    /* Disconnect after receiving the reply to GET */
+    redisAsyncDisconnect(c);
+}
+
+void connectCallback(const redisAsyncContext *c) {
+    ((void)c);
+    printf("connected...\n");
+}
+
+void disconnectCallback(const redisAsyncContext *c, int status) {
+    if (status != REDIS_OK) {
+        printf("Error: %s\n", c->errstr);
+    }
+    printf("disconnected...\n");
+}
+
+int main (int argc, char **argv) {
+    signal(SIGPIPE, SIG_IGN);
+
+    redisAsyncContext *c = redisAsyncConnect("127.0.0.1", 6379);
+    if (c->err) {
+        /* Let *c leak for now... */
+        printf("Error: %s\n", c->errstr);
+        return 1;
+    }
+
+    redisLibevAttach(EV_DEFAULT_ c);
+    redisAsyncSetConnectCallback(c,connectCallback);
+    redisAsyncSetDisconnectCallback(c,disconnectCallback);
+    redisAsyncCommand(c, NULL, NULL, "SET key %b", argv[argc-1], strlen(argv[argc-1]));
+    redisAsyncCommand(c, getCallback, (char*)"end-1", "GET key");
+    ev_loop(EV_DEFAULT_ 0);
+    return 0;
+}

deps/hiredis/example-libevent.c

@@ -0,0 +1,48 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <signal.h>
+#include "hiredis.h"
+#include "async.h"
+#include "adapters/libevent.h"
+
+void getCallback(redisAsyncContext *c, void *r, void *privdata) {
+    redisReply *reply = r;
+    if (reply == NULL) return;
+    printf("argv[%s]: %s\n", (char*)privdata, reply->str);
+
+    /* Disconnect after receiving the reply to GET */
+    redisAsyncDisconnect(c);
+}
+
+void connectCallback(const redisAsyncContext *c) {
+    ((void)c);
+    printf("connected...\n");
+}
+
+void disconnectCallback(const redisAsyncContext *c, int status) {
+    if (status != REDIS_OK) {
+        printf("Error: %s\n", c->errstr);
+    }
+    printf("disconnected...\n");
+}
+
+int main (int argc, char **argv) {
+    signal(SIGPIPE, SIG_IGN);
+    struct event_base *base = event_base_new();
+
+    redisAsyncContext *c = redisAsyncConnect("127.0.0.1", 6379);
+    if (c->err) {
+        /* Let *c leak for now... */
+        printf("Error: %s\n", c->errstr);
+        return 1;
+    }
+
+    redisLibeventAttach(c,base);
+    redisAsyncSetConnectCallback(c,connectCallback);
+    redisAsyncSetDisconnectCallback(c,disconnectCallback);
+    redisAsyncCommand(c, NULL, NULL, "SET key %b", argv[argc-1], strlen(argv[argc-1]));
+    redisAsyncCommand(c, getCallback, (char*)"end-1", "GET key");
+    event_base_dispatch(base);
+    return 0;
+}

deps/hiredis/example.c

@@ -0,0 +1,68 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "hiredis.h"
+
+int main(void) {
+    unsigned int j;
+    redisContext *c;
+    redisReply *reply;
+
+    struct timeval timeout = { 1, 500000 }; // 1.5 seconds
+    c = redisConnectWithTimeout((char*)"127.0.0.2", 6379, timeout);
+    if (c->err) {
+        printf("Connection error: %s\n", c->errstr);
+        exit(1);
+    }
+
+    /* PING server */
+    reply = redisCommand(c,"PING");
+    printf("PING: %s\n", reply->str);
+    freeReplyObject(reply);
+
+    /* Set a key */
+    reply = redisCommand(c,"SET %s %s", "foo", "hello world");
+    printf("SET: %s\n", reply->str);
+    freeReplyObject(reply);
+
+    /* Set a key using binary safe API */
+    reply = redisCommand(c,"SET %b %b", "bar", 3, "hello", 5);
+    printf("SET (binary API): %s\n", reply->str);
+    freeReplyObject(reply);
+
+    /* Try a GET and two INCR */
+    reply = redisCommand(c,"GET foo");
+    printf("GET foo: %s\n", reply->str);
+    freeReplyObject(reply);
+
+    reply = redisCommand(c,"INCR counter");
+    printf("INCR counter: %lld\n", reply->integer);
+    freeReplyObject(reply);
+    /* again ... */
+    reply = redisCommand(c,"INCR counter");
+    printf("INCR counter: %lld\n", reply->integer);
+    freeReplyObject(reply);
+
+    /* Create a list of numbers, from 0 to 9 */
+    reply = redisCommand(c,"DEL mylist");
+    freeReplyObject(reply);
+    for (j = 0; j < 10; j++) {
+        char buf[64];
+
+        snprintf(buf,64,"%d",j);
+        reply = redisCommand(c,"LPUSH mylist element-%s", buf);
+        freeReplyObject(reply);
+    }
+
+    /* Let's check what we have inside the list */
+    reply = redisCommand(c,"LRANGE mylist 0 -1");
+    if (reply->type == REDIS_REPLY_ARRAY) {
+        for (j = 0; j < reply->elements; j++) {
+            printf("%u) %s\n", j, reply->element[j]->str);
+        }
+    }
+    freeReplyObject(reply);
+
+    return 0;
+}

deps/hiredis/fmacros.h

@@ -0,0 +1,14 @@
+#ifndef __HIREDIS_FMACRO_H
+#define __HIREDIS_FMACRO_H
+
+#ifndef _BSD_SOURCE
+#define _BSD_SOURCE
+#endif
+
+#ifdef __linux__
+#define _XOPEN_SOURCE 700
+#else
+#define _XOPEN_SOURCE
+#endif
+
+#endif

deps/hiredis/hiredis.h

@@ -0,0 +1,184 @@
+/*
+ * Copyright (c) 2009-2010, Salvatore Sanfilippo <antirez at gmail dot com>
+ * Copyright (c) 2010, Pieter Noordhuis <pcnoordhuis at gmail dot com>
+ *
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ *   * Redistributions of source code must retain the above copyright notice,
+ *     this list of conditions and the following disclaimer.
+ *   * Redistributions in binary form must reproduce the above copyright
+ *     notice, this list of conditions and the following disclaimer in the
+ *     documentation and/or other materials provided with the distribution.
+ *   * Neither the name of Redis nor the names of its contributors may be used
+ *     to endorse or promote products derived from this software without
+ *     specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#ifndef __HIREDIS_H
+#define __HIREDIS_H
+#include <stdio.h> /* for size_t */
+#include <stdarg.h> /* for va_list */
+#include <sys/time.h> /* for struct timeval */
+
+#define HIREDIS_MAJOR 0
+#define HIREDIS_MINOR 9
+#define HIREDIS_PATCH 2
+
+#define REDIS_ERR -1
+#define REDIS_OK 0
+
+/* When an error occurs, the err flag in a context is set to hold the type of
+ * error that occured. REDIS_ERR_IO means there was an I/O error and you
+ * should use the "errno" variable to find out what is wrong.
+ * For other values, the "errstr" field will hold a description. */
+#define REDIS_ERR_IO 1 /* error in read or write */
+#define REDIS_ERR_EOF 3 /* eof */
+#define REDIS_ERR_PROTOCOL 4 /* protocol error */
+#define REDIS_ERR_OTHER 2 /* something else */
+
+/* Connection type can be blocking or non-blocking and is set in the
+ * least significant bit of the flags field in redisContext. */
+#define REDIS_BLOCK 0x1
+
+/* Connection may be disconnected before being free'd. The second bit
+ * in the flags field is set when the context is connected. */
+#define REDIS_CONNECTED 0x2
+
+/* The async API might try to disconnect cleanly and flush the output
+ * buffer and read all subsequent replies before disconnecting.
+ * This flag means no new commands can come in and the connection
+ * should be terminated once all replies have been read. */
+#define REDIS_DISCONNECTING 0x4
+
+/* Flag specific to the async API which means that the context should be clean
+ * up as soon as possible. */
+#define REDIS_FREEING 0x8
+
+/* Flag that is set when an async callback is executed. */
+#define REDIS_IN_CALLBACK 0x10
+
+/* Flag that is set when the async context has one or more subscriptions. */
+#define REDIS_SUBSCRIBED 0x20
+
+#define REDIS_REPLY_STRING 1
+#define REDIS_REPLY_ARRAY 2
+#define REDIS_REPLY_INTEGER 3
+#define REDIS_REPLY_NIL 4
+#define REDIS_REPLY_STATUS 5
+#define REDIS_REPLY_ERROR 6
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/* This is the reply object returned by redisCommand() */
+typedef struct redisReply {
+    int type; /* REDIS_REPLY_* */
+    long long integer; /* The integer when type is REDIS_REPLY_INTEGER */
+    int len; /* Length of string */
+    char *str; /* Used for both REDIS_REPLY_ERROR and REDIS_REPLY_STRING */
+    size_t elements; /* number of elements, for REDIS_REPLY_ARRAY */
+    struct redisReply **element; /* elements vector for REDIS_REPLY_ARRAY */
+} redisReply;
+
+typedef struct redisReadTask {
+    int type;
+    int elements; /* number of elements in multibulk container */
+    int idx; /* index in parent (array) object */
+    void *obj; /* holds user-generated value for a read task */
+    struct redisReadTask *parent; /* parent task */
+    void *privdata; /* user-settable arbitrary field */
+} redisReadTask;
+
+typedef struct redisReplyObjectFunctions {
+    void *(*createString)(const redisReadTask*, char*, size_t);
+    void *(*createArray)(const redisReadTask*, int);
+    void *(*createInteger)(const redisReadTask*, long long);
+    void *(*createNil)(const redisReadTask*);
+    void (*freeObject)(void*);
+} redisReplyObjectFunctions;
+
+struct redisContext; /* need forward declaration of redisContext */
+
+/* Context for a connection to Redis */
+typedef struct redisContext {
+    int fd;
+    int flags;
+    char *obuf; /* Write buffer */
+    int err; /* Error flags, 0 when there is no error */
+    char *errstr; /* String representation of error when applicable */
+
+    /* Function set for reply buildup and reply reader */
+    redisReplyObjectFunctions *fn;
+    void *reader;
+} redisContext;
+
+void freeReplyObject(void *reply);
+void *redisReplyReaderCreate(void);
+int redisReplyReaderSetReplyObjectFunctions(void *reader, redisReplyObjectFunctions *fn);
+int redisReplyReaderSetPrivdata(void *reader, void *privdata);
+void *redisReplyReaderGetObject(void *reader);
+char *redisReplyReaderGetError(void *reader);
+void redisReplyReaderFree(void *ptr);
+void redisReplyReaderFeed(void *reader, const char *buf, size_t len);
+int redisReplyReaderGetReply(void *reader, void **reply);
+
+/* Functions to format a command according to the protocol. */
+int redisvFormatCommand(char **target, const char *format, va_list ap);
+int redisFormatCommand(char **target, const char *format, ...);
+int redisFormatCommandArgv(char **target, int argc, const char **argv, const size_t *argvlen);
+
+redisContext *redisConnect(const char *ip, int port);
+redisContext *redisConnectWithTimeout(const char *ip, int port, struct timeval tv);
+redisContext *redisConnectNonBlock(const char *ip, int port);
+redisContext *redisConnectUnix(const char *path);
+redisContext *redisConnectUnixWithTimeout(const char *path, struct timeval tv);
+redisContext *redisConnectUnixNonBlock(const char *path);
+int redisSetTimeout(redisContext *c, struct timeval tv);
+int redisSetReplyObjectFunctions(redisContext *c, redisReplyObjectFunctions *fn);
+void redisFree(redisContext *c);
+int redisBufferRead(redisContext *c);
+int redisBufferWrite(redisContext *c, int *done);
+
+/* In a blocking context, this function first checks if there are unconsumed
+ * replies to return and returns one if so. Otherwise, it flushes the output
+ * buffer to the socket and reads until it has a reply. In a non-blocking
+ * context, it will return unconsumed replies until there are no more. */
+int redisGetReply(redisContext *c, void **reply);
+int redisGetReplyFromReader(redisContext *c, void **reply);
+
+/* Write a command to the output buffer. Use these functions in blocking mode
+ * to get a pipeline of commands. */
+void redisvAppendCommand(redisContext *c, const char *format, va_list ap);
+void redisAppendCommand(redisContext *c, const char *format, ...);
+void redisAppendCommandArgv(redisContext *c, int argc, const char **argv, const size_t *argvlen);
+
+/* Issue a command to Redis. In a blocking context, it is identical to calling
+ * redisAppendCommand, followed by redisGetReply. The function will return
+ * NULL if there was an error in performing the request, otherwise it will
+ * return the reply. In a non-blocking context, it is identical to calling
+ * only redisAppendCommand and will always return NULL. */
+void *redisvCommand(redisContext *c, const char *format, va_list ap);
+void *redisCommand(redisContext *c, const char *format, ...);
+void *redisCommandArgv(redisContext *c, int argc, const char **argv, const size_t *argvlen);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif

deps/hiredis/net.c

@@ -0,0 +1,252 @@
+/* Extracted from anet.c to work properly with Hiredis error reporting.
+ *
+ * Copyright (c) 2006-2010, Salvatore Sanfilippo <antirez at gmail dot com>
+ * Copyright (c) 2010, Pieter Noordhuis <pcnoordhuis at gmail dot com>
+ *
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ *   * Redistributions of source code must retain the above copyright notice,
+ *     this list of conditions and the following disclaimer.
+ *   * Redistributions in binary form must reproduce the above copyright
+ *     notice, this list of conditions and the following disclaimer in the
+ *     documentation and/or other materials provided with the distribution.
+ *   * Neither the name of Redis nor the names of its contributors may be used
+ *     to endorse or promote products derived from this software without
+ *     specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "fmacros.h"
+#include <sys/types.h>
+#include <sys/socket.h>
+#include <sys/select.h>
+#include <sys/un.h>
+#include <netinet/in.h>
+#include <netinet/tcp.h>
+#include <arpa/inet.h>
+#include <unistd.h>
+#include <fcntl.h>
+#include <string.h>
+#include <netdb.h>
+#include <errno.h>
+#include <stdarg.h>
+#include <stdio.h>
+
+#include "net.h"
+#include "sds.h"
+
+/* Forward declaration */
+void __redisSetError(redisContext *c, int type, sds err);
+
+static int redisCreateSocket(redisContext *c, int type) {
+    int s, on = 1;
+    if ((s = socket(type, SOCK_STREAM, 0)) == -1) {
+        __redisSetError(c,REDIS_ERR_IO,NULL);
+        return REDIS_ERR;
+    }
+    if (type == AF_INET) {
+        if (setsockopt(s, SOL_SOCKET, SO_REUSEADDR, &on, sizeof(on)) == -1) {
+            __redisSetError(c,REDIS_ERR_IO,NULL);
+            close(s);
+            return REDIS_ERR;
+        }
+    }
+    return s;
+}
+
+static int redisSetBlocking(redisContext *c, int fd, int blocking) {
+    int flags;
+
+    /* Set the socket nonblocking.
+     * Note that fcntl(2) for F_GETFL and F_SETFL can't be
+     * interrupted by a signal. */
+    if ((flags = fcntl(fd, F_GETFL)) == -1) {
+        __redisSetError(c,REDIS_ERR_IO,
+            sdscatprintf(sdsempty(), "fcntl(F_GETFL): %s", strerror(errno)));
+        close(fd);
+        return REDIS_ERR;
+    }
+
+    if (blocking)
+        flags &= ~O_NONBLOCK;
+    else
+        flags |= O_NONBLOCK;
+
+    if (fcntl(fd, F_SETFL, flags) == -1) {
+        __redisSetError(c,REDIS_ERR_IO,
+            sdscatprintf(sdsempty(), "fcntl(F_SETFL): %s", strerror(errno)));
+        close(fd);
+        return REDIS_ERR;
+    }
+    return REDIS_OK;
+}
+
+static int redisSetTcpNoDelay(redisContext *c, int fd) {
+    int yes = 1;
+    if (setsockopt(fd, IPPROTO_TCP, TCP_NODELAY, &yes, sizeof(yes)) == -1) {
+        __redisSetError(c,REDIS_ERR_IO,
+            sdscatprintf(sdsempty(), "setsockopt(TCP_NODELAY): %s", strerror(errno)));
+        close(fd);
+        return REDIS_ERR;
+    }
+    return REDIS_OK;
+}
+
+static int redisContextWaitReady(redisContext *c, int fd, const struct timeval *timeout) {
+    struct timeval to;
+    struct timeval *toptr = NULL;
+    fd_set wfd;
+    int err;
+    socklen_t errlen;
+
+    /* Only use timeout when not NULL. */
+    if (timeout != NULL) {
+        to = *timeout;
+        toptr = &to;
+    }
+
+    if (errno == EINPROGRESS) {
+        FD_ZERO(&wfd);
+        FD_SET(fd, &wfd);
+
+        if (select(FD_SETSIZE, NULL, &wfd, NULL, toptr) == -1) {
+            __redisSetError(c,REDIS_ERR_IO,
+                sdscatprintf(sdsempty(), "select(2): %s", strerror(errno)));
+            close(fd);
+            return REDIS_ERR;
+        }
+
+        if (!FD_ISSET(fd, &wfd)) {
+            errno = ETIMEDOUT;
+            __redisSetError(c,REDIS_ERR_IO,NULL);
+            close(fd);
+            return REDIS_ERR;
+        }
+
+        err = 0;
+        errlen = sizeof(err);
+        if (getsockopt(fd, SOL_SOCKET, SO_ERROR, &err, &errlen) == -1) {
+            __redisSetError(c,REDIS_ERR_IO,
+                sdscatprintf(sdsempty(), "getsockopt(SO_ERROR): %s", strerror(errno)));
+            close(fd);
+            return REDIS_ERR;
+        }
+
+        if (err) {
+            errno = err;
+            __redisSetError(c,REDIS_ERR_IO,NULL);
+            close(fd);
+            return REDIS_ERR;
+        }
+
+        return REDIS_OK;
+    }
+
+    __redisSetError(c,REDIS_ERR_IO,NULL);
+    close(fd);
+    return REDIS_ERR;
+}
+
+int redisContextSetTimeout(redisContext *c, struct timeval tv) {
+    if (setsockopt(c->fd,SOL_SOCKET,SO_RCVTIMEO,&tv,sizeof(tv)) == -1) {
+        __redisSetError(c,REDIS_ERR_IO,
+            sdscatprintf(sdsempty(), "setsockopt(SO_RCVTIMEO): %s", strerror(errno)));
+        return REDIS_ERR;
+    }
+    if (setsockopt(c->fd,SOL_SOCKET,SO_SNDTIMEO,&tv,sizeof(tv)) == -1) {
+        __redisSetError(c,REDIS_ERR_IO,
+            sdscatprintf(sdsempty(), "setsockopt(SO_SNDTIMEO): %s", strerror(errno)));
+        return REDIS_ERR;
+    }
+    return REDIS_OK;
+}
+
+int redisContextConnectTcp(redisContext *c, const char *addr, int port, struct timeval *timeout) {
+    int s;
+    int blocking = (c->flags & REDIS_BLOCK);
+    struct sockaddr_in sa;
+
+    if ((s = redisCreateSocket(c,AF_INET)) < 0)
+        return REDIS_ERR;
+    if (redisSetBlocking(c,s,0) != REDIS_OK)
+        return REDIS_ERR;
+
+    sa.sin_family = AF_INET;
+    sa.sin_port = htons(port);
+    if (inet_aton(addr, &sa.sin_addr) == 0) {
+        struct hostent *he;
+
+        he = gethostbyname(addr);
+        if (he == NULL) {
+            __redisSetError(c,REDIS_ERR_OTHER,
+                sdscatprintf(sdsempty(),"Can't resolve: %s",addr));
+            close(s);
+            return REDIS_ERR;
+        }
+        memcpy(&sa.sin_addr, he->h_addr, sizeof(struct in_addr));
+    }
+
+    if (connect(s, (struct sockaddr*)&sa, sizeof(sa)) == -1) {
+        if (errno == EINPROGRESS && !blocking) {
+            /* This is ok. */
+        } else {
+            if (redisContextWaitReady(c,s,timeout) != REDIS_OK)
+                return REDIS_ERR;
+        }
+    }
+
+    /* Reset socket to be blocking after connect(2). */
+    if (blocking && redisSetBlocking(c,s,1) != REDIS_OK)
+        return REDIS_ERR;
+
+    if (redisSetTcpNoDelay(c,s) != REDIS_OK)
+        return REDIS_ERR;
+
+    c->fd = s;
+    c->flags |= REDIS_CONNECTED;
+    return REDIS_OK;
+}
+
+int redisContextConnectUnix(redisContext *c, const char *path, struct timeval *timeout) {
+    int s;
+    int blocking = (c->flags & REDIS_BLOCK);
+    struct sockaddr_un sa;
+
+    if ((s = redisCreateSocket(c,AF_LOCAL)) < 0)
+        return REDIS_ERR;
+    if (redisSetBlocking(c,s,0) != REDIS_OK)
+        return REDIS_ERR;
+
+    sa.sun_family = AF_LOCAL;
+    strncpy(sa.sun_path,path,sizeof(sa.sun_path)-1);
+    if (connect(s, (struct sockaddr*)&sa, sizeof(sa)) == -1) {
+        if (errno == EINPROGRESS && !blocking) {
+            /* This is ok. */
+        } else {
+            if (redisContextWaitReady(c,s,timeout) != REDIS_OK)
+                return REDIS_ERR;
+        }
+    }
+
+    /* Reset socket to be blocking after connect(2). */
+    if (blocking && redisSetBlocking(c,s,1) != REDIS_OK)
+        return REDIS_ERR;
+
+    c->fd = s;
+    c->flags |= REDIS_CONNECTED;
+    return REDIS_OK;
+}

deps/hiredis/net.h

@@ -1,5 +1,8 @@
-/*
- * Copyright (c) 2009-2010, Salvatore Sanfilippo <antirez at gmail dot com>
+/* Extracted from anet.c to work properly with Hiredis error reporting.
+ *
+ * Copyright (c) 2006-2010, Salvatore Sanfilippo <antirez at gmail dot com>
+ * Copyright (c) 2010, Pieter Noordhuis <pcnoordhuis at gmail dot com>
+ *
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
@@ -27,49 +30,17 @@
  * POSSIBILITY OF SUCH DAMAGE.
  */
 
-#ifndef __REDIS_H__
-#define __REDIS_H__
+#ifndef __NET_H
+#define __NET_H
 
-enum
-{
-  REG_GS = 0,
-# define REG_GS		REG_GS
-  REG_FS,
-# define REG_FS		REG_FS
-  REG_ES,
-# define REG_ES		REG_ES
-  REG_DS,
-# define REG_DS		REG_DS
-  REG_EDI,
-# define REG_EDI	REG_EDI
-  REG_ESI,
-# define REG_ESI	REG_ESI
-  REG_EBP,
-# define REG_EBP	REG_EBP
-  REG_ESP,
-# define REG_ESP	REG_ESP
-  REG_EBX,
-# define REG_EBX	REG_EBX
-  REG_EDX,
-# define REG_EDX	REG_EDX
-  REG_ECX,
-# define REG_ECX	REG_ECX
-  REG_EAX,
-# define REG_EAX	REG_EAX
-  REG_TRAPNO,
-# define REG_TRAPNO	REG_TRAPNO
-  REG_ERR,
-# define REG_ERR	REG_ERR
-  REG_EIP,
-# define REG_EIP	REG_EIP
-  REG_CS,
-# define REG_CS		REG_CS
-  REG_EFL,
-# define REG_EFL	REG_EFL
-  REG_UESP,
-# define REG_UESP	REG_UESP
-  REG_SS
-# define REG_SS	REG_SS
-};
+#include "hiredis.h"
+
+#if defined(__sun)
+#define AF_LOCAL AF_UNIX
+#endif
+
+int redisContextSetTimeout(redisContext *c, struct timeval tv);
+int redisContextConnectTcp(redisContext *c, const char *addr, int port, struct timeval *timeout);
+int redisContextConnectUnix(redisContext *c, const char *path, struct timeval *timeout);
 
 #endif

deps/hiredis/sds.c

@@ -0,0 +1,594 @@
+/* SDSLib, A C dynamic strings library
+ *
+ * Copyright (c) 2006-2010, Salvatore Sanfilippo <antirez at gmail dot com>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ *   * Redistributions of source code must retain the above copyright notice,
+ *     this list of conditions and the following disclaimer.
+ *   * Redistributions in binary form must reproduce the above copyright
+ *     notice, this list of conditions and the following disclaimer in the
+ *     documentation and/or other materials provided with the distribution.
+ *   * Neither the name of Redis nor the names of its contributors may be used
+ *     to endorse or promote products derived from this software without
+ *     specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#define SDS_ABORT_ON_OOM
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <ctype.h>
+#include "sds.h"
+
+static void sdsOomAbort(void) {
+    fprintf(stderr,"SDS: Out Of Memory (SDS_ABORT_ON_OOM defined)\n");
+    abort();
+}
+
+sds sdsnewlen(const void *init, size_t initlen) {
+    struct sdshdr *sh;
+
+    sh = malloc(sizeof(struct sdshdr)+initlen+1);
+#ifdef SDS_ABORT_ON_OOM
+    if (sh == NULL) sdsOomAbort();
+#else
+    if (sh == NULL) return NULL;
+#endif
+    sh->len = initlen;
+    sh->free = 0;
+    if (initlen) {
+        if (init) memcpy(sh->buf, init, initlen);
+        else memset(sh->buf,0,initlen);
+    }
+    sh->buf[initlen] = '\0';
+    return (char*)sh->buf;
+}
+
+sds sdsempty(void) {
+    return sdsnewlen("",0);
+}
+
+sds sdsnew(const char *init) {
+    size_t initlen = (init == NULL) ? 0 : strlen(init);
+    return sdsnewlen(init, initlen);
+}
+
+sds sdsdup(const sds s) {
+    return sdsnewlen(s, sdslen(s));
+}
+
+void sdsfree(sds s) {
+    if (s == NULL) return;
+    free(s-sizeof(struct sdshdr));
+}
+
+void sdsupdatelen(sds s) {
+    struct sdshdr *sh = (void*) (s-(sizeof(struct sdshdr)));
+    int reallen = strlen(s);
+    sh->free += (sh->len-reallen);
+    sh->len = reallen;
+}
+
+static sds sdsMakeRoomFor(sds s, size_t addlen) {
+    struct sdshdr *sh, *newsh;
+    size_t free = sdsavail(s);
+    size_t len, newlen;
+
+    if (free >= addlen) return s;
+    len = sdslen(s);
+    sh = (void*) (s-(sizeof(struct sdshdr)));
+    newlen = (len+addlen)*2;
+    newsh = realloc(sh, sizeof(struct sdshdr)+newlen+1);
+#ifdef SDS_ABORT_ON_OOM
+    if (newsh == NULL) sdsOomAbort();
+#else
+    if (newsh == NULL) return NULL;
+#endif
+
+    newsh->free = newlen - len;
+    return newsh->buf;
+}
+
+/* Grow the sds to have the specified length. Bytes that were not part of
+ * the original length of the sds will be set to zero. */
+sds sdsgrowzero(sds s, size_t len) {
+    struct sdshdr *sh = (void*)(s-(sizeof(struct sdshdr)));
+    size_t totlen, curlen = sh->len;
+
+    if (len <= curlen) return s;
+    s = sdsMakeRoomFor(s,len-curlen);
+    if (s == NULL) return NULL;
+
+    /* Make sure added region doesn't contain garbage */
+    sh = (void*)(s-(sizeof(struct sdshdr)));
+    memset(s+curlen,0,(len-curlen+1)); /* also set trailing \0 byte */
+    totlen = sh->len+sh->free;
+    sh->len = len;
+    sh->free = totlen-sh->len;
+    return s;
+}
+
+sds sdscatlen(sds s, const void *t, size_t len) {
+    struct sdshdr *sh;
+    size_t curlen = sdslen(s);
+
+    s = sdsMakeRoomFor(s,len);
+    if (s == NULL) return NULL;
+    sh = (void*) (s-(sizeof(struct sdshdr)));
+    memcpy(s+curlen, t, len);
+    sh->len = curlen+len;
+    sh->free = sh->free-len;
+    s[curlen+len] = '\0';
+    return s;
+}
+
+sds sdscat(sds s, const char *t) {
+    return sdscatlen(s, t, strlen(t));
+}
+
+sds sdscpylen(sds s, char *t, size_t len) {
+    struct sdshdr *sh = (void*) (s-(sizeof(struct sdshdr)));
+    size_t totlen = sh->free+sh->len;
+
+    if (totlen < len) {
+        s = sdsMakeRoomFor(s,len-sh->len);
+        if (s == NULL) return NULL;
+        sh = (void*) (s-(sizeof(struct sdshdr)));
+        totlen = sh->free+sh->len;
+    }
+    memcpy(s, t, len);
+    s[len] = '\0';
+    sh->len = len;
+    sh->free = totlen-len;
+    return s;
+}
+
+sds sdscpy(sds s, char *t) {
+    return sdscpylen(s, t, strlen(t));
+}
+
+sds sdscatvprintf(sds s, const char *fmt, va_list ap) {
+    va_list cpy;
+    char *buf, *t;
+    size_t buflen = 16;
+
+    while(1) {
+        buf = malloc(buflen);
+#ifdef SDS_ABORT_ON_OOM
+        if (buf == NULL) sdsOomAbort();
+#else
+        if (buf == NULL) return NULL;
+#endif
+        buf[buflen-2] = '\0';
+        va_copy(cpy,ap);
+        vsnprintf(buf, buflen, fmt, cpy);
+        if (buf[buflen-2] != '\0') {
+            free(buf);
+            buflen *= 2;
+            continue;
+        }
+        break;
+    }
+    t = sdscat(s, buf);
+    free(buf);
+    return t;
+}
+
+sds sdscatprintf(sds s, const char *fmt, ...) {
+    va_list ap;
+    char *t;
+    va_start(ap, fmt);
+    t = sdscatvprintf(s,fmt,ap);
+    va_end(ap);
+    return t;
+}
+
+sds sdstrim(sds s, const char *cset) {
+    struct sdshdr *sh = (void*) (s-(sizeof(struct sdshdr)));
+    char *start, *end, *sp, *ep;
+    size_t len;
+
+    sp = start = s;
+    ep = end = s+sdslen(s)-1;
+    while(sp <= end && strchr(cset, *sp)) sp++;
+    while(ep > start && strchr(cset, *ep)) ep--;
+    len = (sp > ep) ? 0 : ((ep-sp)+1);
+    if (sh->buf != sp) memmove(sh->buf, sp, len);
+    sh->buf[len] = '\0';
+    sh->free = sh->free+(sh->len-len);
+    sh->len = len;
+    return s;
+}
+
+sds sdsrange(sds s, int start, int end) {
+    struct sdshdr *sh = (void*) (s-(sizeof(struct sdshdr)));
+    size_t newlen, len = sdslen(s);
+
+    if (len == 0) return s;
+    if (start < 0) {
+        start = len+start;
+        if (start < 0) start = 0;
+    }
+    if (end < 0) {
+        end = len+end;
+        if (end < 0) end = 0;
+    }
+    newlen = (start > end) ? 0 : (end-start)+1;
+    if (newlen != 0) {
+        if (start >= (signed)len) {
+            newlen = 0;
+        } else if (end >= (signed)len) {
+            end = len-1;
+            newlen = (start > end) ? 0 : (end-start)+1;
+        }
+    } else {
+        start = 0;
+    }
+    if (start && newlen) memmove(sh->buf, sh->buf+start, newlen);
+    sh->buf[newlen] = 0;
+    sh->free = sh->free+(sh->len-newlen);
+    sh->len = newlen;
+    return s;
+}
+
+void sdstolower(sds s) {
+    int len = sdslen(s), j;
+
+    for (j = 0; j < len; j++) s[j] = tolower(s[j]);
+}
+
+void sdstoupper(sds s) {
+    int len = sdslen(s), j;
+
+    for (j = 0; j < len; j++) s[j] = toupper(s[j]);
+}
+
+int sdscmp(sds s1, sds s2) {
+    size_t l1, l2, minlen;
+    int cmp;
+
+    l1 = sdslen(s1);
+    l2 = sdslen(s2);
+    minlen = (l1 < l2) ? l1 : l2;
+    cmp = memcmp(s1,s2,minlen);
+    if (cmp == 0) return l1-l2;
+    return cmp;
+}
+
+/* Split 's' with separator in 'sep'. An array
+ * of sds strings is returned. *count will be set
+ * by reference to the number of tokens returned.
+ *
+ * On out of memory, zero length string, zero length
+ * separator, NULL is returned.
+ *
+ * Note that 'sep' is able to split a string using
+ * a multi-character separator. For example
+ * sdssplit("foo_-_bar","_-_"); will return two
+ * elements "foo" and "bar".
+ *
+ * This version of the function is binary-safe but
+ * requires length arguments. sdssplit() is just the
+ * same function but for zero-terminated strings.
+ */
+sds *sdssplitlen(char *s, int len, char *sep, int seplen, int *count) {
+    int elements = 0, slots = 5, start = 0, j;
+
+    sds *tokens = malloc(sizeof(sds)*slots);
+#ifdef SDS_ABORT_ON_OOM
+    if (tokens == NULL) sdsOomAbort();
+#endif
+    if (seplen < 1 || len < 0 || tokens == NULL) return NULL;
+    if (len == 0) {
+        *count = 0;
+        return tokens;
+    }
+    for (j = 0; j < (len-(seplen-1)); j++) {
+        /* make sure there is room for the next element and the final one */
+        if (slots < elements+2) {
+            sds *newtokens;
+
+            slots *= 2;
+            newtokens = realloc(tokens,sizeof(sds)*slots);
+            if (newtokens == NULL) {
+#ifdef SDS_ABORT_ON_OOM
+                sdsOomAbort();
+#else
+                goto cleanup;
+#endif
+            }
+            tokens = newtokens;
+        }
+        /* search the separator */
+        if ((seplen == 1 && *(s+j) == sep[0]) || (memcmp(s+j,sep,seplen) == 0)) {
+            tokens[elements] = sdsnewlen(s+start,j-start);
+            if (tokens[elements] == NULL) {
+#ifdef SDS_ABORT_ON_OOM
+                sdsOomAbort();
+#else
+                goto cleanup;
+#endif
+            }
+            elements++;
+            start = j+seplen;
+            j = j+seplen-1; /* skip the separator */
+        }
+    }
+    /* Add the final element. We are sure there is room in the tokens array. */
+    tokens[elements] = sdsnewlen(s+start,len-start);
+    if (tokens[elements] == NULL) {
+#ifdef SDS_ABORT_ON_OOM
+                sdsOomAbort();
+#else
+                goto cleanup;
+#endif
+    }
+    elements++;
+    *count = elements;
+    return tokens;
+
+#ifndef SDS_ABORT_ON_OOM
+cleanup:
+    {
+        int i;
+        for (i = 0; i < elements; i++) sdsfree(tokens[i]);
+        free(tokens);
+        return NULL;
+    }
+#endif
+}
+
+void sdsfreesplitres(sds *tokens, int count) {
+    if (!tokens) return;
+    while(count--)
+        sdsfree(tokens[count]);
+    free(tokens);
+}
+
+sds sdsfromlonglong(long long value) {
+    char buf[32], *p;
+    unsigned long long v;
+
+    v = (value < 0) ? -value : value;
+    p = buf+31; /* point to the last character */
+    do {
+        *p-- = '0'+(v%10);
+        v /= 10;
+    } while(v);
+    if (value < 0) *p-- = '-';
+    p++;
+    return sdsnewlen(p,32-(p-buf));
+}
+
+sds sdscatrepr(sds s, char *p, size_t len) {
+    s = sdscatlen(s,"\"",1);
+    while(len--) {
+        switch(*p) {
+        case '\\':
+        case '"':
+            s = sdscatprintf(s,"\\%c",*p);
+            break;
+        case '\n': s = sdscatlen(s,"\\n",1); break;
+        case '\r': s = sdscatlen(s,"\\r",1); break;
+        case '\t': s = sdscatlen(s,"\\t",1); break;
+        case '\a': s = sdscatlen(s,"\\a",1); break;
+        case '\b': s = sdscatlen(s,"\\b",1); break;
+        default:
+            if (isprint(*p))
+                s = sdscatprintf(s,"%c",*p);
+            else
+                s = sdscatprintf(s,"\\x%02x",(unsigned char)*p);
+            break;
+        }
+        p++;
+    }
+    return sdscatlen(s,"\"",1);
+}
+
+/* Split a line into arguments, where every argument can be in the
+ * following programming-language REPL-alike form:
+ *
+ * foo bar "newline are supported\n" and "\xff\x00otherstuff"
+ *
+ * The number of arguments is stored into *argc, and an array
+ * of sds is returned. The caller should sdsfree() all the returned
+ * strings and finally free() the array itself.
+ *
+ * Note that sdscatrepr() is able to convert back a string into
+ * a quoted string in the same format sdssplitargs() is able to parse.
+ */
+sds *sdssplitargs(char *line, int *argc) {
+    char *p = line;
+    char *current = NULL;
+    char **vector = NULL;
+
+    *argc = 0;
+    while(1) {
+        /* skip blanks */
+        while(*p && isspace(*p)) p++;
+        if (*p) {
+            /* get a token */
+            int inq=0; /* set to 1 if we are in "quotes" */
+            int done=0;
+
+            if (current == NULL) current = sdsempty();
+            while(!done) {
+                if (inq) {
+                    if (*p == '\\' && *(p+1)) {
+                        char c;
+
+                        p++;
+                        switch(*p) {
+                        case 'n': c = '\n'; break;
+                        case 'r': c = '\r'; break;
+                        case 't': c = '\t'; break;
+                        case 'b': c = '\b'; break;
+                        case 'a': c = '\a'; break;
+                        default: c = *p; break;
+                        }
+                        current = sdscatlen(current,&c,1);
+                    } else if (*p == '"') {
+                        /* closing quote must be followed by a space */
+                        if (*(p+1) && !isspace(*(p+1))) goto err;
+                        done=1;
+                    } else if (!*p) {
+                        /* unterminated quotes */
+                        goto err;
+                    } else {
+                        current = sdscatlen(current,p,1);
+                    }
+                } else {
+                    switch(*p) {
+                    case ' ':
+                    case '\n':
+                    case '\r':
+                    case '\t':
+                    case '\0':
+                        done=1;
+                        break;
+                    case '"':
+                        inq=1;
+                        break;
+                    default:
+                        current = sdscatlen(current,p,1);
+                        break;
+                    }
+                }
+                if (*p) p++;
+            }
+            /* add the token to the vector */
+            vector = realloc(vector,((*argc)+1)*sizeof(char*));
+            vector[*argc] = current;
+            (*argc)++;
+            current = NULL;
+        } else {
+            return vector;
+        }
+    }
+
+err:
+    while((*argc)--)
+        sdsfree(vector[*argc]);
+    free(vector);
+    if (current) sdsfree(current);
+    return NULL;
+}
+
+#ifdef SDS_TEST_MAIN
+#include <stdio.h>
+
+int __failed_tests = 0;
+int __test_num = 0;
+#define test_cond(descr,_c) do { \
+    __test_num++; printf("%d - %s: ", __test_num, descr); \
+    if(_c) printf("PASSED\n"); else {printf("FAILED\n"); __failed_tests++;} \
+} while(0);
+#define test_report() do { \
+    printf("%d tests, %d passed, %d failed\n", __test_num, \
+                    __test_num-__failed_tests, __failed_tests); \
+    if (__failed_tests) { \
+        printf("=== WARNING === We have failed tests here...\n"); \
+    } \
+} while(0);
+
+int main(void) {
+    {
+        sds x = sdsnew("foo"), y;
+
+        test_cond("Create a string and obtain the length",
+            sdslen(x) == 3 && memcmp(x,"foo\0",4) == 0)
+
+        sdsfree(x);
+        x = sdsnewlen("foo",2);
+        test_cond("Create a string with specified length",
+            sdslen(x) == 2 && memcmp(x,"fo\0",3) == 0)
+
+        x = sdscat(x,"bar");
+        test_cond("Strings concatenation",
+            sdslen(x) == 5 && memcmp(x,"fobar\0",6) == 0);
+
+        x = sdscpy(x,"a");
+        test_cond("sdscpy() against an originally longer string",
+            sdslen(x) == 1 && memcmp(x,"a\0",2) == 0)
+
+        x = sdscpy(x,"xyzxxxxxxxxxxyyyyyyyyyykkkkkkkkkk");
+        test_cond("sdscpy() against an originally shorter string",
+            sdslen(x) == 33 &&
+            memcmp(x,"xyzxxxxxxxxxxyyyyyyyyyykkkkkkkkkk\0",33) == 0)
+
+        sdsfree(x);
+        x = sdscatprintf(sdsempty(),"%d",123);
+        test_cond("sdscatprintf() seems working in the base case",
+            sdslen(x) == 3 && memcmp(x,"123\0",4) ==0)
+
+        sdsfree(x);
+        x = sdstrim(sdsnew("xxciaoyyy"),"xy");
+        test_cond("sdstrim() correctly trims characters",
+            sdslen(x) == 4 && memcmp(x,"ciao\0",5) == 0)
+
+        y = sdsrange(sdsdup(x),1,1);
+        test_cond("sdsrange(...,1,1)",
+            sdslen(y) == 1 && memcmp(y,"i\0",2) == 0)
+
+        sdsfree(y);
+        y = sdsrange(sdsdup(x),1,-1);
+        test_cond("sdsrange(...,1,-1)",
+            sdslen(y) == 3 && memcmp(y,"iao\0",4) == 0)
+
+        sdsfree(y);
+        y = sdsrange(sdsdup(x),-2,-1);
+        test_cond("sdsrange(...,-2,-1)",
+            sdslen(y) == 2 && memcmp(y,"ao\0",3) == 0)
+
+        sdsfree(y);
+        y = sdsrange(sdsdup(x),2,1);
+        test_cond("sdsrange(...,2,1)",
+            sdslen(y) == 0 && memcmp(y,"\0",1) == 0)
+
+        sdsfree(y);
+        y = sdsrange(sdsdup(x),1,100);
+        test_cond("sdsrange(...,1,100)",
+            sdslen(y) == 3 && memcmp(y,"iao\0",4) == 0)
+
+        sdsfree(y);
+        y = sdsrange(sdsdup(x),100,100);
+        test_cond("sdsrange(...,100,100)",
+            sdslen(y) == 0 && memcmp(y,"\0",1) == 0)
+
+        sdsfree(y);
+        sdsfree(x);
+        x = sdsnew("foo");
+        y = sdsnew("foa");
+        test_cond("sdscmp(foo,foa)", sdscmp(x,y) > 0)
+
+        sdsfree(y);
+        sdsfree(x);
+        x = sdsnew("bar");
+        y = sdsnew("bar");
+        test_cond("sdscmp(bar,bar)", sdscmp(x,y) == 0)
+
+        sdsfree(y);
+        sdsfree(x);
+        x = sdsnew("aar");
+        y = sdsnew("bar");
+        test_cond("sdscmp(bar,bar)", sdscmp(x,y) < 0)
+    }
+    test_report()
+}
+#endif

deps/hiredis/sds.h

@@ -0,0 +1,88 @@
+/* SDSLib, A C dynamic strings library
+ *
+ * Copyright (c) 2006-2010, Salvatore Sanfilippo <antirez at gmail dot com>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ *   * Redistributions of source code must retain the above copyright notice,
+ *     this list of conditions and the following disclaimer.
+ *   * Redistributions in binary form must reproduce the above copyright
+ *     notice, this list of conditions and the following disclaimer in the
+ *     documentation and/or other materials provided with the distribution.
+ *   * Neither the name of Redis nor the names of its contributors may be used
+ *     to endorse or promote products derived from this software without
+ *     specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#ifndef __SDS_H
+#define __SDS_H
+
+#include <sys/types.h>
+#include <stdarg.h>
+
+typedef char *sds;
+
+struct sdshdr {
+    int len;
+    int free;
+    char buf[];
+};
+
+static inline size_t sdslen(const sds s) {
+    struct sdshdr *sh = (void*)(s-(sizeof(struct sdshdr)));
+    return sh->len;
+}
+
+static inline size_t sdsavail(const sds s) {
+    struct sdshdr *sh = (void*)(s-(sizeof(struct sdshdr)));
+    return sh->free;
+}
+
+sds sdsnewlen(const void *init, size_t initlen);
+sds sdsnew(const char *init);
+sds sdsempty(void);
+size_t sdslen(const sds s);
+sds sdsdup(const sds s);
+void sdsfree(sds s);
+size_t sdsavail(sds s);
+sds sdsgrowzero(sds s, size_t len);
+sds sdscatlen(sds s, const void *t, size_t len);
+sds sdscat(sds s, const char *t);
+sds sdscpylen(sds s, char *t, size_t len);
+sds sdscpy(sds s, char *t);
+
+sds sdscatvprintf(sds s, const char *fmt, va_list ap);
+#ifdef __GNUC__
+sds sdscatprintf(sds s, const char *fmt, ...)
+    __attribute__((format(printf, 2, 3)));
+#else
+sds sdscatprintf(sds s, const char *fmt, ...);
+#endif
+
+sds sdstrim(sds s, const char *cset);
+sds sdsrange(sds s, int start, int end);
+void sdsupdatelen(sds s);
+int sdscmp(sds s1, sds s2);
+sds *sdssplitlen(char *s, int len, char *sep, int seplen, int *count);
+void sdsfreesplitres(sds *tokens, int count);
+void sdstolower(sds s);
+void sdstoupper(sds s);
+sds sdsfromlonglong(long long value);
+sds sdscatrepr(sds s, char *p, size_t len);
+sds *sdssplitargs(char *line, int *argc);
+
+#endif

deps/hiredis/test.c

@@ -0,0 +1,532 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <strings.h>
+#include <sys/time.h>
+#include <assert.h>
+#include <unistd.h>
+#include <signal.h>
+#include <errno.h>
+
+#include "hiredis.h"
+
+/* The following lines make up our testing "framework" :) */
+static int tests = 0, fails = 0;
+#define test(_s) { printf("#%02d ", ++tests); printf(_s); }
+#define test_cond(_c) if(_c) printf("PASSED\n"); else {printf("FAILED\n"); fails++;}
+
+static long long usec(void) {
+    struct timeval tv;
+    gettimeofday(&tv,NULL);
+    return (((long long)tv.tv_sec)*1000000)+tv.tv_usec;
+}
+
+static int use_unix = 0;
+static redisContext *blocking_context = NULL;
+static void __connect(redisContext **target) {
+    *target = blocking_context = (use_unix ?
+        redisConnectUnix("/tmp/redis.sock") : redisConnect((char*)"127.0.0.1", 6379));
+    if (blocking_context->err) {
+        printf("Connection error: %s\n", blocking_context->errstr);
+        exit(1);
+    }
+}
+
+static void test_format_commands(void) {
+    char *cmd;
+    int len;
+
+    test("Format command without interpolation: ");
+    len = redisFormatCommand(&cmd,"SET foo bar");
+    test_cond(strncmp(cmd,"*3\r\n$3\r\nSET\r\n$3\r\nfoo\r\n$3\r\nbar\r\n",len) == 0 &&
+        len == 4+4+(3+2)+4+(3+2)+4+(3+2));
+    free(cmd);
+
+    test("Format command with %%s string interpolation: ");
+    len = redisFormatCommand(&cmd,"SET %s %s","foo","bar");
+    test_cond(strncmp(cmd,"*3\r\n$3\r\nSET\r\n$3\r\nfoo\r\n$3\r\nbar\r\n",len) == 0 &&
+        len == 4+4+(3+2)+4+(3+2)+4+(3+2));
+    free(cmd);
+
+    test("Format command with %%s and an empty string: ");
+    len = redisFormatCommand(&cmd,"SET %s %s","foo","");
+    test_cond(strncmp(cmd,"*3\r\n$3\r\nSET\r\n$3\r\nfoo\r\n$0\r\n\r\n",len) == 0 &&
+        len == 4+4+(3+2)+4+(3+2)+4+(0+2));
+    free(cmd);
+
+    test("Format command with an empty string in between proper interpolations: ");
+    len = redisFormatCommand(&cmd,"SET %s %s","","foo");
+    test_cond(strncmp(cmd,"*3\r\n$3\r\nSET\r\n$0\r\n\r\n$3\r\nfoo\r\n",len) == 0 &&
+        len == 4+4+(3+2)+4+(0+2)+4+(3+2));
+    free(cmd);
+
+    test("Format command with %%b string interpolation: ");
+    len = redisFormatCommand(&cmd,"SET %b %b","foo",3,"b\0r",3);
+    test_cond(strncmp(cmd,"*3\r\n$3\r\nSET\r\n$3\r\nfoo\r\n$3\r\nb\0r\r\n",len) == 0 &&
+        len == 4+4+(3+2)+4+(3+2)+4+(3+2));
+    free(cmd);
+
+    test("Format command with %%b and an empty string: ");
+    len = redisFormatCommand(&cmd,"SET %b %b","foo",3,"",0);
+    test_cond(strncmp(cmd,"*3\r\n$3\r\nSET\r\n$3\r\nfoo\r\n$0\r\n\r\n",len) == 0 &&
+        len == 4+4+(3+2)+4+(3+2)+4+(0+2));
+    free(cmd);
+
+    test("Format command with literal %%: ");
+    len = redisFormatCommand(&cmd,"SET %% %%");
+    test_cond(strncmp(cmd,"*3\r\n$3\r\nSET\r\n$1\r\n%\r\n$1\r\n%\r\n",len) == 0 &&
+        len == 4+4+(3+2)+4+(1+2)+4+(1+2));
+    free(cmd);
+
+    test("Format command with printf-delegation (long long): ");
+    len = redisFormatCommand(&cmd,"key:%08lld",1234ll);
+    test_cond(strncmp(cmd,"*1\r\n$12\r\nkey:00001234\r\n",len) == 0 &&
+        len == 4+5+(12+2));
+    free(cmd);
+
+    test("Format command with printf-delegation (float): ");
+    len = redisFormatCommand(&cmd,"v:%06.1f",12.34f);
+    test_cond(strncmp(cmd,"*1\r\n$8\r\nv:0012.3\r\n",len) == 0 &&
+        len == 4+4+(8+2));
+    free(cmd);
+
+    test("Format command with printf-delegation and extra interpolation: ");
+    len = redisFormatCommand(&cmd,"key:%d %b",1234,"foo",3);
+    test_cond(strncmp(cmd,"*2\r\n$8\r\nkey:1234\r\n$3\r\nfoo\r\n",len) == 0 &&
+        len == 4+4+(8+2)+4+(3+2));
+    free(cmd);
+
+    test("Format command with wrong printf format and extra interpolation: ");
+    len = redisFormatCommand(&cmd,"key:%08p %b",1234,"foo",3);
+    test_cond(strncmp(cmd,"*2\r\n$6\r\nkey:8p\r\n$3\r\nfoo\r\n",len) == 0 &&
+        len == 4+4+(6+2)+4+(3+2));
+    free(cmd);
+
+    const char *argv[3];
+    argv[0] = "SET";
+    argv[1] = "foo\0xxx";
+    argv[2] = "bar";
+    size_t lens[3] = { 3, 7, 3 };
+    int argc = 3;
+
+    test("Format command by passing argc/argv without lengths: ");
+    len = redisFormatCommandArgv(&cmd,argc,argv,NULL);
+    test_cond(strncmp(cmd,"*3\r\n$3\r\nSET\r\n$3\r\nfoo\r\n$3\r\nbar\r\n",len) == 0 &&
+        len == 4+4+(3+2)+4+(3+2)+4+(3+2));
+    free(cmd);
+
+    test("Format command by passing argc/argv with lengths: ");
+    len = redisFormatCommandArgv(&cmd,argc,argv,lens);
+    test_cond(strncmp(cmd,"*3\r\n$3\r\nSET\r\n$7\r\nfoo\0xxx\r\n$3\r\nbar\r\n",len) == 0 &&
+        len == 4+4+(3+2)+4+(7+2)+4+(3+2));
+    free(cmd);
+}
+
+static void test_blocking_connection(void) {
+    redisContext *c;
+    redisReply *reply;
+    int major, minor;
+
+    test("Returns error when host cannot be resolved: ");
+    c = redisConnect((char*)"idontexist.local", 6379);
+    test_cond(c->err == REDIS_ERR_OTHER &&
+        strcmp(c->errstr,"Can't resolve: idontexist.local") == 0);
+    redisFree(c);
+
+    test("Returns error when the port is not open: ");
+    c = redisConnect((char*)"localhost", 56380);
+    test_cond(c->err == REDIS_ERR_IO &&
+        strcmp(c->errstr,"Connection refused") == 0);
+    redisFree(c);
+
+    __connect(&c);
+    test("Is able to deliver commands: ");
+    reply = redisCommand(c,"PING");
+    test_cond(reply->type == REDIS_REPLY_STATUS &&
+        strcasecmp(reply->str,"pong") == 0)
+    freeReplyObject(reply);
+
+    /* Switch to DB 9 for testing, now that we know we can chat. */
+    reply = redisCommand(c,"SELECT 9");
+    freeReplyObject(reply);
+
+    /* Make sure the DB is emtpy */
+    reply = redisCommand(c,"DBSIZE");
+    if (reply->type != REDIS_REPLY_INTEGER || reply->integer != 0) {
+        printf("Database #9 is not empty, test can not continue\n");
+        exit(1);
+    }
+    freeReplyObject(reply);
+
+    test("Is a able to send commands verbatim: ");
+    reply = redisCommand(c,"SET foo bar");
+    test_cond (reply->type == REDIS_REPLY_STATUS &&
+        strcasecmp(reply->str,"ok") == 0)
+    freeReplyObject(reply);
+
+    test("%%s String interpolation works: ");
+    reply = redisCommand(c,"SET %s %s","foo","hello world");
+    freeReplyObject(reply);
+    reply = redisCommand(c,"GET foo");
+    test_cond(reply->type == REDIS_REPLY_STRING &&
+        strcmp(reply->str,"hello world") == 0);
+    freeReplyObject(reply);
+
+    test("%%b String interpolation works: ");
+    reply = redisCommand(c,"SET %b %b","foo",3,"hello\x00world",11);
+    freeReplyObject(reply);
+    reply = redisCommand(c,"GET foo");
+    test_cond(reply->type == REDIS_REPLY_STRING &&
+        memcmp(reply->str,"hello\x00world",11) == 0)
+
+    test("Binary reply length is correct: ");
+    test_cond(reply->len == 11)
+    freeReplyObject(reply);
+
+    test("Can parse nil replies: ");
+    reply = redisCommand(c,"GET nokey");
+    test_cond(reply->type == REDIS_REPLY_NIL)
+    freeReplyObject(reply);
+
+    /* test 7 */
+    test("Can parse integer replies: ");
+    reply = redisCommand(c,"INCR mycounter");
+    test_cond(reply->type == REDIS_REPLY_INTEGER && reply->integer == 1)
+    freeReplyObject(reply);
+
+    test("Can parse multi bulk replies: ");
+    freeReplyObject(redisCommand(c,"LPUSH mylist foo"));
+    freeReplyObject(redisCommand(c,"LPUSH mylist bar"));
+    reply = redisCommand(c,"LRANGE mylist 0 -1");
+    test_cond(reply->type == REDIS_REPLY_ARRAY &&
+              reply->elements == 2 &&
+              !memcmp(reply->element[0]->str,"bar",3) &&
+              !memcmp(reply->element[1]->str,"foo",3))
+    freeReplyObject(reply);
+
+    /* m/e with multi bulk reply *before* other reply.
+     * specifically test ordering of reply items to parse. */
+    test("Can handle nested multi bulk replies: ");
+    freeReplyObject(redisCommand(c,"MULTI"));
+    freeReplyObject(redisCommand(c,"LRANGE mylist 0 -1"));
+    freeReplyObject(redisCommand(c,"PING"));
+    reply = (redisCommand(c,"EXEC"));
+    test_cond(reply->type == REDIS_REPLY_ARRAY &&
+              reply->elements == 2 &&
+              reply->element[0]->type == REDIS_REPLY_ARRAY &&
+              reply->element[0]->elements == 2 &&
+              !memcmp(reply->element[0]->element[0]->str,"bar",3) &&
+              !memcmp(reply->element[0]->element[1]->str,"foo",3) &&
+              reply->element[1]->type == REDIS_REPLY_STATUS &&
+              strcasecmp(reply->element[1]->str,"pong") == 0);
+    freeReplyObject(reply);
+
+    {
+        /* Find out Redis version to determine the path for the next test */
+        const char *field = "redis_version:";
+        char *p, *eptr;
+
+        reply = redisCommand(c,"INFO");
+        p = strstr(reply->str,field);
+        major = strtol(p+strlen(field),&eptr,10);
+        p = eptr+1; /* char next to the first "." */
+        minor = strtol(p,&eptr,10);
+        freeReplyObject(reply);
+    }
+
+    test("Returns I/O error when the connection is lost: ");
+    reply = redisCommand(c,"QUIT");
+    if (major >= 2 && minor > 0) {
+        /* > 2.0 returns OK on QUIT and read() should be issued once more
+         * to know the descriptor is at EOF. */
+        test_cond(strcasecmp(reply->str,"OK") == 0 &&
+            redisGetReply(c,(void**)&reply) == REDIS_ERR);
+        freeReplyObject(reply);
+    } else {
+        test_cond(reply == NULL);
+    }
+
+    /* On 2.0, QUIT will cause the connection to be closed immediately and
+     * the read(2) for the reply on QUIT will set the error to EOF.
+     * On >2.0, QUIT will return with OK and another read(2) needed to be
+     * issued to find out the socket was closed by the server. In both
+     * conditions, the error will be set to EOF. */
+    assert(c->err == REDIS_ERR_EOF &&
+        strcmp(c->errstr,"Server closed the connection") == 0);
+    redisFree(c);
+
+    __connect(&c);
+    test("Returns I/O error on socket timeout: ");
+    struct timeval tv = { 0, 1000 };
+    assert(redisSetTimeout(c,tv) == REDIS_OK);
+    test_cond(redisGetReply(c,(void**)&reply) == REDIS_ERR &&
+        c->err == REDIS_ERR_IO && errno == EAGAIN);
+    redisFree(c);
+
+    /* Context should be connected */
+    __connect(&c);
+}
+
+static void test_reply_reader(void) {
+    void *reader;
+    void *reply;
+    char *err;
+    int ret;
+
+    test("Error handling in reply parser: ");
+    reader = redisReplyReaderCreate();
+    redisReplyReaderFeed(reader,(char*)"@foo\r\n",6);
+    ret = redisReplyReaderGetReply(reader,NULL);
+    err = redisReplyReaderGetError(reader);
+    test_cond(ret == REDIS_ERR &&
+              strcasecmp(err,"Protocol error, got \"@\" as reply type byte") == 0);
+    redisReplyReaderFree(reader);
+
+    /* when the reply already contains multiple items, they must be free'd
+     * on an error. valgrind will bark when this doesn't happen. */
+    test("Memory cleanup in reply parser: ");
+    reader = redisReplyReaderCreate();
+    redisReplyReaderFeed(reader,(char*)"*2\r\n",4);
+    redisReplyReaderFeed(reader,(char*)"$5\r\nhello\r\n",11);
+    redisReplyReaderFeed(reader,(char*)"@foo\r\n",6);
+    ret = redisReplyReaderGetReply(reader,NULL);
+    err = redisReplyReaderGetError(reader);
+    test_cond(ret == REDIS_ERR &&
+              strcasecmp(err,"Protocol error, got \"@\" as reply type byte") == 0);
+    redisReplyReaderFree(reader);
+
+    test("Set error on nested multi bulks with depth > 1: ");
+    reader = redisReplyReaderCreate();
+    redisReplyReaderFeed(reader,(char*)"*1\r\n",4);
+    redisReplyReaderFeed(reader,(char*)"*1\r\n",4);
+    redisReplyReaderFeed(reader,(char*)"*1\r\n",4);
+    ret = redisReplyReaderGetReply(reader,NULL);
+    err = redisReplyReaderGetError(reader);
+    test_cond(ret == REDIS_ERR &&
+              strncasecmp(err,"No support for",14) == 0);
+    redisReplyReaderFree(reader);
+
+    test("Works with NULL functions for reply: ");
+    reader = redisReplyReaderCreate();
+    redisReplyReaderSetReplyObjectFunctions(reader,NULL);
+    redisReplyReaderFeed(reader,(char*)"+OK\r\n",5);
+    ret = redisReplyReaderGetReply(reader,&reply);
+    test_cond(ret == REDIS_OK && reply == (void*)REDIS_REPLY_STATUS);
+    redisReplyReaderFree(reader);
+
+    test("Works when a single newline (\\r\\n) covers two calls to feed: ");
+    reader = redisReplyReaderCreate();
+    redisReplyReaderSetReplyObjectFunctions(reader,NULL);
+    redisReplyReaderFeed(reader,(char*)"+OK\r",4);
+    ret = redisReplyReaderGetReply(reader,&reply);
+    assert(ret == REDIS_OK && reply == NULL);
+    redisReplyReaderFeed(reader,(char*)"\n",1);
+    ret = redisReplyReaderGetReply(reader,&reply);
+    test_cond(ret == REDIS_OK && reply == (void*)REDIS_REPLY_STATUS);
+    redisReplyReaderFree(reader);
+
+    test("Properly reset state after protocol error: ");
+    reader = redisReplyReaderCreate();
+    redisReplyReaderSetReplyObjectFunctions(reader,NULL);
+    redisReplyReaderFeed(reader,(char*)"x",1);
+    ret = redisReplyReaderGetReply(reader,&reply);
+    assert(ret == REDIS_ERR);
+    ret = redisReplyReaderGetReply(reader,&reply);
+    test_cond(ret == REDIS_OK && reply == NULL)
+}
+
+static void test_throughput(void) {
+    int i, num;
+    long long t1, t2;
+    redisContext *c = blocking_context;
+    redisReply **replies;
+
+    test("Throughput:\n");
+    for (i = 0; i < 500; i++)
+        freeReplyObject(redisCommand(c,"LPUSH mylist foo"));
+
+    num = 1000;
+    replies = malloc(sizeof(redisReply*)*num);
+    t1 = usec();
+    for (i = 0; i < num; i++) {
+        replies[i] = redisCommand(c,"PING");
+        assert(replies[i] != NULL && replies[i]->type == REDIS_REPLY_STATUS);
+    }
+    t2 = usec();
+    for (i = 0; i < num; i++) freeReplyObject(replies[i]);
+    free(replies);
+    printf("\t(%dx PING: %.3fs)\n", num, (t2-t1)/1000000.0);
+
+    replies = malloc(sizeof(redisReply*)*num);
+    t1 = usec();
+    for (i = 0; i < num; i++) {
+        replies[i] = redisCommand(c,"LRANGE mylist 0 499");
+        assert(replies[i] != NULL && replies[i]->type == REDIS_REPLY_ARRAY);
+        assert(replies[i] != NULL && replies[i]->elements == 500);
+    }
+    t2 = usec();
+    for (i = 0; i < num; i++) freeReplyObject(replies[i]);
+    free(replies);
+    printf("\t(%dx LRANGE with 500 elements: %.3fs)\n", num, (t2-t1)/1000000.0);
+
+    num = 10000;
+    replies = malloc(sizeof(redisReply*)*num);
+    for (i = 0; i < num; i++)
+        redisAppendCommand(c,"PING");
+    t1 = usec();
+    for (i = 0; i < num; i++) {
+        assert(redisGetReply(c, (void*)&replies[i]) == REDIS_OK);
+        assert(replies[i] != NULL && replies[i]->type == REDIS_REPLY_STATUS);
+    }
+    t2 = usec();
+    for (i = 0; i < num; i++) freeReplyObject(replies[i]);
+    free(replies);
+    printf("\t(%dx PING (pipelined): %.3fs)\n", num, (t2-t1)/1000000.0);
+
+    replies = malloc(sizeof(redisReply*)*num);
+    for (i = 0; i < num; i++)
+        redisAppendCommand(c,"LRANGE mylist 0 499");
+    t1 = usec();
+    for (i = 0; i < num; i++) {
+        assert(redisGetReply(c, (void*)&replies[i]) == REDIS_OK);
+        assert(replies[i] != NULL && replies[i]->type == REDIS_REPLY_ARRAY);
+        assert(replies[i] != NULL && replies[i]->elements == 500);
+    }
+    t2 = usec();
+    for (i = 0; i < num; i++) freeReplyObject(replies[i]);
+    free(replies);
+    printf("\t(%dx LRANGE with 500 elements (pipelined): %.3fs)\n", num, (t2-t1)/1000000.0);
+}
+
+static void cleanup(void) {
+    redisContext *c = blocking_context;
+    redisReply *reply;
+
+    /* Make sure we're on DB 9 */
+    reply = redisCommand(c,"SELECT 9");
+    assert(reply != NULL); freeReplyObject(reply);
+    reply = redisCommand(c,"FLUSHDB");
+    assert(reply != NULL); freeReplyObject(reply);
+    redisFree(c);
+}
+
+// static long __test_callback_flags = 0;
+// static void __test_callback(redisContext *c, void *privdata) {
+//     ((void)c);
+//     /* Shift to detect execution order */
+//     __test_callback_flags <<= 8;
+//     __test_callback_flags |= (long)privdata;
+// }
+//
+// static void __test_reply_callback(redisContext *c, redisReply *reply, void *privdata) {
+//     ((void)c);
+//     /* Shift to detect execution order */
+//     __test_callback_flags <<= 8;
+//     __test_callback_flags |= (long)privdata;
+//     if (reply) freeReplyObject(reply);
+// }
+//
+// static redisContext *__connect_nonblock() {
+//     /* Reset callback flags */
+//     __test_callback_flags = 0;
+//     return redisConnectNonBlock("127.0.0.1", 6379, NULL);
+// }
+//
+// static void test_nonblocking_connection() {
+//     redisContext *c;
+//     int wdone = 0;
+//
+//     test("Calls command callback when command is issued: ");
+//     c = __connect_nonblock();
+//     redisSetCommandCallback(c,__test_callback,(void*)1);
+//     redisCommand(c,"PING");
+//     test_cond(__test_callback_flags == 1);
+//     redisFree(c);
+//
+//     test("Calls disconnect callback on redisDisconnect: ");
+//     c = __connect_nonblock();
+//     redisSetDisconnectCallback(c,__test_callback,(void*)2);
+//     redisDisconnect(c);
+//     test_cond(__test_callback_flags == 2);
+//     redisFree(c);
+//
+//     test("Calls disconnect callback and free callback on redisFree: ");
+//     c = __connect_nonblock();
+//     redisSetDisconnectCallback(c,__test_callback,(void*)2);
+//     redisSetFreeCallback(c,__test_callback,(void*)4);
+//     redisFree(c);
+//     test_cond(__test_callback_flags == ((2 << 8) | 4));
+//
+//     test("redisBufferWrite against empty write buffer: ");
+//     c = __connect_nonblock();
+//     test_cond(redisBufferWrite(c,&wdone) == REDIS_OK && wdone == 1);
+//     redisFree(c);
+//
+//     test("redisBufferWrite against not yet connected fd: ");
+//     c = __connect_nonblock();
+//     redisCommand(c,"PING");
+//     test_cond(redisBufferWrite(c,NULL) == REDIS_ERR &&
+//               strncmp(c->error,"write:",6) == 0);
+//     redisFree(c);
+//
+//     test("redisBufferWrite against closed fd: ");
+//     c = __connect_nonblock();
+//     redisCommand(c,"PING");
+//     redisDisconnect(c);
+//     test_cond(redisBufferWrite(c,NULL) == REDIS_ERR &&
+//               strncmp(c->error,"write:",6) == 0);
+//     redisFree(c);
+//
+//     test("Process callbacks in the right sequence: ");
+//     c = __connect_nonblock();
+//     redisCommandWithCallback(c,__test_reply_callback,(void*)1,"PING");
+//     redisCommandWithCallback(c,__test_reply_callback,(void*)2,"PING");
+//     redisCommandWithCallback(c,__test_reply_callback,(void*)3,"PING");
+//
+//     /* Write output buffer */
+//     wdone = 0;
+//     while(!wdone) {
+//         usleep(500);
+//         redisBufferWrite(c,&wdone);
+//     }
+//
+//     /* Read until at least one callback is executed (the 3 replies will
+//      * arrive in a single packet, causing all callbacks to be executed in
+//      * a single pass). */
+//     while(__test_callback_flags == 0) {
+//         assert(redisBufferRead(c) == REDIS_OK);
+//         redisProcessCallbacks(c);
+//     }
+//     test_cond(__test_callback_flags == 0x010203);
+//     redisFree(c);
+//
+//     test("redisDisconnect executes pending callbacks with NULL reply: ");
+//     c = __connect_nonblock();
+//     redisSetDisconnectCallback(c,__test_callback,(void*)1);
+//     redisCommandWithCallback(c,__test_reply_callback,(void*)2,"PING");
+//     redisDisconnect(c);
+//     test_cond(__test_callback_flags == 0x0201);
+//     redisFree(c);
+// }
+
+int main(int argc, char **argv) {
+    if (argc > 1) {
+        if (strcmp(argv[1],"-s") == 0)
+            use_unix = 1;
+    }
+
+    signal(SIGPIPE, SIG_IGN);
+    test_format_commands();
+    test_blocking_connection();
+    test_reply_reader();
+    // test_nonblocking_connection();
+    test_throughput();
+    cleanup();
+
+    if (fails == 0) {
+        printf("ALL TESTS PASSED\n");
+    } else {
+        printf("*** %d TESTS FAILED ***\n", fails);
+    }
+    return 0;
+}

deps/hiredis/util.h

@@ -1,5 +1,4 @@
-/* zmalloc - total amount of allocated memory aware version of malloc()
- *
+/*
  * Copyright (c) 2009-2010, Salvatore Sanfilippo <antirez at gmail dot com>
  * All rights reserved.
  *
@@ -28,14 +27,14 @@
  * POSSIBILITY OF SUCH DAMAGE.
  */
 
-#ifndef _ZMALLOC_H
-#define _ZMALLOC_H
+#ifndef __UTIL_H
+#define __UTIL_H
+#include <stdlib.h>
 
-void *zmalloc(size_t size);
-void *zrealloc(void *ptr, size_t size);
-void zfree(void *ptr);
-char *zstrdup(const char *s);
-size_t zmalloc_used_memory(void);
-void zmalloc_enable_thread_safeness(void);
+/* Abort on out of memory */
+static void redisOOM(void) {
+    fprintf(stderr,"Out of memory in hiredis");
+    exit(1);
+}
 
-#endif /* _ZMALLOC_H */
+#endif

deps/jemalloc/.gitignore

@@ -0,0 +1,15 @@
+/Makefile
+/config.stamp
+/config.status
+
+/doc/html.xsl
+/doc/jemalloc.xml
+/doc/manpages.xsl
+
+/src/*.d
+/lib
+
+# The following headers are generated
+/include/jemalloc/*.h
+/include/jemalloc/internal/jemalloc_internal.h
+/test/jemalloc_test.h

deps/jemalloc/COPYING

@@ -0,0 +1,51 @@
+Unless otherwise specified, files in the jemalloc source distribution are
+subject to the following licenses:
+--------------------------------------------------------------------------------
+Copyright (C) 2002-2010 Jason Evans <jasone@canonware.com>.
+All rights reserved.
+Copyright (C) 2007-2010 Mozilla Foundation.  All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are met:
+1. Redistributions of source code must retain the above copyright notice(s),
+   this list of conditions and the following disclaimer.
+2. Redistributions in binary form must reproduce the above copyright notice(s),
+   this list of conditions and the following disclaimer in the documentation
+   and/or other materials provided with the distribution.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDER(S) ``AS IS'' AND ANY EXPRESS
+OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO
+EVENT SHALL THE COPYRIGHT HOLDER(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE
+OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
+ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+--------------------------------------------------------------------------------
+Copyright (C) 2009-2010 Facebook, Inc.
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without modification,
+are permitted provided that the following conditions are met:
+* Redistributions of source code must retain the above copyright notice, this
+  list of conditions and the following disclaimer.
+* Redistributions in binary form must reproduce the above copyright notice, this
+  list of conditions and the following disclaimer in the documentation and/or
+  other materials provided with the distribution.
+* Neither the name of Facebook, Inc. nor the names of its contributors may be
+  used to endorse or promote products derived from this software without
+  specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
+ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+--------------------------------------------------------------------------------

deps/jemalloc/ChangeLog

@@ -0,0 +1,213 @@
+Following are change highlights associated with official releases.  Important
+bug fixes are all mentioned, but internal enhancements are omitted here for
+brevity (even though they are more fun to write about).  Much more detail can be
+found in the git revision history:
+
+    http://www.canonware.com/cgi-bin/gitweb.cgi?p=jemalloc.git
+    git://canonware.com/jemalloc.git
+
+* 2.2.1 (March 30, 2011)
+
+  Bug fixes:
+  - Implement atomic operations for x86/x64.  This fixes compilation failures
+    for versions of gcc that are still in wide use.
+  - Fix an assertion in arena_purge().
+
+* 2.2.0 (March 22, 2011)
+
+  This version incorporates several improvements to algorithms and data
+  structures that tend to reduce fragmentation and increase speed.
+
+  New features:
+  - Add the "stats.cactive" mallctl.
+  - Update pprof (from google-perftools 1.7).
+  - Improve backtracing-related configuration logic, and add the
+    --disable-prof-libgcc option.
+
+  Bug fixes:
+  - Change default symbol visibility from "internal", to "hidden", which
+    decreases the overhead of library-internal function calls.
+  - Fix symbol visibility so that it is also set on OS X.
+  - Fix a build dependency regression caused by the introduction of the .pic.o
+    suffix for PIC object files.
+  - Add missing checks for mutex initialization failures.
+  - Don't use libgcc-based backtracing except on x64, where it is known to work.
+  - Fix deadlocks on OS X that were due to memory allocation in
+    pthread_mutex_lock().
+  - Heap profiling-specific fixes:
+    + Fix memory corruption due to integer overflow in small region index
+      computation, when using a small enough sample interval that profiling
+      context pointers are stored in small run headers.
+    + Fix a bootstrap ordering bug that only occurred with TLS disabled.
+    + Fix a rallocm() rsize bug.
+    + Fix error detection bugs for aligned memory allocation.
+
+* 2.1.3 (March 14, 2011)
+
+  Bug fixes:
+  - Fix a cpp logic regression (due to the "thread.{de,}allocatedp" mallctl fix
+    for OS X in 2.1.2).
+  - Fix a "thread.arena" mallctl bug.
+  - Fix a thread cache stats merging bug.
+
+* 2.1.2 (March 2, 2011)
+
+  Bug fixes:
+  - Fix "thread.{de,}allocatedp" mallctl for OS X.
+  - Add missing jemalloc.a to build system.
+
+* 2.1.1 (January 31, 2011)
+
+  Bug fixes:
+  - Fix aligned huge reallocation (affected allocm()).
+  - Fix the ALLOCM_LG_ALIGN macro definition.
+  - Fix a heap dumping deadlock.
+  - Fix a "thread.arena" mallctl bug.
+
+* 2.1.0 (December 3, 2010)
+
+  This version incorporates some optimizations that can't quite be considered
+  bug fixes.
+
+  New features:
+  - Use Linux's mremap(2) for huge object reallocation when possible.
+  - Avoid locking in mallctl*() when possible.
+  - Add the "thread.[de]allocatedp" mallctl's.
+  - Convert the manual page source from roff to DocBook, and generate both roff
+    and HTML manuals.
+
+  Bug fixes:
+  - Fix a crash due to incorrect bootstrap ordering.  This only impacted
+    --enable-debug --enable-dss configurations.
+  - Fix a minor statistics bug for mallctl("swap.avail", ...).
+
+* 2.0.1 (October 29, 2010)
+
+  Bug fixes:
+  - Fix a race condition in heap profiling that could cause undefined behavior
+    if "opt.prof_accum" were disabled.
+  - Add missing mutex unlocks for some OOM error paths in the heap profiling
+    code.
+  - Fix a compilation error for non-C99 builds.
+
+* 2.0.0 (October 24, 2010)
+
+  This version focuses on the experimental *allocm() API, and on improved
+  run-time configuration/introspection.  Nonetheless, numerous performance
+  improvements are also included.
+
+  New features:
+  - Implement the experimental {,r,s,d}allocm() API, which provides a superset
+    of the functionality available via malloc(), calloc(), posix_memalign(),
+    realloc(), malloc_usable_size(), and free().  These functions can be used to
+    allocate/reallocate aligned zeroed memory, ask for optional extra memory
+    during reallocation, prevent object movement during reallocation, etc.
+  - Replace JEMALLOC_OPTIONS/JEMALLOC_PROF_PREFIX with MALLOC_CONF, which is
+    more human-readable, and more flexible.  For example:
+      JEMALLOC_OPTIONS=AJP
+    is now:
+      MALLOC_CONF=abort:true,fill:true,stats_print:true
+  - Port to Apple OS X.  Sponsored by Mozilla.
+  - Make it possible for the application to control thread-->arena mappings via
+    the "thread.arena" mallctl.
+  - Add compile-time support for all TLS-related functionality via pthreads TSD.
+    This is mainly of interest for OS X, which does not support TLS, but has a
+    TSD implementation with similar performance.
+  - Override memalign() and valloc() if they are provided by the system.
+  - Add the "arenas.purge" mallctl, which can be used to synchronously purge all
+    dirty unused pages.
+  - Make cumulative heap profiling data optional, so that it is possible to
+    limit the amount of memory consumed by heap profiling data structures.
+  - Add per thread allocation counters that can be accessed via the
+    "thread.allocated" and "thread.deallocated" mallctls.
+
+  Incompatible changes:
+  - Remove JEMALLOC_OPTIONS and malloc_options (see MALLOC_CONF above).
+  - Increase default backtrace depth from 4 to 128 for heap profiling.
+  - Disable interval-based profile dumps by default.
+
+  Bug fixes:
+  - Remove bad assertions in fork handler functions.  These assertions could
+    cause aborts for some combinations of configure settings.
+  - Fix strerror_r() usage to deal with non-standard semantics in GNU libc.
+  - Fix leak context reporting.  This bug tended to cause the number of contexts
+    to be underreported (though the reported number of objects and bytes were
+    correct).
+  - Fix a realloc() bug for large in-place growing reallocation.  This bug could
+    cause memory corruption, but it was hard to trigger.
+  - Fix an allocation bug for small allocations that could be triggered if
+    multiple threads raced to create a new run of backing pages.
+  - Enhance the heap profiler to trigger samples based on usable size, rather
+    than request size.
+  - Fix a heap profiling bug due to sometimes losing track of requested object
+    size for sampled objects.
+
+* 1.0.3 (August 12, 2010)
+
+  Bug fixes:
+  - Fix the libunwind-based implementation of stack backtracing (used for heap
+    profiling).  This bug could cause zero-length backtraces to be reported.
+  - Add a missing mutex unlock in library initialization code.  If multiple
+    threads raced to initialize malloc, some of them could end up permanently
+    blocked.
+
+* 1.0.2 (May 11, 2010)
+
+  Bug fixes:
+  - Fix junk filling of large objects, which could cause memory corruption.
+  - Add MAP_NORESERVE support for chunk mapping, because otherwise virtual
+    memory limits could cause swap file configuration to fail.  Contributed by
+    Jordan DeLong.
+
+* 1.0.1 (April 14, 2010)
+
+  Bug fixes:
+  - Fix compilation when --enable-fill is specified.
+  - Fix threads-related profiling bugs that affected accuracy and caused memory
+    to be leaked during thread exit.
+  - Fix dirty page purging race conditions that could cause crashes.
+  - Fix crash in tcache flushing code during thread destruction.
+
+* 1.0.0 (April 11, 2010)
+
+  This release focuses on speed and run-time introspection.  Numerous
+  algorithmic improvements make this release substantially faster than its
+  predecessors.
+
+  New features:
+  - Implement autoconf-based configuration system.
+  - Add mallctl*(), for the purposes of introspection and run-time
+    configuration.
+  - Make it possible for the application to manually flush a thread's cache, via
+    the "tcache.flush" mallctl.
+  - Base maximum dirty page count on proportion of active memory.
+  - Compute various addtional run-time statistics, including per size class
+    statistics for large objects.
+  - Expose malloc_stats_print(), which can be called repeatedly by the
+    application.
+  - Simplify the malloc_message() signature to only take one string argument,
+    and incorporate an opaque data pointer argument for use by the application
+    in combination with malloc_stats_print().
+  - Add support for allocation backed by one or more swap files, and allow the
+    application to disable over-commit if swap files are in use.
+  - Implement allocation profiling and leak checking.
+
+  Removed features:
+  - Remove the dynamic arena rebalancing code, since thread-specific caching
+    reduces its utility.
+
+  Bug fixes:
+  - Modify chunk allocation to work when address space layout randomization
+    (ASLR) is in use.
+  - Fix thread cleanup bugs related to TLS destruction.
+  - Handle 0-size allocation requests in posix_memalign().
+  - Fix a chunk leak.  The leaked chunks were never touched, so this impacted
+    virtual memory usage, but not physical memory usage.
+
+* linux_2008082[78]a (August 27/28, 2008)
+
+  These snapshot releases are the simple result of incorporating Linux-specific
+  support into the FreeBSD malloc sources.
+
+--------------------------------------------------------------------------------
+vim:filetype=text:textwidth=80

deps/jemalloc/INSTALL

@@ -0,0 +1,251 @@
+Building and installing jemalloc can be as simple as typing the following while
+in the root directory of the source tree:
+
+    ./configure
+    make
+    make install
+
+=== 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-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-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.
+
+--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.
+
+--enable-cc-silence
+    Enable code that silences non-useful compiler warnings.  This is helpful
+    when trying to tell serious warnings from those due to compiler
+    limitations, but it potentially incurs a performance penalty.
+
+--enable-debug
+    Enable assertions and validation code.  This incurs a substantial
+    performance hit, but is very useful during application development.
+
+--enable-stats
+    Enable statistics gathering functionality.  See the "opt.stats_print"
+    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-tiny
+    Disable tiny (sub-quantum-sized) object support.  Technically it is not
+    legal for a malloc implementation to allocate objects with less than
+    quantum alignment (8 or 16 bytes, depending on architecture), but in
+    practice it never causes any problems if, for example, 4-byte allocations
+    are 4-byte-aligned.
+
+--disable-tcache
+    Disable thread-specific caches for small objects.  Objects are cached and
+    released in bulk, thus reducing the total number of mutex operations.  See
+    the "opt.tcache" option for usage details.
+
+--enable-swap
+    Enable mmap()ed swap file support.  When this feature is built in, it is
+    possible to specify one or more files that act as backing store.  This
+    effectively allows for per application swap files.
+
+--enable-dss
+    Enable support for page allocation/deallocation via sbrk(2), in addition to
+    mmap(2).
+
+--enable-fill
+    Enable support for junk/zero filling of memory.  See the "opt.junk"/
+    "opt.zero" option documentation for usage details.
+
+--enable-xmalloc
+    Enable support for optional immediate termination due to out-of-memory
+    errors, as is commonly implemented by "xmalloc" wrapper function for malloc.
+    See the "opt.xmalloc" option documentation for usage details.
+
+--enable-sysv
+    Enable support for System V semantics, wherein malloc(0) returns NULL
+    rather than a minimal allocation.  See the "opt.sysv" option documentation
+    for usage details.
+
+--enable-dynamic-page-shift
+    Under most conditions, the system page size never changes (usually 4KiB or
+    8KiB, depending on architecture and configuration), and unless this option
+    is enabled, jemalloc assumes that page size can safely be determined during
+    configuration and hard-coded.  Enabling dynamic page size determination has
+    a measurable impact on performance, since the compiler is forced to load
+    the page size from memory rather than embedding immediate values.
+
+--disable-lazy-lock
+    Disable code that wraps pthread_create() to detect when an application
+    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-tls
+    Disable thread-local storage (TLS), which allows for fast access to
+    thread-local variables via the __thread keyword.  If TLS is available,
+    jemalloc uses it for several purposes.
+
+--with-xslroot=<path>
+    Specify where to find DocBook XSL stylesheets when building the
+    documentation.
+
+The following environment variables (not a definitive list) impact configure's
+behavior:
+
+CFLAGS="?"
+    Pass these flags to the compiler.  You probably shouldn't define this unless
+    you know what you are doing.  (Use EXTRA_CFLAGS instead.)
+
+EXTRA_CFLAGS="?"
+    Append these flags to CFLAGS.  This makes it possible to add flags such as
+    -Werror, while allowing the configure script to determine what other flags
+    are appropriate for the specified configuration.
+
+    The configure script specifically checks whether an optimization flag (-O*)
+    is specified in EXTRA_CFLAGS, and refrains from specifying an optimization
+    level if it finds that one has already been specified.
+
+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.
+
+=== Advanced compilation =======================================================
+
+To install only parts of jemalloc, use the following targets:
+
+    install_bin
+    install_include
+    install_lib
+    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 any of the following commands:
+
+    nroff -man -t doc/jemalloc.3
+
+    groff -man -t -Tps doc/jemalloc.3 | ps2pdf - doc/jemalloc.3.pdf
+
+    (cd doc; groff -man -man-ext -t -Thtml jemalloc.3 > jemalloc.3.html)

deps/jemalloc/Makefile.in

@@ -0,0 +1,259 @@
+# Clear out all vpaths, then set just one (default vpath) for the main build
+# directory.
+vpath
+vpath % .
+
+# Clear the default suffixes, so that built-in rules are not used.
+.SUFFIXES :
+
+SHELL := /bin/sh
+
+CC := @CC@
+
+# Configuration parameters.
+DESTDIR =
+BINDIR := $(DESTDIR)@BINDIR@
+INCLUDEDIR := $(DESTDIR)@INCLUDEDIR@
+LIBDIR := $(DESTDIR)@LIBDIR@
+DATADIR := $(DESTDIR)@DATADIR@
+MANDIR := $(DESTDIR)@MANDIR@
+
+# Build parameters.
+CPPFLAGS := @CPPFLAGS@ -I@srcroot@include -I@objroot@include
+CFLAGS := @CFLAGS@
+ifeq (macho, @abi@)
+CFLAGS += -dynamic
+endif
+LDFLAGS := @LDFLAGS@
+LIBS := @LIBS@
+RPATH_EXTRA := @RPATH_EXTRA@
+ifeq (macho, @abi@)
+SO := dylib
+WL_SONAME := dylib_install_name
+else
+SO := so
+WL_SONAME := soname
+endif
+REV := 1
+ifeq (macho, @abi@)
+TEST_LIBRARY_PATH := DYLD_FALLBACK_LIBRARY_PATH=@objroot@lib
+else
+TEST_LIBRARY_PATH :=
+endif
+
+# Lists of files.
+BINS := @srcroot@bin/pprof
+CHDRS := @objroot@include/jemalloc/jemalloc@install_suffix@.h \
+	@objroot@include/jemalloc/jemalloc_defs@install_suffix@.h
+CSRCS := @srcroot@src/jemalloc.c @srcroot@src/arena.c @srcroot@src/atomic.c \
+	@srcroot@src/base.c @srcroot@src/bitmap.c @srcroot@src/chunk.c \
+	@srcroot@src/chunk_dss.c @srcroot@src/chunk_mmap.c \
+	@srcroot@src/chunk_swap.c @srcroot@src/ckh.c @srcroot@src/ctl.c \
+	@srcroot@src/extent.c @srcroot@src/hash.c @srcroot@src/huge.c \
+	@srcroot@src/mb.c @srcroot@src/mutex.c @srcroot@src/prof.c \
+	@srcroot@src/rtree.c @srcroot@src/stats.c @srcroot@src/tcache.c
+ifeq (macho, @abi@)
+CSRCS += @srcroot@src/zone.c
+endif
+STATIC_LIBS :=	@objroot@lib/libjemalloc@install_suffix@.a
+DSOS := @objroot@lib/libjemalloc@install_suffix@.$(SO).$(REV) \
+	@objroot@lib/libjemalloc@install_suffix@.$(SO) \
+	@objroot@lib/libjemalloc@install_suffix@_pic.a
+MAN3 := @objroot@doc/jemalloc@install_suffix@.3
+DOCS_XML := @objroot@doc/jemalloc@install_suffix@.xml
+DOCS_HTML := $(DOCS_XML:@objroot@%.xml=@srcroot@%.html)
+DOCS_MAN3 := $(DOCS_XML:@objroot@%.xml=@srcroot@%.3)
+DOCS := $(DOCS_HTML) $(DOCS_MAN3)
+CTESTS := @srcroot@test/allocated.c @srcroot@test/allocm.c \
+	@srcroot@test/bitmap.c @srcroot@test/mremap.c \
+	@srcroot@test/posix_memalign.c @srcroot@test/rallocm.c \
+	@srcroot@test/thread_arena.c
+
+.PHONY: all dist doc_html doc_man doc
+.PHONY: install_bin install_include install_lib
+.PHONY: install_html install_man install_doc install
+.PHONY: tests check clean distclean relclean
+
+.SECONDARY : $(CTESTS:@srcroot@%.c=@objroot@%.o)
+
+# Default target.
+all: $(DSOS) $(STATIC_LIBS)
+
+dist: doc
+
+@srcroot@doc/%.html : @objroot@doc/%.xml @srcroot@doc/stylesheet.xsl @objroot@doc/html.xsl
+	@XSLTPROC@ -o $@ @objroot@doc/html.xsl $<
+
+@srcroot@doc/%.3 : @objroot@doc/%.xml @srcroot@doc/stylesheet.xsl @objroot@doc/manpages.xsl
+	@XSLTPROC@ -o $@ @objroot@doc/manpages.xsl $<
+
+doc_html: $(DOCS_HTML)
+doc_man: $(DOCS_MAN3)
+doc: $(DOCS)
+
+#
+# Include generated dependency files.
+#
+-include $(CSRCS:@srcroot@%.c=@objroot@%.d)
+-include $(CSRCS:@srcroot@%.c=@objroot@%.pic.d)
+-include $(CTESTS:@srcroot@%.c=@objroot@%.d)
+
+@objroot@src/%.o: @srcroot@src/%.c
+	@mkdir -p $(@D)
+	$(CC) $(CFLAGS) -c $(CPPFLAGS) -o $@ $<
+	@$(SHELL) -ec "$(CC) -MM $(CPPFLAGS) $< | sed \"s/\($(subst /,\/,$(notdir $(basename $@)))\)\.o\([ :]*\)/$(subst /,\/,$(strip $(dir $@)))\1.o \2/g\" > $(@:%.o=%.d)"
+
+@objroot@src/%.pic.o: @srcroot@src/%.c
+	@mkdir -p $(@D)
+	$(CC) $(CFLAGS) -fPIC -DPIC -c $(CPPFLAGS) -o $@ $<
+	@$(SHELL) -ec "$(CC) -MM $(CPPFLAGS) $< | sed \"s/\($(subst /,\/,$(notdir $(basename $(basename $@))))\)\.o\([ :]*\)/$(subst /,\/,$(strip $(dir $@)))\1.pic.o \2/g\" > $(@:%.o=%.d)"
+
+%.$(SO) : %.$(SO).$(REV)
+	@mkdir -p $(@D)
+	ln -sf $(<F) $@
+
+@objroot@lib/libjemalloc@install_suffix@.$(SO).$(REV) : $(CSRCS:@srcroot@%.c=@objroot@%.pic.o)
+	@mkdir -p $(@D)
+	$(CC) -shared -Wl,-$(WL_SONAME),$(@F) $(RPATH_EXTRA:%=@RPATH@%) -o $@ $+ $(LDFLAGS) $(LIBS)
+
+@objroot@lib/libjemalloc@install_suffix@_pic.a : $(CSRCS:@srcroot@%.c=@objroot@%.pic.o)
+	@mkdir -p $(@D)
+	ar crus $@ $+
+
+@objroot@lib/libjemalloc@install_suffix@.a : $(CSRCS:@srcroot@%.c=@objroot@%.o)
+	@mkdir -p $(@D)
+	ar crus $@ $+
+
+@objroot@test/%.o: @srcroot@test/%.c
+	@mkdir -p $(@D)
+	$(CC) $(CFLAGS) -c $(CPPFLAGS) -I@objroot@test -o $@ $<
+	@$(SHELL) -ec "$(CC) -MM $(CPPFLAGS) -I@objroot@test $< | sed \"s/\($(subst /,\/,$(notdir $(basename $@)))\)\.o\([ :]*\)/$(subst /,\/,$(strip $(dir $@)))\1.o \2/g\" > $(@:%.o=%.d)"
+
+# Automatic dependency generation misses #include "*.c".
+@objroot@test/bitmap.o : @objroot@src/bitmap.o
+
+@objroot@test/%: @objroot@test/%.o \
+		 @objroot@lib/libjemalloc@install_suffix@.$(SO)
+	@mkdir -p $(@D)
+ifneq (@RPATH@, )
+	$(CC) -o $@ $< @RPATH@@objroot@lib -L@objroot@lib -ljemalloc@install_suffix@
+else
+	$(CC) -o $@ $< -L@objroot@lib -ljemalloc@install_suffix@
+endif
+
+install_bin:
+	install -d $(BINDIR)
+	@for b in $(BINS); do \
+	echo "install -m 755 $$b $(BINDIR)"; \
+	install -m 755 $$b $(BINDIR); \
+done
+
+install_include:
+	install -d $(INCLUDEDIR)/jemalloc
+	@for h in $(CHDRS); do \
+	echo "install -m 644 $$h $(INCLUDEDIR)/jemalloc"; \
+	install -m 644 $$h $(INCLUDEDIR)/jemalloc; \
+done
+
+install_lib: $(DSOS) $(STATIC_LIBS)
+	install -d $(LIBDIR)
+	install -m 755 @objroot@lib/libjemalloc@install_suffix@.$(SO).$(REV) $(LIBDIR)
+	ln -sf libjemalloc@install_suffix@.$(SO).$(REV) $(LIBDIR)/libjemalloc@install_suffix@.$(SO)
+	install -m 755 @objroot@lib/libjemalloc@install_suffix@_pic.a $(LIBDIR)
+	install -m 755 @objroot@lib/libjemalloc@install_suffix@.a $(LIBDIR)
+
+install_html:
+	install -d $(DATADIR)/doc/jemalloc@install_suffix@
+	@for d in $(DOCS_HTML); do \
+	echo "install -m 644 $$d $(DATADIR)/doc/jemalloc@install_suffix@"; \
+	install -m 644 $$d $(DATADIR)/doc/jemalloc@install_suffix@; \
+done
+
+install_man:
+	install -d $(MANDIR)/man3
+	@for d in $(DOCS_MAN3); do \
+	echo "install -m 644 $$d $(MANDIR)/man3"; \
+	install -m 644 $$d $(MANDIR)/man3; \
+done
+
+install_doc: install_html install_man
+
+install: install_bin install_include install_lib install_doc
+
+tests: $(CTESTS:@srcroot@%.c=@objroot@%)
+
+check: tests
+	@mkdir -p @objroot@test
+	@$(SHELL) -c 'total=0; \
+		failures=0; \
+		echo "========================================="; \
+		for t in $(CTESTS:@srcroot@%.c=@objroot@%); do \
+			total=`expr $$total + 1`; \
+			/bin/echo -n "$${t} ... "; \
+			$(TEST_LIBRARY_PATH) $${t} @abs_srcroot@ @abs_objroot@ \
+			  > @objroot@$${t}.out 2>&1; \
+			if test -e "@srcroot@$${t}.exp"; then \
+				diff -u @srcroot@$${t}.exp \
+				  @objroot@$${t}.out >/dev/null 2>&1; \
+				fail=$$?; \
+				if test "$${fail}" -eq "1" ; then \
+					failures=`expr $${failures} + 1`; \
+					echo "*** FAIL ***"; \
+				else \
+					echo "pass"; \
+				fi; \
+			else \
+				echo "*** FAIL *** (.exp file is missing)"; \
+				failures=`expr $${failures} + 1`; \
+			fi; \
+		done; \
+		echo "========================================="; \
+		echo "Failures: $${failures}/$${total}"'
+
+clean:
+	rm -f $(CSRCS:@srcroot@%.c=@objroot@%.o)
+	rm -f $(CSRCS:@srcroot@%.c=@objroot@%.pic.o)
+	rm -f $(CSRCS:@srcroot@%.c=@objroot@%.d)
+	rm -f $(CSRCS:@srcroot@%.c=@objroot@%.pic.d)
+	rm -f $(CTESTS:@srcroot@%.c=@objroot@%)
+	rm -f $(CTESTS:@srcroot@%.c=@objroot@%.o)
+	rm -f $(CTESTS:@srcroot@%.c=@objroot@%.d)
+	rm -f $(CTESTS:@srcroot@%.c=@objroot@%.out)
+	rm -f $(DSOS) $(STATIC_LIBS)
+
+distclean: clean
+	rm -rf @objroot@autom4te.cache
+	rm -f @objroot@config.log
+	rm -f @objroot@config.status
+	rm -f @objroot@config.stamp
+	rm -f @cfghdrs_out@
+	rm -f @cfgoutputs_out@
+
+relclean: distclean
+	rm -f @objroot@configure
+	rm -f @srcroot@VERSION
+	rm -f $(DOCS_HTML)
+	rm -f $(DOCS_MAN3)
+
+#===============================================================================
+# Re-configuration rules.
+
+ifeq (@enable_autogen@, 1)
+@srcroot@configure : @srcroot@configure.ac
+	cd ./@srcroot@ && @AUTOCONF@
+
+@objroot@config.status : @srcroot@configure
+	./@objroot@config.status --recheck
+
+@srcroot@config.stamp.in : @srcroot@configure.ac
+	echo stamp > @srcroot@config.stamp.in
+
+@objroot@config.stamp : @cfgoutputs_in@ @cfghdrs_in@ @srcroot@configure
+	./@objroot@config.status
+	@touch $@
+
+# There must be some action in order for make to re-read Makefile when it is
+# out of date.
+@cfgoutputs_out@ @cfghdrs_out@ : @objroot@config.stamp
+	@true
+endif

deps/jemalloc/README

@@ -0,0 +1,16 @@
+jemalloc is a general-purpose scalable concurrent malloc(3) implementation.
+This distribution is a stand-alone "portable" implementation that currently
+targets Linux and Apple OS X.  jemalloc is included as the default allocator in
+the FreeBSD and NetBSD operating systems, and it is used by the Mozilla Firefox
+web browser on Microsoft Windows-related platforms.  Depending on your needs,
+one of the other divergent versions may suit your needs better than this
+distribution.
+
+The COPYING file contains copyright and licensing information.
+
+The INSTALL file contains information on how to configure, build, and install
+jemalloc.
+
+The ChangeLog file contains a brief summary of changes for each release.
+
+URL: http://www.canonware.com/jemalloc/

deps/jemalloc/VERSION

@@ -0,0 +1 @@
+2.2.1-0-g5ef7abf6d846720fb3fb8c737861c99b5ad1d862

deps/jemalloc/autogen.sh

@@ -0,0 +1,17 @@
+#!/bin/sh
+
+for i in autoconf; do
+    echo "$i"
+    $i
+    if [ $? -ne 0 ]; then
+	echo "Error $? in $i"
+	exit 1
+    fi
+done
+
+echo "./configure --enable-autogen $@"
+./configure --enable-autogen $@
+if [ $? -ne 0 ]; then
+    echo "Error $? in ./configure"
+    exit 1
+fi

deps/jemalloc/configure.ac

@@ -0,0 +1,927 @@
+dnl Process this file with autoconf to produce a configure script.
+AC_INIT([Makefile.in])
+
+dnl ============================================================================
+dnl Custom macro definitions.
+
+dnl JE_CFLAGS_APPEND(cflag)
+AC_DEFUN([JE_CFLAGS_APPEND],
+[
+AC_MSG_CHECKING([whether compiler supports $1])
+TCFLAGS="${CFLAGS}"
+if test "x${CFLAGS}" = "x" ; then
+  CFLAGS="$1"
+else
+  CFLAGS="${CFLAGS} $1"
+fi
+AC_RUN_IFELSE([AC_LANG_PROGRAM(
+[[
+]], [[
+    return 0;
+]])],
+              AC_MSG_RESULT([yes]),
+              AC_MSG_RESULT([no])
+              [CFLAGS="${TCFLAGS}"]
+)
+])
+
+dnl JE_COMPILABLE(label, hcode, mcode, rvar)
+AC_DEFUN([JE_COMPILABLE],
+[
+AC_MSG_CHECKING([whether $1 is compilable])
+AC_RUN_IFELSE([AC_LANG_PROGRAM(
+[$2], [$3])],
+              AC_MSG_RESULT([yes])
+              [$4="yes"],
+              AC_MSG_RESULT([no])
+              [$4="no"]
+)
+])
+
+dnl ============================================================================
+
+srcroot=$srcdir
+if test "x${srcroot}" = "x." ; then
+  srcroot=""
+else
+  srcroot="${srcroot}/"
+fi
+AC_SUBST([srcroot])
+abs_srcroot="`cd \"${srcdir}\"; pwd`/"
+AC_SUBST([abs_srcroot])
+
+objroot=""
+AC_SUBST([objroot])
+abs_objroot="`pwd`/"
+AC_SUBST([abs_objroot])
+
+dnl Munge install path variables.
+if test "x$prefix" = "xNONE" ; then
+  prefix="/usr/local"
+fi
+if test "x$exec_prefix" = "xNONE" ; then
+  exec_prefix=$prefix
+fi
+PREFIX=$prefix
+AC_SUBST([PREFIX])
+BINDIR=`eval echo $bindir`
+BINDIR=`eval echo $BINDIR`
+AC_SUBST([BINDIR])
+INCLUDEDIR=`eval echo $includedir`
+INCLUDEDIR=`eval echo $INCLUDEDIR`
+AC_SUBST([INCLUDEDIR])
+LIBDIR=`eval echo $libdir`
+LIBDIR=`eval echo $LIBDIR`
+AC_SUBST([LIBDIR])
+DATADIR=`eval echo $datadir`
+DATADIR=`eval echo $DATADIR`
+AC_SUBST([DATADIR])
+MANDIR=`eval echo $mandir`
+MANDIR=`eval echo $MANDIR`
+AC_SUBST([MANDIR])
+
+dnl Support for building documentation.
+AC_PATH_PROG([XSLTPROC], [xsltproc], , [$PATH])
+AC_ARG_WITH([xslroot],
+  [AS_HELP_STRING([--with-xslroot=<path>], [XSL stylesheet root path])],
+if test "x$with_xslroot" = "xno" ; then
+  XSLROOT="/usr/share/xml/docbook/stylesheet/docbook-xsl"
+else
+  XSLROOT="${with_xslroot}"
+fi,
+  XSLROOT="/usr/share/xml/docbook/stylesheet/docbook-xsl"
+)
+AC_SUBST([XSLROOT])
+
+dnl If CFLAGS isn't defined, set CFLAGS to something reasonable.  Otherwise,
+dnl just prevent autoconf from molesting CFLAGS.
+CFLAGS=$CFLAGS
+AC_PROG_CC
+if test "x$CFLAGS" = "x" ; then
+  no_CFLAGS="yes"
+  if test "x$GCC" = "xyes" ; then
+    JE_CFLAGS_APPEND([-std=gnu99])
+    JE_CFLAGS_APPEND([-Wall])
+    JE_CFLAGS_APPEND([-pipe])
+    JE_CFLAGS_APPEND([-g3])
+  fi
+fi
+dnl Append EXTRA_CFLAGS to CFLAGS, if defined.
+if test "x$EXTRA_CFLAGS" != "x" ; then
+  JE_CFLAGS_APPEND([$EXTRA_CFLAGS])
+fi
+AC_PROG_CPP
+
+AC_CHECK_SIZEOF([void *])
+if test "x${ac_cv_sizeof_void_p}" = "x8" ; then
+  LG_SIZEOF_PTR=3
+elif test "x${ac_cv_sizeof_void_p}" = "x4" ; then
+  LG_SIZEOF_PTR=2
+else
+  AC_MSG_ERROR([Unsupported pointer size: ${ac_cv_sizeof_void_p}])
+fi
+AC_DEFINE_UNQUOTED([LG_SIZEOF_PTR], [$LG_SIZEOF_PTR])
+
+AC_CHECK_SIZEOF([int])
+if test "x${ac_cv_sizeof_int}" = "x8" ; then
+  LG_SIZEOF_INT=3
+elif test "x${ac_cv_sizeof_int}" = "x4" ; then
+  LG_SIZEOF_INT=2
+else
+  AC_MSG_ERROR([Unsupported int size: ${ac_cv_sizeof_int}])
+fi
+AC_DEFINE_UNQUOTED([LG_SIZEOF_INT], [$LG_SIZEOF_INT])
+
+AC_CHECK_SIZEOF([long])
+if test "x${ac_cv_sizeof_long}" = "x8" ; then
+  LG_SIZEOF_LONG=3
+elif test "x${ac_cv_sizeof_long}" = "x4" ; then
+  LG_SIZEOF_LONG=2
+else
+  AC_MSG_ERROR([Unsupported long size: ${ac_cv_sizeof_long}])
+fi
+AC_DEFINE_UNQUOTED([LG_SIZEOF_LONG], [$LG_SIZEOF_LONG])
+
+AC_CANONICAL_HOST
+dnl CPU-specific settings.
+CPU_SPINWAIT=""
+case "${host_cpu}" in
+  i[[345]]86)
+	;;
+  i686)
+	JE_COMPILABLE([__asm__], [], [[__asm__ volatile("pause"); return 0;]],
+	              [asm])
+	if test "x${asm}" = "xyes" ; then
+	    CPU_SPINWAIT='__asm__ volatile("pause")'
+	fi
+	;;
+  x86_64)
+	JE_COMPILABLE([__asm__ syntax], [],
+	              [[__asm__ volatile("pause"); return 0;]], [asm])
+	if test "x${asm}" = "xyes" ; then
+	    CPU_SPINWAIT='__asm__ volatile("pause")'
+	fi
+	;;
+  *)
+	;;
+esac
+AC_DEFINE_UNQUOTED([CPU_SPINWAIT], [$CPU_SPINWAIT])
+
+dnl Platform-specific settings.  abi and RPATH can probably be determined
+dnl programmatically, but doing so is error-prone, which makes it generally
+dnl not worth the trouble.
+dnl 
+dnl Define cpp macros in CPPFLAGS, rather than doing AC_DEFINE(macro), since the
+dnl definitions need to be seen before any headers are included, which is a pain
+dnl to make happen otherwise.
+case "${host}" in
+  *-*-darwin*)
+	CFLAGS="$CFLAGS -fno-common -no-cpp-precomp"
+	abi="macho"
+	AC_DEFINE([JEMALLOC_PURGE_MADVISE_FREE])
+	RPATH=""
+	;;
+  *-*-freebsd*)
+	CFLAGS="$CFLAGS"
+	abi="elf"
+	AC_DEFINE([JEMALLOC_PURGE_MADVISE_FREE])
+	RPATH="-Wl,-rpath,"
+	;;
+  *-*-linux*)
+	CFLAGS="$CFLAGS"
+	CPPFLAGS="$CPPFLAGS -D_GNU_SOURCE"
+	abi="elf"
+	AC_DEFINE([JEMALLOC_PURGE_MADVISE_DONTNEED])
+	RPATH="-Wl,-rpath,"
+	;;
+  *-*-netbsd*)
+	AC_MSG_CHECKING([ABI])
+        AC_COMPILE_IFELSE([AC_LANG_PROGRAM(
+[[#ifdef __ELF__
+/* ELF */
+#else
+#error aout
+#endif
+]])],
+                          [CFLAGS="$CFLAGS"; abi="elf"],
+                          [abi="aout"])
+	AC_MSG_RESULT([$abi])
+	AC_DEFINE([JEMALLOC_PURGE_MADVISE_FREE])
+	RPATH="-Wl,-rpath,"
+	;;
+  *-*-solaris2*)
+	CFLAGS="$CFLAGS"
+	abi="elf"
+	RPATH="-Wl,-R,"
+	dnl Solaris needs this for sigwait().
+	CPPFLAGS="$CPPFLAGS -D_POSIX_PTHREAD_SEMANTICS"
+	LIBS="$LIBS -lposix4 -lsocket -lnsl"
+	;;
+  *)
+	AC_MSG_RESULT([Unsupported operating system: ${host}])
+	abi="elf"
+	RPATH="-Wl,-rpath,"
+	;;
+esac
+AC_SUBST([abi])
+AC_SUBST([RPATH])
+
+JE_COMPILABLE([__attribute__ syntax],
+              [static __attribute__((unused)) void foo(void){}],
+              [],
+              [attribute])
+if test "x${attribute}" = "xyes" ; then
+  AC_DEFINE([JEMALLOC_HAVE_ATTR], [ ])
+  if test "x${GCC}" = "xyes" -a "x${abi}" = "xelf"; then
+    JE_CFLAGS_APPEND([-fvisibility=hidden])
+  fi
+fi
+
+JE_COMPILABLE([mremap(...MREMAP_FIXED...)], [
+#define _GNU_SOURCE
+#include <sys/mman.h>
+], [
+void *p = mremap((void *)0, 0, 0, MREMAP_MAYMOVE|MREMAP_FIXED, (void *)0);
+], [mremap_fixed])
+if test "x${mremap_fixed}" = "xyes" ; then
+  AC_DEFINE([JEMALLOC_MREMAP_FIXED])
+fi
+
+dnl Support optional additions to rpath.
+AC_ARG_WITH([rpath],
+  [AS_HELP_STRING([--with-rpath=<rpath>], [Colon-separated rpath (ELF systems only)])],
+if test "x$with_rpath" = "xno" ; then
+  RPATH_EXTRA=
+else
+  RPATH_EXTRA="`echo $with_rpath | tr \":\" \" \"`"
+fi,
+  RPATH_EXTRA=
+)
+AC_SUBST([RPATH_EXTRA])
+
+dnl Disable rules that do automatic regeneration of configure output by default.
+AC_ARG_ENABLE([autogen],
+  [AS_HELP_STRING([--enable-autogen], [Automatically regenerate configure output])],
+if test "x$enable_autogen" = "xno" ; then
+  enable_autogen="0"
+else
+  enable_autogen="1"
+fi
+,
+enable_autogen="0"
+)
+AC_SUBST([enable_autogen])
+
+AC_PROG_INSTALL
+AC_PROG_RANLIB
+AC_PATH_PROG([AR], [ar], , [$PATH])
+AC_PATH_PROG([LD], [ld], , [$PATH])
+AC_PATH_PROG([AUTOCONF], [autoconf], , [$PATH])
+
+dnl Do not prefix public APIs by default.
+AC_ARG_WITH([jemalloc_prefix],
+  [AS_HELP_STRING([--with-jemalloc-prefix=<prefix>], [Prefix to prepend to all public APIs])],
+  [JEMALLOC_PREFIX="$with_jemalloc_prefix"],
+  [if test "x$abi" != "xmacho" ; then
+  JEMALLOC_PREFIX=""
+else
+  JEMALLOC_PREFIX="je_"
+fi]
+)
+if test "x$JEMALLOC_PREFIX" != "x" ; then
+  JEMALLOC_CPREFIX=`echo ${JEMALLOC_PREFIX} | tr "a-z" "A-Z"`
+  AC_DEFINE_UNQUOTED([JEMALLOC_PREFIX], ["$JEMALLOC_PREFIX"])
+  AC_DEFINE_UNQUOTED([JEMALLOC_CPREFIX], ["$JEMALLOC_CPREFIX"])
+  jemalloc_prefix="$JEMALLOC_PREFIX"
+  jemalloc_cprefix="$JEMALLOC_CPREFIX"
+  AC_SUBST([jemalloc_prefix])
+  AC_SUBST([jemalloc_cprefix])
+  AC_DEFINE_UNQUOTED([JEMALLOC_P(string_that_no_one_should_want_to_use_as_a_jemalloc_API_prefix)], [${JEMALLOC_PREFIX}##string_that_no_one_should_want_to_use_as_a_jemalloc_API_prefix])
+fi
+
+dnl Do not add suffix to installed files by default.
+AC_ARG_WITH([install_suffix],
+  [AS_HELP_STRING([--with-install-suffix=<suffix>], [Suffix to append to all installed files])],
+  [INSTALL_SUFFIX="$with_install_suffix"],
+  [INSTALL_SUFFIX=]
+)
+install_suffix="$INSTALL_SUFFIX"
+AC_SUBST([install_suffix])
+
+cfgoutputs_in="${srcroot}Makefile.in"
+cfgoutputs_in="${cfgoutputs_in} ${srcroot}doc/html.xsl.in"
+cfgoutputs_in="${cfgoutputs_in} ${srcroot}doc/manpages.xsl.in"
+cfgoutputs_in="${cfgoutputs_in} ${srcroot}doc/jemalloc.xml.in"
+cfgoutputs_in="${cfgoutputs_in} ${srcroot}include/jemalloc/jemalloc.h.in"
+cfgoutputs_in="${cfgoutputs_in} ${srcroot}include/jemalloc/internal/jemalloc_internal.h.in"
+cfgoutputs_in="${cfgoutputs_in} ${srcroot}test/jemalloc_test.h.in"
+
+cfgoutputs_out="Makefile"
+cfgoutputs_out="${cfgoutputs_out} doc/html.xsl"
+cfgoutputs_out="${cfgoutputs_out} doc/manpages.xsl"
+cfgoutputs_out="${cfgoutputs_out} doc/jemalloc${install_suffix}.xml"
+cfgoutputs_out="${cfgoutputs_out} include/jemalloc/jemalloc${install_suffix}.h"
+cfgoutputs_out="${cfgoutputs_out} include/jemalloc/internal/jemalloc_internal.h"
+cfgoutputs_out="${cfgoutputs_out} test/jemalloc_test.h"
+
+cfgoutputs_tup="Makefile"
+cfgoutputs_tup="${cfgoutputs_tup} doc/html.xsl:doc/html.xsl.in"
+cfgoutputs_tup="${cfgoutputs_tup} doc/manpages.xsl:doc/manpages.xsl.in"
+cfgoutputs_tup="${cfgoutputs_tup} doc/jemalloc${install_suffix}.xml:doc/jemalloc.xml.in"
+cfgoutputs_tup="${cfgoutputs_tup} include/jemalloc/jemalloc${install_suffix}.h:include/jemalloc/jemalloc.h.in"
+cfgoutputs_tup="${cfgoutputs_tup} include/jemalloc/internal/jemalloc_internal.h"
+cfgoutputs_tup="${cfgoutputs_tup} test/jemalloc_test.h:test/jemalloc_test.h.in"
+
+cfghdrs_in="${srcroot}include/jemalloc/jemalloc_defs.h.in"
+
+cfghdrs_out="include/jemalloc/jemalloc_defs${install_suffix}.h"
+
+cfghdrs_tup="include/jemalloc/jemalloc_defs${install_suffix}.h:include/jemalloc/jemalloc_defs.h.in"
+
+dnl Do not silence irrelevant compiler warnings by default, since enabling this
+dnl option incurs a performance penalty.
+AC_ARG_ENABLE([cc-silence],
+  [AS_HELP_STRING([--enable-cc-silence],
+                  [Silence irrelevant compiler warnings])],
+[if test "x$enable_cc_silence" = "xno" ; then
+  enable_cc_silence="0"
+else
+  enable_cc_silence="1"
+fi
+],
+[enable_cc_silence="0"]
+)
+if test "x$enable_cc_silence" = "x1" ; then
+  AC_DEFINE([JEMALLOC_CC_SILENCE])
+fi
+
+dnl Do not compile with debugging by default.
+AC_ARG_ENABLE([debug],
+  [AS_HELP_STRING([--enable-debug], [Build debugging code])],
+[if test "x$enable_debug" = "xno" ; then
+  enable_debug="0"
+else
+  enable_debug="1"
+fi
+],
+[enable_debug="0"]
+)
+if test "x$enable_debug" = "x1" ; then
+  AC_DEFINE([JEMALLOC_DEBUG], [ ])
+  AC_DEFINE([JEMALLOC_IVSALLOC], [ ])
+fi
+AC_SUBST([enable_debug])
+
+dnl Only optimize if not debugging.
+if test "x$enable_debug" = "x0" -a "x$no_CFLAGS" = "xyes" ; then
+  dnl Make sure that an optimization flag was not specified in EXTRA_CFLAGS.
+  optimize="no"
+  echo "$EXTRA_CFLAGS" | grep "\-O" >/dev/null || optimize="yes"
+  if test "x${optimize}" = "xyes" ; then
+    if test "x$GCC" = "xyes" ; then
+      JE_CFLAGS_APPEND([-O3])
+      JE_CFLAGS_APPEND([-funroll-loops])
+    else
+      JE_CFLAGS_APPEND([-O])
+    fi
+  fi
+fi
+
+dnl Do not enable statistics calculation by default.
+AC_ARG_ENABLE([stats],
+  [AS_HELP_STRING([--enable-stats], [Enable statistics calculation/reporting])],
+[if test "x$enable_stats" = "xno" ; then
+  enable_stats="0"
+else
+  enable_stats="1"
+fi
+],
+[enable_stats="0"]
+)
+if test "x$enable_stats" = "x1" ; then
+  AC_DEFINE([JEMALLOC_STATS], [ ])
+fi
+AC_SUBST([enable_stats])
+
+dnl Do not enable profiling by default.
+AC_ARG_ENABLE([prof],
+  [AS_HELP_STRING([--enable-prof], [Enable allocation profiling])],
+[if test "x$enable_prof" = "xno" ; then
+  enable_prof="0"
+else
+  enable_prof="1"
+fi
+],
+[enable_prof="0"]
+)
+if test "x$enable_prof" = "x1" ; then
+  backtrace_method=""
+else
+  backtrace_method="N/A"
+fi
+
+AC_ARG_ENABLE([prof-libunwind],
+  [AS_HELP_STRING([--enable-prof-libunwind], [Use libunwind for backtracing])],
+[if test "x$enable_prof_libunwind" = "xno" ; then
+  enable_prof_libunwind="0"
+else
+  enable_prof_libunwind="1"
+fi
+],
+[enable_prof_libunwind="0"]
+)
+AC_ARG_WITH([static_libunwind],
+  [AS_HELP_STRING([--with-static-libunwind=<libunwind.a>],
+  [Path to static libunwind library; use rather than dynamically linking])],
+if test "x$with_static_libunwind" = "xno" ; then
+  LUNWIND="-lunwind"
+else
+  if test ! -f "$with_static_libunwind" ; then
+    AC_MSG_ERROR([Static libunwind not found: $with_static_libunwind])
+  fi
+  LUNWIND="$with_static_libunwind"
+fi,
+  LUNWIND="-lunwind"
+)
+if test "x$backtrace_method" = "x" -a "x$enable_prof_libunwind" = "x1" ; then
+  AC_CHECK_HEADERS([libunwind.h], , [enable_prof_libunwind="0"])
+  if test "x$LUNWIND" = "x-lunwind" ; then
+    AC_CHECK_LIB([unwind], [backtrace], [LIBS="$LIBS $LUNWIND"],
+                 [enable_prof_libunwind="0"])
+  else
+    LIBS="$LIBS $LUNWIND"
+  fi
+  if test "x${enable_prof_libunwind}" = "x1" ; then
+    backtrace_method="libunwind"
+    AC_DEFINE([JEMALLOC_PROF_LIBUNWIND], [ ])
+  fi
+fi
+
+AC_ARG_ENABLE([prof-libgcc],
+  [AS_HELP_STRING([--disable-prof-libgcc],
+  [Do not use libgcc for backtracing])],
+[if test "x$enable_prof_libgcc" = "xno" ; then
+  enable_prof_libgcc="0"
+else
+  enable_prof_libgcc="1"
+fi
+],
+[enable_prof_libgcc="1"]
+)
+if test "x$backtrace_method" = "x" -a "x$enable_prof_libgcc" = "x1" \
+     -a "x$GCC" = "xyes" ; then
+  AC_CHECK_HEADERS([unwind.h], , [enable_prof_libgcc="0"])
+  AC_CHECK_LIB([gcc], [_Unwind_Backtrace], [LIBS="$LIBS -lgcc"], [enable_prof_libgcc="0"])
+  dnl The following is conservative, in that it only has entries for CPUs on
+  dnl which jemalloc has been tested.
+  AC_MSG_CHECKING([libgcc-based backtracing reliability on ${host_cpu}])
+  case "${host_cpu}" in
+    i[[3456]]86)
+      AC_MSG_RESULT([unreliable])
+      enable_prof_libgcc="0";
+      ;;
+    x86_64)
+      AC_MSG_RESULT([reliable])
+      ;;
+    *)
+      AC_MSG_RESULT([unreliable])
+      enable_prof_libgcc="0";
+      ;;
+  esac
+  if test "x${enable_prof_libgcc}" = "x1" ; then
+    backtrace_method="libgcc"
+    AC_DEFINE([JEMALLOC_PROF_LIBGCC], [ ])
+  fi
+else
+  enable_prof_libgcc="0"
+fi
+
+AC_ARG_ENABLE([prof-gcc],
+  [AS_HELP_STRING([--disable-prof-gcc],
+  [Do not use gcc intrinsics for backtracing])],
+[if test "x$enable_prof_gcc" = "xno" ; then
+  enable_prof_gcc="0"
+else
+  enable_prof_gcc="1"
+fi
+],
+[enable_prof_gcc="1"]
+)
+if test "x$backtrace_method" = "x" -a "x$enable_prof_gcc" = "x1" \
+     -a "x$GCC" = "xyes" ; then
+  backtrace_method="gcc intrinsics"
+  AC_DEFINE([JEMALLOC_PROF_GCC], [ ])
+else
+  enable_prof_gcc="0"
+fi
+
+if test "x$backtrace_method" = "x" ; then
+  backtrace_method="none (disabling profiling)"
+  enable_prof="0"
+fi
+AC_MSG_CHECKING([configured backtracing method])
+AC_MSG_RESULT([$backtrace_method])
+if test "x$enable_prof" = "x1" ; then
+  LIBS="$LIBS -lm"
+  AC_DEFINE([JEMALLOC_PROF], [ ])
+fi
+AC_SUBST([enable_prof])
+
+dnl Enable tiny allocations by default.
+AC_ARG_ENABLE([tiny],
+  [AS_HELP_STRING([--disable-tiny], [Disable tiny (sub-quantum) allocations])],
+[if test "x$enable_tiny" = "xno" ; then
+  enable_tiny="0"
+else
+  enable_tiny="1"
+fi
+],
+[enable_tiny="1"]
+)
+if test "x$enable_tiny" = "x1" ; then
+  AC_DEFINE([JEMALLOC_TINY], [ ])
+fi
+AC_SUBST([enable_tiny])
+
+dnl Enable thread-specific caching by default.
+AC_ARG_ENABLE([tcache],
+  [AS_HELP_STRING([--disable-tcache], [Disable per thread caches])],
+[if test "x$enable_tcache" = "xno" ; then
+  enable_tcache="0"
+else
+  enable_tcache="1"
+fi
+],
+[enable_tcache="1"]
+)
+if test "x$enable_tcache" = "x1" ; then
+  AC_DEFINE([JEMALLOC_TCACHE], [ ])
+fi
+AC_SUBST([enable_tcache])
+
+dnl Do not enable mmap()ped swap files by default.
+AC_ARG_ENABLE([swap],
+  [AS_HELP_STRING([--enable-swap], [Enable mmap()ped swap files])],
+[if test "x$enable_swap" = "xno" ; then
+  enable_swap="0"
+else
+  enable_swap="1"
+fi
+],
+[enable_swap="0"]
+)
+if test "x$enable_swap" = "x1" ; then
+  AC_DEFINE([JEMALLOC_SWAP], [ ])
+fi
+AC_SUBST([enable_swap])
+
+dnl Do not enable allocation from DSS by default.
+AC_ARG_ENABLE([dss],
+  [AS_HELP_STRING([--enable-dss], [Enable allocation from DSS])],
+[if test "x$enable_dss" = "xno" ; then
+  enable_dss="0"
+else
+  enable_dss="1"
+fi
+],
+[enable_dss="0"]
+)
+if test "x$enable_dss" = "x1" ; then
+  AC_DEFINE([JEMALLOC_DSS], [ ])
+fi
+AC_SUBST([enable_dss])
+
+dnl Do not support the junk/zero filling option by default.
+AC_ARG_ENABLE([fill],
+  [AS_HELP_STRING([--enable-fill], [Support junk/zero filling option])],
+[if test "x$enable_fill" = "xno" ; then
+  enable_fill="0"
+else
+  enable_fill="1"
+fi
+],
+[enable_fill="0"]
+)
+if test "x$enable_fill" = "x1" ; then
+  AC_DEFINE([JEMALLOC_FILL], [ ])
+fi
+AC_SUBST([enable_fill])
+
+dnl Do not support the xmalloc option by default.
+AC_ARG_ENABLE([xmalloc],
+  [AS_HELP_STRING([--enable-xmalloc], [Support xmalloc option])],
+[if test "x$enable_xmalloc" = "xno" ; then
+  enable_xmalloc="0"
+else
+  enable_xmalloc="1"
+fi
+],
+[enable_xmalloc="0"]
+)
+if test "x$enable_xmalloc" = "x1" ; then
+  AC_DEFINE([JEMALLOC_XMALLOC], [ ])
+fi
+AC_SUBST([enable_xmalloc])
+
+dnl Do not support the SYSV option by default.
+AC_ARG_ENABLE([sysv],
+  [AS_HELP_STRING([--enable-sysv], [Support SYSV semantics option])],
+[if test "x$enable_sysv" = "xno" ; then
+  enable_sysv="0"
+else
+  enable_sysv="1"
+fi
+],
+[enable_sysv="0"]
+)
+if test "x$enable_sysv" = "x1" ; then
+  AC_DEFINE([JEMALLOC_SYSV], [ ])
+fi
+AC_SUBST([enable_sysv])
+
+dnl Do not determine page shift at run time by default.
+AC_ARG_ENABLE([dynamic_page_shift],
+  [AS_HELP_STRING([--enable-dynamic-page-shift],
+  [Determine page size at run time (don't trust configure result)])],
+[if test "x$enable_dynamic_page_shift" = "xno" ; then
+  enable_dynamic_page_shift="0"
+else
+  enable_dynamic_page_shift="1"
+fi
+],
+[enable_dynamic_page_shift="0"]
+)
+if test "x$enable_dynamic_page_shift" = "x1" ; then
+  AC_DEFINE([DYNAMIC_PAGE_SHIFT], [ ])
+fi
+AC_SUBST([enable_dynamic_page_shift])
+
+AC_MSG_CHECKING([STATIC_PAGE_SHIFT])
+AC_RUN_IFELSE([AC_LANG_PROGRAM(
+[[#include <stdio.h>
+#include <unistd.h>
+#include <strings.h>
+]], [[
+    long result;
+    FILE *f;
+
+    result = sysconf(_SC_PAGESIZE);
+    if (result == -1) {
+	return 1;
+    }
+    f = fopen("conftest.out", "w");
+    if (f == NULL) {
+	return 1;
+    }
+    fprintf(f, "%u\n", ffs((int)result) - 1);
+    close(f);
+
+    return 0;
+]])],
+              [STATIC_PAGE_SHIFT=`cat conftest.out`]
+              AC_MSG_RESULT([$STATIC_PAGE_SHIFT])
+              AC_DEFINE_UNQUOTED([STATIC_PAGE_SHIFT], [$STATIC_PAGE_SHIFT]),
+              AC_MSG_RESULT([error]))
+
+dnl ============================================================================
+dnl jemalloc configuration.
+dnl 
+
+dnl Set VERSION if source directory has an embedded git repository.
+if test -d "${srcroot}../.git" ; then
+  git describe --long --abbrev=40 > ${srcroot}VERSION
+fi
+jemalloc_version=`cat ${srcroot}VERSION`
+jemalloc_version_major=`echo ${jemalloc_version} | tr ".g-" " " | awk '{print [$]1}'`
+jemalloc_version_minor=`echo ${jemalloc_version} | tr ".g-" " " | awk '{print [$]2}'`
+jemalloc_version_bugfix=`echo ${jemalloc_version} | tr ".g-" " " | awk '{print [$]3}'`
+jemalloc_version_nrev=`echo ${jemalloc_version} | tr ".g-" " " | awk '{print [$]4}'`
+jemalloc_version_gid=`echo ${jemalloc_version} | tr ".g-" " " | awk '{print [$]5}'`
+AC_SUBST([jemalloc_version])
+AC_SUBST([jemalloc_version_major])
+AC_SUBST([jemalloc_version_minor])
+AC_SUBST([jemalloc_version_bugfix])
+AC_SUBST([jemalloc_version_nrev])
+AC_SUBST([jemalloc_version_gid])
+
+dnl ============================================================================
+dnl Configure pthreads.
+
+AC_CHECK_HEADERS([pthread.h], , [AC_MSG_ERROR([pthread.h is missing])])
+AC_CHECK_LIB([pthread], [pthread_create], [LIBS="$LIBS -lpthread"],
+             [AC_MSG_ERROR([libpthread is missing])])
+
+CPPFLAGS="$CPPFLAGS -D_REENTRANT"
+
+dnl Enable lazy locking by default.
+AC_ARG_ENABLE([lazy_lock],
+  [AS_HELP_STRING([--disable-lazy-lock],
+  [Disable lazy locking (always lock, even when single-threaded)])],
+[if test "x$enable_lazy_lock" = "xno" ; then
+  enable_lazy_lock="0"
+else
+  enable_lazy_lock="1"
+fi
+],
+[enable_lazy_lock="1"]
+)
+if test "x$enable_lazy_lock" = "x1" ; then
+  AC_CHECK_HEADERS([dlfcn.h], , [AC_MSG_ERROR([dlfcn.h is missing])])
+  AC_CHECK_LIB([dl], [dlopen], [LIBS="$LIBS -ldl"],
+               [AC_MSG_ERROR([libdl is missing])])
+  AC_DEFINE([JEMALLOC_LAZY_LOCK], [ ])
+fi
+AC_SUBST([enable_lazy_lock])
+
+AC_ARG_ENABLE([tls],
+  [AS_HELP_STRING([--disable-tls], [Disable thread-local storage (__thread keyword)])],
+if test "x$enable_tls" = "xno" ; then
+  enable_tls="0"
+else
+  enable_tls="1"
+fi
+,
+enable_tls="1"
+)
+if test "x${enable_tls}" = "x1" ; then
+AC_MSG_CHECKING([for TLS])
+AC_RUN_IFELSE([AC_LANG_PROGRAM(
+[[
+    __thread int x;
+]], [[
+    x = 42;
+
+    return 0;
+]])],
+              AC_MSG_RESULT([yes]),
+              AC_MSG_RESULT([no])
+              enable_tls="0")
+fi
+AC_SUBST([enable_tls])
+if test "x${enable_tls}" = "x0" ; then
+  AC_DEFINE_UNQUOTED([NO_TLS], [ ])
+fi
+
+dnl ============================================================================
+dnl Check for ffsl(3), and fail if not found.  This function exists on all
+dnl platforms that jemalloc currently has a chance of functioning on without
+dnl modification.
+
+AC_CHECK_FUNC([ffsl], [],
+	      [AC_MSG_ERROR([Cannot build without ffsl(3)])])
+
+dnl ============================================================================
+dnl Check for atomic(3) operations as provided on Darwin.
+
+JE_COMPILABLE([Darwin OSAtomic*()], [
+#include <libkern/OSAtomic.h>
+#include <inttypes.h>
+], [
+	{
+		int32_t x32 = 0;
+		volatile int32_t *x32p = &x32;
+		OSAtomicAdd32(1, x32p);
+	}
+	{
+		int64_t x64 = 0;
+		volatile int64_t *x64p = &x64;
+		OSAtomicAdd64(1, x64p);
+	}
+], [osatomic])
+if test "x${osatomic}" = "xyes" ; then
+  AC_DEFINE([JEMALLOC_OSATOMIC])
+fi
+
+dnl ============================================================================
+dnl Check for spinlock(3) operations as provided on Darwin.
+
+JE_COMPILABLE([Darwin OSSpin*()], [
+#include <libkern/OSAtomic.h>
+#include <inttypes.h>
+], [
+	OSSpinLock lock = 0;
+	OSSpinLockLock(&lock);
+	OSSpinLockUnlock(&lock);
+], [osspin])
+if test "x${osspin}" = "xyes" ; then
+  AC_DEFINE([JEMALLOC_OSSPIN])
+fi
+
+dnl ============================================================================
+dnl Check for allocator-related functions that should be wrapped.
+
+AC_CHECK_FUNC([memalign],
+	      [AC_DEFINE([JEMALLOC_OVERRIDE_MEMALIGN])])
+AC_CHECK_FUNC([valloc],
+	      [AC_DEFINE([JEMALLOC_OVERRIDE_VALLOC])])
+
+dnl ============================================================================
+dnl Darwin-related configuration.
+
+if test "x${abi}" = "xmacho" ; then
+  AC_DEFINE([JEMALLOC_IVSALLOC])
+  AC_DEFINE([JEMALLOC_ZONE])
+
+  dnl The szone version jumped from 3 to 6 between the OS X 10.5.x and 10.6
+  dnl releases.  malloc_zone_t and malloc_introspection_t have new fields in
+  dnl 10.6, which is the only source-level indication of the change.
+  AC_MSG_CHECKING([malloc zone version])
+  AC_TRY_COMPILE([#include <stdlib.h>
+#include <malloc/malloc.h>], [
+	static malloc_zone_t zone;
+	static struct malloc_introspection_t zone_introspect;
+
+	zone.size = NULL;
+	zone.malloc = NULL;
+	zone.calloc = NULL;
+	zone.valloc = NULL;
+	zone.free = NULL;
+	zone.realloc = NULL;
+	zone.destroy = NULL;
+	zone.zone_name = "jemalloc_zone";
+	zone.batch_malloc = NULL;
+	zone.batch_free = NULL;
+	zone.introspect = &zone_introspect;
+	zone.version = 6;
+	zone.memalign = NULL;
+	zone.free_definite_size = NULL;
+
+	zone_introspect.enumerator = NULL;
+	zone_introspect.good_size = NULL;
+	zone_introspect.check = NULL;
+	zone_introspect.print = NULL;
+	zone_introspect.log = NULL;
+	zone_introspect.force_lock = NULL;
+	zone_introspect.force_unlock = NULL;
+	zone_introspect.statistics = NULL;
+	zone_introspect.zone_locked = NULL;
+], [AC_DEFINE_UNQUOTED([JEMALLOC_ZONE_VERSION], [6])
+    AC_MSG_RESULT([6])],
+   [AC_DEFINE_UNQUOTED([JEMALLOC_ZONE_VERSION], [3])
+   AC_MSG_RESULT([3])])
+fi
+
+dnl ============================================================================
+dnl Check for typedefs, structures, and compiler characteristics.
+AC_HEADER_STDBOOL
+
+dnl Process .in files.
+AC_SUBST([cfghdrs_in])
+AC_SUBST([cfghdrs_out])
+AC_CONFIG_HEADERS([$cfghdrs_tup])
+
+dnl ============================================================================
+dnl Generate outputs.
+AC_CONFIG_FILES([$cfgoutputs_tup config.stamp])
+AC_SUBST([cfgoutputs_in])
+AC_SUBST([cfgoutputs_out])
+AC_OUTPUT
+
+dnl ============================================================================
+dnl Print out the results of configuration.
+AC_MSG_RESULT([===============================================================================])
+AC_MSG_RESULT([jemalloc version   : $jemalloc_version])
+AC_MSG_RESULT([])
+AC_MSG_RESULT([CC                 : ${CC}])
+AC_MSG_RESULT([CPPFLAGS           : ${CPPFLAGS}])
+AC_MSG_RESULT([CFLAGS             : ${CFLAGS}])
+AC_MSG_RESULT([LDFLAGS            : ${LDFLAGS}])
+AC_MSG_RESULT([LIBS               : ${LIBS}])
+AC_MSG_RESULT([RPATH_EXTRA        : ${RPATH_EXTRA}])
+AC_MSG_RESULT([])
+AC_MSG_RESULT([XSLTPROC           : ${XSLTPROC}])
+AC_MSG_RESULT([XSLROOT            : ${XSLROOT}])
+AC_MSG_RESULT([])
+AC_MSG_RESULT([PREFIX             : ${PREFIX}])
+AC_MSG_RESULT([BINDIR             : ${BINDIR}])
+AC_MSG_RESULT([INCLUDEDIR         : ${INCLUDEDIR}])
+AC_MSG_RESULT([LIBDIR             : ${LIBDIR}])
+AC_MSG_RESULT([DATADIR            : ${DATADIR}])
+AC_MSG_RESULT([MANDIR             : ${MANDIR}])
+AC_MSG_RESULT([])
+AC_MSG_RESULT([srcroot            : ${srcroot}])
+AC_MSG_RESULT([abs_srcroot        : ${abs_srcroot}])
+AC_MSG_RESULT([objroot            : ${objroot}])
+AC_MSG_RESULT([abs_objroot        : ${abs_objroot}])
+AC_MSG_RESULT([])
+AC_MSG_RESULT([JEMALLOC_PREFIX    : ${JEMALLOC_PREFIX}])
+AC_MSG_RESULT([install_suffix     : ${install_suffix}])
+AC_MSG_RESULT([autogen            : ${enable_autogen}])
+AC_MSG_RESULT([cc-silence         : ${enable_cc_silence}])
+AC_MSG_RESULT([debug              : ${enable_debug}])
+AC_MSG_RESULT([stats              : ${enable_stats}])
+AC_MSG_RESULT([prof               : ${enable_prof}])
+AC_MSG_RESULT([prof-libunwind     : ${enable_prof_libunwind}])
+AC_MSG_RESULT([prof-libgcc        : ${enable_prof_libgcc}])
+AC_MSG_RESULT([prof-gcc           : ${enable_prof_gcc}])
+AC_MSG_RESULT([tiny               : ${enable_tiny}])
+AC_MSG_RESULT([tcache             : ${enable_tcache}])
+AC_MSG_RESULT([fill               : ${enable_fill}])
+AC_MSG_RESULT([xmalloc            : ${enable_xmalloc}])
+AC_MSG_RESULT([sysv               : ${enable_sysv}])
+AC_MSG_RESULT([swap               : ${enable_swap}])
+AC_MSG_RESULT([dss                : ${enable_dss}])
+AC_MSG_RESULT([dynamic_page_shift : ${enable_dynamic_page_shift}])
+AC_MSG_RESULT([lazy_lock          : ${enable_lazy_lock}])
+AC_MSG_RESULT([tls                : ${enable_tls}])
+AC_MSG_RESULT([===============================================================================])

deps/jemalloc/doc/html.xsl.in

@@ -0,0 +1,4 @@
+<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:stylesheet>

deps/jemalloc/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>

deps/jemalloc/doc/stylesheet.xsl

@@ -0,0 +1,7 @@
+<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">
+  <xsl:param name="funcsynopsis.style">ansi</xsl:param>
+  <xsl:param name="function.parens" select="1"/>
+  <xsl:template match="mallctl">
+    "<xsl:call-template name="inline.monoseq"/>"
+  </xsl:template>
+</xsl:stylesheet>

deps/jemalloc/include/jemalloc/internal/arena.h

@@ -0,0 +1,743 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+/*
+ * Subpages are an artificially designated partitioning of pages.  Their only
+ * purpose is to support subpage-spaced size classes.
+ *
+ * There must be at least 4 subpages per page, due to the way size classes are
+ * handled.
+ */
+#define	LG_SUBPAGE		8
+#define	SUBPAGE			((size_t)(1U << LG_SUBPAGE))
+#define	SUBPAGE_MASK		(SUBPAGE - 1)
+
+/* Return the smallest subpage multiple that is >= s. */
+#define	SUBPAGE_CEILING(s)						\
+	(((s) + SUBPAGE_MASK) & ~SUBPAGE_MASK)
+
+#ifdef JEMALLOC_TINY
+   /* Smallest size class to support. */
+#  define LG_TINY_MIN		LG_SIZEOF_PTR
+#  define TINY_MIN		(1U << LG_TINY_MIN)
+#endif
+
+/*
+ * Maximum size class that is a multiple of the quantum, but not (necessarily)
+ * a power of 2.  Above this size, allocations are rounded up to the nearest
+ * power of 2.
+ */
+#define	LG_QSPACE_MAX_DEFAULT	7
+
+/*
+ * Maximum size class that is a multiple of the cacheline, but not (necessarily)
+ * a power of 2.  Above this size, allocations are rounded up to the nearest
+ * power of 2.
+ */
+#define	LG_CSPACE_MAX_DEFAULT	9
+
+/*
+ * RUN_MAX_OVRHD indicates maximum desired run header overhead.  Runs are sized
+ * as small as possible such that this setting is still honored, without
+ * violating other constraints.  The goal is to make runs as small as possible
+ * without exceeding a per run external fragmentation threshold.
+ *
+ * We use binary fixed point math for overhead computations, where the binary
+ * point is implicitly RUN_BFP bits to the left.
+ *
+ * Note that it is possible to set RUN_MAX_OVRHD low enough that it cannot be
+ * honored for some/all object sizes, since when heap profiling is enabled
+ * there is one pointer of header overhead per object (plus a constant).  This
+ * constraint is relaxed (ignored) for runs that are so small that the
+ * per-region overhead is greater than:
+ *
+ *   (RUN_MAX_OVRHD / (reg_size << (3+RUN_BFP))
+ */
+#define	RUN_BFP			12
+/*                                    \/   Implicit binary fixed point. */
+#define	RUN_MAX_OVRHD		0x0000003dU
+#define	RUN_MAX_OVRHD_RELAX	0x00001800U
+
+/* Maximum number of regions in one run. */
+#define	LG_RUN_MAXREGS		11
+#define	RUN_MAXREGS		(1U << LG_RUN_MAXREGS)
+
+/*
+ * The minimum ratio of active:dirty pages per arena is computed as:
+ *
+ *   (nactive >> opt_lg_dirty_mult) >= ndirty
+ *
+ * So, supposing that opt_lg_dirty_mult is 5, there can be no less than 32
+ * times as many active pages as dirty pages.
+ */
+#define	LG_DIRTY_MULT_DEFAULT	5
+
+typedef struct arena_chunk_map_s arena_chunk_map_t;
+typedef struct arena_chunk_s arena_chunk_t;
+typedef struct arena_run_s arena_run_t;
+typedef struct arena_bin_info_s arena_bin_info_t;
+typedef struct arena_bin_s arena_bin_t;
+typedef struct arena_s arena_t;
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+/* Each element of the chunk map corresponds to one page within the chunk. */
+struct arena_chunk_map_s {
+	union {
+		/*
+		 * Linkage for run trees.  There are two disjoint uses:
+		 *
+		 * 1) arena_t's runs_avail_{clean,dirty} trees.
+		 * 2) arena_run_t conceptually uses this linkage for in-use
+		 *    non-full runs, rather than directly embedding linkage.
+		 */
+		rb_node(arena_chunk_map_t)	rb_link;
+		/*
+		 * List of runs currently in purgatory.  arena_chunk_purge()
+		 * temporarily allocates runs that contain dirty pages while
+		 * purging, so that other threads cannot use the runs while the
+		 * purging thread is operating without the arena lock held.
+		 */
+		ql_elm(arena_chunk_map_t)	ql_link;
+	}				u;
+
+#ifdef JEMALLOC_PROF
+	/* Profile counters, used for large object runs. */
+	prof_ctx_t			*prof_ctx;
+#endif
+
+	/*
+	 * Run address (or size) and various flags are stored together.  The bit
+	 * layout looks like (assuming 32-bit system):
+	 *
+	 *   ???????? ???????? ????---- ----dula
+	 *
+	 * ? : Unallocated: Run address for first/last pages, unset for internal
+	 *                  pages.
+	 *     Small: Run page offset.
+	 *     Large: Run size for first page, unset for trailing pages.
+	 * - : Unused.
+	 * d : dirty?
+	 * u : unzeroed?
+	 * l : large?
+	 * a : allocated?
+	 *
+	 * Following are example bit patterns for the three types of runs.
+	 *
+	 * p : run page offset
+	 * s : run size
+	 * c : (binind+1) for size class (used only if prof_promote is true)
+	 * x : don't care
+	 * - : 0
+	 * + : 1
+	 * [DULA] : bit set
+	 * [dula] : bit unset
+	 *
+	 *   Unallocated (clean):
+	 *     ssssssss ssssssss ssss---- ----du-a
+	 *     xxxxxxxx xxxxxxxx xxxx---- -----Uxx
+	 *     ssssssss ssssssss ssss---- ----dU-a
+	 *
+	 *   Unallocated (dirty):
+	 *     ssssssss ssssssss ssss---- ----D--a
+	 *     xxxxxxxx xxxxxxxx xxxx---- ----xxxx
+	 *     ssssssss ssssssss ssss---- ----D--a
+	 *
+	 *   Small:
+	 *     pppppppp pppppppp pppp---- ----d--A
+	 *     pppppppp pppppppp pppp---- -------A
+	 *     pppppppp pppppppp pppp---- ----d--A
+	 *
+	 *   Large:
+	 *     ssssssss ssssssss ssss---- ----D-LA
+	 *     xxxxxxxx xxxxxxxx xxxx---- ----xxxx
+	 *     -------- -------- -------- ----D-LA
+	 *
+	 *   Large (sampled, size <= PAGE_SIZE):
+	 *     ssssssss ssssssss sssscccc ccccD-LA
+	 *
+	 *   Large (not sampled, size == PAGE_SIZE):
+	 *     ssssssss ssssssss ssss---- ----D-LA
+	 */
+	size_t				bits;
+#ifdef JEMALLOC_PROF
+#define	CHUNK_MAP_CLASS_SHIFT	4
+#define	CHUNK_MAP_CLASS_MASK	((size_t)0xff0U)
+#endif
+#define	CHUNK_MAP_FLAGS_MASK	((size_t)0xfU)
+#define	CHUNK_MAP_DIRTY		((size_t)0x8U)
+#define	CHUNK_MAP_UNZEROED	((size_t)0x4U)
+#define	CHUNK_MAP_LARGE		((size_t)0x2U)
+#define	CHUNK_MAP_ALLOCATED	((size_t)0x1U)
+#define	CHUNK_MAP_KEY		CHUNK_MAP_ALLOCATED
+};
+typedef rb_tree(arena_chunk_map_t) arena_avail_tree_t;
+typedef rb_tree(arena_chunk_map_t) arena_run_tree_t;
+
+/* Arena chunk header. */
+struct arena_chunk_s {
+	/* Arena that owns the chunk. */
+	arena_t		*arena;
+
+	/* Linkage for the arena's chunks_dirty list. */
+	ql_elm(arena_chunk_t) link_dirty;
+
+	/*
+	 * True if the chunk is currently in the chunks_dirty list, due to
+	 * having at some point contained one or more dirty pages.  Removal
+	 * from chunks_dirty is lazy, so (dirtied && ndirty == 0) is possible.
+	 */
+	bool		dirtied;
+
+	/* Number of dirty pages. */
+	size_t		ndirty;
+
+	/*
+	 * Map of pages within chunk that keeps track of free/large/small.  The
+	 * first map_bias entries are omitted, since the chunk header does not
+	 * need to be tracked in the map.  This omission saves a header page
+	 * for common chunk sizes (e.g. 4 MiB).
+	 */
+	arena_chunk_map_t map[1]; /* Dynamically sized. */
+};
+typedef rb_tree(arena_chunk_t) arena_chunk_tree_t;
+
+struct arena_run_s {
+#ifdef JEMALLOC_DEBUG
+	uint32_t	magic;
+#  define ARENA_RUN_MAGIC 0x384adf93
+#endif
+
+	/* Bin this run is associated with. */
+	arena_bin_t	*bin;
+
+	/* Index of next region that has never been allocated, or nregs. */
+	uint32_t	nextind;
+
+	/* Number of free regions in run. */
+	unsigned	nfree;
+};
+
+/*
+ * Read-only information associated with each element of arena_t's bins array
+ * is stored separately, partly to reduce memory usage (only one copy, rather
+ * than one per arena), but mainly to avoid false cacheline sharing.
+ */
+struct arena_bin_info_s {
+	/* Size of regions in a run for this bin's size class. */
+	size_t		reg_size;
+
+	/* Total size of a run for this bin's size class. */
+	size_t		run_size;
+
+	/* Total number of regions in a run for this bin's size class. */
+	uint32_t	nregs;
+
+	/*
+	 * Offset of first bitmap_t element in a run header for this bin's size
+	 * class.
+	 */
+	uint32_t	bitmap_offset;
+
+	/*
+	 * Metadata used to manipulate bitmaps for runs associated with this
+	 * bin.
+	 */
+	bitmap_info_t	bitmap_info;
+
+#ifdef JEMALLOC_PROF
+	/*
+	 * Offset of first (prof_ctx_t *) in a run header for this bin's size
+	 * class, or 0 if (opt_prof == false).
+	 */
+	uint32_t	ctx0_offset;
+#endif
+
+	/* Offset of first region in a run for this bin's size class. */
+	uint32_t	reg0_offset;
+};
+
+struct arena_bin_s {
+	/*
+	 * All operations on runcur, runs, and stats require that lock be
+	 * locked.  Run allocation/deallocation are protected by the arena lock,
+	 * which may be acquired while holding one or more bin locks, but not
+	 * vise versa.
+	 */
+	malloc_mutex_t	lock;
+
+	/*
+	 * Current run being used to service allocations of this bin's size
+	 * class.
+	 */
+	arena_run_t	*runcur;
+
+	/*
+	 * Tree of non-full runs.  This tree is used when looking for an
+	 * existing run when runcur is no longer usable.  We choose the
+	 * non-full run that is lowest in memory; this policy tends to keep
+	 * objects packed well, and it can also help reduce the number of
+	 * almost-empty chunks.
+	 */
+	arena_run_tree_t runs;
+
+#ifdef JEMALLOC_STATS
+	/* Bin statistics. */
+	malloc_bin_stats_t stats;
+#endif
+};
+
+struct arena_s {
+#ifdef JEMALLOC_DEBUG
+	uint32_t		magic;
+#  define ARENA_MAGIC 0x947d3d24
+#endif
+
+	/* This arena's index within the arenas array. */
+	unsigned		ind;
+
+	/*
+	 * Number of threads currently assigned to this arena.  This field is
+	 * protected by arenas_lock.
+	 */
+	unsigned		nthreads;
+
+	/*
+	 * There are three classes of arena operations from a locking
+	 * perspective:
+	 * 1) Thread asssignment (modifies nthreads) is protected by
+	 *    arenas_lock.
+	 * 2) Bin-related operations are protected by bin locks.
+	 * 3) Chunk- and run-related operations are protected by this mutex.
+	 */
+	malloc_mutex_t		lock;
+
+#ifdef JEMALLOC_STATS
+	arena_stats_t		stats;
+#  ifdef JEMALLOC_TCACHE
+	/*
+	 * List of tcaches for extant threads associated with this arena.
+	 * Stats from these are merged incrementally, and at exit.
+	 */
+	ql_head(tcache_t)	tcache_ql;
+#  endif
+#endif
+
+#ifdef JEMALLOC_PROF
+	uint64_t		prof_accumbytes;
+#endif
+
+	/* List of dirty-page-containing chunks this arena manages. */
+	ql_head(arena_chunk_t)	chunks_dirty;
+
+	/*
+	 * In order to avoid rapid chunk allocation/deallocation when an arena
+	 * oscillates right on the cusp of needing a new chunk, cache the most
+	 * recently freed chunk.  The spare is left in the arena's chunk trees
+	 * until it is deleted.
+	 *
+	 * There is one spare chunk per arena, rather than one spare total, in
+	 * order to avoid interactions between multiple threads that could make
+	 * a single spare inadequate.
+	 */
+	arena_chunk_t		*spare;
+
+	/* Number of pages in active runs. */
+	size_t			nactive;
+
+	/*
+	 * Current count of pages within unused runs that are potentially
+	 * dirty, and for which madvise(... MADV_DONTNEED) has not been called.
+	 * By tracking this, we can institute a limit on how much dirty unused
+	 * memory is mapped for each arena.
+	 */
+	size_t			ndirty;
+
+	/*
+	 * Approximate number of pages being purged.  It is possible for
+	 * multiple threads to purge dirty pages concurrently, and they use
+	 * npurgatory to indicate the total number of pages all threads are
+	 * attempting to purge.
+	 */
+	size_t			npurgatory;
+
+	/*
+	 * Size/address-ordered trees of this arena's available runs.  The trees
+	 * are used for first-best-fit run allocation.  The dirty tree contains
+	 * runs with dirty pages (i.e. very likely to have been touched and
+	 * therefore have associated physical pages), whereas the clean tree
+	 * contains runs with pages that either have no associated physical
+	 * pages, or have pages that the kernel may recycle at any time due to
+	 * previous madvise(2) calls.  The dirty tree is used in preference to
+	 * the clean tree for allocations, because using dirty pages reduces
+	 * the amount of dirty purging necessary to keep the active:dirty page
+	 * ratio below the purge threshold.
+	 */
+	arena_avail_tree_t	runs_avail_clean;
+	arena_avail_tree_t	runs_avail_dirty;
+
+	/*
+	 * bins is used to store trees of free regions of the following sizes,
+	 * assuming a 64-bit system with 16-byte quantum, 4 KiB page size, and
+	 * default MALLOC_CONF.
+	 *
+	 *   bins[i] |   size |
+	 *   --------+--------+
+	 *        0  |      8 |
+	 *   --------+--------+
+	 *        1  |     16 |
+	 *        2  |     32 |
+	 *        3  |     48 |
+	 *           :        :
+	 *        6  |     96 |
+	 *        7  |    112 |
+	 *        8  |    128 |
+	 *   --------+--------+
+	 *        9  |    192 |
+	 *       10  |    256 |
+	 *       11  |    320 |
+	 *       12  |    384 |
+	 *       13  |    448 |
+	 *       14  |    512 |
+	 *   --------+--------+
+	 *       15  |    768 |
+	 *       16  |   1024 |
+	 *       17  |   1280 |
+	 *           :        :
+	 *       25  |   3328 |
+	 *       26  |   3584 |
+	 *       27  |   3840 |
+	 *   --------+--------+
+	 */
+	arena_bin_t		bins[1]; /* Dynamically sized. */
+};
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+extern size_t	opt_lg_qspace_max;
+extern size_t	opt_lg_cspace_max;
+extern ssize_t	opt_lg_dirty_mult;
+/*
+ * small_size2bin is a compact lookup table that rounds request sizes up to
+ * size classes.  In order to reduce cache footprint, the table is compressed,
+ * and all accesses are via the SMALL_SIZE2BIN macro.
+ */
+extern uint8_t const	*small_size2bin;
+#define	SMALL_SIZE2BIN(s)	(small_size2bin[(s-1) >> LG_TINY_MIN])
+
+extern arena_bin_info_t	*arena_bin_info;
+
+/* Various bin-related settings. */
+#ifdef JEMALLOC_TINY		/* Number of (2^n)-spaced tiny bins. */
+#  define		ntbins	((unsigned)(LG_QUANTUM - LG_TINY_MIN))
+#else
+#  define		ntbins	0
+#endif
+extern unsigned		nqbins; /* Number of quantum-spaced bins. */
+extern unsigned		ncbins; /* Number of cacheline-spaced bins. */
+extern unsigned		nsbins; /* Number of subpage-spaced bins. */
+extern unsigned		nbins;
+#ifdef JEMALLOC_TINY
+#  define		tspace_max	((size_t)(QUANTUM >> 1))
+#endif
+#define			qspace_min	QUANTUM
+extern size_t		qspace_max;
+extern size_t		cspace_min;
+extern size_t		cspace_max;
+extern size_t		sspace_min;
+extern size_t		sspace_max;
+#define			small_maxclass	sspace_max
+
+#define			nlclasses (chunk_npages - map_bias)
+
+void	arena_purge_all(arena_t *arena);
+#ifdef JEMALLOC_PROF
+void	arena_prof_accum(arena_t *arena, uint64_t accumbytes);
+#endif
+#ifdef JEMALLOC_TCACHE
+void	arena_tcache_fill_small(arena_t *arena, tcache_bin_t *tbin,
+    size_t binind
+#  ifdef JEMALLOC_PROF
+    , uint64_t prof_accumbytes
+#  endif
+    );
+#endif
+void	*arena_malloc_small(arena_t *arena, size_t size, bool zero);
+void	*arena_malloc_large(arena_t *arena, size_t size, bool zero);
+void	*arena_malloc(size_t size, bool zero);
+void	*arena_palloc(arena_t *arena, size_t size, size_t alloc_size,
+    size_t alignment, bool zero);
+size_t	arena_salloc(const void *ptr);
+#ifdef JEMALLOC_PROF
+void	arena_prof_promoted(const void *ptr, size_t size);
+size_t	arena_salloc_demote(const void *ptr);
+#endif
+void	arena_dalloc_bin(arena_t *arena, arena_chunk_t *chunk, void *ptr,
+    arena_chunk_map_t *mapelm);
+void	arena_dalloc_large(arena_t *arena, arena_chunk_t *chunk, void *ptr);
+#ifdef JEMALLOC_STATS
+void	arena_stats_merge(arena_t *arena, size_t *nactive, size_t *ndirty,
+    arena_stats_t *astats, malloc_bin_stats_t *bstats,
+    malloc_large_stats_t *lstats);
+#endif
+void	*arena_ralloc_no_move(void *ptr, size_t oldsize, size_t size,
+    size_t extra, bool zero);
+void	*arena_ralloc(void *ptr, size_t oldsize, size_t size, size_t extra,
+    size_t alignment, bool zero);
+bool	arena_new(arena_t *arena, unsigned ind);
+bool	arena_boot(void);
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#ifndef JEMALLOC_ENABLE_INLINE
+size_t	arena_bin_index(arena_t *arena, arena_bin_t *bin);
+unsigned	arena_run_regind(arena_run_t *run, arena_bin_info_t *bin_info,
+    const void *ptr);
+#  ifdef JEMALLOC_PROF
+prof_ctx_t	*arena_prof_ctx_get(const void *ptr);
+void	arena_prof_ctx_set(const void *ptr, prof_ctx_t *ctx);
+#  endif
+void	arena_dalloc(arena_t *arena, arena_chunk_t *chunk, void *ptr);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_ARENA_C_))
+JEMALLOC_INLINE size_t
+arena_bin_index(arena_t *arena, arena_bin_t *bin)
+{
+	size_t binind = bin - arena->bins;
+	assert(binind < nbins);
+	return (binind);
+}
+
+JEMALLOC_INLINE unsigned
+arena_run_regind(arena_run_t *run, arena_bin_info_t *bin_info, const void *ptr)
+{
+	unsigned shift, diff, regind;
+	size_t size;
+
+	dassert(run->magic == ARENA_RUN_MAGIC);
+	/*
+	 * Freeing a pointer lower than region zero can cause assertion
+	 * failure.
+	 */
+	assert((uintptr_t)ptr >= (uintptr_t)run +
+	    (uintptr_t)bin_info->reg0_offset);
+
+	/*
+	 * Avoid doing division with a variable divisor if possible.  Using
+	 * actual division here can reduce allocator throughput by over 20%!
+	 */
+	diff = (unsigned)((uintptr_t)ptr - (uintptr_t)run -
+	    bin_info->reg0_offset);
+
+	/* Rescale (factor powers of 2 out of the numerator and denominator). */
+	size = bin_info->reg_size;
+	shift = ffs(size) - 1;
+	diff >>= shift;
+	size >>= shift;
+
+	if (size == 1) {
+		/* The divisor was a power of 2. */
+		regind = diff;
+	} else {
+		/*
+		 * To divide by a number D that is not a power of two we
+		 * multiply by (2^21 / D) and then right shift by 21 positions.
+		 *
+		 *   X / D
+		 *
+		 * becomes
+		 *
+		 *   (X * size_invs[D - 3]) >> SIZE_INV_SHIFT
+		 *
+		 * We can omit the first three elements, because we never
+		 * divide by 0, and 1 and 2 are both powers of two, which are
+		 * handled above.
+		 */
+#define	SIZE_INV_SHIFT	((sizeof(unsigned) << 3) - LG_RUN_MAXREGS)
+#define	SIZE_INV(s)	(((1U << SIZE_INV_SHIFT) / (s)) + 1)
+		static const unsigned size_invs[] = {
+		    SIZE_INV(3),
+		    SIZE_INV(4), SIZE_INV(5), SIZE_INV(6), SIZE_INV(7),
+		    SIZE_INV(8), SIZE_INV(9), SIZE_INV(10), SIZE_INV(11),
+		    SIZE_INV(12), SIZE_INV(13), SIZE_INV(14), SIZE_INV(15),
+		    SIZE_INV(16), SIZE_INV(17), SIZE_INV(18), SIZE_INV(19),
+		    SIZE_INV(20), SIZE_INV(21), SIZE_INV(22), SIZE_INV(23),
+		    SIZE_INV(24), SIZE_INV(25), SIZE_INV(26), SIZE_INV(27),
+		    SIZE_INV(28), SIZE_INV(29), SIZE_INV(30), SIZE_INV(31)
+		};
+
+		if (size <= ((sizeof(size_invs) / sizeof(unsigned)) + 2))
+			regind = (diff * size_invs[size - 3]) >> SIZE_INV_SHIFT;
+		else
+			regind = diff / size;
+#undef SIZE_INV
+#undef SIZE_INV_SHIFT
+	}
+	assert(diff == regind * size);
+	assert(regind < bin_info->nregs);
+
+	return (regind);
+}
+
+#ifdef JEMALLOC_PROF
+JEMALLOC_INLINE prof_ctx_t *
+arena_prof_ctx_get(const void *ptr)
+{
+	prof_ctx_t *ret;
+	arena_chunk_t *chunk;
+	size_t pageind, mapbits;
+
+	assert(ptr != NULL);
+	assert(CHUNK_ADDR2BASE(ptr) != ptr);
+
+	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
+	pageind = ((uintptr_t)ptr - (uintptr_t)chunk) >> PAGE_SHIFT;
+	mapbits = chunk->map[pageind-map_bias].bits;
+	assert((mapbits & CHUNK_MAP_ALLOCATED) != 0);
+	if ((mapbits & CHUNK_MAP_LARGE) == 0) {
+		if (prof_promote)
+			ret = (prof_ctx_t *)(uintptr_t)1U;
+		else {
+			arena_run_t *run = (arena_run_t *)((uintptr_t)chunk +
+			    (uintptr_t)((pageind - (mapbits >> PAGE_SHIFT)) <<
+			    PAGE_SHIFT));
+			size_t binind = arena_bin_index(chunk->arena, run->bin);
+			arena_bin_info_t *bin_info = &arena_bin_info[binind];
+			unsigned regind;
+
+			dassert(run->magic == ARENA_RUN_MAGIC);
+			regind = arena_run_regind(run, bin_info, ptr);
+			ret = *(prof_ctx_t **)((uintptr_t)run +
+			    bin_info->ctx0_offset + (regind *
+			    sizeof(prof_ctx_t *)));
+		}
+	} else
+		ret = chunk->map[pageind-map_bias].prof_ctx;
+
+	return (ret);
+}
+
+JEMALLOC_INLINE void
+arena_prof_ctx_set(const void *ptr, prof_ctx_t *ctx)
+{
+	arena_chunk_t *chunk;
+	size_t pageind, mapbits;
+
+	assert(ptr != NULL);
+	assert(CHUNK_ADDR2BASE(ptr) != ptr);
+
+	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
+	pageind = ((uintptr_t)ptr - (uintptr_t)chunk) >> PAGE_SHIFT;
+	mapbits = chunk->map[pageind-map_bias].bits;
+	assert((mapbits & CHUNK_MAP_ALLOCATED) != 0);
+	if ((mapbits & CHUNK_MAP_LARGE) == 0) {
+		if (prof_promote == false) {
+			arena_run_t *run = (arena_run_t *)((uintptr_t)chunk +
+			    (uintptr_t)((pageind - (mapbits >> PAGE_SHIFT)) <<
+			    PAGE_SHIFT));
+			arena_bin_t *bin = run->bin;
+			size_t binind;
+			arena_bin_info_t *bin_info;
+			unsigned regind;
+
+			dassert(run->magic == ARENA_RUN_MAGIC);
+			binind = arena_bin_index(chunk->arena, bin);
+			bin_info = &arena_bin_info[binind];
+			regind = arena_run_regind(run, bin_info, ptr);
+
+			*((prof_ctx_t **)((uintptr_t)run + bin_info->ctx0_offset
+			    + (regind * sizeof(prof_ctx_t *)))) = ctx;
+		} else
+			assert((uintptr_t)ctx == (uintptr_t)1U);
+	} else
+		chunk->map[pageind-map_bias].prof_ctx = ctx;
+}
+#endif
+
+JEMALLOC_INLINE void
+arena_dalloc(arena_t *arena, arena_chunk_t *chunk, void *ptr)
+{
+	size_t pageind;
+	arena_chunk_map_t *mapelm;
+
+	assert(arena != NULL);
+	dassert(arena->magic == ARENA_MAGIC);
+	assert(chunk->arena == arena);
+	assert(ptr != NULL);
+	assert(CHUNK_ADDR2BASE(ptr) != ptr);
+
+	pageind = ((uintptr_t)ptr - (uintptr_t)chunk) >> PAGE_SHIFT;
+	mapelm = &chunk->map[pageind-map_bias];
+	assert((mapelm->bits & CHUNK_MAP_ALLOCATED) != 0);
+	if ((mapelm->bits & CHUNK_MAP_LARGE) == 0) {
+		/* Small allocation. */
+#ifdef JEMALLOC_TCACHE
+		tcache_t *tcache;
+
+		if ((tcache = tcache_get()) != NULL)
+			tcache_dalloc_small(tcache, ptr);
+		else {
+#endif
+			arena_run_t *run;
+			arena_bin_t *bin;
+
+			run = (arena_run_t *)((uintptr_t)chunk +
+			    (uintptr_t)((pageind - (mapelm->bits >>
+			    PAGE_SHIFT)) << PAGE_SHIFT));
+			dassert(run->magic == ARENA_RUN_MAGIC);
+			bin = run->bin;
+#ifdef JEMALLOC_DEBUG
+			{
+				size_t binind = arena_bin_index(arena, bin);
+				arena_bin_info_t *bin_info =
+				    &arena_bin_info[binind];
+				assert(((uintptr_t)ptr - ((uintptr_t)run +
+				    (uintptr_t)bin_info->reg0_offset)) %
+				    bin_info->reg_size == 0);
+			}
+#endif
+			malloc_mutex_lock(&bin->lock);
+			arena_dalloc_bin(arena, chunk, ptr, mapelm);
+			malloc_mutex_unlock(&bin->lock);
+#ifdef JEMALLOC_TCACHE
+		}
+#endif
+	} else {
+#ifdef JEMALLOC_TCACHE
+		size_t size = mapelm->bits & ~PAGE_MASK;
+
+		assert(((uintptr_t)ptr & PAGE_MASK) == 0);
+		if (size <= tcache_maxclass) {
+			tcache_t *tcache;
+
+			if ((tcache = tcache_get()) != NULL)
+				tcache_dalloc_large(tcache, ptr, size);
+			else {
+				malloc_mutex_lock(&arena->lock);
+				arena_dalloc_large(arena, chunk, ptr);
+				malloc_mutex_unlock(&arena->lock);
+			}
+		} else {
+			malloc_mutex_lock(&arena->lock);
+			arena_dalloc_large(arena, chunk, ptr);
+			malloc_mutex_unlock(&arena->lock);
+		}
+#else
+		assert(((uintptr_t)ptr & PAGE_MASK) == 0);
+		malloc_mutex_lock(&arena->lock);
+		arena_dalloc_large(arena, chunk, ptr);
+		malloc_mutex_unlock(&arena->lock);
+#endif
+	}
+}
+#endif
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/

deps/jemalloc/include/jemalloc/internal/atomic.h

@@ -0,0 +1,169 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+#define	atomic_read_uint64(p)	atomic_add_uint64(p, 0)
+#define	atomic_read_uint32(p)	atomic_add_uint32(p, 0)
+
+#if (LG_SIZEOF_PTR == 3)
+#  define atomic_read_z(p)						\
+    (size_t)atomic_add_uint64((uint64_t *)p, (uint64_t)0)
+#  define atomic_add_z(p, x)						\
+    (size_t)atomic_add_uint64((uint64_t *)p, (uint64_t)x)
+#  define atomic_sub_z(p, x)						\
+    (size_t)atomic_sub_uint64((uint64_t *)p, (uint64_t)x)
+#elif (LG_SIZEOF_PTR == 2)
+#  define atomic_read_z(p)						\
+    (size_t)atomic_add_uint32((uint32_t *)p, (uint32_t)0)
+#  define atomic_add_z(p, x)						\
+    (size_t)atomic_add_uint32((uint32_t *)p, (uint32_t)x)
+#  define atomic_sub_z(p, x)						\
+    (size_t)atomic_sub_uint32((uint32_t *)p, (uint32_t)x)
+#endif
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#ifndef JEMALLOC_ENABLE_INLINE
+uint64_t	atomic_add_uint64(uint64_t *p, uint64_t x);
+uint64_t	atomic_sub_uint64(uint64_t *p, uint64_t x);
+uint32_t	atomic_add_uint32(uint32_t *p, uint32_t x);
+uint32_t	atomic_sub_uint32(uint32_t *p, uint32_t x);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_ATOMIC_C_))
+/******************************************************************************/
+/* 64-bit operations. */
+#ifdef __GCC_HAVE_SYNC_COMPARE_AND_SWAP_8
+JEMALLOC_INLINE uint64_t
+atomic_add_uint64(uint64_t *p, uint64_t x)
+{
+
+	return (__sync_add_and_fetch(p, x));
+}
+
+JEMALLOC_INLINE uint64_t
+atomic_sub_uint64(uint64_t *p, uint64_t x)
+{
+
+	return (__sync_sub_and_fetch(p, x));
+}
+#elif (defined(JEMALLOC_OSATOMIC))
+JEMALLOC_INLINE uint64_t
+atomic_add_uint64(uint64_t *p, uint64_t x)
+{
+
+	return (OSAtomicAdd64((int64_t)x, (int64_t *)p));
+}
+
+JEMALLOC_INLINE uint64_t
+atomic_sub_uint64(uint64_t *p, uint64_t x)
+{
+
+	return (OSAtomicAdd64(-((int64_t)x), (int64_t *)p));
+}
+#elif (defined(__amd64_) || defined(__x86_64__))
+JEMALLOC_INLINE uint64_t
+atomic_add_uint64(uint64_t *p, uint64_t x)
+{
+
+	asm volatile (
+	    "lock; xaddq %0, %1;"
+	    : "+r" (x), "=m" (*p) /* Outputs. */
+	    : "m" (*p) /* Inputs. */
+	    );
+
+	return (x);
+}
+
+JEMALLOC_INLINE uint64_t
+atomic_sub_uint64(uint64_t *p, uint64_t x)
+{
+
+	x = (uint64_t)(-(int64_t)x);
+	asm volatile (
+	    "lock; xaddq %0, %1;"
+	    : "+r" (x), "=m" (*p) /* Outputs. */
+	    : "m" (*p) /* Inputs. */
+	    );
+
+	return (x);
+}
+#else
+#  if (LG_SIZEOF_PTR == 3)
+#    error "Missing implementation for 64-bit atomic operations"
+#  endif
+#endif
+
+/******************************************************************************/
+/* 32-bit operations. */
+#ifdef __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4
+JEMALLOC_INLINE uint32_t
+atomic_add_uint32(uint32_t *p, uint32_t x)
+{
+
+	return (__sync_add_and_fetch(p, x));
+}
+
+JEMALLOC_INLINE uint32_t
+atomic_sub_uint32(uint32_t *p, uint32_t x)
+{
+
+	return (__sync_sub_and_fetch(p, x));
+}
+#elif (defined(JEMALLOC_OSATOMIC))
+JEMALLOC_INLINE uint32_t
+atomic_add_uint32(uint32_t *p, uint32_t x)
+{
+
+	return (OSAtomicAdd32((int32_t)x, (int32_t *)p));
+}
+
+JEMALLOC_INLINE uint32_t
+atomic_sub_uint32(uint32_t *p, uint32_t x)
+{
+
+	return (OSAtomicAdd32(-((int32_t)x), (int32_t *)p));
+}
+#elif (defined(__i386__) || defined(__amd64_) || defined(__x86_64__))
+JEMALLOC_INLINE uint32_t
+atomic_add_uint32(uint32_t *p, uint32_t x)
+{
+
+	asm volatile (
+	    "lock; xaddl %0, %1;"
+	    : "+r" (x), "=m" (*p) /* Outputs. */
+	    : "m" (*p) /* Inputs. */
+	    );
+
+	return (x);
+}
+
+JEMALLOC_INLINE uint32_t
+atomic_sub_uint32(uint32_t *p, uint32_t x)
+{
+
+	x = (uint32_t)(-(int32_t)x);
+	asm volatile (
+	    "lock; xaddl %0, %1;"
+	    : "+r" (x), "=m" (*p) /* Outputs. */
+	    : "m" (*p) /* Inputs. */
+	    );
+
+	return (x);
+}
+#else
+#  error "Missing implementation for 32-bit atomic operations"
+#endif
+#endif
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/

deps/jemalloc/include/jemalloc/internal/base.h

@@ -0,0 +1,24 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+extern malloc_mutex_t	base_mtx;
+
+void	*base_alloc(size_t size);
+extent_node_t *base_node_alloc(void);
+void	base_node_dealloc(extent_node_t *node);
+bool	base_boot(void);
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/

deps/jemalloc/include/jemalloc/internal/bitmap.h

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

deps/jemalloc/include/jemalloc/internal/chunk.h

@@ -0,0 +1,65 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+/*
+ * Size and alignment of memory chunks that are allocated by the OS's virtual
+ * memory system.
+ */
+#define	LG_CHUNK_DEFAULT	22
+
+/* Return the chunk address for allocation address a. */
+#define	CHUNK_ADDR2BASE(a)						\
+	((void *)((uintptr_t)(a) & ~chunksize_mask))
+
+/* Return the chunk offset of address a. */
+#define	CHUNK_ADDR2OFFSET(a)						\
+	((size_t)((uintptr_t)(a) & chunksize_mask))
+
+/* Return the smallest chunk multiple that is >= s. */
+#define	CHUNK_CEILING(s)						\
+	(((s) + chunksize_mask) & ~chunksize_mask)
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+extern size_t		opt_lg_chunk;
+#ifdef JEMALLOC_SWAP
+extern bool		opt_overcommit;
+#endif
+
+#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
+/* Protects stats_chunks; currently not used for any other purpose. */
+extern malloc_mutex_t	chunks_mtx;
+/* Chunk statistics. */
+extern chunk_stats_t	stats_chunks;
+#endif
+
+#ifdef JEMALLOC_IVSALLOC
+extern rtree_t		*chunks_rtree;
+#endif
+
+extern size_t		chunksize;
+extern size_t		chunksize_mask; /* (chunksize - 1). */
+extern size_t		chunk_npages;
+extern size_t		map_bias; /* Number of arena chunk header pages. */
+extern size_t		arena_maxclass; /* Max size class for arenas. */
+
+void	*chunk_alloc(size_t size, bool base, bool *zero);
+void	chunk_dealloc(void *chunk, size_t size);
+bool	chunk_boot(void);
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
+
+#include "jemalloc/internal/chunk_swap.h"
+#include "jemalloc/internal/chunk_dss.h"
+#include "jemalloc/internal/chunk_mmap.h"

deps/jemalloc/include/jemalloc/internal/chunk_dss.h

@@ -0,0 +1,30 @@
+#ifdef JEMALLOC_DSS
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+/*
+ * Protects sbrk() calls.  This avoids malloc races among threads, though it
+ * does not protect against races with threads that call sbrk() directly.
+ */
+extern malloc_mutex_t	dss_mtx;
+
+void	*chunk_alloc_dss(size_t size, bool *zero);
+bool	chunk_in_dss(void *chunk);
+bool	chunk_dealloc_dss(void *chunk, size_t size);
+bool	chunk_dss_boot(void);
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
+#endif /* JEMALLOC_DSS */

deps/jemalloc/include/jemalloc/internal/chunk_mmap.h

@@ -0,0 +1,23 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+void	*chunk_alloc_mmap(size_t size);
+void	*chunk_alloc_mmap_noreserve(size_t size);
+void	chunk_dealloc_mmap(void *chunk, size_t size);
+
+bool	chunk_mmap_boot(void);
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/

deps/jemalloc/include/jemalloc/internal/chunk_swap.h

@@ -0,0 +1,34 @@
+#ifdef JEMALLOC_SWAP
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+extern malloc_mutex_t	swap_mtx;
+extern bool		swap_enabled;
+extern bool		swap_prezeroed;
+extern size_t		swap_nfds;
+extern int		*swap_fds;
+#ifdef JEMALLOC_STATS
+extern size_t		swap_avail;
+#endif
+
+void	*chunk_alloc_swap(size_t size, bool *zero);
+bool	chunk_in_swap(void *chunk);
+bool	chunk_dealloc_swap(void *chunk, size_t size);
+bool	chunk_swap_enable(const int *fds, unsigned nfds, bool prezeroed);
+bool	chunk_swap_boot(void);
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
+#endif /* JEMALLOC_SWAP */

deps/jemalloc/include/jemalloc/internal/ckh.h

@@ -0,0 +1,95 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+typedef struct ckh_s ckh_t;
+typedef struct ckhc_s ckhc_t;
+
+/* Typedefs to allow easy function pointer passing. */
+typedef void ckh_hash_t (const void *, unsigned, size_t *, size_t *);
+typedef bool ckh_keycomp_t (const void *, const void *);
+
+/* 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)
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+/* Hash table cell. */
+struct ckhc_s {
+	const void	*key;
+	const void	*data;
+};
+
+struct ckh_s {
+#ifdef JEMALLOC_DEBUG
+#define	CKH_MAGIC	0x3af2489d
+	uint32_t	magic;
+#endif
+
+#ifdef CKH_COUNT
+	/* Counters used to get an idea of performance. */
+	uint64_t	ngrows;
+	uint64_t	nshrinks;
+	uint64_t	nshrinkfails;
+	uint64_t	ninserts;
+	uint64_t	nrelocs;
+#endif
+
+	/* Used for pseudo-random number generation. */
+#define	CKH_A		1103515241
+#define	CKH_C		12347
+	uint32_t	prn_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;
+};
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+bool	ckh_new(ckh_t *ckh, size_t minitems, ckh_hash_t *hash,
+    ckh_keycomp_t *keycomp);
+void	ckh_delete(ckh_t *ckh);
+size_t	ckh_count(ckh_t *ckh);
+bool	ckh_iter(ckh_t *ckh, size_t *tabind, void **key, void **data);
+bool	ckh_insert(ckh_t *ckh, const void *key, const void *data);
+bool	ckh_remove(ckh_t *ckh, const void *searchkey, void **key,
+    void **data);
+bool	ckh_search(ckh_t *ckh, const void *seachkey, void **key, void **data);
+void	ckh_string_hash(const void *key, unsigned minbits, size_t *hash1,
+    size_t *hash2);
+bool	ckh_string_keycomp(const void *k1, const void *k2);
+void	ckh_pointer_hash(const void *key, unsigned minbits, size_t *hash1,
+    size_t *hash2);
+bool	ckh_pointer_keycomp(const void *k1, const void *k2);
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/

deps/jemalloc/include/jemalloc/internal/ctl.h

@@ -0,0 +1,118 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+typedef struct ctl_node_s ctl_node_t;
+typedef struct ctl_arena_stats_s ctl_arena_stats_t;
+typedef struct ctl_stats_s ctl_stats_t;
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+struct ctl_node_s {
+	bool			named;
+	union {
+		struct {
+			const char	*name;
+			/* If (nchildren == 0), this is a terminal node. */
+			unsigned	nchildren;
+			const	ctl_node_t *children;
+		} named;
+		struct {
+			const ctl_node_t *(*index)(const size_t *, size_t,
+			    size_t);
+		} indexed;
+	} u;
+	int	(*ctl)(const size_t *, size_t, void *, size_t *, void *,
+	    size_t);
+};
+
+struct ctl_arena_stats_s {
+	bool			initialized;
+	unsigned		nthreads;
+	size_t			pactive;
+	size_t			pdirty;
+#ifdef JEMALLOC_STATS
+	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 elements. */
+	malloc_large_stats_t	*lstats;	/* nlclasses elements. */
+#endif
+};
+
+struct ctl_stats_s {
+#ifdef JEMALLOC_STATS
+	size_t			allocated;
+	size_t			active;
+	size_t			mapped;
+	struct {
+		size_t		current;	/* stats_chunks.curchunks */
+		uint64_t	total;		/* stats_chunks.nchunks */
+		size_t		high;		/* stats_chunks.highchunks */
+	} chunks;
+	struct {
+		size_t		allocated;	/* huge_allocated */
+		uint64_t	nmalloc;	/* huge_nmalloc */
+		uint64_t	ndalloc;	/* huge_ndalloc */
+	} huge;
+#endif
+	ctl_arena_stats_t	*arenas;	/* (narenas + 1) elements. */
+#ifdef JEMALLOC_SWAP
+	size_t			swap_avail;
+#endif
+};
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+int	ctl_byname(const char *name, void *oldp, size_t *oldlenp, void *newp,
+    size_t newlen);
+int	ctl_nametomib(const char *name, size_t *mibp, size_t *miblenp);
+
+int	ctl_bymib(const size_t *mib, size_t miblen, void *oldp, size_t *oldlenp,
+    void *newp, size_t newlen);
+bool	ctl_boot(void);
+
+#define	xmallctl(name, oldp, oldlenp, newp, newlen) do {		\
+	if (JEMALLOC_P(mallctl)(name, oldp, oldlenp, newp, newlen)	\
+	    != 0) {							\
+		malloc_write("<jemalloc>: Failure in xmallctl(\"");	\
+		malloc_write(name);					\
+		malloc_write("\", ...)\n");				\
+		abort();						\
+	}								\
+} while (0)
+
+#define	xmallctlnametomib(name, mibp, miblenp) do {			\
+	if (JEMALLOC_P(mallctlnametomib)(name, mibp, miblenp) != 0) {	\
+		malloc_write(						\
+		    "<jemalloc>: Failure in xmallctlnametomib(\"");	\
+		malloc_write(name);					\
+		malloc_write("\", ...)\n");				\
+		abort();						\
+	}								\
+} while (0)
+
+#define	xmallctlbymib(mib, miblen, oldp, oldlenp, newp, newlen) do {	\
+	if (JEMALLOC_P(mallctlbymib)(mib, miblen, oldp, oldlenp, newp,	\
+	    newlen) != 0) {						\
+		malloc_write(						\
+		    "<jemalloc>: Failure in xmallctlbymib()\n");	\
+		abort();						\
+	}								\
+} while (0)
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
+

deps/jemalloc/include/jemalloc/internal/extent.h

@@ -0,0 +1,49 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+typedef struct extent_node_s extent_node_t;
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+/* Tree of extents. */
+struct extent_node_s {
+#if (defined(JEMALLOC_SWAP) || defined(JEMALLOC_DSS))
+	/* Linkage for the size/address-ordered tree. */
+	rb_node(extent_node_t)	link_szad;
+#endif
+
+	/* Linkage for the address-ordered tree. */
+	rb_node(extent_node_t)	link_ad;
+
+#ifdef JEMALLOC_PROF
+	/* Profile counters, used for huge objects. */
+	prof_ctx_t		*prof_ctx;
+#endif
+
+	/* Pointer to the extent that this tree node is responsible for. */
+	void			*addr;
+
+	/* Total region size. */
+	size_t			size;
+};
+typedef rb_tree(extent_node_t) extent_tree_t;
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+#if (defined(JEMALLOC_SWAP) || defined(JEMALLOC_DSS))
+rb_proto(, extent_tree_szad_, extent_tree_t, extent_node_t)
+#endif
+
+rb_proto(, extent_tree_ad_, extent_tree_t, extent_node_t)
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
+

deps/jemalloc/include/jemalloc/internal/hash.h

@@ -0,0 +1,70 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#ifndef JEMALLOC_ENABLE_INLINE
+uint64_t	hash(const void *key, size_t len, uint64_t seed);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_HASH_C_))
+/*
+ * The following hash function is based on MurmurHash64A(), placed into the
+ * public domain by Austin Appleby.  See http://murmurhash.googlepages.com/ for
+ * details.
+ */
+JEMALLOC_INLINE uint64_t
+hash(const void *key, size_t len, uint64_t seed)
+{
+	const uint64_t m = 0xc6a4a7935bd1e995;
+	const int r = 47;
+	uint64_t h = seed ^ (len * m);
+	const uint64_t *data = (const uint64_t *)key;
+	const uint64_t *end = data + (len/8);
+	const unsigned char *data2;
+
+	assert(((uintptr_t)key & 0x7) == 0);
+
+	while(data != end) {
+		uint64_t k = *data++;
+
+		k *= m;
+		k ^= k >> r;
+		k *= m;
+
+		h ^= k;
+		h *= m;
+	}
+
+	data2 = (const unsigned char *)data;
+	switch(len & 7) {
+		case 7: h ^= ((uint64_t)(data2[6])) << 48;
+		case 6: h ^= ((uint64_t)(data2[5])) << 40;
+		case 5: h ^= ((uint64_t)(data2[4])) << 32;
+		case 4: h ^= ((uint64_t)(data2[3])) << 24;
+		case 3: h ^= ((uint64_t)(data2[2])) << 16;
+		case 2: h ^= ((uint64_t)(data2[1])) << 8;
+		case 1: h ^= ((uint64_t)(data2[0]));
+			h *= m;
+	}
+
+	h ^= h >> r;
+	h *= m;
+	h ^= h >> r;
+
+	return (h);
+}
+#endif
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/

deps/jemalloc/include/jemalloc/internal/huge.h

@@ -0,0 +1,41 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+#ifdef JEMALLOC_STATS
+/* Huge allocation statistics. */
+extern uint64_t		huge_nmalloc;
+extern uint64_t		huge_ndalloc;
+extern size_t		huge_allocated;
+#endif
+
+/* Protects chunk-related data structures. */
+extern malloc_mutex_t	huge_mtx;
+
+void	*huge_malloc(size_t size, bool zero);
+void	*huge_palloc(size_t size, size_t alignment, bool zero);
+void	*huge_ralloc_no_move(void *ptr, size_t oldsize, size_t size,
+    size_t extra);
+void	*huge_ralloc(void *ptr, size_t oldsize, size_t size, size_t extra,
+    size_t alignment, bool zero);
+void	huge_dalloc(void *ptr, bool unmap);
+size_t	huge_salloc(const void *ptr);
+#ifdef JEMALLOC_PROF
+prof_ctx_t	*huge_prof_ctx_get(const void *ptr);
+void	huge_prof_ctx_set(const void *ptr, prof_ctx_t *ctx);
+#endif
+bool	huge_boot(void);
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/

deps/jemalloc/include/jemalloc/internal/jemalloc_internal.h.in

@@ -0,0 +1,782 @@
+#include <sys/mman.h>
+#include <sys/param.h>
+#include <sys/time.h>
+#include <sys/types.h>
+#include <sys/sysctl.h>
+#include <sys/uio.h>
+
+#include <errno.h>
+#include <limits.h>
+#ifndef SIZE_T_MAX
+#  define SIZE_T_MAX	SIZE_MAX
+#endif
+#include <pthread.h>
+#include <sched.h>
+#include <stdarg.h>
+#include <stdbool.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdint.h>
+#include <stddef.h>
+#ifndef offsetof
+#  define offsetof(type, member)	((size_t)&(((type *)NULL)->member))
+#endif
+#include <inttypes.h>
+#include <string.h>
+#include <strings.h>
+#include <ctype.h>
+#include <unistd.h>
+#include <fcntl.h>
+#include <pthread.h>
+#include <math.h>
+
+#define	JEMALLOC_MANGLE
+#include "../jemalloc@install_suffix@.h"
+
+#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>
+#include <malloc/malloc.h>
+#endif
+
+#ifdef JEMALLOC_LAZY_LOCK
+#include <dlfcn.h>
+#endif
+
+#define	RB_COMPACT
+#include "jemalloc/internal/rb.h"
+#include "jemalloc/internal/qr.h"
+#include "jemalloc/internal/ql.h"
+
+extern void	(*JEMALLOC_P(malloc_message))(void *wcbopaque, const char *s);
+
+/*
+ * Define a custom assert() in order to reduce the chances of deadlock during
+ * assertion failure.
+ */
+#ifndef assert
+#  ifdef JEMALLOC_DEBUG
+#    define assert(e) do {						\
+	if (!(e)) {							\
+		char line_buf[UMAX2S_BUFSIZE];				\
+		malloc_write("<jemalloc>: ");				\
+		malloc_write(__FILE__);					\
+		malloc_write(":");					\
+		malloc_write(u2s(__LINE__, 10, line_buf));		\
+		malloc_write(": Failed assertion: ");			\
+		malloc_write("\"");					\
+		malloc_write(#e);					\
+		malloc_write("\"\n");					\
+		abort();						\
+	}								\
+} while (0)
+#  else
+#    define assert(e)
+#  endif
+#endif
+
+#ifdef JEMALLOC_DEBUG
+#  define dassert(e) assert(e)
+#else
+#  define dassert(e)
+#endif
+
+/*
+ * jemalloc can conceptually be broken into components (arena, tcache, etc.),
+ * but there are circular dependencies that cannot be broken without
+ * substantial performance degradation.  In order to reduce the effect on
+ * visual code flow, read the header files in multiple passes, with one of the
+ * following cpp variables defined during each pass:
+ *
+ *   JEMALLOC_H_TYPES   : Preprocessor-defined constants and psuedo-opaque data
+ *                        types.
+ *   JEMALLOC_H_STRUCTS : Data structures.
+ *   JEMALLOC_H_EXTERNS : Extern data declarations and function prototypes.
+ *   JEMALLOC_H_INLINES : Inline functions.
+ */
+/******************************************************************************/
+#define JEMALLOC_H_TYPES
+
+#define	ALLOCM_LG_ALIGN_MASK	((int)0x3f)
+
+#define	ZU(z)	((size_t)z)
+
+#ifndef __DECONST
+#  define	__DECONST(type, var)	((type)(uintptr_t)(const void *)(var))
+#endif
+
+#ifdef JEMALLOC_DEBUG
+   /* Disable inlining to make debugging easier. */
+#  define JEMALLOC_INLINE
+#  define inline
+#else
+#  define JEMALLOC_ENABLE_INLINE
+#  define JEMALLOC_INLINE static inline
+#endif
+
+/* Size of stack-allocated buffer passed to buferror(). */
+#define	BUFERROR_BUF		64
+
+/* Minimum alignment of allocations is 2^LG_QUANTUM bytes. */
+#ifdef __i386__
+#  define LG_QUANTUM		4
+#endif
+#ifdef __ia64__
+#  define LG_QUANTUM		4
+#endif
+#ifdef __alpha__
+#  define LG_QUANTUM		4
+#endif
+#ifdef __sparc64__
+#  define LG_QUANTUM		4
+#endif
+#if (defined(__amd64__) || defined(__x86_64__))
+#  define LG_QUANTUM		4
+#endif
+#ifdef __arm__
+#  define LG_QUANTUM		3
+#endif
+#ifdef __mips__
+#  define LG_QUANTUM		3
+#endif
+#ifdef __powerpc__
+#  define LG_QUANTUM		4
+#endif
+#ifdef __s390x__
+#  define LG_QUANTUM		4
+#endif
+
+#define	QUANTUM			((size_t)(1U << LG_QUANTUM))
+#define	QUANTUM_MASK		(QUANTUM - 1)
+
+/* Return the smallest quantum multiple that is >= a. */
+#define	QUANTUM_CEILING(a)						\
+	(((a) + QUANTUM_MASK) & ~QUANTUM_MASK)
+
+#define	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.
+ */
+#define	LG_CACHELINE		6
+#define	CACHELINE		((size_t)(1U << LG_CACHELINE))
+#define	CACHELINE_MASK		(CACHELINE - 1)
+
+/* Return the smallest cacheline multiple that is >= s. */
+#define	CACHELINE_CEILING(s)						\
+	(((s) + CACHELINE_MASK) & ~CACHELINE_MASK)
+
+/*
+ * Page size.  STATIC_PAGE_SHIFT is determined by the configure script.  If
+ * DYNAMIC_PAGE_SHIFT is enabled, only use the STATIC_PAGE_* macros where
+ * compile-time values are required for the purposes of defining data
+ * structures.
+ */
+#define	STATIC_PAGE_SIZE ((size_t)(1U << STATIC_PAGE_SHIFT))
+#define	STATIC_PAGE_MASK ((size_t)(STATIC_PAGE_SIZE - 1))
+
+#ifdef PAGE_SHIFT
+#  undef PAGE_SHIFT
+#endif
+#ifdef PAGE_SIZE
+#  undef PAGE_SIZE
+#endif
+#ifdef PAGE_MASK
+#  undef PAGE_MASK
+#endif
+
+#ifdef DYNAMIC_PAGE_SHIFT
+#  define PAGE_SHIFT	lg_pagesize
+#  define PAGE_SIZE	pagesize
+#  define PAGE_MASK	pagesize_mask
+#else
+#  define PAGE_SHIFT	STATIC_PAGE_SHIFT
+#  define PAGE_SIZE	STATIC_PAGE_SIZE
+#  define PAGE_MASK	STATIC_PAGE_MASK
+#endif
+
+/* Return the smallest pagesize multiple that is >= s. */
+#define	PAGE_CEILING(s)							\
+	(((s) + PAGE_MASK) & ~PAGE_MASK)
+
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/prn.h"
+#include "jemalloc/internal/ckh.h"
+#include "jemalloc/internal/stats.h"
+#include "jemalloc/internal/ctl.h"
+#include "jemalloc/internal/mutex.h"
+#include "jemalloc/internal/mb.h"
+#include "jemalloc/internal/extent.h"
+#include "jemalloc/internal/arena.h"
+#include "jemalloc/internal/bitmap.h"
+#include "jemalloc/internal/base.h"
+#include "jemalloc/internal/chunk.h"
+#include "jemalloc/internal/huge.h"
+#include "jemalloc/internal/rtree.h"
+#include "jemalloc/internal/tcache.h"
+#include "jemalloc/internal/hash.h"
+#ifdef JEMALLOC_ZONE
+#include "jemalloc/internal/zone.h"
+#endif
+#include "jemalloc/internal/prof.h"
+
+#undef JEMALLOC_H_TYPES
+/******************************************************************************/
+#define JEMALLOC_H_STRUCTS
+
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/prn.h"
+#include "jemalloc/internal/ckh.h"
+#include "jemalloc/internal/stats.h"
+#include "jemalloc/internal/ctl.h"
+#include "jemalloc/internal/mutex.h"
+#include "jemalloc/internal/mb.h"
+#include "jemalloc/internal/bitmap.h"
+#include "jemalloc/internal/extent.h"
+#include "jemalloc/internal/arena.h"
+#include "jemalloc/internal/base.h"
+#include "jemalloc/internal/chunk.h"
+#include "jemalloc/internal/huge.h"
+#include "jemalloc/internal/rtree.h"
+#include "jemalloc/internal/tcache.h"
+#include "jemalloc/internal/hash.h"
+#ifdef JEMALLOC_ZONE
+#include "jemalloc/internal/zone.h"
+#endif
+#include "jemalloc/internal/prof.h"
+
+#ifdef JEMALLOC_STATS
+typedef struct {
+	uint64_t	allocated;
+	uint64_t	deallocated;
+} thread_allocated_t;
+#endif
+
+#undef JEMALLOC_H_STRUCTS
+/******************************************************************************/
+#define JEMALLOC_H_EXTERNS
+
+extern bool	opt_abort;
+#ifdef JEMALLOC_FILL
+extern bool	opt_junk;
+#endif
+#ifdef JEMALLOC_SYSV
+extern bool	opt_sysv;
+#endif
+#ifdef JEMALLOC_XMALLOC
+extern bool	opt_xmalloc;
+#endif
+#ifdef JEMALLOC_FILL
+extern bool	opt_zero;
+#endif
+extern size_t	opt_narenas;
+
+#ifdef DYNAMIC_PAGE_SHIFT
+extern size_t		pagesize;
+extern size_t		pagesize_mask;
+extern size_t		lg_pagesize;
+#endif
+
+/* Number of CPUs. */
+extern unsigned		ncpus;
+
+extern malloc_mutex_t	arenas_lock; /* Protects arenas initialization. */
+extern pthread_key_t	arenas_tsd;
+#ifndef NO_TLS
+/*
+ * Map of pthread_self() --> arenas[???], used for selecting an arena to use
+ * for allocations.
+ */
+extern __thread arena_t	*arenas_tls JEMALLOC_ATTR(tls_model("initial-exec"));
+#  define ARENA_GET()	arenas_tls
+#  define ARENA_SET(v)	do {						\
+	arenas_tls = (v);						\
+	pthread_setspecific(arenas_tsd, (void *)(v));			\
+} while (0)
+#else
+#  define ARENA_GET()	((arena_t *)pthread_getspecific(arenas_tsd))
+#  define ARENA_SET(v)	do {						\
+	pthread_setspecific(arenas_tsd, (void *)(v));			\
+} while (0)
+#endif
+
+/*
+ * Arenas that are used to service external requests.  Not all elements of the
+ * arenas array are necessarily used; arenas are created lazily as needed.
+ */
+extern arena_t		**arenas;
+extern unsigned		narenas;
+
+#ifdef JEMALLOC_STATS
+#  ifndef NO_TLS
+extern __thread thread_allocated_t	thread_allocated_tls;
+#    define ALLOCATED_GET() (thread_allocated_tls.allocated)
+#    define ALLOCATEDP_GET() (&thread_allocated_tls.allocated)
+#    define DEALLOCATED_GET() (thread_allocated_tls.deallocated)
+#    define DEALLOCATEDP_GET() (&thread_allocated_tls.deallocated)
+#    define ALLOCATED_ADD(a, d) do {					\
+	thread_allocated_tls.allocated += a;				\
+	thread_allocated_tls.deallocated += d;				\
+} while (0)
+#  else
+extern pthread_key_t	thread_allocated_tsd;
+thread_allocated_t	*thread_allocated_get_hard(void);
+
+#    define ALLOCATED_GET() (thread_allocated_get()->allocated)
+#    define ALLOCATEDP_GET() (&thread_allocated_get()->allocated)
+#    define DEALLOCATED_GET() (thread_allocated_get()->deallocated)
+#    define DEALLOCATEDP_GET() (&thread_allocated_get()->deallocated)
+#    define ALLOCATED_ADD(a, d) do {					\
+	thread_allocated_t *thread_allocated = thread_allocated_get();	\
+	thread_allocated->allocated += (a);				\
+	thread_allocated->deallocated += (d);				\
+} while (0)
+#  endif
+#endif
+
+arena_t	*arenas_extend(unsigned ind);
+arena_t	*choose_arena_hard(void);
+int	buferror(int errnum, char *buf, size_t buflen);
+void	jemalloc_prefork(void);
+void	jemalloc_postfork(void);
+
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/prn.h"
+#include "jemalloc/internal/ckh.h"
+#include "jemalloc/internal/stats.h"
+#include "jemalloc/internal/ctl.h"
+#include "jemalloc/internal/mutex.h"
+#include "jemalloc/internal/mb.h"
+#include "jemalloc/internal/bitmap.h"
+#include "jemalloc/internal/extent.h"
+#include "jemalloc/internal/arena.h"
+#include "jemalloc/internal/base.h"
+#include "jemalloc/internal/chunk.h"
+#include "jemalloc/internal/huge.h"
+#include "jemalloc/internal/rtree.h"
+#include "jemalloc/internal/tcache.h"
+#include "jemalloc/internal/hash.h"
+#ifdef JEMALLOC_ZONE
+#include "jemalloc/internal/zone.h"
+#endif
+#include "jemalloc/internal/prof.h"
+
+#undef JEMALLOC_H_EXTERNS
+/******************************************************************************/
+#define JEMALLOC_H_INLINES
+
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/prn.h"
+#include "jemalloc/internal/ckh.h"
+#include "jemalloc/internal/stats.h"
+#include "jemalloc/internal/ctl.h"
+#include "jemalloc/internal/mutex.h"
+#include "jemalloc/internal/mb.h"
+#include "jemalloc/internal/extent.h"
+#include "jemalloc/internal/base.h"
+#include "jemalloc/internal/chunk.h"
+#include "jemalloc/internal/huge.h"
+
+#ifndef JEMALLOC_ENABLE_INLINE
+size_t	pow2_ceil(size_t x);
+size_t	s2u(size_t size);
+size_t	sa2u(size_t size, size_t alignment, size_t *run_size_p);
+void	malloc_write(const char *s);
+arena_t	*choose_arena(void);
+#  if (defined(JEMALLOC_STATS) && defined(NO_TLS))
+thread_allocated_t	*thread_allocated_get(void);
+#  endif
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_C_))
+/* Compute the smallest power of 2 that is >= x. */
+JEMALLOC_INLINE size_t
+pow2_ceil(size_t x)
+{
+
+	x--;
+	x |= x >> 1;
+	x |= x >> 2;
+	x |= x >> 4;
+	x |= x >> 8;
+	x |= x >> 16;
+#if (LG_SIZEOF_PTR == 3)
+	x |= x >> 32;
+#endif
+	x++;
+	return (x);
+}
+
+/*
+ * Compute usable size that would result from allocating an object with the
+ * specified size.
+ */
+JEMALLOC_INLINE size_t
+s2u(size_t size)
+{
+
+	if (size <= small_maxclass)
+		return (arena_bin_info[SMALL_SIZE2BIN(size)].reg_size);
+	if (size <= arena_maxclass)
+		return (PAGE_CEILING(size));
+	return (CHUNK_CEILING(size));
+}
+
+/*
+ * Compute usable size that would result from allocating an object with the
+ * specified size and alignment.
+ */
+JEMALLOC_INLINE size_t
+sa2u(size_t size, size_t alignment, size_t *run_size_p)
+{
+	size_t usize;
+
+	/*
+	 * Round size up to the nearest multiple of alignment.
+	 *
+	 * This done, we can take advantage of the fact that for each small
+	 * size class, every object is aligned at the smallest power of two
+	 * that is non-zero in the base two representation of the size.  For
+	 * example:
+	 *
+	 *   Size |   Base 2 | Minimum alignment
+	 *   -----+----------+------------------
+	 *     96 |  1100000 |  32
+	 *    144 | 10100000 |  32
+	 *    192 | 11000000 |  64
+	 *
+	 * Depending on runtime settings, it is possible that arena_malloc()
+	 * will further round up to a power of two, but that never causes
+	 * correctness issues.
+	 */
+	usize = (size + (alignment - 1)) & (-alignment);
+	/*
+	 * (usize < size) protects against the combination of maximal
+	 * alignment and size greater than maximal alignment.
+	 */
+	if (usize < size) {
+		/* size_t overflow. */
+		return (0);
+	}
+
+	if (usize <= arena_maxclass && alignment <= PAGE_SIZE) {
+		if (usize <= small_maxclass)
+			return (arena_bin_info[SMALL_SIZE2BIN(usize)].reg_size);
+		return (PAGE_CEILING(usize));
+	} else {
+		size_t run_size;
+
+		/*
+		 * We can't achieve subpage alignment, so round up alignment
+		 * permanently; it makes later calculations simpler.
+		 */
+		alignment = PAGE_CEILING(alignment);
+		usize = PAGE_CEILING(size);
+		/*
+		 * (usize < size) protects against very large sizes within
+		 * PAGE_SIZE of SIZE_T_MAX.
+		 *
+		 * (usize + alignment < usize) protects against the
+		 * combination of maximal alignment and usize large enough
+		 * to cause overflow.  This is similar to the first overflow
+		 * check above, but it needs to be repeated due to the new
+		 * usize value, which may now be *equal* to maximal
+		 * alignment, whereas before we only detected overflow if the
+		 * original size was *greater* than maximal alignment.
+		 */
+		if (usize < size || usize + alignment < usize) {
+			/* size_t overflow. */
+			return (0);
+		}
+
+		/*
+		 * Calculate the size of the over-size run that arena_palloc()
+		 * would need to allocate in order to guarantee the alignment.
+		 */
+		if (usize >= alignment)
+			run_size = usize + alignment - PAGE_SIZE;
+		else {
+			/*
+			 * It is possible that (alignment << 1) will cause
+			 * overflow, but it doesn't matter because we also
+			 * subtract PAGE_SIZE, which in the case of overflow
+			 * leaves us with a very large run_size.  That causes
+			 * the first conditional below to fail, which means
+			 * that the bogus run_size value never gets used for
+			 * anything important.
+			 */
+			run_size = (alignment << 1) - PAGE_SIZE;
+		}
+		if (run_size_p != NULL)
+			*run_size_p = run_size;
+
+		if (run_size <= arena_maxclass)
+			return (PAGE_CEILING(usize));
+		return (CHUNK_CEILING(usize));
+	}
+}
+
+/*
+ * Wrapper around malloc_message() that avoids the need for
+ * JEMALLOC_P(malloc_message)(...) throughout the code.
+ */
+JEMALLOC_INLINE void
+malloc_write(const char *s)
+{
+
+	JEMALLOC_P(malloc_message)(NULL, s);
+}
+
+/*
+ * Choose an arena based on a per-thread value (fast-path code, calls slow-path
+ * code if necessary).
+ */
+JEMALLOC_INLINE arena_t *
+choose_arena(void)
+{
+	arena_t *ret;
+
+	ret = ARENA_GET();
+	if (ret == NULL) {
+		ret = choose_arena_hard();
+		assert(ret != NULL);
+	}
+
+	return (ret);
+}
+
+#if (defined(JEMALLOC_STATS) && defined(NO_TLS))
+JEMALLOC_INLINE thread_allocated_t *
+thread_allocated_get(void)
+{
+	thread_allocated_t *thread_allocated = (thread_allocated_t *)
+	    pthread_getspecific(thread_allocated_tsd);
+
+	if (thread_allocated == NULL)
+		return (thread_allocated_get_hard());
+	return (thread_allocated);
+}
+#endif
+#endif
+
+#include "jemalloc/internal/bitmap.h"
+#include "jemalloc/internal/rtree.h"
+#include "jemalloc/internal/tcache.h"
+#include "jemalloc/internal/arena.h"
+#include "jemalloc/internal/hash.h"
+#ifdef JEMALLOC_ZONE
+#include "jemalloc/internal/zone.h"
+#endif
+
+#ifndef JEMALLOC_ENABLE_INLINE
+void	*imalloc(size_t size);
+void	*icalloc(size_t size);
+void	*ipalloc(size_t usize, size_t alignment, bool zero);
+size_t	isalloc(const void *ptr);
+#  ifdef JEMALLOC_IVSALLOC
+size_t	ivsalloc(const void *ptr);
+#  endif
+void	idalloc(void *ptr);
+void	*iralloc(void *ptr, size_t size, size_t extra, size_t alignment,
+    bool zero, bool no_move);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_C_))
+JEMALLOC_INLINE void *
+imalloc(size_t size)
+{
+
+	assert(size != 0);
+
+	if (size <= arena_maxclass)
+		return (arena_malloc(size, false));
+	else
+		return (huge_malloc(size, false));
+}
+
+JEMALLOC_INLINE void *
+icalloc(size_t size)
+{
+
+	if (size <= arena_maxclass)
+		return (arena_malloc(size, true));
+	else
+		return (huge_malloc(size, true));
+}
+
+JEMALLOC_INLINE void *
+ipalloc(size_t usize, size_t alignment, bool zero)
+{
+	void *ret;
+
+	assert(usize != 0);
+	assert(usize == sa2u(usize, alignment, NULL));
+
+	if (usize <= arena_maxclass && alignment <= PAGE_SIZE)
+		ret = arena_malloc(usize, zero);
+	else {
+		size_t run_size = 0;
+
+		/*
+		 * Ideally we would only ever call sa2u() once per aligned
+		 * allocation request, and the caller of this function has
+		 * already done so once.  However, it's rather burdensome to
+		 * require every caller to pass in run_size, especially given
+		 * that it's only relevant to large allocations.  Therefore,
+		 * just call it again here in order to get run_size.
+		 */
+		sa2u(usize, alignment, &run_size);
+		if (run_size <= arena_maxclass) {
+			ret = arena_palloc(choose_arena(), usize, run_size,
+			    alignment, zero);
+		} else if (alignment <= chunksize)
+			ret = huge_malloc(usize, zero);
+		else
+			ret = huge_palloc(usize, alignment, zero);
+	}
+
+	assert(((uintptr_t)ret & (alignment - 1)) == 0);
+	return (ret);
+}
+
+JEMALLOC_INLINE size_t
+isalloc(const void *ptr)
+{
+	size_t ret;
+	arena_chunk_t *chunk;
+
+	assert(ptr != NULL);
+
+	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
+	if (chunk != ptr) {
+		/* Region. */
+		dassert(chunk->arena->magic == ARENA_MAGIC);
+
+#ifdef JEMALLOC_PROF
+		ret = arena_salloc_demote(ptr);
+#else
+		ret = arena_salloc(ptr);
+#endif
+	} else
+		ret = huge_salloc(ptr);
+
+	return (ret);
+}
+
+#ifdef JEMALLOC_IVSALLOC
+JEMALLOC_INLINE size_t
+ivsalloc(const void *ptr)
+{
+
+	/* Return 0 if ptr is not within a chunk managed by jemalloc. */
+	if (rtree_get(chunks_rtree, (uintptr_t)CHUNK_ADDR2BASE(ptr)) == NULL)
+		return (0);
+
+	return (isalloc(ptr));
+}
+#endif
+
+JEMALLOC_INLINE void
+idalloc(void *ptr)
+{
+	arena_chunk_t *chunk;
+
+	assert(ptr != NULL);
+
+	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
+	if (chunk != ptr)
+		arena_dalloc(chunk->arena, chunk, ptr);
+	else
+		huge_dalloc(ptr, true);
+}
+
+JEMALLOC_INLINE void *
+iralloc(void *ptr, size_t size, size_t extra, size_t alignment, bool zero,
+    bool no_move)
+{
+	void *ret;
+	size_t oldsize;
+
+	assert(ptr != NULL);
+	assert(size != 0);
+
+	oldsize = isalloc(ptr);
+
+	if (alignment != 0 && ((uintptr_t)ptr & ((uintptr_t)alignment-1))
+	    != 0) {
+		size_t usize, copysize;
+
+		/*
+		 * Existing object alignment is inadquate; allocate new space
+		 * and copy.
+		 */
+		if (no_move)
+			return (NULL);
+		usize = sa2u(size + extra, alignment, NULL);
+		if (usize == 0)
+			return (NULL);
+		ret = ipalloc(usize, alignment, zero);
+		if (ret == NULL) {
+			if (extra == 0)
+				return (NULL);
+			/* Try again, without extra this time. */
+			usize = sa2u(size, alignment, NULL);
+			if (usize == 0)
+				return (NULL);
+			ret = ipalloc(usize, alignment, zero);
+			if (ret == NULL)
+				return (NULL);
+		}
+		/*
+		 * Copy at most size bytes (not size+extra), since the caller
+		 * has no expectation that the extra bytes will be reliably
+		 * preserved.
+		 */
+		copysize = (size < oldsize) ? size : oldsize;
+		memcpy(ret, ptr, copysize);
+		idalloc(ptr);
+		return (ret);
+	}
+
+	if (no_move) {
+		if (size <= arena_maxclass) {
+			return (arena_ralloc_no_move(ptr, oldsize, size,
+			    extra, zero));
+		} else {
+			return (huge_ralloc_no_move(ptr, oldsize, size,
+			    extra));
+		}
+	} else {
+		if (size + extra <= arena_maxclass) {
+			return (arena_ralloc(ptr, oldsize, size, extra,
+			    alignment, zero));
+		} else {
+			return (huge_ralloc(ptr, oldsize, size, extra,
+			    alignment, zero));
+		}
+	}
+}
+#endif
+
+#include "jemalloc/internal/prof.h"
+
+#undef JEMALLOC_H_INLINES
+/******************************************************************************/

deps/jemalloc/include/jemalloc/internal/mb.h

@@ -0,0 +1,108 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#ifndef JEMALLOC_ENABLE_INLINE
+void	mb_write(void);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_MB_C_))
+#ifdef __i386__
+/*
+ * According to the Intel Architecture Software Developer's Manual, current
+ * processors execute instructions in order from the perspective of other
+ * processors in a multiprocessor system, but 1) Intel reserves the right to
+ * change that, and 2) the compiler's optimizer could re-order instructions if
+ * there weren't some form of barrier.  Therefore, even if running on an
+ * architecture that does not need memory barriers (everything through at least
+ * i686), an "optimizer barrier" is necessary.
+ */
+JEMALLOC_INLINE void
+mb_write(void)
+{
+
+#  if 0
+	/* This is a true memory barrier. */
+	asm volatile ("pusha;"
+	    "xor  %%eax,%%eax;"
+	    "cpuid;"
+	    "popa;"
+	    : /* Outputs. */
+	    : /* Inputs. */
+	    : "memory" /* Clobbers. */
+	    );
+#else
+	/*
+	 * This is hopefully enough to keep the compiler from reordering
+	 * instructions around this one.
+	 */
+	asm volatile ("nop;"
+	    : /* Outputs. */
+	    : /* Inputs. */
+	    : "memory" /* Clobbers. */
+	    );
+#endif
+}
+#elif (defined(__amd64_) || defined(__x86_64__))
+JEMALLOC_INLINE void
+mb_write(void)
+{
+
+	asm volatile ("sfence"
+	    : /* Outputs. */
+	    : /* Inputs. */
+	    : "memory" /* Clobbers. */
+	    );
+}
+#elif defined(__powerpc__)
+JEMALLOC_INLINE void
+mb_write(void)
+{
+
+	asm volatile ("eieio"
+	    : /* Outputs. */
+	    : /* Inputs. */
+	    : "memory" /* Clobbers. */
+	    );
+}
+#elif defined(__sparc64__)
+JEMALLOC_INLINE void
+mb_write(void)
+{
+
+	asm volatile ("membar #StoreStore"
+	    : /* Outputs. */
+	    : /* Inputs. */
+	    : "memory" /* Clobbers. */
+	    );
+}
+#else
+/*
+ * This is much slower than a simple memory barrier, but the semantics of mutex
+ * unlock make this work.
+ */
+JEMALLOC_INLINE void
+mb_write(void)
+{
+	malloc_mutex_t mtx;
+
+	malloc_mutex_init(&mtx);
+	malloc_mutex_lock(&mtx);
+	malloc_mutex_unlock(&mtx);
+}
+#endif
+#endif
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/

deps/jemalloc/include/jemalloc/internal/mutex.h

@@ -0,0 +1,86 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+#ifdef JEMALLOC_OSSPIN
+typedef OSSpinLock malloc_mutex_t;
+#else
+typedef pthread_mutex_t malloc_mutex_t;
+#endif
+
+#ifdef PTHREAD_ADAPTIVE_MUTEX_INITIALIZER_NP
+#  define MALLOC_MUTEX_INITIALIZER PTHREAD_ADAPTIVE_MUTEX_INITIALIZER_NP
+#else
+#  define MALLOC_MUTEX_INITIALIZER PTHREAD_MUTEX_INITIALIZER
+#endif
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+#ifdef JEMALLOC_LAZY_LOCK
+extern bool isthreaded;
+#else
+#  define isthreaded true
+#endif
+
+bool	malloc_mutex_init(malloc_mutex_t *mutex);
+void	malloc_mutex_destroy(malloc_mutex_t *mutex);
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#ifndef JEMALLOC_ENABLE_INLINE
+void	malloc_mutex_lock(malloc_mutex_t *mutex);
+bool	malloc_mutex_trylock(malloc_mutex_t *mutex);
+void	malloc_mutex_unlock(malloc_mutex_t *mutex);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_MUTEX_C_))
+JEMALLOC_INLINE void
+malloc_mutex_lock(malloc_mutex_t *mutex)
+{
+
+	if (isthreaded) {
+#ifdef JEMALLOC_OSSPIN
+		OSSpinLockLock(mutex);
+#else
+		pthread_mutex_lock(mutex);
+#endif
+	}
+}
+
+JEMALLOC_INLINE bool
+malloc_mutex_trylock(malloc_mutex_t *mutex)
+{
+
+	if (isthreaded) {
+#ifdef JEMALLOC_OSSPIN
+		return (OSSpinLockTry(mutex) == false);
+#else
+		return (pthread_mutex_trylock(mutex) != 0);
+#endif
+	} else
+		return (false);
+}
+
+JEMALLOC_INLINE void
+malloc_mutex_unlock(malloc_mutex_t *mutex)
+{
+
+	if (isthreaded) {
+#ifdef JEMALLOC_OSSPIN
+		OSSpinLockUnlock(mutex);
+#else
+		pthread_mutex_unlock(mutex);
+#endif
+	}
+}
+#endif
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/

deps/jemalloc/include/jemalloc/internal/prn.h

@@ -0,0 +1,60 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+/*
+ * Simple linear congruential pseudo-random number generator:
+ *
+ *   prn(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.
+ *
+ * Macro parameters:
+ *   uint32_t r          : Result.
+ *   unsigned lg_range   : (0..32], number of least significant bits to return.
+ *   uint32_t state      : Seed value.
+ *   const uint32_t a, c : See above discussion.
+ */
+#define prn32(r, lg_range, state, a, c) do {				\
+	assert(lg_range > 0);						\
+	assert(lg_range <= 32);						\
+									\
+	r = (state * (a)) + (c);					\
+	state = r;							\
+	r >>= (32 - lg_range);						\
+} while (false)
+
+/* Same as prn32(), but 64 bits of pseudo-randomness, using uint64_t. */
+#define prn64(r, lg_range, state, a, c) do {				\
+	assert(lg_range > 0);						\
+	assert(lg_range <= 64);						\
+									\
+	r = (state * (a)) + (c);					\
+	state = r;							\
+	r >>= (64 - lg_range);						\
+} while (false)
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/

deps/jemalloc/include/jemalloc/internal/prof.h

@@ -0,0 +1,561 @@
+#ifdef JEMALLOC_PROF
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+typedef struct prof_bt_s prof_bt_t;
+typedef struct prof_cnt_s prof_cnt_t;
+typedef struct prof_thr_cnt_s prof_thr_cnt_t;
+typedef struct prof_ctx_s prof_ctx_t;
+typedef struct prof_tdata_s prof_tdata_t;
+
+/* Option defaults. */
+#define	PROF_PREFIX_DEFAULT		"jeprof"
+#define	LG_PROF_BT_MAX_DEFAULT		7
+#define	LG_PROF_SAMPLE_DEFAULT		0
+#define	LG_PROF_INTERVAL_DEFAULT	-1
+#define	LG_PROF_TCMAX_DEFAULT		-1
+
+/*
+ * Hard limit on stack backtrace depth.  Note that the version of
+ * prof_backtrace() that is based on __builtin_return_address() necessarily has
+ * a hard-coded number of backtrace frame handlers.
+ */
+#if (defined(JEMALLOC_PROF_LIBGCC) || defined(JEMALLOC_PROF_LIBUNWIND))
+#  define LG_PROF_BT_MAX	((ZU(1) << (LG_SIZEOF_PTR+3)) - 1)
+#else
+#  define LG_PROF_BT_MAX	7 /* >= LG_PROF_BT_MAX_DEFAULT */
+#endif
+#define	PROF_BT_MAX		(1U << LG_PROF_BT_MAX)
+
+/* Initial hash table size. */
+#define	PROF_CKH_MINITEMS	64
+
+/* Size of memory buffer to use when writing dump files. */
+#define	PROF_DUMP_BUF_SIZE	65536
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+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	nignore;
+	unsigned	max;
+} prof_unwind_data_t;
+#endif
+
+struct prof_cnt_s {
+	/*
+	 * Profiling counters.  An allocation/deallocation pair can operate on
+	 * different prof_thr_cnt_t objects that are linked into the same
+	 * prof_ctx_t cnts_ql, so it is possible for the cur* counters to go
+	 * negative.  In principle it is possible for the *bytes counters to
+	 * overflow/underflow, but a general solution would require something
+	 * like 128-bit counters; this implementation doesn't bother to solve
+	 * that problem.
+	 */
+	int64_t		curobjs;
+	int64_t		curbytes;
+	uint64_t	accumobjs;
+	uint64_t	accumbytes;
+};
+
+struct prof_thr_cnt_s {
+	/* Linkage into prof_ctx_t's cnts_ql. */
+	ql_elm(prof_thr_cnt_t)	cnts_link;
+
+	/* Linkage into thread's LRU. */
+	ql_elm(prof_thr_cnt_t)	lru_link;
+
+	/*
+	 * Associated context.  If a thread frees an object that it did not
+	 * allocate, it is possible that the context is not cached in the
+	 * thread's hash table, in which case it must be able to look up the
+	 * context, insert a new prof_thr_cnt_t into the thread's hash table,
+	 * and link it into the prof_ctx_t's cnts_ql.
+	 */
+	prof_ctx_t		*ctx;
+
+	/*
+	 * Threads use memory barriers to update the counters.  Since there is
+	 * only ever one writer, the only challenge is for the reader to get a
+	 * consistent read of the counters.
+	 *
+	 * The writer uses this series of operations:
+	 *
+	 * 1) Increment epoch to an odd number.
+	 * 2) Update counters.
+	 * 3) Increment epoch to an even number.
+	 *
+	 * The reader must assure 1) that the epoch is even while it reads the
+	 * counters, and 2) that the epoch doesn't change between the time it
+	 * starts and finishes reading the counters.
+	 */
+	unsigned		epoch;
+
+	/* Profiling counters. */
+	prof_cnt_t		cnts;
+};
+
+struct prof_ctx_s {
+	/* Associated backtrace. */
+	prof_bt_t		*bt;
+
+	/* Protects cnt_merged and cnts_ql. */
+	malloc_mutex_t		lock;
+
+	/* Temporary storage for summation during dump. */
+	prof_cnt_t		cnt_summed;
+
+	/* When threads exit, they merge their stats into cnt_merged. */
+	prof_cnt_t		cnt_merged;
+
+	/*
+	 * List of profile counters, one for each thread that has allocated in
+	 * this context.
+	 */
+	ql_head(prof_thr_cnt_t)	cnts_ql;
+};
+
+struct prof_tdata_s {
+	/*
+	 * Hash of (prof_bt_t *)-->(prof_thr_cnt_t *).  Each thread keeps a
+	 * cache of backtraces, with associated thread-specific prof_thr_cnt_t
+	 * objects.  Other threads may read the prof_thr_cnt_t contents, but no
+	 * others will ever write them.
+	 *
+	 * Upon thread exit, the thread must merge all the prof_thr_cnt_t
+	 * counter data into the associated prof_ctx_t objects, and unlink/free
+	 * the prof_thr_cnt_t objects.
+	 */
+	ckh_t			bt2cnt;
+
+	/* LRU for contents of bt2cnt. */
+	ql_head(prof_thr_cnt_t)	lru_ql;
+
+	/* Backtrace vector, used for calls to prof_backtrace(). */
+	void			**vec;
+
+	/* Sampling state. */
+	uint64_t		prn_state;
+	uint64_t		threshold;
+	uint64_t		accum;
+};
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+extern bool	opt_prof;
+/*
+ * Even if opt_prof is true, sampling can be temporarily disabled by setting
+ * opt_prof_active to false.  No locking is used when updating opt_prof_active,
+ * so there are no guarantees regarding how long it will take for all threads
+ * to notice state changes.
+ */
+extern bool	opt_prof_active;
+extern size_t	opt_lg_prof_bt_max;   /* Maximum backtrace depth. */
+extern size_t	opt_lg_prof_sample;   /* Mean bytes between samples. */
+extern ssize_t	opt_lg_prof_interval; /* lg(prof_interval). */
+extern bool	opt_prof_gdump;       /* High-water memory dumping. */
+extern bool	opt_prof_leak;        /* Dump leak summary at exit. */
+extern bool	opt_prof_accum;       /* Report cumulative bytes. */
+extern ssize_t	opt_lg_prof_tcmax;    /* lg(max per thread bactrace cache) */
+extern char	opt_prof_prefix[PATH_MAX + 1];
+
+/*
+ * 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;
+
+/*
+ * If true, promote small sampled objects to large objects, since small run
+ * headers do not have embedded profile context pointers.
+ */
+extern bool	prof_promote;
+
+/* (1U << opt_lg_prof_bt_max). */
+extern unsigned	prof_bt_max;
+
+/* Thread-specific backtrace cache, used to reduce bt2ctx contention. */
+#ifndef NO_TLS
+extern __thread prof_tdata_t	*prof_tdata_tls
+    JEMALLOC_ATTR(tls_model("initial-exec"));
+#  define PROF_TCACHE_GET()	prof_tdata_tls
+#  define PROF_TCACHE_SET(v)	do {					\
+	prof_tdata_tls = (v);						\
+	pthread_setspecific(prof_tdata_tsd, (void *)(v));		\
+} while (0)
+#else
+#  define PROF_TCACHE_GET()						\
+	((prof_tdata_t *)pthread_getspecific(prof_tdata_tsd))
+#  define PROF_TCACHE_SET(v)	do {					\
+	pthread_setspecific(prof_tdata_tsd, (void *)(v));		\
+} while (0)
+#endif
+/*
+ * Same contents as b2cnt_tls, but initialized such that the TSD destructor is
+ * called when a thread exits, so that prof_tdata_tls contents can be merged,
+ * unlinked, and deallocated.
+ */
+extern pthread_key_t	prof_tdata_tsd;
+
+void	bt_init(prof_bt_t *bt, void **vec);
+void	prof_backtrace(prof_bt_t *bt, unsigned nignore, unsigned max);
+prof_thr_cnt_t	*prof_lookup(prof_bt_t *bt);
+void	prof_idump(void);
+bool	prof_mdump(const char *filename);
+void	prof_gdump(void);
+prof_tdata_t	*prof_tdata_init(void);
+void	prof_boot0(void);
+void	prof_boot1(void);
+bool	prof_boot2(void);
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#ifndef JEMALLOC_ENABLE_INLINE
+void	prof_sample_threshold_update(prof_tdata_t *prof_tdata);
+prof_thr_cnt_t	*prof_alloc_prep(size_t size);
+prof_ctx_t	*prof_ctx_get(const void *ptr);
+void	prof_ctx_set(const void *ptr, prof_ctx_t *ctx);
+bool	prof_sample_accum_update(size_t size);
+void	prof_malloc(const void *ptr, size_t size, prof_thr_cnt_t *cnt);
+void	prof_realloc(const void *ptr, size_t size, prof_thr_cnt_t *cnt,
+    size_t old_size, prof_ctx_t *old_ctx);
+void	prof_free(const void *ptr, size_t size);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_PROF_C_))
+JEMALLOC_INLINE void
+prof_sample_threshold_update(prof_tdata_t *prof_tdata)
+{
+	uint64_t r;
+	double u;
+
+	/*
+	 * Compute sample threshold as a geometrically distributed random
+	 * variable with mean (2^opt_lg_prof_sample).
+	 *
+	 *                         __        __
+	 *                         |  log(u)  |                     1
+	 * prof_tdata->threshold = | -------- |, where p = -------------------
+	 *                         | log(1-p) |             opt_lg_prof_sample
+	 *                                                 2
+	 *
+	 * For more information on the math, see:
+	 *
+	 *   Non-Uniform Random Variate Generation
+	 *   Luc Devroye
+	 *   Springer-Verlag, New York, 1986
+	 *   pp 500
+	 *   (http://cg.scs.carleton.ca/~luc/rnbookindex.html)
+	 */
+	prn64(r, 53, prof_tdata->prn_state,
+	    (uint64_t)6364136223846793005LLU, (uint64_t)1442695040888963407LLU);
+	u = (double)r * (1.0/9007199254740992.0L);
+	prof_tdata->threshold = (uint64_t)(log(u) /
+	    log(1.0 - (1.0 / (double)((uint64_t)1U << opt_lg_prof_sample))))
+	    + (uint64_t)1U;
+}
+
+JEMALLOC_INLINE prof_thr_cnt_t *
+prof_alloc_prep(size_t size)
+{
+#ifdef JEMALLOC_ENABLE_INLINE
+   /* This function does not have its own stack frame, because it is inlined. */
+#  define NIGNORE 1
+#else
+#  define NIGNORE 2
+#endif
+	prof_thr_cnt_t *ret;
+	prof_tdata_t *prof_tdata;
+	prof_bt_t bt;
+
+	assert(size == s2u(size));
+
+	prof_tdata = PROF_TCACHE_GET();
+	if (prof_tdata == NULL) {
+		prof_tdata = prof_tdata_init();
+		if (prof_tdata == NULL)
+			return (NULL);
+	}
+
+	if (opt_prof_active == false) {
+		/* Sampling is currently inactive, so avoid sampling. */
+		ret = (prof_thr_cnt_t *)(uintptr_t)1U;
+	} else if (opt_lg_prof_sample == 0) {
+		/*
+		 * Don't bother with sampling logic, since sampling interval is
+		 * 1.
+		 */
+		bt_init(&bt, prof_tdata->vec);
+		prof_backtrace(&bt, NIGNORE, prof_bt_max);
+		ret = prof_lookup(&bt);
+	} else {
+		if (prof_tdata->threshold == 0) {
+			/*
+			 * Initialize.  Seed the prng differently for each
+			 * thread.
+			 */
+			prof_tdata->prn_state = (uint64_t)(uintptr_t)&size;
+			prof_sample_threshold_update(prof_tdata);
+		}
+
+		/*
+		 * Determine whether to capture a backtrace based on whether
+		 * size is enough for prof_accum to reach
+		 * prof_tdata->threshold.  However, delay updating these
+		 * variables until prof_{m,re}alloc(), because we don't know
+		 * for sure that the allocation will succeed.
+		 *
+		 * Use subtraction rather than addition to avoid potential
+		 * integer overflow.
+		 */
+		if (size >= prof_tdata->threshold - prof_tdata->accum) {
+			bt_init(&bt, prof_tdata->vec);
+			prof_backtrace(&bt, NIGNORE, prof_bt_max);
+			ret = prof_lookup(&bt);
+		} else
+			ret = (prof_thr_cnt_t *)(uintptr_t)1U;
+	}
+
+	return (ret);
+#undef NIGNORE
+}
+
+JEMALLOC_INLINE prof_ctx_t *
+prof_ctx_get(const void *ptr)
+{
+	prof_ctx_t *ret;
+	arena_chunk_t *chunk;
+
+	assert(ptr != NULL);
+
+	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
+	if (chunk != ptr) {
+		/* Region. */
+		dassert(chunk->arena->magic == ARENA_MAGIC);
+
+		ret = arena_prof_ctx_get(ptr);
+	} else
+		ret = huge_prof_ctx_get(ptr);
+
+	return (ret);
+}
+
+JEMALLOC_INLINE void
+prof_ctx_set(const void *ptr, prof_ctx_t *ctx)
+{
+	arena_chunk_t *chunk;
+
+	assert(ptr != NULL);
+
+	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
+	if (chunk != ptr) {
+		/* Region. */
+		dassert(chunk->arena->magic == ARENA_MAGIC);
+
+		arena_prof_ctx_set(ptr, ctx);
+	} else
+		huge_prof_ctx_set(ptr, ctx);
+}
+
+JEMALLOC_INLINE bool
+prof_sample_accum_update(size_t size)
+{
+	prof_tdata_t *prof_tdata;
+
+	/* Sampling logic is unnecessary if the interval is 1. */
+	assert(opt_lg_prof_sample != 0);
+
+	prof_tdata = PROF_TCACHE_GET();
+	assert(prof_tdata != NULL);
+
+	/* Take care to avoid integer overflow. */
+	if (size >= prof_tdata->threshold - prof_tdata->accum) {
+		prof_tdata->accum -= (prof_tdata->threshold - size);
+		/* Compute new sample threshold. */
+		prof_sample_threshold_update(prof_tdata);
+		while (prof_tdata->accum >= prof_tdata->threshold) {
+			prof_tdata->accum -= prof_tdata->threshold;
+			prof_sample_threshold_update(prof_tdata);
+		}
+		return (false);
+	} else {
+		prof_tdata->accum += size;
+		return (true);
+	}
+}
+
+JEMALLOC_INLINE void
+prof_malloc(const void *ptr, size_t size, prof_thr_cnt_t *cnt)
+{
+
+	assert(ptr != NULL);
+	assert(size == isalloc(ptr));
+
+	if (opt_lg_prof_sample != 0) {
+		if (prof_sample_accum_update(size)) {
+			/*
+			 * Don't sample.  For malloc()-like allocation, it is
+			 * always possible to tell in advance how large an
+			 * object's usable size will be, so there should never
+			 * be a difference between the size passed to
+			 * prof_alloc_prep() and prof_malloc().
+			 */
+			assert((uintptr_t)cnt == (uintptr_t)1U);
+		}
+	}
+
+	if ((uintptr_t)cnt > (uintptr_t)1U) {
+		prof_ctx_set(ptr, cnt->ctx);
+
+		cnt->epoch++;
+		/*********/
+		mb_write();
+		/*********/
+		cnt->cnts.curobjs++;
+		cnt->cnts.curbytes += size;
+		if (opt_prof_accum) {
+			cnt->cnts.accumobjs++;
+			cnt->cnts.accumbytes += size;
+		}
+		/*********/
+		mb_write();
+		/*********/
+		cnt->epoch++;
+		/*********/
+		mb_write();
+		/*********/
+	} else
+		prof_ctx_set(ptr, (prof_ctx_t *)(uintptr_t)1U);
+}
+
+JEMALLOC_INLINE void
+prof_realloc(const void *ptr, size_t size, prof_thr_cnt_t *cnt,
+    size_t old_size, prof_ctx_t *old_ctx)
+{
+	prof_thr_cnt_t *told_cnt;
+
+	assert(ptr != NULL || (uintptr_t)cnt <= (uintptr_t)1U);
+
+	if (ptr != NULL) {
+		assert(size == isalloc(ptr));
+		if (opt_lg_prof_sample != 0) {
+			if (prof_sample_accum_update(size)) {
+				/*
+				 * Don't sample.  The size passed to
+				 * prof_alloc_prep() was larger than what
+				 * actually got allocated, so a backtrace was
+				 * captured for this allocation, even though
+				 * its actual size was insufficient to cross
+				 * the sample threshold.
+				 */
+				cnt = (prof_thr_cnt_t *)(uintptr_t)1U;
+			}
+		}
+	}
+
+	if ((uintptr_t)old_ctx > (uintptr_t)1U) {
+		told_cnt = prof_lookup(old_ctx->bt);
+		if (told_cnt == NULL) {
+			/*
+			 * It's too late to propagate OOM for this realloc(),
+			 * so operate directly on old_cnt->ctx->cnt_merged.
+			 */
+			malloc_mutex_lock(&old_ctx->lock);
+			old_ctx->cnt_merged.curobjs--;
+			old_ctx->cnt_merged.curbytes -= old_size;
+			malloc_mutex_unlock(&old_ctx->lock);
+			told_cnt = (prof_thr_cnt_t *)(uintptr_t)1U;
+		}
+	} else
+		told_cnt = (prof_thr_cnt_t *)(uintptr_t)1U;
+
+	if ((uintptr_t)told_cnt > (uintptr_t)1U)
+		told_cnt->epoch++;
+	if ((uintptr_t)cnt > (uintptr_t)1U) {
+		prof_ctx_set(ptr, cnt->ctx);
+		cnt->epoch++;
+	} else
+		prof_ctx_set(ptr, (prof_ctx_t *)(uintptr_t)1U);
+	/*********/
+	mb_write();
+	/*********/
+	if ((uintptr_t)told_cnt > (uintptr_t)1U) {
+		told_cnt->cnts.curobjs--;
+		told_cnt->cnts.curbytes -= old_size;
+	}
+	if ((uintptr_t)cnt > (uintptr_t)1U) {
+		cnt->cnts.curobjs++;
+		cnt->cnts.curbytes += size;
+		if (opt_prof_accum) {
+			cnt->cnts.accumobjs++;
+			cnt->cnts.accumbytes += size;
+		}
+	}
+	/*********/
+	mb_write();
+	/*********/
+	if ((uintptr_t)told_cnt > (uintptr_t)1U)
+		told_cnt->epoch++;
+	if ((uintptr_t)cnt > (uintptr_t)1U)
+		cnt->epoch++;
+	/*********/
+	mb_write(); /* Not strictly necessary. */
+}
+
+JEMALLOC_INLINE void
+prof_free(const void *ptr, size_t size)
+{
+	prof_ctx_t *ctx = prof_ctx_get(ptr);
+
+	if ((uintptr_t)ctx > (uintptr_t)1) {
+		assert(size == isalloc(ptr));
+		prof_thr_cnt_t *tcnt = prof_lookup(ctx->bt);
+
+		if (tcnt != NULL) {
+			tcnt->epoch++;
+			/*********/
+			mb_write();
+			/*********/
+			tcnt->cnts.curobjs--;
+			tcnt->cnts.curbytes -= size;
+			/*********/
+			mb_write();
+			/*********/
+			tcnt->epoch++;
+			/*********/
+			mb_write();
+			/*********/
+		} else {
+			/*
+			 * OOM during free() cannot be propagated, so operate
+			 * directly on cnt->ctx->cnt_merged.
+			 */
+			malloc_mutex_lock(&ctx->lock);
+			ctx->cnt_merged.curobjs--;
+			ctx->cnt_merged.curbytes -= size;
+			malloc_mutex_unlock(&ctx->lock);
+		}
+	}
+}
+#endif
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
+#endif /* JEMALLOC_PROF */

deps/jemalloc/include/jemalloc/internal/ql.h

@@ -0,0 +1,83 @@
+/*
+ * List definitions.
+ */
+#define ql_head(a_type)							\
+struct {								\
+	a_type *qlh_first;						\
+}
+
+#define ql_head_initializer(a_head) {NULL}
+
+#define ql_elm(a_type)	qr(a_type)
+
+/* List functions. */
+#define ql_new(a_head) do {						\
+	(a_head)->qlh_first = NULL;					\
+} while (0)
+
+#define ql_elm_new(a_elm, a_field) qr_new((a_elm), a_field)
+
+#define ql_first(a_head) ((a_head)->qlh_first)
+
+#define ql_last(a_head, a_field)					\
+	((ql_first(a_head) != NULL)					\
+	    ? qr_prev(ql_first(a_head), a_field) : NULL)
+
+#define ql_next(a_head, a_elm, a_field)					\
+	((ql_last(a_head, a_field) != (a_elm))				\
+	    ? qr_next((a_elm), a_field)	: NULL)
+
+#define ql_prev(a_head, a_elm, a_field)					\
+	((ql_first(a_head) != (a_elm)) ? qr_prev((a_elm), a_field)	\
+				       : NULL)
+
+#define ql_before_insert(a_head, a_qlelm, a_elm, a_field) do {		\
+	qr_before_insert((a_qlelm), (a_elm), a_field);			\
+	if (ql_first(a_head) == (a_qlelm)) {				\
+		ql_first(a_head) = (a_elm);				\
+	}								\
+} while (0)
+
+#define ql_after_insert(a_qlelm, a_elm, a_field)			\
+	qr_after_insert((a_qlelm), (a_elm), a_field)
+
+#define ql_head_insert(a_head, a_elm, a_field) do {			\
+	if (ql_first(a_head) != NULL) {					\
+		qr_before_insert(ql_first(a_head), (a_elm), a_field);	\
+	}								\
+	ql_first(a_head) = (a_elm);					\
+} while (0)
+
+#define ql_tail_insert(a_head, a_elm, a_field) do {			\
+	if (ql_first(a_head) != NULL) {					\
+		qr_before_insert(ql_first(a_head), (a_elm), a_field);	\
+	}								\
+	ql_first(a_head) = qr_next((a_elm), a_field);			\
+} while (0)
+
+#define ql_remove(a_head, a_elm, a_field) do {				\
+	if (ql_first(a_head) == (a_elm)) {				\
+		ql_first(a_head) = qr_next(ql_first(a_head), a_field);	\
+	}								\
+	if (ql_first(a_head) != (a_elm)) {				\
+		qr_remove((a_elm), a_field);				\
+	} else {							\
+		ql_first(a_head) = NULL;				\
+	}								\
+} while (0)
+
+#define ql_head_remove(a_head, a_type, a_field) do {			\
+	a_type *t = ql_first(a_head);					\
+	ql_remove((a_head), t, a_field);				\
+} while (0)
+
+#define ql_tail_remove(a_head, a_type, a_field) do {			\
+	a_type *t = ql_last(a_head, a_field);				\
+	ql_remove((a_head), t, a_field);				\
+} while (0)
+
+#define ql_foreach(a_var, a_head, a_field)				\
+	qr_foreach((a_var), ql_first(a_head), a_field)
+
+#define ql_reverse_foreach(a_var, a_head, a_field)			\
+	qr_reverse_foreach((a_var), ql_first(a_head), a_field)

deps/jemalloc/include/jemalloc/internal/qr.h

@@ -0,0 +1,67 @@
+/* Ring definitions. */
+#define qr(a_type)							\
+struct {								\
+	a_type	*qre_next;						\
+	a_type	*qre_prev;						\
+}
+
+/* Ring functions. */
+#define qr_new(a_qr, a_field) do {					\
+	(a_qr)->a_field.qre_next = (a_qr);				\
+	(a_qr)->a_field.qre_prev = (a_qr);				\
+} while (0)
+
+#define qr_next(a_qr, a_field) ((a_qr)->a_field.qre_next)
+
+#define qr_prev(a_qr, a_field) ((a_qr)->a_field.qre_prev)
+
+#define qr_before_insert(a_qrelm, a_qr, a_field) do {			\
+	(a_qr)->a_field.qre_prev = (a_qrelm)->a_field.qre_prev;		\
+	(a_qr)->a_field.qre_next = (a_qrelm);				\
+	(a_qr)->a_field.qre_prev->a_field.qre_next = (a_qr);		\
+	(a_qrelm)->a_field.qre_prev = (a_qr);				\
+} while (0)
+
+#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)
+
+#define qr_meld(a_qr_a, a_qr_b, a_field) do {				\
+	void *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;					\
+	(a_qr_a)->a_field.qre_prev = (a_qr_b)->a_field.qre_prev;	\
+	(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)
+
+#define qr_remove(a_qr, a_field) do {					\
+	(a_qr)->a_field.qre_prev->a_field.qre_next			\
+	    = (a_qr)->a_field.qre_next;					\
+	(a_qr)->a_field.qre_next->a_field.qre_prev			\
+	    = (a_qr)->a_field.qre_prev;					\
+	(a_qr)->a_field.qre_next = (a_qr);				\
+	(a_qr)->a_field.qre_prev = (a_qr);				\
+} while (0)
+
+#define qr_foreach(var, a_qr, a_field)					\
+	for ((var) = (a_qr);						\
+	    (var) != NULL;						\
+	    (var) = (((var)->a_field.qre_next != (a_qr))		\
+	    ? (var)->a_field.qre_next : NULL))
+
+#define qr_reverse_foreach(var, a_qr, a_field)				\
+	for ((var) = ((a_qr) != NULL) ? qr_prev(a_qr, a_field) : NULL;	\
+	    (var) != NULL;						\
+	    (var) = (((var) != (a_qr))					\
+	    ? (var)->a_field.qre_prev : NULL))

deps/jemalloc/include/jemalloc/internal/rb.h

@@ -0,0 +1,973 @@
+/*-
+ *******************************************************************************
+ *
+ * cpp macro implementation of left-leaning 2-3 red-black trees.  Parent
+ * pointers are not used, and color bits are stored in the least significant
+ * bit of right-child pointers (if RB_COMPACT is defined), thus making node
+ * linkage as compact as is possible for red-black trees.
+ *
+ * Usage:
+ *
+ *   #include <stdint.h>
+ *   #include <stdbool.h>
+ *   #define NDEBUG // (Optional, see assert(3).)
+ *   #include <assert.h>
+ *   #define RB_COMPACT // (Optional, embed color bits in right-child pointers.)
+ *   #include <rb.h>
+ *   ...
+ *
+ *******************************************************************************
+ */
+
+#ifndef RB_H_
+#define	RB_H_
+
+#if 0
+__FBSDID("$FreeBSD: head/lib/libc/stdlib/rb.h 204493 2010-02-28 22:57:13Z jasone $");
+#endif
+
+#ifdef RB_COMPACT
+/* Node structure. */
+#define	rb_node(a_type)							\
+struct {								\
+    a_type *rbn_left;							\
+    a_type *rbn_right_red;						\
+}
+#else
+#define	rb_node(a_type)							\
+struct {								\
+    a_type *rbn_left;							\
+    a_type *rbn_right;							\
+    bool rbn_red;							\
+}
+#endif
+
+/* Root structure. */
+#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)				\
+    ((a_node)->a_field.rbn_left)
+#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)				\
+    ((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 {		\
+    (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)				\
+    ((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 {		\
+    (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 {			\
+    (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 {			\
+    (a_node)->a_field.rbn_right_red = (a_type *) (((intptr_t)		\
+      (a_node)->a_field.rbn_right_red) & ((ssize_t)-2));		\
+} while (0)
+#else
+/* Right accessors. */
+#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 {		\
+    (a_node)->a_field.rbn_right = a_right;				\
+} while (0)
+
+/* Color accessors. */
+#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 {		\
+    (a_node)->a_field.rbn_red = (a_red);				\
+} while (0)
+#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 {			\
+    (a_node)->a_field.rbn_red = false;					\
+} 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);			\
+} while (0)
+
+/* Internal utility macros. */
+#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) {				\
+	for (;								\
+	  rbtn_left_get(a_type, a_field, (r_node)) != &(a_rbt)->rbt_nil;\
+	  (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 {		\
+    (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))) {							\
+	}								\
+    }									\
+} while (0)
+
+#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 {		\
+    (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)));			\
+    rbtn_right_set(a_type, a_field, (r_node), (a_node));		\
+} while (0)
+
+/*
+ * The rb_proto() macro generates function prototypes that correspond to the
+ * functions generated by an equivalently parameterized call to rb_gen().
+ */
+
+#define	rb_proto(a_attr, a_prefix, a_rbt_type, a_type)			\
+a_attr void								\
+a_prefix##new(a_rbt_type *rbtree);					\
+a_attr a_type *								\
+a_prefix##first(a_rbt_type *rbtree);					\
+a_attr a_type *								\
+a_prefix##last(a_rbt_type *rbtree);					\
+a_attr a_type *								\
+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_attr a_type *								\
+a_prefix##nsearch(a_rbt_type *rbtree, a_type *key);			\
+a_attr a_type *								\
+a_prefix##psearch(a_rbt_type *rbtree, a_type *key);			\
+a_attr void								\
+a_prefix##insert(a_rbt_type *rbtree, a_type *node);			\
+a_attr void								\
+a_prefix##remove(a_rbt_type *rbtree, a_type *node);			\
+a_attr a_type *								\
+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);
+
+/*
+ * The rb_gen() macro generates a type-specific red-black tree implementation,
+ * based on the above cpp macros.
+ *
+ * Arguments:
+ *
+ *   a_attr    : Function attribute for generated functions (ex: static).
+ *   a_prefix  : Prefix for generated functions (ex: ex_).
+ *   a_rb_type : Type for red-black tree data structure (ex: ex_t).
+ *   a_type    : Type for red-black tree node data structure (ex: ex_node_t).
+ *   a_field   : Name of red-black tree node linkage (ex: ex_link).
+ *   a_cmp     : Node comparison function name, with the following prototype:
+ *                 int (a_cmp *)(a_type *a_node, a_type *a_other);
+ *                                       ^^^^^^
+ *                                    or a_key
+ *               Interpretation of comparision function return values:
+ *                 -1 : a_node <  a_other
+ *                  0 : a_node == a_other
+ *                  1 : a_node >  a_other
+ *               In all cases, the a_node or a_key macro argument is the first
+ *               argument to the comparison function, which makes it possible
+ *               to write comparison functions that treat the first argument
+ *               specially.
+ *
+ * Assuming the following setup:
+ *
+ *   typedef struct ex_node_s ex_node_t;
+ *   struct ex_node_s {
+ *       rb_node(ex_node_t) ex_link;
+ *   };
+ *   typedef rb_tree(ex_node_t) ex_t;
+ *   rb_gen(static, ex_, ex_t, ex_node_t, ex_link, ex_cmp)
+ *
+ * The following API is generated:
+ *
+ *   static void
+ *   ex_new(ex_t *extree);
+ *       Description: Initialize a red-black tree structure.
+ *       Args:
+ *         extree: 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.
+ *       Args:
+ *         extree: Pointer to an initialized red-black tree object.
+ *       Ret: First/last node in extree, or NULL if extree is empty.
+ *
+ *   static ex_node_t *
+ *   ex_next(ex_t *extree, ex_node_t *node);
+ *   static ex_node_t *
+ *   ex_prev(ex_t *extree, 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
+ *            last/first.
+ *
+ *   static ex_node_t *
+ *   ex_search(ex_t *extree, 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.
+ *
+ *   static ex_node_t *
+ *   ex_nsearch(ex_t *extree, ex_node_t *key);
+ *   static ex_node_t *
+ *   ex_psearch(ex_t *extree, 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.
+ *       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).
+ *
+ *   static void
+ *   ex_insert(ex_t *extree, ex_node_t *node);
+ *       Description: Insert node into extree.
+ *       Args:
+ *         extree: Pointer to an initialized red-black tree object.
+ *         node  : Node to be inserted into extree.
+ *
+ *   static void
+ *   ex_remove(ex_t *extree, ex_node_t *node);
+ *       Description: Remove node from extree.
+ *       Args:
+ *         extree: Pointer to an initialized red-black tree object.
+ *         node  : Node in extree to be removed.
+ *
+ *   static ex_node_t *
+ *   ex_iter(ex_t *extree, 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_node_t *, void *), void *arg);
+ *       Description: Iterate forward/backward over extree, starting at node.
+ *                    If extree 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().
+ *       Ret: NULL if iteration completed, or the non-NULL callback return value
+ *            that caused termination of the iteration.
+ */
+#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 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);							\
+}									\
+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);							\
+}									\
+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) {	\
+	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;						\
+	while (true) {							\
+	    int cmp = (a_cmp)(node, tnode);				\
+	    if (cmp < 0) {						\
+		ret = tnode;						\
+		tnode = rbtn_left_get(a_type, a_field, tnode);		\
+	    } else if (cmp > 0) {					\
+		tnode = rbtn_right_get(a_type, a_field, tnode);		\
+	    } else {							\
+		break;							\
+	    }								\
+	    assert(tnode != &rbtree->rbt_nil);				\
+	}								\
+    }									\
+    if (ret == &rbtree->rbt_nil) {					\
+	ret = (NULL);							\
+    }									\
+    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) {	\
+	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;						\
+	while (true) {							\
+	    int cmp = (a_cmp)(node, tnode);				\
+	    if (cmp < 0) {						\
+		tnode = rbtn_left_get(a_type, a_field, tnode);		\
+	    } else if (cmp > 0) {					\
+		ret = tnode;						\
+		tnode = rbtn_right_get(a_type, a_field, tnode);		\
+	    } else {							\
+		break;							\
+	    }								\
+	    assert(tnode != &rbtree->rbt_nil);				\
+	}								\
+    }									\
+    if (ret == &rbtree->rbt_nil) {					\
+	ret = (NULL);							\
+    }									\
+    return (ret);							\
+}									\
+a_attr a_type *								\
+a_prefix##search(a_rbt_type *rbtree, a_type *key) {			\
+    a_type *ret;							\
+    int cmp;								\
+    ret = rbtree->rbt_root;						\
+    while (ret != &rbtree->rbt_nil					\
+      && (cmp = (a_cmp)(key, ret)) != 0) {				\
+	if (cmp < 0) {							\
+	    ret = rbtn_left_get(a_type, a_field, ret);			\
+	} else {							\
+	    ret = rbtn_right_get(a_type, a_field, ret);			\
+	}								\
+    }									\
+    if (ret == &rbtree->rbt_nil) {					\
+	ret = (NULL);							\
+    }									\
+    return (ret);							\
+}									\
+a_attr a_type *								\
+a_prefix##nsearch(a_rbt_type *rbtree, a_type *key) {			\
+    a_type *ret;							\
+    a_type *tnode = rbtree->rbt_root;					\
+    ret = &rbtree->rbt_nil;						\
+    while (tnode != &rbtree->rbt_nil) {					\
+	int cmp = (a_cmp)(key, tnode);					\
+	if (cmp < 0) {							\
+	    ret = tnode;						\
+	    tnode = rbtn_left_get(a_type, a_field, tnode);		\
+	} else if (cmp > 0) {						\
+	    tnode = rbtn_right_get(a_type, a_field, tnode);		\
+	} else {							\
+	    ret = tnode;						\
+	    break;							\
+	}								\
+    }									\
+    if (ret == &rbtree->rbt_nil) {					\
+	ret = (NULL);							\
+    }									\
+    return (ret);							\
+}									\
+a_attr a_type *								\
+a_prefix##psearch(a_rbt_type *rbtree, a_type *key) {			\
+    a_type *ret;							\
+    a_type *tnode = rbtree->rbt_root;					\
+    ret = &rbtree->rbt_nil;						\
+    while (tnode != &rbtree->rbt_nil) {					\
+	int cmp = (a_cmp)(key, tnode);					\
+	if (cmp < 0) {							\
+	    tnode = rbtn_left_get(a_type, a_field, tnode);		\
+	} else if (cmp > 0) {						\
+	    ret = tnode;						\
+	    tnode = rbtn_right_get(a_type, a_field, tnode);		\
+	} else {							\
+	    ret = tnode;						\
+	    break;							\
+	}								\
+    }									\
+    if (ret == &rbtree->rbt_nil) {					\
+	ret = (NULL);							\
+    }									\
+    return (ret);							\
+}									\
+a_attr void								\
+a_prefix##insert(a_rbt_type *rbtree, a_type *node) {			\
+    struct {								\
+	a_type *node;							\
+	int cmp;							\
+    } path[sizeof(void *) << 4], *pathp;				\
+    rbt_node_new(a_type, a_field, rbtree, node);			\
+    /* Wind. */								\
+    path->node = rbtree->rbt_root;					\
+    for (pathp = path; pathp->node != &rbtree->rbt_nil; pathp++) {	\
+	int cmp = pathp->cmp = a_cmp(node, pathp->node);		\
+	assert(cmp != 0);						\
+	if (cmp < 0) {							\
+	    pathp[1].node = rbtn_left_get(a_type, a_field,		\
+	      pathp->node);						\
+	} else {							\
+	    pathp[1].node = rbtn_right_get(a_type, a_field,		\
+	      pathp->node);						\
+	}								\
+    }									\
+    pathp->node = node;							\
+    /* Unwind. */							\
+    for (pathp--; (uintptr_t)pathp >= (uintptr_t)path; pathp--) {	\
+	a_type *cnode = pathp->node;					\
+	if (pathp->cmp < 0) {						\
+	    a_type *left = pathp[1].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)) {		\
+		    /* Fix up 4-node. */				\
+		    a_type *tnode;					\
+		    rbtn_black_set(a_type, a_field, leftleft);		\
+		    rbtn_rotate_right(a_type, a_field, cnode, tnode);	\
+		    cnode = tnode;					\
+		}							\
+	    } else {							\
+		return;							\
+	    }								\
+	} else {							\
+	    a_type *right = pathp[1].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)) {		\
+		    /* Split 4-node. */					\
+		    rbtn_black_set(a_type, a_field, left);		\
+		    rbtn_black_set(a_type, a_field, right);		\
+		    rbtn_red_set(a_type, a_field, cnode);		\
+		} else {						\
+		    /* Lean left. */					\
+		    a_type *tnode;					\
+		    bool tred = rbtn_red_get(a_type, a_field, cnode);	\
+		    rbtn_rotate_left(a_type, a_field, cnode, tnode);	\
+		    rbtn_color_set(a_type, a_field, tnode, tred);	\
+		    rbtn_red_set(a_type, a_field, cnode);		\
+		    cnode = tnode;					\
+		}							\
+	    } else {							\
+		return;							\
+	    }								\
+	}								\
+	pathp->node = cnode;						\
+    }									\
+    /* Set root, and make it black. */					\
+    rbtree->rbt_root = path->node;					\
+    rbtn_black_set(a_type, a_field, rbtree->rbt_root);			\
+}									\
+a_attr void								\
+a_prefix##remove(a_rbt_type *rbtree, a_type *node) {			\
+    struct {								\
+	a_type *node;							\
+	int cmp;							\
+    } *pathp, *nodep, path[sizeof(void *) << 4];			\
+    /* Wind. */								\
+    nodep = NULL; /* Silence compiler warning. */			\
+    path->node = rbtree->rbt_root;					\
+    for (pathp = path; pathp->node != &rbtree->rbt_nil; pathp++) {	\
+	int cmp = pathp->cmp = a_cmp(node, pathp->node);		\
+	if (cmp < 0) {							\
+	    pathp[1].node = rbtn_left_get(a_type, a_field,		\
+	      pathp->node);						\
+	} else {							\
+	    pathp[1].node = rbtn_right_get(a_type, a_field,		\
+	      pathp->node);						\
+	    if (cmp == 0) {						\
+	        /* Find node's successor, in preparation for swap. */	\
+		pathp->cmp = 1;						\
+		nodep = pathp;						\
+		for (pathp++; pathp->node != &rbtree->rbt_nil;		\
+		  pathp++) {						\
+		    pathp->cmp = -1;					\
+		    pathp[1].node = rbtn_left_get(a_type, a_field,	\
+		      pathp->node);					\
+		}							\
+		break;							\
+	    }								\
+	}								\
+    }									\
+    assert(nodep->node == node);					\
+    pathp--;								\
+    if (pathp->node != node) {						\
+	/* Swap node with its successor. */				\
+	bool tred = rbtn_red_get(a_type, a_field, pathp->node);		\
+	rbtn_color_set(a_type, a_field, pathp->node,			\
+	  rbtn_red_get(a_type, a_field, node));				\
+	rbtn_left_set(a_type, a_field, pathp->node,			\
+	  rbtn_left_get(a_type, a_field, node));			\
+	/* If node's successor is its right child, the following code */\
+	/* will do the wrong thing for the right child pointer.       */\
+	/* However, it doesn't matter, because the pointer will be    */\
+	/* properly set when the successor is pruned.                 */\
+	rbtn_right_set(a_type, a_field, pathp->node,			\
+	  rbtn_right_get(a_type, a_field, node));			\
+	rbtn_color_set(a_type, a_field, node, tred);			\
+	/* The pruned leaf node's child pointers are never accessed   */\
+	/* again, so don't bother setting them to nil.                */\
+	nodep->node = pathp->node;					\
+	pathp->node = node;						\
+	if (nodep == path) {						\
+	    rbtree->rbt_root = nodep->node;				\
+	} else {							\
+	    if (nodep[-1].cmp < 0) {					\
+		rbtn_left_set(a_type, a_field, nodep[-1].node,		\
+		  nodep->node);						\
+	    } else {							\
+		rbtn_right_set(a_type, a_field, nodep[-1].node,		\
+		  nodep->node);						\
+	    }								\
+	}								\
+    } else {								\
+	a_type *left = rbtn_left_get(a_type, a_field, node);		\
+	if (left != &rbtree->rbt_nil) {					\
+	    /* 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, left));		\
+	    rbtn_black_set(a_type, a_field, left);			\
+	    if (pathp == path) {					\
+		rbtree->rbt_root = left;				\
+	    } else {							\
+		if (pathp[-1].cmp < 0) {				\
+		    rbtn_left_set(a_type, a_field, pathp[-1].node,	\
+		      left);						\
+		} else {						\
+		    rbtn_right_set(a_type, a_field, pathp[-1].node,	\
+		      left);						\
+		}							\
+	    }								\
+	    return;							\
+	} else if (pathp == path) {					\
+	    /* The tree only contained one node. */			\
+	    rbtree->rbt_root = &rbtree->rbt_nil;			\
+	    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);						\
+	return;								\
+    }									\
+    /* The node to be pruned is black, so unwind until balance is     */\
+    /* restored.                                                      */\
+    pathp->node = &rbtree->rbt_nil;					\
+    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)) {		\
+		    /* In the following diagrams, ||, //, and \\      */\
+		    /* indicate the path to the removed node.         */\
+		    /*                                                */\
+		    /*      ||                                        */\
+		    /*    pathp(r)                                    */\
+		    /*  //        \                                   */\
+		    /* (b)        (b)                                 */\
+		    /*           /                                    */\
+		    /*          (r)                                   */\
+		    /*                                                */\
+		    rbtn_black_set(a_type, a_field, pathp->node);	\
+		    rbtn_rotate_right(a_type, a_field, right, tnode);	\
+		    rbtn_right_set(a_type, a_field, pathp->node, tnode);\
+		    rbtn_rotate_left(a_type, a_field, pathp->node,	\
+		      tnode);						\
+		} else {						\
+		    /*      ||                                        */\
+		    /*    pathp(r)                                    */\
+		    /*  //        \                                   */\
+		    /* (b)        (b)                                 */\
+		    /*           /                                    */\
+		    /*          (b)                                   */\
+		    /*                                                */\
+		    rbtn_rotate_left(a_type, a_field, pathp->node,	\
+		      tnode);						\
+		}							\
+		/* Balance restored, but rotation modified subtree    */\
+		/* root.                                              */\
+		assert((uintptr_t)pathp > (uintptr_t)path);		\
+		if (pathp[-1].cmp < 0) {				\
+		    rbtn_left_set(a_type, a_field, pathp[-1].node,	\
+		      tnode);						\
+		} else {						\
+		    rbtn_right_set(a_type, a_field, pathp[-1].node,	\
+		      tnode);						\
+		}							\
+		return;							\
+	    } else {							\
+		a_type *right = rbtn_right_get(a_type, a_field,		\
+		  pathp->node);						\
+		a_type *rightleft = rbtn_left_get(a_type, a_field,	\
+		  right);						\
+		if (rbtn_red_get(a_type, a_field, rightleft)) {		\
+		    /*      ||                                        */\
+		    /*    pathp(b)                                    */\
+		    /*  //        \                                   */\
+		    /* (b)        (b)                                 */\
+		    /*           /                                    */\
+		    /*          (r)                                   */\
+		    a_type *tnode;					\
+		    rbtn_black_set(a_type, a_field, rightleft);		\
+		    rbtn_rotate_right(a_type, a_field, right, tnode);	\
+		    rbtn_right_set(a_type, a_field, pathp->node, tnode);\
+		    rbtn_rotate_left(a_type, a_field, pathp->node,	\
+		      tnode);						\
+		    /* Balance restored, but rotation modified        */\
+		    /* subree root, which may actually be the tree    */\
+		    /* root.                                          */\
+		    if (pathp == path) {				\
+			/* Set root. */					\
+			rbtree->rbt_root = tnode;			\
+		    } else {						\
+			if (pathp[-1].cmp < 0) {			\
+			    rbtn_left_set(a_type, a_field,		\
+			      pathp[-1].node, tnode);			\
+			} else {					\
+			    rbtn_right_set(a_type, a_field,		\
+			      pathp[-1].node, tnode);			\
+			}						\
+		    }							\
+		    return;						\
+		} else {						\
+		    /*      ||                                        */\
+		    /*    pathp(b)                                    */\
+		    /*  //        \                                   */\
+		    /* (b)        (b)                                 */\
+		    /*           /                                    */\
+		    /*          (b)                                   */\
+		    a_type *tnode;					\
+		    rbtn_red_set(a_type, a_field, pathp->node);		\
+		    rbtn_rotate_left(a_type, a_field, pathp->node,	\
+		      tnode);						\
+		    pathp->node = tnode;				\
+		}							\
+	    }								\
+	} else {							\
+	    a_type *left;						\
+	    rbtn_right_set(a_type, a_field, pathp->node,		\
+	      pathp[1].node);						\
+	    left = rbtn_left_get(a_type, a_field, pathp->node);		\
+	    if (rbtn_red_get(a_type, a_field, left)) {			\
+		a_type *tnode;						\
+		a_type *leftright = rbtn_right_get(a_type, a_field,	\
+		  left);						\
+		a_type *leftrightleft = rbtn_left_get(a_type, a_field,	\
+		  leftright);						\
+		if (rbtn_red_get(a_type, a_field, leftrightleft)) {	\
+		    /*      ||                                        */\
+		    /*    pathp(b)                                    */\
+		    /*   /        \\                                  */\
+		    /* (r)        (b)                                 */\
+		    /*   \                                            */\
+		    /*   (b)                                          */\
+		    /*   /                                            */\
+		    /* (r)                                            */\
+		    a_type *unode;					\
+		    rbtn_black_set(a_type, a_field, leftrightleft);	\
+		    rbtn_rotate_right(a_type, a_field, pathp->node,	\
+		      unode);						\
+		    rbtn_rotate_right(a_type, a_field, pathp->node,	\
+		      tnode);						\
+		    rbtn_right_set(a_type, a_field, unode, tnode);	\
+		    rbtn_rotate_left(a_type, a_field, unode, tnode);	\
+		} else {						\
+		    /*      ||                                        */\
+		    /*    pathp(b)                                    */\
+		    /*   /        \\                                  */\
+		    /* (r)        (b)                                 */\
+		    /*   \                                            */\
+		    /*   (b)                                          */\
+		    /*   /                                            */\
+		    /* (b)                                            */\
+		    assert(leftright != &rbtree->rbt_nil);		\
+		    rbtn_red_set(a_type, a_field, leftright);		\
+		    rbtn_rotate_right(a_type, a_field, pathp->node,	\
+		      tnode);						\
+		    rbtn_black_set(a_type, a_field, tnode);		\
+		}							\
+		/* Balance restored, but rotation modified subtree    */\
+		/* root, which may actually be the tree root.         */\
+		if (pathp == path) {					\
+		    /* Set root. */					\
+		    rbtree->rbt_root = tnode;				\
+		} else {						\
+		    if (pathp[-1].cmp < 0) {				\
+			rbtn_left_set(a_type, a_field, pathp[-1].node,	\
+			  tnode);					\
+		    } else {						\
+			rbtn_right_set(a_type, a_field, pathp[-1].node,	\
+			  tnode);					\
+		    }							\
+		}							\
+		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)) {		\
+		    /*        ||                                      */\
+		    /*      pathp(r)                                  */\
+		    /*     /        \\                                */\
+		    /*   (b)        (b)                               */\
+		    /*   /                                            */\
+		    /* (r)                                            */\
+		    a_type *tnode;					\
+		    rbtn_black_set(a_type, a_field, pathp->node);	\
+		    rbtn_red_set(a_type, a_field, left);		\
+		    rbtn_black_set(a_type, a_field, leftleft);		\
+		    rbtn_rotate_right(a_type, a_field, pathp->node,	\
+		      tnode);						\
+		    /* Balance restored, but rotation modified        */\
+		    /* subtree root.                                  */\
+		    assert((uintptr_t)pathp > (uintptr_t)path);		\
+		    if (pathp[-1].cmp < 0) {				\
+			rbtn_left_set(a_type, a_field, pathp[-1].node,	\
+			  tnode);					\
+		    } else {						\
+			rbtn_right_set(a_type, a_field, pathp[-1].node,	\
+			  tnode);					\
+		    }							\
+		    return;						\
+		} else {						\
+		    /*        ||                                      */\
+		    /*      pathp(r)                                  */\
+		    /*     /        \\                                */\
+		    /*   (b)        (b)                               */\
+		    /*   /                                            */\
+		    /* (b)                                            */\
+		    rbtn_red_set(a_type, a_field, left);		\
+		    rbtn_black_set(a_type, a_field, pathp->node);	\
+		    /* Balance restored. */				\
+		    return;						\
+		}							\
+	    } else {							\
+		a_type *leftleft = rbtn_left_get(a_type, a_field, left);\
+		if (rbtn_red_get(a_type, a_field, leftleft)) {		\
+		    /*               ||                               */\
+		    /*             pathp(b)                           */\
+		    /*            /        \\                         */\
+		    /*          (b)        (b)                        */\
+		    /*          /                                     */\
+		    /*        (r)                                     */\
+		    a_type *tnode;					\
+		    rbtn_black_set(a_type, a_field, leftleft);		\
+		    rbtn_rotate_right(a_type, a_field, pathp->node,	\
+		      tnode);						\
+		    /* Balance restored, but rotation modified        */\
+		    /* subtree root, which may actually be the tree   */\
+		    /* root.                                          */\
+		    if (pathp == path) {				\
+			/* Set root. */					\
+			rbtree->rbt_root = tnode;			\
+		    } else {						\
+			if (pathp[-1].cmp < 0) {			\
+			    rbtn_left_set(a_type, a_field,		\
+			      pathp[-1].node, tnode);			\
+			} else {					\
+			    rbtn_right_set(a_type, a_field,		\
+			      pathp[-1].node, tnode);			\
+			}						\
+		    }							\
+		    return;						\
+		} else {						\
+		    /*               ||                               */\
+		    /*             pathp(b)                           */\
+		    /*            /        \\                         */\
+		    /*          (b)        (b)                        */\
+		    /*          /                                     */\
+		    /*        (b)                                     */\
+		    rbtn_red_set(a_type, a_field, left);		\
+		}							\
+	    }								\
+	}								\
+    }									\
+    /* Set root. */							\
+    rbtree->rbt_root = path->node;					\
+    assert(rbtn_red_get(a_type, a_field, rbtree->rbt_root) == false);	\
+}									\
+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);					\
+    } 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);						\
+	}								\
+	return (a_prefix##iter_recurse(rbtree, rbtn_right_get(a_type,	\
+	  a_field, node), cb, arg));					\
+    }									\
+}									\
+a_attr a_type *								\
+a_prefix##iter_start(a_rbt_type *rbtree, a_type *start, a_type *node,	\
+  a_type *(*cb)(a_rbt_type *, a_type *, void *), void *arg) {		\
+    int cmp = a_cmp(start, 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);						\
+	}								\
+	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));		\
+    } else {								\
+	a_type *ret;							\
+	if ((ret = cb(rbtree, node, arg)) != NULL) {			\
+	    return (ret);						\
+	}								\
+	return (a_prefix##iter_recurse(rbtree, rbtn_right_get(a_type,	\
+	  a_field, node), cb, arg));					\
+    }									\
+}									\
+a_attr a_type *								\
+a_prefix##iter(a_rbt_type *rbtree, a_type *start, a_type *(*cb)(	\
+  a_rbt_type *, a_type *, void *), void *arg) {				\
+    a_type *ret;							\
+    if (start != NULL) {						\
+	ret = a_prefix##iter_start(rbtree, start, rbtree->rbt_root,	\
+	  cb, arg);							\
+    } else {								\
+	ret = a_prefix##iter_recurse(rbtree, rbtree->rbt_root, cb, arg);\
+    }									\
+    if (ret == &rbtree->rbt_nil) {					\
+	ret = NULL;							\
+    }									\
+    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);					\
+    } 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);						\
+	}								\
+	return (a_prefix##reverse_iter_recurse(rbtree,			\
+	  rbtn_left_get(a_type, a_field, node), cb, arg));		\
+    }									\
+}									\
+a_attr a_type *								\
+a_prefix##reverse_iter_start(a_rbt_type *rbtree, a_type *start,		\
+  a_type *node, a_type *(*cb)(a_rbt_type *, a_type *, void *),		\
+  void *arg) {								\
+    int cmp = a_cmp(start, node);					\
+    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);						\
+	}								\
+	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));		\
+    } else {								\
+	a_type *ret;							\
+	if ((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));		\
+    }									\
+}									\
+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 *ret;							\
+    if (start != NULL) {						\
+	ret = a_prefix##reverse_iter_start(rbtree, start,		\
+	  rbtree->rbt_root, cb, arg);					\
+    } else {								\
+	ret = a_prefix##reverse_iter_recurse(rbtree, rbtree->rbt_root,	\
+	  cb, arg);							\
+    }									\
+    if (ret == &rbtree->rbt_nil) {					\
+	ret = NULL;							\
+    }									\
+    return (ret);							\
+}
+
+#endif /* RB_H_ */

deps/jemalloc/include/jemalloc/internal/rtree.h

@@ -0,0 +1,161 @@
+/*
+ * This radix tree implementation is tailored to the singular purpose of
+ * tracking which chunks are currently owned by jemalloc.  This functionality
+ * is mandatory for OS X, where jemalloc must be able to respond to object
+ * ownership queries.
+ *
+ *******************************************************************************
+ */
+#ifdef JEMALLOC_H_TYPES
+
+typedef struct rtree_s rtree_t;
+
+/*
+ * Size of each radix tree node (must be a power of 2).  This impacts tree
+ * depth.
+ */
+#if (LG_SIZEOF_PTR == 2)
+#  define RTREE_NODESIZE (1U << 14)
+#else
+#  define RTREE_NODESIZE CACHELINE
+#endif
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+struct rtree_s {
+	malloc_mutex_t	mutex;
+	void		**root;
+	unsigned	height;
+	unsigned	level2bits[1]; /* Dynamically sized. */
+};
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+rtree_t	*rtree_new(unsigned bits);
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#ifndef JEMALLOC_ENABLE_INLINE
+#ifndef JEMALLOC_DEBUG
+void	*rtree_get_locked(rtree_t *rtree, uintptr_t key);
+#endif
+void	*rtree_get(rtree_t *rtree, uintptr_t key);
+bool	rtree_set(rtree_t *rtree, uintptr_t key, void *val);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_RTREE_C_))
+#define	RTREE_GET_GENERATE(f)						\
+/* The least significant bits of the key are ignored. */		\
+JEMALLOC_INLINE void *							\
+f(rtree_t *rtree, uintptr_t key)					\
+{									\
+	void *ret;							\
+	uintptr_t subkey;						\
+	unsigned i, lshift, height, bits;				\
+	void **node, **child;						\
+									\
+	RTREE_LOCK(&rtree->mutex);					\
+	for (i = lshift = 0, height = rtree->height, node = rtree->root;\
+	    i < height - 1;						\
+	    i++, lshift += bits, node = child) {			\
+		bits = rtree->level2bits[i];				\
+		subkey = (key << lshift) >> ((ZU(1) << (LG_SIZEOF_PTR + \
+		    3)) - bits);					\
+		child = (void**)node[subkey];				\
+		if (child == NULL) {					\
+			RTREE_UNLOCK(&rtree->mutex);			\
+			return (NULL);					\
+		}							\
+	}								\
+									\
+	/*								\
+	 * node is a leaf, so it contains values rather than node	\
+	 * pointers.							\
+	 */								\
+	bits = rtree->level2bits[i];					\
+	subkey = (key << lshift) >> ((ZU(1) << (LG_SIZEOF_PTR+3)) -	\
+	    bits);							\
+	ret = node[subkey];						\
+	RTREE_UNLOCK(&rtree->mutex);					\
+									\
+	RTREE_GET_VALIDATE						\
+	return (ret);							\
+}
+
+#ifdef JEMALLOC_DEBUG
+#  define RTREE_LOCK(l)		malloc_mutex_lock(l)
+#  define RTREE_UNLOCK(l)	malloc_mutex_unlock(l)
+#  define RTREE_GET_VALIDATE
+RTREE_GET_GENERATE(rtree_get_locked)
+#  undef RTREE_LOCK
+#  undef RTREE_UNLOCK
+#  undef RTREE_GET_VALIDATE
+#endif
+
+#define	RTREE_LOCK(l)
+#define	RTREE_UNLOCK(l)
+#ifdef JEMALLOC_DEBUG
+   /*
+    * Suppose that it were possible for a jemalloc-allocated chunk to be
+    * munmap()ped, followed by a different allocator in another thread re-using
+    * overlapping virtual memory, all without invalidating the cached rtree
+    * value.  The result would be a false positive (the rtree would claim that
+    * jemalloc owns memory that it had actually discarded).  This scenario
+    * seems impossible, but the following assertion is a prudent sanity check.
+    */
+#  define RTREE_GET_VALIDATE						\
+	assert(rtree_get_locked(rtree, key) == ret);
+#else
+#  define RTREE_GET_VALIDATE
+#endif
+RTREE_GET_GENERATE(rtree_get)
+#undef RTREE_LOCK
+#undef RTREE_UNLOCK
+#undef RTREE_GET_VALIDATE
+
+JEMALLOC_INLINE bool
+rtree_set(rtree_t *rtree, uintptr_t key, void *val)
+{
+	uintptr_t subkey;
+	unsigned i, lshift, height, bits;
+	void **node, **child;
+
+	malloc_mutex_lock(&rtree->mutex);
+	for (i = lshift = 0, height = rtree->height, node = rtree->root;
+	    i < height - 1;
+	    i++, lshift += bits, node = child) {
+		bits = rtree->level2bits[i];
+		subkey = (key << lshift) >> ((ZU(1) << (LG_SIZEOF_PTR+3)) -
+		    bits);
+		child = (void**)node[subkey];
+		if (child == NULL) {
+			child = (void**)base_alloc(sizeof(void *) <<
+			    rtree->level2bits[i+1]);
+			if (child == NULL) {
+				malloc_mutex_unlock(&rtree->mutex);
+				return (true);
+			}
+			memset(child, 0, sizeof(void *) <<
+			    rtree->level2bits[i+1]);
+			node[subkey] = child;
+		}
+	}
+
+	/* node is a leaf, so it contains values rather than node pointers. */
+	bits = rtree->level2bits[i];
+	subkey = (key << lshift) >> ((ZU(1) << (LG_SIZEOF_PTR+3)) - bits);
+	node[subkey] = val;
+	malloc_mutex_unlock(&rtree->mutex);
+
+	return (false);
+}
+#endif
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/

deps/jemalloc/include/jemalloc/internal/stats.h

@@ -0,0 +1,207 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+#define	UMAX2S_BUFSIZE	65
+
+#ifdef JEMALLOC_STATS
+typedef struct tcache_bin_stats_s tcache_bin_stats_t;
+typedef struct malloc_bin_stats_s malloc_bin_stats_t;
+typedef struct malloc_large_stats_s malloc_large_stats_t;
+typedef struct arena_stats_s arena_stats_t;
+#endif
+#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
+typedef struct chunk_stats_s chunk_stats_t;
+#endif
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+#ifdef JEMALLOC_STATS
+
+#ifdef JEMALLOC_TCACHE
+struct tcache_bin_stats_s {
+	/*
+	 * Number of allocation requests that corresponded to the size of this
+	 * bin.
+	 */
+	uint64_t	nrequests;
+};
+#endif
+
+struct malloc_bin_stats_s {
+	/*
+	 * Current number of bytes allocated, including objects currently
+	 * cached by tcache.
+	 */
+	size_t		allocated;
+
+	/*
+	 * 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;
+
+#ifdef JEMALLOC_TCACHE
+	/* Number of tcache fills from this bin. */
+	uint64_t	nfills;
+
+	/* Number of tcache flushes to this bin. */
+	uint64_t	nflushes;
+#endif
+
+	/* Total number of runs created for this bin's size class. */
+	uint64_t	nruns;
+
+	/*
+	 * Total number of runs reused by extracting them from the runs tree for
+	 * this bin's size class.
+	 */
+	uint64_t	reruns;
+
+	/* High-water mark for this bin. */
+	size_t		highruns;
+
+	/* Current number of runs in this bin. */
+	size_t		curruns;
+};
+
+struct malloc_large_stats_s {
+	/*
+	 * Total number of allocation/deallocation requests served directly by
+	 * the arena.  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 this size class.
+	 * This includes requests served by tcache, though tcache only
+	 * periodically merges into this counter.
+	 */
+	uint64_t	nrequests;
+
+	/* High-water mark for this size class. */
+	size_t		highruns;
+
+	/* Current number of runs of this size class. */
+	size_t		curruns;
+};
+
+struct arena_stats_s {
+	/* Number of bytes currently mapped. */
+	size_t		mapped;
+
+	/*
+	 * Total number of purge sweeps, total number of madvise calls made,
+	 * and total pages purged in order to keep dirty unused memory under
+	 * control.
+	 */
+	uint64_t	npurge;
+	uint64_t	nmadvise;
+	uint64_t	purged;
+
+	/* Per-size-category statistics. */
+	size_t		allocated_large;
+	uint64_t	nmalloc_large;
+	uint64_t	ndalloc_large;
+	uint64_t	nrequests_large;
+
+	/*
+	 * One element for each possible size class, including sizes that
+	 * overlap with bin size classes.  This is necessary because ipalloc()
+	 * sometimes has to use such large objects in order to assure proper
+	 * alignment.
+	 */
+	malloc_large_stats_t	*lstats;
+};
+#endif /* JEMALLOC_STATS */
+
+#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
+struct chunk_stats_s {
+#  ifdef JEMALLOC_STATS
+	/* Number of chunks that were allocated. */
+	uint64_t	nchunks;
+#  endif
+
+	/* High-water mark for number of chunks allocated. */
+	size_t		highchunks;
+
+	/*
+	 * Current number of chunks allocated.  This value isn't maintained for
+	 * any other purpose, so keep track of it in order to be able to set
+	 * highchunks.
+	 */
+	size_t		curchunks;
+};
+#endif /* JEMALLOC_STATS */
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+extern bool	opt_stats_print;
+
+#ifdef JEMALLOC_STATS
+extern size_t	stats_cactive;
+#endif
+
+char	*u2s(uint64_t x, unsigned base, char *s);
+#ifdef JEMALLOC_STATS
+void malloc_cprintf(void (*write)(void *, const char *), void *cbopaque,
+    const char *format, ...) JEMALLOC_ATTR(format(printf, 3, 4));
+void	malloc_printf(const char *format, ...)
+    JEMALLOC_ATTR(format(printf, 1, 2));
+#endif
+void	stats_print(void (*write)(void *, const char *), void *cbopaque,
+    const char *opts);
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+#ifdef JEMALLOC_STATS
+
+#ifndef JEMALLOC_ENABLE_INLINE
+size_t	stats_cactive_get(void);
+void	stats_cactive_add(size_t size);
+void	stats_cactive_sub(size_t size);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_STATS_C_))
+JEMALLOC_INLINE size_t
+stats_cactive_get(void)
+{
+
+	return (atomic_read_z(&stats_cactive));
+}
+
+JEMALLOC_INLINE void
+stats_cactive_add(size_t size)
+{
+
+	atomic_add_z(&stats_cactive, size);
+}
+
+JEMALLOC_INLINE void
+stats_cactive_sub(size_t size)
+{
+
+	atomic_sub_z(&stats_cactive, size);
+}
+#endif
+
+#endif /* JEMALLOC_STATS */
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/

deps/jemalloc/include/jemalloc/internal/tcache.h

@@ -0,0 +1,431 @@
+#ifdef JEMALLOC_TCACHE
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+typedef struct tcache_bin_info_s tcache_bin_info_t;
+typedef struct tcache_bin_s tcache_bin_t;
+typedef struct tcache_s tcache_t;
+
+/*
+ * Absolute maximum number of cache slots for each small bin in the thread
+ * cache.  This is an additional constraint beyond that imposed as: twice the
+ * number of regions per run for this size class.
+ *
+ * This constant must be an even number.
+ */
+#define	TCACHE_NSLOTS_SMALL_MAX		200
+
+/* Number of cache slots for large size classes. */
+#define	TCACHE_NSLOTS_LARGE		20
+
+/* (1U << opt_lg_tcache_max) is used to compute tcache_maxclass. */
+#define	LG_TCACHE_MAXCLASS_DEFAULT	15
+
+/*
+ * (1U << opt_lg_tcache_gc_sweep) is the approximate number of allocation
+ * events between full GC sweeps (-1: disabled).  Integer rounding may cause
+ * the actual number to be slightly higher, since GC is performed
+ * incrementally.
+ */
+#define	LG_TCACHE_GC_SWEEP_DEFAULT	13
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+/*
+ * 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 {
+#  ifdef JEMALLOC_STATS
+	tcache_bin_stats_t tstats;
+#  endif
+	int		low_water;	/* Min # cached since last GC. */
+	unsigned	lg_fill_div;	/* Fill (ncached_max >> lg_fill_div). */
+	unsigned	ncached;	/* # of cached objects. */
+	void		**avail;	/* Stack of available objects. */
+};
+
+struct tcache_s {
+#  ifdef JEMALLOC_STATS
+	ql_elm(tcache_t) link;		/* Used for aggregating stats. */
+#  endif
+#  ifdef JEMALLOC_PROF
+	uint64_t	prof_accumbytes;/* Cleared after arena_prof_accum() */
+#  endif
+	arena_t		*arena;		/* This thread's arena. */
+	unsigned	ev_cnt;		/* Event count since incremental GC. */
+	unsigned	next_gc_bin;	/* Next bin to GC. */
+	tcache_bin_t	tbins[1];	/* Dynamically sized. */
+	/*
+	 * The pointer stacks associated with tbins follow as a contiguous
+	 * array.  During tcache initialization, the avail pointer in each
+	 * element of tbins is initialized to point to the proper offset within
+	 * this array.
+	 */
+};
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+extern bool	opt_tcache;
+extern ssize_t	opt_lg_tcache_max;
+extern ssize_t	opt_lg_tcache_gc_sweep;
+
+extern tcache_bin_info_t	*tcache_bin_info;
+
+/* Map of thread-specific caches. */
+#ifndef NO_TLS
+extern __thread tcache_t	*tcache_tls
+    JEMALLOC_ATTR(tls_model("initial-exec"));
+#  define TCACHE_GET()	tcache_tls
+#  define TCACHE_SET(v)	do {						\
+	tcache_tls = (tcache_t *)(v);					\
+	pthread_setspecific(tcache_tsd, (void *)(v));			\
+} while (0)
+#else
+#  define TCACHE_GET()	((tcache_t *)pthread_getspecific(tcache_tsd))
+#  define TCACHE_SET(v)	do {						\
+	pthread_setspecific(tcache_tsd, (void *)(v));			\
+} while (0)
+#endif
+extern pthread_key_t		tcache_tsd;
+
+/*
+ * Number of tcache bins.  There are nbins small-object bins, plus 0 or more
+ * large-object bins.
+ */
+extern size_t			nhbins;
+
+/* Maximum cached size class. */
+extern size_t			tcache_maxclass;
+
+/* Number of tcache allocation/deallocation events between incremental GCs. */
+extern unsigned			tcache_gc_incr;
+
+void	tcache_bin_flush_small(tcache_bin_t *tbin, size_t binind, unsigned rem
+#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
+    , tcache_t *tcache
+#endif
+    );
+void	tcache_bin_flush_large(tcache_bin_t *tbin, size_t binind, unsigned rem
+#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
+    , tcache_t *tcache
+#endif
+    );
+tcache_t *tcache_create(arena_t *arena);
+void	*tcache_alloc_small_hard(tcache_t *tcache, tcache_bin_t *tbin,
+    size_t binind);
+void	tcache_destroy(tcache_t *tcache);
+#ifdef JEMALLOC_STATS
+void	tcache_stats_merge(tcache_t *tcache, arena_t *arena);
+#endif
+bool	tcache_boot(void);
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#ifndef JEMALLOC_ENABLE_INLINE
+void	tcache_event(tcache_t *tcache);
+tcache_t *tcache_get(void);
+void	*tcache_alloc_easy(tcache_bin_t *tbin);
+void	*tcache_alloc_small(tcache_t *tcache, size_t size, bool zero);
+void	*tcache_alloc_large(tcache_t *tcache, size_t size, bool zero);
+void	tcache_dalloc_small(tcache_t *tcache, void *ptr);
+void	tcache_dalloc_large(tcache_t *tcache, void *ptr, size_t size);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_TCACHE_C_))
+JEMALLOC_INLINE tcache_t *
+tcache_get(void)
+{
+	tcache_t *tcache;
+
+	if ((isthreaded & opt_tcache) == false)
+		return (NULL);
+
+	tcache = TCACHE_GET();
+	if ((uintptr_t)tcache <= (uintptr_t)2) {
+		if (tcache == NULL) {
+			tcache = tcache_create(choose_arena());
+			if (tcache == NULL)
+				return (NULL);
+		} else {
+			if (tcache == (void *)(uintptr_t)1) {
+				/*
+				 * Make a note that an allocator function was
+				 * called after the tcache_thread_cleanup() was
+				 * called.
+				 */
+				TCACHE_SET((uintptr_t)2);
+			}
+			return (NULL);
+		}
+	}
+
+	return (tcache);
+}
+
+JEMALLOC_INLINE void
+tcache_event(tcache_t *tcache)
+{
+
+	if (tcache_gc_incr == 0)
+		return;
+
+	tcache->ev_cnt++;
+	assert(tcache->ev_cnt <= tcache_gc_incr);
+	if (tcache->ev_cnt == tcache_gc_incr) {
+		size_t binind = tcache->next_gc_bin;
+		tcache_bin_t *tbin = &tcache->tbins[binind];
+		tcache_bin_info_t *tbin_info = &tcache_bin_info[binind];
+
+		if (tbin->low_water > 0) {
+			/*
+			 * Flush (ceiling) 3/4 of the objects below the low
+			 * water mark.
+			 */
+			if (binind < nbins) {
+				tcache_bin_flush_small(tbin, binind,
+				    tbin->ncached - tbin->low_water +
+				    (tbin->low_water >> 2)
+#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
+				    , tcache
+#endif
+				    );
+			} else {
+				tcache_bin_flush_large(tbin, binind,
+				    tbin->ncached - tbin->low_water +
+				    (tbin->low_water >> 2)
+#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
+				    , tcache
+#endif
+				    );
+			}
+			/*
+			 * Reduce fill count by 2X.  Limit lg_fill_div such that
+			 * the fill count is always at least 1.
+			 */
+			if ((tbin_info->ncached_max >> (tbin->lg_fill_div+1))
+			    >= 1)
+				tbin->lg_fill_div++;
+		} else if (tbin->low_water < 0) {
+			/*
+			 * Increase fill count by 2X.  Make sure lg_fill_div
+			 * stays greater than 0.
+			 */
+			if (tbin->lg_fill_div > 1)
+				tbin->lg_fill_div--;
+		}
+		tbin->low_water = tbin->ncached;
+
+		tcache->next_gc_bin++;
+		if (tcache->next_gc_bin == nhbins)
+			tcache->next_gc_bin = 0;
+		tcache->ev_cnt = 0;
+	}
+}
+
+JEMALLOC_INLINE void *
+tcache_alloc_easy(tcache_bin_t *tbin)
+{
+	void *ret;
+
+	if (tbin->ncached == 0) {
+		tbin->low_water = -1;
+		return (NULL);
+	}
+	tbin->ncached--;
+	if ((int)tbin->ncached < tbin->low_water)
+		tbin->low_water = tbin->ncached;
+	ret = tbin->avail[tbin->ncached];
+	return (ret);
+}
+
+JEMALLOC_INLINE void *
+tcache_alloc_small(tcache_t *tcache, size_t size, bool zero)
+{
+	void *ret;
+	size_t binind;
+	tcache_bin_t *tbin;
+
+	binind = SMALL_SIZE2BIN(size);
+	assert(binind < nbins);
+	tbin = &tcache->tbins[binind];
+	ret = tcache_alloc_easy(tbin);
+	if (ret == NULL) {
+		ret = tcache_alloc_small_hard(tcache, tbin, binind);
+		if (ret == NULL)
+			return (NULL);
+	}
+	assert(arena_salloc(ret) == arena_bin_info[binind].reg_size);
+
+	if (zero == false) {
+#ifdef JEMALLOC_FILL
+		if (opt_junk)
+			memset(ret, 0xa5, size);
+		else if (opt_zero)
+			memset(ret, 0, size);
+#endif
+	} else
+		memset(ret, 0, size);
+
+#ifdef JEMALLOC_STATS
+	tbin->tstats.nrequests++;
+#endif
+#ifdef JEMALLOC_PROF
+	tcache->prof_accumbytes += arena_bin_info[binind].reg_size;
+#endif
+	tcache_event(tcache);
+	return (ret);
+}
+
+JEMALLOC_INLINE void *
+tcache_alloc_large(tcache_t *tcache, size_t size, bool zero)
+{
+	void *ret;
+	size_t binind;
+	tcache_bin_t *tbin;
+
+	size = PAGE_CEILING(size);
+	assert(size <= tcache_maxclass);
+	binind = nbins + (size >> PAGE_SHIFT) - 1;
+	assert(binind < nhbins);
+	tbin = &tcache->tbins[binind];
+	ret = tcache_alloc_easy(tbin);
+	if (ret == NULL) {
+		/*
+		 * Only allocate one large object at a time, because it's quite
+		 * expensive to create one and not use it.
+		 */
+		ret = arena_malloc_large(tcache->arena, size, zero);
+		if (ret == NULL)
+			return (NULL);
+	} else {
+#ifdef JEMALLOC_PROF
+		arena_chunk_t *chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ret);
+		size_t pageind = (((uintptr_t)ret - (uintptr_t)chunk) >>
+		    PAGE_SHIFT);
+		chunk->map[pageind-map_bias].bits &= ~CHUNK_MAP_CLASS_MASK;
+#endif
+		if (zero == false) {
+#ifdef JEMALLOC_FILL
+			if (opt_junk)
+				memset(ret, 0xa5, size);
+			else if (opt_zero)
+				memset(ret, 0, size);
+#endif
+		} else
+			memset(ret, 0, size);
+
+#ifdef JEMALLOC_STATS
+		tbin->tstats.nrequests++;
+#endif
+#ifdef JEMALLOC_PROF
+		tcache->prof_accumbytes += size;
+#endif
+	}
+
+	tcache_event(tcache);
+	return (ret);
+}
+
+JEMALLOC_INLINE void
+tcache_dalloc_small(tcache_t *tcache, void *ptr)
+{
+	arena_t *arena;
+	arena_chunk_t *chunk;
+	arena_run_t *run;
+	arena_bin_t *bin;
+	tcache_bin_t *tbin;
+	tcache_bin_info_t *tbin_info;
+	size_t pageind, binind;
+	arena_chunk_map_t *mapelm;
+
+	assert(arena_salloc(ptr) <= small_maxclass);
+
+	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
+	arena = chunk->arena;
+	pageind = ((uintptr_t)ptr - (uintptr_t)chunk) >> PAGE_SHIFT;
+	mapelm = &chunk->map[pageind-map_bias];
+	run = (arena_run_t *)((uintptr_t)chunk + (uintptr_t)((pageind -
+	    (mapelm->bits >> PAGE_SHIFT)) << PAGE_SHIFT));
+	dassert(run->magic == ARENA_RUN_MAGIC);
+	bin = run->bin;
+	binind = ((uintptr_t)bin - (uintptr_t)&arena->bins) /
+	    sizeof(arena_bin_t);
+	assert(binind < nbins);
+
+#ifdef JEMALLOC_FILL
+	if (opt_junk)
+		memset(ptr, 0x5a, arena_bin_info[binind].reg_size);
+#endif
+
+	tbin = &tcache->tbins[binind];
+	tbin_info = &tcache_bin_info[binind];
+	if (tbin->ncached == tbin_info->ncached_max) {
+		tcache_bin_flush_small(tbin, binind, (tbin_info->ncached_max >>
+		    1)
+#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
+		    , tcache
+#endif
+		    );
+	}
+	assert(tbin->ncached < tbin_info->ncached_max);
+	tbin->avail[tbin->ncached] = ptr;
+	tbin->ncached++;
+
+	tcache_event(tcache);
+}
+
+JEMALLOC_INLINE void
+tcache_dalloc_large(tcache_t *tcache, void *ptr, size_t size)
+{
+	arena_t *arena;
+	arena_chunk_t *chunk;
+	size_t pageind, binind;
+	tcache_bin_t *tbin;
+	tcache_bin_info_t *tbin_info;
+
+	assert((size & PAGE_MASK) == 0);
+	assert(arena_salloc(ptr) > small_maxclass);
+	assert(arena_salloc(ptr) <= tcache_maxclass);
+
+	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
+	arena = chunk->arena;
+	pageind = ((uintptr_t)ptr - (uintptr_t)chunk) >> PAGE_SHIFT;
+	binind = nbins + (size >> PAGE_SHIFT) - 1;
+
+#ifdef JEMALLOC_FILL
+	if (opt_junk)
+		memset(ptr, 0x5a, size);
+#endif
+
+	tbin = &tcache->tbins[binind];
+	tbin_info = &tcache_bin_info[binind];
+	if (tbin->ncached == tbin_info->ncached_max) {
+		tcache_bin_flush_large(tbin, binind, (tbin_info->ncached_max >>
+		    1)
+#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
+		    , tcache
+#endif
+		    );
+	}
+	assert(tbin->ncached < tbin_info->ncached_max);
+	tbin->avail[tbin->ncached] = ptr;
+	tbin->ncached++;
+
+	tcache_event(tcache);
+}
+#endif
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
+#endif /* JEMALLOC_TCACHE */

deps/jemalloc/include/jemalloc/internal/zone.h

@@ -0,0 +1,23 @@
+#ifndef JEMALLOC_ZONE
+#  error "This source file is for zones on Darwin (OS X)."
+#endif
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+malloc_zone_t *create_zone(void);
+void	szone2ozone(malloc_zone_t *zone);
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/

deps/jemalloc/include/jemalloc/jemalloc.h.in

@@ -0,0 +1,66 @@
+#ifndef JEMALLOC_H_
+#define	JEMALLOC_H_
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#include <limits.h>
+#include <strings.h>
+
+#define	JEMALLOC_VERSION "@jemalloc_version@"
+#define	JEMALLOC_VERSION_MAJOR @jemalloc_version_major@
+#define	JEMALLOC_VERSION_MINOR @jemalloc_version_minor@
+#define	JEMALLOC_VERSION_BUGFIX @jemalloc_version_bugfix@
+#define	JEMALLOC_VERSION_NREV @jemalloc_version_nrev@
+#define	JEMALLOC_VERSION_GID "@jemalloc_version_gid@"
+
+#include "jemalloc_defs@install_suffix@.h"
+#ifndef JEMALLOC_P
+#  define JEMALLOC_P(s) s
+#endif
+
+#define	ALLOCM_LG_ALIGN(la)	(la)
+#if LG_SIZEOF_PTR == 2
+#define	ALLOCM_ALIGN(a)	(ffs(a)-1)
+#else
+#define	ALLOCM_ALIGN(a)	((a < (size_t)INT_MAX) ? ffs(a)-1 : ffs(a>>32)+31)
+#endif
+#define	ALLOCM_ZERO	((int)0x40)
+#define	ALLOCM_NO_MOVE	((int)0x80)
+
+#define	ALLOCM_SUCCESS		0
+#define	ALLOCM_ERR_OOM		1
+#define	ALLOCM_ERR_NOT_MOVED	2
+
+extern const char	*JEMALLOC_P(malloc_conf);
+extern void		(*JEMALLOC_P(malloc_message))(void *, const char *);
+
+void	*JEMALLOC_P(malloc)(size_t size) JEMALLOC_ATTR(malloc);
+void	*JEMALLOC_P(calloc)(size_t num, size_t size) JEMALLOC_ATTR(malloc);
+int	JEMALLOC_P(posix_memalign)(void **memptr, size_t alignment, size_t size)
+    JEMALLOC_ATTR(nonnull(1));
+void	*JEMALLOC_P(realloc)(void *ptr, size_t size);
+void	JEMALLOC_P(free)(void *ptr);
+
+size_t	JEMALLOC_P(malloc_usable_size)(const void *ptr);
+void	JEMALLOC_P(malloc_stats_print)(void (*write_cb)(void *, const char *),
+    void *cbopaque, const char *opts);
+int	JEMALLOC_P(mallctl)(const char *name, void *oldp, size_t *oldlenp,
+    void *newp, size_t newlen);
+int	JEMALLOC_P(mallctlnametomib)(const char *name, size_t *mibp,
+    size_t *miblenp);
+int	JEMALLOC_P(mallctlbymib)(const size_t *mib, size_t miblen, void *oldp,
+    size_t *oldlenp, void *newp, size_t newlen);
+
+int	JEMALLOC_P(allocm)(void **ptr, size_t *rsize, size_t size, int flags)
+    JEMALLOC_ATTR(nonnull(1));
+int	JEMALLOC_P(rallocm)(void **ptr, size_t *rsize, size_t size,
+    size_t extra, int flags) JEMALLOC_ATTR(nonnull(1));
+int	JEMALLOC_P(sallocm)(const void *ptr, size_t *rsize, int flags)
+    JEMALLOC_ATTR(nonnull(1));
+int	JEMALLOC_P(dallocm)(void *ptr, int flags) JEMALLOC_ATTR(nonnull(1));
+
+#ifdef __cplusplus
+};
+#endif
+#endif /* JEMALLOC_H_ */

deps/jemalloc/include/jemalloc/jemalloc_defs.h.in

@@ -0,0 +1,158 @@
+#ifndef JEMALLOC_DEFS_H_
+#define	JEMALLOC_DEFS_H_
+
+/*
+ * If JEMALLOC_PREFIX is defined, it will cause all public APIs to be prefixed.
+ * This makes it possible, with some care, to use multiple allocators
+ * simultaneously.
+ *
+ * In many cases it is more convenient to manually prefix allocator function
+ * calls than to let macros do it automatically, particularly when using
+ * multiple allocators simultaneously.  Define JEMALLOC_MANGLE before
+ * #include'ing jemalloc.h in order to cause name mangling that corresponds to
+ * the API prefixing.
+ */
+#undef JEMALLOC_PREFIX
+#undef JEMALLOC_CPREFIX
+#if (defined(JEMALLOC_PREFIX) && defined(JEMALLOC_MANGLE))
+#undef JEMALLOC_P
+#endif
+
+/*
+ * Hyper-threaded CPUs may need a special instruction inside spin loops in
+ * order to yield to another virtual CPU.
+ */
+#undef CPU_SPINWAIT
+
+/*
+ * Defined if OSAtomic*() functions are available, as provided by Darwin, and
+ * documented in the atomic(3) manual page.
+ */
+#undef JEMALLOC_OSATOMIC
+
+/*
+ * Defined if OSSpin*() functions are available, as provided by Darwin, and
+ * documented in the spinlock(3) manual page.
+ */
+#undef JEMALLOC_OSSPIN
+
+/* Defined if __attribute__((...)) syntax is supported. */
+#undef JEMALLOC_HAVE_ATTR
+#ifdef JEMALLOC_HAVE_ATTR
+#  define JEMALLOC_ATTR(s) __attribute__((s))
+#else
+#  define JEMALLOC_ATTR(s)
+#endif
+
+/* JEMALLOC_CC_SILENCE enables code that silences unuseful compiler warnings. */
+#undef JEMALLOC_CC_SILENCE
+
+/*
+ * JEMALLOC_DEBUG enables assertions and other sanity checks, and disables
+ * inline functions.
+ */
+#undef JEMALLOC_DEBUG
+
+/* JEMALLOC_STATS enables statistics calculation. */
+#undef JEMALLOC_STATS
+
+/* JEMALLOC_PROF enables allocation profiling. */
+#undef JEMALLOC_PROF
+
+/* Use libunwind for profile backtracing if defined. */
+#undef JEMALLOC_PROF_LIBUNWIND
+
+/* Use libgcc for profile backtracing if defined. */
+#undef JEMALLOC_PROF_LIBGCC
+
+/* Use gcc intrinsics for profile backtracing if defined. */
+#undef JEMALLOC_PROF_GCC
+
+/*
+ * JEMALLOC_TINY enables support for tiny objects, which are smaller than one
+ * quantum.
+ */
+#undef JEMALLOC_TINY
+
+/*
+ * JEMALLOC_TCACHE enables a thread-specific caching layer for small objects.
+ * This makes it possible to allocate/deallocate objects without any locking
+ * when the cache is in the steady state.
+ */
+#undef JEMALLOC_TCACHE
+
+/*
+ * JEMALLOC_DSS enables use of sbrk(2) to allocate chunks from the data storage
+ * segment (DSS).
+ */
+#undef JEMALLOC_DSS
+
+/* JEMALLOC_SWAP enables mmap()ed swap file support. */
+#undef JEMALLOC_SWAP
+
+/* Support memory filling (junk/zero). */
+#undef JEMALLOC_FILL
+
+/* Support optional abort() on OOM. */
+#undef JEMALLOC_XMALLOC
+
+/* Support SYSV semantics. */
+#undef JEMALLOC_SYSV
+
+/* Support lazy locking (avoid locking unless a second thread is launched). */
+#undef JEMALLOC_LAZY_LOCK
+
+/* Determine page size at run time if defined. */
+#undef DYNAMIC_PAGE_SHIFT
+
+/* One page is 2^STATIC_PAGE_SHIFT bytes. */
+#undef STATIC_PAGE_SHIFT
+
+/* TLS is used to map arenas and magazine caches to threads. */
+#undef NO_TLS
+
+/*
+ * JEMALLOC_IVSALLOC enables ivsalloc(), which verifies that pointers reside
+ * within jemalloc-owned chunks before dereferencing them.
+ */
+#undef JEMALLOC_IVSALLOC
+
+/*
+ * Define overrides for non-standard allocator-related functions if they
+ * are present on the system.
+ */
+#undef JEMALLOC_OVERRIDE_MEMALIGN
+#undef JEMALLOC_OVERRIDE_VALLOC
+
+/*
+ * Darwin (OS X) uses zones to work around Mach-O symbol override shortcomings.
+ */
+#undef JEMALLOC_ZONE
+#undef JEMALLOC_ZONE_VERSION
+
+/* If defined, use mremap(...MREMAP_FIXED...) for huge realloc(). */
+#undef JEMALLOC_MREMAP_FIXED
+
+/*
+ * Methods for purging unused pages differ between operating systems.
+ *
+ *   madvise(..., MADV_DONTNEED) : On Linux, this immediately discards pages,
+ *                                 such that new pages will be demand-zeroed if
+ *                                 the address region is later touched.
+ *   madvise(..., MADV_FREE) : On FreeBSD and Darwin, this marks pages as being
+ *                             unused, such that they will be discarded rather
+ *                             than swapped out.
+ */
+#undef JEMALLOC_PURGE_MADVISE_DONTNEED
+#undef JEMALLOC_PURGE_MADVISE_FREE
+
+/* sizeof(void *) == 2^LG_SIZEOF_PTR. */
+#undef LG_SIZEOF_PTR
+
+/* sizeof(int) == 2^LG_SIZEOF_INT. */
+#undef LG_SIZEOF_INT
+
+/* sizeof(long) == 2^LG_SIZEOF_LONG. */
+#undef LG_SIZEOF_LONG
+
+#endif /* JEMALLOC_DEFS_H_ */

deps/jemalloc/install-sh

@@ -0,0 +1,250 @@
+#! /bin/sh
+#
+# install - install a program, script, or datafile
+# This comes from X11R5 (mit/util/scripts/install.sh).
+#
+# Copyright 1991 by the Massachusetts Institute of Technology
+#
+# Permission to use, copy, modify, distribute, and sell this software and its
+# documentation for any purpose is hereby granted without fee, provided that
+# the above copyright notice appear in all copies and that both that
+# copyright notice and this permission notice appear in supporting
+# documentation, and that the name of M.I.T. not be used in advertising or
+# publicity pertaining to distribution of the software without specific,
+# written prior permission.  M.I.T. makes no representations about the
+# suitability of this software for any purpose.  It is provided "as is"
+# without express or implied warranty.
+#
+# Calling this script install-sh is preferred over install.sh, to prevent
+# `make' implicit rules from creating a file called install from it
+# when there is no Makefile.
+#
+# This script is compatible with the BSD install script, but was written
+# from scratch.  It can only install one file at a time, a restriction
+# shared with many OS's install programs.
+
+
+# set DOITPROG to echo to test this script
+
+# Don't use :- since 4.3BSD and earlier shells don't like it.
+doit="${DOITPROG-}"
+
+
+# put in absolute paths if you don't have them in your path; or use env. vars.
+
+mvprog="${MVPROG-mv}"
+cpprog="${CPPROG-cp}"
+chmodprog="${CHMODPROG-chmod}"
+chownprog="${CHOWNPROG-chown}"
+chgrpprog="${CHGRPPROG-chgrp}"
+stripprog="${STRIPPROG-strip}"
+rmprog="${RMPROG-rm}"
+mkdirprog="${MKDIRPROG-mkdir}"
+
+transformbasename=""
+transform_arg=""
+instcmd="$mvprog"
+chmodcmd="$chmodprog 0755"
+chowncmd=""
+chgrpcmd=""
+stripcmd=""
+rmcmd="$rmprog -f"
+mvcmd="$mvprog"
+src=""
+dst=""
+dir_arg=""
+
+while [ x"$1" != x ]; do
+    case $1 in
+	-c) instcmd="$cpprog"
+	    shift
+	    continue;;
+
+	-d) dir_arg=true
+	    shift
+	    continue;;
+
+	-m) chmodcmd="$chmodprog $2"
+	    shift
+	    shift
+	    continue;;
+
+	-o) chowncmd="$chownprog $2"
+	    shift
+	    shift
+	    continue;;
+
+	-g) chgrpcmd="$chgrpprog $2"
+	    shift
+	    shift
+	    continue;;
+
+	-s) stripcmd="$stripprog"
+	    shift
+	    continue;;
+
+	-t=*) transformarg=`echo $1 | sed 's/-t=//'`
+	    shift
+	    continue;;
+
+	-b=*) transformbasename=`echo $1 | sed 's/-b=//'`
+	    shift
+	    continue;;
+
+	*)  if [ x"$src" = x ]
+	    then
+		src=$1
+	    else
+		# this colon is to work around a 386BSD /bin/sh bug
+		:
+		dst=$1
+	    fi
+	    shift
+	    continue;;
+    esac
+done
+
+if [ x"$src" = x ]
+then
+	echo "install:	no input file specified"
+	exit 1
+else
+	true
+fi
+
+if [ x"$dir_arg" != x ]; then
+	dst=$src
+	src=""
+	
+	if [ -d $dst ]; then
+		instcmd=:
+	else
+		instcmd=mkdir
+	fi
+else
+
+# Waiting for this to be detected by the "$instcmd $src $dsttmp" command
+# might cause directories to be created, which would be especially bad 
+# if $src (and thus $dsttmp) contains '*'.
+
+	if [ -f $src -o -d $src ]
+	then
+		true
+	else
+		echo "install:  $src does not exist"
+		exit 1
+	fi
+	
+	if [ x"$dst" = x ]
+	then
+		echo "install:	no destination specified"
+		exit 1
+	else
+		true
+	fi
+
+# If destination is a directory, append the input filename; if your system
+# does not like double slashes in filenames, you may need to add some logic
+
+	if [ -d $dst ]
+	then
+		dst="$dst"/`basename $src`
+	else
+		true
+	fi
+fi
+
+## this sed command emulates the dirname command
+dstdir=`echo $dst | sed -e 's,[^/]*$,,;s,/$,,;s,^$,.,'`
+
+# Make sure that the destination directory exists.
+#  this part is taken from Noah Friedman's mkinstalldirs script
+
+# Skip lots of stat calls in the usual case.
+if [ ! -d "$dstdir" ]; then
+defaultIFS='	
+'
+IFS="${IFS-${defaultIFS}}"
+
+oIFS="${IFS}"
+# Some sh's can't handle IFS=/ for some reason.
+IFS='%'
+set - `echo ${dstdir} | sed -e 's@/@%@g' -e 's@^%@/@'`
+IFS="${oIFS}"
+
+pathcomp=''
+
+while [ $# -ne 0 ] ; do
+	pathcomp="${pathcomp}${1}"
+	shift
+
+	if [ ! -d "${pathcomp}" ] ;
+        then
+		$mkdirprog "${pathcomp}"
+	else
+		true
+	fi
+
+	pathcomp="${pathcomp}/"
+done
+fi
+
+if [ x"$dir_arg" != x ]
+then
+	$doit $instcmd $dst &&
+
+	if [ x"$chowncmd" != x ]; then $doit $chowncmd $dst; else true ; fi &&
+	if [ x"$chgrpcmd" != x ]; then $doit $chgrpcmd $dst; else true ; fi &&
+	if [ x"$stripcmd" != x ]; then $doit $stripcmd $dst; else true ; fi &&
+	if [ x"$chmodcmd" != x ]; then $doit $chmodcmd $dst; else true ; fi
+else
+
+# If we're going to rename the final executable, determine the name now.
+
+	if [ x"$transformarg" = x ] 
+	then
+		dstfile=`basename $dst`
+	else
+		dstfile=`basename $dst $transformbasename | 
+			sed $transformarg`$transformbasename
+	fi
+
+# don't allow the sed command to completely eliminate the filename
+
+	if [ x"$dstfile" = x ] 
+	then
+		dstfile=`basename $dst`
+	else
+		true
+	fi
+
+# Make a temp file name in the proper directory.
+
+	dsttmp=$dstdir/#inst.$$#
+
+# Move or copy the file name to the temp name
+
+	$doit $instcmd $src $dsttmp &&
+
+	trap "rm -f ${dsttmp}" 0 &&
+
+# and set any options; do chmod last to preserve setuid bits
+
+# If any of these fail, we abort the whole thing.  If we want to
+# ignore errors from any of these, just make sure not to ignore
+# errors from the above "$doit $instcmd $src $dsttmp" command.
+
+	if [ x"$chowncmd" != x ]; then $doit $chowncmd $dsttmp; else true;fi &&
+	if [ x"$chgrpcmd" != x ]; then $doit $chgrpcmd $dsttmp; else true;fi &&
+	if [ x"$stripcmd" != x ]; then $doit $stripcmd $dsttmp; else true;fi &&
+	if [ x"$chmodcmd" != x ]; then $doit $chmodcmd $dsttmp; else true;fi &&
+
+# Now rename the file to the real destination.
+
+	$doit $rmcmd -f $dstdir/$dstfile &&
+	$doit $mvcmd $dsttmp $dstdir/$dstfile 
+
+fi &&
+
+
+exit 0

deps/jemalloc/src/atomic.c

@@ -0,0 +1,2 @@
+#define	JEMALLOC_ATOMIC_C_
+#include "jemalloc/internal/jemalloc_internal.h"

deps/jemalloc/src/base.c

@@ -0,0 +1,106 @@
+#define	JEMALLOC_BASE_C_
+#include "jemalloc/internal/jemalloc_internal.h"
+
+/******************************************************************************/
+/* Data. */
+
+malloc_mutex_t	base_mtx;
+
+/*
+ * Current pages that are being used for internal memory allocations.  These
+ * pages are carved up in cacheline-size quanta, so that there is no chance of
+ * false cache line sharing.
+ */
+static void		*base_pages;
+static void		*base_next_addr;
+static void		*base_past_addr; /* Addr immediately past base_pages. */
+static extent_node_t	*base_nodes;
+
+/******************************************************************************/
+/* Function prototypes for non-inline static functions. */
+
+static bool	base_pages_alloc(size_t minsize);
+
+/******************************************************************************/
+
+static bool
+base_pages_alloc(size_t minsize)
+{
+	size_t csize;
+	bool zero;
+
+	assert(minsize != 0);
+	csize = CHUNK_CEILING(minsize);
+	zero = false;
+	base_pages = chunk_alloc(csize, true, &zero);
+	if (base_pages == NULL)
+		return (true);
+	base_next_addr = base_pages;
+	base_past_addr = (void *)((uintptr_t)base_pages + csize);
+
+	return (false);
+}
+
+void *
+base_alloc(size_t size)
+{
+	void *ret;
+	size_t csize;
+
+	/* Round size up to nearest multiple of the cacheline size. */
+	csize = CACHELINE_CEILING(size);
+
+	malloc_mutex_lock(&base_mtx);
+	/* Make sure there's enough space for the allocation. */
+	if ((uintptr_t)base_next_addr + csize > (uintptr_t)base_past_addr) {
+		if (base_pages_alloc(csize)) {
+			malloc_mutex_unlock(&base_mtx);
+			return (NULL);
+		}
+	}
+	/* Allocate. */
+	ret = base_next_addr;
+	base_next_addr = (void *)((uintptr_t)base_next_addr + csize);
+	malloc_mutex_unlock(&base_mtx);
+
+	return (ret);
+}
+
+extent_node_t *
+base_node_alloc(void)
+{
+	extent_node_t *ret;
+
+	malloc_mutex_lock(&base_mtx);
+	if (base_nodes != NULL) {
+		ret = base_nodes;
+		base_nodes = *(extent_node_t **)ret;
+		malloc_mutex_unlock(&base_mtx);
+	} else {
+		malloc_mutex_unlock(&base_mtx);
+		ret = (extent_node_t *)base_alloc(sizeof(extent_node_t));
+	}
+
+	return (ret);
+}
+
+void
+base_node_dealloc(extent_node_t *node)
+{
+
+	malloc_mutex_lock(&base_mtx);
+	*(extent_node_t **)node = base_nodes;
+	base_nodes = node;
+	malloc_mutex_unlock(&base_mtx);
+}
+
+bool
+base_boot(void)
+{
+
+	base_nodes = NULL;
+	if (malloc_mutex_init(&base_mtx))
+		return (true);
+
+	return (false);
+}

deps/jemalloc/src/bitmap.c

@@ -0,0 +1,90 @@
+#define JEMALLOC_BITMAP_C_
+#include "jemalloc/internal/jemalloc_internal.h"
+
+/******************************************************************************/
+/* Function prototypes for non-inline static functions. */
+
+static size_t	bits2groups(size_t nbits);
+
+/******************************************************************************/
+
+static size_t
+bits2groups(size_t nbits)
+{
+
+	return ((nbits >> LG_BITMAP_GROUP_NBITS) +
+	    !!(nbits & BITMAP_GROUP_NBITS_MASK));
+}
+
+void
+bitmap_info_init(bitmap_info_t *binfo, size_t nbits)
+{
+	unsigned i;
+	size_t group_count;
+
+	assert(nbits > 0);
+	assert(nbits <= (ZU(1) << LG_BITMAP_MAXBITS));
+
+	/*
+	 * Compute the number of groups necessary to store nbits bits, and
+	 * progressively work upward through the levels until reaching a level
+	 * that requires only one group.
+	 */
+	binfo->levels[0].group_offset = 0;
+	group_count = bits2groups(nbits);
+	for (i = 1; group_count > 1; i++) {
+		assert(i < BITMAP_MAX_LEVELS);
+		binfo->levels[i].group_offset = binfo->levels[i-1].group_offset
+		    + group_count;
+		group_count = bits2groups(group_count);
+	}
+	binfo->levels[i].group_offset = binfo->levels[i-1].group_offset
+	    + group_count;
+	binfo->nlevels = i;
+	binfo->nbits = nbits;
+}
+
+size_t
+bitmap_info_ngroups(const bitmap_info_t *binfo)
+{
+
+	return (binfo->levels[binfo->nlevels].group_offset << LG_SIZEOF_BITMAP);
+}
+
+size_t
+bitmap_size(size_t nbits)
+{
+	bitmap_info_t binfo;
+
+	bitmap_info_init(&binfo, nbits);
+	return (bitmap_info_ngroups(&binfo));
+}
+
+void
+bitmap_init(bitmap_t *bitmap, const bitmap_info_t *binfo)
+{
+	size_t extra;
+	unsigned i;
+
+	/*
+	 * Bits are actually inverted with regard to the external bitmap
+	 * interface, so the bitmap starts out with all 1 bits, except for
+	 * trailing unused bits (if any).  Note that each group uses bit 0 to
+	 * correspond to the first logical bit in the group, so extra bits
+	 * are the most significant bits of the last group.
+	 */
+	memset(bitmap, 0xffU, binfo->levels[binfo->nlevels].group_offset <<
+	    LG_SIZEOF_BITMAP);
+	extra = (BITMAP_GROUP_NBITS - (binfo->nbits & BITMAP_GROUP_NBITS_MASK))
+	    & BITMAP_GROUP_NBITS_MASK;
+	if (extra != 0)
+		bitmap[binfo->levels[1].group_offset - 1] >>= extra;
+	for (i = 1; i < binfo->nlevels; i++) {
+		size_t group_count = binfo->levels[i].group_offset -
+		    binfo->levels[i-1].group_offset;
+		extra = (BITMAP_GROUP_NBITS - (group_count &
+		    BITMAP_GROUP_NBITS_MASK)) & BITMAP_GROUP_NBITS_MASK;
+		if (extra != 0)
+			bitmap[binfo->levels[i+1].group_offset - 1] >>= extra;
+	}
+}

deps/jemalloc/src/chunk.c

@@ -0,0 +1,171 @@
+#define	JEMALLOC_CHUNK_C_
+#include "jemalloc/internal/jemalloc_internal.h"
+
+/******************************************************************************/
+/* Data. */
+
+size_t	opt_lg_chunk = LG_CHUNK_DEFAULT;
+#ifdef JEMALLOC_SWAP
+bool	opt_overcommit = true;
+#endif
+
+#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
+malloc_mutex_t	chunks_mtx;
+chunk_stats_t	stats_chunks;
+#endif
+
+#ifdef JEMALLOC_IVSALLOC
+rtree_t		*chunks_rtree;
+#endif
+
+/* Various chunk-related settings. */
+size_t		chunksize;
+size_t		chunksize_mask; /* (chunksize - 1). */
+size_t		chunk_npages;
+size_t		map_bias;
+size_t		arena_maxclass; /* Max size class for arenas. */
+
+/******************************************************************************/
+
+/*
+ * If the caller specifies (*zero == false), it is still possible to receive
+ * zeroed memory, in which case *zero is toggled to true.  arena_chunk_alloc()
+ * takes advantage of this to avoid demanding zeroed chunks, but taking
+ * advantage of them if they are returned.
+ */
+void *
+chunk_alloc(size_t size, bool base, bool *zero)
+{
+	void *ret;
+
+	assert(size != 0);
+	assert((size & chunksize_mask) == 0);
+
+#ifdef JEMALLOC_SWAP
+	if (swap_enabled) {
+		ret = chunk_alloc_swap(size, zero);
+		if (ret != NULL)
+			goto RETURN;
+	}
+
+	if (swap_enabled == false || opt_overcommit) {
+#endif
+#ifdef JEMALLOC_DSS
+		ret = chunk_alloc_dss(size, zero);
+		if (ret != NULL)
+			goto RETURN;
+#endif
+		ret = chunk_alloc_mmap(size);
+		if (ret != NULL) {
+			*zero = true;
+			goto RETURN;
+		}
+#ifdef JEMALLOC_SWAP
+	}
+#endif
+
+	/* All strategies for allocation failed. */
+	ret = NULL;
+RETURN:
+#ifdef JEMALLOC_IVSALLOC
+	if (base == false && ret != NULL) {
+		if (rtree_set(chunks_rtree, (uintptr_t)ret, ret)) {
+			chunk_dealloc(ret, size);
+			return (NULL);
+		}
+	}
+#endif
+#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
+	if (ret != NULL) {
+#  ifdef JEMALLOC_PROF
+		bool gdump;
+#  endif
+		malloc_mutex_lock(&chunks_mtx);
+#  ifdef JEMALLOC_STATS
+		stats_chunks.nchunks += (size / chunksize);
+#  endif
+		stats_chunks.curchunks += (size / chunksize);
+		if (stats_chunks.curchunks > stats_chunks.highchunks) {
+			stats_chunks.highchunks = stats_chunks.curchunks;
+#  ifdef JEMALLOC_PROF
+			gdump = true;
+#  endif
+		}
+#  ifdef JEMALLOC_PROF
+		else
+			gdump = false;
+#  endif
+		malloc_mutex_unlock(&chunks_mtx);
+#  ifdef JEMALLOC_PROF
+		if (opt_prof && opt_prof_gdump && gdump)
+			prof_gdump();
+#  endif
+	}
+#endif
+
+	assert(CHUNK_ADDR2BASE(ret) == ret);
+	return (ret);
+}
+
+void
+chunk_dealloc(void *chunk, size_t size)
+{
+
+	assert(chunk != NULL);
+	assert(CHUNK_ADDR2BASE(chunk) == chunk);
+	assert(size != 0);
+	assert((size & chunksize_mask) == 0);
+
+#ifdef JEMALLOC_IVSALLOC
+	rtree_set(chunks_rtree, (uintptr_t)chunk, NULL);
+#endif
+#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
+	malloc_mutex_lock(&chunks_mtx);
+	stats_chunks.curchunks -= (size / chunksize);
+	malloc_mutex_unlock(&chunks_mtx);
+#endif
+
+#ifdef JEMALLOC_SWAP
+	if (swap_enabled && chunk_dealloc_swap(chunk, size) == false)
+		return;
+#endif
+#ifdef JEMALLOC_DSS
+	if (chunk_dealloc_dss(chunk, size) == false)
+		return;
+#endif
+	chunk_dealloc_mmap(chunk, size);
+}
+
+bool
+chunk_boot(void)
+{
+
+	/* Set variables according to the value of opt_lg_chunk. */
+	chunksize = (ZU(1) << opt_lg_chunk);
+	assert(chunksize >= PAGE_SIZE);
+	chunksize_mask = chunksize - 1;
+	chunk_npages = (chunksize >> PAGE_SHIFT);
+
+#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
+	if (malloc_mutex_init(&chunks_mtx))
+		return (true);
+	memset(&stats_chunks, 0, sizeof(chunk_stats_t));
+#endif
+#ifdef JEMALLOC_SWAP
+	if (chunk_swap_boot())
+		return (true);
+#endif
+	if (chunk_mmap_boot())
+		return (true);
+#ifdef JEMALLOC_DSS
+	if (chunk_dss_boot())
+		return (true);
+#endif
+#ifdef JEMALLOC_IVSALLOC
+	chunks_rtree = rtree_new((ZU(1) << (LG_SIZEOF_PTR+3)) - opt_lg_chunk);
+	if (chunks_rtree == NULL)
+		return (true);
+#endif
+
+	return (false);
+}

deps/jemalloc/src/chunk_dss.c

@@ -0,0 +1,284 @@
+#define	JEMALLOC_CHUNK_DSS_C_
+#include "jemalloc/internal/jemalloc_internal.h"
+#ifdef JEMALLOC_DSS
+/******************************************************************************/
+/* Data. */
+
+malloc_mutex_t	dss_mtx;
+
+/* Base address of the DSS. */
+static void	*dss_base;
+/* Current end of the DSS, or ((void *)-1) if the DSS is exhausted. */
+static void	*dss_prev;
+/* Current upper limit on DSS addresses. */
+static void	*dss_max;
+
+/*
+ * Trees of chunks that were previously allocated (trees differ only in node
+ * ordering).  These are used when allocating chunks, in an attempt to re-use
+ * address space.  Depending on function, different tree orderings are needed,
+ * which is why there are two trees with the same contents.
+ */
+static extent_tree_t	dss_chunks_szad;
+static extent_tree_t	dss_chunks_ad;
+
+/******************************************************************************/
+/* Function prototypes for non-inline static functions. */
+
+static void	*chunk_recycle_dss(size_t size, bool *zero);
+static extent_node_t *chunk_dealloc_dss_record(void *chunk, size_t size);
+
+/******************************************************************************/
+
+static void *
+chunk_recycle_dss(size_t size, bool *zero)
+{
+	extent_node_t *node, key;
+
+	key.addr = NULL;
+	key.size = size;
+	malloc_mutex_lock(&dss_mtx);
+	node = extent_tree_szad_nsearch(&dss_chunks_szad, &key);
+	if (node != NULL) {
+		void *ret = node->addr;
+
+		/* Remove node from the tree. */
+		extent_tree_szad_remove(&dss_chunks_szad, node);
+		if (node->size == size) {
+			extent_tree_ad_remove(&dss_chunks_ad, node);
+			base_node_dealloc(node);
+		} else {
+			/*
+			 * Insert the remainder of node's address range as a
+			 * smaller chunk.  Its position within dss_chunks_ad
+			 * does not change.
+			 */
+			assert(node->size > size);
+			node->addr = (void *)((uintptr_t)node->addr + size);
+			node->size -= size;
+			extent_tree_szad_insert(&dss_chunks_szad, node);
+		}
+		malloc_mutex_unlock(&dss_mtx);
+
+		if (*zero)
+			memset(ret, 0, size);
+		return (ret);
+	}
+	malloc_mutex_unlock(&dss_mtx);
+
+	return (NULL);
+}
+
+void *
+chunk_alloc_dss(size_t size, bool *zero)
+{
+	void *ret;
+
+	ret = chunk_recycle_dss(size, zero);
+	if (ret != NULL)
+		return (ret);
+
+	/*
+	 * sbrk() uses a signed increment argument, so take care not to
+	 * interpret a huge allocation request as a negative increment.
+	 */
+	if ((intptr_t)size < 0)
+		return (NULL);
+
+	malloc_mutex_lock(&dss_mtx);
+	if (dss_prev != (void *)-1) {
+		intptr_t incr;
+
+		/*
+		 * The loop is necessary to recover from races with other
+		 * threads that are using the DSS for something other than
+		 * malloc.
+		 */
+		do {
+			/* Get the current end of the DSS. */
+			dss_max = sbrk(0);
+
+			/*
+			 * Calculate how much padding is necessary to
+			 * chunk-align the end of the DSS.
+			 */
+			incr = (intptr_t)size
+			    - (intptr_t)CHUNK_ADDR2OFFSET(dss_max);
+			if (incr == (intptr_t)size)
+				ret = dss_max;
+			else {
+				ret = (void *)((intptr_t)dss_max + incr);
+				incr += size;
+			}
+
+			dss_prev = sbrk(incr);
+			if (dss_prev == dss_max) {
+				/* Success. */
+				dss_max = (void *)((intptr_t)dss_prev + incr);
+				malloc_mutex_unlock(&dss_mtx);
+				*zero = true;
+				return (ret);
+			}
+		} while (dss_prev != (void *)-1);
+	}
+	malloc_mutex_unlock(&dss_mtx);
+
+	return (NULL);
+}
+
+static extent_node_t *
+chunk_dealloc_dss_record(void *chunk, size_t size)
+{
+	extent_node_t *xnode, *node, *prev, key;
+
+	xnode = NULL;
+	while (true) {
+		key.addr = (void *)((uintptr_t)chunk + size);
+		node = extent_tree_ad_nsearch(&dss_chunks_ad, &key);
+		/* Try to coalesce forward. */
+		if (node != NULL && node->addr == key.addr) {
+			/*
+			 * Coalesce chunk with the following address range.
+			 * This does not change the position within
+			 * dss_chunks_ad, so only remove/insert from/into
+			 * dss_chunks_szad.
+			 */
+			extent_tree_szad_remove(&dss_chunks_szad, node);
+			node->addr = chunk;
+			node->size += size;
+			extent_tree_szad_insert(&dss_chunks_szad, node);
+			break;
+		} else if (xnode == NULL) {
+			/*
+			 * It is possible that base_node_alloc() will cause a
+			 * new base chunk to be allocated, so take care not to
+			 * deadlock on dss_mtx, and recover if another thread
+			 * deallocates an adjacent chunk while this one is busy
+			 * allocating xnode.
+			 */
+			malloc_mutex_unlock(&dss_mtx);
+			xnode = base_node_alloc();
+			malloc_mutex_lock(&dss_mtx);
+			if (xnode == NULL)
+				return (NULL);
+		} else {
+			/* Coalescing forward failed, so insert a new node. */
+			node = xnode;
+			xnode = NULL;
+			node->addr = chunk;
+			node->size = size;
+			extent_tree_ad_insert(&dss_chunks_ad, node);
+			extent_tree_szad_insert(&dss_chunks_szad, node);
+			break;
+		}
+	}
+	/* Discard xnode if it ended up unused do to a race. */
+	if (xnode != NULL)
+		base_node_dealloc(xnode);
+
+	/* Try to coalesce backward. */
+	prev = extent_tree_ad_prev(&dss_chunks_ad, node);
+	if (prev != NULL && (void *)((uintptr_t)prev->addr + prev->size) ==
+	    chunk) {
+		/*
+		 * Coalesce chunk with the previous address range.  This does
+		 * not change the position within dss_chunks_ad, so only
+		 * remove/insert node from/into dss_chunks_szad.
+		 */
+		extent_tree_szad_remove(&dss_chunks_szad, prev);
+		extent_tree_ad_remove(&dss_chunks_ad, prev);
+
+		extent_tree_szad_remove(&dss_chunks_szad, node);
+		node->addr = prev->addr;
+		node->size += prev->size;
+		extent_tree_szad_insert(&dss_chunks_szad, node);
+
+		base_node_dealloc(prev);
+	}
+
+	return (node);
+}
+
+bool
+chunk_in_dss(void *chunk)
+{
+	bool ret;
+
+	malloc_mutex_lock(&dss_mtx);
+	if ((uintptr_t)chunk >= (uintptr_t)dss_base
+	    && (uintptr_t)chunk < (uintptr_t)dss_max)
+		ret = true;
+	else
+		ret = false;
+	malloc_mutex_unlock(&dss_mtx);
+
+	return (ret);
+}
+
+bool
+chunk_dealloc_dss(void *chunk, size_t size)
+{
+	bool ret;
+
+	malloc_mutex_lock(&dss_mtx);
+	if ((uintptr_t)chunk >= (uintptr_t)dss_base
+	    && (uintptr_t)chunk < (uintptr_t)dss_max) {
+		extent_node_t *node;
+
+		/* Try to coalesce with other unused chunks. */
+		node = chunk_dealloc_dss_record(chunk, size);
+		if (node != NULL) {
+			chunk = node->addr;
+			size = node->size;
+		}
+
+		/* Get the current end of the DSS. */
+		dss_max = sbrk(0);
+
+		/*
+		 * Try to shrink the DSS if this chunk is at the end of the
+		 * DSS.  The sbrk() call here is subject to a race condition
+		 * with threads that use brk(2) or sbrk(2) directly, but the
+		 * alternative would be to leak memory for the sake of poorly
+		 * designed multi-threaded programs.
+		 */
+		if ((void *)((uintptr_t)chunk + size) == dss_max
+		    && (dss_prev = sbrk(-(intptr_t)size)) == dss_max) {
+			/* Success. */
+			dss_max = (void *)((intptr_t)dss_prev - (intptr_t)size);
+
+			if (node != NULL) {
+				extent_tree_szad_remove(&dss_chunks_szad, node);
+				extent_tree_ad_remove(&dss_chunks_ad, node);
+				base_node_dealloc(node);
+			}
+		} else
+			madvise(chunk, size, MADV_DONTNEED);
+
+		ret = false;
+		goto RETURN;
+	}
+
+	ret = true;
+RETURN:
+	malloc_mutex_unlock(&dss_mtx);
+	return (ret);
+}
+
+bool
+chunk_dss_boot(void)
+{
+
+	if (malloc_mutex_init(&dss_mtx))
+		return (true);
+	dss_base = sbrk(0);
+	dss_prev = dss_base;
+	dss_max = dss_base;
+	extent_tree_szad_new(&dss_chunks_szad);
+	extent_tree_ad_new(&dss_chunks_ad);
+
+	return (false);
+}
+
+/******************************************************************************/
+#endif /* JEMALLOC_DSS */

deps/jemalloc/src/chunk_mmap.c

@@ -0,0 +1,239 @@
+#define	JEMALLOC_CHUNK_MMAP_C_
+#include "jemalloc/internal/jemalloc_internal.h"
+
+/******************************************************************************/
+/* Data. */
+
+/*
+ * Used by chunk_alloc_mmap() to decide whether to attempt the fast path and
+ * potentially avoid some system calls.
+ */
+#ifndef NO_TLS
+static __thread bool	mmap_unaligned_tls
+    JEMALLOC_ATTR(tls_model("initial-exec"));
+#define	MMAP_UNALIGNED_GET()	mmap_unaligned_tls
+#define	MMAP_UNALIGNED_SET(v)	do {					\
+	mmap_unaligned_tls = (v);					\
+} while (0)
+#else
+static pthread_key_t	mmap_unaligned_tsd;
+#define	MMAP_UNALIGNED_GET()	((bool)pthread_getspecific(mmap_unaligned_tsd))
+#define	MMAP_UNALIGNED_SET(v)	do {					\
+	pthread_setspecific(mmap_unaligned_tsd, (void *)(v));		\
+} while (0)
+#endif
+
+/******************************************************************************/
+/* Function prototypes for non-inline static functions. */
+
+static void	*pages_map(void *addr, size_t size, bool noreserve);
+static void	pages_unmap(void *addr, size_t size);
+static void	*chunk_alloc_mmap_slow(size_t size, bool unaligned,
+    bool noreserve);
+static void	*chunk_alloc_mmap_internal(size_t size, bool noreserve);
+
+/******************************************************************************/
+
+static void *
+pages_map(void *addr, size_t size, bool noreserve)
+{
+	void *ret;
+
+	/*
+	 * We don't use MAP_FIXED here, because it can cause the *replacement*
+	 * of existing mappings, and we only want to create new mappings.
+	 */
+	int flags = MAP_PRIVATE | MAP_ANON;
+#ifdef MAP_NORESERVE
+	if (noreserve)
+		flags |= MAP_NORESERVE;
+#endif
+	ret = mmap(addr, size, PROT_READ | PROT_WRITE, flags, -1, 0);
+	assert(ret != NULL);
+
+	if (ret == MAP_FAILED)
+		ret = NULL;
+	else if (addr != NULL && ret != addr) {
+		/*
+		 * We succeeded in mapping memory, but not in the right place.
+		 */
+		if (munmap(ret, size) == -1) {
+			char buf[BUFERROR_BUF];
+
+			buferror(errno, buf, sizeof(buf));
+			malloc_write("<jemalloc>: Error in munmap(): ");
+			malloc_write(buf);
+			malloc_write("\n");
+			if (opt_abort)
+				abort();
+		}
+		ret = NULL;
+	}
+
+	assert(ret == NULL || (addr == NULL && ret != addr)
+	    || (addr != NULL && ret == addr));
+	return (ret);
+}
+
+static void
+pages_unmap(void *addr, size_t size)
+{
+
+	if (munmap(addr, size) == -1) {
+		char buf[BUFERROR_BUF];
+
+		buferror(errno, buf, sizeof(buf));
+		malloc_write("<jemalloc>: Error in munmap(): ");
+		malloc_write(buf);
+		malloc_write("\n");
+		if (opt_abort)
+			abort();
+	}
+}
+
+static void *
+chunk_alloc_mmap_slow(size_t size, bool unaligned, bool noreserve)
+{
+	void *ret;
+	size_t offset;
+
+	/* Beware size_t wrap-around. */
+	if (size + chunksize <= size)
+		return (NULL);
+
+	ret = pages_map(NULL, size + chunksize, noreserve);
+	if (ret == NULL)
+		return (NULL);
+
+	/* Clean up unneeded leading/trailing space. */
+	offset = CHUNK_ADDR2OFFSET(ret);
+	if (offset != 0) {
+		/* Note that mmap() returned an unaligned mapping. */
+		unaligned = true;
+
+		/* Leading space. */
+		pages_unmap(ret, chunksize - offset);
+
+		ret = (void *)((uintptr_t)ret +
+		    (chunksize - offset));
+
+		/* Trailing space. */
+		pages_unmap((void *)((uintptr_t)ret + size),
+		    offset);
+	} else {
+		/* Trailing space only. */
+		pages_unmap((void *)((uintptr_t)ret + size),
+		    chunksize);
+	}
+
+	/*
+	 * If mmap() returned an aligned mapping, reset mmap_unaligned so that
+	 * the next chunk_alloc_mmap() execution tries the fast allocation
+	 * method.
+	 */
+	if (unaligned == false)
+		MMAP_UNALIGNED_SET(false);
+
+	return (ret);
+}
+
+static void *
+chunk_alloc_mmap_internal(size_t size, bool noreserve)
+{
+	void *ret;
+
+	/*
+	 * Ideally, there would be a way to specify alignment to mmap() (like
+	 * NetBSD has), but in the absence of such a feature, we have to work
+	 * hard to efficiently create aligned mappings.  The reliable, but
+	 * slow method is to create a mapping that is over-sized, then trim the
+	 * excess.  However, that always results in at least one call to
+	 * pages_unmap().
+	 *
+	 * A more optimistic approach is to try mapping precisely the right
+	 * amount, then try to append another mapping if alignment is off.  In
+	 * practice, this works out well as long as the application is not
+	 * interleaving mappings via direct mmap() calls.  If we do run into a
+	 * situation where there is an interleaved mapping and we are unable to
+	 * extend an unaligned mapping, our best option is to switch to the
+	 * slow method until mmap() returns another aligned mapping.  This will
+	 * tend to leave a gap in the memory map that is too small to cause
+	 * later problems for the optimistic method.
+	 *
+	 * Another possible confounding factor is address space layout
+	 * randomization (ASLR), which causes mmap(2) to disregard the
+	 * requested address.  mmap_unaligned tracks whether the previous
+	 * chunk_alloc_mmap() execution received any unaligned or relocated
+	 * mappings, and if so, the current execution will immediately fall
+	 * back to the slow method.  However, we keep track of whether the fast
+	 * method would have succeeded, and if so, we make a note to try the
+	 * fast method next time.
+	 */
+
+	if (MMAP_UNALIGNED_GET() == false) {
+		size_t offset;
+
+		ret = pages_map(NULL, size, noreserve);
+		if (ret == NULL)
+			return (NULL);
+
+		offset = CHUNK_ADDR2OFFSET(ret);
+		if (offset != 0) {
+			MMAP_UNALIGNED_SET(true);
+			/* Try to extend chunk boundary. */
+			if (pages_map((void *)((uintptr_t)ret + size),
+			    chunksize - offset, noreserve) == NULL) {
+				/*
+				 * Extension failed.  Clean up, then revert to
+				 * the reliable-but-expensive method.
+				 */
+				pages_unmap(ret, size);
+				ret = chunk_alloc_mmap_slow(size, true,
+				    noreserve);
+			} else {
+				/* Clean up unneeded leading space. */
+				pages_unmap(ret, chunksize - offset);
+				ret = (void *)((uintptr_t)ret + (chunksize -
+				    offset));
+			}
+		}
+	} else
+		ret = chunk_alloc_mmap_slow(size, false, noreserve);
+
+	return (ret);
+}
+
+void *
+chunk_alloc_mmap(size_t size)
+{
+
+	return (chunk_alloc_mmap_internal(size, false));
+}
+
+void *
+chunk_alloc_mmap_noreserve(size_t size)
+{
+
+	return (chunk_alloc_mmap_internal(size, true));
+}
+
+void
+chunk_dealloc_mmap(void *chunk, size_t size)
+{
+
+	pages_unmap(chunk, size);
+}
+
+bool
+chunk_mmap_boot(void)
+{
+
+#ifdef NO_TLS
+	if (pthread_key_create(&mmap_unaligned_tsd, NULL) != 0) {
+		malloc_write("<jemalloc>: Error in pthread_key_create()\n");
+		return (true);
+	}
+#endif
+
+	return (false);
+}

deps/jemalloc/src/chunk_swap.c

@@ -0,0 +1,402 @@
+#define	JEMALLOC_CHUNK_SWAP_C_
+#include "jemalloc/internal/jemalloc_internal.h"
+#ifdef JEMALLOC_SWAP
+/******************************************************************************/
+/* Data. */
+
+malloc_mutex_t	swap_mtx;
+bool		swap_enabled;
+bool		swap_prezeroed;
+size_t		swap_nfds;
+int		*swap_fds;
+#ifdef JEMALLOC_STATS
+size_t		swap_avail;
+#endif
+
+/* Base address of the mmap()ed file(s). */
+static void	*swap_base;
+/* Current end of the space in use (<= swap_max). */
+static void	*swap_end;
+/* Absolute upper limit on file-backed addresses. */
+static void	*swap_max;
+
+/*
+ * Trees of chunks that were previously allocated (trees differ only in node
+ * ordering).  These are used when allocating chunks, in an attempt to re-use
+ * address space.  Depending on function, different tree orderings are needed,
+ * which is why there are two trees with the same contents.
+ */
+static extent_tree_t	swap_chunks_szad;
+static extent_tree_t	swap_chunks_ad;
+
+/******************************************************************************/
+/* Function prototypes for non-inline static functions. */
+
+static void	*chunk_recycle_swap(size_t size, bool *zero);
+static extent_node_t *chunk_dealloc_swap_record(void *chunk, size_t size);
+
+/******************************************************************************/
+
+static void *
+chunk_recycle_swap(size_t size, bool *zero)
+{
+	extent_node_t *node, key;
+
+	key.addr = NULL;
+	key.size = size;
+	malloc_mutex_lock(&swap_mtx);
+	node = extent_tree_szad_nsearch(&swap_chunks_szad, &key);
+	if (node != NULL) {
+		void *ret = node->addr;
+
+		/* Remove node from the tree. */
+		extent_tree_szad_remove(&swap_chunks_szad, node);
+		if (node->size == size) {
+			extent_tree_ad_remove(&swap_chunks_ad, node);
+			base_node_dealloc(node);
+		} else {
+			/*
+			 * Insert the remainder of node's address range as a
+			 * smaller chunk.  Its position within swap_chunks_ad
+			 * does not change.
+			 */
+			assert(node->size > size);
+			node->addr = (void *)((uintptr_t)node->addr + size);
+			node->size -= size;
+			extent_tree_szad_insert(&swap_chunks_szad, node);
+		}
+#ifdef JEMALLOC_STATS
+		swap_avail -= size;
+#endif
+		malloc_mutex_unlock(&swap_mtx);
+
+		if (*zero)
+			memset(ret, 0, size);
+		return (ret);
+	}
+	malloc_mutex_unlock(&swap_mtx);
+
+	return (NULL);
+}
+
+void *
+chunk_alloc_swap(size_t size, bool *zero)
+{
+	void *ret;
+
+	assert(swap_enabled);
+
+	ret = chunk_recycle_swap(size, zero);
+	if (ret != NULL)
+		return (ret);
+
+	malloc_mutex_lock(&swap_mtx);
+	if ((uintptr_t)swap_end + size <= (uintptr_t)swap_max) {
+		ret = swap_end;
+		swap_end = (void *)((uintptr_t)swap_end + size);
+#ifdef JEMALLOC_STATS
+		swap_avail -= size;
+#endif
+		malloc_mutex_unlock(&swap_mtx);
+
+		if (swap_prezeroed)
+			*zero = true;
+		else if (*zero)
+			memset(ret, 0, size);
+	} else {
+		malloc_mutex_unlock(&swap_mtx);
+		return (NULL);
+	}
+
+	return (ret);
+}
+
+static extent_node_t *
+chunk_dealloc_swap_record(void *chunk, size_t size)
+{
+	extent_node_t *xnode, *node, *prev, key;
+
+	xnode = NULL;
+	while (true) {
+		key.addr = (void *)((uintptr_t)chunk + size);
+		node = extent_tree_ad_nsearch(&swap_chunks_ad, &key);
+		/* Try to coalesce forward. */
+		if (node != NULL && node->addr == key.addr) {
+			/*
+			 * Coalesce chunk with the following address range.
+			 * This does not change the position within
+			 * swap_chunks_ad, so only remove/insert from/into
+			 * swap_chunks_szad.
+			 */
+			extent_tree_szad_remove(&swap_chunks_szad, node);
+			node->addr = chunk;
+			node->size += size;
+			extent_tree_szad_insert(&swap_chunks_szad, node);
+			break;
+		} else if (xnode == NULL) {
+			/*
+			 * It is possible that base_node_alloc() will cause a
+			 * new base chunk to be allocated, so take care not to
+			 * deadlock on swap_mtx, and recover if another thread
+			 * deallocates an adjacent chunk while this one is busy
+			 * allocating xnode.
+			 */
+			malloc_mutex_unlock(&swap_mtx);
+			xnode = base_node_alloc();
+			malloc_mutex_lock(&swap_mtx);
+			if (xnode == NULL)
+				return (NULL);
+		} else {
+			/* Coalescing forward failed, so insert a new node. */
+			node = xnode;
+			xnode = NULL;
+			node->addr = chunk;
+			node->size = size;
+			extent_tree_ad_insert(&swap_chunks_ad, node);
+			extent_tree_szad_insert(&swap_chunks_szad, node);
+			break;
+		}
+	}
+	/* Discard xnode if it ended up unused do to a race. */
+	if (xnode != NULL)
+		base_node_dealloc(xnode);
+
+	/* Try to coalesce backward. */
+	prev = extent_tree_ad_prev(&swap_chunks_ad, node);
+	if (prev != NULL && (void *)((uintptr_t)prev->addr + prev->size) ==
+	    chunk) {
+		/*
+		 * Coalesce chunk with the previous address range.  This does
+		 * not change the position within swap_chunks_ad, so only
+		 * remove/insert node from/into swap_chunks_szad.
+		 */
+		extent_tree_szad_remove(&swap_chunks_szad, prev);
+		extent_tree_ad_remove(&swap_chunks_ad, prev);
+
+		extent_tree_szad_remove(&swap_chunks_szad, node);
+		node->addr = prev->addr;
+		node->size += prev->size;
+		extent_tree_szad_insert(&swap_chunks_szad, node);
+
+		base_node_dealloc(prev);
+	}
+
+	return (node);
+}
+
+bool
+chunk_in_swap(void *chunk)
+{
+	bool ret;
+
+	assert(swap_enabled);
+
+	malloc_mutex_lock(&swap_mtx);
+	if ((uintptr_t)chunk >= (uintptr_t)swap_base
+	    && (uintptr_t)chunk < (uintptr_t)swap_max)
+		ret = true;
+	else
+		ret = false;
+	malloc_mutex_unlock(&swap_mtx);
+
+	return (ret);
+}
+
+bool
+chunk_dealloc_swap(void *chunk, size_t size)
+{
+	bool ret;
+
+	assert(swap_enabled);
+
+	malloc_mutex_lock(&swap_mtx);
+	if ((uintptr_t)chunk >= (uintptr_t)swap_base
+	    && (uintptr_t)chunk < (uintptr_t)swap_max) {
+		extent_node_t *node;
+
+		/* Try to coalesce with other unused chunks. */
+		node = chunk_dealloc_swap_record(chunk, size);
+		if (node != NULL) {
+			chunk = node->addr;
+			size = node->size;
+		}
+
+		/*
+		 * Try to shrink the in-use memory if this chunk is at the end
+		 * of the in-use memory.
+		 */
+		if ((void *)((uintptr_t)chunk + size) == swap_end) {
+			swap_end = (void *)((uintptr_t)swap_end - size);
+
+			if (node != NULL) {
+				extent_tree_szad_remove(&swap_chunks_szad,
+				    node);
+				extent_tree_ad_remove(&swap_chunks_ad, node);
+				base_node_dealloc(node);
+			}
+		} else
+			madvise(chunk, size, MADV_DONTNEED);
+
+#ifdef JEMALLOC_STATS
+		swap_avail += size;
+#endif
+		ret = false;
+		goto RETURN;
+	}
+
+	ret = true;
+RETURN:
+	malloc_mutex_unlock(&swap_mtx);
+	return (ret);
+}
+
+bool
+chunk_swap_enable(const int *fds, unsigned nfds, bool prezeroed)
+{
+	bool ret;
+	unsigned i;
+	off_t off;
+	void *vaddr;
+	size_t cumsize, voff;
+	size_t sizes[nfds];
+
+	malloc_mutex_lock(&swap_mtx);
+
+	/* Get file sizes. */
+	for (i = 0, cumsize = 0; i < nfds; i++) {
+		off = lseek(fds[i], 0, SEEK_END);
+		if (off == ((off_t)-1)) {
+			ret = true;
+			goto RETURN;
+		}
+		if (PAGE_CEILING(off) != off) {
+			/* Truncate to a multiple of the page size. */
+			off &= ~PAGE_MASK;
+			if (ftruncate(fds[i], off) != 0) {
+				ret = true;
+				goto RETURN;
+			}
+		}
+		sizes[i] = off;
+		if (cumsize + off < cumsize) {
+			/*
+			 * Cumulative file size is greater than the total
+			 * address space.  Bail out while it's still obvious
+			 * what the problem is.
+			 */
+			ret = true;
+			goto RETURN;
+		}
+		cumsize += off;
+	}
+
+	/* Round down to a multiple of the chunk size. */
+	cumsize &= ~chunksize_mask;
+	if (cumsize == 0) {
+		ret = true;
+		goto RETURN;
+	}
+
+	/*
+	 * Allocate a chunk-aligned region of anonymous memory, which will
+	 * be the final location for the memory-mapped files.
+	 */
+	vaddr = chunk_alloc_mmap_noreserve(cumsize);
+	if (vaddr == NULL) {
+		ret = true;
+		goto RETURN;
+	}
+
+	/* Overlay the files onto the anonymous mapping. */
+	for (i = 0, voff = 0; i < nfds; i++) {
+		void *addr = mmap((void *)((uintptr_t)vaddr + voff), sizes[i],
+		    PROT_READ | PROT_WRITE, MAP_SHARED | MAP_FIXED, fds[i], 0);
+		if (addr == MAP_FAILED) {
+			char buf[BUFERROR_BUF];
+
+
+			buferror(errno, buf, sizeof(buf));
+			malloc_write(
+			    "<jemalloc>: Error in mmap(..., MAP_FIXED, ...): ");
+			malloc_write(buf);
+			malloc_write("\n");
+			if (opt_abort)
+				abort();
+			if (munmap(vaddr, voff) == -1) {
+				buferror(errno, buf, sizeof(buf));
+				malloc_write("<jemalloc>: Error in munmap(): ");
+				malloc_write(buf);
+				malloc_write("\n");
+			}
+			ret = true;
+			goto RETURN;
+		}
+		assert(addr == (void *)((uintptr_t)vaddr + voff));
+
+		/*
+		 * Tell the kernel that the mapping will be accessed randomly,
+		 * and that it should not gratuitously sync pages to the
+		 * filesystem.
+		 */
+#ifdef MADV_RANDOM
+		madvise(addr, sizes[i], MADV_RANDOM);
+#endif
+#ifdef MADV_NOSYNC
+		madvise(addr, sizes[i], MADV_NOSYNC);
+#endif
+
+		voff += sizes[i];
+	}
+
+	swap_prezeroed = prezeroed;
+	swap_base = vaddr;
+	swap_end = swap_base;
+	swap_max = (void *)((uintptr_t)vaddr + cumsize);
+
+	/* Copy the fds array for mallctl purposes. */
+	swap_fds = (int *)base_alloc(nfds * sizeof(int));
+	if (swap_fds == NULL) {
+		ret = true;
+		goto RETURN;
+	}
+	memcpy(swap_fds, fds, nfds * sizeof(int));
+	swap_nfds = nfds;
+
+#ifdef JEMALLOC_STATS
+	swap_avail = cumsize;
+#endif
+
+	swap_enabled = true;
+
+	ret = false;
+RETURN:
+	malloc_mutex_unlock(&swap_mtx);
+	return (ret);
+}
+
+bool
+chunk_swap_boot(void)
+{
+
+	if (malloc_mutex_init(&swap_mtx))
+		return (true);
+
+	swap_enabled = false;
+	swap_prezeroed = false; /* swap.* mallctl's depend on this. */
+	swap_nfds = 0;
+	swap_fds = NULL;
+#ifdef JEMALLOC_STATS
+	swap_avail = 0;
+#endif
+	swap_base = NULL;
+	swap_end = NULL;
+	swap_max = NULL;
+
+	extent_tree_szad_new(&swap_chunks_szad);
+	extent_tree_ad_new(&swap_chunks_ad);
+
+	return (false);
+}
+
+/******************************************************************************/
+#endif /* JEMALLOC_SWAP */

deps/jemalloc/src/ckh.c

@@ -0,0 +1,619 @@
+/*
+ *******************************************************************************
+ * Implementation of (2^1+,2) cuckoo hashing, where 2^1+ indicates that each
+ * hash bucket contains 2^n cells, for n >= 1, and 2 indicates that two hash
+ * functions are employed.  The original cuckoo hashing algorithm was described
+ * in:
+ *
+ *   Pagh, R., F.F. Rodler (2004) Cuckoo Hashing.  Journal of Algorithms
+ *     51(2):122-144.
+ *
+ * Generalization of cuckoo hashing was discussed in:
+ *
+ *   Erlingsson, U., M. Manasse, F. McSherry (2006) A cool and practical
+ *     alternative to traditional hash tables.  In Proceedings of the 7th
+ *     Workshop on Distributed Data and Structures (WDAS'06), Santa Clara, CA,
+ *     January 2006.
+ *
+ * This implementation uses precisely two hash functions because that is the
+ * fewest that can work, and supporting multiple hashes is an implementation
+ * burden.  Here is a reproduction of Figure 1 from Erlingsson et al. (2006)
+ * that shows approximate expected maximum load factors for various
+ * configurations:
+ *
+ *           |         #cells/bucket         |
+ *   #hashes |   1   |   2   |   4   |   8   |
+ *   --------+-------+-------+-------+-------+
+ *         1 | 0.006 | 0.006 | 0.03  | 0.12  |
+ *         2 | 0.49  | 0.86  |>0.93< |>0.96< |
+ *         3 | 0.91  | 0.97  | 0.98  | 0.999 |
+ *         4 | 0.97  | 0.99  | 0.999 |       |
+ *
+ * The number of cells per bucket is chosen such that a bucket fits in one cache
+ * line.  So, on 32- and 64-bit systems, we use (8,2) and (4,2) cuckoo hashing,
+ * respectively.
+ *
+ ******************************************************************************/
+#define	JEMALLOC_CKH_C_
+#include "jemalloc/internal/jemalloc_internal.h"
+
+/******************************************************************************/
+/* Function prototypes for non-inline static functions. */
+
+static bool	ckh_grow(ckh_t *ckh);
+static void	ckh_shrink(ckh_t *ckh);
+
+/******************************************************************************/
+
+/*
+ * Search bucket for key and return the cell number if found; SIZE_T_MAX
+ * otherwise.
+ */
+JEMALLOC_INLINE size_t
+ckh_bucket_search(ckh_t *ckh, size_t bucket, const void *key)
+{
+	ckhc_t *cell;
+	unsigned i;
+
+	for (i = 0; i < (ZU(1) << LG_CKH_BUCKET_CELLS); i++) {
+		cell = &ckh->tab[(bucket << LG_CKH_BUCKET_CELLS) + i];
+		if (cell->key != NULL && ckh->keycomp(key, cell->key))
+			return ((bucket << LG_CKH_BUCKET_CELLS) + i);
+	}
+
+	return (SIZE_T_MAX);
+}
+
+/*
+ * Search table for key and return cell number if found; SIZE_T_MAX otherwise.
+ */
+JEMALLOC_INLINE size_t
+ckh_isearch(ckh_t *ckh, const void *key)
+{
+	size_t hash1, hash2, bucket, cell;
+
+	assert(ckh != NULL);
+	dassert(ckh->magic == CKH_MAGIC);
+
+	ckh->hash(key, ckh->lg_curbuckets, &hash1, &hash2);
+
+	/* Search primary bucket. */
+	bucket = hash1 & ((ZU(1) << ckh->lg_curbuckets) - 1);
+	cell = ckh_bucket_search(ckh, bucket, key);
+	if (cell != SIZE_T_MAX)
+		return (cell);
+
+	/* Search secondary bucket. */
+	bucket = hash2 & ((ZU(1) << ckh->lg_curbuckets) - 1);
+	cell = ckh_bucket_search(ckh, bucket, key);
+	return (cell);
+}
+
+JEMALLOC_INLINE bool
+ckh_try_bucket_insert(ckh_t *ckh, size_t bucket, const void *key,
+    const void *data)
+{
+	ckhc_t *cell;
+	unsigned offset, i;
+
+	/*
+	 * Cycle through the cells in the bucket, starting at a random position.
+	 * The randomness avoids worst-case search overhead as buckets fill up.
+	 */
+	prn32(offset, LG_CKH_BUCKET_CELLS, ckh->prn_state, CKH_A, CKH_C);
+	for (i = 0; i < (ZU(1) << LG_CKH_BUCKET_CELLS); i++) {
+		cell = &ckh->tab[(bucket << LG_CKH_BUCKET_CELLS) +
+		    ((i + offset) & ((ZU(1) << LG_CKH_BUCKET_CELLS) - 1))];
+		if (cell->key == NULL) {
+			cell->key = key;
+			cell->data = data;
+			ckh->count++;
+			return (false);
+		}
+	}
+
+	return (true);
+}
+
+/*
+ * No space is available in bucket.  Randomly evict an item, then try to find an
+ * alternate location for that item.  Iteratively repeat this
+ * eviction/relocation procedure until either success or detection of an
+ * eviction/relocation bucket cycle.
+ */
+JEMALLOC_INLINE bool
+ckh_evict_reloc_insert(ckh_t *ckh, size_t argbucket, void const **argkey,
+    void const **argdata)
+{
+	const void *key, *data, *tkey, *tdata;
+	ckhc_t *cell;
+	size_t hash1, hash2, bucket, tbucket;
+	unsigned i;
+
+	bucket = argbucket;
+	key = *argkey;
+	data = *argdata;
+	while (true) {
+		/*
+		 * Choose a random item within the bucket to evict.  This is
+		 * critical to correct function, because without (eventually)
+		 * evicting all items within a bucket during iteration, it
+		 * would be possible to get stuck in an infinite loop if there
+		 * were an item for which both hashes indicated the same
+		 * bucket.
+		 */
+		prn32(i, LG_CKH_BUCKET_CELLS, ckh->prn_state, CKH_A, CKH_C);
+		cell = &ckh->tab[(bucket << LG_CKH_BUCKET_CELLS) + i];
+		assert(cell->key != NULL);
+
+		/* Swap cell->{key,data} and {key,data} (evict). */
+		tkey = cell->key; tdata = cell->data;
+		cell->key = key; cell->data = data;
+		key = tkey; data = tdata;
+
+#ifdef CKH_COUNT
+		ckh->nrelocs++;
+#endif
+
+		/* Find the alternate bucket for the evicted item. */
+		ckh->hash(key, ckh->lg_curbuckets, &hash1, &hash2);
+		tbucket = hash2 & ((ZU(1) << ckh->lg_curbuckets) - 1);
+		if (tbucket == bucket) {
+			tbucket = hash1 & ((ZU(1) << ckh->lg_curbuckets) - 1);
+			/*
+			 * It may be that (tbucket == bucket) still, if the
+			 * item's hashes both indicate this bucket.  However,
+			 * we are guaranteed to eventually escape this bucket
+			 * during iteration, assuming pseudo-random item
+			 * selection (true randomness would make infinite
+			 * looping a remote possibility).  The reason we can
+			 * never get trapped forever is that there are two
+			 * cases:
+			 *
+			 * 1) This bucket == argbucket, so we will quickly
+			 *    detect an eviction cycle and terminate.
+			 * 2) An item was evicted to this bucket from another,
+			 *    which means that at least one item in this bucket
+			 *    has hashes that indicate distinct buckets.
+			 */
+		}
+		/* Check for a cycle. */
+		if (tbucket == argbucket) {
+			*argkey = key;
+			*argdata = data;
+			return (true);
+		}
+
+		bucket = tbucket;
+		if (ckh_try_bucket_insert(ckh, bucket, key, data) == false)
+			return (false);
+	}
+}
+
+JEMALLOC_INLINE bool
+ckh_try_insert(ckh_t *ckh, void const**argkey, void const**argdata)
+{
+	size_t hash1, hash2, bucket;
+	const void *key = *argkey;
+	const void *data = *argdata;
+
+	ckh->hash(key, ckh->lg_curbuckets, &hash1, &hash2);
+
+	/* Try to insert in primary bucket. */
+	bucket = hash1 & ((ZU(1) << ckh->lg_curbuckets) - 1);
+	if (ckh_try_bucket_insert(ckh, bucket, key, data) == false)
+		return (false);
+
+	/* Try to insert in secondary bucket. */
+	bucket = hash2 & ((ZU(1) << ckh->lg_curbuckets) - 1);
+	if (ckh_try_bucket_insert(ckh, bucket, key, data) == false)
+		return (false);
+
+	/*
+	 * Try to find a place for this item via iterative eviction/relocation.
+	 */
+	return (ckh_evict_reloc_insert(ckh, bucket, argkey, argdata));
+}
+
+/*
+ * Try to rebuild the hash table from scratch by inserting all items from the
+ * old table into the new.
+ */
+JEMALLOC_INLINE bool
+ckh_rebuild(ckh_t *ckh, ckhc_t *aTab)
+{
+	size_t count, i, nins;
+	const void *key, *data;
+
+	count = ckh->count;
+	ckh->count = 0;
+	for (i = nins = 0; nins < count; i++) {
+		if (aTab[i].key != NULL) {
+			key = aTab[i].key;
+			data = aTab[i].data;
+			if (ckh_try_insert(ckh, &key, &data)) {
+				ckh->count = count;
+				return (true);
+			}
+			nins++;
+		}
+	}
+
+	return (false);
+}
+
+static bool
+ckh_grow(ckh_t *ckh)
+{
+	bool ret;
+	ckhc_t *tab, *ttab;
+	size_t lg_curcells;
+	unsigned lg_prevbuckets;
+
+#ifdef CKH_COUNT
+	ckh->ngrows++;
+#endif
+
+	/*
+	 * It is possible (though unlikely, given well behaved hashes) that the
+	 * table will have to be doubled more than once in order to create a
+	 * usable table.
+	 */
+	lg_prevbuckets = ckh->lg_curbuckets;
+	lg_curcells = ckh->lg_curbuckets + LG_CKH_BUCKET_CELLS;
+	while (true) {
+		size_t usize;
+
+		lg_curcells++;
+		usize = sa2u(sizeof(ckhc_t) << lg_curcells, CACHELINE, NULL);
+		if (usize == 0) {
+			ret = true;
+			goto RETURN;
+		}
+		tab = (ckhc_t *)ipalloc(usize, CACHELINE, true);
+		if (tab == NULL) {
+			ret = true;
+			goto RETURN;
+		}
+		/* Swap in new table. */
+		ttab = ckh->tab;
+		ckh->tab = tab;
+		tab = ttab;
+		ckh->lg_curbuckets = lg_curcells - LG_CKH_BUCKET_CELLS;
+
+		if (ckh_rebuild(ckh, tab) == false) {
+			idalloc(tab);
+			break;
+		}
+
+		/* Rebuilding failed, so back out partially rebuilt table. */
+		idalloc(ckh->tab);
+		ckh->tab = tab;
+		ckh->lg_curbuckets = lg_prevbuckets;
+	}
+
+	ret = false;
+RETURN:
+	return (ret);
+}
+
+static void
+ckh_shrink(ckh_t *ckh)
+{
+	ckhc_t *tab, *ttab;
+	size_t lg_curcells, usize;
+	unsigned lg_prevbuckets;
+
+	/*
+	 * It is possible (though unlikely, given well behaved hashes) that the
+	 * table rebuild will fail.
+	 */
+	lg_prevbuckets = ckh->lg_curbuckets;
+	lg_curcells = ckh->lg_curbuckets + LG_CKH_BUCKET_CELLS - 1;
+	usize = sa2u(sizeof(ckhc_t) << lg_curcells, CACHELINE, NULL);
+	if (usize == 0)
+		return;
+	tab = (ckhc_t *)ipalloc(usize, CACHELINE, true);
+	if (tab == NULL) {
+		/*
+		 * An OOM error isn't worth propagating, since it doesn't
+		 * prevent this or future operations from proceeding.
+		 */
+		return;
+	}
+	/* Swap in new table. */
+	ttab = ckh->tab;
+	ckh->tab = tab;
+	tab = ttab;
+	ckh->lg_curbuckets = lg_curcells - LG_CKH_BUCKET_CELLS;
+
+	if (ckh_rebuild(ckh, tab) == false) {
+		idalloc(tab);
+#ifdef CKH_COUNT
+		ckh->nshrinks++;
+#endif
+		return;
+	}
+
+	/* Rebuilding failed, so back out partially rebuilt table. */
+	idalloc(ckh->tab);
+	ckh->tab = tab;
+	ckh->lg_curbuckets = lg_prevbuckets;
+#ifdef CKH_COUNT
+	ckh->nshrinkfails++;
+#endif
+}
+
+bool
+ckh_new(ckh_t *ckh, size_t minitems, ckh_hash_t *hash, ckh_keycomp_t *keycomp)
+{
+	bool ret;
+	size_t mincells, usize;
+	unsigned lg_mincells;
+
+	assert(minitems > 0);
+	assert(hash != NULL);
+	assert(keycomp != NULL);
+
+#ifdef CKH_COUNT
+	ckh->ngrows = 0;
+	ckh->nshrinks = 0;
+	ckh->nshrinkfails = 0;
+	ckh->ninserts = 0;
+	ckh->nrelocs = 0;
+#endif
+	ckh->prn_state = 42; /* Value doesn't really matter. */
+	ckh->count = 0;
+
+	/*
+	 * Find the minimum power of 2 that is large enough to fit aBaseCount
+	 * entries.  We are using (2+,2) cuckoo hashing, which has an expected
+	 * maximum load factor of at least ~0.86, so 0.75 is a conservative load
+	 * factor that will typically allow 2^aLgMinItems to fit without ever
+	 * growing the table.
+	 */
+	assert(LG_CKH_BUCKET_CELLS > 0);
+	mincells = ((minitems + (3 - (minitems % 3))) / 3) << 2;
+	for (lg_mincells = LG_CKH_BUCKET_CELLS;
+	    (ZU(1) << lg_mincells) < mincells;
+	    lg_mincells++)
+		; /* Do nothing. */
+	ckh->lg_minbuckets = lg_mincells - LG_CKH_BUCKET_CELLS;
+	ckh->lg_curbuckets = lg_mincells - LG_CKH_BUCKET_CELLS;
+	ckh->hash = hash;
+	ckh->keycomp = keycomp;
+
+	usize = sa2u(sizeof(ckhc_t) << lg_mincells, CACHELINE, NULL);
+	if (usize == 0) {
+		ret = true;
+		goto RETURN;
+	}
+	ckh->tab = (ckhc_t *)ipalloc(usize, CACHELINE, true);
+	if (ckh->tab == NULL) {
+		ret = true;
+		goto RETURN;
+	}
+
+#ifdef JEMALLOC_DEBUG
+	ckh->magic = CKH_MAGIC;
+#endif
+
+	ret = false;
+RETURN:
+	return (ret);
+}
+
+void
+ckh_delete(ckh_t *ckh)
+{
+
+	assert(ckh != NULL);
+	dassert(ckh->magic == CKH_MAGIC);
+
+#ifdef CKH_VERBOSE
+	malloc_printf(
+	    "%s(%p): ngrows: %"PRIu64", nshrinks: %"PRIu64","
+	    " nshrinkfails: %"PRIu64", ninserts: %"PRIu64","
+	    " nrelocs: %"PRIu64"\n", __func__, ckh,
+	    (unsigned long long)ckh->ngrows,
+	    (unsigned long long)ckh->nshrinks,
+	    (unsigned long long)ckh->nshrinkfails,
+	    (unsigned long long)ckh->ninserts,
+	    (unsigned long long)ckh->nrelocs);
+#endif
+
+	idalloc(ckh->tab);
+#ifdef JEMALLOC_DEBUG
+	memset(ckh, 0x5a, sizeof(ckh_t));
+#endif
+}
+
+size_t
+ckh_count(ckh_t *ckh)
+{
+
+	assert(ckh != NULL);
+	dassert(ckh->magic == CKH_MAGIC);
+
+	return (ckh->count);
+}
+
+bool
+ckh_iter(ckh_t *ckh, size_t *tabind, void **key, void **data)
+{
+	size_t i, ncells;
+
+	for (i = *tabind, ncells = (ZU(1) << (ckh->lg_curbuckets +
+	    LG_CKH_BUCKET_CELLS)); i < ncells; i++) {
+		if (ckh->tab[i].key != NULL) {
+			if (key != NULL)
+				*key = (void *)ckh->tab[i].key;
+			if (data != NULL)
+				*data = (void *)ckh->tab[i].data;
+			*tabind = i + 1;
+			return (false);
+		}
+	}
+
+	return (true);
+}
+
+bool
+ckh_insert(ckh_t *ckh, const void *key, const void *data)
+{
+	bool ret;
+
+	assert(ckh != NULL);
+	dassert(ckh->magic == CKH_MAGIC);
+	assert(ckh_search(ckh, key, NULL, NULL));
+
+#ifdef CKH_COUNT
+	ckh->ninserts++;
+#endif
+
+	while (ckh_try_insert(ckh, &key, &data)) {
+		if (ckh_grow(ckh)) {
+			ret = true;
+			goto RETURN;
+		}
+	}
+
+	ret = false;
+RETURN:
+	return (ret);
+}
+
+bool
+ckh_remove(ckh_t *ckh, const void *searchkey, void **key, void **data)
+{
+	size_t cell;
+
+	assert(ckh != NULL);
+	dassert(ckh->magic == CKH_MAGIC);
+
+	cell = ckh_isearch(ckh, searchkey);
+	if (cell != SIZE_T_MAX) {
+		if (key != NULL)
+			*key = (void *)ckh->tab[cell].key;
+		if (data != NULL)
+			*data = (void *)ckh->tab[cell].data;
+		ckh->tab[cell].key = NULL;
+		ckh->tab[cell].data = NULL; /* Not necessary. */
+
+		ckh->count--;
+		/* Try to halve the table if it is less than 1/4 full. */
+		if (ckh->count < (ZU(1) << (ckh->lg_curbuckets
+		    + LG_CKH_BUCKET_CELLS - 2)) && ckh->lg_curbuckets
+		    > ckh->lg_minbuckets) {
+			/* Ignore error due to OOM. */
+			ckh_shrink(ckh);
+		}
+
+		return (false);
+	}
+
+	return (true);
+}
+
+bool
+ckh_search(ckh_t *ckh, const void *searchkey, void **key, void **data)
+{
+	size_t cell;
+
+	assert(ckh != NULL);
+	dassert(ckh->magic == CKH_MAGIC);
+
+	cell = ckh_isearch(ckh, searchkey);
+	if (cell != SIZE_T_MAX) {
+		if (key != NULL)
+			*key = (void *)ckh->tab[cell].key;
+		if (data != NULL)
+			*data = (void *)ckh->tab[cell].data;
+		return (false);
+	}
+
+	return (true);
+}
+
+void
+ckh_string_hash(const void *key, unsigned minbits, size_t *hash1, size_t *hash2)
+{
+	size_t ret1, ret2;
+	uint64_t h;
+
+	assert(minbits <= 32 || (SIZEOF_PTR == 8 && minbits <= 64));
+	assert(hash1 != NULL);
+	assert(hash2 != NULL);
+
+	h = hash(key, strlen((const char *)key), 0x94122f335b332aeaLLU);
+	if (minbits <= 32) {
+		/*
+		 * Avoid doing multiple hashes, since a single hash provides
+		 * enough bits.
+		 */
+		ret1 = h & ZU(0xffffffffU);
+		ret2 = h >> 32;
+	} else {
+		ret1 = h;
+		ret2 = hash(key, strlen((const char *)key),
+		    0x8432a476666bbc13U);
+	}
+
+	*hash1 = ret1;
+	*hash2 = ret2;
+}
+
+bool
+ckh_string_keycomp(const void *k1, const void *k2)
+{
+
+    assert(k1 != NULL);
+    assert(k2 != NULL);
+
+    return (strcmp((char *)k1, (char *)k2) ? false : true);
+}
+
+void
+ckh_pointer_hash(const void *key, unsigned minbits, size_t *hash1,
+    size_t *hash2)
+{
+	size_t ret1, ret2;
+	uint64_t h;
+	union {
+		const void	*v;
+		uint64_t	i;
+	} u;
+
+	assert(minbits <= 32 || (SIZEOF_PTR == 8 && minbits <= 64));
+	assert(hash1 != NULL);
+	assert(hash2 != NULL);
+
+	assert(sizeof(u.v) == sizeof(u.i));
+#if (LG_SIZEOF_PTR != LG_SIZEOF_INT)
+	u.i = 0;
+#endif
+	u.v = key;
+	h = hash(&u.i, sizeof(u.i), 0xd983396e68886082LLU);
+	if (minbits <= 32) {
+		/*
+		 * Avoid doing multiple hashes, since a single hash provides
+		 * enough bits.
+		 */
+		ret1 = h & ZU(0xffffffffU);
+		ret2 = h >> 32;
+	} else {
+		assert(SIZEOF_PTR == 8);
+		ret1 = h;
+		ret2 = hash(&u.i, sizeof(u.i), 0x5e2be9aff8709a5dLLU);
+	}
+
+	*hash1 = ret1;
+	*hash2 = ret2;
+}
+
+bool
+ckh_pointer_keycomp(const void *k1, const void *k2)
+{
+
+	return ((k1 == k2) ? true : false);
+}

deps/jemalloc/src/extent.c

@@ -0,0 +1,41 @@
+#define	JEMALLOC_EXTENT_C_
+#include "jemalloc/internal/jemalloc_internal.h"
+
+/******************************************************************************/
+
+#if (defined(JEMALLOC_SWAP) || defined(JEMALLOC_DSS))
+static inline int
+extent_szad_comp(extent_node_t *a, extent_node_t *b)
+{
+	int ret;
+	size_t a_size = a->size;
+	size_t b_size = b->size;
+
+	ret = (a_size > b_size) - (a_size < b_size);
+	if (ret == 0) {
+		uintptr_t a_addr = (uintptr_t)a->addr;
+		uintptr_t b_addr = (uintptr_t)b->addr;
+
+		ret = (a_addr > b_addr) - (a_addr < b_addr);
+	}
+
+	return (ret);
+}
+
+/* Generate red-black tree functions. */
+rb_gen(, extent_tree_szad_, extent_tree_t, extent_node_t, link_szad,
+    extent_szad_comp)
+#endif
+
+static inline int
+extent_ad_comp(extent_node_t *a, extent_node_t *b)
+{
+	uintptr_t a_addr = (uintptr_t)a->addr;
+	uintptr_t b_addr = (uintptr_t)b->addr;
+
+	return ((a_addr > b_addr) - (a_addr < b_addr));
+}
+
+/* Generate red-black tree functions. */
+rb_gen(, extent_tree_ad_, extent_tree_t, extent_node_t, link_ad,
+    extent_ad_comp)

deps/jemalloc/src/hash.c

@@ -0,0 +1,2 @@
+#define	JEMALLOC_HASH_C_
+#include "jemalloc/internal/jemalloc_internal.h"

deps/jemalloc/src/huge.c

@@ -0,0 +1,379 @@
+#define	JEMALLOC_HUGE_C_
+#include "jemalloc/internal/jemalloc_internal.h"
+
+/******************************************************************************/
+/* Data. */
+
+#ifdef JEMALLOC_STATS
+uint64_t	huge_nmalloc;
+uint64_t	huge_ndalloc;
+size_t		huge_allocated;
+#endif
+
+malloc_mutex_t	huge_mtx;
+
+/******************************************************************************/
+
+/* Tree of chunks that are stand-alone huge allocations. */
+static extent_tree_t	huge;
+
+void *
+huge_malloc(size_t size, bool zero)
+{
+	void *ret;
+	size_t csize;
+	extent_node_t *node;
+
+	/* Allocate one or more contiguous chunks for this request. */
+
+	csize = CHUNK_CEILING(size);
+	if (csize == 0) {
+		/* size is large enough to cause size_t wrap-around. */
+		return (NULL);
+	}
+
+	/* Allocate an extent node with which to track the chunk. */
+	node = base_node_alloc();
+	if (node == NULL)
+		return (NULL);
+
+	ret = chunk_alloc(csize, false, &zero);
+	if (ret == NULL) {
+		base_node_dealloc(node);
+		return (NULL);
+	}
+
+	/* Insert node into huge. */
+	node->addr = ret;
+	node->size = csize;
+
+	malloc_mutex_lock(&huge_mtx);
+	extent_tree_ad_insert(&huge, node);
+#ifdef JEMALLOC_STATS
+	stats_cactive_add(csize);
+	huge_nmalloc++;
+	huge_allocated += csize;
+#endif
+	malloc_mutex_unlock(&huge_mtx);
+
+#ifdef JEMALLOC_FILL
+	if (zero == false) {
+		if (opt_junk)
+			memset(ret, 0xa5, csize);
+		else if (opt_zero)
+			memset(ret, 0, csize);
+	}
+#endif
+
+	return (ret);
+}
+
+/* Only handles large allocations that require more than chunk alignment. */
+void *
+huge_palloc(size_t size, size_t alignment, bool zero)
+{
+	void *ret;
+	size_t alloc_size, chunk_size, offset;
+	extent_node_t *node;
+
+	/*
+	 * This allocation requires alignment that is even larger than chunk
+	 * alignment.  This means that huge_malloc() isn't good enough.
+	 *
+	 * Allocate almost twice as many chunks as are demanded by the size or
+	 * alignment, in order to assure the alignment can be achieved, then
+	 * unmap leading and trailing chunks.
+	 */
+	assert(alignment > chunksize);
+
+	chunk_size = CHUNK_CEILING(size);
+
+	if (size >= alignment)
+		alloc_size = chunk_size + alignment - chunksize;
+	else
+		alloc_size = (alignment << 1) - chunksize;
+
+	/* Allocate an extent node with which to track the chunk. */
+	node = base_node_alloc();
+	if (node == NULL)
+		return (NULL);
+
+	ret = chunk_alloc(alloc_size, false, &zero);
+	if (ret == NULL) {
+		base_node_dealloc(node);
+		return (NULL);
+	}
+
+	offset = (uintptr_t)ret & (alignment - 1);
+	assert((offset & chunksize_mask) == 0);
+	assert(offset < alloc_size);
+	if (offset == 0) {
+		/* Trim trailing space. */
+		chunk_dealloc((void *)((uintptr_t)ret + chunk_size), alloc_size
+		    - chunk_size);
+	} else {
+		size_t trailsize;
+
+		/* Trim leading space. */
+		chunk_dealloc(ret, alignment - offset);
+
+		ret = (void *)((uintptr_t)ret + (alignment - offset));
+
+		trailsize = alloc_size - (alignment - offset) - chunk_size;
+		if (trailsize != 0) {
+		    /* Trim trailing space. */
+		    assert(trailsize < alloc_size);
+		    chunk_dealloc((void *)((uintptr_t)ret + chunk_size),
+			trailsize);
+		}
+	}
+
+	/* Insert node into huge. */
+	node->addr = ret;
+	node->size = chunk_size;
+
+	malloc_mutex_lock(&huge_mtx);
+	extent_tree_ad_insert(&huge, node);
+#ifdef JEMALLOC_STATS
+	stats_cactive_add(chunk_size);
+	huge_nmalloc++;
+	huge_allocated += chunk_size;
+#endif
+	malloc_mutex_unlock(&huge_mtx);
+
+#ifdef JEMALLOC_FILL
+	if (zero == false) {
+		if (opt_junk)
+			memset(ret, 0xa5, chunk_size);
+		else if (opt_zero)
+			memset(ret, 0, chunk_size);
+	}
+#endif
+
+	return (ret);
+}
+
+void *
+huge_ralloc_no_move(void *ptr, size_t oldsize, size_t size, size_t extra)
+{
+
+	/*
+	 * Avoid moving the allocation if the size class can be left the same.
+	 */
+	if (oldsize > arena_maxclass
+	    && CHUNK_CEILING(oldsize) >= CHUNK_CEILING(size)
+	    && CHUNK_CEILING(oldsize) <= CHUNK_CEILING(size+extra)) {
+		assert(CHUNK_CEILING(oldsize) == oldsize);
+#ifdef JEMALLOC_FILL
+		if (opt_junk && size < oldsize) {
+			memset((void *)((uintptr_t)ptr + size), 0x5a,
+			    oldsize - size);
+		}
+#endif
+		return (ptr);
+	}
+
+	/* Reallocation would require a move. */
+	return (NULL);
+}
+
+void *
+huge_ralloc(void *ptr, size_t oldsize, size_t size, size_t extra,
+    size_t alignment, bool zero)
+{
+	void *ret;
+	size_t copysize;
+
+	/* Try to avoid moving the allocation. */
+	ret = huge_ralloc_no_move(ptr, oldsize, size, extra);
+	if (ret != NULL)
+		return (ret);
+
+	/*
+	 * size and oldsize are different enough that we need to use a
+	 * different size class.  In that case, fall back to allocating new
+	 * space and copying.
+	 */
+	if (alignment > chunksize)
+		ret = huge_palloc(size + extra, alignment, zero);
+	else
+		ret = huge_malloc(size + extra, zero);
+
+	if (ret == NULL) {
+		if (extra == 0)
+			return (NULL);
+		/* Try again, this time without extra. */
+		if (alignment > chunksize)
+			ret = huge_palloc(size, alignment, zero);
+		else
+			ret = huge_malloc(size, zero);
+
+		if (ret == NULL)
+			return (NULL);
+	}
+
+	/*
+	 * Copy at most size bytes (not size+extra), since the caller has no
+	 * expectation that the extra bytes will be reliably preserved.
+	 */
+	copysize = (size < oldsize) ? size : oldsize;
+
+	/*
+	 * Use mremap(2) if this is a huge-->huge reallocation, and neither the
+	 * source nor the destination are in swap or dss.
+	 */
+#ifdef JEMALLOC_MREMAP_FIXED
+	if (oldsize >= chunksize
+#  ifdef JEMALLOC_SWAP
+	    && (swap_enabled == false || (chunk_in_swap(ptr) == false &&
+	    chunk_in_swap(ret) == false))
+#  endif
+#  ifdef JEMALLOC_DSS
+	    && chunk_in_dss(ptr) == false && chunk_in_dss(ret) == false
+#  endif
+	    ) {
+		size_t newsize = huge_salloc(ret);
+
+		if (mremap(ptr, oldsize, newsize, MREMAP_MAYMOVE|MREMAP_FIXED,
+		    ret) == MAP_FAILED) {
+			/*
+			 * Assuming no chunk management bugs in the allocator,
+			 * the only documented way an error can occur here is
+			 * if the application changed the map type for a
+			 * portion of the old allocation.  This is firmly in
+			 * undefined behavior territory, so write a diagnostic
+			 * message, and optionally abort.
+			 */
+			char buf[BUFERROR_BUF];
+
+			buferror(errno, buf, sizeof(buf));
+			malloc_write("<jemalloc>: Error in mremap(): ");
+			malloc_write(buf);
+			malloc_write("\n");
+			if (opt_abort)
+				abort();
+			memcpy(ret, ptr, copysize);
+			idalloc(ptr);
+		} else
+			huge_dalloc(ptr, false);
+	} else
+#endif
+	{
+		memcpy(ret, ptr, copysize);
+		idalloc(ptr);
+	}
+	return (ret);
+}
+
+void
+huge_dalloc(void *ptr, bool unmap)
+{
+	extent_node_t *node, key;
+
+	malloc_mutex_lock(&huge_mtx);
+
+	/* Extract from tree of huge allocations. */
+	key.addr = ptr;
+	node = extent_tree_ad_search(&huge, &key);
+	assert(node != NULL);
+	assert(node->addr == ptr);
+	extent_tree_ad_remove(&huge, node);
+
+#ifdef JEMALLOC_STATS
+	stats_cactive_sub(node->size);
+	huge_ndalloc++;
+	huge_allocated -= node->size;
+#endif
+
+	malloc_mutex_unlock(&huge_mtx);
+
+	if (unmap) {
+	/* Unmap chunk. */
+#ifdef JEMALLOC_FILL
+#if (defined(JEMALLOC_SWAP) || defined(JEMALLOC_DSS))
+		if (opt_junk)
+			memset(node->addr, 0x5a, node->size);
+#endif
+#endif
+		chunk_dealloc(node->addr, node->size);
+	}
+
+	base_node_dealloc(node);
+}
+
+size_t
+huge_salloc(const void *ptr)
+{
+	size_t ret;
+	extent_node_t *node, key;
+
+	malloc_mutex_lock(&huge_mtx);
+
+	/* Extract from tree of huge allocations. */
+	key.addr = __DECONST(void *, ptr);
+	node = extent_tree_ad_search(&huge, &key);
+	assert(node != NULL);
+
+	ret = node->size;
+
+	malloc_mutex_unlock(&huge_mtx);
+
+	return (ret);
+}
+
+#ifdef JEMALLOC_PROF
+prof_ctx_t *
+huge_prof_ctx_get(const void *ptr)
+{
+	prof_ctx_t *ret;
+	extent_node_t *node, key;
+
+	malloc_mutex_lock(&huge_mtx);
+
+	/* Extract from tree of huge allocations. */
+	key.addr = __DECONST(void *, ptr);
+	node = extent_tree_ad_search(&huge, &key);
+	assert(node != NULL);
+
+	ret = node->prof_ctx;
+
+	malloc_mutex_unlock(&huge_mtx);
+
+	return (ret);
+}
+
+void
+huge_prof_ctx_set(const void *ptr, prof_ctx_t *ctx)
+{
+	extent_node_t *node, key;
+
+	malloc_mutex_lock(&huge_mtx);
+
+	/* Extract from tree of huge allocations. */
+	key.addr = __DECONST(void *, ptr);
+	node = extent_tree_ad_search(&huge, &key);
+	assert(node != NULL);
+
+	node->prof_ctx = ctx;
+
+	malloc_mutex_unlock(&huge_mtx);
+}
+#endif
+
+bool
+huge_boot(void)
+{
+
+	/* Initialize chunks data. */
+	if (malloc_mutex_init(&huge_mtx))
+		return (true);
+	extent_tree_ad_new(&huge);
+
+#ifdef JEMALLOC_STATS
+	huge_nmalloc = 0;
+	huge_ndalloc = 0;
+	huge_allocated = 0;
+#endif
+
+	return (false);
+}

deps/jemalloc/src/mb.c

@@ -0,0 +1,2 @@
+#define	JEMALLOC_MB_C_
+#include "jemalloc/internal/jemalloc_internal.h"

deps/jemalloc/src/mutex.c

@@ -0,0 +1,90 @@
+#define	JEMALLOC_MUTEX_C_
+#include "jemalloc/internal/jemalloc_internal.h"
+
+/******************************************************************************/
+/* Data. */
+
+#ifdef JEMALLOC_LAZY_LOCK
+bool isthreaded = false;
+#endif
+
+#ifdef JEMALLOC_LAZY_LOCK
+static void	pthread_create_once(void);
+#endif
+
+/******************************************************************************/
+/*
+ * We intercept pthread_create() calls in order to toggle isthreaded if the
+ * process goes multi-threaded.
+ */
+
+#ifdef JEMALLOC_LAZY_LOCK
+static int (*pthread_create_fptr)(pthread_t *__restrict, const pthread_attr_t *,
+    void *(*)(void *), void *__restrict);
+
+static void
+pthread_create_once(void)
+{
+
+	pthread_create_fptr = dlsym(RTLD_NEXT, "pthread_create");
+	if (pthread_create_fptr == NULL) {
+		malloc_write("<jemalloc>: Error in dlsym(RTLD_NEXT, "
+		    "\"pthread_create\")\n");
+		abort();
+	}
+
+	isthreaded = true;
+}
+
+JEMALLOC_ATTR(visibility("default"))
+int
+pthread_create(pthread_t *__restrict thread,
+    const pthread_attr_t *__restrict attr, void *(*start_routine)(void *),
+    void *__restrict arg)
+{
+	static pthread_once_t once_control = PTHREAD_ONCE_INIT;
+
+	pthread_once(&once_control, pthread_create_once);
+
+	return (pthread_create_fptr(thread, attr, start_routine, arg));
+}
+#endif
+
+/******************************************************************************/
+
+bool
+malloc_mutex_init(malloc_mutex_t *mutex)
+{
+#ifdef JEMALLOC_OSSPIN
+	*mutex = 0;
+#else
+	pthread_mutexattr_t attr;
+
+	if (pthread_mutexattr_init(&attr) != 0)
+		return (true);
+#ifdef PTHREAD_MUTEX_ADAPTIVE_NP
+	pthread_mutexattr_settype(&attr, PTHREAD_MUTEX_ADAPTIVE_NP);
+#else
+	pthread_mutexattr_settype(&attr, PTHREAD_MUTEX_DEFAULT);
+#endif
+	if (pthread_mutex_init(mutex, &attr) != 0) {
+		pthread_mutexattr_destroy(&attr);
+		return (true);
+	}
+	pthread_mutexattr_destroy(&attr);
+
+#endif
+	return (false);
+}
+
+void
+malloc_mutex_destroy(malloc_mutex_t *mutex)
+{
+
+#ifndef JEMALLOC_OSSPIN
+	if (pthread_mutex_destroy(mutex) != 0) {
+		malloc_write("<jemalloc>: Error in pthread_mutex_destroy()\n");
+		abort();
+	}
+#endif
+}