diff options
| author | Satish Patel <satish.patel@linaro.org> | 2016-04-01 18:03:14 +0530 |
|---|---|---|
| committer | Satish Patel <satish.patel@linaro.org> | 2016-04-01 18:03:14 +0530 |
| commit | 042b63cd8d14ecf23f3ae832f35a0f1d0e2ed47f (patch) | |
| tree | bf26ffd29997656092b698deb0a6c4c8cc7e878d | |
| parent | 8303aab78a91ca5909bfb70fdf8ec5b0010891fc (diff) | |
| download | nedmalloc-042b63cd8d14ecf23f3ae832f35a0f1d0e2ed47f.tar.gz | |
Initial (original) sources of nedmalloc
- required source files are taken from
https://github.com/ned14/nedmalloc
Note: Android build support yet to be added
Signed-off-by: Satish Patel <satish.patel@linaro.org>
| -rw-r--r-- | Readme.html | 706 | ||||
| -rw-r--r-- | malloc.c.h | 6186 | ||||
| -rw-r--r-- | nedmalloc.c | 2275 | ||||
| -rw-r--r-- | nedmalloc.h | 1620 |
4 files changed, 10787 insertions, 0 deletions
diff --git a/Readme.html b/Readme.html new file mode 100644 index 0000000..03d42c0 --- /dev/null +++ b/Readme.html @@ -0,0 +1,706 @@ +<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN" "http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">
+<html xmlns="http://www.w3.org/1999/xhtml">
+
+<head>
+<meta content="text/html; charset=utf-8" http-equiv="Content-Type" />
+<title>nedalloc Readme</title>
+<style type="text/css">
+<!--
+body {
+ text-align: justify;
+}
+h1, h2, h3, h4, h5, h6 {
+ margin-bottom: -0.5em;
+}
+h1 {
+ text-align: center;
+}
+h2 {
+ text-decoration: underline;
+ margin-bottom: -0.25em;
+}
+p {
+ margin-top: 0.5em;
+ margin-bottom: 0.5em;
+}
+ul li, ol li {
+ margin-top: 0.2em;
+ margin-bottom: 0.2em;
+}
+dl {
+ margin-left: 2em;
+}
+dl dt {
+ font-weight: bold;
+}
+dt + dd {
+ margin-bottom: 1em;
+}
+.gitcommit {
+ font-family: "Courier New", Courier, monospace;
+ font-size: smaller;
+}
+-->
+</style>
+</head>
+
+<body>
+
+<div style="text-align: center">
+ <h1 style="text-decoration: underline">nedalloc v1.10 beta 4 (?)</h1>
+ <h2 style="text-decoration: none;">by Niall Douglas</h2>
+ <p>Web site: <a href="http://www.nedprod.com/programs/portable/nedmalloc/">http://www.nedprod.com/programs/portable/nedmalloc/</a></p>
+ <p>Trunk build status: <a href="https://travis-ci.org/ned14/nedmalloc"><img style="vertical-align:middle;border:none" src="https://travis-ci.org/ned14/nedmalloc.png?branch=master"/></a></p>
+ <hr /></div>
+<p>Enclosed is nedalloc, an alternative malloc implementation for multiple threads
+without lock contention based on <a href="http://g.oswego.edu/" target="_blank">
+dlmalloc</a> v2.8.4 and a specialised user mode page allocator (Windows Vista or
+later only). It has the following features:</p>
+<ol>
+ <li>A per-thread small block cache for maximum CPU scalability.</li>
+ <li>A per-thread arena to minimise lock contention.</li>
+ <li>The ability to patch Windows binaries to replace the C memory allocation
+ API malloc, realloc(), free() et al such that by simply inserting nedmalloc.dll
+ into a process one realises performance improvements without recompilation.</li>
+ <li>On POSIX, it knows how to talk to valgrind so you can track memory
+ corruption and/or memory leaks.</li>
+ <li>A unique user mode page allocator implementation which delivers O(1) scaling
+ for blocks of any size, including an O(1) very fast realloc(). Improves medium
+ sized block (~1Mb) allocation speeds by about 25 times on current hardware.
+ Requires Windows Vista or later only, and requires Administrator privileges
+ as well as either UAC disabled or a UAC prompt at the start of each program
+ run.</li>
+ <li>A malloc v2 API which enables considerable improvements in efficiency by
+ allowing client code to better inform the allocator on what (not) to do.</li>
+ <li>An enhanced C++ STL allocator implementation to enable super-fast std::vector<>
+ <strong>[unfinished]</strong></li>
+</ol>
+<p>It is licensed under the
+<a href="http://www.boost.org/LICENSE_1_0.txt" target="_blank">Boost Software License</a>
+which basically means you can do anything you like with it. This does not apply
+to the malloc.c.h file which remains copyright to others. Commercial support is
+available from <a href="http://www.nedproductions.biz/" target="_blank">ned Productions
+Limited</a>.</p>
+<p>It has been tested on win32 (x86), win64 (x64), Linux (x64), FreeBSD (x64) and
+Apple Mac OS X (x86). It works very well on all of these and is very significantly
+faster than the system allocator on Windows XP and FreeBSD <v7. If you are using
+>= 10.6 Apple Mac OS X or you are on Windows 7 or later then you probably won't
+see much improvement without modifying your source to use the v2 malloc API (and
+kudos to Apple and Microsoft for adopting excellent allocators).</p>
+<p>The user mode page allocator returns jaw dropping real world performance improvements
+but requires running the process as the superuser. Without, it still offers sizeable
+gains on all older operating systems and through the v2 malloc API modest gains
+on all very recent operating systems, especially in these situations:</p>
+<ol>
+ <li>If you are repeatedly extending large vector arrays, you will see a LARGE
+ improvement if you use the address space reservation features.</li>
+ <li>If you do a lot of work with 16 byte aligned vectors e.g. SSE or AVX vector
+ arrays, you will find the v2 malloc API a godsend.</li>
+</ol>
+<p style="text-decoration: underline"><strong>Table of Contents: </strong></p>
+<ol style="list-style-type: upper-alpha; position: relative; margin-top: -0.5em;">
+ <li><a href="#touse">How to use</a><ul style="list-style-type: none; margin-left: 0; padding-left: 0">
+ <li>A1. <a href="#CPPAPI">The C++ API</a></li>
+ <li>A2. <a href="#v2mallocAPI">The v2 malloc C API</a></li>
+ </ul>
+ </li>
+ <li><a href="#notes">Notes</a><ul style="list-style-type: none; margin-left: 0; padding-left: 0">
+ <li>B1. <a href="#memorybloat">Memory Bloating</a></li>
+ <li>B2. <a href="#memoryleaks">Memory Leakage</a></li>
+ <li>B3. <a href="#threadcache">The Threadcache</a></li>
+ <li>B4. <a href="#largepages">Large Page support</a></li>
+ <li>B5. <a href="#logger">Memory operation logging</a></li>
+ <li>B6. <a href="#windowsonly">Windows-only features</a></li>
+ </ul>
+ </li>
+ <li><a href="#speedcomparisons">Speed Comparisons</a></li>
+ <li><a href="#troubleshooting">Troubleshooting</a></li>
+ <li><a href="#changelog">Changelog</a></li>
+</ol>
+<h2><a name="touse">A. To use:</a></h2>
+<p>The quickest way is to drop nedmalloc.h, nedmalloc.c and malloc.c.h into your
+project. Call nedmalloc(), nedcalloc(), nedrealloc() and nedfree() instead of your
+normal allocator, or nedpmalloc(), nedpcalloc(), nedprealloc() and nedpfree() if
+you want to segment your memory usage into pools. Make sure that you call neddisablethreadcache()
+for every pool you use on thread exit, and don't forget neddisablethreadcache(0)
+for the system pool if necessary. Run and enjoy!</p>
+<p>To test, compile <a href="test.c">test.c</a> (C) and <a href="test.cpp">test.cpp</a>
+(C++). Both will run a comparison between your system allocator and nedalloc and
+tell you how much faster nedalloc is. They also serve as examples of usage.</p>
+<p>If you'd like nedalloc as a Windows DLL or POSIX ELF shared object, the easiest
+thing to do is to use <a href="http://www.scons.org/" target="_blank">scons</a>
+which comes with a myriad of build options listed using scons -h. <b>If you want
+to build some MSVC project files for use with Microsoft Visual Studio</b> then what
+you do is (i) install <a href="http://www.python.org/" target="_blank">python</a>
+(ii) install <a href="http://www.scons.org/" target="_blank">scons</a> (iii) open
+a Visual Studio Command Box for the Visual Studio you wish to use via Start Menu
+=> Programs => Microsoft Visual Studio XXXX => Visual Studio Tools => Visual Studio
+XXXX Command Prompt (iv) change directory to the nedmalloc directory (e.g. by dragging
+in its folder) (v) type "!MakeMSVCProjs" and hit Return. Note that for Visual Studio
+2008 and later support you need scons v2.1 or later.</p>
+<p>nedalloc comes with two new memory allocator APIs: one is for C++, and the other
+is for C. <strong>Full documentation</strong> for all nedalloc's APIs and features
+is provided in the enclosed <a href="nedalloc.chm">nedalloc.chm</a> which is in
+Microsoft HTML Help format (Linux and Apple Mac OS X will happily read this format
+too). If you don't want to use the CHM documentation, <a href="nedmalloc.h">nedmalloc.h</a>
+is extensively commented with <a href="http://www.doxygen.org/" target="_blank">
+doxygen markup</a>.</p>
+<h3><a name="CPPAPI">A1: The C++ API:</a></h3>
+<p>For the v1.10 release which was generously sponsored by
+<a href="http://www.ara.com/" target="_blank">Applied Research Associates (USA)</a>,
+a C++ metaprogrammed STL allocator was designed which makes use of advanced nedalloc
+features to remedy many of the long standing problems and inefficiencies caused
+by C++'s traditional over-fondness for copying things. While its implementation
+is complex, usage is extremely easy - simply supply nedallocator<> as the custom
+allocator to STL container classes.</p>
+<p>As nedmalloc can do even better for vector extension, nedmalloc.h also contains
+a nedvector<> implementation which is the standard STL vector<> implementation except
+that it makes use of the non-relocating facilities of realloc2() (see below). This
+allows nedvector<> to not need to overallocate memory (most STL vector<> implementations
+will overallocate by 50%) which saves a lot of memory as well as <strong>completely
+avoiding array copy construction</strong> which make std::vector<>::resize() so
+very, very slow.</p>
+<p>Even without nedalloc's major speed improvements as a simple C style allocator,
+the improvements to the C++ memory infrastructure alone can generate huge performance
+gains.</p>
+<h3><a name="v2mallocAPI">A2: The v2 malloc C API:</a></h3>
+<p><strong>[Note: This API will be completely replaced in v1.2]</strong></p>
+<p>For the v1.10 release which was generously sponsored by
+<a href="http://www.ara.com/" target="_blank">Applied Research Associates (USA)</a>,
+a new general purpose allocator API was designed which is intended to remedy many
+of the long standing problems and inefficiencies introduced by the ISO C allocator
+API. Internally nedalloc's implementations of nedmalloc(), nedcalloc(), nedmemalign()
+and nedrealloc() all call into this API:</p>
+<ul>
+ <li><code>void* malloc2(size_t bytes, size_t alignment, unsigned flags)</code></li>
+ <li><code>void* realloc2(void* mem, size_t bytes, size_t alignment, unsigned
+ flags)</code></li>
+</ul>
+<p>If nedmalloc.h is being included by C++ code, the alignment and flags parameters
+default to zero which makes the new API identical to the old API (roll on the introduction
+of default parameters to C!). The ability for realloc2() to take an alignment is
+<em>particularly</em> useful for extending aligned vector arrays such as SSE/AVX
+vector arrays. Hitherto SSE/AVX vector code had to jump through all sorts of unpleasant
+hoops to maintain alignment during array extension :(.</p>
+<p>The flags supported include the ability to zero memory, to prevent realloc2()
+from moving a memory block, to force mmap() to be used from the beginning (useful
+when you know an array will be repeatedly extended) and to cause malloc2() to reserve
+additional address space after the allocation such that a realloc2() up to that
+reserved space will be very quick. On 32 bit Windows and Linux this reservation
+costs no address space in your process, so using it will NOT cause premature address
+space exhaustion.</p>
+<p>You should note that realloc()'s thunk to realloc2() defaults the flags to M2_RESERVE_MULT(8)
+i.e. if realloc() needs to allocate a block larger than mmap_threshold, it will
+also reserve eight times the address space of that allocation in order to make future
+realloc()'s up to that point much faster. This catches the vast majority of situations
+where large arrays are repeatedly extended.</p>
+<h2><a name="notes">B. Notes:</a></h2>
+<p>If you want the very latest version of this allocator, get it from the TnFOX
+GIT repository at either of (both are identical mirrors):</p>
+<ul>
+ <li>
+ <a href="git://nedmalloc.git.sourceforge.net/gitroot/nedmalloc/nedmalloc">git://nedmalloc.git.sourceforge.net/gitroot/nedmalloc/nedmalloc</a></li>
+ <li><a href="git://github.com/ned14/nedmalloc.git">git://github.com/ned14/nedmalloc.git</a></li>
+</ul>
+<p>IF YOU THINK YOU HAVE FOUND A BUG, PLEASE CHECK ONE OF THESE REPOS FIRST BEFORE
+REPORTING IT!</p>
+<h3><a name="memorybloat">B1: Memory Bloating</a></h3>
+<p>Because of how nedalloc allocates an mspace per thread, it <em>can</em> cause
+severe bloating of memory usage under certain allocation patterns. You can substantially
+reduce this wastage by setting DEFAULTMAXTHREADSINPOOL or the threads parameter
+to nedcreatepool() to a fraction of the number of threads which would normally be
+in a pool at once. This will reduce bloating at the cost of an increase in lock
+contention, with DEFAULTMAXTHREADSINPOOL=1 removing almost all bloating. If the
+block sizes typically allocated are less than THREADCACHEMAX, locking is avoided
+90-99% of the time and if most of your allocations are below this value, you can
+safely set DEFAULTMAXTHREADSINPOOL or even MAXTHREADSINPOOL to one.</p>
+<p>If you have LOTS of threads you may find that the threadcache held per thread
+is causing memory bloating. You can call nedtrimthreadcache() to trim the cache
+in a thread when you know that it won't be doing memory allocation (e.g. just before
+going to sleep), or alternatively you can set THREADCACHEMAXFREESPACE to something
+smaller than its default of 1Mb.</p>
+<p>Lastly, some people find that memory is not returned to the system when they
+think it ought to be. dlmalloc only returns free memory to the system when there
+is DEFAULT_TRIM_THRESHOLD (default=2Mb) free in a mspace, and it only checks how
+much there is free outside the topmost segment every MAX_RELEASE_CHECK_RATE free()'s.
+In other words, if your program very rapidly deallocates an awful lot of memory
+and then does not call free() for some time thereafter, dlmalloc will not release
+memory to the system. Generally in any real world code scenario free() will be called
+fairly frequently, and if not then you can always force release using nedmalloc_trim().</p>
+<h3><a name="memoryleaks">B2: Memory Leakage</a></h3>
+<p>You will suffer memory leakage unless you call neddisablethreadcache() per pool
+for every thread which exits (unless you are using nedalloc from its DLL on Windows).
+This is because nedalloc cannot portably know when a thread exits and thus when
+its thread cache can be returned for use by other code. Don't forget pool zero,
+the system pool. On some POSIX threads implementations there exists a pthread_atexit()
+which registers a termination handler for thread exit - if you don't have one of
+these then you'll have to do it manually.</p>
+<p>Equally if you use nedalloc from a dynamically loaded DLL or shared object which
+you later kick out of memory, you will leak memory if you don't disable all thread
+caches for all pools (as per the preceding paragraph), destroy all thread pools
+using neddestroypool() and destroy the system pool using neddestroysyspool().</p>
+<h3><a name="threadcache">B3: The Threadcache</a></h3>
+<p>For C++ type allocation patterns (where the same small sizes of memory are regularly
+allocated and deallocated as objects are created and destroyed), the threadcache
+always benefits performance as it will cache all malloc/free allocations under THREADCACHEMAX
+in size. If however your allocation patterns are different, searching the threadcache
+may significantly slow down your code - as a rule of thumb, if cache utilisation
+is below 80% (see the source for neddisablethreadcache() for how to enable debug
+printing in release mode) then you should disable the thread cache for that thread.
+You can compile out the threadcache code by setting THREADCACHEMAX to zero.</p>
+<h3><a name="largepages">B4: Large Page support</a></h3>
+<p>For some applications defining ENABLE_LARGE_PAGES can give a 10-15% performance
+increase by having nedalloc allocate using large pages only (which are 2Mb on x86/x64).
+Large pages take much less space in the TLB cache and can greatly benefit programs
+with a large working set, particularly on 64 bit systems.</p>
+<p>Support for large pages is limited to Linux and Windows. On Linux one must employ
+the libhugetlbfs library anyway as this is the "official" form of large page support,
+and setting it up and configuring it involves mounting a special hugetlbfs filing
+system. dlmalloc does not require a dependency on the libhugetlbfs headers, rather
+it searches for the library in the current process and if not found it silently
+disables support.</p>
+<p>On Windows, large page support is only implemented on Windows Server 2003/Vista
+or later and they are only permitted to be allocated by users holding the "Lock
+pages in memory" local security setting which is DISABLED by default. Furthermore,
+the process using nedalloc must hold the SeLockMemoryPrivilege privilege. If you
+are using the DLL then the DLL attempts to enable the SeLockMemoryPrivilege during
+initialisation - therefore if you are not using the DLL you will have to do this
+manually yourself. As with Linux support, if at any stage large pages cannot be
+allocated, then dlmalloc silently disables support - this allows one binary to function
+correctly in any environment. <strong>Note that on Windows</strong> if your process
+allocates a lot of memory at once when the machine has been running for an extended
+period, then the whole computer may hang for several seconds as the Windows kernel
+copies memory around in order to coalesce a large page. This is a problem with the
+Windows kernel and its VM design, not nedmalloc! If you would like to see how large
+pages ought to be implemented, research how FreeBSD implemented them.</p>
+<h3><a name="logger">B5: Memory operation logging</a></h3>
+<p>It is often very useful to have a log of the memory operations which an application
+performs - you would be amazed at the inefficiencies in memory usage that this can
+reveal. nedalloc contains a very fast memory operation logger which keeps a per-thread
+log of selected operations, including an optional stack backtrace. On pool destruction,
+or nedflushlogs(), nedalloc will write out the log as a Comma Separated Value format
+file which can be loaded into applications such as Excel for analysis.</p>
+<p>To use, define ENABLE_LOGGING to the bitmask of enum LogEntryType items in which
+you are interested, so 0xffffffff would log absolutely everything. The macro NEDMALLOC_TESTLOGENTRY,
+whose default is (ENABLE_LOGGING & logentrytype), is then used to determine which
+items should be logged. You can also enable stack backtracing on MSVC and GCC using
+NEDMALLOC_STACKBACKTRACEDEPTH.</p>
+<h3><a name="windowsonly">B6: Windows-only features</a></h3>
+<p>If you are running on Windows, there are quite a few extra options available
+thanks to work generously sponsored by
+<a href="http://www.ara.com/" target="_blank">Applied Research Associates (USA)</a>:</p>
+<dl>
+ <dt>Automatic threadcache cleanup and log output</dt>
+ <dd>If you build nedalloc as a DLL and link that into your application, then
+ the DLL can trap thread exits in your application and call neddisablethreadcache()
+ on all currently existing nedpool's for you. On process exit, the DLL will also
+ call nedflushlogs() for you on all still extant nedpool's.</dd>
+ <dt>Replacing the system allocator in the whole process</dt>
+ <dd>
+ <p>If you define REPLACE_SYSTEM_ALLOCATOR when building the DLL then the DLL
+ will replace <em>most</em> usage of the MSVCRT allocator (release MSVCRT,<strong>
+ not</strong> debug MSVCRTD)<strong> </strong>within any process it is loaded
+ into with nedalloc's routines instead, whilst remaining able to handle the odd
+ free() of a MSVCRT allocated block allocated during CRT init. This very conveniently
+ allows you to simply link with the nedalloc DLL and your application magically
+ now uses it with no code changes required, and because the MSVC implementation
+ of operators new and delete both call malloc() and free() it also covers all
+ C++ code. The following code is suggested:</p>
+ <code>#pragma comment(lib, "nedmalloc.lib")</code>
+ <p>This asks the linker to link against nedmalloc.lib during linking - without
+ this pragma the linker will generally leave out nedmalloc as there are no explicitly
+ imported routines that it understands. This auto-patching feature can also be
+ combined with
+ <a href="http://research.microsoft.com/en-us/projects/detours/" target="_blank">
+ Microsoft's Detours</a> to run any arbitrary application using nedalloc instead
+ of the system allocator:</p>
+ <code>withdll /d:nedmalloc.dll program.exe</code>
+ <p>For those not able to use Microsoft Detours, there is an enclosed unsupported/nedmalloc_loader
+ program which does one variant of the same thing. It may or may not be useful
+ to you - it is not intended to be maintained, and it probably doesn't work on
+ newer systems.</p>
+ <p>The reason that only the release MSVCRT not the debug MSVCRTD is patched
+ is twofold: (i) usually one <em>wants</em> the debug heap in debug builds so
+ it does memory corruption checking and reports memory leaks and (ii) the MSVC
+ CRT actually implements operator new and malloc using a completely different
+ implementation based on the Windows kernel HeapAlloc() function and it does
+ a lot of hoop jumping to handle mismatching CRT versions and lots of other stuff.
+ You can enable patching of the debug memory allocation functions in winpatcher.c
+ by uncommenting the relevant lines.</p>
+ </dd>
+ <dt>User mode page allocation</dt>
+ <dd>The user mode page allocator is a user space implementation of kernel memory
+ page allocation made possible by misusing the Address Windowing Extensions (AWE)
+ provided by newer versions of Microsoft Windows. AWE allows - with a bit of
+ persuasion - direct control of the Memory Management Unit of the CPU, thus allowing
+ memory pages to be arbitrarily remapped from one address to another. The user
+ mode page allocator can therefore allocate memory in microseconds by simply
+ mapping it into where it needs to be, or it can realloc() gigabytes of memory
+ from its old location into a new bigger space in microseconds. This O(1) scaling
+ gives processes running on the user mode page allocator an <strong>unholy</strong>
+ speed increase which gets exponentially better the larger the data set.<br />
+ <br />
+ Want to know more in lots of detail? Here are two academic papers on the topic:
+ <ol>
+ <li>Douglas, N, (2011-May), '<a href="http://arxiv.org/abs/1105.1815">User Mode Memory Page Management: An old idea applied anew to the memory wall problem</a>', ArXiv e-prints, vol: 1105.1815.</li>
+ <li>Douglas, N, (2011-May), '<a href="http://arxiv.org/abs/1105.1811">User Mode Memory Page Allocation: A Silver Bullet For Memory Allocation?</a>', ArXiv e-prints, vol: 1105.1811.</li>
+ </ol>
+ </dd>
+</dl>
+<h2><a name="speedcomparisons">C. Speed comparisons:</a></h2>
+<p>See Benchmarks.xls for details.</p>
+<p>The enclosed test.c can do one of two things: it can be a torture test which
+mostly hammers realloc() or it can be a pure speed test which sticks to simple malloc()
+and free(). If you enable C++ mode, half of the allocation sizes will be a two power
+multiple less than 512 bytes (to mimic C++ stack instantiated objects) which are
+extremely common in C++ code.</p>
+<p>The torture test is designed to mercilessly work realloc() which is the most
+complex and complete code path in any memory allocator. Most allocators have
+<strong>very</strong> poor realloc() performance - not so nedalloc which makes use
+of mremap() support on Linux and Windows. Even without mremap() support nedalloc's
+realloc() tends to be significantly faster than any standard allocator.</p>
+<p>The speed test is designed to be a representative synthetic memory allocator
+test where most allocations follow a stack pattern. It works by randomly mixing
+allocations with frees with sizes being a random value less than 16Kb. </p>
+<p>The C++ test.cpp simply benchmarks how much difference nedalloc::nedallocatorise<>
+makes to std::vector<> performance, particularly the performance of push_back(),
+pop_back() and vector assignment all of which are very common in real world code.
+As you will see, the STL - even with C++0x move constructor support - does not perform
+anywhere close to nedalloc's version which achieves its gains by simply avoiding
+copy and move construction completely.</p>
+<p>The real world code results are from Tn's TestIO benchmark. This is a heavily
+multithreaded and memory intensive benchmark with a lot of branching and other stuff
+modern processors don't like so much. As you'll note, the test doesn't show the
+benefits of the threadcache mostly due to the saturation of the memory bus being
+the limiting factor.</p>
+<h2><a name="troubleshooting">D. Troubleshooting:</a></h2>
+<p>I get a quite a few bug reports about code not working properly under nedalloc.
+I do not wish to sound presumptuous, however in an overwhelming majority of cases
+the problem is in your application code and not nedalloc (see below for all the
+bugs reported and fixed since 2006). Some of the largest corporations and IT deployments
+in the world use nedalloc pre-v1.10, and pre-v1.10 has been very heavily stress
+tested on everything from 32 processor SMP clusters right through to root DNS servers,
+ATM machine networks and embedded operating systems requiring a very high uptime.
+The v1.10 release adds a LOT of new code and features, and hence there are quite
+likely a lot of new bugs in the new code.</p>
+<p>In particular, just because it just happens to appear to work under the system
+allocator does not mean that your application is not riddled with memory corruption
+and non-ANSI usage of the API! And usually<strong> this is not your code's fault,
+but rather it is usually the third party libraries being used which sadly often
+include system libraries</strong>.</p>
+<p>Even though debugging an application for memory errors is a true black art made
+possible only with a great deal of patience, intuition and skill, here is a checklist
+for things to do before reporting a bug in nedalloc:</p>
+<ol>
+ <li>Make SURE you try nedalloc from GIT HEAD. For around six months of 2007
+ I kept getting the same report of a bug long fixed in GIT HEAD.</li>
+ <li>Make SURE you try nedalloc v1.06. If it works in v1.06 but isn't working
+ in nedalloc >= v1.10, then it's probably a bug in the new code (please report
+ it to me!)</li>
+ <li>Make use of nedalloc's internal debug routines. Try turning on full sanity
+ checks by #define FULLSANITYCHECKS 1. Also make use of all the assertion checking
+ performed when DEBUG is defined as 1. A lot of bug reports are made before running
+ under a debug build where an assertion trip clearly showed the problem. Lastly,
+ try changing the thread cache by #defining THREADCACHEMAX - this fundamentally
+ changes how the memory allocator behaves: if everything is fine with the thread
+ cache fully on or fully off, then this strongly suggests the source of your
+ problem.</li>
+ <li>Make SURE you are matching allocations and frees belonging to nedalloc if
+ you are not defining REPLACE_SYSTEM_ALLOCATOR. Attempting to free a block not
+ allocated by nedalloc will end badly, similarly passing one of nedalloc's blocks
+ to another allocator will likely also end badly. I have inserted as many assertion
+ and debug checks for this possibility as I can think of (further suggestions
+ are welcome), but no system can ever be watertight. If you're using C++, make
+ use of the C++ nedallocatorise API provided or else use some form of strong
+ template type system to have the compiler guarantee membership of a memory pointer
+ - see <a href="http://www.boost.org/" target="_blank">the Boost libraries</a>,
+ or indeed <a href="http://www.nedprod.com/TnFOX/" target="_blank">my own TnFOX
+ portability toolkit</a>.</li>
+ <li>If you're still having problems, or more likely your code runs absolutely
+ fine under debug builds but trips up under release which suggests a timing bug,
+ it is time to deploy heavyweight tools. Under Linux, you should use
+ <a href="http://valgrind.org/" target="_blank">valgrind</a>. Under Windows,
+ there is an excellent commercial tool called
+ <a href="http://www.glowcode.com/" target="_blank">Glowcode</a>. Any programming
+ team serious on quality should ALWAYS run their projects through these tools
+ before each and every release anyway - you would be amazed at what you miss
+ during all other testing.</li>
+ <li>Lastly, in the worst case scenario, consider hiring in a memory debugging
+ expert. There are quite a few on the market and they often are authors of memory
+ allocators. <a href="http://www.malloc.de/en/" target="_blank">Wolfram Gloger
+ (the author of ptmalloc) provides consulting services</a>.
+ <a href="http://www.nedproductions.biz/" target="_blank">My own consulting company
+ ned Productions Limited</a> may be able to provide such a service depending
+ on our current workload.</li>
+</ol>
+<p>I hope that these tips help. And I urge anyone considering simply dropping back
+to the system allocator as a quick fix to reconsider: squashing memory bugs often
+brings with it <strong>significant</strong> extra benefits in performance and reliability.
+It may cost what appears to be a lot extra now, but it usually will save itself
+many times its cost over the next few years. I know of one large multinational corporation
+who <strong>saved hundreds of millions of dollars</strong> due to the debugging
+of their system software performed when trying to get it working with nedalloc -
+they found one bug in nedalloc but over a hundred in their own code, and in the
+process improved performance <strong>threefold</strong> which saved an expensive
+hardware upgrade and deployment. The conclusion can only be that fixing memory bugs
+now tends to be worth it in the long run.</p>
+<h2><a name="changelog">E. ChangeLog:</a></h2>
+<h3>v1.10 beta 4 ?:</h3>
+<ul>
+ <li><span class="gitcommit">[master 726d9c7]</span> Fixed memory corruption
+ introduced when creating more than two nedpool's (issue #7). Thanks to mxmauro
+ for reporting this.</li>
+ <li><span class="gitcommit">[master c191ea9]</span> Merged dlmalloc v2.8.6.</li>
+ <li><span class="gitcommit">[master 06f1c70]</span> Added support for clang,
+ plus fixed up some compile errors in C++11.</li>
+ <li><span class="gitcommit">[master 8f8256c]</span> Added support for
+ valgrind instrumentation so valgrind can track programs using nedmalloc.</li>
+ <li><span class="gitcommit">[master 69825ca]</span> Fixed issue #8 where
+ memory allocated via the independent_*() functions was being incorrectly
+ identified as system allocated. Thanks to Geri for reporting this.</li>
+ <li><span class="gitcommit">[master a6a0dec]</span> Fixed issue #10 where
+ a failure to allocate memory on POSIX was not being trapped correctly. Thanks
+ to btaudul for reporting this.</li>
+ <li><span class="gitcommit">[master a559f9e]</span> Fixed issue #12 where
+ RTLD_DEFAULT was undefined. Replaced this code entirely with new code which
+ parses /proc/meminfo for huge page size. Thanks to Geri for reporting this.</li>
+ <li><span class="gitcommit">[master 9119158]</span> Added Travis CI build bot
+ support to nedalloc, testing gcc, clang and clang static analyser.</li>
+ <li><span class="gitcommit">[master xxxxxxx]</span> Fixed issue #14 where
+ nedalloc was using is_pod<> instead of is_trivially_copyable<>.
+ Thanks to JustSid for reporting this.</li>
+</ul>
+<h3>v1.10 beta 3 17th July 2012:</h3>
+<ul>
+ <li><span class="gitcommit">[master 5f26c1a]</span> Due to a bug introduced
+ in sha 7a9dd5c (17th April 2010), nedmalloc has never allocated more than a
+ single mspace when using the system pool. This effectively had disabled
+ concurrency for any allocation > THREADCACHEMAX (8Kb) which no doubt made
+ nedmalloc v1.10 betas 1 and 2 appear no faster than system allocators. My
+ thanks to the eagle eyes of Gavin Lambert for spotting this.</li>
+</ul>
+<h3>v1.10 beta 2 10th July 2012:</h3>
+<ul>
+ <li><span class="gitcommit">[master 51ab2a2]</span> scons now tests for C++0x
+ support before turning it on and tries multiple libraries for clock_gettime()
+ rather than assuming it lives in librt. This ought to fix miscompilation on
+ Mac OS X. Thanks to Robert D. Blanchet Jr. for reporting this.</li>
+ <li><span class="gitcommit">[master b2c3517]</span> Mac defines malloc_size
+ to be const void *ptr, not void *ptr</li>
+ <li><span class="gitcommit">[master 9333e50]</span> Updated to use the new
+ O(1) Cfind(rounds=1) feature in nedtries</li>
+ <li><span class="gitcommit">[master 54c7e44]</span> Avoid overflowing allocation
+ size. Thanks to Xi Wang for supplying a patch fixing this.</li>
+ <li><span class="gitcommit">[master 5b614a0]</span> Removed __try1 and __finally1
+ from MinGW support as x64 target no longer supports SEH. Thanks to Geri for
+ reporting this.</li>
+ <li><span class="gitcommit">[master 48f1aa9]</span> Tidied up bitrot which
+ had broken compilation due to mismatched #if...#endif.</li>
+</ul>
+<h3>v1.10 beta 1 19th May 2011:</h3>
+<ul>
+ <li><span class="gitcommit">[master 89f1806]</span> Moved from SVN to GIT. Bumped
+ version to v1.10 as new ARA contract will involve significant further improvements
+ mainly centering around realloc() performance.</li>
+ <li><span class="gitcommit">[master 254fe7c]</span> Added nedmemsize() for API
+ compatibility with other allocators. Added DEFAULTMAXTHREADSINPOOL and set it
+ to FOUR which is a BREAKING CHANGE from previous versions of nedalloc (which
+ set it to 16).</li>
+ <li><span class="gitcommit">[nedmalloc_fast_realloc 97d1420]</span> Added win32mremap()
+ implementation.</li>
+ <li><span class="gitcommit">[nedmalloc_fast_realloc 8a1001e]</span> Significantly
+ improved test.c with new test options TESTCPLUSPLUS, BLOCKSIZE, TESTTYPE and
+ MAXMEMORY.</li>
+ <li><span class="gitcommit">[nedmalloc_fast_realloc 7ea606d]</span> Implemented
+ two variants of direct mremap() on Windows, one using file mappings and the
+ other using over-reservation. The former is used on 32 bit and the latter on
+ 64 bit.</li>
+ <li><span class="gitcommit">[nedmalloc_fast_realloc 26ff9a7]</span> Added the
+ malloc2() interface to nedalloc.</li>
+ <li><span class="gitcommit">[nedmalloc_fast_realloc 5bc5d97]</span> Rewrote
+ Readme.txt to become Readme.html which makes it much clearer to read.</li>
+ <li><span class="gitcommit">[nedmalloc_fast_realloc 2efa595]</span> Added doxygen
+ markup to nedmalloc.h and a first go at a policy driven STL allocator class.</li>
+ <li><span class="gitcommit">[nedmalloc_fast_realloc d851bde]</span> Added a
+ CHM documenting the nedalloc API.</li>
+ <li><span class="gitcommit">[nedmalloc_fast_realloc dbd3991]</span> Added a
+ fast malloc operations logger which outputs a CSV log on process exit.</li>
+ <li><span class="gitcommit">[nedmalloc_fast_realloc d6a8585]</span> Added stack
+ backtracing to the logger.</li>
+ <li><span class="gitcommit">[master c7ea06d]</span> Finished user mode page
+ allocator, so merged nedmalloc_fast_realloc branch.</li>
+ <li><span class="gitcommit">[master 9a8800f]</span> Fixed small bug which was
+ preventing the windows patcher from correctly finding the proper MSVCRT.</li>
+ <li><span class="gitcommit">[master 37c58b1]</span> Fixed leak of mutexes when
+ using pthread or win32 mutexs as locks. Thanks to Gavin Lambert for reporting
+ this.</li>
+ <li><span class="gitcommit">[master f67e284]</span> Fixed nedflushlogs() not
+ actually flushing data and/or causing a segfault. Thanks to Roman Tatkin for
+ reporting this.</li>
+ <li><span class="gitcommit">[master 1324bf3]</span> Finally got round to retiring
+ the MSVC project files as they were sources of never ending hassle due to being
+ out of sync with the SConstruct config. Rebuilt scons build system to be fully
+ compatible with MSVC instead (long overdue!)</li>
+ <li><span class="gitcommit">[master 068494e]</span> As the release of v1.10
+ RC1 approaches, fixed a long standing problem with the binary patcher where
+ multiple MSVCRT versions in the process weren't handled - everything was sent
+ to one MSVCRT only, and needless to say that sorta worked sometimes and sometimes
+ not. Now when nedmalloc passes a foreign block to the system allocator, it runs
+ a stack backtrace to figure out what MSVCRT in the process it ought to pass
+ it to. It's slow, but fixes a very common segfault on process exit on VS2010.</li>
+ <li><span class="gitcommit">[master 4cca52c]</span> Very embarrassingly, nedmalloc
+ has been severely but unpredictably broken on POSIX for over a year now when built with DEBUG defined.
+ This was turning on DEFAULT_GRANULARITY_ALIGNED whose POSIX implementation
+ was causing random segfaults so mysterious that neither gdb nor valgrind
+ could pick them up - in other words, the very worst kind of memory
+ corruption: undetectable, untraceable and undebuggable. I only found them
+ myself due to a recent bug report for TnFOX on POSIX where due to luck, very
+ recent Linux kernels just happened by pure accident to cause this bug to
+ manifest itself as preventing process init right at the very start - so
+ early that no debugger could attach. After over a week of trial & error I
+ narrowed it down to being somewhere in nedmalloc, then having something to
+ do with DEBUG being defined or not, then two hours ago the eureka moment
+ arrived and I quite literally did a jig around the room in joy. Problem is
+ now fixed thank the heavens!!!</li>
+ <li><span class="gitcommit">[master 3d55a01]</span> Fixed a problem where the
+ binary patcher was early outing too soon and therefore failing to patch all
+ the binaries properly. It would seem that the Microsoft linker doesn't sort
+ the import table like I had thought it did - I would guess it sorts per DLL
+ location, otherwise is unsorted. Thanks to Roman Tatkin for reporting this bug.</li>
+ <li><span class="gitcommit">[master 6c74071]</span> Added override of _GNU_SOURCE
+ for when HAVE_MREMAP is auto-detected. Thanks to Maxim Zakharov for reporting
+ this issue.</li>
+ <li><span class="gitcommit">[master dee2d27]</span> Marked off the v2 malloc API
+ as deprecated in preparation for beta release. Updated CHM documentation.</li>
+</ul>
+<h3>v1.06 beta 2 21st March 2010:</h3>
+<ul>
+ <li>{ 1153 } Added detection of whether host process is using MSVCRT or MSVCRTD
+ and the fixing up of which runtime tolerant nedalloc should use if nedalloc
+ was linked differently. This ought to save a great deal of hassle later on by
+ preventing failed-to-RTM user bug reports :)</li>
+ <li>{ 1154 } Fixed nedalloc trying to use MLOCK_T even when USE_LOCKS=0. Thanks
+ to Ariel Manzur for reporting this.</li>
+ <li>{ 1155 } Fixed USE_SPIN_LOCKS=0 not compiling on Windows.</li>
+ <li>{ 1157 } Fixed bug where foreign blocks entering the threadcache weren't
+ being marked as such, thus typically causing a segfault on process exit.</li>
+ <li>{ 1158 } Fixed compilation problems on mingw. Thanks to Amanieu d'Antras
+ for reporting these.</li>
+ <li>{ 1159 } Released as beta2.</li>
+</ul>
+<h3>v1.06 beta 1 13th January 2010:</h3>
+<ul>
+ <li>{ 1079 } Fixed misdeclaration of struct mallinfo as C++ type. Thanks to
+ James Mansion for reporting this.</li>
+ <li>{ 1082 } Fixed dlmalloc bug which caused header corruption to mmap() allocations
+ when running under multiple threads.</li>
+ <li>{ 1088 } Fixed assertion failure for nedblksize() with latest dlmalloc.
+ Thanks to Anteru for reporting this.</li>
+ <li>{ 1088 } Added neddestroysyspool(). Thanks to Lars Wehmeyer for suggesting
+ this.</li>
+ <li>{ 1088 } Fixed thread id high bit set bug causing SIGABRT on Mac OS X. Thanks
+ to Chris Dillman for reporting this.</li>
+ <li>{ 1094 } Integrated dlmalloc v2.8.4 final.</li>
+ <li>{ 1095 } Added nedtrimthreadcache(). Thanks to Hayim Hendeles for suggesting
+ this.</li>
+ <li>{ 1095 } Fixed silly assertion of null pointer dereference. Thanks to Ullrich
+ Heinemann for reporting this.</li>
+ <li>{ 1096 } Fixed lots of level 4 warnings on MSVC. Thanks to Anteru for suggesting
+ this.</li>
+ <li>{ 1098 } Improved non-nedalloc block detection to 6.25% probability of being
+ wrong. Thanks to Applied Research Associates for sponsoring this.</li>
+ <li>{ 1099 } Added USE_MAGIC_HEADERS which allows nedalloc to handle freeing
+ a system allocated block. Added USE_ALLOCATOR which allows the changing of which
+ backend allocator to use (with choices between the system allocator and dlmalloc
+ - choosing the system allocator is intended for debug situations only e.g. valgrind).
+ Thanks to Applied Research Associates for sponsoring this.</li>
+ <li>{ 1105 } Added ability to build nedalloc as a DLL. Added support for a run
+ time PE binary patcher which can patch all usage of the system allocator replacing
+ it with nedalloc. Thanks to Applied Research Associates for sponsoring this.</li>
+ <li>{ 1108 } Added patcher loader which can load any arbitrary program injecting
+ the nedalloc DLL which then patches in its replacement for the system allocator.
+ Doesn't work on all programs, but does on most e.g. Microsoft Word. Thanks to
+ Applied Research Associates for sponsoring this.</li>
+ <li>{ 1116 } Finished debugging and optimising the latest additions to the codebase.
+ The patcher now works well on x64 as well as x86. Added support for large pages
+ on Windows. Thanks to Applied Research Associates for sponsoring this.</li>
+ <li>{ 1125 } Added nedpoollist() which returns a snapshot of the nedpool's currently
+ existing. The Windows DLL thread exit code now disables the thread cache for
+ all currently existing nedpool's. Thanks to Applied Research Associates for
+ sponsoring this.</li>
+ <li>{ 1126 } Added ENABLE_TOLERANT_NEDMALLOC which allows nedalloc to recognise
+ system allocator blocks and to do the right thing with them.</li>
+ <li>{ 1139 } Added link time code generation support for Windows builds. This
+ currently has zero performance improvement on x64 (on MSVC9) but can add 15%
+ to x86 performance (on MSVC9). Also added scons SConstruct and SConscript files.</li>
+</ul>
+<h3>v1.05 15th June 2008:</h3>
+<ul>
+ <li>{ 1042 } Added error check for TLSSET() and TLSFREE() macros. Thanks to
+ Markus Elfring for reporting this.</li>
+ <li>{ 1043 } Fixed a segfault when freeing memory allocated using nedindependent_comalloc().
+ Thanks to Pavel Vozenilek for reporting this.</li>
+</ul>
+<h3>v1.04 14th July 2007:</h3>
+<ul>
+ <li>Fixed a bug with the new optimised implementation that failed to lock on
+ a realloc under certain conditions.</li>
+ <li>Fixed lack of thread synchronisation in InitPool() causing pool corruption.</li>
+ <li>Fixed a memory leak of thread cache contents on disabling. Thanks to Earl
+ Chew for reporting this.</li>
+ <li>Added a sanity check for freed blocks being valid.</li>
+ <li>Reworked test.c into being a torture test.</li>
+ <li>Fixed GCC assembler optimisation misspecification.</li>
+</ul>
+<h3>v1.04alpha_svn915 7th October 2006:</h3>
+<ul>
+ <li>Fixed failure to unlock thread cache list if allocating a new list failed.
+ Thanks to Dmitry Chichkov for reporting this. Futher thanks to Aleksey Sanin.</li>
+ <li>Fixed realloc(0, <size>) segfaulting. Thanks to Dmitry Chichkov for reporting
+ this.</li>
+ <li>Made config defines #ifndef so they can be overriden by the build system.
+ Thanks to Aleksey Sanin for suggesting this.</li>
+ <li>Fixed deadlock in nedprealloc() due to unnecessary locking of preferred
+ thread mspace when mspace_realloc() always uses the original block's mspace
+ anyway. Thanks to Aleksey Sanin for reporting this.</li>
+ <li>Made some speed improvements by hacking mspace_malloc() to no longer lock
+ its mspace, thus allowing the recursive mutex implementation to be removed with
+ an associated speed increase. Thanks to Aleksey Sanin for suggesting this.</li>
+ <li>Fixed a bug where allocating mspaces overran its max limit. Thanks to Aleksey
+ Sanin for reporting this.</li>
+</ul>
+<h3>v1.03 10th July 2006:</h3>
+<ul>
+ <li>Fixed memory corruption bug in threadcache code which only appeared with
+ >4 threads and in heavy use of the threadcache.</li>
+</ul>
+<h3>v1.02 15th May 2006:</h3>
+<ul>
+ <li>Integrated dlmalloc v2.8.4, fixing the win32 memory release problem and
+ improving performance still further. Speed is now up to twice the speed of v1.01
+ (average is 67% faster).</li>
+ <li>Fixed win32 critical section implementation. Thanks to Pavel Kuznetsov for
+ reporting this.</li>
+ <li>Wasn't locking mspace if all mspaces were locked. Thanks to Pavel Kuznetsov
+ for reporting this.</li>
+ <li>Added Apple Mac OS X support.</li>
+</ul>
+<h3>v1.01 24th February 2006:</h3>
+<ul>
+ <li>Fixed multiprocessor scaling problems by removing sources of cache sloshing.</li>
+ <li>Earl Chew <earl_chew <at> agilent <dot> com> sent patches for the following:
+ <ol>
+ <li>size2binidx() wasn't working for default code path (non x86).</li>
+ <li>Fixed failure to release mspace lock under certain circumstances which
+ caused a deadlock.</li>
+ </ol>
+ </li>
+</ul>
+<h3>v1.00 1st January 2006:</h3>
+<ul>
+ <li>First release</li>
+</ul>
+
+</body>
+
+</html>
diff --git a/malloc.c.h b/malloc.c.h new file mode 100644 index 0000000..ff33c94 --- /dev/null +++ b/malloc.c.h @@ -0,0 +1,6186 @@ +/*
+ This is a version (aka dlmalloc) of malloc/free/realloc written by
+ Doug Lea and released to the public domain, as explained at
+ http://creativecommons.org/licenses/publicdomain. Send questions,
+ comments, complaints, performance data, etc to dl@cs.oswego.edu
+
+* Version 2.8.4 Wed May 27 09:56:23 2009 Doug Lea (dl at gee)
+
+ Note: There may be an updated version of this malloc obtainable at
+ ftp://gee.cs.oswego.edu/pub/misc/malloc.c
+ Check before installing!
+
+* Quickstart
+
+ This library is all in one file to simplify the most common usage:
+ ftp it, compile it (-O3), and link it into another program. All of
+ the compile-time options default to reasonable values for use on
+ most platforms. You might later want to step through various
+ compile-time and dynamic tuning options.
+
+ For convenience, an include file for code using this malloc is at:
+ ftp://gee.cs.oswego.edu/pub/misc/malloc-2.8.4.h
+ You don't really need this .h file unless you call functions not
+ defined in your system include files. The .h file contains only the
+ excerpts from this file needed for using this malloc on ANSI C/C++
+ systems, so long as you haven't changed compile-time options about
+ naming and tuning parameters. If you do, then you can create your
+ own malloc.h that does include all settings by cutting at the point
+ indicated below. Note that you may already by default be using a C
+ library containing a malloc that is based on some version of this
+ malloc (for example in linux). You might still want to use the one
+ in this file to customize settings or to avoid overheads associated
+ with library versions.
+
+* Vital statistics:
+
+ Supported pointer/size_t representation: 4 or 8 bytes
+ size_t MUST be an unsigned type of the same width as
+ pointers. (If you are using an ancient system that declares
+ size_t as a signed type, or need it to be a different width
+ than pointers, you can use a previous release of this malloc
+ (e.g. 2.7.2) supporting these.)
+
+ Alignment: 8 bytes (default)
+ This suffices for nearly all current machines and C compilers.
+ However, you can define MALLOC_ALIGNMENT to be wider than this
+ if necessary (up to 128bytes), at the expense of using more space.
+
+ Minimum overhead per allocated chunk: 4 or 8 bytes (if 4byte sizes)
+ 8 or 16 bytes (if 8byte sizes)
+ Each malloced chunk has a hidden word of overhead holding size
+ and status information, and additional cross-check word
+ if FOOTERS is defined.
+
+ Minimum allocated size: 4-byte ptrs: 16 bytes (including overhead)
+ 8-byte ptrs: 32 bytes (including overhead)
+
+ Even a request for zero bytes (i.e., malloc(0)) returns a
+ pointer to something of the minimum allocatable size.
+ The maximum overhead wastage (i.e., number of extra bytes
+ allocated than were requested in malloc) is less than or equal
+ to the minimum size, except for requests >= mmap_threshold that
+ are serviced via mmap(), where the worst case wastage is about
+ 32 bytes plus the remainder from a system page (the minimal
+ mmap unit); typically 4096 or 8192 bytes.
+
+ Security: static-safe; optionally more or less
+ The "security" of malloc refers to the ability of malicious
+ code to accentuate the effects of errors (for example, freeing
+ space that is not currently malloc'ed or overwriting past the
+ ends of chunks) in code that calls malloc. This malloc
+ guarantees not to modify any memory locations below the base of
+ heap, i.e., static variables, even in the presence of usage
+ errors. The routines additionally detect most improper frees
+ and reallocs. All this holds as long as the static bookkeeping
+ for malloc itself is not corrupted by some other means. This
+ is only one aspect of security -- these checks do not, and
+ cannot, detect all possible programming errors.
+
+ If FOOTERS is defined nonzero, then each allocated chunk
+ carries an additional check word to verify that it was malloced
+ from its space. These check words are the same within each
+ execution of a program using malloc, but differ across
+ executions, so externally crafted fake chunks cannot be
+ freed. This improves security by rejecting frees/reallocs that
+ could corrupt heap memory, in addition to the checks preventing
+ writes to statics that are always on. This may further improve
+ security at the expense of time and space overhead. (Note that
+ FOOTERS may also be worth using with MSPACES.)
+
+ By default detected errors cause the program to abort (calling
+ "abort()"). You can override this to instead proceed past
+ errors by defining PROCEED_ON_ERROR. In this case, a bad free
+ has no effect, and a malloc that encounters a bad address
+ caused by user overwrites will ignore the bad address by
+ dropping pointers and indices to all known memory. This may
+ be appropriate for programs that should continue if at all
+ possible in the face of programming errors, although they may
+ run out of memory because dropped memory is never reclaimed.
+
+ If you don't like either of these options, you can define
+ CORRUPTION_ERROR_ACTION and USAGE_ERROR_ACTION to do anything
+ else. And if if you are sure that your program using malloc has
+ no errors or vulnerabilities, you can define INSECURE to 1,
+ which might (or might not) provide a small performance improvement.
+
+ Thread-safety: NOT thread-safe unless USE_LOCKS defined
+ When USE_LOCKS is defined, each public call to malloc, free,
+ etc is surrounded with either a pthread mutex or a win32
+ spinlock (depending on WIN32). This is not especially fast, and
+ can be a major bottleneck. It is designed only to provide
+ minimal protection in concurrent environments, and to provide a
+ basis for extensions. If you are using malloc in a concurrent
+ program, consider instead using nedmalloc
+ (http://www.nedprod.com/programs/portable/nedmalloc/) or
+ ptmalloc (See http://www.malloc.de), which are derived
+ from versions of this malloc.
+
+ System requirements: Any combination of MORECORE and/or MMAP/MUNMAP
+ This malloc can use unix sbrk or any emulation (invoked using
+ the CALL_MORECORE macro) and/or mmap/munmap or any emulation
+ (invoked using CALL_MMAP/CALL_MUNMAP) to get and release system
+ memory. On most unix systems, it tends to work best if both
+ MORECORE and MMAP are enabled. On Win32, it uses emulations
+ based on VirtualAlloc. It also uses common C library functions
+ like memset.
+
+ Compliance: I believe it is compliant with the Single Unix Specification
+ (See http://www.unix.org). Also SVID/XPG, ANSI C, and probably
+ others as well.
+
+* Overview of algorithms
+
+ This is not the fastest, most space-conserving, most portable, or
+ most tunable malloc ever written. However it is among the fastest
+ while also being among the most space-conserving, portable and
+ tunable. Consistent balance across these factors results in a good
+ general-purpose allocator for malloc-intensive programs.
+
+ In most ways, this malloc is a best-fit allocator. Generally, it
+ chooses the best-fitting existing chunk for a request, with ties
+ broken in approximately least-recently-used order. (This strategy
+ normally maintains low fragmentation.) However, for requests less
+ than 256bytes, it deviates from best-fit when there is not an
+ exactly fitting available chunk by preferring to use space adjacent
+ to that used for the previous small request, as well as by breaking
+ ties in approximately most-recently-used order. (These enhance
+ locality of series of small allocations.) And for very large requests
+ (>= 256Kb by default), it relies on system memory mapping
+ facilities, if supported. (This helps avoid carrying around and
+ possibly fragmenting memory used only for large chunks.)
+
+ All operations (except malloc_stats and mallinfo) have execution
+ times that are bounded by a constant factor of the number of bits in
+ a size_t, not counting any clearing in calloc or copying in realloc,
+ or actions surrounding MORECORE and MMAP that have times
+ proportional to the number of non-contiguous regions returned by
+ system allocation routines, which is often just 1. In real-time
+ applications, you can optionally suppress segment traversals using
+ NO_SEGMENT_TRAVERSAL, which assures bounded execution even when
+ system allocators return non-contiguous spaces, at the typical
+ expense of carrying around more memory and increased fragmentation.
+
+ The implementation is not very modular and seriously overuses
+ macros. Perhaps someday all C compilers will do as good a job
+ inlining modular code as can now be done by brute-force expansion,
+ but now, enough of them seem not to.
+
+ Some compilers issue a lot of warnings about code that is
+ dead/unreachable only on some platforms, and also about intentional
+ uses of negation on unsigned types. All known cases of each can be
+ ignored.
+
+ For a longer but out of date high-level description, see
+ http://gee.cs.oswego.edu/dl/html/malloc.html
+
+* MSPACES
+ If MSPACES is defined, then in addition to malloc, free, etc.,
+ this file also defines mspace_malloc, mspace_free, etc. These
+ are versions of malloc routines that take an "mspace" argument
+ obtained using create_mspace, to control all internal bookkeeping.
+ If ONLY_MSPACES is defined, only these versions are compiled.
+ So if you would like to use this allocator for only some allocations,
+ and your system malloc for others, you can compile with
+ ONLY_MSPACES and then do something like...
+ static mspace mymspace = create_mspace(0,0); // for example
+ #define mymalloc(bytes) mspace_malloc(mymspace, bytes)
+
+ (Note: If you only need one instance of an mspace, you can instead
+ use "USE_DL_PREFIX" to relabel the global malloc.)
+
+ You can similarly create thread-local allocators by storing
+ mspaces as thread-locals. For example:
+ static __thread mspace tlms = 0;
+ void* tlmalloc(size_t bytes) {
+ if (tlms == 0) tlms = create_mspace(0, 0);
+ return mspace_malloc(tlms, bytes);
+ }
+ void tlfree(void* mem) { mspace_free(tlms, mem); }
+
+ Unless FOOTERS is defined, each mspace is completely independent.
+ You cannot allocate from one and free to another (although
+ conformance is only weakly checked, so usage errors are not always
+ caught). If FOOTERS is defined, then each chunk carries around a tag
+ indicating its originating mspace, and frees are directed to their
+ originating spaces.
+
+ ------------------------- Compile-time options ---------------------------
+
+Be careful in setting #define values for numerical constants of type
+size_t. On some systems, literal values are not automatically extended
+to size_t precision unless they are explicitly casted. You can also
+use the symbolic values MAX_SIZE_T, SIZE_T_ONE, etc below.
+
+WIN32 default: defined if _WIN32 defined
+ Defining WIN32 sets up defaults for MS environment and compilers.
+ Otherwise defaults are for unix. Beware that there seem to be some
+ cases where this malloc might not be a pure drop-in replacement for
+ Win32 malloc: Random-looking failures from Win32 GDI API's (eg;
+ SetDIBits()) may be due to bugs in some video driver implementations
+ when pixel buffers are malloc()ed, and the region spans more than
+ one VirtualAlloc()ed region. Because dlmalloc uses a small (64Kb)
+ default granularity, pixel buffers may straddle virtual allocation
+ regions more often than when using the Microsoft allocator. You can
+ avoid this by using VirtualAlloc() and VirtualFree() for all pixel
+ buffers rather than using malloc(). If this is not possible,
+ recompile this malloc with a larger DEFAULT_GRANULARITY.
+
+MALLOC_ALIGNMENT default: (size_t)8
+ Controls the minimum alignment for malloc'ed chunks. It must be a
+ power of two and at least 8, even on machines for which smaller
+ alignments would suffice. It may be defined as larger than this
+ though. Note however that code and data structures are optimized for
+ the case of 8-byte alignment.
+
+MSPACES default: 0 (false)
+ If true, compile in support for independent allocation spaces.
+ This is only supported if HAVE_MMAP is true.
+
+ONLY_MSPACES default: 0 (false)
+ If true, only compile in mspace versions, not regular versions.
+
+USE_LOCKS default: 0 (false)
+ Causes each call to each public routine to be surrounded with
+ pthread or WIN32 mutex lock/unlock. (If set true, this can be
+ overridden on a per-mspace basis for mspace versions.) If set to a
+ non-zero value other than 1, locks are used, but their
+ implementation is left out, so lock functions must be supplied manually,
+ as described below.
+
+USE_SPIN_LOCKS default: 1 iff USE_LOCKS and on x86 using gcc or MSC
+ If true, uses custom spin locks for locking. This is currently
+ supported only for x86 platforms using gcc or recent MS compilers.
+ Otherwise, posix locks or win32 critical sections are used.
+
+FOOTERS default: 0
+ If true, provide extra checking and dispatching by placing
+ information in the footers of allocated chunks. This adds
+ space and time overhead.
+
+INSECURE default: 0
+ If true, omit checks for usage errors and heap space overwrites.
+
+USE_DL_PREFIX default: NOT defined
+ Causes compiler to prefix all public routines with the string 'dl'.
+ This can be useful when you only want to use this malloc in one part
+ of a program, using your regular system malloc elsewhere.
+
+ABORT default: defined as abort()
+ Defines how to abort on failed checks. On most systems, a failed
+ check cannot die with an "assert" or even print an informative
+ message, because the underlying print routines in turn call malloc,
+ which will fail again. Generally, the best policy is to simply call
+ abort(). It's not very useful to do more than this because many
+ errors due to overwriting will show up as address faults (null, odd
+ addresses etc) rather than malloc-triggered checks, so will also
+ abort. Also, most compilers know that abort() does not return, so
+ can better optimize code conditionally calling it.
+
+PROCEED_ON_ERROR default: defined as 0 (false)
+ Controls whether detected bad addresses cause them to bypassed
+ rather than aborting. If set, detected bad arguments to free and
+ realloc are ignored. And all bookkeeping information is zeroed out
+ upon a detected overwrite of freed heap space, thus losing the
+ ability to ever return it from malloc again, but enabling the
+ application to proceed. If PROCEED_ON_ERROR is defined, the
+ static variable malloc_corruption_error_count is compiled in
+ and can be examined to see if errors have occurred. This option
+ generates slower code than the default abort policy.
+
+DEBUG default: NOT defined
+ The DEBUG setting is mainly intended for people trying to modify
+ this code or diagnose problems when porting to new platforms.
+ However, it may also be able to better isolate user errors than just
+ using runtime checks. The assertions in the check routines spell
+ out in more detail the assumptions and invariants underlying the
+ algorithms. The checking is fairly extensive, and will slow down
+ execution noticeably. Calling malloc_stats or mallinfo with DEBUG
+ set will attempt to check every non-mmapped allocated and free chunk
+ in the course of computing the summaries.
+
+ABORT_ON_ASSERT_FAILURE default: defined as 1 (true)
+ Debugging assertion failures can be nearly impossible if your
+ version of the assert macro causes malloc to be called, which will
+ lead to a cascade of further failures, blowing the runtime stack.
+ ABORT_ON_ASSERT_FAILURE cause assertions failures to call abort(),
+ which will usually make debugging easier.
+
+MALLOC_FAILURE_ACTION default: sets errno to ENOMEM, or no-op on win32
+ The action to take before "return 0" when malloc fails to be able to
+ return memory because there is none available.
+
+HAVE_MORECORE default: 1 (true) unless win32 or ONLY_MSPACES
+ True if this system supports sbrk or an emulation of it.
+
+MORECORE default: sbrk
+ The name of the sbrk-style system routine to call to obtain more
+ memory. See below for guidance on writing custom MORECORE
+ functions. The type of the argument to sbrk/MORECORE varies across
+ systems. It cannot be size_t, because it supports negative
+ arguments, so it is normally the signed type of the same width as
+ size_t (sometimes declared as "intptr_t"). It doesn't much matter
+ though. Internally, we only call it with arguments less than half
+ the max value of a size_t, which should work across all reasonable
+ possibilities, although sometimes generating compiler warnings.
+
+MORECORE_CONTIGUOUS default: 1 (true) if HAVE_MORECORE
+ If true, take advantage of fact that consecutive calls to MORECORE
+ with positive arguments always return contiguous increasing
+ addresses. This is true of unix sbrk. It does not hurt too much to
+ set it true anyway, since malloc copes with non-contiguities.
+ Setting it false when definitely non-contiguous saves time
+ and possibly wasted space it would take to discover this though.
+
+MORECORE_CANNOT_TRIM default: NOT defined
+ True if MORECORE cannot release space back to the system when given
+ negative arguments. This is generally necessary only if you are
+ using a hand-crafted MORECORE function that cannot handle negative
+ arguments.
+
+NO_SEGMENT_TRAVERSAL default: 0
+ If non-zero, suppresses traversals of memory segments
+ returned by either MORECORE or CALL_MMAP. This disables
+ merging of segments that are contiguous, and selectively
+ releasing them to the OS if unused, but bounds execution times.
+
+HAVE_MMAP default: 1 (true)
+ True if this system supports mmap or an emulation of it. If so, and
+ HAVE_MORECORE is not true, MMAP is used for all system
+ allocation. If set and HAVE_MORECORE is true as well, MMAP is
+ primarily used to directly allocate very large blocks. It is also
+ used as a backup strategy in cases where MORECORE fails to provide
+ space from system. Note: A single call to MUNMAP is assumed to be
+ able to unmap memory that may have be allocated using multiple calls
+ to MMAP, so long as they are adjacent.
+
+HAVE_MREMAP default: 1 on linux, else 0
+ If true realloc() uses mremap() to re-allocate large blocks and
+ extend or shrink allocation spaces.
+
+MMAP_CLEARS default: 1 except on WINCE.
+ True if mmap clears memory so calloc doesn't need to. This is true
+ for standard unix mmap using /dev/zero and on WIN32 except for WINCE.
+
+USE_BUILTIN_FFS default: 0 (i.e., not used)
+ Causes malloc to use the builtin ffs() function to compute indices.
+ Some compilers may recognize and intrinsify ffs to be faster than the
+ supplied C version. Also, the case of x86 using gcc is special-cased
+ to an asm instruction, so is already as fast as it can be, and so
+ this setting has no effect. Similarly for Win32 under recent MS compilers.
+ (On most x86s, the asm version is only slightly faster than the C version.)
+
+malloc_getpagesize default: derive from system includes, or 4096.
+ The system page size. To the extent possible, this malloc manages
+ memory from the system in page-size units. This may be (and
+ usually is) a function rather than a constant. This is ignored
+ if WIN32, where page size is determined using getSystemInfo during
+ initialization. This may be several megabytes if ENABLE_LARGE_PAGES
+ is enabled.
+
+ENABLE_LARGE_PAGES default: NOT defined
+ Causes the system page size to be the value of GetLargePageMinimum()
+ if that function is available (Windows Server 2003/Vista or later)
+ or gethugepagesize() if the libhugetlbfs library is loaded into this
+ process (Linux, BSD and others, see http://libhugetlbfs.sourceforge.net/).
+ This allows the use of large page entries in the MMU which can
+ significantly improve performance in large working set applications
+ as TLB cache load is reduced by a factor of three. Note that on Windows
+ enabling this option is equal to locking the process' memory in current
+ implementations of Windows and requires the SE_LOCK_MEMORY_PRIVILEGE
+ to be held by the process in order to succeed. For POSIX to make
+ large page support work you need to link against libhugetlbfs, otherwise
+ support silently disables itself.
+
+USE_DEV_RANDOM default: 0 (i.e., not used)
+ Causes malloc to use /dev/random to initialize secure magic seed for
+ stamping footers. Otherwise, the current time is used.
+
+NO_MALLINFO default: 0
+ If defined, don't compile "mallinfo". This can be a simple way
+ of dealing with mismatches between system declarations and
+ those in this file.
+
+MALLINFO_FIELD_TYPE default: size_t
+ The type of the fields in the mallinfo struct. This was originally
+ defined as "int" in SVID etc, but is more usefully defined as
+ size_t. The value is used only if HAVE_USR_INCLUDE_MALLOC_H is not set
+
+REALLOC_ZERO_BYTES_FREES default: not defined
+ This should be set if a call to realloc with zero bytes should
+ be the same as a call to free. Some people think it should. Otherwise,
+ since this malloc returns a unique pointer for malloc(0), so does
+ realloc(p, 0).
+
+LACKS_UNISTD_H, LACKS_FCNTL_H, LACKS_SYS_PARAM_H, LACKS_SYS_MMAN_H
+LACKS_STRINGS_H, LACKS_STRING_H, LACKS_SYS_TYPES_H, LACKS_ERRNO_H
+LACKS_STDLIB_H default: NOT defined unless on WIN32
+ Define these if your system does not have these header files.
+ You might need to manually insert some of the declarations they provide.
+
+DEFAULT_GRANULARITY default: page size if MORECORE_CONTIGUOUS,
+ system_info.dwAllocationGranularity in WIN32,
+ GetLargePageMinimum() if ENABLE_LARGE_PAGES,
+ otherwise 64K.
+ Also settable using mallopt(M_GRANULARITY, x)
+ The unit for allocating and deallocating memory from the system. On
+ most systems with contiguous MORECORE, there is no reason to
+ make this more than a page. However, systems with MMAP tend to
+ either require or encourage larger granularities. You can increase
+ this value to prevent system allocation functions to be called so
+ often, especially if they are slow. The value must be at least one
+ page and must be a power of two. Setting to 0 causes initialization
+ to either page size or win32 region size. (Note: In previous
+ versions of malloc, the equivalent of this option was called
+ "TOP_PAD")
+
+DEFAULT_GRANULARITY_ALIGNED default: undefined (which means page size)
+ Affects Win32 only (POSIX doesn't have the necessary API support).
+ Whether to enforce alignment when allocating and deallocating memory
+ from the system i.e. the base address of all allocations will be
+ aligned to DEFAULT_GRANULARITY if it is set. Note that enabling this carries
+ some overhead as multiple calls must now be made when probing for a valid
+ aligned value, however it does greatly ease the checking for whether
+ a given memory pointer was allocated by this allocator rather than
+ some other.
+
+DEFAULT_TRIM_THRESHOLD default: 2MB
+ Also settable using mallopt(M_TRIM_THRESHOLD, x)
+ The maximum amount of unused top-most memory to keep before
+ releasing via malloc_trim in free(). Automatic trimming is mainly
+ useful in long-lived programs using contiguous MORECORE. Because
+ trimming via sbrk can be slow on some systems, and can sometimes be
+ wasteful (in cases where programs immediately afterward allocate
+ more large chunks) the value should be high enough so that your
+ overall system performance would improve by releasing this much
+ memory. As a rough guide, you might set to a value close to the
+ average size of a process (program) running on your system.
+ Releasing this much memory would allow such a process to run in
+ memory. Generally, it is worth tuning trim thresholds when a
+ program undergoes phases where several large chunks are allocated
+ and released in ways that can reuse each other's storage, perhaps
+ mixed with phases where there are no such chunks at all. The trim
+ value must be greater than page size to have any useful effect. To
+ disable trimming completely, you can set to MAX_SIZE_T. Note that the trick
+ some people use of mallocing a huge space and then freeing it at
+ program startup, in an attempt to reserve system memory, doesn't
+ have the intended effect under automatic trimming, since that memory
+ will immediately be returned to the system.
+
+DEFAULT_MMAP_THRESHOLD default: 256K
+ Also settable using mallopt(M_MMAP_THRESHOLD, x)
+ The request size threshold for using MMAP to directly service a
+ request. Requests of at least this size that cannot be allocated
+ using already-existing space will be serviced via mmap. (If enough
+ normal freed space already exists it is used instead.) Using mmap
+ segregates relatively large chunks of memory so that they can be
+ individually obtained and released from the host system. A request
+ serviced through mmap is never reused by any other request (at least
+ not directly; the system may just so happen to remap successive
+ requests to the same locations). Segregating space in this way has
+ the benefits that: Mmapped space can always be individually released
+ back to the system, which helps keep the system level memory demands
+ of a long-lived program low. Also, mapped memory doesn't become
+ `locked' between other chunks, as can happen with normally allocated
+ chunks, which means that even trimming via malloc_trim would not
+ release them. However, it has the disadvantage that the space
+ cannot be reclaimed, consolidated, and then used to service later
+ requests, as happens with normal chunks. The advantages of mmap
+ nearly always outweigh disadvantages for "large" chunks, but the
+ value of "large" may vary across systems. The default is an
+ empirically derived value that works well in most systems. You can
+ disable mmap by setting to MAX_SIZE_T.
+
+MAX_RELEASE_CHECK_RATE default: 4095 unless not HAVE_MMAP
+ The number of consolidated frees between checks to release
+ unused segments when freeing. When using non-contiguous segments,
+ especially with multiple mspaces, checking only for topmost space
+ doesn't always suffice to trigger trimming. To compensate for this,
+ free() will, with a period of MAX_RELEASE_CHECK_RATE (or the
+ current number of segments, if greater) try to release unused
+ segments to the OS when freeing chunks that result in
+ consolidation. The best value for this parameter is a compromise
+ between slowing down frees with relatively costly checks that
+ rarely trigger versus holding on to unused memory. To effectively
+ disable, set to MAX_SIZE_T. This may lead to a very slight speed
+ improvement at the expense of carrying around more memory.
+*/
+
+/* Version identifier to allow people to support multiple versions */
+#ifndef DLMALLOC_VERSION
+#define DLMALLOC_VERSION 20804
+#endif /* DLMALLOC_VERSION */
+
+#ifndef WIN32
+#ifdef _WIN32
+#define WIN32 1
+#endif /* _WIN32 */
+#ifdef _WIN32_WCE
+#define LACKS_FCNTL_H
+#define WIN32 1
+#endif /* _WIN32_WCE */
+#endif /* WIN32 */
+#ifdef WIN32
+#define WIN32_LEAN_AND_MEAN
+#include <windows.h>
+#include <tchar.h>
+#define HAVE_MMAP 1
+#define HAVE_MORECORE 0
+#define LACKS_UNISTD_H
+#define LACKS_SYS_PARAM_H
+#define LACKS_SYS_MMAN_H
+#define LACKS_STRING_H
+#define LACKS_STRINGS_H
+#define LACKS_SYS_TYPES_H
+#define LACKS_ERRNO_H
+#ifndef MALLOC_FAILURE_ACTION
+#define MALLOC_FAILURE_ACTION
+#endif /* MALLOC_FAILURE_ACTION */
+#ifndef MMAP_CLEARS
+#ifdef _WIN32_WCE /* WINCE reportedly does not clear */
+#define MMAP_CLEARS 0
+#else
+#define MMAP_CLEARS 1
+#endif /* _WIN32_WCE */
+#endif
+#endif /* WIN32 */
+
+#if defined(DARWIN) || defined(_DARWIN)
+/* Mac OSX docs advise not to use sbrk; it seems better to use mmap */
+#ifndef HAVE_MORECORE
+#define HAVE_MORECORE 0
+#define HAVE_MMAP 1
+/* OSX allocators provide 16 byte alignment */
+#ifndef MALLOC_ALIGNMENT
+#define MALLOC_ALIGNMENT ((size_t)16U)
+#endif
+#endif /* HAVE_MORECORE */
+#endif /* DARWIN */
+
+#ifndef LACKS_SYS_TYPES_H
+#include <sys/types.h> /* For size_t */
+#endif /* LACKS_SYS_TYPES_H */
+
+#if (defined(__GNUC__) && ((__GNUC__ * 10000 + __GNUC_MINOR__ * 100 + __GNUC_PATCHLEVEL__) > 40100)) || (defined(_MSC_VER) && _MSC_VER>=1310)
+#define SPIN_LOCKS_AVAILABLE 1
+#else
+#define SPIN_LOCKS_AVAILABLE 0
+#endif
+
+/* The maximum possible size_t value has all bits set */
+#define MAX_SIZE_T (~(size_t)0)
+
+#ifndef ONLY_MSPACES
+#define ONLY_MSPACES 0 /* define to a value */
+#else
+#define ONLY_MSPACES 1
+#endif /* ONLY_MSPACES */
+#ifndef MSPACES
+#if ONLY_MSPACES
+#define MSPACES 1
+#else /* ONLY_MSPACES */
+#define MSPACES 0
+#endif /* ONLY_MSPACES */
+#endif /* MSPACES */
+#ifndef MALLOC_ALIGNMENT
+#define MALLOC_ALIGNMENT ((size_t)8U)
+#endif /* MALLOC_ALIGNMENT */
+#ifndef FOOTERS
+#define FOOTERS 0
+#endif /* FOOTERS */
+#ifndef ABORT
+#define ABORT abort()
+#endif /* ABORT */
+#ifndef ABORT_ON_ASSERT_FAILURE
+#define ABORT_ON_ASSERT_FAILURE 1
+#endif /* ABORT_ON_ASSERT_FAILURE */
+#ifndef PROCEED_ON_ERROR
+#define PROCEED_ON_ERROR 0
+#endif /* PROCEED_ON_ERROR */
+#ifndef USE_LOCKS
+#define USE_LOCKS 0
+#endif /* USE_LOCKS */
+#ifndef USE_SPIN_LOCKS
+#if USE_LOCKS && SPIN_LOCKS_AVAILABLE
+#define USE_SPIN_LOCKS 1
+#else
+#define USE_SPIN_LOCKS 0
+#endif /* USE_LOCKS && SPIN_LOCKS_AVAILABLE. */
+#endif /* USE_SPIN_LOCKS */
+#ifndef INSECURE
+#define INSECURE 0
+#endif /* INSECURE */
+#ifndef HAVE_MMAP
+#define HAVE_MMAP 1
+#endif /* HAVE_MMAP */
+#ifndef MMAP_CLEARS
+#define MMAP_CLEARS 1
+#endif /* MMAP_CLEARS */
+#ifndef HAVE_MREMAP
+#ifdef linux
+#define HAVE_MREMAP 1
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE /* Turns on mremap() definition */
+#endif
+#else /* linux */
+#define HAVE_MREMAP 0
+#endif /* linux */
+#endif /* HAVE_MREMAP */
+#ifndef MALLOC_FAILURE_ACTION
+#define MALLOC_FAILURE_ACTION errno = ENOMEM;
+#endif /* MALLOC_FAILURE_ACTION */
+#ifndef HAVE_MORECORE
+#if ONLY_MSPACES
+#define HAVE_MORECORE 0
+#else /* ONLY_MSPACES */
+#define HAVE_MORECORE 1
+#endif /* ONLY_MSPACES */
+#endif /* HAVE_MORECORE */
+#if !HAVE_MORECORE
+#define MORECORE_CONTIGUOUS 0
+#else /* !HAVE_MORECORE */
+#define MORECORE_DEFAULT sbrk
+#ifndef MORECORE_CONTIGUOUS
+#define MORECORE_CONTIGUOUS 1
+#endif /* MORECORE_CONTIGUOUS */
+#endif /* HAVE_MORECORE */
+#ifndef DEFAULT_GRANULARITY
+#if (MORECORE_CONTIGUOUS || defined(WIN32))
+#define DEFAULT_GRANULARITY (0) /* 0 means to compute in init_mparams */
+#else /* MORECORE_CONTIGUOUS */
+#define DEFAULT_GRANULARITY ((size_t)64U * (size_t)1024U)
+#endif /* MORECORE_CONTIGUOUS */
+#endif /* DEFAULT_GRANULARITY */
+#ifndef DEFAULT_TRIM_THRESHOLD
+#ifndef MORECORE_CANNOT_TRIM
+#define DEFAULT_TRIM_THRESHOLD ((size_t)2U * (size_t)1024U * (size_t)1024U)
+#else /* MORECORE_CANNOT_TRIM */
+#define DEFAULT_TRIM_THRESHOLD MAX_SIZE_T
+#endif /* MORECORE_CANNOT_TRIM */
+#endif /* DEFAULT_TRIM_THRESHOLD */
+#ifndef DEFAULT_MMAP_THRESHOLD
+#if HAVE_MMAP
+#define DEFAULT_MMAP_THRESHOLD ((size_t)256U * (size_t)1024U)
+#else /* HAVE_MMAP */
+#define DEFAULT_MMAP_THRESHOLD MAX_SIZE_T
+#endif /* HAVE_MMAP */
+#endif /* DEFAULT_MMAP_THRESHOLD */
+#ifndef MAX_RELEASE_CHECK_RATE
+#if HAVE_MMAP
+#define MAX_RELEASE_CHECK_RATE 4095
+#else
+#define MAX_RELEASE_CHECK_RATE MAX_SIZE_T
+#endif /* HAVE_MMAP */
+#endif /* MAX_RELEASE_CHECK_RATE */
+#ifndef USE_BUILTIN_FFS
+#define USE_BUILTIN_FFS 0
+#endif /* USE_BUILTIN_FFS */
+#ifndef USE_DEV_RANDOM
+#define USE_DEV_RANDOM 0
+#endif /* USE_DEV_RANDOM */
+#ifndef NO_MALLINFO
+#define NO_MALLINFO 0
+#endif /* NO_MALLINFO */
+#ifndef MALLINFO_FIELD_TYPE
+#define MALLINFO_FIELD_TYPE size_t
+#endif /* MALLINFO_FIELD_TYPE */
+#ifndef NO_SEGMENT_TRAVERSAL
+#define NO_SEGMENT_TRAVERSAL 0
+#endif /* NO_SEGMENT_TRAVERSAL */
+
+/*
+ malloc2 flag options. These non-portable functions provide
+ additional functionality beyond those traditionally provided by
+ a memory allocator. You may pass custom flags in the bits
+ specified by M2_CUSTOM_FLAGS_MASK.
+*/
+
+#ifndef M2_FLAGS_DEFINED
+#define M2_FLAGS_DEFINED
+
+#define M2_ZERO_MEMORY (1<<0)
+#define M2_PREVENT_MOVE (1<<1)
+#define M2_ALWAYS_MMAP (1<<2)
+#define M2_RESERVED1 (1<<3)
+#define M2_RESERVED2 (1<<4)
+#define M2_RESERVED3 (1<<5)
+#define M2_RESERVED4 (1<<6)
+#define M2_RESERVED5 (1<<7)
+#define M2_RESERVE_ISMULTIPLIER (1<<15)
+/* 7 bits is given to the address reservation specifier.
+This lets you set a multiplier (bit 15 set) or a 1<< shift value.
+*/
+#define M2_RESERVE_MASK 0x00007f00
+#define M2_RESERVE_MULT(n) (M2_RESERVE_ISMULTIPLIER|(((n)<<8)&M2_RESERVE_MASK))
+#define M2_RESERVE_SHIFT(n) (((n)<<8)&M2_RESERVE_MASK)
+#define M2_FLAGS_MASK 0x0000ffff
+#define M2_CUSTOM_FLAGS_BEGIN (1<<16)
+#define M2_CUSTOM_FLAGS_MASK 0xffff0000
+
+#endif /* M2_FLAGS_DEFINED */
+
+/*
+ mallopt tuning options. SVID/XPG defines four standard parameter
+ numbers for mallopt, normally defined in malloc.h. None of these
+ are used in this malloc, so setting them has no effect. But this
+ malloc does support the following options.
+*/
+
+#define M_TRIM_THRESHOLD (-1)
+#define M_GRANULARITY (-2)
+#define M_MMAP_THRESHOLD (-3)
+
+/* ------------------------ Mallinfo declarations ------------------------ */
+
+#if !NO_MALLINFO
+/*
+ This version of malloc supports the standard SVID/XPG mallinfo
+ routine that returns a struct containing usage properties and
+ statistics. It should work on any system that has a
+ /usr/include/malloc.h defining struct mallinfo. The main
+ declaration needed is the mallinfo struct that is returned (by-copy)
+ by mallinfo(). The malloinfo struct contains a bunch of fields that
+ are not even meaningful in this version of malloc. These fields are
+ are instead filled by mallinfo() with other numbers that might be of
+ interest.
+
+ HAVE_USR_INCLUDE_MALLOC_H should be set if you have a
+ /usr/include/malloc.h file that includes a declaration of struct
+ mallinfo. If so, it is included; else a compliant version is
+ declared below. These must be precisely the same for mallinfo() to
+ work. The original SVID version of this struct, defined on most
+ systems with mallinfo, declares all fields as ints. But some others
+ define as unsigned long. If your system defines the fields using a
+ type of different width than listed here, you MUST #include your
+ system version and #define HAVE_USR_INCLUDE_MALLOC_H.
+*/
+
+/* #define HAVE_USR_INCLUDE_MALLOC_H */
+
+#ifdef HAVE_USR_INCLUDE_MALLOC_H
+#include "/usr/include/malloc.h"
+#else /* HAVE_USR_INCLUDE_MALLOC_H */
+#ifndef STRUCT_MALLINFO_DECLARED
+#define STRUCT_MALLINFO_DECLARED 1
+struct mallinfo {
+ MALLINFO_FIELD_TYPE arena; /* non-mmapped space allocated from system */
+ MALLINFO_FIELD_TYPE ordblks; /* number of free chunks */
+ MALLINFO_FIELD_TYPE smblks; /* always 0 */
+ MALLINFO_FIELD_TYPE hblks; /* always 0 */
+ MALLINFO_FIELD_TYPE hblkhd; /* space in mmapped regions */
+ MALLINFO_FIELD_TYPE usmblks; /* maximum total allocated space */
+ MALLINFO_FIELD_TYPE fsmblks; /* always 0 */
+ MALLINFO_FIELD_TYPE uordblks; /* total allocated space */
+ MALLINFO_FIELD_TYPE fordblks; /* total free space */
+ MALLINFO_FIELD_TYPE keepcost; /* releasable (via malloc_trim) space */
+};
+#endif /* STRUCT_MALLINFO_DECLARED */
+#endif /* HAVE_USR_INCLUDE_MALLOC_H */
+#endif /* NO_MALLINFO */
+
+/*
+ Try to persuade compilers to inline. The most critical functions for
+ inlining are defined as macros, so these aren't used for them.
+*/
+
+#ifndef FORCEINLINE
+ #if defined(__GNUC__)
+#define FORCEINLINE __inline __attribute__ ((always_inline))
+ #elif defined(_MSC_VER)
+ #define FORCEINLINE __forceinline
+ #endif
+#endif
+#ifndef NOINLINE
+ #if defined(__GNUC__)
+ #define NOINLINE __attribute__ ((noinline))
+ #elif defined(_MSC_VER)
+ #define NOINLINE __declspec(noinline)
+ #else
+ #define NOINLINE
+ #endif
+#endif
+
+#ifdef __cplusplus
+extern "C" {
+#ifndef FORCEINLINE
+ #define FORCEINLINE inline
+#endif
+#endif /* __cplusplus */
+#ifndef FORCEINLINE
+ #define FORCEINLINE
+#endif
+
+#if !ONLY_MSPACES
+
+/* ------------------- Declarations of public routines ------------------- */
+
+#ifndef USE_DL_PREFIX
+#define dlcalloc calloc
+#define dlfree free
+#define dlmalloc malloc
+#define dlmemalign memalign
+#define dlrealloc realloc
+#define dlvalloc valloc
+#define dlpvalloc pvalloc
+#define dlmallinfo mallinfo
+#define dlmallopt mallopt
+#define dlmalloc_trim malloc_trim
+#define dlmalloc_stats malloc_stats
+#define dlmalloc_usable_size malloc_usable_size
+#define dlmalloc_footprint malloc_footprint
+#define dlmalloc_max_footprint malloc_max_footprint
+#define dlindependent_calloc independent_calloc
+#define dlindependent_comalloc independent_comalloc
+#endif /* USE_DL_PREFIX */
+
+
+/*
+ malloc(size_t n)
+ Returns a pointer to a newly allocated chunk of at least n bytes, or
+ null if no space is available, in which case errno is set to ENOMEM
+ on ANSI C systems.
+
+ If n is zero, malloc returns a minimum-sized chunk. (The minimum
+ size is 16 bytes on most 32bit systems, and 32 bytes on 64bit
+ systems.) Note that size_t is an unsigned type, so calls with
+ arguments that would be negative if signed are interpreted as
+ requests for huge amounts of space, which will often fail. The
+ maximum supported value of n differs across systems, but is in all
+ cases less than the maximum representable value of a size_t.
+*/
+void* dlmalloc(size_t);
+
+/*
+ free(void* p)
+ Releases the chunk of memory pointed to by p, that had been previously
+ allocated using malloc or a related routine such as realloc.
+ It has no effect if p is null. If p was not malloced or already
+ freed, free(p) will by default cause the current program to abort.
+*/
+void dlfree(void*);
+
+/*
+ calloc(size_t n_elements, size_t element_size);
+ Returns a pointer to n_elements * element_size bytes, with all locations
+ set to zero.
+*/
+void* dlcalloc(size_t, size_t);
+
+/*
+ realloc(void* p, size_t n)
+ Returns a pointer to a chunk of size n that contains the same data
+ as does chunk p up to the minimum of (n, p's size) bytes, or null
+ if no space is available.
+
+ The returned pointer may or may not be the same as p. The algorithm
+ prefers extending p in most cases when possible, otherwise it
+ employs the equivalent of a malloc-copy-free sequence.
+
+ If p is null, realloc is equivalent to malloc.
+
+ If space is not available, realloc returns null, errno is set (if on
+ ANSI) and p is NOT freed.
+
+ if n is for fewer bytes than already held by p, the newly unused
+ space is lopped off and freed if possible. realloc with a size
+ argument of zero (re)allocates a minimum-sized chunk.
+
+ The old unix realloc convention of allowing the last-free'd chunk
+ to be used as an argument to realloc is not supported.
+*/
+
+void* dlrealloc(void*, size_t);
+
+/*
+ memalign(size_t alignment, size_t n);
+ Returns a pointer to a newly allocated chunk of n bytes, aligned
+ in accord with the alignment argument.
+
+ The alignment argument should be a power of two. If the argument is
+ not a power of two, the nearest greater power is used.
+ 8-byte alignment is guaranteed by normal malloc calls, so don't
+ bother calling memalign with an argument of 8 or less.
+
+ Overreliance on memalign is a sure way to fragment space.
+*/
+void* dlmemalign(size_t, size_t);
+
+/*
+ valloc(size_t n);
+ Equivalent to memalign(pagesize, n), where pagesize is the page
+ size of the system. If the pagesize is unknown, 4096 is used.
+*/
+void* dlvalloc(size_t);
+
+/*
+ mallopt(int parameter_number, int parameter_value)
+ Sets tunable parameters The format is to provide a
+ (parameter-number, parameter-value) pair. mallopt then sets the
+ corresponding parameter to the argument value if it can (i.e., so
+ long as the value is meaningful), and returns 1 if successful else
+ 0. To workaround the fact that mallopt is specified to use int,
+ not size_t parameters, the value -1 is specially treated as the
+ maximum unsigned size_t value.
+
+ SVID/XPG/ANSI defines four standard param numbers for mallopt,
+ normally defined in malloc.h. None of these are use in this malloc,
+ so setting them has no effect. But this malloc also supports other
+ options in mallopt. See below for details. Briefly, supported
+ parameters are as follows (listed defaults are for "typical"
+ configurations).
+
+ Symbol param # default allowed param values
+ M_TRIM_THRESHOLD -1 2*1024*1024 any (-1 disables)
+ M_GRANULARITY -2 page size any power of 2 >= page size
+ M_MMAP_THRESHOLD -3 256*1024 any (or 0 if no MMAP support)
+*/
+int dlmallopt(int, int);
+
+/*
+ malloc_footprint();
+ Returns the number of bytes obtained from the system. The total
+ number of bytes allocated by malloc, realloc etc., is less than this
+ value. Unlike mallinfo, this function returns only a precomputed
+ result, so can be called frequently to monitor memory consumption.
+ Even if locks are otherwise defined, this function does not use them,
+ so results might not be up to date.
+*/
+size_t dlmalloc_footprint(void);
+
+/*
+ malloc_max_footprint();
+ Returns the maximum number of bytes obtained from the system. This
+ value will be greater than current footprint if deallocated space
+ has been reclaimed by the system. The peak number of bytes allocated
+ by malloc, realloc etc., is less than this value. Unlike mallinfo,
+ this function returns only a precomputed result, so can be called
+ frequently to monitor memory consumption. Even if locks are
+ otherwise defined, this function does not use them, so results might
+ not be up to date.
+*/
+size_t dlmalloc_max_footprint(void);
+
+#if !NO_MALLINFO
+/*
+ mallinfo()
+ Returns (by copy) a struct containing various summary statistics:
+
+ arena: current total non-mmapped bytes allocated from system
+ ordblks: the number of free chunks
+ smblks: always zero.
+ hblks: current number of mmapped regions
+ hblkhd: total bytes held in mmapped regions
+ usmblks: the maximum total allocated space. This will be greater
+ than current total if trimming has occurred.
+ fsmblks: always zero
+ uordblks: current total allocated space (normal or mmapped)
+ fordblks: total free space
+ keepcost: the maximum number of bytes that could ideally be released
+ back to system via malloc_trim. ("ideally" means that
+ it ignores page restrictions etc.)
+
+ Because these fields are ints, but internal bookkeeping may
+ be kept as longs, the reported values may wrap around zero and
+ thus be inaccurate.
+*/
+struct mallinfo dlmallinfo(void);
+#endif /* NO_MALLINFO */
+
+/*
+ independent_calloc(size_t n_elements, size_t element_size, void* chunks[]);
+
+ independent_calloc is similar to calloc, but instead of returning a
+ single cleared space, it returns an array of pointers to n_elements
+ independent elements that can hold contents of size elem_size, each
+ of which starts out cleared, and can be independently freed,
+ realloc'ed etc. The elements are guaranteed to be adjacently
+ allocated (this is not guaranteed to occur with multiple callocs or
+ mallocs), which may also improve cache locality in some
+ applications.
+
+ The "chunks" argument is optional (i.e., may be null, which is
+ probably the most typical usage). If it is null, the returned array
+ is itself dynamically allocated and should also be freed when it is
+ no longer needed. Otherwise, the chunks array must be of at least
+ n_elements in length. It is filled in with the pointers to the
+ chunks.
+
+ In either case, independent_calloc returns this pointer array, or
+ null if the allocation failed. If n_elements is zero and "chunks"
+ is null, it returns a chunk representing an array with zero elements
+ (which should be freed if not wanted).
+
+ Each element must be individually freed when it is no longer
+ needed. If you'd like to instead be able to free all at once, you
+ should instead use regular calloc and assign pointers into this
+ space to represent elements. (In this case though, you cannot
+ independently free elements.)
+
+ independent_calloc simplifies and speeds up implementations of many
+ kinds of pools. It may also be useful when constructing large data
+ structures that initially have a fixed number of fixed-sized nodes,
+ but the number is not known at compile time, and some of the nodes
+ may later need to be freed. For example:
+
+ struct Node { int item; struct Node* next; };
+
+ struct Node* build_list() {
+ struct Node** pool;
+ int n = read_number_of_nodes_needed();
+ if (n <= 0) return 0;
+ pool = (struct Node**)(independent_calloc(n, sizeof(struct Node), 0);
+ if (pool == 0) die();
+ // organize into a linked list...
+ struct Node* first = pool[0];
+ for (i = 0; i < n-1; ++i)
+ pool[i]->next = pool[i+1];
+ free(pool); // Can now free the array (or not, if it is needed later)
+ return first;
+ }
+*/
+void** dlindependent_calloc(size_t, size_t, void**);
+
+/*
+ independent_comalloc(size_t n_elements, size_t sizes[], void* chunks[]);
+
+ independent_comalloc allocates, all at once, a set of n_elements
+ chunks with sizes indicated in the "sizes" array. It returns
+ an array of pointers to these elements, each of which can be
+ independently freed, realloc'ed etc. The elements are guaranteed to
+ be adjacently allocated (this is not guaranteed to occur with
+ multiple callocs or mallocs), which may also improve cache locality
+ in some applications.
+
+ The "chunks" argument is optional (i.e., may be null). If it is null
+ the returned array is itself dynamically allocated and should also
+ be freed when it is no longer needed. Otherwise, the chunks array
+ must be of at least n_elements in length. It is filled in with the
+ pointers to the chunks.
+
+ In either case, independent_comalloc returns this pointer array, or
+ null if the allocation failed. If n_elements is zero and chunks is
+ null, it returns a chunk representing an array with zero elements
+ (which should be freed if not wanted).
+
+ Each element must be individually freed when it is no longer
+ needed. If you'd like to instead be able to free all at once, you
+ should instead use a single regular malloc, and assign pointers at
+ particular offsets in the aggregate space. (In this case though, you
+ cannot independently free elements.)
+
+ independent_comallac differs from independent_calloc in that each
+ element may have a different size, and also that it does not
+ automatically clear elements.
+
+ independent_comalloc can be used to speed up allocation in cases
+ where several structs or objects must always be allocated at the
+ same time. For example:
+
+ struct Head { ... }
+ struct Foot { ... }
+
+ void send_message(char* msg) {
+ int msglen = strlen(msg);
+ size_t sizes[3] = { sizeof(struct Head), msglen, sizeof(struct Foot) };
+ void* chunks[3];
+ if (independent_comalloc(3, sizes, chunks) == 0)
+ die();
+ struct Head* head = (struct Head*)(chunks[0]);
+ char* body = (char*)(chunks[1]);
+ struct Foot* foot = (struct Foot*)(chunks[2]);
+ // ...
+ }
+
+ In general though, independent_comalloc is worth using only for
+ larger values of n_elements. For small values, you probably won't
+ detect enough difference from series of malloc calls to bother.
+
+ Overuse of independent_comalloc can increase overall memory usage,
+ since it cannot reuse existing noncontiguous small chunks that
+ might be available for some of the elements.
+*/
+void** dlindependent_comalloc(size_t, size_t*, void**);
+
+
+/*
+ pvalloc(size_t n);
+ Equivalent to valloc(minimum-page-that-holds(n)), that is,
+ round up n to nearest pagesize.
+ */
+void* dlpvalloc(size_t);
+
+/*
+ malloc_trim(size_t pad);
+
+ If possible, gives memory back to the system (via negative arguments
+ to sbrk) if there is unused memory at the `high' end of the malloc
+ pool or in unused MMAP segments. You can call this after freeing
+ large blocks of memory to potentially reduce the system-level memory
+ requirements of a program. However, it cannot guarantee to reduce
+ memory. Under some allocation patterns, some large free blocks of
+ memory will be locked between two used chunks, so they cannot be
+ given back to the system.
+
+ The `pad' argument to malloc_trim represents the amount of free
+ trailing space to leave untrimmed. If this argument is zero, only
+ the minimum amount of memory to maintain internal data structures
+ will be left. Non-zero arguments can be supplied to maintain enough
+ trailing space to service future expected allocations without having
+ to re-obtain memory from the system.
+
+ Malloc_trim returns 1 if it actually released any memory, else 0.
+*/
+int dlmalloc_trim(size_t);
+
+/*
+ malloc_stats();
+ Prints on stderr the amount of space obtained from the system (both
+ via sbrk and mmap), the maximum amount (which may be more than
+ current if malloc_trim and/or munmap got called), and the current
+ number of bytes allocated via malloc (or realloc, etc) but not yet
+ freed. Note that this is the number of bytes allocated, not the
+ number requested. It will be larger than the number requested
+ because of alignment and bookkeeping overhead. Because it includes
+ alignment wastage as being in use, this figure may be greater than
+ zero even when no user-level chunks are allocated.
+
+ The reported current and maximum system memory can be inaccurate if
+ a program makes other calls to system memory allocation functions
+ (normally sbrk) outside of malloc.
+
+ malloc_stats prints only the most commonly interesting statistics.
+ More information can be obtained by calling mallinfo.
+*/
+void dlmalloc_stats(void);
+
+#endif /* ONLY_MSPACES */
+
+/*
+ malloc_usable_size(void* p);
+
+ Returns the number of bytes you can actually use in
+ an allocated chunk, which may be more than you requested (although
+ often not) due to alignment and minimum size constraints.
+ You can use this many bytes without worrying about
+ overwriting other allocated objects. This is not a particularly great
+ programming practice. malloc_usable_size can be more useful in
+ debugging and assertions, for example:
+
+ p = malloc(n);
+ assert(malloc_usable_size(p) >= 256);
+*/
+size_t dlmalloc_usable_size(void*);
+
+
+#if MSPACES
+
+/*
+ mspace is an opaque type representing an independent
+ region of space that supports mspace_malloc, etc.
+*/
+typedef void* mspace;
+
+/*
+ create_mspace creates and returns a new independent space with the
+ given initial capacity, or, if 0, the default granularity size. It
+ returns null if there is no system memory available to create the
+ space. If argument locked is non-zero, the space uses a separate
+ lock to control access. The capacity of the space will grow
+ dynamically as needed to service mspace_malloc requests. You can
+ control the sizes of incremental increases of this space by
+ compiling with a different DEFAULT_GRANULARITY or dynamically
+ setting with mallopt(M_GRANULARITY, value).
+*/
+mspace create_mspace(size_t capacity, int locked);
+
+/*
+ destroy_mspace destroys the given space, and attempts to return all
+ of its memory back to the system, returning the total number of
+ bytes freed. After destruction, the results of access to all memory
+ used by the space become undefined.
+*/
+size_t destroy_mspace(mspace msp);
+
+/*
+ create_mspace_with_base uses the memory supplied as the initial base
+ of a new mspace. Part (less than 128*sizeof(size_t) bytes) of this
+ space is used for bookkeeping, so the capacity must be at least this
+ large. (Otherwise 0 is returned.) When this initial space is
+ exhausted, additional memory will be obtained from the system.
+ Destroying this space will deallocate all additionally allocated
+ space (if possible) but not the initial base.
+*/
+mspace create_mspace_with_base(void* base, size_t capacity, int locked);
+
+/*
+ mspace_track_large_chunks controls whether requests for large chunks
+ are allocated in their own untracked mmapped regions, separate from
+ others in this mspace. By default large chunks are not tracked,
+ which reduces fragmentation. However, such chunks are not
+ necessarily released to the system upon destroy_mspace. Enabling
+ tracking by setting to true may increase fragmentation, but avoids
+ leakage when relying on destroy_mspace to release all memory
+ allocated using this space. The function returns the previous
+ setting.
+*/
+int mspace_track_large_chunks(mspace msp, int enable);
+
+
+/*
+ mspace_malloc behaves as malloc, but operates within
+ the given space.
+*/
+void* mspace_malloc(mspace msp, size_t bytes);
+
+/*
+ mspace_free behaves as free, but operates within
+ the given space.
+
+ If compiled with FOOTERS==1, mspace_free is not actually needed.
+ free may be called instead of mspace_free because freed chunks from
+ any space are handled by their originating spaces.
+*/
+void mspace_free(mspace msp, void* mem);
+
+/*
+ mspace_realloc behaves as realloc, but operates within
+ the given space.
+
+ If compiled with FOOTERS==1, mspace_realloc is not actually
+ needed. realloc may be called instead of mspace_realloc because
+ realloced chunks from any space are handled by their originating
+ spaces.
+*/
+void* mspace_realloc(mspace msp, void* mem, size_t newsize);
+
+/*
+ mspace_calloc behaves as calloc, but operates within
+ the given space.
+*/
+void* mspace_calloc(mspace msp, size_t n_elements, size_t elem_size);
+
+/*
+ mspace_memalign behaves as memalign, but operates within
+ the given space.
+*/
+void* mspace_memalign(mspace msp, size_t alignment, size_t bytes);
+
+/*
+ mspace_independent_calloc behaves as independent_calloc, but
+ operates within the given space.
+*/
+void** mspace_independent_calloc(mspace msp, size_t n_elements,
+ size_t elem_size, void* chunks[]);
+
+/*
+ mspace_independent_comalloc behaves as independent_comalloc, but
+ operates within the given space.
+*/
+void** mspace_independent_comalloc(mspace msp, size_t n_elements,
+ size_t sizes[], void* chunks[]);
+
+/*
+ mspace_malloc2 behaves as mspace_malloc, but provides additional
+ functionality. Setting alignment to a non-zero value is
+ identical to using mspace_memalign(). Flags may be set to:
+
+ * M2_ZERO_MEMORY: Sets the contents of the allocated chunk to
+ zero.
+ * M2_ALWAYS_MMAP: Always allocate as though mmap_threshold
+ were being exceeded. This is useful for large
+ arrays which frequently extend.
+ * M2_RESERVE_MULT(n): Reserve n times as much address space such
+ that mmapped realloc() is much faster.
+ * M2_RESERVE_SHIFT(n): Reserve (1<<n) bytes of address space such
+ that mmapped realloc() is much faster.
+
+ Note when setting RESERVE sizes that on some platforms (e.g. Windows)
+ page tables are constructed for the reservation size. On x86/x64
+ Windows this costs 2Kb of kernel memory per Mb reserved, and as on
+ x86 kernel memory is not abundant you should not be excessive.
+*/
+void* mspace_malloc2(mspace msp, size_t bytes, size_t alignment, unsigned flags);
+
+/*
+ mspace_realloc2 behaves as mspace_realloc, but provides additional
+ functionality. Setting alignment to a non-zero value is
+ identical to using mspace_memalign(). Flags may be set to:
+
+ * M2_ZERO_MEMORY: Sets any increase in the allocated chunk to
+ zero. Note that this zeroes only the increase
+ from what dlmalloc thinks the chunk's size is,
+ so if you didn't use this flag when allocating
+ with malloc2 (which zeroes up to chunk size)
+ then you may have garbage just before the new
+ space.
+ * M2_PREVENT_MOVE: Prevent moves in realloc2() which is very
+ useful for C++ container objects.
+ * M2_ALWAYS_MMAP: Always allocate as though mmap_threshold
+ were being exceeded. Note that setting this
+ bit will not necessarily mmap a chunk which
+ isn't already mmapped, but it will force a
+ mmapped chunk if new memory needs allocating.
+ * M2_RESERVE_MULT(n): Reserve n times as much address space such
+ that mmapped realloc() is much faster.
+ * M2_RESERVE_SHIFT(n): Reserve (1<<n) bytes of address space such
+ that mmapped realloc() is much faster.
+
+ Note when setting RESERVE sizes that on some platforms (e.g. Windows)
+ page tables are constructed for the reservation size. On x86/x64
+ Windows this costs 2Kb of kernel memory per Mb reserved, and as on
+ x86 kernel memory is not abundant you should not be excessive.
+ With regard to M2_RESERVE_*, these only take effect when the
+ mmapped chunk has exceeded its reservation space and a new
+ reservation space needs to be created.
+*/
+void* mspace_realloc2(mspace msp, void* mem, size_t newsize, size_t alignment, unsigned flags);
+
+/*
+ mspace_footprint() returns the number of bytes obtained from the
+ system for this space.
+*/
+size_t mspace_footprint(mspace msp);
+
+/*
+ mspace_max_footprint() returns the peak number of bytes obtained from the
+ system for this space.
+*/
+size_t mspace_max_footprint(mspace msp);
+
+
+#if !NO_MALLINFO
+/*
+ mspace_mallinfo behaves as mallinfo, but reports properties of
+ the given space.
+*/
+struct mallinfo mspace_mallinfo(mspace msp);
+#endif /* NO_MALLINFO */
+
+/*
+ malloc_usable_size(void* p) behaves the same as malloc_usable_size;
+*/
+ size_t mspace_usable_size(void* mem);
+
+/*
+ mspace_malloc_stats behaves as malloc_stats, but reports
+ properties of the given space.
+*/
+void mspace_malloc_stats(mspace msp);
+
+/*
+ mspace_trim behaves as malloc_trim, but
+ operates within the given space.
+*/
+int mspace_trim(mspace msp, size_t pad);
+
+/*
+ An alias for mallopt.
+*/
+int mspace_mallopt(int, int);
+
+#endif /* MSPACES */
+
+#ifdef __cplusplus
+} /* end of extern "C" */
+#endif /* __cplusplus */
+
+/*
+ ========================================================================
+ To make a fully customizable malloc.h header file, cut everything
+ above this line, put into file malloc.h, edit to suit, and #include it
+ on the next line, as well as in programs that use this malloc.
+ ========================================================================
+*/
+
+/* #include "malloc.h" */
+
+/*------------------------------ internal #includes ---------------------- */
+
+#ifdef _MSC_VER
+#pragma warning( disable : 4146 ) /* no "unsigned" warnings */
+#endif /* WIN32 */
+
+#include <stdio.h> /* for printing in malloc_stats */
+
+#ifndef LACKS_ERRNO_H
+#include <errno.h> /* for MALLOC_FAILURE_ACTION */
+#endif /* LACKS_ERRNO_H */
+#if FOOTERS || DEBUG
+#include <time.h> /* for magic initialization */
+#endif /* FOOTERS */
+#ifndef LACKS_STDLIB_H
+#include <stdlib.h> /* for abort() */
+#endif /* LACKS_STDLIB_H */
+#ifdef DEBUG
+#if ABORT_ON_ASSERT_FAILURE
+#undef assert
+#define assert(x) if(!(x)) ABORT
+#else /* ABORT_ON_ASSERT_FAILURE */
+#include <assert.h>
+#endif /* ABORT_ON_ASSERT_FAILURE */
+#else /* DEBUG */
+#ifndef assert
+#define assert(x)
+#endif
+#define DEBUG 0
+#endif /* DEBUG */
+#ifndef LACKS_STRING_H
+#include <string.h> /* for memset etc */
+#endif /* LACKS_STRING_H */
+#if USE_BUILTIN_FFS
+#ifndef LACKS_STRINGS_H
+#include <strings.h> /* for ffs */
+#endif /* LACKS_STRINGS_H */
+#endif /* USE_BUILTIN_FFS */
+#if HAVE_MMAP
+#ifndef LACKS_SYS_MMAN_H
+/* On some versions of linux, mremap decl in mman.h needs __USE_GNU set */
+#if (defined(linux) && !defined(__USE_GNU))
+#define __USE_GNU 1
+#include <sys/mman.h> /* for mmap */
+#undef __USE_GNU
+#else
+#include <sys/mman.h> /* for mmap */
+#endif /* linux */
+#endif /* LACKS_SYS_MMAN_H */
+#ifndef LACKS_FCNTL_H
+#include <fcntl.h>
+#endif /* LACKS_FCNTL_H */
+#endif /* HAVE_MMAP */
+#ifndef LACKS_UNISTD_H
+#include <unistd.h> /* for sbrk, sysconf */
+#else /* LACKS_UNISTD_H */
+#if !defined(__FreeBSD__) && !defined(__OpenBSD__) && !defined(__NetBSD__)
+extern void* sbrk(ptrdiff_t);
+#endif /* FreeBSD etc */
+#endif /* LACKS_UNISTD_H */
+
+/* Declarations for locking */
+#if USE_LOCKS
+#ifndef WIN32
+#include <pthread.h>
+#if defined (__SVR4) && defined (__sun) /* solaris */
+#include <thread.h>
+#endif /* solaris */
+#elif defined(_MSC_VER)
+#ifndef _M_AMD64
+/* These are already defined on AMD64 builds */
+#ifdef __cplusplus
+extern "C" {
+#endif /* __cplusplus */
+LONG __cdecl _InterlockedCompareExchange(LONG volatile *Dest, LONG Exchange, LONG Comp);
+LONG __cdecl _InterlockedExchange(LONG volatile *Target, LONG Value);
+#ifdef __cplusplus
+}
+#endif /* __cplusplus */
+#endif /* _M_AMD64 */
+#pragma intrinsic (_InterlockedCompareExchange)
+#pragma intrinsic (_InterlockedExchange)
+#define interlockedcompareexchange _InterlockedCompareExchange
+#define interlockedexchange _InterlockedExchange
+#elif defined(WIN32) && defined(__GNUC__)
+/* MinGW or something like it */
+#define interlockedcompareexchange(a, b, c) __sync_val_compare_and_swap(a, c, b)
+#define interlockedexchange __sync_lock_test_and_set
+#endif /* Win32 */
+#endif /* USE_LOCKS */
+
+/* Declarations for bit scanning on win32 */
+#if defined(_MSC_VER) && _MSC_VER>=1300
+#ifndef BitScanForward /* Try to avoid pulling in WinNT.h */
+#ifdef __cplusplus
+extern "C" {
+#endif /* __cplusplus */
+unsigned char _BitScanForward(unsigned long *index, unsigned long mask);
+unsigned char _BitScanReverse(unsigned long *index, unsigned long mask);
+#ifdef __cplusplus
+}
+#endif /* __cplusplus */
+
+#define BitScanForward _BitScanForward
+#define BitScanReverse _BitScanReverse
+#pragma intrinsic(_BitScanForward)
+#pragma intrinsic(_BitScanReverse)
+#endif /* BitScanForward */
+#endif /* defined(_MSC_VER) && _MSC_VER>=1300 */
+
+#ifndef WIN32
+#ifndef malloc_getpagesize
+# ifdef _SC_PAGESIZE /* some SVR4 systems omit an underscore */
+# ifndef _SC_PAGE_SIZE
+# define _SC_PAGE_SIZE _SC_PAGESIZE
+# endif
+# endif
+# ifdef _SC_PAGE_SIZE
+# define malloc_getpagesize sysconf(_SC_PAGE_SIZE)
+# else
+# if defined(BSD) || defined(DGUX) || defined(HAVE_GETPAGESIZE)
+ extern size_t getpagesize();
+# define malloc_getpagesize getpagesize()
+# else
+# ifdef WIN32 /* use supplied emulation of getpagesize */
+# define malloc_getpagesize getpagesize()
+# else
+# ifndef LACKS_SYS_PARAM_H
+# include <sys/param.h>
+# endif
+# ifdef EXEC_PAGESIZE
+# define malloc_getpagesize EXEC_PAGESIZE
+# else
+# ifdef NBPG
+# ifndef CLSIZE
+# define malloc_getpagesize NBPG
+# else
+# define malloc_getpagesize (NBPG * CLSIZE)
+# endif
+# else
+# ifdef NBPC
+# define malloc_getpagesize NBPC
+# else
+# ifdef PAGESIZE
+# define malloc_getpagesize PAGESIZE
+# else /* just guess */
+# define malloc_getpagesize ((size_t)4096U)
+# endif
+# endif
+# endif
+# endif
+# endif
+# endif
+# endif
+#endif
+#endif
+
+
+
+/* ------------------- size_t and alignment properties -------------------- */
+
+/* The byte and bit size of a size_t */
+#define SIZE_T_SIZE (sizeof(size_t))
+#define SIZE_T_BITSIZE (sizeof(size_t) << 3)
+
+/* Some constants coerced to size_t */
+/* Annoying but necessary to avoid errors on some platforms */
+#define SIZE_T_ZERO ((size_t)0)
+#define SIZE_T_ONE ((size_t)1)
+#define SIZE_T_TWO ((size_t)2)
+#define SIZE_T_FOUR ((size_t)4)
+#define TWO_SIZE_T_SIZES (SIZE_T_SIZE<<1)
+#define THREE_SIZE_T_SIZES (TWO_SIZE_T_SIZES+SIZE_T_SIZE)
+#define FOUR_SIZE_T_SIZES (SIZE_T_SIZE<<2)
+#define SEVEN_SIZE_T_SIZES (FOUR_SIZE_T_SIZES+TWO_SIZE_T_SIZES+SIZE_T_SIZE)
+#define HALF_MAX_SIZE_T (MAX_SIZE_T / 2U)
+
+/* The bit mask value corresponding to MALLOC_ALIGNMENT */
+#define CHUNK_ALIGN_MASK (MALLOC_ALIGNMENT - SIZE_T_ONE)
+
+/* True if address a has acceptable alignment */
+#define is_aligned(A) (((size_t)((A)) & (CHUNK_ALIGN_MASK)) == 0)
+
+/* the number of bytes to offset an address to align it */
+#define align_offset(A)\
+ ((((size_t)(A) & CHUNK_ALIGN_MASK) == 0)? 0 :\
+ ((MALLOC_ALIGNMENT - ((size_t)(A) & CHUNK_ALIGN_MASK)) & CHUNK_ALIGN_MASK))
+
+/*
+ malloc_params holds global properties, including those that can be
+ dynamically set using mallopt. There is a single instance, mparams,
+ initialized in init_mparams. Note that the non-zeroness of "magic"
+ also serves as an initialization flag.
+*/
+typedef unsigned int flag_t;
+struct malloc_params {
+ volatile size_t magic;
+ size_t page_size;
+ size_t granularity;
+ size_t mmap_threshold;
+ size_t trim_threshold;
+ flag_t default_mflags;
+};
+
+static struct malloc_params mparams;
+
+/* Ensure mparams initialized */
+#define ensure_initialization() (void)(mparams.magic != 0 || init_mparams())
+
+/* -------------------------- MMAP preliminaries ------------------------- */
+
+/*
+ If HAVE_MORECORE or HAVE_MMAP are false, we just define calls and
+ checks to fail so compiler optimizer can delete code rather than
+ using so many "#if"s.
+*/
+
+
+/* MORECORE and MMAP must return MFAIL on failure */
+#define MFAIL ((void*)(MAX_SIZE_T))
+#define CMFAIL ((char*)(MFAIL)) /* defined for convenience */
+
+#if HAVE_MMAP
+
+#ifndef MREMAP_MAYMOVE
+#define MREMAP_MAYMOVE 1
+#endif /* MREMAP_MAYMOVE */
+
+#ifdef ENABLE_LARGE_PAGES
+static size_t largepagesize = 0;
+#endif /* ENABLE_LARGE_PAGES */
+static size_t mmapped_granularity;
+#ifndef WIN32
+#if !defined(MAP_ANONYMOUS) && defined(MAP_ANON)
+#define MAP_ANONYMOUS MAP_ANON
+#endif /* MAP_ANON */
+#define MMAP_PROT (PROT_READ|PROT_WRITE)
+#ifdef MAP_ANONYMOUS
+#define MMAP_FLAGS (MAP_PRIVATE|MAP_ANONYMOUS)
+#else /* MAP_ANONYMOUS */
+#define MMAP_FLAGS (MAP_PRIVATE)
+static int dev_zero_fd = -1; /* Cached file descriptor for /dev/zero. */
+#endif /* MAP_ANONYMOUS */
+
+static FORCEINLINE void* posix_mmap(size_t size) {
+ void* baseaddress = 0;
+ void* ptr = MFAIL;
+ int flags = MMAP_FLAGS, fd = -1;
+#ifndef MAP_ANONYMOUS
+ if (dev_zero_fd < 0)
+ dev_zero_fd = open("/dev/zero", O_RDWR);
+ fd = dev_zero_fd;
+#endif
+#ifdef ENABLE_LARGE_PAGES
+#ifdef MAP_LARGEPAGE
+#define MMAP_FLAGS_LARGEPAGE MAP_LARGEPAGE
+#elif defined(MAP_HUGETLB)
+#define MMAP_FLAGS_LARGEPAGE MAP_HUGETLB
+#else
+#error Cannot figure out how to enable large page support for this system!
+#endif
+ if(largepagesize && size >= largepagesize && !(size & (largepagesize-1)))
+ ptr = mmap(baseaddress, size, MMAP_PROT, flags|MMAP_FLAGS_LARGEPAGE, fd, 0);
+#endif
+ if (MFAIL==ptr) {
+ ptr = mmap(baseaddress, size, MMAP_PROT, flags, fd, 0);
+ }
+#if DEBUG && 0
+ printf("mmap returns %p size %u\n", ptr, (unsigned)size);
+#endif
+ return ptr;
+}
+
+/* For direct MMAP, use MAP_GROWSDOWN (linux)|MAP_STACK (bsd) to minimize interference */
+static FORCEINLINE void* posix_direct_mmap(size_t size) {
+ void* ptr = 0;
+ int flags = MMAP_FLAGS, fd = -1;
+#ifndef MAP_ANONYMOUS
+ if (dev_zero_fd < 0)
+ dev_zero_fd = open("/dev/zero", O_RDWR);
+ fd = dev_zero_fd;
+#endif
+#ifdef MAP_GROWSDOWN
+ flags |= MAP_GROWSDOWN;
+#elif defined(MAP_STACK)
+ flags |= MAP_STACK;
+#else
+#warning Cannot figure out how to request memory from the top of the address space!
+#endif
+ ptr = mmap(0, size, MMAP_PROT, flags, fd, 0);
+#if DEBUG && 0
+ printf("Direct mmap returns %p size %u\n", ptr, (unsigned)size);
+#endif
+ return ptr;
+}
+
+#define MMAP_DEFAULT(s) posix_mmap(s)
+#define MUNMAP_DEFAULT(h, a, s) munmap((a), (s))
+#define DIRECT_MMAP_DEFAULT(h, s, f) posix_direct_mmap(s)
+#if HAVE_MREMAP
+#define MREMAP_DEFAULT(addr, osz, nsz, mv) mremap((addr), (osz), (nsz), (mv))
+#define DIRECT_MREMAP_DEFAULT(h, addr, osz, nsz, mv, f) mremap((addr), (osz), (nsz), (mv))
+#endif /* HAVE_MREMAP */
+
+#else /* WIN32 */
+
+/* Win32 MMAP via VirtualAlloc */
+#ifdef DEFAULT_GRANULARITY_ALIGNED
+static void* lastWin32mmap; /* Used as a hint */
+#endif /* DEFAULT_GRANULARITY_ALIGNED */
+static FORCEINLINE void* win32mmap(size_t size) {
+ void* baseaddress = 0;
+ void* ptr = 0;
+#ifdef ENABLE_LARGE_PAGES
+ /* Note that large pages are *always* allocated on a large page boundary */
+ if(largepagesize && size >= largepagesize && !(size & (largepagesize-1))) {
+ ptr = VirtualAlloc(baseaddress, size, MEM_RESERVE|MEM_COMMIT|MEM_LARGE_PAGES, PAGE_READWRITE);
+ if(!ptr) {
+ if (ERROR_PRIVILEGE_NOT_HELD==GetLastError()) {
+ /*fprintf(stderr, "nedmalloc: Failed to allocate large memory pages (does the user running this process have the right to lock pages in memory?). Large pages will not be used.\n");*/
+ OutputDebugStringA("nedmalloc: Failed to allocate large memory pages (does the user running this process have the right to lock pages in memory?). Large pages will not be used.\n");
+ CreateEvent(NULL, FALSE, FALSE, __T("LargePagesDisabled"));
+ largepagesize=0;
+ }
+ }
+ }
+#endif
+ if(!ptr) {
+#ifdef DEFAULT_GRANULARITY_ALIGNED
+ /* We try to avoid overhead by speculatively reserving at aligned
+ addresses until we succeed */
+ void* originalbaseaddress;
+ baseaddress = originalbaseaddress = lastWin32mmap;
+ for(;;) {
+ void* reserveaddr = VirtualAlloc(baseaddress, size, MEM_RESERVE, PAGE_READWRITE);
+ if (!reserveaddr)
+ baseaddress = (void*)((size_t)baseaddress + mparams.granularity);
+ else if ((size_t)reserveaddr & (mparams.granularity - SIZE_T_ONE)) {
+ VirtualFree(reserveaddr, 0, MEM_RELEASE);
+ baseaddress = (void*)(((size_t)reserveaddr + mparams.granularity) & ~(mparams.granularity - SIZE_T_ONE));
+ }
+ else break;
+ if (baseaddress == originalbaseaddress) /* If this wraps then we are out of address space */
+ return MFAIL;
+ }
+#endif
+ if (!ptr) ptr = VirtualAlloc(baseaddress, size, baseaddress ? MEM_COMMIT : MEM_RESERVE|MEM_COMMIT, PAGE_READWRITE);
+#ifdef DEFAULT_GRANULARITY_ALIGNED
+#if DEBUG
+ if (lastWin32mmap && ptr!=lastWin32mmap) printf("Non-contiguous VirtualAlloc between %p and %p\n", ptr, lastWin32mmap);
+#endif
+ if (ptr) lastWin32mmap = (void*)((size_t) ptr + mparams.granularity);
+#endif
+ }
+#if DEBUG && 0
+#ifdef ENABLE_LARGE_PAGES
+ printf("VirtualAlloc returns %p size %u. LargePagesAvailable=%d\n", ptr, size, largepagesavailable);
+#else
+ printf("VirtualAlloc returns %p size %u\n", ptr, size);
+#endif
+#endif
+ return (ptr != 0)? ptr: MFAIL;
+}
+
+/* For direct MMAP, we have two allocation methods, one which is
+unresizeable and the other which is resizeable. The resizeable
+method is slower to allocate and free, but is much quicker to
+reallocate. Benchmarking shows that the crossover is at around
+512Kb-1Mb, so if you are reallocating blocks 1Mb or above then you
+ought to specify one of the M2_RESERVE_* flags which can specify
+a multiple of the allocation size or a fixed two power size.
+
+The mremap() Linux call is then effectively emulated by
+reserving the address space specified by M2_RESERVE_* and committing
+or decommitting memory as the block is resized. If the block
+exceeds its reservation size, the mremap() emulation fails and
+dlmalloc will allocate a new block, copy its contents and delete
+the old one.
+
+************ The WIN32_DIRECT_USE_FILE_MAPPINGS option ************
+Reserving excessive address space is not an issue on 64 bit
+machines, but on 32 bit one can very quickly expend 2Gb. A
+solution is to use a win32 file mapping of the swap file as the
+reservation and then to map views of differing sizes which exactly
+is what mremap() does on Linux.
+
+Unfortunately the file mapping method is benchmarked at around
+40-45% slower than using VirtualAlloc to over-reserve. It is still
+50% faster for avrg. 2Mb block sizes than no mremap() at all, so
+we turn it on on 32 bit and disable it on 64 bit.
+*/
+#ifndef WIN32_DIRECT_USE_FILE_MAPPINGS
+#if defined(_M_IA64) || defined(_M_X64) || defined(WIN64)
+#define WIN32_DIRECT_USE_FILE_MAPPINGS 0
+#else
+#define WIN32_DIRECT_USE_FILE_MAPPINGS 1
+#endif
+#endif
+
+static FORCEINLINE void* win32direct_mmap(void **handle, size_t size, unsigned flags) {
+ void* ptr = 0;
+ unsigned mremapvalue = (flags & M2_RESERVE_MASK)>>8;
+#if 0
+ mremapvalue=4;
+ flags|=M2_RESERVE_ISMULTIPLIER;
+#endif
+#if 0
+ mremapvalue=22;/*4Mb*/
+#endif
+ if (!mremapvalue) {
+ ptr = VirtualAlloc(0, size, MEM_RESERVE|MEM_TOP_DOWN|MEM_COMMIT, PAGE_READWRITE);
+ }
+ else {
+ size_t reservesize = (flags & M2_RESERVE_ISMULTIPLIER) ? size*mremapvalue : SIZE_T_ONE<<mremapvalue;
+#if WIN32_DIRECT_USE_FILE_MAPPINGS
+ HANDLE fmh;
+ if (reservesize < size)
+ reservesize = size;
+ fmh = CreateFileMapping(INVALID_HANDLE_VALUE, NULL, PAGE_READWRITE|SEC_RESERVE,
+#if defined(_M_IA64) || defined(_M_X64) || defined(WIN64)
+ (DWORD)(reservesize>>32),
+#else
+ 0,
+#endif
+ (DWORD)(reservesize&((DWORD)-1)), NULL);
+ if (!fmh)
+ return MFAIL;
+ *handle = (void*)fmh;
+ ptr = MapViewOfFile(fmh, FILE_MAP_ALL_ACCESS, 0, 0, size);
+ if (ptr)
+ ptr = VirtualAlloc(ptr, size, MEM_COMMIT, PAGE_READWRITE);
+ if (!ptr) {
+ CloseHandle(fmh);
+ return MFAIL;
+ }
+#else
+ void* ptr2;
+ /* If on 32 bit, cap reservation to 1Gb */
+ if (sizeof(size_t) == 4 && reservesize > 1*1024*1024*1024)
+ reservesize = 1*1024*1024*1024;
+ ptr = VirtualAlloc(0, reservesize, MEM_RESERVE|MEM_TOP_DOWN, PAGE_READWRITE);
+ if (ptr) {
+ ptr2 = VirtualAlloc(ptr, size, MEM_COMMIT, PAGE_READWRITE);
+ if (ptr2)
+ *handle = (void*)reservesize;
+ else
+ VirtualFree(ptr, 0, MEM_RELEASE);
+ }
+#endif
+ }
+#if DEBUG && 0
+ printf("VirtualAlloc returns %p size %u\n", ptr, size);
+#endif
+ return (ptr != 0)? ptr: MFAIL;
+}
+
+/* Implementation of mremap for direct MMAP */
+static FORCEINLINE void* win32direct_mremap(void **handle, void *ptr, size_t oldsize, size_t newsize, int flags, unsigned flags2) {
+ void* newptr = 0;
+ if (!*handle)
+ return MFAIL; /* We only resize file mappings reserved with M2_RESERVE_* */
+ if (newsize == oldsize)
+ return ptr;
+ {
+#if WIN32_DIRECT_USE_FILE_MAPPINGS
+ HANDLE fmh = (HANDLE) *handle;
+ /* It is VERY important to map a new view of a file mapping before
+ unmapping the old view. Otherwise the NT kernel will start writing
+ your file mapping in its entirety to the swap file which is otherwise
+ avoidable. Hence fail if can't move. */
+ if (!(flags & MREMAP_MAYMOVE))
+ return MFAIL;
+ newptr = MapViewOfFile(fmh, FILE_MAP_ALL_ACCESS, 0, 0, newsize);
+ if (newptr && newsize > oldsize)
+ newptr = VirtualAlloc(newptr, newsize, MEM_COMMIT, PAGE_READWRITE);
+ if (!newptr)
+ return MFAIL;
+ UnmapViewOfFile(ptr);
+#else
+ size_t reservesize = (size_t) *handle;
+ if (newsize > reservesize)
+ return MFAIL;
+ if (newsize > oldsize) {
+ if (VirtualAlloc((char*)ptr + oldsize, newsize - oldsize, MEM_COMMIT, PAGE_READWRITE) == 0)
+ return MFAIL;
+ }
+ else {
+ if (VirtualFree((char*)ptr + newsize, oldsize - newsize, MEM_DECOMMIT) == 0)
+ return MFAIL;
+ }
+ newptr = ptr;
+#endif
+ }
+#if DEBUG && 0
+ printf("win32remap(%p, %u, %u, %d) returns %p!\n", ptr, oldsize, newsize, flags, newptr);
+#endif
+ return newptr;
+}
+
+/* This function supports releasing coalesed segments */
+static FORCEINLINE int win32munmap(void *handle, void* ptr, size_t size) {
+ if (!handle) {
+ MEMORY_BASIC_INFORMATION minfo;
+ char* cptr = (char*)ptr;
+ while (size) {
+ if (VirtualQuery(cptr, &minfo, sizeof(minfo)) == 0)
+ return -1;
+ if (minfo.BaseAddress != cptr || minfo.AllocationBase != cptr ||
+ minfo.State != MEM_COMMIT || minfo.RegionSize > size)
+ return -1;
+ if (VirtualFree(cptr, 0, MEM_RELEASE) == 0)
+ return -1;
+ cptr += minfo.RegionSize;
+ size -= minfo.RegionSize;
+ }
+ }
+ else {
+#if WIN32_DIRECT_USE_FILE_MAPPINGS
+ HANDLE fmh = (HANDLE) handle;
+ /* As noted above in win32direct_mremap(), it is VERY important
+ to destroy the file mapping object before unmapping views as so
+ to avoid causing the system to write the file mapping to swap. */
+ if (CloseHandle(fmh) == 0)
+ return -1;
+ if (UnmapViewOfFile(ptr) == 0)
+ return -1;
+#else
+ /* We know it's contiguous, so avoid the VirtualQuery() */
+ if (VirtualFree(ptr, 0, MEM_RELEASE) == 0)
+ return -1;
+#endif
+ }
+ return 0;
+}
+
+#define MMAP_DEFAULT(s) win32mmap(s)
+#define MUNMAP_DEFAULT(h, a, s) win32munmap((h), (a), (s))
+#define DIRECT_MMAP_DEFAULT(h, s, f) win32direct_mmap((h), (s), (f))
+#define DIRECT_MREMAP_DEFAULT(h, a, os, ns, f, f2) win32direct_mremap((h), (a), (os), (ns), (f), (f2))
+#endif /* WIN32 */
+
+#ifndef MREMAP_DEFAULT
+#define MREMAP_DEFAULT(addr, osz, nsz, mv) MFAIL
+#endif /* MREMAP_DEFAULT */
+#ifndef DIRECT_MREMAP_DEFAULT
+#define DIRECT_MREMAP_DEFAULT(h, addr, osz, nsz, mv, f2) MFAIL
+#endif /* DIRECT_MREMAP_DEFAULT */
+
+#endif /* HAVE_MMAP */
+
+
+/**
+ * Define CALL_MORECORE
+ */
+#if HAVE_MORECORE
+ #ifdef MORECORE
+ #define CALL_MORECORE(S) MORECORE(S)
+ #else /* MORECORE */
+ #define CALL_MORECORE(S) MORECORE_DEFAULT(S)
+ #endif /* MORECORE */
+#else /* HAVE_MORECORE */
+ #define CALL_MORECORE(S) MFAIL
+#endif /* HAVE_MORECORE */
+
+/**
+ * Define CALL_MMAP/CALL_MUNMAP/CALL_DIRECT_MMAP
+ */
+#if HAVE_MMAP
+ #define USE_MMAP_BIT (SIZE_T_ONE)
+
+ #ifdef MUNMAP
+ #define CALL_MUNMAP(h, a, s) MUNMAP((h), (a), (s))
+ #else /* MUNMAP */
+ #define CALL_MUNMAP(h, a, s) MUNMAP_DEFAULT((h), (a), (s))
+ #endif /* MUNMAP */
+ #ifdef MMAP
+ #define CALL_MMAP(s, f) MMAP((s), (f))
+ #else /* MMAP */
+ #define CALL_MMAP(s, f) MMAP_DEFAULT(s)
+ #endif /* MMAP */
+ #ifdef MREMAP
+ #define CALL_MREMAP(a, os, ns, f) MREMAP((a), (os), (ns), (f))
+ #else /* MREMAP */
+ #define CALL_MREMAP(a, os, ns, f) MREMAP_DEFAULT((a), (os), (ns), (f))
+ #endif /* MREMAP */
+
+ #ifdef DIRECT_MMAP
+ #define CALL_DIRECT_MMAP(h, s, f) DIRECT_MMAP((h), (s), (f))
+ #else /* DIRECT_MMAP */
+ #define CALL_DIRECT_MMAP(h, s, f) DIRECT_MMAP_DEFAULT((h), (s), (f))
+ #endif /* DIRECT_MMAP */
+ #ifdef DIRECT_MREMAP
+ #define CALL_DIRECT_MREMAP(h, a, os, ns, f, f2) DIRECT_MREMAP((h), (a), (os), (ns), (f), (f2))
+ #else /* DIRECT_MMAP */
+ #define CALL_DIRECT_MREMAP(h, a, os, ns, f, f2) DIRECT_MREMAP_DEFAULT((h), (a), (os), (ns), (f), (f2))
+ #endif /* DIRECT_MMAP */
+#else /* HAVE_MMAP */
+ #define USE_MMAP_BIT (SIZE_T_ZERO)
+
+ #define MUNMAP(h, a, s) (-1)
+ #define MMAP(s, f) MFAIL
+ #define MREMAP(a, os, ns, f) MFAIL
+ #define DIRECT_MMAP(h, s, f) MFAIL
+ #define DIRECT_MREMAP(h, a, os, ns, f, f2) MFAIL
+
+ #define CALL_MUNMAP(h, a, s) MUNMAP((h), (a), (s))
+ #define CALL_MMAP(s, f) MMAP((s), (f))
+ #define CALL_MREMAP(a, os, ns, f) MREMAP((a), (os), (ns), (f))
+ #define CALL_DIRECT_MMAP(h, s, f) DIRECT_MMAP((h), (s), (f))
+ #define CALL_DIRECT_MREMAP(h, a, os, ns, f, f2) DIRECT_MREMAP((h), (a), (os), (ns), (f), (f2))
+#endif /* HAVE_MMAP */
+
+/* mstate bit set if continguous morecore disabled or failed */
+#define USE_NONCONTIGUOUS_BIT (4U)
+
+/* segment bit set in create_mspace_with_base */
+#define EXTERN_BIT (8U)
+
+
+/* --------------------------- Lock preliminaries ------------------------ */
+
+/*
+ When locks are defined, there is one global lock, plus
+ one per-mspace lock.
+
+ The global lock_ensures that mparams.magic and other unique
+ mparams values are initialized only once. It also protects
+ sequences of calls to MORECORE. In many cases sys_alloc requires
+ two calls, that should not be interleaved with calls by other
+ threads. This does not protect against direct calls to MORECORE
+ by other threads not using this lock, so there is still code to
+ cope the best we can on interference.
+
+ Per-mspace locks surround calls to malloc, free, etc. To enable use
+ in layered extensions, per-mspace locks are reentrant.
+
+ Because lock-protected regions generally have bounded times, it is
+ OK to use the supplied simple spinlocks. Spinlocks are likely to
+ improve performance for lightly contended applications, but worsen
+ performance under heavy contention.
+
+ If USE_LOCKS is > 1, the definitions of lock routines here are
+ bypassed, in which case you will need to define the type MLOCK_T,
+ and at least INITIAL_LOCK, DESTROY_LOCK, ACQUIRE_LOCK, RELEASE_LOCK
+ and possibly TRY_LOCK (which is not used in this malloc, but commonly
+ needed in extensions.) You must also declare a
+ static MLOCK_T malloc_global_mutex = { initialization values };.
+
+*/
+
+#if USE_LOCKS == 1
+
+#if USE_SPIN_LOCKS && SPIN_LOCKS_AVAILABLE
+#ifndef WIN32
+
+/* Custom pthread-style spin locks on x86 and x64 for gcc */
+struct pthread_mlock_t {
+ volatile unsigned int l;
+ char cachelinepadding[64];
+ unsigned int c;
+ pthread_t threadid;
+};
+#define MLOCK_T struct pthread_mlock_t
+#define CURRENT_THREAD pthread_self()
+#define INITIAL_LOCK(sl) ((sl)->threadid = 0, (sl)->l = (sl)->c = 0, 0)
+#define DESTROY_LOCK(sl) (0)
+#define ACQUIRE_LOCK(sl) pthread_acquire_lock(sl)
+#define RELEASE_LOCK(sl) pthread_release_lock(sl)
+#define TRY_LOCK(sl) pthread_try_lock(sl)
+#define SPINS_PER_YIELD 63
+
+static MLOCK_T malloc_global_mutex = { 0, "", 0, 0};
+
+static FORCEINLINE int pthread_acquire_lock (MLOCK_T *sl) {
+ pthread_t mythreadid = CURRENT_THREAD;
+ int spins = 0;
+ volatile unsigned int* lp = &sl->l;
+ for (;;) {
+ if (*lp != 0) {
+ if (sl->threadid == mythreadid) {
+ ++sl->c;
+ return 0;
+ }
+ }
+ else {
+ /* place args to cmpxchgl in locals to evade oddities in some gccs */
+ int cmp = 0;
+ int val = 1;
+ int ret;
+ ret = __sync_bool_compare_and_swap(lp, cmp, val);
+ if (ret) {
+ assert(!sl->threadid);
+ sl->threadid = mythreadid;
+ sl->c = 1;
+ return 0;
+ }
+ }
+ if ((++spins & SPINS_PER_YIELD) == 0) {
+#if defined (__SVR4) && defined (__sun) /* solaris */
+ thr_yield();
+#else
+#if defined(__linux__) || defined(__FreeBSD__) || defined(__APPLE__)
+ sched_yield();
+#else /* no-op yield on unknown systems */
+ ;
+#endif /* __linux__ || __FreeBSD__ || __APPLE__ */
+#endif /* solaris */
+ }
+ }
+}
+
+static FORCEINLINE void pthread_release_lock (MLOCK_T *sl) {
+ volatile unsigned int* lp = &sl->l;
+ assert(*lp != 0);
+ assert(sl->threadid == CURRENT_THREAD);
+ if (--sl->c == 0) {
+ sl->threadid = 0;
+ __sync_lock_release(lp, 0);
+ }
+}
+
+static FORCEINLINE int pthread_try_lock (MLOCK_T *sl) {
+ pthread_t mythreadid = CURRENT_THREAD;
+ volatile unsigned int* lp = &sl->l;
+ if (*lp != 0) {
+ if (sl->threadid == mythreadid) {
+ ++sl->c;
+ return 1;
+ }
+ }
+ else {
+ int cmp = 0;
+ int val = 1;
+ int ret;
+ ret=__sync_bool_compare_and_swap(lp, cmp, val);
+ if (ret) {
+ assert(!sl->threadid);
+ sl->threadid = mythreadid;
+ sl->c = 1;
+ return 1;
+ }
+ }
+ return 0;
+}
+
+
+#else /* WIN32 */
+/* Custom win32-style spin locks on x86 and x64 for MSC */
+struct win32_mlock_t {
+ volatile long l;
+ char cachelinepadding[64];
+ unsigned int c;
+ long threadid;
+};
+
+#define MLOCK_T struct win32_mlock_t
+#define CURRENT_THREAD ((long)GetCurrentThreadId())
+#define INITIAL_LOCK(sl) ((sl)->threadid = 0, (sl)->l = (sl)->c = 0, 0)
+#define DESTROY_LOCK(sl) (0)
+#define ACQUIRE_LOCK(sl) win32_acquire_lock(sl)
+#define RELEASE_LOCK(sl) win32_release_lock(sl)
+#define TRY_LOCK(sl) win32_try_lock(sl)
+#define SPINS_PER_YIELD 63
+
+static MLOCK_T malloc_global_mutex = {0, "", 0, 0};
+
+static FORCEINLINE int win32_acquire_lock (MLOCK_T *sl) {
+ long mythreadid = CURRENT_THREAD;
+ int spins = 0;
+ for (;;) {
+ if (sl->l != 0) {
+ if (sl->threadid == mythreadid) {
+ ++sl->c;
+ return 0;
+ }
+ }
+ else {
+ if (!interlockedexchange(&sl->l, 1)) {
+ assert(!sl->threadid);
+ sl->threadid = mythreadid;
+ sl->c = 1;
+ return 0;
+ }
+ }
+ if ((++spins & SPINS_PER_YIELD) == 0)
+ SleepEx(0, FALSE);
+ }
+}
+
+static FORCEINLINE void win32_release_lock (MLOCK_T *sl) {
+ assert(sl->threadid == CURRENT_THREAD);
+ assert(sl->l != 0);
+ if (--sl->c == 0) {
+ sl->threadid = 0;
+ interlockedexchange (&sl->l, 0);
+ }
+}
+
+static FORCEINLINE int win32_try_lock (MLOCK_T *sl) {
+ long mythreadid = CURRENT_THREAD;
+ if (sl->l != 0) {
+ if (sl->threadid == mythreadid) {
+ ++sl->c;
+ return 1;
+ }
+ }
+ else {
+ if (!interlockedexchange(&sl->l, 1)){
+ assert(!sl->threadid);
+ sl->threadid = mythreadid;
+ sl->c = 1;
+ return 1;
+ }
+ }
+ return 0;
+}
+
+#endif /* WIN32 */
+#else /* USE_SPIN_LOCKS */
+
+#ifndef WIN32
+/* pthreads-based locks */
+
+#define MLOCK_T pthread_mutex_t
+#define CURRENT_THREAD pthread_self()
+#define INITIAL_LOCK(sl) pthread_init_lock(sl)
+#define DESTROY_LOCK(sl) pthread_mutex_destroy(sl)
+#define ACQUIRE_LOCK(sl) pthread_mutex_lock(sl)
+#define RELEASE_LOCK(sl) pthread_mutex_unlock(sl)
+#define TRY_LOCK(sl) (!pthread_mutex_trylock(sl))
+
+static MLOCK_T malloc_global_mutex = PTHREAD_MUTEX_INITIALIZER;
+
+/* Cope with old-style linux recursive lock initialization by adding */
+/* skipped internal declaration from pthread.h */
+#ifdef linux
+#ifndef PTHREAD_MUTEX_RECURSIVE
+extern "C" int pthread_mutexattr_setkind_np __P ((pthread_mutexattr_t *__attr,
+ int __kind));
+#define PTHREAD_MUTEX_RECURSIVE PTHREAD_MUTEX_RECURSIVE_NP
+#define pthread_mutexattr_settype(x,y) pthread_mutexattr_setkind_np(x,y)
+#endif
+#endif
+
+static int pthread_init_lock (MLOCK_T *sl) {
+ pthread_mutexattr_t attr;
+ if (pthread_mutexattr_init(&attr)) return 1;
+ if (pthread_mutexattr_settype(&attr, PTHREAD_MUTEX_RECURSIVE)) return 1;
+ if (pthread_mutex_init(sl, &attr)) return 1;
+ if (pthread_mutexattr_destroy(&attr)) return 1;
+ return 0;
+}
+
+#else /* WIN32 */
+/* Win32 critical sections */
+#define MLOCK_T CRITICAL_SECTION
+#define CURRENT_THREAD GetCurrentThreadId()
+#define INITIAL_LOCK(sl) (!InitializeCriticalSectionAndSpinCount((sl), 0x80000000|4000))
+#define DESTROY_LOCK(sl) (DeleteCriticalSection(sl), 0)
+#define ACQUIRE_LOCK(sl) (EnterCriticalSection(sl), 0)
+#define RELEASE_LOCK(sl) LeaveCriticalSection(sl)
+#define TRY_LOCK(sl) TryEnterCriticalSection(sl)
+#define NEED_GLOBAL_LOCK_INIT
+
+static MLOCK_T malloc_global_mutex;
+static volatile long malloc_global_mutex_status;
+
+/* Use spin loop to initialize global lock */
+static void init_malloc_global_mutex() {
+ for (;;) {
+ long stat = malloc_global_mutex_status;
+ if (stat > 0)
+ return;
+ /* transition to < 0 while initializing, then to > 0) */
+ if (stat == 0 &&
+ interlockedcompareexchange(&malloc_global_mutex_status, -1, 0) == 0) {
+ InitializeCriticalSection(&malloc_global_mutex);
+ interlockedexchange(&malloc_global_mutex_status,1);
+ return;
+ }
+ SleepEx(0, FALSE);
+ }
+}
+
+#endif /* WIN32 */
+#endif /* USE_SPIN_LOCKS */
+#endif /* USE_LOCKS == 1 */
+
+/* ----------------------- User-defined locks ------------------------ */
+
+#if USE_LOCKS > 1
+/* Define your own lock implementation here */
+/* #define INITIAL_LOCK(sl) ... */
+/* #define DESTROY_LOCK(sl) ... */
+/* #define ACQUIRE_LOCK(sl) ... */
+/* #define RELEASE_LOCK(sl) ... */
+/* #define TRY_LOCK(sl) ... */
+/* static MLOCK_T malloc_global_mutex = ... */
+#endif /* USE_LOCKS > 1 */
+
+/* ----------------------- Lock-based state ------------------------ */
+
+#if USE_LOCKS
+#define USE_LOCK_BIT (2U)
+#else /* USE_LOCKS */
+#define USE_LOCK_BIT (0U)
+#define INITIAL_LOCK(l)
+#define DESTROY_LOCK(l)
+#endif /* USE_LOCKS */
+
+#if USE_LOCKS
+#ifndef ACQUIRE_MALLOC_GLOBAL_LOCK
+#define ACQUIRE_MALLOC_GLOBAL_LOCK() ACQUIRE_LOCK(&malloc_global_mutex);
+#endif
+#ifndef RELEASE_MALLOC_GLOBAL_LOCK
+#define RELEASE_MALLOC_GLOBAL_LOCK() RELEASE_LOCK(&malloc_global_mutex);
+#endif
+#else /* USE_LOCKS */
+#define ACQUIRE_MALLOC_GLOBAL_LOCK()
+#define RELEASE_MALLOC_GLOBAL_LOCK()
+#endif /* USE_LOCKS */
+
+
+/* ----------------------- Chunk representations ------------------------ */
+
+/*
+ (The following includes lightly edited explanations by Colin Plumb.)
+
+ The malloc_chunk declaration below is misleading (but accurate and
+ necessary). It declares a "view" into memory allowing access to
+ necessary fields at known offsets from a given base.
+
+ Chunks of memory are maintained using a `boundary tag' method as
+ originally described by Knuth. (See the paper by Paul Wilson
+ ftp://ftp.cs.utexas.edu/pub/garbage/allocsrv.ps for a survey of such
+ techniques.) Sizes of free chunks are stored both in the front of
+ each chunk and at the end. This makes consolidating fragmented
+ chunks into bigger chunks fast. The head fields also hold bits
+ representing whether chunks are free or in use.
+
+ Here are some pictures to make it clearer. They are "exploded" to
+ show that the state of a chunk can be thought of as extending from
+ the high 31 bits of the head field of its header through the
+ prev_foot and PINUSE_BIT bit of the following chunk header.
+
+ A chunk that's in use looks like:
+
+ chunk-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ | Size of previous chunk (if P = 0) |
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |P|
+ | Size of this chunk 1| +-+
+ mem-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ | |
+ +- -+
+ | |
+ +- -+
+ | :
+ +- size - sizeof(size_t) available payload bytes -+
+ : |
+ chunk-> +- -+
+ | |
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |1|
+ | Size of next chunk (may or may not be in use) | +-+
+ mem-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+
+ And if it's free, it looks like this:
+
+ chunk-> +- -+
+ | User payload (must be in use, or we would have merged!) |
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |P|
+ | Size of this chunk 0| +-+
+ mem-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ | Next pointer |
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ | Prev pointer |
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ | :
+ +- size - sizeof(struct chunk) unused bytes -+
+ : |
+ chunk-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ | Size of this chunk |
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0|
+ | Size of next chunk (must be in use, or we would have merged)| +-+
+ mem-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ | :
+ +- User payload -+
+ : |
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ |0|
+ +-+
+ Note that since we always merge adjacent free chunks, the chunks
+ adjacent to a free chunk must be in use.
+
+ Given a pointer to a chunk (which can be derived trivially from the
+ payload pointer) we can, in O(1) time, find out whether the adjacent
+ chunks are free, and if so, unlink them from the lists that they
+ are on and merge them with the current chunk.
+
+ Chunks always begin on even word boundaries, so the mem portion
+ (which is returned to the user) is also on an even word boundary, and
+ thus at least double-word aligned.
+
+ The P (PINUSE_BIT) bit, stored in the unused low-order bit of the
+ chunk size (which is always a multiple of two words), is an in-use
+ bit for the *previous* chunk. If that bit is *clear*, then the
+ word before the current chunk size contains the previous chunk
+ size, and can be used to find the front of the previous chunk.
+ The very first chunk allocated always has this bit set, preventing
+ access to non-existent (or non-owned) memory. If pinuse is set for
+ any given chunk, then you CANNOT determine the size of the
+ previous chunk, and might even get a memory addressing fault when
+ trying to do so.
+
+ The C (CINUSE_BIT) bit, stored in the unused second-lowest bit of
+ the chunk size redundantly records whether the current chunk is
+ inuse (unless the chunk is mmapped). This redundancy enables usage
+ checks within free and realloc, and reduces indirection when freeing
+ and consolidating chunks.
+
+ Each freshly allocated chunk must have both cinuse and pinuse set.
+ That is, each allocated chunk borders either a previously allocated
+ and still in-use chunk, or the base of its memory arena. This is
+ ensured by making all allocations from the the `lowest' part of any
+ found chunk. Further, no free chunk physically borders another one,
+ so each free chunk is known to be preceded and followed by either
+ inuse chunks or the ends of memory.
+
+ Note that the `foot' of the current chunk is actually represented
+ as the prev_foot of the NEXT chunk. This makes it easier to
+ deal with alignments etc but can be very confusing when trying
+ to extend or adapt this code.
+
+ The exceptions to all this are
+
+ 1. The special chunk `top' is the top-most available chunk (i.e.,
+ the one bordering the end of available memory). It is treated
+ specially. Top is never included in any bin, is used only if
+ no other chunk is available, and is released back to the
+ system if it is very large (see M_TRIM_THRESHOLD). In effect,
+ the top chunk is treated as larger (and thus less well
+ fitting) than any other available chunk. The top chunk
+ doesn't update its trailing size field since there is no next
+ contiguous chunk that would have to index off it. However,
+ space is still allocated for it (TOP_FOOT_SIZE) to enable
+ separation or merging when space is extended.
+
+ 3. Chunks allocated via mmap, have both cinuse and pinuse bits
+ cleared in their head fields. Because they are allocated
+ one-by-one, each must carry its own prev_foot field, which is
+ also used to hold the offset this chunk has within its mmapped
+ region, which is needed to preserve alignment. Each mmapped
+ chunk is trailed by the first two fields of a fake next-chunk
+ for sake of usage checks.
+
+*/
+
+struct malloc_chunk {
+ size_t prev_foot; /* Size of previous chunk (if free). */
+ size_t head; /* Size and inuse bits. */
+ struct malloc_chunk* fd; /* double links -- used only if free. */
+ struct malloc_chunk* bk;
+};
+
+typedef struct malloc_chunk mchunk;
+typedef struct malloc_chunk* mchunkptr;
+typedef struct malloc_chunk* sbinptr; /* The type of bins of chunks */
+typedef unsigned int bindex_t; /* Described below */
+typedef unsigned int binmap_t; /* Described below */
+
+/* ------------------- Chunks sizes and alignments ----------------------- */
+
+#define MCHUNK_SIZE (sizeof(mchunk))
+
+#if FOOTERS
+#define CHUNK_OVERHEAD (TWO_SIZE_T_SIZES)
+#else /* FOOTERS */
+#define CHUNK_OVERHEAD (SIZE_T_SIZE)
+#endif /* FOOTERS */
+
+/* MMapped chunks need three words of overhead, two for normal
+chunk and one for handle ... */
+#define MMAP_CHUNK_OVERHEAD (THREE_SIZE_T_SIZES)
+/* ... and additional padding for fake next-chunk at foot */
+#define MMAP_FOOT_PAD (FOUR_SIZE_T_SIZES)
+
+/* The smallest size we can malloc is an aligned minimal chunk */
+#define MIN_CHUNK_SIZE\
+ ((MCHUNK_SIZE + CHUNK_ALIGN_MASK) & ~CHUNK_ALIGN_MASK)
+
+/* conversion from malloc headers to user pointers, and back */
+#define chunk2mem(p) ((void*)((char*)(p) + TWO_SIZE_T_SIZES))
+#define mem2chunk(mem) ((mchunkptr)((char*)(mem) - TWO_SIZE_T_SIZES))
+/* chunk associated with aligned address A */
+#define align_as_chunk(A) (mchunkptr)((A) + align_offset(chunk2mem(A)))
+
+/* Bounds on request (not chunk) sizes. */
+#define MAX_REQUEST ((-MIN_CHUNK_SIZE) << 2)
+#define MIN_REQUEST (MIN_CHUNK_SIZE - CHUNK_OVERHEAD - SIZE_T_ONE)
+
+/* pad request bytes into a usable size */
+#define pad_request(req) \
+ (((req) + CHUNK_OVERHEAD + CHUNK_ALIGN_MASK) & ~CHUNK_ALIGN_MASK)
+
+/* pad request, checking for minimum (but not maximum) */
+#define request2size(req) \
+ (((req) < MIN_REQUEST)? MIN_CHUNK_SIZE : pad_request(req))
+
+
+/* ------------------ Operations on head and foot fields ----------------- */
+
+/*
+ The head field of a chunk is or'ed with PINUSE_BIT when previous
+ adjacent chunk in use, and or'ed with CINUSE_BIT if this chunk is in
+ use, unless mmapped, in which case both bits are cleared.
+
+ FLAG4_BIT is not used by this malloc, but might be useful in extensions.
+*/
+
+#define PINUSE_BIT (SIZE_T_ONE)
+#define CINUSE_BIT (SIZE_T_TWO)
+#define FLAG4_BIT (SIZE_T_FOUR)
+#define INUSE_BITS (PINUSE_BIT|CINUSE_BIT)
+#define FLAG_BITS (PINUSE_BIT|CINUSE_BIT|FLAG4_BIT)
+
+/* Head value for fenceposts */
+#define FENCEPOST_HEAD (INUSE_BITS|SIZE_T_SIZE)
+
+/* extraction of fields from head words */
+#define cinuse(p) ((p)->head & CINUSE_BIT)
+#define pinuse(p) ((p)->head & PINUSE_BIT)
+#define flag4inuse(p) ((p)->head & FLAG4_BIT)
+#define is_inuse(p) (((p)->head & INUSE_BITS) != PINUSE_BIT)
+#define is_mmapped(p) (((p)->head & INUSE_BITS) == 0)
+
+#define chunksize(p) ((p)->head & ~(FLAG_BITS))
+
+#define clear_pinuse(p) ((p)->head &= ~PINUSE_BIT)
+#define set_flag4(p) ((p)->head |= FLAG4_BIT)
+#define clear_flag4(p) ((p)->head &= ~FLAG4_BIT)
+
+/* Treat space at ptr +/- offset as a chunk */
+#define chunk_plus_offset(p, s) ((mchunkptr)(((char*)(p)) + (s)))
+#define chunk_minus_offset(p, s) ((mchunkptr)(((char*)(p)) - (s)))
+
+/* Ptr to next or previous physical malloc_chunk. */
+#define next_chunk(p) ((mchunkptr)( ((char*)(p)) + ((p)->head & ~FLAG_BITS)))
+#define prev_chunk(p) ((mchunkptr)( ((char*)(p)) - ((p)->prev_foot) ))
+
+/* extract next chunk's pinuse bit */
+#define next_pinuse(p) ((next_chunk(p)->head) & PINUSE_BIT)
+
+/* Get/set size at footer */
+#define get_foot(p, s) (((mchunkptr)((char*)(p) + (s)))->prev_foot)
+#define set_foot(p, s) (((mchunkptr)((char*)(p) + (s)))->prev_foot = (s))
+
+/* Set size, pinuse bit, and foot */
+#define set_size_and_pinuse_of_free_chunk(p, s)\
+ ((p)->head = (s|PINUSE_BIT), set_foot(p, s))
+
+/* Set size, pinuse bit, foot, and clear next pinuse */
+#define set_free_with_pinuse(p, s, n)\
+ (clear_pinuse(n), set_size_and_pinuse_of_free_chunk(p, s))
+
+/* Get the internal overhead associated with chunk p */
+#define overhead_for(p)\
+ (is_mmapped(p)? MMAP_CHUNK_OVERHEAD : CHUNK_OVERHEAD)
+
+/* Return true if malloced space is not necessarily cleared */
+#if MMAP_CLEARS
+#define calloc_must_clear(p) (!is_mmapped(p))
+#else /* MMAP_CLEARS */
+#define calloc_must_clear(p) (1)
+#endif /* MMAP_CLEARS */
+
+/* ---------------------- Overlaid data structures ----------------------- */
+
+/*
+ When chunks are not in use, they are treated as nodes of either
+ lists or trees.
+
+ "Small" chunks are stored in circular doubly-linked lists, and look
+ like this:
+
+ chunk-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ | Size of previous chunk |
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ `head:' | Size of chunk, in bytes |P|
+ mem-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ | Forward pointer to next chunk in list |
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ | Back pointer to previous chunk in list |
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ | Unused space (may be 0 bytes long) .
+ . .
+ . |
+nextchunk-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ `foot:' | Size of chunk, in bytes |
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+
+ Larger chunks are kept in a form of bitwise digital trees (aka
+ tries) keyed on chunksizes. Because malloc_tree_chunks are only for
+ free chunks greater than 256 bytes, their size doesn't impose any
+ constraints on user chunk sizes. Each node looks like:
+
+ chunk-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ | Size of previous chunk |
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ `head:' | Size of chunk, in bytes |P|
+ mem-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ | Forward pointer to next chunk of same size |
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ | Back pointer to previous chunk of same size |
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ | Pointer to left child (child[0]) |
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ | Pointer to right child (child[1]) |
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ | Pointer to parent |
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ | bin index of this chunk |
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ | Unused space .
+ . |
+nextchunk-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ `foot:' | Size of chunk, in bytes |
+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+
+ Each tree holding treenodes is a tree of unique chunk sizes. Chunks
+ of the same size are arranged in a circularly-linked list, with only
+ the oldest chunk (the next to be used, in our FIFO ordering)
+ actually in the tree. (Tree members are distinguished by a non-null
+ parent pointer.) If a chunk with the same size an an existing node
+ is inserted, it is linked off the existing node using pointers that
+ work in the same way as fd/bk pointers of small chunks.
+
+ Each tree contains a power of 2 sized range of chunk sizes (the
+ smallest is 0x100 <= x < 0x180), which is is divided in half at each
+ tree level, with the chunks in the smaller half of the range (0x100
+ <= x < 0x140 for the top nose) in the left subtree and the larger
+ half (0x140 <= x < 0x180) in the right subtree. This is, of course,
+ done by inspecting individual bits.
+
+ Using these rules, each node's left subtree contains all smaller
+ sizes than its right subtree. However, the node at the root of each
+ subtree has no particular ordering relationship to either. (The
+ dividing line between the subtree sizes is based on trie relation.)
+ If we remove the last chunk of a given size from the interior of the
+ tree, we need to replace it with a leaf node. The tree ordering
+ rules permit a node to be replaced by any leaf below it.
+
+ The smallest chunk in a tree (a common operation in a best-fit
+ allocator) can be found by walking a path to the leftmost leaf in
+ the tree. Unlike a usual binary tree, where we follow left child
+ pointers until we reach a null, here we follow the right child
+ pointer any time the left one is null, until we reach a leaf with
+ both child pointers null. The smallest chunk in the tree will be
+ somewhere along that path.
+
+ The worst case number of steps to add, find, or remove a node is
+ bounded by the number of bits differentiating chunks within
+ bins. Under current bin calculations, this ranges from 6 up to 21
+ (for 32 bit sizes) or up to 53 (for 64 bit sizes). The typical case
+ is of course much better.
+*/
+
+struct malloc_tree_chunk {
+ /* The first four fields must be compatible with malloc_chunk */
+ size_t prev_foot;
+ size_t head;
+ struct malloc_tree_chunk* fd;
+ struct malloc_tree_chunk* bk;
+
+ struct malloc_tree_chunk* child[2];
+ struct malloc_tree_chunk* parent;
+ bindex_t index;
+};
+
+typedef struct malloc_tree_chunk tchunk;
+typedef struct malloc_tree_chunk* tchunkptr;
+typedef struct malloc_tree_chunk* tbinptr; /* The type of bins of trees */
+
+/* A little helper macro for trees */
+#define leftmost_child(t) ((t)->child[0] != 0? (t)->child[0] : (t)->child[1])
+
+/* ----------------------------- Segments -------------------------------- */
+
+/*
+ Each malloc space may include non-contiguous segments, held in a
+ list headed by an embedded malloc_segment record representing the
+ top-most space. Segments also include flags holding properties of
+ the space. Large chunks that are directly allocated by mmap are not
+ included in this list. They are instead independently created and
+ destroyed without otherwise keeping track of them.
+
+ Segment management mainly comes into play for spaces allocated by
+ MMAP. Any call to MMAP might or might not return memory that is
+ adjacent to an existing segment. MORECORE normally contiguously
+ extends the current space, so this space is almost always adjacent,
+ which is simpler and faster to deal with. (This is why MORECORE is
+ used preferentially to MMAP when both are available -- see
+ sys_alloc.) When allocating using MMAP, we don't use any of the
+ hinting mechanisms (inconsistently) supported in various
+ implementations of unix mmap, or distinguish reserving from
+ committing memory. Instead, we just ask for space, and exploit
+ contiguity when we get it. It is probably possible to do
+ better than this on some systems, but no general scheme seems
+ to be significantly better.
+
+ Management entails a simpler variant of the consolidation scheme
+ used for chunks to reduce fragmentation -- new adjacent memory is
+ normally prepended or appended to an existing segment. However,
+ there are limitations compared to chunk consolidation that mostly
+ reflect the fact that segment processing is relatively infrequent
+ (occurring only when getting memory from system) and that we
+ don't expect to have huge numbers of segments:
+
+ * Segments are not indexed, so traversal requires linear scans. (It
+ would be possible to index these, but is not worth the extra
+ overhead and complexity for most programs on most platforms.)
+ * New segments are only appended to old ones when holding top-most
+ memory; if they cannot be prepended to others, they are held in
+ different segments.
+
+ Except for the top-most segment of an mstate, each segment record
+ is kept at the tail of its segment. Segments are added by pushing
+ segment records onto the list headed by &mstate.seg for the
+ containing mstate.
+
+ Segment flags control allocation/merge/deallocation policies:
+ * If EXTERN_BIT set, then we did not allocate this segment,
+ and so should not try to deallocate or merge with others.
+ (This currently holds only for the initial segment passed
+ into create_mspace_with_base.)
+ * If USE_MMAP_BIT set, the segment may be merged with
+ other surrounding mmapped segments and trimmed/de-allocated
+ using munmap.
+ * If neither bit is set, then the segment was obtained using
+ MORECORE so can be merged with surrounding MORECORE'd segments
+ and deallocated/trimmed using MORECORE with negative arguments.
+*/
+
+struct malloc_segment {
+ char* base; /* base address */
+ size_t size; /* allocated size */
+ struct malloc_segment* next; /* ptr to next segment */
+ flag_t sflags; /* mmap and extern flag */
+};
+
+#define is_mmapped_segment(S) ((S)->sflags & USE_MMAP_BIT)
+#define is_extern_segment(S) ((S)->sflags & EXTERN_BIT)
+
+typedef struct malloc_segment msegment;
+typedef struct malloc_segment* msegmentptr;
+
+/* ---------------------------- malloc_state ----------------------------- */
+
+/*
+ A malloc_state holds all of the bookkeeping for a space.
+ The main fields are:
+
+ Top
+ The topmost chunk of the currently active segment. Its size is
+ cached in topsize. The actual size of topmost space is
+ topsize+TOP_FOOT_SIZE, which includes space reserved for adding
+ fenceposts and segment records if necessary when getting more
+ space from the system. The size at which to autotrim top is
+ cached from mparams in trim_check, except that it is disabled if
+ an autotrim fails.
+
+ Designated victim (dv)
+ This is the preferred chunk for servicing small requests that
+ don't have exact fits. It is normally the chunk split off most
+ recently to service another small request. Its size is cached in
+ dvsize. The link fields of this chunk are not maintained since it
+ is not kept in a bin.
+
+ SmallBins
+ An array of bin headers for free chunks. These bins hold chunks
+ with sizes less than MIN_LARGE_SIZE bytes. Each bin contains
+ chunks of all the same size, spaced 8 bytes apart. To simplify
+ use in double-linked lists, each bin header acts as a malloc_chunk
+ pointing to the real first node, if it exists (else pointing to
+ itself). This avoids special-casing for headers. But to avoid
+ waste, we allocate only the fd/bk pointers of bins, and then use
+ repositioning tricks to treat these as the fields of a chunk.
+
+ TreeBins
+ Treebins are pointers to the roots of trees holding a range of
+ sizes. There are 2 equally spaced treebins for each power of two
+ from TREE_SHIFT to TREE_SHIFT+16. The last bin holds anything
+ larger.
+
+ Bin maps
+ There is one bit map for small bins ("smallmap") and one for
+ treebins ("treemap). Each bin sets its bit when non-empty, and
+ clears the bit when empty. Bit operations are then used to avoid
+ bin-by-bin searching -- nearly all "search" is done without ever
+ looking at bins that won't be selected. The bit maps
+ conservatively use 32 bits per map word, even if on 64bit system.
+ For a good description of some of the bit-based techniques used
+ here, see Henry S. Warren Jr's book "Hacker's Delight" (and
+ supplement at http://hackersdelight.org/). Many of these are
+ intended to reduce the branchiness of paths through malloc etc, as
+ well as to reduce the number of memory locations read or written.
+
+ Segments
+ A list of segments headed by an embedded malloc_segment record
+ representing the initial space.
+
+ Address check support
+ The least_addr field is the least address ever obtained from
+ MORECORE or MMAP. Attempted frees and reallocs of any address less
+ than this are trapped (unless INSECURE is defined).
+
+ Magic tag
+ A cross-check field that should always hold same value as mparams.magic.
+
+ Flags
+ Bits recording whether to use MMAP, locks, or contiguous MORECORE
+
+ Statistics
+ Each space keeps track of current and maximum system memory
+ obtained via MORECORE or MMAP.
+
+ Trim support
+ Fields holding the amount of unused topmost memory that should trigger
+ timming, and a counter to force periodic scanning to release unused
+ non-topmost segments.
+
+ Locking
+ If USE_LOCKS is defined, the "mutex" lock is acquired and released
+ around every public call using this mspace.
+
+ Extension support
+ A void* pointer and a size_t field that can be used to help implement
+ extensions to this malloc.
+*/
+
+/* Bin types, widths and sizes */
+#define NSMALLBINS (32U)
+#define NTREEBINS (32U)
+#define SMALLBIN_SHIFT (3U)
+#define SMALLBIN_WIDTH (SIZE_T_ONE << SMALLBIN_SHIFT)
+#define TREEBIN_SHIFT (8U)
+#define MIN_LARGE_SIZE (SIZE_T_ONE << TREEBIN_SHIFT)
+#define MAX_SMALL_SIZE (MIN_LARGE_SIZE - SIZE_T_ONE)
+#define MAX_SMALL_REQUEST (MAX_SMALL_SIZE - CHUNK_ALIGN_MASK - CHUNK_OVERHEAD)
+
+struct malloc_state {
+ binmap_t smallmap;
+ binmap_t treemap;
+ size_t dvsize;
+ size_t topsize;
+ char* least_addr;
+ mchunkptr dv;
+ mchunkptr top;
+ size_t trim_check;
+ size_t release_checks;
+ size_t magic;
+ mchunkptr smallbins[(NSMALLBINS+1)*2];
+ tbinptr treebins[NTREEBINS];
+ size_t footprint;
+ size_t max_footprint;
+ flag_t mflags;
+ msegment seg;
+#if USE_LOCKS
+ MLOCK_T mutex; /* locate lock among fields that rarely change */
+#endif /* USE_LOCKS */
+ void* extp; /* Unused but available for extensions */
+ size_t exts;
+};
+
+typedef struct malloc_state* mstate;
+
+/* ------------- Global malloc_state and malloc_params ------------------- */
+
+#if !ONLY_MSPACES
+
+/* The global malloc_state used for all non-"mspace" calls */
+static struct malloc_state _gm_;
+#define gm (&_gm_)
+#define is_global(M) ((M) == &_gm_)
+
+#endif /* !ONLY_MSPACES */
+
+#define is_initialized(M) ((M)->top != 0)
+
+/* -------------------------- system alloc setup ------------------------- */
+
+/* Operations on mflags */
+
+#define use_lock(M) ((M)->mflags & USE_LOCK_BIT)
+#define enable_lock(M) ((M)->mflags |= USE_LOCK_BIT)
+#define disable_lock(M) ((M)->mflags &= ~USE_LOCK_BIT)
+
+#define use_mmap(M) ((M)->mflags & USE_MMAP_BIT)
+#define enable_mmap(M) ((M)->mflags |= USE_MMAP_BIT)
+#define disable_mmap(M) ((M)->mflags &= ~USE_MMAP_BIT)
+
+#define use_noncontiguous(M) ((M)->mflags & USE_NONCONTIGUOUS_BIT)
+#define disable_contiguous(M) ((M)->mflags |= USE_NONCONTIGUOUS_BIT)
+
+#define set_lock(M,L)\
+ ((M)->mflags = (L)?\
+ ((M)->mflags | USE_LOCK_BIT) :\
+ ((M)->mflags & ~USE_LOCK_BIT))
+
+/* page-align a size */
+#define page_align(S)\
+ (((S) + (mparams.page_size - SIZE_T_ONE)) & ~(mparams.page_size - SIZE_T_ONE))
+
+/* granularity-align a size */
+#define granularity_align(S)\
+ (((S) + (mparams.granularity - SIZE_T_ONE))\
+ & ~(mparams.granularity - SIZE_T_ONE))
+
+
+/* For mmapped allocations align the size specially such that large
+page sizes don't cause excessive allocation wastage. */
+#define mmap_align_size(S) (((S) + (mmapped_granularity - SIZE_T_ONE)) & ~(mmapped_granularity - SIZE_T_ONE))
+
+/* For sys_alloc, enough padding to ensure can malloc request on success */
+#define SYS_ALLOC_PADDING (TOP_FOOT_SIZE + MALLOC_ALIGNMENT)
+
+#define is_page_aligned(S)\
+ (((size_t)(S) & (mparams.page_size - SIZE_T_ONE)) == 0)
+#define is_granularity_aligned(S)\
+ (((size_t)(S) & (mparams.granularity - SIZE_T_ONE)) == 0)
+
+/* True if segment S holds address A */
+#define segment_holds(S, A)\
+ ((char*)(A) >= S->base && (char*)(A) < S->base + S->size)
+
+/* Return segment holding given address */
+static msegmentptr segment_holding(mstate m, char* addr) {
+ msegmentptr sp = &m->seg;
+ for (;;) {
+ if (addr >= sp->base && addr < sp->base + sp->size)
+ return sp;
+ if ((sp = sp->next) == 0)
+ return 0;
+ }
+}
+
+/* Return true if segment contains a segment link */
+static int has_segment_link(mstate m, msegmentptr ss) {
+ msegmentptr sp = &m->seg;
+ for (;;) {
+ if ((char*)sp >= ss->base && (char*)sp < ss->base + ss->size)
+ return 1;
+ if ((sp = sp->next) == 0)
+ return 0;
+ }
+}
+
+#ifndef MORECORE_CANNOT_TRIM
+#define should_trim(M,s) ((s) > (M)->trim_check)
+#else /* MORECORE_CANNOT_TRIM */
+#define should_trim(M,s) (0)
+#endif /* MORECORE_CANNOT_TRIM */
+
+/*
+ TOP_FOOT_SIZE is padding at the end of a segment, including space
+ that may be needed to place segment records and fenceposts when new
+ noncontiguous segments are added.
+*/
+#define TOP_FOOT_SIZE\
+ (align_offset(chunk2mem(0))+pad_request(sizeof(struct malloc_segment))+MIN_CHUNK_SIZE)
+
+
+/* ------------------------------- Hooks -------------------------------- */
+
+/*
+ PREACTION should be defined to return 0 on success, and nonzero on
+ failure. If you are not using locking, you can redefine these to do
+ anything you like.
+*/
+
+#if USE_LOCKS
+
+#define PREACTION(M) ((use_lock(M))? ACQUIRE_LOCK(&(M)->mutex) : 0)
+#define POSTACTION(M) { if (use_lock(M)) RELEASE_LOCK(&(M)->mutex); }
+#else /* USE_LOCKS */
+
+#ifndef PREACTION
+#define PREACTION(M) (0)
+#endif /* PREACTION */
+
+#ifndef POSTACTION
+#define POSTACTION(M)
+#endif /* POSTACTION */
+
+#endif /* USE_LOCKS */
+
+/*
+ CORRUPTION_ERROR_ACTION is triggered upon detected bad addresses.
+ USAGE_ERROR_ACTION is triggered on detected bad frees and
+ reallocs. The argument p is an address that might have triggered the
+ fault. It is ignored by the two predefined actions, but might be
+ useful in custom actions that try to help diagnose errors.
+*/
+
+#if PROCEED_ON_ERROR
+
+/* A count of the number of corruption errors causing resets */
+int malloc_corruption_error_count;
+
+/* default corruption action */
+static void reset_on_error(mstate m);
+
+#define CORRUPTION_ERROR_ACTION(m) reset_on_error(m)
+#define USAGE_ERROR_ACTION(m, p)
+
+#else /* PROCEED_ON_ERROR */
+
+#ifndef CORRUPTION_ERROR_ACTION
+#define CORRUPTION_ERROR_ACTION(m) ABORT
+#endif /* CORRUPTION_ERROR_ACTION */
+
+#ifndef USAGE_ERROR_ACTION
+#define USAGE_ERROR_ACTION(m,p) ABORT
+#endif /* USAGE_ERROR_ACTION */
+
+#endif /* PROCEED_ON_ERROR */
+
+/* -------------------------- Debugging setup ---------------------------- */
+
+#if ! DEBUG
+
+#define check_free_chunk(M,P)
+#define check_inuse_chunk(M,P)
+#define check_malloced_chunk(M,P,N)
+#define check_mmapped_chunk(M,P)
+#define check_malloc_state(M)
+#define check_top_chunk(M,P)
+
+#else /* DEBUG */
+#define check_free_chunk(M,P) do_check_free_chunk(M,P)
+#define check_inuse_chunk(M,P) do_check_inuse_chunk(M,P)
+#define check_top_chunk(M,P) do_check_top_chunk(M,P)
+#define check_malloced_chunk(M,P,N) do_check_malloced_chunk(M,P,N)
+#define check_mmapped_chunk(M,P) do_check_mmapped_chunk(M,P)
+#define check_malloc_state(M) do_check_malloc_state(M)
+
+static void do_check_any_chunk(mstate m, mchunkptr p);
+static void do_check_top_chunk(mstate m, mchunkptr p);
+static void do_check_mmapped_chunk(mstate m, mchunkptr p);
+static void do_check_inuse_chunk(mstate m, mchunkptr p);
+static void do_check_free_chunk(mstate m, mchunkptr p);
+static void do_check_malloced_chunk(mstate m, void* mem, size_t s);
+static void do_check_tree(mstate m, tchunkptr t);
+static void do_check_treebin(mstate m, bindex_t i);
+static void do_check_smallbin(mstate m, bindex_t i);
+static void do_check_malloc_state(mstate m);
+static int bin_find(mstate m, mchunkptr x);
+static size_t traverse_and_check(mstate m);
+#endif /* DEBUG */
+
+/* ---------------------------- Indexing Bins ---------------------------- */
+
+#define is_small(s) (((s) >> SMALLBIN_SHIFT) < NSMALLBINS)
+#define small_index(s) (bindex_t)((s) >> SMALLBIN_SHIFT)
+#define small_index2size(i) ((i) << SMALLBIN_SHIFT)
+#define MIN_SMALL_INDEX (small_index(MIN_CHUNK_SIZE))
+
+/* addressing by index. See above about smallbin repositioning */
+#define smallbin_at(M, i) ((sbinptr)((char*)&((M)->smallbins[(i)<<1])))
+#define treebin_at(M,i) (&((M)->treebins[i]))
+
+/* assign tree index for size S to variable I. Use x86 asm if possible */
+#if defined(__GNUC__) && (defined(__i386__) || defined(__x86_64__))
+#define compute_tree_index(S, I)\
+{\
+ unsigned int X = S >> TREEBIN_SHIFT;\
+ if (X == 0)\
+ I = 0;\
+ else if (X > 0xFFFF)\
+ I = NTREEBINS-1;\
+ else {\
+ unsigned int K = (unsigned) sizeof(X)*__CHAR_BIT__ - 1 - (unsigned) __builtin_clz(X); \
+ I = (bindex_t)((K << 1) + ((S >> (K + (TREEBIN_SHIFT-1)) & 1)));\
+ }\
+}
+
+#elif defined (__INTEL_COMPILER)
+#define compute_tree_index(S, I)\
+{\
+ size_t X = S >> TREEBIN_SHIFT;\
+ if (X == 0)\
+ I = 0;\
+ else if (X > 0xFFFF)\
+ I = NTREEBINS-1;\
+ else {\
+ unsigned int K = _bit_scan_reverse (X); \
+ I = (bindex_t)((K << 1) + ((S >> (K + (TREEBIN_SHIFT-1)) & 1)));\
+ }\
+}
+
+#elif defined(_MSC_VER) && _MSC_VER>=1300
+#define compute_tree_index(S, I)\
+{\
+ size_t X = S >> TREEBIN_SHIFT;\
+ if (X == 0)\
+ I = 0;\
+ else if (X > 0xFFFF)\
+ I = NTREEBINS-1;\
+ else {\
+ unsigned int K;\
+ _BitScanReverse((DWORD *) &K, (DWORD) X);\
+ I = (bindex_t)((K << 1) + ((S >> (K + (TREEBIN_SHIFT-1)) & 1)));\
+ }\
+}
+
+#else /* GNUC */
+#define compute_tree_index(S, I)\
+{\
+ size_t X = S >> TREEBIN_SHIFT;\
+ if (X == 0)\
+ I = 0;\
+ else if (X > 0xFFFF)\
+ I = NTREEBINS-1;\
+ else {\
+ unsigned int Y = (unsigned int)X;\
+ unsigned int N = ((Y - 0x100) >> 16) & 8;\
+ unsigned int K = (((Y <<= N) - 0x1000) >> 16) & 4;\
+ N += K;\
+ N += K = (((Y <<= K) - 0x4000) >> 16) & 2;\
+ K = 14 - N + ((Y <<= K) >> 15);\
+ I = (K << 1) + ((S >> (K + (TREEBIN_SHIFT-1)) & 1));\
+ }\
+}
+#endif /* GNUC */
+
+/* Bit representing maximum resolved size in a treebin at i */
+#define bit_for_tree_index(i) \
+ (i == NTREEBINS-1)? (SIZE_T_BITSIZE-1) : (((i) >> 1) + TREEBIN_SHIFT - 2)
+
+/* Shift placing maximum resolved bit in a treebin at i as sign bit */
+#define leftshift_for_tree_index(i) \
+ ((i == NTREEBINS-1)? 0 : \
+ ((SIZE_T_BITSIZE-SIZE_T_ONE) - (((i) >> 1) + TREEBIN_SHIFT - 2)))
+
+/* The size of the smallest chunk held in bin with index i */
+#define minsize_for_tree_index(i) \
+ ((SIZE_T_ONE << (((i) >> 1) + TREEBIN_SHIFT)) | \
+ (((size_t)((i) & SIZE_T_ONE)) << (((i) >> 1) + TREEBIN_SHIFT - 1)))
+
+
+/* ------------------------ Operations on bin maps ----------------------- */
+
+/* bit corresponding to given index */
+#define idx2bit(i) ((binmap_t)(1) << (i))
+
+/* Mark/Clear bits with given index */
+#define mark_smallmap(M,i) ((M)->smallmap |= idx2bit(i))
+#define clear_smallmap(M,i) ((M)->smallmap &= ~idx2bit(i))
+#define smallmap_is_marked(M,i) ((M)->smallmap & idx2bit(i))
+
+#define mark_treemap(M,i) ((M)->treemap |= idx2bit(i))
+#define clear_treemap(M,i) ((M)->treemap &= ~idx2bit(i))
+#define treemap_is_marked(M,i) ((M)->treemap & idx2bit(i))
+
+/* isolate the least set bit of a bitmap */
+#define least_bit(x) ((x) & -(x))
+
+/* mask with all bits to left of least bit of x on */
+#define left_bits(x) ((x<<1) | -(x<<1))
+
+/* mask with all bits to left of or equal to least bit of x on */
+#define same_or_left_bits(x) ((x) | -(x))
+
+/* index corresponding to given bit. Use x86 asm if possible */
+
+#if defined(__GNUC__) && (defined(__i386__) || defined(__x86_64__))
+#define compute_bit2idx(X, I)\
+{\
+ unsigned int J;\
+ J = __builtin_ctz(X); \
+ I = (bindex_t)J;\
+}
+
+#elif defined (__INTEL_COMPILER)
+#define compute_bit2idx(X, I)\
+{\
+ unsigned int J;\
+ J = _bit_scan_forward (X); \
+ I = (bindex_t)J;\
+}
+
+#elif defined(_MSC_VER) && _MSC_VER>=1300
+#define compute_bit2idx(X, I)\
+{\
+ unsigned int J;\
+ _BitScanForward((DWORD *) &J, X);\
+ I = (bindex_t)J;\
+}
+
+#elif USE_BUILTIN_FFS
+#define compute_bit2idx(X, I) I = ffs(X)-1
+
+#else
+#define compute_bit2idx(X, I)\
+{\
+ unsigned int Y = X - 1;\
+ unsigned int K = Y >> (16-4) & 16;\
+ unsigned int N = K; Y >>= K;\
+ N += K = Y >> (8-3) & 8; Y >>= K;\
+ N += K = Y >> (4-2) & 4; Y >>= K;\
+ N += K = Y >> (2-1) & 2; Y >>= K;\
+ N += K = Y >> (1-0) & 1; Y >>= K;\
+ I = (bindex_t)(N + Y);\
+}
+#endif /* GNUC */
+
+
+/* ----------------------- Runtime Check Support ------------------------- */
+
+/*
+ For security, the main invariant is that malloc/free/etc never
+ writes to a static address other than malloc_state, unless static
+ malloc_state itself has been corrupted, which cannot occur via
+ malloc (because of these checks). In essence this means that we
+ believe all pointers, sizes, maps etc held in malloc_state, but
+ check all of those linked or offsetted from other embedded data
+ structures. These checks are interspersed with main code in a way
+ that tends to minimize their run-time cost.
+
+ When FOOTERS is defined, in addition to range checking, we also
+ verify footer fields of inuse chunks, which can be used guarantee
+ that the mstate controlling malloc/free is intact. This is a
+ streamlined version of the approach described by William Robertson
+ et al in "Run-time Detection of Heap-based Overflows" LISA'03
+ http://www.usenix.org/events/lisa03/tech/robertson.html The footer
+ of an inuse chunk holds the xor of its mstate and a random seed,
+ that is checked upon calls to free() and realloc(). This is
+ (probablistically) unguessable from outside the program, but can be
+ computed by any code successfully malloc'ing any chunk, so does not
+ itself provide protection against code that has already broken
+ security through some other means. Unlike Robertson et al, we
+ always dynamically check addresses of all offset chunks (previous,
+ next, etc). This turns out to be cheaper than relying on hashes.
+*/
+
+#if !INSECURE
+/* Check if address a is at least as high as any from MORECORE or MMAP */
+#define ok_address(M, a) ((char*)(a) >= (M)->least_addr)
+/* Check if address of next chunk n is higher than base chunk p */
+#define ok_next(p, n) ((char*)(p) < (char*)(n))
+/* Check if p has inuse status */
+#define ok_inuse(p) is_inuse(p)
+/* Check if p has its pinuse bit on */
+#define ok_pinuse(p) pinuse(p)
+
+#else /* !INSECURE */
+#define ok_address(M, a) (1)
+#define ok_next(b, n) (1)
+#define ok_inuse(p) (1)
+#define ok_pinuse(p) (1)
+#endif /* !INSECURE */
+
+#if (FOOTERS && !INSECURE)
+/* Check if (alleged) mstate m has expected magic field */
+#define ok_magic(M) ((M)->magic == mparams.magic)
+#else /* (FOOTERS && !INSECURE) */
+#define ok_magic(M) (1)
+#endif /* (FOOTERS && !INSECURE) */
+
+
+/* In gcc, use __builtin_expect to minimize impact of checks */
+#if !INSECURE
+#if defined(__GNUC__) && __GNUC__ >= 3
+#define RTCHECK(e) __builtin_expect(e, 1)
+#else /* GNUC */
+#define RTCHECK(e) (e)
+#endif /* GNUC */
+#else /* !INSECURE */
+#define RTCHECK(e) (1)
+#endif /* !INSECURE */
+
+/* macros to set up inuse chunks with or without footers */
+
+#if !FOOTERS
+
+#define mark_inuse_foot(M,p,s)
+
+/* Macros for setting head/foot of non-mmapped chunks */
+
+/* Set cinuse bit and pinuse bit of next chunk */
+#define set_inuse(M,p,s)\
+ ((p)->head = (((p)->head & PINUSE_BIT)|s|CINUSE_BIT),\
+ ((mchunkptr)(((char*)(p)) + (s)))->head |= PINUSE_BIT)
+
+/* Set cinuse and pinuse of this chunk and pinuse of next chunk */
+#define set_inuse_and_pinuse(M,p,s)\
+ ((p)->head = (s|PINUSE_BIT|CINUSE_BIT),\
+ ((mchunkptr)(((char*)(p)) + (s)))->head |= PINUSE_BIT)
+
+/* Set size, cinuse and pinuse bit of this chunk */
+#define set_size_and_pinuse_of_inuse_chunk(M, p, s)\
+ ((p)->head = (s|PINUSE_BIT|CINUSE_BIT))
+
+#else /* FOOTERS */
+
+/* Set foot of inuse chunk to be xor of mstate and seed */
+#define mark_inuse_foot(M,p,s)\
+ (((mchunkptr)((char*)(p) + (s)))->prev_foot = ((size_t)(M) ^ mparams.magic))
+
+#define get_mstate_for(p)\
+ ((mstate)(((mchunkptr)((char*)(p) +\
+ (chunksize(p))))->prev_foot ^ mparams.magic))
+
+#define set_inuse(M,p,s)\
+ ((p)->head = (((p)->head & PINUSE_BIT)|s|CINUSE_BIT),\
+ (((mchunkptr)(((char*)(p)) + (s)))->head |= PINUSE_BIT), \
+ mark_inuse_foot(M,p,s))
+
+#define set_inuse_and_pinuse(M,p,s)\
+ ((p)->head = (s|PINUSE_BIT|CINUSE_BIT),\
+ (((mchunkptr)(((char*)(p)) + (s)))->head |= PINUSE_BIT),\
+ mark_inuse_foot(M,p,s))
+
+#define set_size_and_pinuse_of_inuse_chunk(M, p, s)\
+ ((p)->head = (s|PINUSE_BIT|CINUSE_BIT),\
+ mark_inuse_foot(M, p, s))
+
+#endif /* !FOOTERS */
+
+/* ---------------------------- setting mparams -------------------------- */
+
+#ifdef ENABLE_LARGE_PAGES
+#ifndef WIN32
+typedef long (*gethugepagesize_t)(void);
+#else
+typedef size_t (WINAPI *GetLargePageMinimum_t)(void);
+#endif
+#endif
+
+/* Initialize mparams */
+static int init_mparams(void) {
+#ifdef NEED_GLOBAL_LOCK_INIT
+ if (malloc_global_mutex_status <= 0)
+ init_malloc_global_mutex();
+#endif
+
+ ACQUIRE_MALLOC_GLOBAL_LOCK();
+ if (mparams.magic == 0) {
+ size_t magic;
+ size_t psize;
+ size_t gsize;
+
+#ifndef WIN32
+ psize = malloc_getpagesize;
+ gsize = ((DEFAULT_GRANULARITY != 0)? DEFAULT_GRANULARITY : psize);
+ mmapped_granularity = psize;
+#else /* WIN32 */
+ {
+ SYSTEM_INFO system_info;
+ GetSystemInfo(&system_info);
+ psize = system_info.dwPageSize;
+ gsize = ((DEFAULT_GRANULARITY > system_info.dwAllocationGranularity)?
+ DEFAULT_GRANULARITY : system_info.dwAllocationGranularity);
+ mmapped_granularity = system_info.dwAllocationGranularity;
+ }
+#endif /* WIN32 */
+#ifdef ENABLE_LARGE_PAGES
+#ifndef WIN32
+ {
+ int ih=open("/proc/meminfo", O_RDONLY);
+ if(-1!=ih)
+ {
+ char buffer[4096], *hugepagesize, *hugepages;
+ buffer[read(ih, buffer, sizeof(buffer)-1)]=0;
+ close(ih);
+ hugepagesize=strstr(buffer, "Hugepagesize:");
+ hugepages=strstr(buffer, "HugePages_Total:");
+ if(hugepagesize && hugepages)
+ {
+ unsigned _hugepages=0, _hugepagesize=0;
+ while((*++hugepagesize!=' '));
+ while((*++hugepages!=' '));
+ while((*++hugepagesize==' '));
+ while((*++hugepages==' '));
+ sscanf(hugepagesize, "%u", &_hugepagesize);
+ sscanf(hugepages, "%u", &_hugepages);
+#if DEBUG
+ printf("Hugepages=%u, size=%u\n", _hugepages, _hugepagesize);
+#endif
+ if(_hugepages && _hugepagesize)
+ largepagesize = ((size_t)_hugepagesize)*1024;
+ }
+ }
+ }
+#else /* WIN32 */
+ {
+ GetLargePageMinimum_t GetLargePageMinimum_ = (GetLargePageMinimum_t) GetProcAddress(GetModuleHandle(__T("kernel32.dll")), "GetLargePageMinimum");
+ if(GetLargePageMinimum_)
+ largepagesize = GetLargePageMinimum_();
+ }
+#endif /* WIN32 */
+ if(largepagesize) {
+ psize = largepagesize;
+ gsize = ((DEFAULT_GRANULARITY > largepagesize)?
+ DEFAULT_GRANULARITY : largepagesize);
+ if(gsize < largepagesize) gsize = largepagesize;
+ }
+#endif /* ENABLE_LARGE_PAGES */
+
+ /* Sanity-check configuration:
+ size_t must be unsigned and as wide as pointer type.
+ ints must be at least 4 bytes.
+ alignment must be at least 8.
+ Alignment, min chunk size, and page size must all be powers of 2.
+ */
+ if ((sizeof(size_t) != sizeof(char*)) ||
+ (MAX_SIZE_T < MIN_CHUNK_SIZE) ||
+ (sizeof(int) < 4) ||
+ (MALLOC_ALIGNMENT < (size_t)8U) ||
+ ((MALLOC_ALIGNMENT & (MALLOC_ALIGNMENT-SIZE_T_ONE)) != 0) ||
+ ((MCHUNK_SIZE & (MCHUNK_SIZE-SIZE_T_ONE)) != 0) ||
+ ((gsize & (gsize-SIZE_T_ONE)) != 0) ||
+ ((psize & (psize-SIZE_T_ONE)) != 0))
+ ABORT;
+
+ mparams.granularity = gsize;
+ mparams.page_size = psize;
+ mparams.mmap_threshold = DEFAULT_MMAP_THRESHOLD;
+ mparams.trim_threshold = DEFAULT_TRIM_THRESHOLD;
+#if MORECORE_CONTIGUOUS
+ mparams.default_mflags = USE_LOCK_BIT|USE_MMAP_BIT;
+#else /* MORECORE_CONTIGUOUS */
+ mparams.default_mflags = USE_LOCK_BIT|USE_MMAP_BIT|USE_NONCONTIGUOUS_BIT;
+#endif /* MORECORE_CONTIGUOUS */
+
+#if !ONLY_MSPACES
+ /* Set up lock for main malloc area */
+ gm->mflags = mparams.default_mflags;
+ INITIAL_LOCK(&gm->mutex);
+#endif
+
+ {
+#if USE_DEV_RANDOM
+ int fd;
+ unsigned char buf[sizeof(size_t)];
+ /* Try to use /dev/urandom, else fall back on using time */
+ if ((fd = open("/dev/urandom", O_RDONLY)) >= 0 &&
+ read(fd, buf, sizeof(buf)) == sizeof(buf)) {
+ magic = *((size_t *) buf);
+ close(fd);
+ }
+ else
+#endif /* USE_DEV_RANDOM */
+#ifdef WIN32
+ magic = (size_t)(GetTickCount() ^ (size_t)0x55555555U);
+#else
+ magic = (size_t)(time(0) ^ (size_t)0x55555555U);
+#endif
+ magic |= (size_t)8U; /* ensure nonzero */
+ magic &= ~(size_t)7U; /* improve chances of fault for bad values */
+ mparams.magic = magic;
+ }
+ }
+
+ RELEASE_MALLOC_GLOBAL_LOCK();
+ return 1;
+}
+
+/* support for mallopt */
+static int change_mparam(int param_number, int value) {
+ size_t val;
+ ensure_initialization();
+ val = (value == -1)? MAX_SIZE_T : (size_t)value;
+ switch(param_number) {
+ case M_TRIM_THRESHOLD:
+ mparams.trim_threshold = val;
+ return 1;
+ case M_GRANULARITY:
+ if (val >= mparams.page_size && ((val & (val-1)) == 0)) {
+ mparams.granularity = val;
+ return 1;
+ }
+ else
+ return 0;
+ case M_MMAP_THRESHOLD:
+ mparams.mmap_threshold = val;
+ return 1;
+ default:
+ return 0;
+ }
+}
+
+#if DEBUG
+/* ------------------------- Debugging Support --------------------------- */
+
+/* Check properties of any chunk, whether free, inuse, mmapped etc */
+static void do_check_any_chunk(mstate m, mchunkptr p) {
+ assert((is_aligned(chunk2mem(p))) || (p->head == FENCEPOST_HEAD));
+ assert(ok_address(m, p));
+}
+
+/* Check properties of top chunk */
+static void do_check_top_chunk(mstate m, mchunkptr p) {
+ msegmentptr sp = segment_holding(m, (char*)p);
+ size_t sz = p->head & ~INUSE_BITS; /* third-lowest bit can be set! */
+ assert(sp != 0);
+ assert((is_aligned(chunk2mem(p))) || (p->head == FENCEPOST_HEAD));
+ assert(ok_address(m, p));
+ assert(sz == m->topsize);
+ assert(sz > 0);
+ assert(sz == ((sp->base + sp->size) - (char*)p) - TOP_FOOT_SIZE);
+ assert(pinuse(p));
+ assert(!pinuse(chunk_plus_offset(p, sz)));
+}
+
+/* Check properties of (inuse) mmapped chunks */
+static void do_check_mmapped_chunk(mstate m, mchunkptr p) {
+ size_t sz = chunksize(p);
+ size_t len = (sz + (p->prev_foot) + MMAP_FOOT_PAD);
+ assert(is_mmapped(p));
+ assert(use_mmap(m));
+ assert((is_aligned(chunk2mem(p))) || (p->head == FENCEPOST_HEAD));
+ assert(ok_address(m, p));
+ assert(!is_small(sz));
+ assert((len & (mmapped_granularity-SIZE_T_ONE)) == 0);
+ assert(chunk_plus_offset(p, sz)->head == FENCEPOST_HEAD);
+ assert(chunk_plus_offset(p, sz+SIZE_T_SIZE)->head == 0);
+}
+
+/* Check properties of inuse chunks */
+static void do_check_inuse_chunk(mstate m, mchunkptr p) {
+ do_check_any_chunk(m, p);
+ assert(is_inuse(p));
+ assert(next_pinuse(p));
+ /* If not pinuse and not mmapped, previous chunk has OK offset */
+ assert(is_mmapped(p) || pinuse(p) || next_chunk(prev_chunk(p)) == p);
+ if (is_mmapped(p))
+ do_check_mmapped_chunk(m, p);
+}
+
+/* Check properties of free chunks */
+static void do_check_free_chunk(mstate m, mchunkptr p) {
+ size_t sz = chunksize(p);
+ mchunkptr next = chunk_plus_offset(p, sz);
+ do_check_any_chunk(m, p);
+ assert(!is_inuse(p));
+ assert(!next_pinuse(p));
+ assert (!is_mmapped(p));
+ if (p != m->dv && p != m->top) {
+ if (sz >= MIN_CHUNK_SIZE) {
+ assert((sz & CHUNK_ALIGN_MASK) == 0);
+ assert(is_aligned(chunk2mem(p)));
+ assert(next->prev_foot == sz);
+ assert(pinuse(p));
+ assert (next == m->top || is_inuse(next));
+ assert(p->fd->bk == p);
+ assert(p->bk->fd == p);
+ }
+ else /* markers are always of size SIZE_T_SIZE */
+ assert(sz == SIZE_T_SIZE);
+ }
+}
+
+/* Check properties of malloced chunks at the point they are malloced */
+static void do_check_malloced_chunk(mstate m, void* mem, size_t s) {
+ if (mem != 0) {
+ mchunkptr p = mem2chunk(mem);
+ size_t sz = p->head & ~INUSE_BITS;
+ do_check_inuse_chunk(m, p);
+ assert((sz & CHUNK_ALIGN_MASK) == 0);
+ assert(sz >= MIN_CHUNK_SIZE);
+ assert(sz >= s);
+ /* unless mmapped, size is less than MIN_CHUNK_SIZE more than request */
+ assert(is_mmapped(p) || sz < (s + MIN_CHUNK_SIZE));
+ }
+}
+
+/* Check a tree and its subtrees. */
+static void do_check_tree(mstate m, tchunkptr t) {
+ tchunkptr head = 0;
+ tchunkptr u = t;
+ bindex_t tindex = t->index;
+ size_t tsize = chunksize(t);
+ bindex_t idx;
+ compute_tree_index(tsize, idx);
+ assert(tindex == idx);
+ assert(tsize >= MIN_LARGE_SIZE);
+ assert(tsize >= minsize_for_tree_index(idx));
+ assert((idx == NTREEBINS-1) || (tsize < minsize_for_tree_index((idx+1))));
+
+ do { /* traverse through chain of same-sized nodes */
+ do_check_any_chunk(m, ((mchunkptr)u));
+ assert(u->index == tindex);
+ assert(chunksize(u) == tsize);
+ assert(!is_inuse(u));
+ assert(!next_pinuse(u));
+ assert(u->fd->bk == u);
+ assert(u->bk->fd == u);
+ if (u->parent == 0) {
+ assert(u->child[0] == 0);
+ assert(u->child[1] == 0);
+ }
+ else {
+ assert(head == 0); /* only one node on chain has parent */
+ head = u;
+ assert(u->parent != u);
+ assert (u->parent->child[0] == u ||
+ u->parent->child[1] == u ||
+ *((tbinptr*)(u->parent)) == u);
+ if (u->child[0] != 0) {
+ assert(u->child[0]->parent == u);
+ assert(u->child[0] != u);
+ do_check_tree(m, u->child[0]);
+ }
+ if (u->child[1] != 0) {
+ assert(u->child[1]->parent == u);
+ assert(u->child[1] != u);
+ do_check_tree(m, u->child[1]);
+ }
+ if (u->child[0] != 0 && u->child[1] != 0) {
+ assert(chunksize(u->child[0]) < chunksize(u->child[1]));
+ }
+ }
+ u = u->fd;
+ } while (u != t);
+ assert(head != 0);
+}
+
+/* Check all the chunks in a treebin. */
+static void do_check_treebin(mstate m, bindex_t i) {
+ tbinptr* tb = treebin_at(m, i);
+ tchunkptr t = *tb;
+ int empty = (m->treemap & (1U << i)) == 0;
+ if (t == 0)
+ assert(empty);
+ if (!empty)
+ do_check_tree(m, t);
+}
+
+/* Check all the chunks in a smallbin. */
+static void do_check_smallbin(mstate m, bindex_t i) {
+ sbinptr b = smallbin_at(m, i);
+ mchunkptr p = b->bk;
+ unsigned int empty = (m->smallmap & (1U << i)) == 0;
+ if (p == b)
+ assert(empty);
+ if (!empty) {
+ for (; p != b; p = p->bk) {
+ size_t size = chunksize(p);
+ mchunkptr q;
+ /* each chunk claims to be free */
+ do_check_free_chunk(m, p);
+ /* chunk belongs in bin */
+ assert(small_index(size) == i);
+ assert(p->bk == b || chunksize(p->bk) == chunksize(p));
+ /* chunk is followed by an inuse chunk */
+ q = next_chunk(p);
+ if (q->head != FENCEPOST_HEAD)
+ do_check_inuse_chunk(m, q);
+ }
+ }
+}
+
+/* Find x in a bin. Used in other check functions. */
+static int bin_find(mstate m, mchunkptr x) {
+ size_t size = chunksize(x);
+ if (is_small(size)) {
+ bindex_t sidx = small_index(size);
+ sbinptr b = smallbin_at(m, sidx);
+ if (smallmap_is_marked(m, sidx)) {
+ mchunkptr p = b;
+ do {
+ if (p == x)
+ return 1;
+ } while ((p = p->fd) != b);
+ }
+ }
+ else {
+ bindex_t tidx;
+ compute_tree_index(size, tidx);
+ if (treemap_is_marked(m, tidx)) {
+ tchunkptr t = *treebin_at(m, tidx);
+ size_t sizebits = size << leftshift_for_tree_index(tidx);
+ while (t != 0 && chunksize(t) != size) {
+ t = t->child[(sizebits >> (SIZE_T_BITSIZE-SIZE_T_ONE)) & 1];
+ sizebits <<= 1;
+ }
+ if (t != 0) {
+ tchunkptr u = t;
+ do {
+ if (u == (tchunkptr)x)
+ return 1;
+ } while ((u = u->fd) != t);
+ }
+ }
+ }
+ return 0;
+}
+
+/* Traverse each chunk and check it; return total */
+static size_t traverse_and_check(mstate m) {
+ size_t sum = 0;
+ if (is_initialized(m)) {
+ msegmentptr s = &m->seg;
+ sum += m->topsize + TOP_FOOT_SIZE;
+ while (s != 0) {
+ mchunkptr q = align_as_chunk(s->base);
+ mchunkptr lastq = 0;
+ assert(pinuse(q));
+ while (segment_holds(s, q) &&
+ q != m->top && q->head != FENCEPOST_HEAD) {
+ sum += chunksize(q);
+ if (is_inuse(q)) {
+ assert(!bin_find(m, q));
+ do_check_inuse_chunk(m, q);
+ }
+ else {
+ assert(q == m->dv || bin_find(m, q));
+ assert(lastq == 0 || is_inuse(lastq)); /* Not 2 consecutive free */
+ do_check_free_chunk(m, q);
+ }
+ lastq = q;
+ q = next_chunk(q);
+ }
+ s = s->next;
+ }
+ }
+ return sum;
+}
+
+/* Check all properties of malloc_state. */
+static void do_check_malloc_state(mstate m) {
+ bindex_t i;
+ size_t total;
+ /* check bins */
+ for (i = 0; i < NSMALLBINS; ++i)
+ do_check_smallbin(m, i);
+ for (i = 0; i < NTREEBINS; ++i)
+ do_check_treebin(m, i);
+
+ if (m->dvsize != 0) { /* check dv chunk */
+ do_check_any_chunk(m, m->dv);
+ assert(m->dvsize == chunksize(m->dv));
+ assert(m->dvsize >= MIN_CHUNK_SIZE);
+ assert(bin_find(m, m->dv) == 0);
+ }
+
+ if (m->top != 0) { /* check top chunk */
+ do_check_top_chunk(m, m->top);
+ /*assert(m->topsize == chunksize(m->top)); redundant */
+ assert(m->topsize > 0);
+ assert(bin_find(m, m->top) == 0);
+ }
+
+ total = traverse_and_check(m);
+ assert(total <= m->footprint);
+ assert(m->footprint <= m->max_footprint);
+}
+#endif /* DEBUG */
+
+/* ----------------------------- statistics ------------------------------ */
+
+#if !NO_MALLINFO
+static struct mallinfo internal_mallinfo(mstate m) {
+ struct mallinfo nm = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
+ ensure_initialization();
+ if (!PREACTION(m)) {
+ check_malloc_state(m);
+ if (is_initialized(m)) {
+ size_t nfree = SIZE_T_ONE; /* top always free */
+ size_t mfree = m->topsize + TOP_FOOT_SIZE;
+ size_t sum = mfree;
+ msegmentptr s = &m->seg;
+ while (s != 0) {
+ mchunkptr q = align_as_chunk(s->base);
+ while (segment_holds(s, q) &&
+ q != m->top && q->head != FENCEPOST_HEAD) {
+ size_t sz = chunksize(q);
+ sum += sz;
+ if (!is_inuse(q)) {
+ mfree += sz;
+ ++nfree;
+ }
+ q = next_chunk(q);
+ }
+ s = s->next;
+ }
+
+ nm.arena = sum;
+ nm.ordblks = nfree;
+ nm.hblkhd = m->footprint - sum;
+ nm.usmblks = m->max_footprint;
+ nm.uordblks = m->footprint - mfree;
+ nm.fordblks = mfree;
+ nm.keepcost = m->topsize;
+ }
+
+ POSTACTION(m);
+ }
+ return nm;
+}
+#endif /* !NO_MALLINFO */
+
+static void internal_malloc_stats(mstate m) {
+ ensure_initialization();
+ if (!PREACTION(m)) {
+ size_t maxfp = 0;
+ size_t fp = 0;
+ size_t used = 0;
+ check_malloc_state(m);
+ if (is_initialized(m)) {
+ msegmentptr s = &m->seg;
+ maxfp = m->max_footprint;
+ fp = m->footprint;
+ used = fp - (m->topsize + TOP_FOOT_SIZE);
+
+ while (s != 0) {
+ mchunkptr q = align_as_chunk(s->base);
+ while (segment_holds(s, q) &&
+ q != m->top && q->head != FENCEPOST_HEAD) {
+ if (!is_inuse(q))
+ used -= chunksize(q);
+ q = next_chunk(q);
+ }
+ s = s->next;
+ }
+ }
+
+ fprintf(stderr, "max system bytes = %10lu\n", (unsigned long)(maxfp));
+ fprintf(stderr, "system bytes = %10lu\n", (unsigned long)(fp));
+ fprintf(stderr, "in use bytes = %10lu\n", (unsigned long)(used));
+
+ POSTACTION(m);
+ }
+}
+
+/* ----------------------- Operations on smallbins ----------------------- */
+
+/*
+ Various forms of linking and unlinking are defined as macros. Even
+ the ones for trees, which are very long but have very short typical
+ paths. This is ugly but reduces reliance on inlining support of
+ compilers.
+*/
+
+/* Link a free chunk into a smallbin */
+#define insert_small_chunk(M, P, S) {\
+ bindex_t I = small_index(S);\
+ mchunkptr B = smallbin_at(M, I);\
+ mchunkptr F = B;\
+ assert(S >= MIN_CHUNK_SIZE);\
+ if (!smallmap_is_marked(M, I))\
+ mark_smallmap(M, I);\
+ else if (RTCHECK(ok_address(M, B->fd)))\
+ F = B->fd;\
+ else {\
+ CORRUPTION_ERROR_ACTION(M);\
+ }\
+ B->fd = P;\
+ F->bk = P;\
+ P->fd = F;\
+ P->bk = B;\
+}
+
+/* Unlink a chunk from a smallbin */
+#define unlink_small_chunk(M, P, S) {\
+ mchunkptr F = P->fd;\
+ mchunkptr B = P->bk;\
+ bindex_t I = small_index(S);\
+ assert(P != B);\
+ assert(P != F);\
+ assert(chunksize(P) == small_index2size(I));\
+ if (F == B)\
+ clear_smallmap(M, I);\
+ else if (RTCHECK((F == smallbin_at(M,I) || ok_address(M, F)) &&\
+ (B == smallbin_at(M,I) || ok_address(M, B)))) {\
+ F->bk = B;\
+ B->fd = F;\
+ }\
+ else {\
+ CORRUPTION_ERROR_ACTION(M);\
+ }\
+}
+
+/* Unlink the first chunk from a smallbin */
+#define unlink_first_small_chunk(M, B, P, I) {\
+ mchunkptr F = P->fd;\
+ assert(P != B);\
+ assert(P != F);\
+ assert(chunksize(P) == small_index2size(I));\
+ if (B == F)\
+ clear_smallmap(M, I);\
+ else if (RTCHECK(ok_address(M, F))) {\
+ B->fd = F;\
+ F->bk = B;\
+ }\
+ else {\
+ CORRUPTION_ERROR_ACTION(M);\
+ }\
+}
+
+
+
+/* Replace dv node, binning the old one */
+/* Used only when dvsize known to be small */
+#define replace_dv(M, P, S) {\
+ size_t DVS = M->dvsize;\
+ if (DVS != 0) {\
+ mchunkptr DV = M->dv;\
+ assert(is_small(DVS));\
+ insert_small_chunk(M, DV, DVS);\
+ }\
+ M->dvsize = S;\
+ M->dv = P;\
+}
+
+/* ------------------------- Operations on trees ------------------------- */
+
+/* Insert chunk into tree */
+#define insert_large_chunk(M, X, S) {\
+ tbinptr* H;\
+ bindex_t I;\
+ compute_tree_index(S, I);\
+ H = treebin_at(M, I);\
+ X->index = I;\
+ X->child[0] = X->child[1] = 0;\
+ if (!treemap_is_marked(M, I)) {\
+ mark_treemap(M, I);\
+ *H = X;\
+ X->parent = (tchunkptr)H;\
+ X->fd = X->bk = X;\
+ }\
+ else {\
+ tchunkptr T = *H;\
+ size_t K = S << leftshift_for_tree_index(I);\
+ for (;;) {\
+ if (chunksize(T) != S) {\
+ tchunkptr* C = &(T->child[(K >> (SIZE_T_BITSIZE-SIZE_T_ONE)) & 1]);\
+ K <<= 1;\
+ if (*C != 0)\
+ T = *C;\
+ else if (RTCHECK(ok_address(M, C))) {\
+ *C = X;\
+ X->parent = T;\
+ X->fd = X->bk = X;\
+ break;\
+ }\
+ else {\
+ CORRUPTION_ERROR_ACTION(M);\
+ break;\
+ }\
+ }\
+ else {\
+ tchunkptr F = T->fd;\
+ if (RTCHECK(ok_address(M, T) && ok_address(M, F))) {\
+ T->fd = F->bk = X;\
+ X->fd = F;\
+ X->bk = T;\
+ X->parent = 0;\
+ break;\
+ }\
+ else {\
+ CORRUPTION_ERROR_ACTION(M);\
+ break;\
+ }\
+ }\
+ }\
+ }\
+}
+
+/*
+ Unlink steps:
+
+ 1. If x is a chained node, unlink it from its same-sized fd/bk links
+ and choose its bk node as its replacement.
+ 2. If x was the last node of its size, but not a leaf node, it must
+ be replaced with a leaf node (not merely one with an open left or
+ right), to make sure that lefts and rights of descendents
+ correspond properly to bit masks. We use the rightmost descendent
+ of x. We could use any other leaf, but this is easy to locate and
+ tends to counteract removal of leftmosts elsewhere, and so keeps
+ paths shorter than minimally guaranteed. This doesn't loop much
+ because on average a node in a tree is near the bottom.
+ 3. If x is the base of a chain (i.e., has parent links) relink
+ x's parent and children to x's replacement (or null if none).
+*/
+
+#define unlink_large_chunk(M, X) {\
+ tchunkptr XP = X->parent;\
+ tchunkptr R;\
+ if (X->bk != X) {\
+ tchunkptr F = X->fd;\
+ R = X->bk;\
+ if (RTCHECK(ok_address(M, F))) {\
+ F->bk = R;\
+ R->fd = F;\
+ }\
+ else {\
+ CORRUPTION_ERROR_ACTION(M);\
+ }\
+ }\
+ else {\
+ tchunkptr* RP;\
+ if (((R = *(RP = &(X->child[1]))) != 0) ||\
+ ((R = *(RP = &(X->child[0]))) != 0)) {\
+ tchunkptr* CP;\
+ while ((*(CP = &(R->child[1])) != 0) ||\
+ (*(CP = &(R->child[0])) != 0)) {\
+ R = *(RP = CP);\
+ }\
+ if (RTCHECK(ok_address(M, RP)))\
+ *RP = 0;\
+ else {\
+ CORRUPTION_ERROR_ACTION(M);\
+ }\
+ }\
+ }\
+ if (XP != 0) {\
+ tbinptr* H = treebin_at(M, X->index);\
+ if (X == *H) {\
+ if ((*H = R) == 0) \
+ clear_treemap(M, X->index);\
+ }\
+ else if (RTCHECK(ok_address(M, XP))) {\
+ if (XP->child[0] == X) \
+ XP->child[0] = R;\
+ else \
+ XP->child[1] = R;\
+ }\
+ else\
+ CORRUPTION_ERROR_ACTION(M);\
+ if (R != 0) {\
+ if (RTCHECK(ok_address(M, R))) {\
+ tchunkptr C0, C1;\
+ R->parent = XP;\
+ if ((C0 = X->child[0]) != 0) {\
+ if (RTCHECK(ok_address(M, C0))) {\
+ R->child[0] = C0;\
+ C0->parent = R;\
+ }\
+ else\
+ CORRUPTION_ERROR_ACTION(M);\
+ }\
+ if ((C1 = X->child[1]) != 0) {\
+ if (RTCHECK(ok_address(M, C1))) {\
+ R->child[1] = C1;\
+ C1->parent = R;\
+ }\
+ else\
+ CORRUPTION_ERROR_ACTION(M);\
+ }\
+ }\
+ else\
+ CORRUPTION_ERROR_ACTION(M);\
+ }\
+ }\
+}
+
+/* Relays to large vs small bin operations */
+
+#define insert_chunk(M, P, S)\
+ if (is_small(S)) insert_small_chunk(M, P, S)\
+ else { tchunkptr TP = (tchunkptr)(P); insert_large_chunk(M, TP, S); }
+
+#define unlink_chunk(M, P, S)\
+ if (is_small(S)) unlink_small_chunk(M, P, S)\
+ else { tchunkptr TP = (tchunkptr)(P); unlink_large_chunk(M, TP); }
+
+
+/* Relays to internal calls to malloc/free from realloc, memalign etc */
+#if ONLY_MSPACES
+static void* mspace_malloc_implementation(mstate ms, size_t bytes, unsigned flags);
+#define internal_malloc(m, b, f) mspace_malloc_implementation(m, b, f)
+#define internal_free(m, mem) mspace_free(m,mem);
+#else /* ONLY_MSPACES */
+#if MSPACES
+static void* mspace_malloc_implementation(mstate ms, size_t bytes, unsigned flags);
+#define internal_malloc(m, b, f)\
+ (m == gm)? dlmalloc(b) : mspace_malloc_implementation(m, b, f)
+#define internal_free(m, mem)\
+ if (m == gm) dlfree(mem); else mspace_free(m,mem);
+#else /* MSPACES */
+#define internal_malloc(m, b, f) dlmalloc(b)
+#define internal_free(m, mem) dlfree(mem)
+#endif /* MSPACES */
+#endif /* ONLY_MSPACES */
+
+/* ----------------------- Direct-mmapping chunks ----------------------- */
+
+/*
+ Directly mmapped chunks are set up with an offset to the start of
+ the mmapped region stored in the prev_foot field of the chunk. This
+ allows reconstruction of the required argument to MUNMAP when freed,
+ and also allows adjustment of the returned chunk to meet alignment
+ requirements (especially in memalign).
+
+ Directly mmapped chunks store a void* handle which can be used by
+ the DIRECT_MMAP implementation to keep a per-chunk state. This
+ handle is currently stored right at the front of the mmapped region,
+ but do not rely on this - use the supplied parameter.
+*/
+
+/* Malloc using mmap */
+static void* mmap_alloc(mstate m, size_t nb, unsigned flags) {
+ size_t mmsize = mmap_align_size(nb + SEVEN_SIZE_T_SIZES + CHUNK_ALIGN_MASK);
+ if (mmsize > nb) { /* Check for wrap around 0 */
+ void* mmaph = 0;
+ char* mm = (char*)(CALL_DIRECT_MMAP(&mmaph, mmsize, flags));
+ if (mm != CMFAIL) {
+ size_t offset = MALLOC_ALIGNMENT + align_offset(chunk2mem(mm));
+ size_t psize = mmsize - offset - MMAP_FOOT_PAD;
+ mchunkptr p = (mchunkptr)(mm + offset);
+ *(void**)mm = mmaph;
+ p->prev_foot = offset;
+ p->head = psize;
+ mark_inuse_foot(m, p, psize);
+ chunk_plus_offset(p, psize)->head = FENCEPOST_HEAD;
+ chunk_plus_offset(p, psize+SIZE_T_SIZE)->head = 0;
+
+ if (m->least_addr == 0 || mm < m->least_addr)
+ m->least_addr = mm;
+ if ((m->footprint += mmsize) > m->max_footprint)
+ m->max_footprint = m->footprint;
+ assert(is_aligned(chunk2mem(p)));
+ check_mmapped_chunk(m, p);
+ return chunk2mem(p);
+ }
+ }
+ return 0;
+}
+
+/* Realloc using mmap */
+static mchunkptr mmap_resize(mstate m, mchunkptr oldp, size_t nb, unsigned flags) {
+ size_t oldsize = chunksize(oldp);
+ if (is_small(nb)) /* Can't shrink mmap regions below small size */
+ return 0;
+ /* Keep old chunk if big enough but not too big */
+ if (oldsize >= nb + SIZE_T_SIZE &&
+ (oldsize - nb) <= (mparams.granularity << 1))
+ return oldp;
+ else {
+ size_t offset = oldp->prev_foot;
+ size_t oldmmsize = oldsize + offset + MMAP_FOOT_PAD;
+ size_t newmmsize = mmap_align_size(nb + SEVEN_SIZE_T_SIZES + CHUNK_ALIGN_MASK);
+ char* mm = (char*)oldp - offset;
+ void* mmaph = *(void**)mm;
+ char* cp = (char*)CALL_DIRECT_MREMAP(&mmaph, mm, oldmmsize, newmmsize, (flags & M2_PREVENT_MOVE) ? 0 : MREMAP_MAYMOVE, flags);
+ if (cp != CMFAIL) {
+ mchunkptr newp = (mchunkptr)(cp + offset);
+ size_t psize = newmmsize - offset - MMAP_FOOT_PAD;
+ *(void**)cp = mmaph;
+ newp->head = psize;
+ mark_inuse_foot(m, newp, psize);
+ chunk_plus_offset(newp, psize)->head = FENCEPOST_HEAD;
+ chunk_plus_offset(newp, psize+SIZE_T_SIZE)->head = 0;
+
+ if (cp < m->least_addr)
+ m->least_addr = cp;
+ if ((m->footprint += newmmsize - oldmmsize) > m->max_footprint)
+ m->max_footprint = m->footprint;
+ check_mmapped_chunk(m, newp);
+ return newp;
+ }
+ }
+ return 0;
+}
+
+/* -------------------------- mspace management -------------------------- */
+
+/* Initialize top chunk and its size */
+static void init_top(mstate m, mchunkptr p, size_t psize) {
+ /* Ensure alignment */
+ size_t offset = align_offset(chunk2mem(p));
+ p = (mchunkptr)((char*)p + offset);
+ psize -= offset;
+
+ m->top = p;
+ m->topsize = psize;
+ p->head = psize | PINUSE_BIT;
+ /* set size of fake trailing chunk holding overhead space only once */
+ chunk_plus_offset(p, psize)->head = TOP_FOOT_SIZE;
+ m->trim_check = mparams.trim_threshold; /* reset on each update */
+}
+
+/* Initialize bins for a new mstate that is otherwise zeroed out */
+static void init_bins(mstate m) {
+ /* Establish circular links for smallbins */
+ bindex_t i;
+ for (i = 0; i < NSMALLBINS; ++i) {
+ sbinptr bin = smallbin_at(m,i);
+ bin->fd = bin->bk = bin;
+ }
+}
+
+#if PROCEED_ON_ERROR
+
+/* default corruption action */
+static void reset_on_error(mstate m) {
+ int i;
+ ++malloc_corruption_error_count;
+ /* Reinitialize fields to forget about all memory */
+ m->smallbins = m->treebins = 0;
+ m->dvsize = m->topsize = 0;
+ m->seg.base = 0;
+ m->seg.size = 0;
+ m->seg.next = 0;
+ m->top = m->dv = 0;
+ for (i = 0; i < NTREEBINS; ++i)
+ *treebin_at(m, i) = 0;
+ init_bins(m);
+}
+#endif /* PROCEED_ON_ERROR */
+
+/* Allocate chunk and prepend remainder with chunk in successor base. */
+static void* prepend_alloc(mstate m, char* newbase, char* oldbase,
+ size_t nb) {
+ mchunkptr p = align_as_chunk(newbase);
+ mchunkptr oldfirst = align_as_chunk(oldbase);
+ size_t psize = (char*)oldfirst - (char*)p;
+ mchunkptr q = chunk_plus_offset(p, nb);
+ size_t qsize = psize - nb;
+ set_size_and_pinuse_of_inuse_chunk(m, p, nb);
+
+ assert((char*)oldfirst > (char*)q);
+ assert(pinuse(oldfirst));
+ assert(qsize >= MIN_CHUNK_SIZE);
+
+ /* consolidate remainder with first chunk of old base */
+ if (oldfirst == m->top) {
+ size_t tsize = m->topsize += qsize;
+ m->top = q;
+ q->head = tsize | PINUSE_BIT;
+ check_top_chunk(m, q);
+ }
+ else if (oldfirst == m->dv) {
+ size_t dsize = m->dvsize += qsize;
+ m->dv = q;
+ set_size_and_pinuse_of_free_chunk(q, dsize);
+ }
+ else {
+ if (!is_inuse(oldfirst)) {
+ size_t nsize = chunksize(oldfirst);
+ unlink_chunk(m, oldfirst, nsize);
+ oldfirst = chunk_plus_offset(oldfirst, nsize);
+ qsize += nsize;
+ }
+ set_free_with_pinuse(q, qsize, oldfirst);
+ insert_chunk(m, q, qsize);
+ check_free_chunk(m, q);
+ }
+
+ check_malloced_chunk(m, chunk2mem(p), nb);
+ return chunk2mem(p);
+}
+
+/* Add a segment to hold a new noncontiguous region */
+static void add_segment(mstate m, char* tbase, size_t tsize, flag_t mmapped) {
+ /* Determine locations and sizes of segment, fenceposts, old top */
+ char* old_top = (char*)m->top;
+ msegmentptr oldsp = segment_holding(m, old_top);
+ char* old_end = oldsp->base + oldsp->size;
+ size_t ssize = pad_request(sizeof(struct malloc_segment));
+ char* rawsp = old_end - (ssize + FOUR_SIZE_T_SIZES + CHUNK_ALIGN_MASK);
+ size_t offset = align_offset(chunk2mem(rawsp));
+ char* asp = rawsp + offset;
+ char* csp = (asp < (old_top + MIN_CHUNK_SIZE))? old_top : asp;
+ mchunkptr sp = (mchunkptr)csp;
+ msegmentptr ss = (msegmentptr)(chunk2mem(sp));
+ mchunkptr tnext = chunk_plus_offset(sp, ssize);
+ mchunkptr p = tnext;
+ int nfences = 0;
+
+ /* reset top to new space */
+ init_top(m, (mchunkptr)tbase, tsize - TOP_FOOT_SIZE);
+
+ /* Set up segment record */
+ assert(is_aligned(ss));
+ set_size_and_pinuse_of_inuse_chunk(m, sp, ssize);
+ *ss = m->seg; /* Push current record */
+ m->seg.base = tbase;
+ m->seg.size = tsize;
+ m->seg.sflags = mmapped;
+ m->seg.next = ss;
+
+ /* Insert trailing fenceposts */
+ for (;;) {
+ mchunkptr nextp = chunk_plus_offset(p, SIZE_T_SIZE);
+ p->head = FENCEPOST_HEAD;
+ ++nfences;
+ if ((char*)(&(nextp->head)) < old_end)
+ p = nextp;
+ else
+ break;
+ }
+ assert(nfences >= 2);
+
+ /* Insert the rest of old top into a bin as an ordinary free chunk */
+ if (csp != old_top) {
+ mchunkptr q = (mchunkptr)old_top;
+ size_t psize = csp - old_top;
+ mchunkptr tn = chunk_plus_offset(q, psize);
+ set_free_with_pinuse(q, psize, tn);
+ insert_chunk(m, q, psize);
+ }
+
+ check_top_chunk(m, m->top);
+}
+
+/* -------------------------- System allocation -------------------------- */
+
+/* Get memory from system using MORECORE or MMAP */
+static void* sys_alloc(mstate m, size_t nb, unsigned flags) {
+ char* tbase = CMFAIL;
+ size_t tsize = 0;
+ flag_t mmap_flag = 0;
+
+ ensure_initialization();
+
+ /* Directly map large chunks, but only if already initialized */
+ if (use_mmap(m) && (nb >= mparams.mmap_threshold || (flags & M2_ALWAYS_MMAP)) && m->topsize != 0) {
+ void* mem = mmap_alloc(m, nb, flags);
+ if (mem != 0 || (flags & M2_ALWAYS_MMAP))
+ return mem;
+ }
+
+ /*
+ Try getting memory in any of three ways (in most-preferred to
+ least-preferred order):
+ 1. A call to MORECORE that can normally contiguously extend memory.
+ (disabled if not MORECORE_CONTIGUOUS or not HAVE_MORECORE or
+ or main space is mmapped or a previous contiguous call failed)
+ 2. A call to MMAP new space (disabled if not HAVE_MMAP).
+ Note that under the default settings, if MORECORE is unable to
+ fulfill a request, and HAVE_MMAP is true, then mmap is
+ used as a noncontiguous system allocator. This is a useful backup
+ strategy for systems with holes in address spaces -- in this case
+ sbrk cannot contiguously expand the heap, but mmap may be able to
+ find space.
+ 3. A call to MORECORE that cannot usually contiguously extend memory.
+ (disabled if not HAVE_MORECORE)
+
+ In all cases, we need to request enough bytes from system to ensure
+ we can malloc nb bytes upon success, so pad with enough space for
+ top_foot, plus alignment-pad to make sure we don't lose bytes if
+ not on boundary, and round this up to a granularity unit.
+ */
+
+ if (MORECORE_CONTIGUOUS && !use_noncontiguous(m)) {
+ char* br = CMFAIL;
+ msegmentptr ss = (m->top == 0)? 0 : segment_holding(m, (char*)m->top);
+ size_t asize = 0;
+ ACQUIRE_MALLOC_GLOBAL_LOCK();
+
+ if (ss == 0) { /* First time through or recovery */
+ char* base = (char*)CALL_MORECORE(0);
+ if (base != CMFAIL) {
+ asize = granularity_align(nb + SYS_ALLOC_PADDING);
+ /* Adjust to end on a page boundary */
+ if (!is_page_aligned(base))
+ asize += (page_align((size_t)base) - (size_t)base);
+ /* Can't call MORECORE if size is negative when treated as signed */
+ if (asize < HALF_MAX_SIZE_T &&
+ (br = (char*)(CALL_MORECORE(asize))) == base) {
+ tbase = base;
+ tsize = asize;
+ }
+ }
+ }
+ else {
+ /* Subtract out existing available top space from MORECORE request. */
+ asize = granularity_align(nb - m->topsize + SYS_ALLOC_PADDING);
+ /* Use mem here only if it did continuously extend old space */
+ if (asize < HALF_MAX_SIZE_T &&
+ (br = (char*)(CALL_MORECORE(asize))) == ss->base+ss->size) {
+ tbase = br;
+ tsize = asize;
+ }
+ }
+
+ if (tbase == CMFAIL) { /* Cope with partial failure */
+ if (br != CMFAIL) { /* Try to use/extend the space we did get */
+ if (asize < HALF_MAX_SIZE_T &&
+ asize < nb + SYS_ALLOC_PADDING) {
+ size_t esize = granularity_align(nb + SYS_ALLOC_PADDING - asize);
+ if (esize < HALF_MAX_SIZE_T) {
+ char* end = (char*)CALL_MORECORE(esize);
+ if (end != CMFAIL)
+ asize += esize;
+ else { /* Can't use; try to release */
+ (void) CALL_MORECORE(-asize);
+ br = CMFAIL;
+ }
+ }
+ }
+ }
+ if (br != CMFAIL) { /* Use the space we did get */
+ tbase = br;
+ tsize = asize;
+ }
+ else
+ disable_contiguous(m); /* Don't try contiguous path in the future */
+ }
+
+ RELEASE_MALLOC_GLOBAL_LOCK();
+ }
+
+ if (HAVE_MMAP && tbase == CMFAIL) { /* Try MMAP */
+ size_t rsize = granularity_align(nb + SYS_ALLOC_PADDING);
+ if (rsize > nb) { /* Fail if wraps around zero */
+ char* mp = (char*)(CALL_MMAP(rsize, flags));
+ if (mp != CMFAIL) {
+ tbase = mp;
+ tsize = rsize;
+ mmap_flag = USE_MMAP_BIT;
+ }
+ }
+ }
+
+ if (HAVE_MORECORE && tbase == CMFAIL) { /* Try noncontiguous MORECORE */
+ size_t asize = granularity_align(nb + SYS_ALLOC_PADDING);
+ if (asize < HALF_MAX_SIZE_T) {
+ char* br = CMFAIL;
+ char* end = CMFAIL;
+ ACQUIRE_MALLOC_GLOBAL_LOCK();
+ br = (char*)(CALL_MORECORE(asize));
+ end = (char*)(CALL_MORECORE(0));
+ RELEASE_MALLOC_GLOBAL_LOCK();
+ if (br != CMFAIL && end != CMFAIL && br < end) {
+ size_t ssize = end - br;
+ if (ssize > nb + TOP_FOOT_SIZE) {
+ tbase = br;
+ tsize = ssize;
+ }
+ }
+ }
+ }
+
+ if (tbase != CMFAIL) {
+
+ if ((m->footprint += tsize) > m->max_footprint)
+ m->max_footprint = m->footprint;
+
+ if (!is_initialized(m)) { /* first-time initialization */
+ if (m->least_addr == 0 || tbase < m->least_addr)
+ m->least_addr = tbase;
+ m->seg.base = tbase;
+ m->seg.size = tsize;
+ m->seg.sflags = mmap_flag;
+ m->magic = mparams.magic;
+ m->release_checks = MAX_RELEASE_CHECK_RATE;
+ init_bins(m);
+#if !ONLY_MSPACES
+ if (is_global(m))
+ init_top(m, (mchunkptr)tbase, tsize - TOP_FOOT_SIZE);
+ else
+#endif
+ {
+ /* Offset top by embedded malloc_state */
+ mchunkptr mn = next_chunk(mem2chunk(m));
+ init_top(m, mn, (size_t)((tbase + tsize) - (char*)mn) -TOP_FOOT_SIZE);
+ }
+ }
+
+ else {
+ /* Try to merge with an existing segment */
+ msegmentptr sp = &m->seg;
+ /* Only consider most recent segment if traversal suppressed */
+ while (sp != 0 && tbase != sp->base + sp->size)
+ sp = (NO_SEGMENT_TRAVERSAL) ? 0 : sp->next;
+ if (sp != 0 &&
+ !is_extern_segment(sp) &&
+ (sp->sflags & USE_MMAP_BIT) == mmap_flag &&
+ segment_holds(sp, m->top)) { /* append */
+ sp->size += tsize;
+ init_top(m, m->top, m->topsize + tsize);
+ }
+ else {
+ if (tbase < m->least_addr)
+ m->least_addr = tbase;
+ sp = &m->seg;
+ while (sp != 0 && sp->base != tbase + tsize)
+ sp = (NO_SEGMENT_TRAVERSAL) ? 0 : sp->next;
+ if (sp != 0 &&
+ !is_extern_segment(sp) &&
+ (sp->sflags & USE_MMAP_BIT) == mmap_flag) {
+ char* oldbase = sp->base;
+ sp->base = tbase;
+ sp->size += tsize;
+ return prepend_alloc(m, tbase, oldbase, nb);
+ }
+ else
+ add_segment(m, tbase, tsize, mmap_flag);
+ }
+ }
+
+ if (nb < m->topsize) { /* Allocate from new or extended top space */
+ size_t rsize = m->topsize -= nb;
+ mchunkptr p = m->top;
+ mchunkptr r = m->top = chunk_plus_offset(p, nb);
+ r->head = rsize | PINUSE_BIT;
+ set_size_and_pinuse_of_inuse_chunk(m, p, nb);
+ check_top_chunk(m, m->top);
+ check_malloced_chunk(m, chunk2mem(p), nb);
+ return chunk2mem(p);
+ }
+ }
+
+ MALLOC_FAILURE_ACTION;
+ return 0;
+}
+
+/* ----------------------- system deallocation -------------------------- */
+
+/* Unmap and unlink any mmapped segments that don't contain used chunks */
+static size_t release_unused_segments(mstate m) {
+ size_t released = 0;
+ int nsegs = 0;
+ msegmentptr pred = &m->seg;
+ msegmentptr sp = pred->next;
+ while (sp != 0) {
+ char* base = sp->base;
+ size_t size = sp->size;
+ msegmentptr next = sp->next;
+ ++nsegs;
+ if (is_mmapped_segment(sp) && !is_extern_segment(sp)) {
+ mchunkptr p = align_as_chunk(base);
+ size_t psize = chunksize(p);
+ /* Can unmap if first chunk holds entire segment and not pinned */
+ if (!is_inuse(p) && (char*)p + psize >= base + size - TOP_FOOT_SIZE) {
+ tchunkptr tp = (tchunkptr)p;
+ assert(segment_holds(sp, (char*)sp));
+ if (p == m->dv) {
+ m->dv = 0;
+ m->dvsize = 0;
+ }
+ else {
+ unlink_large_chunk(m, tp);
+ }
+ if (CALL_MUNMAP(0/*segment*/, base, size) == 0) {
+ released += size;
+ m->footprint -= size;
+ /* unlink obsoleted record */
+ sp = pred;
+ sp->next = next;
+ }
+ else { /* back out if cannot unmap */
+ insert_large_chunk(m, tp, psize);
+ }
+ }
+ }
+ if (NO_SEGMENT_TRAVERSAL) /* scan only first segment */
+ break;
+ pred = sp;
+ sp = next;
+ }
+ /* Reset check counter */
+ m->release_checks = ((nsegs > MAX_RELEASE_CHECK_RATE)?
+ nsegs : MAX_RELEASE_CHECK_RATE);
+ return released;
+}
+
+static int sys_trim(mstate m, size_t pad) {
+ size_t released = 0;
+ ensure_initialization();
+ if (pad < MAX_REQUEST && is_initialized(m)) {
+ pad += TOP_FOOT_SIZE; /* ensure enough room for segment overhead */
+
+ if (m->topsize > pad) {
+ /* Shrink top space in granularity-size units, keeping at least one */
+ size_t unit = mparams.granularity;
+ size_t extra = ((m->topsize - pad + (unit - SIZE_T_ONE)) / unit -
+ SIZE_T_ONE) * unit;
+ msegmentptr sp = segment_holding(m, (char*)m->top);
+
+ if (!is_extern_segment(sp)) {
+ if (is_mmapped_segment(sp)) {
+ if (HAVE_MMAP &&
+ sp->size >= extra &&
+ !has_segment_link(m, sp)) { /* can't shrink if pinned */
+ size_t newsize = sp->size - extra;
+ /* Prefer mremap, fall back to munmap */
+ if ((CALL_MREMAP(sp->base, sp->size, newsize, 0) != MFAIL) ||
+ (CALL_MUNMAP(0/*segment*/, sp->base + newsize, extra) == 0)) {
+ released = extra;
+ }
+ }
+ }
+ else if (HAVE_MORECORE) {
+ if (extra >= HALF_MAX_SIZE_T) /* Avoid wrapping negative */
+ extra = (HALF_MAX_SIZE_T) + SIZE_T_ONE - unit;
+ ACQUIRE_MALLOC_GLOBAL_LOCK();
+ {
+ /* Make sure end of memory is where we last set it. */
+ char* old_br = (char*)(CALL_MORECORE(0));
+ if (old_br == sp->base + sp->size) {
+ char* rel_br = (char*)(CALL_MORECORE(-extra));
+ char* new_br = (char*)(CALL_MORECORE(0));
+ if (rel_br != CMFAIL && new_br < old_br)
+ released = old_br - new_br;
+ }
+ }
+ RELEASE_MALLOC_GLOBAL_LOCK();
+ }
+ }
+
+ if (released != 0) {
+ sp->size -= released;
+ m->footprint -= released;
+ init_top(m, m->top, m->topsize - released);
+ check_top_chunk(m, m->top);
+ }
+ }
+
+ /* Unmap any unused mmapped segments */
+ if (HAVE_MMAP)
+ released += release_unused_segments(m);
+
+ /* On failure, disable autotrim to avoid repeated failed future calls */
+ if (released == 0 && m->topsize > m->trim_check)
+ m->trim_check = MAX_SIZE_T;
+ }
+
+ return (released != 0)? 1 : 0;
+}
+
+
+/* ---------------------------- malloc support --------------------------- */
+
+/* allocate a large request from the best fitting chunk in a treebin */
+static void* tmalloc_large(mstate m, size_t nb) {
+ tchunkptr v = 0;
+ size_t rsize = -nb; /* Unsigned negation */
+ tchunkptr t;
+ bindex_t idx;
+ compute_tree_index(nb, idx);
+ if ((t = *treebin_at(m, idx)) != 0) {
+ /* Traverse tree for this bin looking for node with size == nb */
+ size_t sizebits = nb << leftshift_for_tree_index(idx);
+ tchunkptr rst = 0; /* The deepest untaken right subtree */
+ for (;;) {
+ tchunkptr rt;
+ size_t trem = chunksize(t) - nb;
+ if (trem < rsize) {
+ v = t;
+ if ((rsize = trem) == 0)
+ break;
+ }
+ rt = t->child[1];
+ t = t->child[(sizebits >> (SIZE_T_BITSIZE-SIZE_T_ONE)) & 1];
+ if (rt != 0 && rt != t)
+ rst = rt;
+ if (t == 0) {
+ t = rst; /* set t to least subtree holding sizes > nb */
+ break;
+ }
+ sizebits <<= 1;
+ }
+ }
+ if (t == 0 && v == 0) { /* set t to root of next non-empty treebin */
+ binmap_t leftbits = left_bits(idx2bit(idx)) & m->treemap;
+ if (leftbits != 0) {
+ bindex_t i;
+ binmap_t leastbit = least_bit(leftbits);
+ compute_bit2idx(leastbit, i);
+ t = *treebin_at(m, i);
+ }
+ }
+
+ while (t != 0) { /* find smallest of tree or subtree */
+ size_t trem = chunksize(t) - nb;
+ if (trem < rsize) {
+ rsize = trem;
+ v = t;
+ }
+ t = leftmost_child(t);
+ }
+
+ /* If dv is a better fit, return 0 so malloc will use it */
+ if (v != 0 && rsize < (size_t)(m->dvsize - nb)) {
+ if (RTCHECK(ok_address(m, v))) { /* split */
+ mchunkptr r = chunk_plus_offset(v, nb);
+ assert(chunksize(v) == rsize + nb);
+ if (RTCHECK(ok_next(v, r))) {
+ unlink_large_chunk(m, v);
+ if (rsize < MIN_CHUNK_SIZE)
+ set_inuse_and_pinuse(m, v, (rsize + nb));
+ else {
+ set_size_and_pinuse_of_inuse_chunk(m, v, nb);
+ set_size_and_pinuse_of_free_chunk(r, rsize);
+ insert_chunk(m, r, rsize);
+ }
+ return chunk2mem(v);
+ }
+ }
+ CORRUPTION_ERROR_ACTION(m);
+ }
+ return 0;
+}
+
+/* allocate a small request from the best fitting chunk in a treebin */
+static void* tmalloc_small(mstate m, size_t nb) {
+ tchunkptr t, v;
+ size_t rsize;
+ bindex_t i;
+ binmap_t leastbit = least_bit(m->treemap);
+ compute_bit2idx(leastbit, i);
+ v = t = *treebin_at(m, i);
+ rsize = chunksize(t) - nb;
+
+ while ((t = leftmost_child(t)) != 0) {
+ size_t trem = chunksize(t) - nb;
+ if (trem < rsize) {
+ rsize = trem;
+ v = t;
+ }
+ }
+
+ if (RTCHECK(ok_address(m, v))) {
+ mchunkptr r = chunk_plus_offset(v, nb);
+ assert(chunksize(v) == rsize + nb);
+ if (RTCHECK(ok_next(v, r))) {
+ unlink_large_chunk(m, v);
+ if (rsize < MIN_CHUNK_SIZE)
+ set_inuse_and_pinuse(m, v, (rsize + nb));
+ else {
+ set_size_and_pinuse_of_inuse_chunk(m, v, nb);
+ set_size_and_pinuse_of_free_chunk(r, rsize);
+ replace_dv(m, r, rsize);
+ }
+ return chunk2mem(v);
+ }
+ }
+
+ CORRUPTION_ERROR_ACTION(m);
+ return 0;
+}
+
+/* --------------------------- realloc support --------------------------- */
+static void* internal_memalign(mstate m, size_t alignment, size_t bytes, unsigned flags);
+
+static void* internal_realloc(mstate m, void* oldmem, size_t bytes, size_t alignment, unsigned flags) {
+ if (bytes >= MAX_REQUEST) {
+ MALLOC_FAILURE_ACTION;
+ return 0;
+ }
+ if (!PREACTION(m)) {
+ mchunkptr oldp = mem2chunk(oldmem);
+ size_t oldsize = chunksize(oldp);
+ mchunkptr next = chunk_plus_offset(oldp, oldsize);
+ mchunkptr newp = 0;
+ void* extra = 0;
+
+ /* Try to either shrink or extend into top. Else malloc-copy-free */
+
+ if (RTCHECK(ok_address(m, oldp) && ok_inuse(oldp) &&
+ ok_next(oldp, next) && ok_pinuse(next))) {
+ size_t nb = request2size(bytes);
+ if (is_mmapped(oldp))
+ newp = mmap_resize(m, oldp, nb, flags);
+ else if (oldsize >= nb) { /* already big enough */
+ size_t rsize = oldsize - nb;
+ newp = oldp;
+ if (rsize >= MIN_CHUNK_SIZE) {
+ mchunkptr remainder = chunk_plus_offset(newp, nb);
+ set_inuse(m, newp, nb);
+ set_inuse_and_pinuse(m, remainder, rsize);
+ extra = chunk2mem(remainder);
+ }
+ }
+ else if (next == m->top && oldsize + m->topsize > nb) {
+ /* Expand into top */
+ size_t newsize = oldsize + m->topsize;
+ size_t newtopsize = newsize - nb;
+ mchunkptr newtop = chunk_plus_offset(oldp, nb);
+ set_inuse(m, oldp, nb);
+ newtop->head = newtopsize |PINUSE_BIT;
+ m->top = newtop;
+ m->topsize = newtopsize;
+ newp = oldp;
+ }
+ }
+ else {
+ USAGE_ERROR_ACTION(m, oldmem);
+ POSTACTION(m);
+ return 0;
+ }
+#if DEBUG
+ if (newp != 0) {
+ check_inuse_chunk(m, newp); /* Check requires lock */
+ }
+#endif
+
+ POSTACTION(m);
+
+ if (newp != 0) {
+ if (extra != 0) {
+ internal_free(m, extra);
+ }
+ return chunk2mem(newp);
+ }
+ else if (!(flags & M2_PREVENT_MOVE)) {
+ void* newmem = (alignment > MALLOC_ALIGNMENT) ?
+ internal_memalign(m, alignment, bytes, flags)
+ : internal_malloc(m, bytes, flags);
+ if (newmem != 0) {
+ size_t oc = oldsize - overhead_for(oldp);
+ memcpy(newmem, oldmem, (oc < bytes)? oc : bytes);
+ internal_free(m, oldmem);
+ }
+ return newmem;
+ }
+ }
+ return 0;
+}
+
+/* --------------------------- memalign support -------------------------- */
+
+static void* internal_memalign(mstate m, size_t alignment, size_t bytes, unsigned flags) {
+ if (alignment <= MALLOC_ALIGNMENT) /* Can just use malloc */
+ return internal_malloc(m, bytes, flags);
+ if (alignment < MIN_CHUNK_SIZE) /* must be at least a minimum chunk size */
+ alignment = MIN_CHUNK_SIZE;
+ if ((alignment & (alignment-SIZE_T_ONE)) != 0) {/* Ensure a power of 2 */
+ size_t a = MALLOC_ALIGNMENT << 1;
+ while (a < alignment) a <<= 1;
+ alignment = a;
+ }
+
+ if (bytes >= MAX_REQUEST - alignment) {
+ if (m != 0) { /* Test isn't needed but avoids compiler warning */
+ MALLOC_FAILURE_ACTION;
+ }
+ }
+ else {
+ size_t nb = request2size(bytes);
+ size_t req = nb + alignment + MIN_CHUNK_SIZE - CHUNK_OVERHEAD;
+ char* mem = (char*)internal_malloc(m, req, flags);
+ if (mem != 0) {
+ void* leader = 0;
+ void* trailer = 0;
+ mchunkptr p = mem2chunk(mem);
+
+ if (PREACTION(m)) return 0;
+ if ((((size_t)(mem)) % alignment) != 0) { /* misaligned */
+ /*
+ Find an aligned spot inside chunk. Since we need to give
+ back leading space in a chunk of at least MIN_CHUNK_SIZE, if
+ the first calculation places us at a spot with less than
+ MIN_CHUNK_SIZE leader, we can move to the next aligned spot.
+ We've allocated enough total room so that this is always
+ possible.
+ */
+ char* br = (char*)mem2chunk((size_t)(((size_t)(mem +
+ alignment -
+ SIZE_T_ONE)) &
+ -alignment));
+ char* pos = ((size_t)(br - (char*)(p)) >= MIN_CHUNK_SIZE)?
+ br : br+alignment;
+ mchunkptr newp = (mchunkptr)pos;
+ size_t leadsize = pos - (char*)(p);
+ size_t newsize = chunksize(p) - leadsize;
+
+ if (is_mmapped(p)) { /* For mmapped chunks, just adjust offset */
+ newp->prev_foot = p->prev_foot + leadsize;
+ newp->head = newsize;
+ }
+ else { /* Otherwise, give back leader, use the rest */
+ set_inuse(m, newp, newsize);
+ set_inuse(m, p, leadsize);
+ leader = chunk2mem(p);
+ }
+ p = newp;
+ }
+
+ /* Give back spare room at the end */
+ if (!is_mmapped(p)) {
+ size_t size = chunksize(p);
+ if (size > nb + MIN_CHUNK_SIZE) {
+ size_t remainder_size = size - nb;
+ mchunkptr remainder = chunk_plus_offset(p, nb);
+ set_inuse(m, p, nb);
+ set_inuse(m, remainder, remainder_size);
+ trailer = chunk2mem(remainder);
+ }
+ }
+
+ assert (chunksize(p) >= nb);
+ assert((((size_t)(chunk2mem(p))) % alignment) == 0);
+ check_inuse_chunk(m, p);
+ POSTACTION(m);
+ if (leader != 0) {
+ internal_free(m, leader);
+ }
+ if (trailer != 0) {
+ internal_free(m, trailer);
+ }
+ return chunk2mem(p);
+ }
+ }
+ return 0;
+}
+
+/* ------------------------ comalloc/coalloc support --------------------- */
+
+static void** ialloc(mstate m,
+ size_t n_elements,
+ size_t* sizes,
+ int opts,
+ void* chunks[]) {
+ /*
+ This provides common support for independent_X routines, handling
+ all of the combinations that can result.
+
+ The opts arg has:
+ bit 0 set if all elements are same size (using sizes[0])
+ bit 1 set if elements should be zeroed
+ */
+
+ size_t element_size; /* chunksize of each element, if all same */
+ size_t contents_size; /* total size of elements */
+ size_t array_size; /* request size of pointer array */
+ void* mem; /* malloced aggregate space */
+ mchunkptr p; /* corresponding chunk */
+ size_t remainder_size; /* remaining bytes while splitting */
+ void** marray; /* either "chunks" or malloced ptr array */
+ mchunkptr array_chunk; /* chunk for malloced ptr array */
+ flag_t was_enabled; /* to disable mmap */
+ size_t size;
+ size_t i;
+
+ ensure_initialization();
+ /* compute array length, if needed */
+ if (chunks != 0) {
+ if (n_elements == 0)
+ return chunks; /* nothing to do */
+ marray = chunks;
+ array_size = 0;
+ }
+ else {
+ /* if empty req, must still return chunk representing empty array */
+ if (n_elements == 0)
+ return (void**)internal_malloc(m, 0, 0);
+ marray = 0;
+ array_size = request2size(n_elements * (sizeof(void*)));
+ }
+
+ /* compute total element size */
+ if (opts & 0x1) { /* all-same-size */
+ element_size = request2size(*sizes);
+ contents_size = n_elements * element_size;
+ }
+ else { /* add up all the sizes */
+ element_size = 0;
+ contents_size = 0;
+ for (i = 0; i != n_elements; ++i)
+ contents_size += request2size(sizes[i]);
+ }
+
+ size = contents_size + array_size;
+
+ /*
+ Allocate the aggregate chunk. First disable direct-mmapping so
+ malloc won't use it, since we would not be able to later
+ free/realloc space internal to a segregated mmap region.
+ */
+ was_enabled = use_mmap(m);
+ disable_mmap(m);
+ mem = internal_malloc(m, size - CHUNK_OVERHEAD, 0);
+ if (was_enabled)
+ enable_mmap(m);
+ if (mem == 0)
+ return 0;
+
+ if (PREACTION(m)) return 0;
+ p = mem2chunk(mem);
+ remainder_size = chunksize(p);
+
+ assert(!is_mmapped(p));
+
+ if (opts & 0x2) { /* optionally clear the elements */
+ memset((size_t*)mem, 0, remainder_size - SIZE_T_SIZE - array_size);
+ }
+
+ /* If not provided, allocate the pointer array as final part of chunk */
+ if (marray == 0) {
+ size_t array_chunk_size;
+ array_chunk = chunk_plus_offset(p, contents_size);
+ array_chunk_size = remainder_size - contents_size;
+ marray = (void**) (chunk2mem(array_chunk));
+ set_size_and_pinuse_of_inuse_chunk(m, array_chunk, array_chunk_size);
+ remainder_size = contents_size;
+ }
+
+ /* split out elements */
+ for (i = 0; ; ++i) {
+ marray[i] = chunk2mem(p);
+ if (i != n_elements-1) {
+ if (element_size != 0)
+ size = element_size;
+ else
+ size = request2size(sizes[i]);
+ remainder_size -= size;
+ set_size_and_pinuse_of_inuse_chunk(m, p, size);
+ p = chunk_plus_offset(p, size);
+ }
+ else { /* the final element absorbs any overallocation slop */
+ set_size_and_pinuse_of_inuse_chunk(m, p, remainder_size);
+ break;
+ }
+ }
+
+#if DEBUG
+ if (marray != chunks) {
+ /* final element must have exactly exhausted chunk */
+ if (element_size != 0) {
+ assert(remainder_size == element_size);
+ }
+ else {
+ assert(remainder_size == request2size(sizes[i]));
+ }
+ check_inuse_chunk(m, mem2chunk(marray));
+ }
+ for (i = 0; i != n_elements; ++i)
+ check_inuse_chunk(m, mem2chunk(marray[i]));
+
+#endif /* DEBUG */
+
+ POSTACTION(m);
+ return marray;
+}
+
+
+/* -------------------------- public routines ---------------------------- */
+
+#if !ONLY_MSPACES
+
+void* dlmalloc(size_t bytes) {
+ /*
+ Basic algorithm:
+ If a small request (< 256 bytes minus per-chunk overhead):
+ 1. If one exists, use a remainderless chunk in associated smallbin.
+ (Remainderless means that there are too few excess bytes to
+ represent as a chunk.)
+ 2. If it is big enough, use the dv chunk, which is normally the
+ chunk adjacent to the one used for the most recent small request.
+ 3. If one exists, split the smallest available chunk in a bin,
+ saving remainder in dv.
+ 4. If it is big enough, use the top chunk.
+ 5. If available, get memory from system and use it
+ Otherwise, for a large request:
+ 1. Find the smallest available binned chunk that fits, and use it
+ if it is better fitting than dv chunk, splitting if necessary.
+ 2. If better fitting than any binned chunk, use the dv chunk.
+ 3. If it is big enough, use the top chunk.
+ 4. If request size >= mmap threshold, try to directly mmap this chunk.
+ 5. If available, get memory from system and use it
+
+ The ugly goto's here ensure that postaction occurs along all paths.
+ */
+
+#if USE_LOCKS
+ ensure_initialization(); /* initialize in sys_alloc if not using locks */
+#endif
+
+ if (!PREACTION(gm)) {
+ void* mem;
+ size_t nb;
+ if (bytes <= MAX_SMALL_REQUEST) {
+ bindex_t idx;
+ binmap_t smallbits;
+ nb = (bytes < MIN_REQUEST)? MIN_CHUNK_SIZE : pad_request(bytes);
+ idx = small_index(nb);
+ smallbits = gm->smallmap >> idx;
+
+ if ((smallbits & 0x3U) != 0) { /* Remainderless fit to a smallbin. */
+ mchunkptr b, p;
+ idx += ~smallbits & 1; /* Uses next bin if idx empty */
+ b = smallbin_at(gm, idx);
+ p = b->fd;
+ assert(chunksize(p) == small_index2size(idx));
+ unlink_first_small_chunk(gm, b, p, idx);
+ set_inuse_and_pinuse(gm, p, small_index2size(idx));
+ mem = chunk2mem(p);
+ check_malloced_chunk(gm, mem, nb);
+ goto postaction;
+ }
+
+ else if (nb > gm->dvsize) {
+ if (smallbits != 0) { /* Use chunk in next nonempty smallbin */
+ mchunkptr b, p, r;
+ size_t rsize;
+ bindex_t i;
+ binmap_t leftbits = (smallbits << idx) & left_bits(idx2bit(idx));
+ binmap_t leastbit = least_bit(leftbits);
+ compute_bit2idx(leastbit, i);
+ b = smallbin_at(gm, i);
+ p = b->fd;
+ assert(chunksize(p) == small_index2size(i));
+ unlink_first_small_chunk(gm, b, p, i);
+ rsize = small_index2size(i) - nb;
+ /* Fit here cannot be remainderless if 4byte sizes */
+ if (SIZE_T_SIZE != 4 && rsize < MIN_CHUNK_SIZE)
+ set_inuse_and_pinuse(gm, p, small_index2size(i));
+ else {
+ set_size_and_pinuse_of_inuse_chunk(gm, p, nb);
+ r = chunk_plus_offset(p, nb);
+ set_size_and_pinuse_of_free_chunk(r, rsize);
+ replace_dv(gm, r, rsize);
+ }
+ mem = chunk2mem(p);
+ check_malloced_chunk(gm, mem, nb);
+ goto postaction;
+ }
+
+ else if (gm->treemap != 0 && (mem = tmalloc_small(gm, nb)) != 0) {
+ check_malloced_chunk(gm, mem, nb);
+ goto postaction;
+ }
+ }
+ }
+ else if (bytes >= MAX_REQUEST)
+ nb = MAX_SIZE_T; /* Too big to allocate. Force failure (in sys alloc) */
+ else {
+ nb = pad_request(bytes);
+ if (gm->treemap != 0 && (mem = tmalloc_large(gm, nb)) != 0) {
+ check_malloced_chunk(gm, mem, nb);
+ goto postaction;
+ }
+ }
+
+ if (nb <= gm->dvsize) {
+ size_t rsize = gm->dvsize - nb;
+ mchunkptr p = gm->dv;
+ if (rsize >= MIN_CHUNK_SIZE) { /* split dv */
+ mchunkptr r = gm->dv = chunk_plus_offset(p, nb);
+ gm->dvsize = rsize;
+ set_size_and_pinuse_of_free_chunk(r, rsize);
+ set_size_and_pinuse_of_inuse_chunk(gm, p, nb);
+ }
+ else { /* exhaust dv */
+ size_t dvs = gm->dvsize;
+ gm->dvsize = 0;
+ gm->dv = 0;
+ set_inuse_and_pinuse(gm, p, dvs);
+ }
+ mem = chunk2mem(p);
+ check_malloced_chunk(gm, mem, nb);
+ goto postaction;
+ }
+
+ else if (nb < gm->topsize) { /* Split top */
+ size_t rsize = gm->topsize -= nb;
+ mchunkptr p = gm->top;
+ mchunkptr r = gm->top = chunk_plus_offset(p, nb);
+ r->head = rsize | PINUSE_BIT;
+ set_size_and_pinuse_of_inuse_chunk(gm, p, nb);
+ mem = chunk2mem(p);
+ check_top_chunk(gm, gm->top);
+ check_malloced_chunk(gm, mem, nb);
+ goto postaction;
+ }
+
+ mem = sys_alloc(gm, nb);
+
+ postaction:
+ POSTACTION(gm);
+ return mem;
+ }
+
+ return 0;
+}
+
+void dlfree(void* mem) {
+ /*
+ Consolidate freed chunks with preceeding or succeeding bordering
+ free chunks, if they exist, and then place in a bin. Intermixed
+ with special cases for top, dv, mmapped chunks, and usage errors.
+ */
+
+ if (mem != 0) {
+ mchunkptr p = mem2chunk(mem);
+#if FOOTERS
+ mstate fm = get_mstate_for(p);
+ if (!ok_magic(fm)) {
+ USAGE_ERROR_ACTION(fm, p);
+ return;
+ }
+#else /* FOOTERS */
+#define fm gm
+#endif /* FOOTERS */
+ if (!PREACTION(fm)) {
+ check_inuse_chunk(fm, p);
+ if (RTCHECK(ok_address(fm, p) && ok_inuse(p))) {
+ size_t psize = chunksize(p);
+ mchunkptr next = chunk_plus_offset(p, psize);
+ if (!pinuse(p)) {
+ size_t prevsize = p->prev_foot;
+ if (is_mmapped(p)) {
+ char* mm = (char*)p - prevsize;
+ psize += prevsize + MMAP_FOOT_PAD;
+ if (CALL_MUNMAP(*(void**)mm, mm, psize) == 0)
+ fm->footprint -= psize;
+ goto postaction;
+ }
+ else {
+ mchunkptr prev = chunk_minus_offset(p, prevsize);
+ psize += prevsize;
+ p = prev;
+ if (RTCHECK(ok_address(fm, prev))) { /* consolidate backward */
+ if (p != fm->dv) {
+ unlink_chunk(fm, p, prevsize);
+ }
+ else if ((next->head & INUSE_BITS) == INUSE_BITS) {
+ fm->dvsize = psize;
+ set_free_with_pinuse(p, psize, next);
+ goto postaction;
+ }
+ }
+ else
+ goto erroraction;
+ }
+ }
+
+ if (RTCHECK(ok_next(p, next) && ok_pinuse(next))) {
+ if (!cinuse(next)) { /* consolidate forward */
+ if (next == fm->top) {
+ size_t tsize = fm->topsize += psize;
+ fm->top = p;
+ p->head = tsize | PINUSE_BIT;
+ if (p == fm->dv) {
+ fm->dv = 0;
+ fm->dvsize = 0;
+ }
+ if (should_trim(fm, tsize))
+ sys_trim(fm, 0);
+ goto postaction;
+ }
+ else if (next == fm->dv) {
+ size_t dsize = fm->dvsize += psize;
+ fm->dv = p;
+ set_size_and_pinuse_of_free_chunk(p, dsize);
+ goto postaction;
+ }
+ else {
+ size_t nsize = chunksize(next);
+ psize += nsize;
+ unlink_chunk(fm, next, nsize);
+ set_size_and_pinuse_of_free_chunk(p, psize);
+ if (p == fm->dv) {
+ fm->dvsize = psize;
+ goto postaction;
+ }
+ }
+ }
+ else
+ set_free_with_pinuse(p, psize, next);
+
+ if (is_small(psize)) {
+ insert_small_chunk(fm, p, psize);
+ check_free_chunk(fm, p);
+ }
+ else {
+ tchunkptr tp = (tchunkptr)p;
+ insert_large_chunk(fm, tp, psize);
+ check_free_chunk(fm, p);
+ if (--fm->release_checks == 0)
+ release_unused_segments(fm);
+ }
+ goto postaction;
+ }
+ }
+ erroraction:
+ USAGE_ERROR_ACTION(fm, p);
+ postaction:
+ POSTACTION(fm);
+ }
+ }
+#if !FOOTERS
+#undef fm
+#endif /* FOOTERS */
+}
+
+void* dlcalloc(size_t n_elements, size_t elem_size) {
+ void* mem;
+ size_t req = 0;
+ if (n_elements != 0) {
+ req = n_elements * elem_size;
+ if (((n_elements | elem_size) & ~(size_t)0xffff) &&
+ (req / n_elements != elem_size))
+ req = MAX_SIZE_T; /* force downstream failure on overflow */
+ }
+ mem = dlmalloc(req);
+ if (mem != 0 && calloc_must_clear(mem2chunk(mem)))
+ memset(mem, 0, req);
+ return mem;
+}
+
+void* dlrealloc(void* oldmem, size_t bytes) {
+ if (oldmem == 0)
+ return dlmalloc(bytes);
+#ifdef REALLOC_ZERO_BYTES_FREES
+ if (bytes == 0) {
+ dlfree(oldmem);
+ return 0;
+ }
+#endif /* REALLOC_ZERO_BYTES_FREES */
+ else {
+#if ! FOOTERS
+ mstate m = gm;
+#else /* FOOTERS */
+ mstate m = get_mstate_for(mem2chunk(oldmem));
+ if (!ok_magic(m)) {
+ USAGE_ERROR_ACTION(m, oldmem);
+ return 0;
+ }
+#endif /* FOOTERS */
+ return internal_realloc(m, oldmem, bytes);
+ }
+}
+
+void* dlmemalign(size_t alignment, size_t bytes) {
+ return internal_memalign(gm, alignment, bytes);
+}
+
+void** dlindependent_calloc(size_t n_elements, size_t elem_size,
+ void* chunks[]) {
+ size_t sz = elem_size; /* serves as 1-element array */
+ return ialloc(gm, n_elements, &sz, 3, chunks);
+}
+
+void** dlindependent_comalloc(size_t n_elements, size_t sizes[],
+ void* chunks[]) {
+ return ialloc(gm, n_elements, sizes, 0, chunks);
+}
+
+void* dlvalloc(size_t bytes) {
+ size_t pagesz;
+ ensure_initialization();
+ pagesz = mparams.page_size;
+ return dlmemalign(pagesz, bytes);
+}
+
+void* dlpvalloc(size_t bytes) {
+ size_t pagesz;
+ ensure_initialization();
+ pagesz = mparams.page_size;
+ return dlmemalign(pagesz, (bytes + pagesz - SIZE_T_ONE) & ~(pagesz - SIZE_T_ONE));
+}
+
+int dlmalloc_trim(size_t pad) {
+ int result = 0;
+ ensure_initialization();
+ if (!PREACTION(gm)) {
+ result = sys_trim(gm, pad);
+ POSTACTION(gm);
+ }
+ return result;
+}
+
+size_t dlmalloc_footprint(void) {
+ return gm->footprint;
+}
+
+size_t dlmalloc_max_footprint(void) {
+ return gm->max_footprint;
+}
+
+#if !NO_MALLINFO
+struct mallinfo dlmallinfo(void) {
+ return internal_mallinfo(gm);
+}
+#endif /* NO_MALLINFO */
+
+void dlmalloc_stats() {
+ internal_malloc_stats(gm);
+}
+
+int dlmallopt(int param_number, int value) {
+ return change_mparam(param_number, value);
+}
+
+#endif /* !ONLY_MSPACES */
+
+size_t dlmalloc_usable_size(void* mem) {
+ if (mem != 0) {
+ mchunkptr p = mem2chunk(mem);
+ if (is_inuse(p))
+ return chunksize(p) - overhead_for(p);
+ }
+ return 0;
+}
+
+/* ----------------------------- user mspaces ---------------------------- */
+
+#if MSPACES
+
+static mstate init_user_mstate(char* tbase, size_t tsize) {
+ size_t msize = pad_request(sizeof(struct malloc_state));
+ mchunkptr mn;
+ mchunkptr msp = align_as_chunk(tbase);
+ mstate m = (mstate)(chunk2mem(msp));
+ memset(m, 0, msize);
+ INITIAL_LOCK(&m->mutex);
+ msp->head = (msize|INUSE_BITS);
+ m->seg.base = m->least_addr = tbase;
+ m->seg.size = m->footprint = m->max_footprint = tsize;
+ m->magic = mparams.magic;
+ m->release_checks = MAX_RELEASE_CHECK_RATE;
+ m->mflags = mparams.default_mflags;
+ m->extp = 0;
+ m->exts = 0;
+ disable_contiguous(m);
+ init_bins(m);
+ mn = next_chunk(mem2chunk(m));
+ init_top(m, mn, (size_t)((tbase + tsize) - (char*)mn) - TOP_FOOT_SIZE);
+ check_top_chunk(m, m->top);
+ return m;
+}
+
+mspace create_mspace(size_t capacity, int locked) {
+ mstate m = 0;
+ size_t msize;
+ ensure_initialization();
+ msize = pad_request(sizeof(struct malloc_state));
+ if (capacity < (size_t) -(msize + TOP_FOOT_SIZE + mparams.page_size)) {
+ size_t rs = ((capacity == 0)? mparams.granularity :
+ (capacity + TOP_FOOT_SIZE + msize));
+ size_t tsize = granularity_align(rs);
+ char* tbase = (char*)(CALL_MMAP(tsize, 0));
+ if (tbase != CMFAIL) {
+ m = init_user_mstate(tbase, tsize);
+ m->seg.sflags = USE_MMAP_BIT;
+ set_lock(m, locked);
+ }
+ }
+ return (mspace)m;
+}
+
+mspace create_mspace_with_base(void* base, size_t capacity, int locked) {
+ mstate m = 0;
+ size_t msize;
+ ensure_initialization();
+ msize = pad_request(sizeof(struct malloc_state));
+ if (capacity > msize + TOP_FOOT_SIZE &&
+ capacity < (size_t) -(msize + TOP_FOOT_SIZE + mparams.page_size)) {
+ m = init_user_mstate((char*)base, capacity);
+ m->seg.sflags = EXTERN_BIT;
+ set_lock(m, locked);
+ }
+ return (mspace)m;
+}
+
+int mspace_track_large_chunks(mspace msp, int enable) {
+ int ret = 0;
+ mstate ms = (mstate)msp;
+ if (!PREACTION(ms)) {
+ if (!use_mmap(ms))
+ ret = 1;
+ if (!enable)
+ enable_mmap(ms);
+ else
+ disable_mmap(ms);
+ POSTACTION(ms);
+ }
+ return ret;
+}
+
+size_t destroy_mspace(mspace msp) {
+ size_t freed = 0;
+ mstate ms = (mstate)msp;
+ if (ok_magic(ms)) {
+ msegmentptr sp = &ms->seg;
+ while (sp != 0) {
+ char* base = sp->base;
+ size_t size = sp->size;
+ flag_t flag = sp->sflags;
+ sp = sp->next;
+ if ((flag & USE_MMAP_BIT) && !(flag & EXTERN_BIT) &&
+ CALL_MUNMAP(0/*segment*/, base, size) == 0)
+ freed += size;
+ }
+ DESTROY_LOCK(&ms->mutex);
+ }
+ else {
+ USAGE_ERROR_ACTION(ms,ms);
+ }
+ return freed;
+}
+
+/*
+ mspace versions of routines are near-clones of the global
+ versions. This is not so nice but better than the alternatives.
+*/
+
+static FORCEINLINE void* mspace_malloc_implementation(mstate ms, size_t bytes, unsigned flags) {
+ if (!PREACTION(ms)) {
+ void* mem;
+ size_t nb = bytes;
+ if (!(flags & M2_ALWAYS_MMAP)) {
+ if (bytes <= MAX_SMALL_REQUEST) {
+ bindex_t idx;
+ binmap_t smallbits;
+ nb = (bytes < MIN_REQUEST)? MIN_CHUNK_SIZE : pad_request(bytes);
+ idx = small_index(nb);
+ smallbits = ms->smallmap >> idx;
+
+ if ((smallbits & 0x3U) != 0) { /* Remainderless fit to a smallbin. */
+ mchunkptr b, p;
+ idx += ~smallbits & 1; /* Uses next bin if idx empty */
+ b = smallbin_at(ms, idx);
+ p = b->fd;
+ assert(chunksize(p) == small_index2size(idx));
+ unlink_first_small_chunk(ms, b, p, idx);
+ set_inuse_and_pinuse(ms, p, small_index2size(idx));
+ mem = chunk2mem(p);
+ check_malloced_chunk(ms, mem, nb);
+ goto postaction;
+ }
+
+ else if (nb > ms->dvsize) {
+ if (smallbits != 0) { /* Use chunk in next nonempty smallbin */
+ mchunkptr b, p, r;
+ size_t rsize;
+ bindex_t i;
+ binmap_t leftbits = (smallbits << idx) & left_bits(idx2bit(idx));
+ binmap_t leastbit = least_bit(leftbits);
+ compute_bit2idx(leastbit, i);
+ b = smallbin_at(ms, i);
+ p = b->fd;
+ assert(chunksize(p) == small_index2size(i));
+ unlink_first_small_chunk(ms, b, p, i);
+ rsize = small_index2size(i) - nb;
+ /* Fit here cannot be remainderless if 4byte sizes */
+ if (SIZE_T_SIZE != 4 && rsize < MIN_CHUNK_SIZE)
+ set_inuse_and_pinuse(ms, p, small_index2size(i));
+ else {
+ set_size_and_pinuse_of_inuse_chunk(ms, p, nb);
+ r = chunk_plus_offset(p, nb);
+ set_size_and_pinuse_of_free_chunk(r, rsize);
+ replace_dv(ms, r, rsize);
+ }
+ mem = chunk2mem(p);
+ check_malloced_chunk(ms, mem, nb);
+ goto postaction;
+ }
+
+ else if (ms->treemap != 0 && (mem = tmalloc_small(ms, nb)) != 0) {
+ check_malloced_chunk(ms, mem, nb);
+ goto postaction;
+ }
+ }
+ }
+ else if (bytes >= MAX_REQUEST)
+ nb = MAX_SIZE_T; /* Too big to allocate. Force failure (in sys alloc) */
+ else {
+ nb = pad_request(bytes);
+ if (ms->treemap != 0 && (mem = tmalloc_large(ms, nb)) != 0) {
+ check_malloced_chunk(ms, mem, nb);
+ goto postaction;
+ }
+ }
+
+ if (nb <= ms->dvsize) {
+ size_t rsize = ms->dvsize - nb;
+ mchunkptr p = ms->dv;
+ if (rsize >= MIN_CHUNK_SIZE) { /* split dv */
+ mchunkptr r = ms->dv = chunk_plus_offset(p, nb);
+ ms->dvsize = rsize;
+ set_size_and_pinuse_of_free_chunk(r, rsize);
+ set_size_and_pinuse_of_inuse_chunk(ms, p, nb);
+ }
+ else { /* exhaust dv */
+ size_t dvs = ms->dvsize;
+ ms->dvsize = 0;
+ ms->dv = 0;
+ set_inuse_and_pinuse(ms, p, dvs);
+ }
+ mem = chunk2mem(p);
+ check_malloced_chunk(ms, mem, nb);
+ goto postaction;
+ }
+
+ else if (nb < ms->topsize) { /* Split top */
+ size_t rsize = ms->topsize -= nb;
+ mchunkptr p = ms->top;
+ mchunkptr r = ms->top = chunk_plus_offset(p, nb);
+ r->head = rsize | PINUSE_BIT;
+ set_size_and_pinuse_of_inuse_chunk(ms, p, nb);
+ mem = chunk2mem(p);
+ check_top_chunk(ms, ms->top);
+ check_malloced_chunk(ms, mem, nb);
+ goto postaction;
+ }
+ }
+
+ mem = sys_alloc(ms, nb, flags);
+
+ postaction:
+ POSTACTION(ms);
+ return mem;
+ }
+
+ return 0;
+}
+void* mspace_malloc(mspace msp, size_t bytes) {
+ mstate ms = (mstate)msp;
+ if (!ok_magic(ms)) {
+ USAGE_ERROR_ACTION(ms,ms);
+ return 0;
+ }
+ return mspace_malloc_implementation(ms, bytes, 0);
+}
+void* mspace_malloc2(mspace msp, size_t bytes, size_t alignment, unsigned flags) {
+ void* mem;
+ mstate ms = (mstate)msp;
+ if (!ok_magic(ms)) {
+ USAGE_ERROR_ACTION(ms,ms);
+ return 0;
+ }
+ if (alignment <= MALLOC_ALIGNMENT)
+ mem = mspace_malloc_implementation(ms, bytes, flags);
+ else
+ mem = internal_memalign(ms, alignment, bytes, flags);
+ if (mem && (flags & M2_ZERO_MEMORY)) {
+ mchunkptr p = mem2chunk(mem);
+ if (calloc_must_clear(p))
+ memset(mem, 0, chunksize(p) - overhead_for(p));
+ }
+ return mem;
+}
+
+void mspace_free(mspace msp, void* mem) {
+ if (mem != 0) {
+ mchunkptr p = mem2chunk(mem);
+#if FOOTERS
+ mstate fm = get_mstate_for(p);
+ msp = msp; /* placate people compiling -Wunused */
+#else /* FOOTERS */
+ mstate fm = (mstate)msp;
+#endif /* FOOTERS */
+ if (!ok_magic(fm)) {
+ USAGE_ERROR_ACTION(fm, p);
+ return;
+ }
+ if (!PREACTION(fm)) {
+ check_inuse_chunk(fm, p);
+ if (RTCHECK(ok_address(fm, p) && ok_inuse(p))) {
+ size_t psize = chunksize(p);
+ mchunkptr next = chunk_plus_offset(p, psize);
+ if (!pinuse(p)) {
+ size_t prevsize = p->prev_foot;
+ if (is_mmapped(p)) {
+ char* mm = (char*)p - prevsize;
+ psize += prevsize + MMAP_FOOT_PAD;
+ if (CALL_MUNMAP(*(void**)mm, mm, psize) == 0)
+ fm->footprint -= psize;
+ goto postaction;
+ }
+ else {
+ mchunkptr prev = chunk_minus_offset(p, prevsize);
+ psize += prevsize;
+ p = prev;
+ if (RTCHECK(ok_address(fm, prev))) { /* consolidate backward */
+ if (p != fm->dv) {
+ unlink_chunk(fm, p, prevsize);
+ }
+ else if ((next->head & INUSE_BITS) == INUSE_BITS) {
+ fm->dvsize = psize;
+ set_free_with_pinuse(p, psize, next);
+ goto postaction;
+ }
+ }
+ else
+ goto erroraction;
+ }
+ }
+
+ if (RTCHECK(ok_next(p, next) && ok_pinuse(next))) {
+ if (!cinuse(next)) { /* consolidate forward */
+ if (next == fm->top) {
+ size_t tsize = fm->topsize += psize;
+ fm->top = p;
+ p->head = tsize | PINUSE_BIT;
+ if (p == fm->dv) {
+ fm->dv = 0;
+ fm->dvsize = 0;
+ }
+ if (should_trim(fm, tsize))
+ sys_trim(fm, 0);
+ goto postaction;
+ }
+ else if (next == fm->dv) {
+ size_t dsize = fm->dvsize += psize;
+ fm->dv = p;
+ set_size_and_pinuse_of_free_chunk(p, dsize);
+ goto postaction;
+ }
+ else {
+ size_t nsize = chunksize(next);
+ psize += nsize;
+ unlink_chunk(fm, next, nsize);
+ set_size_and_pinuse_of_free_chunk(p, psize);
+ if (p == fm->dv) {
+ fm->dvsize = psize;
+ goto postaction;
+ }
+ }
+ }
+ else
+ set_free_with_pinuse(p, psize, next);
+
+ if (is_small(psize)) {
+ insert_small_chunk(fm, p, psize);
+ check_free_chunk(fm, p);
+ }
+ else {
+ tchunkptr tp = (tchunkptr)p;
+ insert_large_chunk(fm, tp, psize);
+ check_free_chunk(fm, p);
+ if (--fm->release_checks == 0)
+ release_unused_segments(fm);
+ }
+ goto postaction;
+ }
+ }
+ erroraction:
+ USAGE_ERROR_ACTION(fm, p);
+ postaction:
+ POSTACTION(fm);
+ }
+ }
+}
+
+void* mspace_calloc(mspace msp, size_t n_elements, size_t elem_size) {
+ void* mem;
+ size_t req = 0;
+ mstate ms = (mstate)msp;
+ if (!ok_magic(ms)) {
+ USAGE_ERROR_ACTION(ms,ms);
+ return 0;
+ }
+ if (n_elements != 0) {
+ req = n_elements * elem_size;
+ if (((n_elements | elem_size) & ~(size_t)0xffff) &&
+ (req / n_elements != elem_size))
+ req = MAX_SIZE_T; /* force downstream failure on overflow */
+ }
+ mem = internal_malloc(ms, req, 0);
+ if (mem != 0) {
+ mchunkptr p = mem2chunk(mem);
+ if (calloc_must_clear(p))
+ memset(mem, 0, chunksize(p) - overhead_for(p));
+ }
+ return mem;
+}
+
+void* mspace_realloc2(mspace msp, void* oldmem, size_t bytes, size_t alignment, unsigned flags) {
+ if (oldmem == 0)
+ return mspace_malloc2(msp, bytes, alignment, flags);
+#ifdef REALLOC_ZERO_BYTES_FREES
+ if (bytes == 0) {
+ mspace_free(msp, oldmem);
+ return 0;
+ }
+#endif /* REALLOC_ZERO_BYTES_FREES */
+ else {
+ void* mem;
+ mchunkptr p = mem2chunk(oldmem);
+ size_t oldsize = chunksize(p) - overhead_for(p);
+#if FOOTERS
+ mstate ms = get_mstate_for(p);
+#else /* FOOTERS */
+ mstate ms = (mstate)msp;
+#endif /* FOOTERS */
+ if (!ok_magic(ms)) {
+ USAGE_ERROR_ACTION(ms,ms);
+ return 0;
+ }
+ mem = internal_realloc(ms, oldmem, bytes, alignment, flags);
+ if (mem && (flags & M2_ZERO_MEMORY) && bytes > oldsize) {
+ p = mem2chunk(mem);
+ if (calloc_must_clear(p)) {
+ size_t newsize = chunksize(p) - overhead_for(p);
+ memset((char*)mem + oldsize, 0, newsize - oldsize);
+ }
+ }
+ return mem;
+ }
+}
+void* mspace_realloc(mspace msp, void* oldmem, size_t bytes) {
+ return mspace_realloc2(msp, oldmem, bytes, 0, 0);
+}
+
+void* mspace_memalign(mspace msp, size_t alignment, size_t bytes) {
+ mstate ms = (mstate)msp;
+ if (!ok_magic(ms)) {
+ USAGE_ERROR_ACTION(ms,ms);
+ return 0;
+ }
+ return internal_memalign(ms, alignment, bytes, 0);
+}
+
+void** mspace_independent_calloc(mspace msp, size_t n_elements,
+ size_t elem_size, void* chunks[]) {
+ size_t sz = elem_size; /* serves as 1-element array */
+ mstate ms = (mstate)msp;
+ if (!ok_magic(ms)) {
+ USAGE_ERROR_ACTION(ms,ms);
+ return 0;
+ }
+ return ialloc(ms, n_elements, &sz, 3, chunks);
+}
+
+void** mspace_independent_comalloc(mspace msp, size_t n_elements,
+ size_t sizes[], void* chunks[]) {
+ mstate ms = (mstate)msp;
+ if (!ok_magic(ms)) {
+ USAGE_ERROR_ACTION(ms,ms);
+ return 0;
+ }
+ return ialloc(ms, n_elements, sizes, 0, chunks);
+}
+
+int mspace_trim(mspace msp, size_t pad) {
+ int result = 0;
+ mstate ms = (mstate)msp;
+ if (ok_magic(ms)) {
+ if (!PREACTION(ms)) {
+ result = sys_trim(ms, pad);
+ POSTACTION(ms);
+ }
+ }
+ else {
+ USAGE_ERROR_ACTION(ms,ms);
+ }
+ return result;
+}
+
+void mspace_malloc_stats(mspace msp) {
+ mstate ms = (mstate)msp;
+ if (ok_magic(ms)) {
+ internal_malloc_stats(ms);
+ }
+ else {
+ USAGE_ERROR_ACTION(ms,ms);
+ }
+}
+
+size_t mspace_footprint(mspace msp) {
+ size_t result = 0;
+ mstate ms = (mstate)msp;
+ if (ok_magic(ms)) {
+ result = ms->footprint;
+ }
+ else {
+ USAGE_ERROR_ACTION(ms,ms);
+ }
+ return result;
+}
+
+
+size_t mspace_max_footprint(mspace msp) {
+ size_t result = 0;
+ mstate ms = (mstate)msp;
+ if (ok_magic(ms)) {
+ result = ms->max_footprint;
+ }
+ else {
+ USAGE_ERROR_ACTION(ms,ms);
+ }
+ return result;
+}
+
+
+#if !NO_MALLINFO
+struct mallinfo mspace_mallinfo(mspace msp) {
+ mstate ms = (mstate)msp;
+ if (!ok_magic(ms)) {
+ USAGE_ERROR_ACTION(ms,ms);
+ }
+ return internal_mallinfo(ms);
+}
+#endif /* NO_MALLINFO */
+
+size_t mspace_usable_size(void* mem) {
+ if (mem != 0) {
+ mchunkptr p = mem2chunk(mem);
+ if (is_inuse(p))
+ return chunksize(p) - overhead_for(p);
+ }
+ return 0;
+}
+
+int mspace_mallopt(int param_number, int value) {
+ return change_mparam(param_number, value);
+}
+
+#endif /* MSPACES */
+
+
+/* -------------------- Alternative MORECORE functions ------------------- */
+
+/*
+ Guidelines for creating a custom version of MORECORE:
+
+ * For best performance, MORECORE should allocate in multiples of pagesize.
+ * MORECORE may allocate more memory than requested. (Or even less,
+ but this will usually result in a malloc failure.)
+ * MORECORE must not allocate memory when given argument zero, but
+ instead return one past the end address of memory from previous
+ nonzero call.
+ * For best performance, consecutive calls to MORECORE with positive
+ arguments should return increasing addresses, indicating that
+ space has been contiguously extended.
+ * Even though consecutive calls to MORECORE need not return contiguous
+ addresses, it must be OK for malloc'ed chunks to span multiple
+ regions in those cases where they do happen to be contiguous.
+ * MORECORE need not handle negative arguments -- it may instead
+ just return MFAIL when given negative arguments.
+ Negative arguments are always multiples of pagesize. MORECORE
+ must not misinterpret negative args as large positive unsigned
+ args. You can suppress all such calls from even occurring by defining
+ MORECORE_CANNOT_TRIM,
+
+ As an example alternative MORECORE, here is a custom allocator
+ kindly contributed for pre-OSX macOS. It uses virtually but not
+ necessarily physically contiguous non-paged memory (locked in,
+ present and won't get swapped out). You can use it by uncommenting
+ this section, adding some #includes, and setting up the appropriate
+ defines above:
+
+ #define MORECORE osMoreCore
+
+ There is also a shutdown routine that should somehow be called for
+ cleanup upon program exit.
+
+ #define MAX_POOL_ENTRIES 100
+ #define MINIMUM_MORECORE_SIZE (64 * 1024U)
+ static int next_os_pool;
+ void *our_os_pools[MAX_POOL_ENTRIES];
+
+ void *osMoreCore(int size)
+ {
+ void *ptr = 0;
+ static void *sbrk_top = 0;
+
+ if (size > 0)
+ {
+ if (size < MINIMUM_MORECORE_SIZE)
+ size = MINIMUM_MORECORE_SIZE;
+ if (CurrentExecutionLevel() == kTaskLevel)
+ ptr = PoolAllocateResident(size + RM_PAGE_SIZE, 0);
+ if (ptr == 0)
+ {
+ return (void *) MFAIL;
+ }
+ // save ptrs so they can be freed during cleanup
+ our_os_pools[next_os_pool] = ptr;
+ next_os_pool++;
+ ptr = (void *) ((((size_t) ptr) + RM_PAGE_MASK) & ~RM_PAGE_MASK);
+ sbrk_top = (char *) ptr + size;
+ return ptr;
+ }
+ else if (size < 0)
+ {
+ // we don't currently support shrink behavior
+ return (void *) MFAIL;
+ }
+ else
+ {
+ return sbrk_top;
+ }
+ }
+
+ // cleanup any allocated memory pools
+ // called as last thing before shutting down driver
+
+ void osCleanupMem(void)
+ {
+ void **ptr;
+
+ for (ptr = our_os_pools; ptr < &our_os_pools[MAX_POOL_ENTRIES]; ptr++)
+ if (*ptr)
+ {
+ PoolDeallocate(*ptr);
+ *ptr = 0;
+ }
+ }
+
+*/
+
+
+/* -----------------------------------------------------------------------
+History:
+ V2.8.4 Wed May 27 09:56:23 2009 Doug Lea (dl at gee)
+ * Use zeros instead of prev foot for is_mmapped
+ * Add mspace_track_large_chunks; thanks to Jean Brouwers
+ * Fix set_inuse in internal_realloc; thanks to Jean Brouwers
+ * Fix insufficient sys_alloc padding when using 16byte alignment
+ * Fix bad error check in mspace_footprint
+ * Adaptations for ptmalloc; thanks to Wolfram Gloger.
+ * Reentrant spin locks; thanks to Earl Chew and others
+ * Win32 improvements; thanks to Niall Douglas and Earl Chew
+ * Add NO_SEGMENT_TRAVERSAL and MAX_RELEASE_CHECK_RATE options
+ * Extension hook in malloc_state
+ * Various small adjustments to reduce warnings on some compilers
+ * Various configuration extensions/changes for more platforms. Thanks
+ to all who contributed these.
+
+ V2.8.3 Thu Sep 22 11:16:32 2005 Doug Lea (dl at gee)
+ * Add max_footprint functions
+ * Ensure all appropriate literals are size_t
+ * Fix conditional compilation problem for some #define settings
+ * Avoid concatenating segments with the one provided
+ in create_mspace_with_base
+ * Rename some variables to avoid compiler shadowing warnings
+ * Use explicit lock initialization.
+ * Better handling of sbrk interference.
+ * Simplify and fix segment insertion, trimming and mspace_destroy
+ * Reinstate REALLOC_ZERO_BYTES_FREES option from 2.7.x
+ * Thanks especially to Dennis Flanagan for help on these.
+
+ V2.8.2 Sun Jun 12 16:01:10 2005 Doug Lea (dl at gee)
+ * Fix memalign brace error.
+
+ V2.8.1 Wed Jun 8 16:11:46 2005 Doug Lea (dl at gee)
+ * Fix improper #endif nesting in C++
+ * Add explicit casts needed for C++
+
+ V2.8.0 Mon May 30 14:09:02 2005 Doug Lea (dl at gee)
+ * Use trees for large bins
+ * Support mspaces
+ * Use segments to unify sbrk-based and mmap-based system allocation,
+ removing need for emulation on most platforms without sbrk.
+ * Default safety checks
+ * Optional footer checks. Thanks to William Robertson for the idea.
+ * Internal code refactoring
+ * Incorporate suggestions and platform-specific changes.
+ Thanks to Dennis Flanagan, Colin Plumb, Niall Douglas,
+ Aaron Bachmann, Emery Berger, and others.
+ * Speed up non-fastbin processing enough to remove fastbins.
+ * Remove useless cfree() to avoid conflicts with other apps.
+ * Remove internal memcpy, memset. Compilers handle builtins better.
+ * Remove some options that no one ever used and rename others.
+
+ V2.7.2 Sat Aug 17 09:07:30 2002 Doug Lea (dl at gee)
+ * Fix malloc_state bitmap array misdeclaration
+
+ V2.7.1 Thu Jul 25 10:58:03 2002 Doug Lea (dl at gee)
+ * Allow tuning of FIRST_SORTED_BIN_SIZE
+ * Use PTR_UINT as type for all ptr->int casts. Thanks to John Belmonte.
+ * Better detection and support for non-contiguousness of MORECORE.
+ Thanks to Andreas Mueller, Conal Walsh, and Wolfram Gloger
+ * Bypass most of malloc if no frees. Thanks To Emery Berger.
+ * Fix freeing of old top non-contiguous chunk im sysmalloc.
+ * Raised default trim and map thresholds to 256K.
+ * Fix mmap-related #defines. Thanks to Lubos Lunak.
+ * Fix copy macros; added LACKS_FCNTL_H. Thanks to Neal Walfield.
+ * Branch-free bin calculation
+ * Default trim and mmap thresholds now 256K.
+
+ V2.7.0 Sun Mar 11 14:14:06 2001 Doug Lea (dl at gee)
+ * Introduce independent_comalloc and independent_calloc.
+ Thanks to Michael Pachos for motivation and help.
+ * Make optional .h file available
+ * Allow > 2GB requests on 32bit systems.
+ * new WIN32 sbrk, mmap, munmap, lock code from <Walter@GeNeSys-e.de>.
+ Thanks also to Andreas Mueller <a.mueller at paradatec.de>,
+ and Anonymous.
+ * Allow override of MALLOC_ALIGNMENT (Thanks to Ruud Waij for
+ helping test this.)
+ * memalign: check alignment arg
+ * realloc: don't try to shift chunks backwards, since this
+ leads to more fragmentation in some programs and doesn't
+ seem to help in any others.
+ * Collect all cases in malloc requiring system memory into sysmalloc
+ * Use mmap as backup to sbrk
+ * Place all internal state in malloc_state
+ * Introduce fastbins (although similar to 2.5.1)
+ * Many minor tunings and cosmetic improvements
+ * Introduce USE_PUBLIC_MALLOC_WRAPPERS, USE_MALLOC_LOCK
+ * Introduce MALLOC_FAILURE_ACTION, MORECORE_CONTIGUOUS
+ Thanks to Tony E. Bennett <tbennett@nvidia.com> and others.
+ * Include errno.h to support default failure action.
+
+ V2.6.6 Sun Dec 5 07:42:19 1999 Doug Lea (dl at gee)
+ * return null for negative arguments
+ * Added Several WIN32 cleanups from Martin C. Fong <mcfong at yahoo.com>
+ * Add 'LACKS_SYS_PARAM_H' for those systems without 'sys/param.h'
+ (e.g. WIN32 platforms)
+ * Cleanup header file inclusion for WIN32 platforms
+ * Cleanup code to avoid Microsoft Visual C++ compiler complaints
+ * Add 'USE_DL_PREFIX' to quickly allow co-existence with existing
+ memory allocation routines
+ * Set 'malloc_getpagesize' for WIN32 platforms (needs more work)
+ * Use 'assert' rather than 'ASSERT' in WIN32 code to conform to
+ usage of 'assert' in non-WIN32 code
+ * Improve WIN32 'sbrk()' emulation's 'findRegion()' routine to
+ avoid infinite loop
+ * Always call 'fREe()' rather than 'free()'
+
+ V2.6.5 Wed Jun 17 15:57:31 1998 Doug Lea (dl at gee)
+ * Fixed ordering problem with boundary-stamping
+
+ V2.6.3 Sun May 19 08:17:58 1996 Doug Lea (dl at gee)
+ * Added pvalloc, as recommended by H.J. Liu
+ * Added 64bit pointer support mainly from Wolfram Gloger
+ * Added anonymously donated WIN32 sbrk emulation
+ * Malloc, calloc, getpagesize: add optimizations from Raymond Nijssen
+ * malloc_extend_top: fix mask error that caused wastage after
+ foreign sbrks
+ * Add linux mremap support code from HJ Liu
+
+ V2.6.2 Tue Dec 5 06:52:55 1995 Doug Lea (dl at gee)
+ * Integrated most documentation with the code.
+ * Add support for mmap, with help from
+ Wolfram Gloger (Gloger@lrz.uni-muenchen.de).
+ * Use last_remainder in more cases.
+ * Pack bins using idea from colin@nyx10.cs.du.edu
+ * Use ordered bins instead of best-fit threshhold
+ * Eliminate block-local decls to simplify tracing and debugging.
+ * Support another case of realloc via move into top
+ * Fix error occuring when initial sbrk_base not word-aligned.
+ * Rely on page size for units instead of SBRK_UNIT to
+ avoid surprises about sbrk alignment conventions.
+ * Add mallinfo, mallopt. Thanks to Raymond Nijssen
+ (raymond@es.ele.tue.nl) for the suggestion.
+ * Add `pad' argument to malloc_trim and top_pad mallopt parameter.
+ * More precautions for cases where other routines call sbrk,
+ courtesy of Wolfram Gloger (Gloger@lrz.uni-muenchen.de).
+ * Added macros etc., allowing use in linux libc from
+ H.J. Lu (hjl@gnu.ai.mit.edu)
+ * Inverted this history list
+
+ V2.6.1 Sat Dec 2 14:10:57 1995 Doug Lea (dl at gee)
+ * Re-tuned and fixed to behave more nicely with V2.6.0 changes.
+ * Removed all preallocation code since under current scheme
+ the work required to undo bad preallocations exceeds
+ the work saved in good cases for most test programs.
+ * No longer use return list or unconsolidated bins since
+ no scheme using them consistently outperforms those that don't
+ given above changes.
+ * Use best fit for very large chunks to prevent some worst-cases.
+ * Added some support for debugging
+
+ V2.6.0 Sat Nov 4 07:05:23 1995 Doug Lea (dl at gee)
+ * Removed footers when chunks are in use. Thanks to
+ Paul Wilson (wilson@cs.texas.edu) for the suggestion.
+
+ V2.5.4 Wed Nov 1 07:54:51 1995 Doug Lea (dl at gee)
+ * Added malloc_trim, with help from Wolfram Gloger
+ (wmglo@Dent.MED.Uni-Muenchen.DE).
+
+ V2.5.3 Tue Apr 26 10:16:01 1994 Doug Lea (dl at g)
+
+ V2.5.2 Tue Apr 5 16:20:40 1994 Doug Lea (dl at g)
+ * realloc: try to expand in both directions
+ * malloc: swap order of clean-bin strategy;
+ * realloc: only conditionally expand backwards
+ * Try not to scavenge used bins
+ * Use bin counts as a guide to preallocation
+ * Occasionally bin return list chunks in first scan
+ * Add a few optimizations from colin@nyx10.cs.du.edu
+
+ V2.5.1 Sat Aug 14 15:40:43 1993 Doug Lea (dl at g)
+ * faster bin computation & slightly different binning
+ * merged all consolidations to one part of malloc proper
+ (eliminating old malloc_find_space & malloc_clean_bin)
+ * Scan 2 returns chunks (not just 1)
+ * Propagate failure in realloc if malloc returns 0
+ * Add stuff to allow compilation on non-ANSI compilers
+ from kpv@research.att.com
+
+ V2.5 Sat Aug 7 07:41:59 1993 Doug Lea (dl at g.oswego.edu)
+ * removed potential for odd address access in prev_chunk
+ * removed dependency on getpagesize.h
+ * misc cosmetics and a bit more internal documentation
+ * anticosmetics: mangled names in macros to evade debugger strangeness
+ * tested on sparc, hp-700, dec-mips, rs6000
+ with gcc & native cc (hp, dec only) allowing
+ Detlefs & Zorn comparison study (in SIGPLAN Notices.)
+
+ Trial version Fri Aug 28 13:14:29 1992 Doug Lea (dl at g.oswego.edu)
+ * Based loosely on libg++-1.2X malloc. (It retains some of the overall
+ structure of old version, but most details differ.)
+
+*/
+
diff --git a/nedmalloc.c b/nedmalloc.c new file mode 100644 index 0000000..999a9c2 --- /dev/null +++ b/nedmalloc.c @@ -0,0 +1,2275 @@ +/* Alternative malloc implementation for multiple threads without
+lock contention based on dlmalloc. (C) 2005-2012 Niall Douglas
+
+Boost Software License - Version 1.0 - August 17th, 2003
+
+Permission is hereby granted, free of charge, to any person or organization
+obtaining a copy of the software and accompanying documentation covered by
+this license (the "Software") to use, reproduce, display, distribute,
+execute, and transmit the Software, and to prepare derivative works of the
+Software, and to permit third-parties to whom the Software is furnished to
+do so, all subject to the following:
+
+The copyright notices in the Software and this entire statement, including
+the above license grant, this restriction and the following disclaimer,
+must be included in all copies of the Software, in whole or in part, and
+all derivative works of the Software, unless such copies or derivative
+works are solely in the form of machine-executable object code generated by
+a source language processor.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE, TITLE AND NON-INFRINGEMENT. IN NO EVENT
+SHALL THE COPYRIGHT HOLDERS OR ANYONE DISTRIBUTING THE SOFTWARE BE LIABLE
+FOR ANY DAMAGES OR OTHER LIABILITY, WHETHER IN CONTRACT, TORT OR OTHERWISE,
+ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+DEALINGS IN THE SOFTWARE.
+*/
+
+#if 0 /* Effectively makes nedmalloc = dlmalloc */
+#define THREADCACHEMAX 0
+#define DEFAULT_GRANULARITY 65536
+#define DEFAULTMAXTHREADSINPOOL 1
+#endif
+
+#ifdef _MSC_VER
+/* Enable full aliasing on MSVC */
+/*#pragma optimize("a", on)*/
+/*#pragma optimize("g", off)*/
+
+#pragma warning(push)
+#pragma warning(disable:4100) /* unreferenced formal parameter */
+#pragma warning(disable:4127) /* conditional expression is constant */
+#pragma warning(disable:4232) /* address of dllimport is not static, identity not guaranteed */
+#pragma warning(disable:4706) /* assignment within conditional expression */
+
+#define _CRT_SECURE_NO_WARNINGS 1 /* Don't care about MSVC warnings on POSIX functions */
+#include <stdio.h>
+#define fopen(f, m) _fsopen((f), (m), 0x40/*_SH_DENYNO*/) /* Have Windows let other programs view the log file as it is written */
+#ifndef UNICODE
+#define UNICODE /* Turn on windows unicode support */
+#endif
+#else
+#include <stdio.h>
+#endif
+
+/*#define NEDMALLOC_DEBUG 1*/
+/*#define ENABLE_LOGGING 7
+#define NEDMALLOC_TESTLOGENTRY(tc, np, type, mspace, size, mem, alignment, flags, returned) (((type)&ENABLE_LOGGING)&&((size)>16*1024))
+#define NEDMALLOC_STACKBACKTRACEDEPTH 16*/
+/*#define NEDMALLOC_FORCERESERVE(p, mem, size) (((size)>=(256*1024)) ? M2_RESERVE_MULT(8) : 0)*/
+/*#define WIN32_DIRECT_USE_FILE_MAPPINGS 0*/
+
+/*#define NEDMALLOC_DEBUG 1*/
+
+/*#define FULLSANITYCHECKS*/
+
+/*#define FORCEINLINE*/
+
+/* There is only support for the user mode page allocator on Windows at present */
+#if !defined(ENABLE_USERMODEPAGEALLOCATOR)
+#define ENABLE_USERMODEPAGEALLOCATOR 0
+#endif
+
+/* If link time code generation is on, don't force or prevent inlining */
+#if defined(_MSC_VER) && defined(NEDMALLOC_DLL_EXPORTS)
+#define FORCEINLINE
+#define NOINLINE
+#endif
+
+#include "nedmalloc.h"
+#include <errno.h>
+#ifdef HAVE_VALGRIND
+#include <valgrind/valgrind.h>
+#include <valgrind/memcheck.h>
+#endif
+#if defined(WIN32)
+ #include <malloc.h>
+#else
+ #if defined(__cplusplus)
+extern "C"
+ #else
+extern
+ #endif
+ #if defined(__linux__) || defined(__FreeBSD__)
+/* Sadly we can't include <malloc.h> as it causes a redefinition error */
+size_t malloc_usable_size(void *);
+ #elif defined(__APPLE__)
+ #if TARGET_OS_IPHONE
+ #include <malloc/malloc.h>
+ #else
+ #include <malloc.h>
+ #endif
+ #else
+ #error Do not know what to do here
+ #endif
+#endif
+
+#if USE_ALLOCATOR==1
+ #define MSPACES 1
+ #define ONLY_MSPACES 1
+#endif
+#define USE_DL_PREFIX 1
+#define USE_RECURSIVE_LOCKS 1
+#ifndef USE_LOCKS
+ #define USE_LOCKS 1
+#endif
+#define FOOTERS 1 /* Need to enable footers so frees lock the right mspace */
+#define INSECURE 0 /* nedblkmstate works a lot better with magic enabled */
+#ifndef NEDMALLOC_DEBUG
+ #if defined(DEBUG) || defined(_DEBUG)
+ #define NEDMALLOC_DEBUG 1
+ #else
+ #define NEDMALLOC_DEBUG 0
+ #endif
+#endif
+/* We need to consistently define DEBUG=0|1, _DEBUG and NDEBUG for dlmalloc */
+#if !defined(DEBUG) && !defined(NDEBUG)
+ #ifdef __GNUC__
+ #warning DEBUG may not be defined but without NDEBUG being defined allocator will run with assert checking! Define NDEBUG to run at full speed.
+ #elif defined(_MSC_VER)
+ #pragma message(__FILE__ ": WARNING: DEBUG may not be defined but without NDEBUG being defined allocator will run with assert checking! Define NDEBUG to run at full speed.")
+ #endif
+#endif
+#undef DEBUG
+#undef _DEBUG
+#if NEDMALLOC_DEBUG
+ #define _DEBUG
+ #define DEBUG 1
+#else
+ #define DEBUG 0
+#endif
+#ifdef NDEBUG /* Disable assert checking on release builds */
+ #undef DEBUG
+ #undef _DEBUG
+#endif
+/* The default of 64Kb means we spend too much time kernel-side */
+#ifndef DEFAULT_GRANULARITY
+ #define DEFAULT_GRANULARITY (1*1024*1024)
+ #if DEBUG
+ #define DEFAULT_GRANULARITY_ALIGNED
+ #endif
+#endif
+/*#define USE_SPIN_LOCKS 0*/
+
+#if ENABLE_USERMODEPAGEALLOCATOR
+extern int OSHavePhysicalPageSupport(void);
+extern void *userpage_malloc(size_t toallocate, unsigned flags);
+extern int userpage_free(void *mem, size_t size);
+extern void *userpage_realloc(void *mem, size_t oldsize, size_t newsize, int flags, unsigned flags2);
+
+#define USERPAGE_TOPDOWN (M2_CUSTOM_FLAGS_BEGIN<<0)
+#define USERPAGE_NOCOMMIT (M2_CUSTOM_FLAGS_BEGIN<<1)
+
+/* This can provide a very significant speed boost */
+#undef MMAP_CLEARS
+#define MMAP_CLEARS 0
+
+#define MUNMAP(h, a, s) (!OSHavePhysicalPageSupport() ? MUNMAP_DEFAULT((h), (a), (s)) : userpage_free((a), (s)))
+#define MMAP(s, f) (!OSHavePhysicalPageSupport() ? MMAP_DEFAULT((s)) : userpage_malloc((s), (f)))
+#define MREMAP(addr, osz, nsz, mv) (!OSHavePhysicalPageSupport() ? MREMAP_DEFAULT((addr), (osz), (nsz), (mv)) : userpage_realloc((addr), (osz), (nsz), (mv), 0))
+#define DIRECT_MMAP(h, s, f) (!OSHavePhysicalPageSupport() ? DIRECT_MMAP_DEFAULT((h), (s), (f)) : userpage_malloc((s), (f)|USERPAGE_TOPDOWN))
+#define DIRECT_MREMAP(h, a, os, ns, f, f2) (!OSHavePhysicalPageSupport() ? DIRECT_MREMAP_DEFAULT((h), (a), (os), (ns), (f), (f2)) : userpage_realloc((a), (os), (ns), (f), (f2)|USERPAGE_TOPDOWN))
+
+/*#undef MREMAP
+#define MREMAP(addr, osz, nsz, mv) (!OSHavePhysicalPageSupport() ? MREMAP_DEFAULT((addr), (osz), (nsz), (mv)) : MFAIL)*/
+/*#undef DIRECT_MREMAP
+#define DIRECT_MREMAP(h, a, os, ns, f, f2) (!OSHavePhysicalPageSupport() ? DIRECT_MREMAP_DEFAULT((h), (a), (os), (ns), (f), (f2)) : MFAIL)*/
+
+#endif
+#include "malloc.c.h"
+#ifdef NDEBUG /* Disable assert checking on release builds */
+ #undef DEBUG
+#endif
+
+/* The default number of threads allowed into a pool at once */
+#ifndef DEFAULTMAXTHREADSINPOOL
+#define DEFAULTMAXTHREADSINPOOL 4
+#endif
+/* The maximum size to be allocated from the thread cache */
+#ifndef THREADCACHEMAX
+#define THREADCACHEMAX 8192
+#elif THREADCACHEMAX && !defined(THREADCACHEMAXBINS)
+ #ifdef __GNUC__
+ #warning If you are changing THREADCACHEMAX, do you also need to change THREADCACHEMAXBINS=(topbitpos(THREADCACHEMAX)-4)?
+ #elif defined(_MSC_VER)
+ #pragma message(__FILE__ ": WARNING: If you are changing THREADCACHEMAX, do you also need to change THREADCACHEMAXBINS=(topbitpos(THREADCACHEMAX)-4)?")
+ #endif
+#endif
+/* The maximum concurrent threads in a pool possible */
+#ifndef MAXTHREADSINPOOL
+#define MAXTHREADSINPOOL 16
+#endif
+/* The maximum number of threadcaches which can be allocated */
+#ifndef THREADCACHEMAXCACHES
+#define THREADCACHEMAXCACHES 256
+#endif
+#ifndef THREADCACHEMAXBINS
+#if 0
+/* The number of cache entries for finer grained bins. This is (topbitpos(THREADCACHEMAX)-4)*2 */
+#define THREADCACHEMAXBINS ((13-4)*2)
+#else
+/* The number of cache entries. This is (topbitpos(THREADCACHEMAX)-4) */
+#define THREADCACHEMAXBINS (13-4)
+#endif
+#endif
+/* Point at which the free space in a thread cache is garbage collected */
+#ifndef THREADCACHEMAXFREESPACE
+#define THREADCACHEMAXFREESPACE (1024*1024)
+#endif
+/* NEDMALLOC_FORCERESERVE is used to force malloc2 flags for normal malloc, calloc et al */
+#ifndef NEDMALLOC_FORCERESERVE
+#define NEDMALLOC_FORCERESERVE(p, mem, size) 0
+#endif
+/* ENABLE_LOGGING is a bitmask of what events to log */
+#if ENABLE_LOGGING
+#ifndef NEDMALLOC_LOGFILE
+#define NEDMALLOC_LOGFILE "nedmalloc.csv"
+#endif
+#endif
+/* NEDMALLOC_TESTLOGENTRY returns non-zero if the entry should be logged */
+#ifndef NEDMALLOC_TESTLOGENTRY
+#define NEDMALLOC_TESTLOGENTRY(tc, np, type, mspace, size, mem, alignment, flags, returned) ((type)&ENABLE_LOGGING)
+#endif
+/* NEDMALLOC_STACKBACKTRACEDEPTH has the logger store a stack backtrace per logged item. Slow! */
+#ifndef NEDMALLOC_STACKBACKTRACEDEPTH
+#define NEDMALLOC_STACKBACKTRACEDEPTH 0
+#endif
+#if NEDMALLOC_STACKBACKTRACEDEPTH
+#include "Dbghelp.h"
+#include "psapi.h"
+#pragma comment(lib, "dbghelp.lib")
+#pragma comment(lib, "psapi.lib")
+#endif
+#define NM_FLAGS_MASK (M2_FLAGS_MASK&~M2_ZERO_MEMORY)
+
+#if USE_LOCKS
+#ifdef WIN32
+ #define TLSVAR DWORD
+ #define TLSALLOC(k) (*(k)=TlsAlloc(), TLS_OUT_OF_INDEXES==*(k))
+ #define TLSFREE(k) (!TlsFree(k))
+ #define TLSGET(k) TlsGetValue(k)
+ #define TLSSET(k, a) (!TlsSetValue(k, a))
+ #ifdef DEBUG
+static LPVOID ChkedTlsGetValue(DWORD idx)
+{
+ LPVOID ret=TlsGetValue(idx);
+ assert(S_OK==GetLastError());
+ return ret;
+}
+ #undef TLSGET
+ #define TLSGET(k) ChkedTlsGetValue(k)
+ #endif
+#else
+ #define TLSVAR pthread_key_t
+ #define TLSALLOC(k) pthread_key_create(k, 0)
+ #define TLSFREE(k) pthread_key_delete(k)
+ #define TLSGET(k) pthread_getspecific(k)
+ #define TLSSET(k, a) pthread_setspecific(k, a)
+#endif
+#else /* Probably if you're not using locks then you don't want ANY pthread stuff at all */
+ #define TLSVAR void *
+ #define TLSALLOC(k) (*k=0)
+ #define TLSFREE(k) (k=0)
+ #define TLSGET(k) k
+ #define TLSSET(k, a) (k=a, 0)
+#endif
+
+#if ENABLE_USERMODEPAGEALLOCATOR
+#include "usermodepageallocator.c"
+#endif
+
+#if defined(__cplusplus)
+#if !defined(NO_NED_NAMESPACE)
+namespace nedalloc {
+#else
+extern "C" {
+#endif
+#endif
+
+#if USE_ALLOCATOR==0
+static void *unsupported_operation(const char *opname) THROWSPEC
+{
+ fprintf(stderr, "nedmalloc: The operation %s is not supported under this build configuration\n", opname);
+ abort();
+ return 0;
+}
+static size_t mspacecounter=(size_t) 0xdeadbeef;
+#endif
+#ifndef ENABLE_FAST_HEAP_DETECTION
+static void *RESTRICT leastusedaddress;
+static size_t largestusedblock;
+#endif
+/* Used to redirect system allocator ops if needed */
+extern void *(*sysmalloc)(size_t);
+extern void *(*syscalloc)(size_t, size_t);
+extern void *(*sysrealloc)(void *, size_t);
+extern void (*sysfree)(void *);
+extern size_t (*sysblksize)(void *);
+
+#if !defined(REPLACE_SYSTEM_ALLOCATOR) || (!defined(_MSC_VER) && !defined(__MINGW32__))
+void *(*sysmalloc)(size_t)=malloc;
+void *(*syscalloc)(size_t, size_t)=calloc;
+void *(*sysrealloc)(void *, size_t)=realloc;
+void (*sysfree)(void *)=free;
+size_t (*sysblksize)(void *)=
+#if defined(_MSC_VER) || defined(__MINGW32__)
+ /* This is the MSVCRT equivalent */
+ _msize;
+#elif defined(__linux__) || defined(__FreeBSD__)
+ /* This is the glibc/ptmalloc2/dlmalloc/BSD equivalent. */
+ malloc_usable_size;
+#elif defined(__APPLE__)
+ /* This is the Apple BSD libc equivalent. */
+ (size_t (*)(void *)) malloc_size;
+#else
+#error Cannot tolerate the memory allocator of an unknown system!
+#endif
+#else
+/* Remove the MSVCRT dependency on the memory functions */
+void *(*sysmalloc)(size_t);
+void *(*syscalloc)(size_t, size_t);
+void *(*sysrealloc)(void *, size_t);
+void (*sysfree)(void *);
+size_t (*sysblksize)(void *);
+#endif
+
+static FORCEINLINE NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void *CallMalloc(void *RESTRICT mspace, size_t size, size_t alignment, unsigned flags) THROWSPEC
+{
+ void *RESTRICT ret=0;
+#if USE_MAGIC_HEADERS
+ size_t _alignment=alignment;
+ size_t *_ret=0;
+ size_t bytes=size+alignment+3*sizeof(size_t);
+ /* Avoid addition overflow. */
+ if(bytes<size)
+ return 0;
+ size=bytes;
+ _alignment=0;
+#endif
+#if USE_ALLOCATOR==0
+ ret=(flags & M2_ZERO_MEMORY) ? syscalloc(1, size) : sysmalloc(size); /* magic headers takes care of alignment */
+#elif USE_ALLOCATOR==1
+ ret=mspace_malloc2((mstate) mspace, size, alignment, flags);
+#ifdef HAVE_VALGRIND
+ if(ret) VALGRIND_MEMPOOL_ALLOC(mspace, ret, size);
+#endif
+#ifndef ENABLE_FAST_HEAP_DETECTION
+ if(ret)
+ {
+ mchunkptr p=mem2chunk(ret);
+ size_t truesize=chunksize(p) - overhead_for(p);
+ if(!leastusedaddress || (void *)((mstate) mspace)->least_addr<leastusedaddress) leastusedaddress=(void *)((mstate) mspace)->least_addr;
+ if(!largestusedblock || truesize>largestusedblock) largestusedblock=(truesize+mparams.page_size) & ~(mparams.page_size-1);
+ }
+#endif
+#endif
+ if(!ret) return 0;
+#if DEBUG
+ if(flags & M2_ZERO_MEMORY)
+ {
+ const char *RESTRICT n;
+ for(n=(const char *)ret; n<(const char *)ret+size; n++)
+ {
+ assert(!*n);
+ }
+ }
+#endif
+#if USE_MAGIC_HEADERS
+ _ret=(size_t *) ret;
+ ret=(void *)(_ret+3);
+ if(alignment) ret=(void *)(((size_t) ret+alignment-1)&~(alignment-1));
+ for(; _ret<(size_t *)ret-2; _ret++) *_ret=*(size_t *)"NEDMALOC";
+ _ret[0]=(size_t) mspace;
+ _ret[1]=size-3*sizeof(size_t);
+#endif
+ return ret;
+}
+
+static FORCEINLINE NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void *CallRealloc(void *RESTRICT mspace, void *RESTRICT mem, int isforeign, size_t oldsize, size_t newsize, size_t alignment, unsigned flags) THROWSPEC
+{
+ void *RESTRICT ret=0;
+#if USE_MAGIC_HEADERS
+ mstate oldmspace=0;
+ size_t *_ret=0, *_mem=(size_t *) mem-3;
+#endif
+ if(isforeign)
+ { /* Transfer */
+#if USE_MAGIC_HEADERS
+ assert(_mem[0]!=*(size_t *) "NEDMALOC");
+#endif
+ if((ret=CallMalloc(mspace, newsize, alignment, flags)))
+ {
+#if defined(DEBUG)
+ printf("*** nedmalloc frees system allocated block %p\n", mem);
+#endif
+ memcpy(ret, mem, oldsize<newsize ? oldsize : newsize);
+ sysfree(mem);
+ }
+ return ret;
+ }
+#if USE_MAGIC_HEADERS
+ assert(_mem[0]==*(size_t *) "NEDMALOC");
+ newsize+=3*sizeof(size_t);
+ oldmspace=(mstate) _mem[1];
+ assert(oldsize>=_mem[2]);
+ for(; *_mem==*(size_t *) "NEDMALOC"; *_mem--=*(size_t *) "nedmaloc");
+ mem=(void *)(++_mem);
+#endif
+#if USE_ALLOCATOR==0
+ ret=sysrealloc(mem, newsize);
+#elif USE_ALLOCATOR==1
+ ret=mspace_realloc2((mstate) mspace, mem, newsize, alignment, flags);
+#ifdef HAVE_VALGRIND
+ if(ret) VALGRIND_MEMPOOL_CHANGE(mspace, mem, ret, newsize);
+#endif
+#ifndef ENABLE_FAST_HEAP_DETECTION
+ if(ret)
+ {
+ mchunkptr p=mem2chunk(ret);
+ size_t truesize=chunksize(p) - overhead_for(p);
+ if(!leastusedaddress || (void *)((mstate) mspace)->least_addr<leastusedaddress) leastusedaddress=(void *)((mstate) mspace)->least_addr;
+ if(!largestusedblock || truesize>largestusedblock) largestusedblock=(truesize+mparams.page_size) & ~(mparams.page_size-1);
+ }
+#endif
+#endif
+ if(!ret)
+ { /* Put it back the way it was */
+#if USE_MAGIC_HEADERS
+ for(; *_mem==0; *_mem++=*(size_t *) "NEDMALOC");
+#endif
+ return 0;
+ }
+#if USE_MAGIC_HEADERS
+ _ret=(size_t *) ret;
+ ret=(void *)(_ret+3);
+ for(; _ret<(size_t *)ret-2; _ret++) *_ret=*(size_t *) "NEDMALOC";
+ _ret[0]=(size_t) mspace;
+ _ret[1]=newsize-3*sizeof(size_t);
+#endif
+ return ret;
+}
+
+static FORCEINLINE void CallFree(void *RESTRICT mspace, void *RESTRICT mem, int isforeign) THROWSPEC
+{
+#if USE_MAGIC_HEADERS
+ mstate oldmspace=0;
+ size_t *_mem=(size_t *) mem-3, oldsize=0;
+#endif
+ if(isforeign)
+ {
+#if USE_MAGIC_HEADERS
+ assert(_mem[0]!=*(size_t *) "NEDMALOC");
+#endif
+#if defined(DEBUG)
+ printf("*** nedmalloc frees system allocated block %p\n", mem);
+#endif
+ sysfree(mem);
+ return;
+ }
+#if USE_MAGIC_HEADERS
+ assert(_mem[0]==*(size_t *) "NEDMALOC");
+ oldmspace=(mstate) _mem[1];
+ oldsize=_mem[2];
+ for(; *_mem==*(size_t *) "NEDMALOC"; *_mem--=*(size_t *) "nedmaloc");
+ mem=(void *)(++_mem);
+#endif
+#if USE_ALLOCATOR==0
+ sysfree(mem);
+#elif USE_ALLOCATOR==1
+#ifdef HAVE_VALGRIND
+ {
+ void *m=mspace ? mspace : get_mstate_for(mem2chunk(mem));
+ mspace_free((mstate) mspace, mem);
+ VALGRIND_MEMPOOL_FREE(m, mem);
+ /* Mark the first two pointers in the newly freed block as used (linked list of freed blocks) */
+ VALGRIND_MAKE_MEM_DEFINED(mem, 2*sizeof(void *));
+ }
+#else
+ mspace_free((mstate) mspace, mem);
+#endif
+#endif
+}
+
+static NEDMALLOCNOALIASATTR mstate nedblkmstate(void *RESTRICT mem) THROWSPEC
+{
+ if(mem)
+ {
+#if USE_MAGIC_HEADERS
+ size_t *_mem=(size_t *) mem-3;
+ if(_mem[0]==*(size_t *) "NEDMALOC")
+ {
+ return (mstate) _mem[1];
+ }
+ else return 0;
+#else
+#if USE_ALLOCATOR==0
+ /* Fail everything */
+ return 0;
+#elif USE_ALLOCATOR==1
+#ifdef ENABLE_FAST_HEAP_DETECTION
+#ifdef WIN32
+ /* On Windows for RELEASE both x86 and x64 the NT heap precedes each block with an eight byte header
+ which looks like:
+ normal: 4 bytes of size, 4 bytes of [char < 64, char < 64, char < 64 bit 0 always set, char random ]
+ mmaped: 4 bytes of size 4 bytes of [zero, zero, 0xb, zero ]
+
+ On Windows for DEBUG both x86 and x64 the preceding four bytes is always 0xfdfdfdfd (no man's land).
+ */
+#pragma pack(push, 1)
+ struct _HEAP_ENTRY
+ {
+ USHORT Size;
+ USHORT PreviousSize;
+ UCHAR Cookie; /* SegmentIndex */
+ UCHAR Flags; /* always bit 0 (HEAP_ENTRY_BUSY). bit 1=(HEAP_ENTRY_EXTRA_PRESENT), bit 2=normal block (HEAP_ENTRY_FILL_PATTERN), bit 3=mmap block (HEAP_ENTRY_VIRTUAL_ALLOC). Bit 4 (HEAP_ENTRY_LAST_ENTRY) could be set */
+ UCHAR UnusedBytes;
+ UCHAR SmallTagIndex; /* fastbin index. Always one of 0x02, 0x03, 0x04 < 0x80 */
+ } *RESTRICT he=((struct _HEAP_ENTRY *) mem)-1;
+#pragma pack(pop)
+ unsigned int header=((unsigned int *)mem)[-1], mask1=0x8080E100, result1, mask2=0xFFFFFF06, result2;
+ result1=header & mask1; /* Positive testing for NT heap */
+ result2=header & mask2; /* Positive testing for dlmalloc */
+ if(result1==0x00000100 && result2!=0x00000102)
+ { /* This is likely a NT heap block */
+ return 0;
+ }
+#endif
+#ifdef __linux__
+ /* On Linux glibc uses ptmalloc2 (really dlmalloc) just as we do, but prev_foot contains rubbish
+ when the preceding block is allocated because ptmalloc2 finds the local mstate by rounding the ptr
+ down to the nearest megabyte. It's like dlmalloc with FOOTERS disabled. */
+ mchunkptr p=mem2chunk(mem);
+ mstate fm=get_mstate_for(p);
+ /* If it's a ptmalloc2 block, fm is likely to be some crazy value */
+ if(!is_aligned(fm)) return 0;
+ if((size_t)mem-(size_t)fm>=(size_t)1<<(SIZE_T_BITSIZE-1)) return 0;
+ if(ok_magic(fm))
+ return fm;
+ else
+ return 0;
+ if(1) { }
+#endif
+ else
+ {
+ mchunkptr p=mem2chunk(mem);
+ mstate fm=get_mstate_for(p);
+ assert(ok_magic(fm)); /* If this fails, someone tried to free a block twice */
+ if(ok_magic(fm))
+ return fm;
+ }
+#else /* ENABLE_FAST_HEAP_DETECTION */
+#ifdef WIN32
+#ifdef _MSC_VER
+ __try
+#else
+#if ENABLE_TOLERANT_NEDMALLOC
+#error Lack of SEH support makes nedmalloc unreliable with foreign memory blocks. Make sure ENABLE_TOLERANT_NEDMALLOC is turned off!
+#endif
+#endif /* defined(_MSC_VER) */
+#elif ENABLE_TOLERANT_NEDMALLOC
+#warning nedmalloc is unreliable with foreign memory blocks, so make sure you really want ENABLE_TOLERANT_NEDMALLOC turned on!
+#endif
+ {
+ /* We try to return zero here if it isn't one of our own blocks, however
+ the current block annotation scheme used by dlmalloc makes it impossible
+ to be absolutely sure of avoiding a segfault.
+
+ mchunkptr->prev_foot = mem-(2*size_t) = mstate ^ mparams.magic for PRECEDING block;
+ mchunkptr->head = mem-(1*size_t) = 8 multiple size of this block with bottom three bits = FLAG_BITS
+ FLAG_BITS = bit 0 is CINUSE (currently in use unless is mmap), bit 1 is PINUSE (previous block currently
+ in use unless mmap), bit 2 is UNUSED and currently is always zero.
+ */
+ register void *RESTRICT leastusedaddress_=leastusedaddress; /* Cache these to avoid register reloading */
+ register size_t largestusedblock_=largestusedblock;
+ if(!is_aligned(mem)) return 0; /* Would fail very rarely as all allocators return aligned blocks */
+ if(mem<leastusedaddress_) return 0; /* Simple but effective */
+ {
+ mchunkptr p=mem2chunk(mem);
+ mstate fm=0;
+ int ismmapped=is_mmapped(p);
+ if((!ismmapped && !is_inuse(p)) || (p->head & FLAG4_BIT)) return 0;
+ /* Reduced uncertainty by 0.5^2 = 25.0% */
+ /* size should never exceed largestusedblock */
+ if(chunksize(p)-overhead_for(p)>largestusedblock_) return 0;
+ /* Reduced uncertainty by a minimum of 0.5^3 = 12.5%, maximum 0.5^16 = 0.0015% */
+ /* Having sanity checked prev_foot and head, check next block */
+ if(!ismmapped && (!next_pinuse(p) || (next_chunk(p)->head & FLAG4_BIT))) return 0;
+ /* Reduced uncertainty by 0.5^5 = 3.13% or 0.5^18 = 0.00038% */
+#if 0
+ /* If previous block is free, check that its next block pointer equals us */
+ if(!ismmapped && !pinuse(p))
+ if(next_chunk(prev_chunk(p))!=p) return 0;
+ /* We could start comparing prev_foot's for similarity but it starts getting slow. */
+#endif
+ fm = get_mstate_for(p);
+ if(!is_aligned(fm) || (void *)fm<leastusedaddress_) return 0;
+#if 0
+ /* See if mem is lower in memory than mem */
+ if((size_t)mem-(size_t)fm>=(size_t)1<<(SIZE_T_BITSIZE-1)) return 0;
+#endif
+ assert(ok_magic(fm)); /* If this fails, someone tried to free a block twice */
+ if(ok_magic(fm))
+ return fm;
+ }
+ }
+#ifdef WIN32
+#ifdef _MSC_VER
+ __except(1) { }
+#endif
+#endif /* WIN32 */
+#endif /* ENABLE_FAST_HEAP_DETECTION */
+#endif /* USE_ALLOCATOR */
+#endif /* USE_MAGIC_HEADERS */
+ }
+ return 0;
+}
+NEDMALLOCNOALIASATTR size_t nedblksize(int *RESTRICT isforeign, void *RESTRICT mem, unsigned flags) THROWSPEC
+{
+ if(mem)
+ {
+ if(isforeign) *isforeign=1;
+#if USE_MAGIC_HEADERS
+ {
+ size_t *_mem=(size_t *) mem-3;
+ if(_mem[0]==*(size_t *) "NEDMALOC")
+ {
+ mstate mspace=(mstate) _mem[1];
+ size_t size=_mem[2];
+ if(isforeign) *isforeign=0;
+ return size;
+ }
+ }
+#elif USE_ALLOCATOR==1
+ if((flags & NM_SKIP_TOLERANCE_CHECKS) || nedblkmstate(mem))
+ {
+ mchunkptr p=mem2chunk(mem);
+ if(isforeign) *isforeign=0;
+ return chunksize(p)-overhead_for(p);
+ }
+#ifdef DEBUG
+ else
+ {
+ int a=1; /* Set breakpoints here if needed */
+ }
+#endif
+#endif
+#if defined(ENABLE_TOLERANT_NEDMALLOC) || USE_ALLOCATOR==0
+ return sysblksize(mem);
+#endif
+ }
+ return 0;
+}
+NEDMALLOCNOALIASATTR size_t nedmemsize(void *RESTRICT mem) THROWSPEC { return nedblksize(0, mem, 0); }
+
+NEDMALLOCNOALIASATTR void nedsetvalue(void *v) THROWSPEC { nedpsetvalue((nedpool *) 0, v); }
+NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedmalloc(size_t size) THROWSPEC { return nedpmalloc((nedpool *) 0, size); }
+NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedcalloc(size_t no, size_t size) THROWSPEC { return nedpcalloc((nedpool *) 0, no, size); }
+NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedrealloc(void *mem, size_t size) THROWSPEC { return nedprealloc((nedpool *) 0, mem, size); }
+NEDMALLOCNOALIASATTR void nedfree(void *mem) THROWSPEC { nedpfree((nedpool *) 0, mem); }
+NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedmemalign(size_t alignment, size_t bytes) THROWSPEC { return nedpmemalign((nedpool *) 0, alignment, bytes); }
+NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedmalloc2(size_t size, size_t alignment, unsigned flags) THROWSPEC { return nedpmalloc2((nedpool *) 0, size, alignment, flags); }
+NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedrealloc2(void *mem, size_t size, size_t alignment, unsigned flags) THROWSPEC { return nedprealloc2((nedpool *) 0, mem, size, alignment, flags); }
+NEDMALLOCNOALIASATTR void nedfree2(void *mem, unsigned flags) THROWSPEC { nedpfree2((nedpool *) 0, mem, flags); }
+NEDMALLOCNOALIASATTR struct nedmallinfo nedmallinfo(void) THROWSPEC { return nedpmallinfo((nedpool *) 0); }
+NEDMALLOCNOALIASATTR int nedmallopt(int parno, int value) THROWSPEC { return nedpmallopt((nedpool *) 0, parno, value); }
+NEDMALLOCNOALIASATTR int nedmalloc_trim(size_t pad) THROWSPEC { return nedpmalloc_trim((nedpool *) 0, pad); }
+void nedmalloc_stats() THROWSPEC { nedpmalloc_stats((nedpool *) 0); }
+NEDMALLOCNOALIASATTR size_t nedmalloc_footprint() THROWSPEC { return nedpmalloc_footprint((nedpool *) 0); }
+NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void **nedindependent_calloc(size_t elemsno, size_t elemsize, void **chunks) THROWSPEC { return nedpindependent_calloc((nedpool *) 0, elemsno, elemsize, chunks); }
+NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void **nedindependent_comalloc(size_t elems, size_t *sizes, void **chunks) THROWSPEC { return nedpindependent_comalloc((nedpool *) 0, elems, sizes, chunks); }
+
+#ifdef WIN32
+typedef unsigned __int64 timeCount;
+static timeCount GetTimestamp()
+{
+ static LARGE_INTEGER ticksPerSec;
+ static double scalefactor;
+ static timeCount baseCount;
+ LARGE_INTEGER val;
+ timeCount ret;
+ if(!scalefactor)
+ {
+ if(QueryPerformanceFrequency(&ticksPerSec))
+ scalefactor=ticksPerSec.QuadPart/1000000000000.0;
+ else
+ scalefactor=1;
+ }
+ if(!QueryPerformanceCounter(&val))
+ return (timeCount) GetTickCount() * 1000000000;
+ ret=(timeCount) (val.QuadPart/scalefactor);
+ if(!baseCount) baseCount=ret;
+ return ret-baseCount;
+}
+#else
+#include <sys/time.h>
+
+typedef unsigned long long timeCount;
+static timeCount GetTimestamp()
+{
+ static timeCount baseCount;
+ timeCount ret;
+#ifdef CLOCK_MONOTONIC
+ struct timespec ts;
+ clock_gettime(CLOCK_MONOTONIC, &ts);
+ ret=((timeCount) ts.tv_sec*1000000000000LL)+ts.tv_nsec*1000LL;
+#else
+ struct timeval tv;
+ gettimeofday(&tv, 0);
+ ret=((timeCount) tv.tv_sec*1000000000000LL)+tv.tv_usec*1000000LL;
+#endif
+ if(!baseCount) baseCount=ret;
+ return ret-baseCount;
+}
+#endif
+
+/* Set ENABLE_LOGGING to an AND mask of which of these you want to
+log, so set it to 0xffffffff for everything */
+typedef enum LogEntryType_t
+{
+ LOGENTRY_MALLOC =(1<<0),
+ LOGENTRY_REALLOC =(1<<1),
+ LOGENTRY_FREE =(1<<2),
+
+ LOGENTRY_THREADCACHE_MALLOC =(1<<3),
+ LOGENTRY_THREADCACHE_FREE =(1<<4),
+ LOGENTRY_THREADCACHE_CLEAN =(1<<5),
+
+ LOGENTRY_POOL_MALLOC =(1<<6),
+ LOGENTRY_POOL_REALLOC =(1<<7),
+ LOGENTRY_POOL_FREE =(1<<8)
+} LogEntryType;
+#if NEDMALLOC_STACKBACKTRACEDEPTH
+typedef struct StackFrameType_t
+{
+ void *pc;
+ char module[64];
+ char functname[256];
+ char file[96];
+ int lineno;
+} StackFrameType;
+#endif
+typedef struct logentry_t
+{
+ timeCount timestamp;
+ nedpool *np;
+ LogEntryType type;
+ int mspace;
+ size_t size;
+ void *mem;
+ size_t alignment;
+ unsigned flags;
+ void *returned;
+#if NEDMALLOC_STACKBACKTRACEDEPTH
+ StackFrameType stack[NEDMALLOC_STACKBACKTRACEDEPTH];
+#endif
+} logentry;
+static const char *LogEntryTypeStrings[]={
+ "LOGENTRY_MALLOC",
+ "LOGENTRY_REALLOC",
+ "LOGENTRY_FREE",
+
+ "LOGENTRY_THREADCACHE_MALLOC",
+ "LOGENTRY_THREADCACHE_FREE",
+ "LOGENTRY_THREADCACHE_CLEAN",
+
+ "LOGENTRY_POOL_MALLOC",
+ "LOGENTRY_POOL_REALLOC",
+ "LOGENTRY_POOL_FREE",
+ "******************"
+};
+
+struct threadcacheblk_t;
+typedef struct threadcacheblk_t threadcacheblk;
+struct threadcacheblk_t
+{ /* Keep less than 32 bytes as sizeof(threadcacheblk) is the minimum allocation size */
+#ifdef FULLSANITYCHECKS
+ unsigned int magic;
+#endif
+ int isforeign;
+ unsigned int lastUsed;
+ size_t size;
+ threadcacheblk *RESTRICT next, *RESTRICT prev;
+};
+typedef struct threadcache_t
+{
+#ifdef FULLSANITYCHECKS
+ unsigned int magic1;
+#endif
+ int mymspace; /* Last mspace entry this thread used */
+ long threadid;
+ unsigned int mallocs, frees, successes;
+#if ENABLE_LOGGING
+ logentry *logentries, *logentriesptr, *logentriesend;
+#endif
+ size_t freeInCache; /* How much free space is stored in this cache */
+ threadcacheblk *RESTRICT bins[(THREADCACHEMAXBINS+1)*2];
+#ifdef FULLSANITYCHECKS
+ unsigned int magic2;
+#endif
+} threadcache;
+struct nedpool_t
+{
+#if USE_LOCKS
+ MLOCK_T mutex;
+#endif
+ void *uservalue;
+ int threads; /* Max entries in m to use */
+ threadcache *RESTRICT caches[THREADCACHEMAXCACHES];
+ TLSVAR mycache; /* Thread cache for this thread. 0 for unset, negative for use mspace-1 directly, otherwise is cache-1 */
+ mstate m[MAXTHREADSINPOOL+1]; /* mspace entries for this pool */
+};
+static nedpool syspool;
+
+#if ENABLE_LOGGING
+#if NEDMALLOC_STACKBACKTRACEDEPTH
+#if defined(WIN32) && defined(_MSC_VER)
+#define COPY_STRING(d, s, maxlen) { size_t len=strlen(s); len=(len>maxlen) ? maxlen-1 : len; memcpy(d, s, len); d[len]=0; }
+
+#pragma optimize("g", off)
+static int ExceptionFilter(unsigned int code, struct _EXCEPTION_POINTERS *ep, CONTEXT *ct) THROWSPEC
+{
+ *ct=*ep->ContextRecord;
+ return EXCEPTION_EXECUTE_HANDLER;
+}
+
+static DWORD64 __stdcall GetModBase(HANDLE hProcess, DWORD64 dwAddr) THROWSPEC
+{
+ DWORD64 modulebase;
+ // Try to get the module base if already loaded, otherwise load the module
+ modulebase=SymGetModuleBase64(hProcess, dwAddr);
+ if(modulebase)
+ return modulebase;
+ else
+ {
+ MEMORY_BASIC_INFORMATION stMBI ;
+ if ( 0 != VirtualQueryEx ( hProcess, (LPCVOID)(size_t)dwAddr, &stMBI, sizeof(stMBI)))
+ {
+ int n;
+ DWORD dwPathLen=0, dwNameLen=0 ;
+ TCHAR szFile[ MAX_PATH ], szModuleName[ MAX_PATH ] ;
+ MODULEINFO mi={0};
+ dwPathLen = GetModuleFileName ( (HMODULE) stMBI.AllocationBase , szFile, MAX_PATH );
+ dwNameLen = GetModuleBaseName (hProcess, (HMODULE) stMBI.AllocationBase , szModuleName, MAX_PATH );
+ for(n=dwNameLen; n>0; n--)
+ {
+ if(szModuleName[n]=='.')
+ {
+ szModuleName[n]=0;
+ break;
+ }
+ }
+ if(!GetModuleInformation(hProcess, (HMODULE) stMBI.AllocationBase, &mi, sizeof(mi)))
+ {
+ //fxmessage("WARNING: GetModuleInformation() returned error code %d\n", GetLastError());
+ }
+ if(!SymLoadModule64 ( hProcess, NULL, (PSTR)( (dwPathLen) ? szFile : 0), (PSTR)( (dwNameLen) ? szModuleName : 0),
+ (DWORD64) mi.lpBaseOfDll, mi.SizeOfImage))
+ {
+ //fxmessage("WARNING: SymLoadModule64() returned error code %d\n", GetLastError());
+ }
+ //fxmessage("%s, %p, %x, %x\n", szFile, mi.lpBaseOfDll, mi.SizeOfImage, (DWORD) mi.lpBaseOfDll+mi.SizeOfImage);
+ modulebase=SymGetModuleBase64(hProcess, dwAddr);
+ return modulebase;
+ }
+ }
+ return 0;
+}
+
+extern HANDLE sym_myprocess;
+extern VOID (WINAPI *RtlCaptureContextAddr)(PCONTEXT);
+extern void DeinitSym(void) THROWSPEC;
+static void DoStackWalk(logentry *p) THROWSPEC
+{
+ int i,i2;
+ HANDLE mythread=(HANDLE) GetCurrentThread();
+ STACKFRAME64 sf={ 0 };
+ CONTEXT ct={ 0 };
+ if(!sym_myprocess)
+ {
+ DWORD symopts;
+ DuplicateHandle(GetCurrentProcess(), GetCurrentProcess(), GetCurrentProcess(), &sym_myprocess, 0, FALSE, DUPLICATE_SAME_ACCESS);
+ symopts=SymGetOptions();
+ SymSetOptions(symopts /*| SYMOPT_DEFERRED_LOADS*/ | SYMOPT_LOAD_LINES);
+ SymInitialize(sym_myprocess, NULL, TRUE);
+ atexit(DeinitSym);
+ }
+ ct.ContextFlags=CONTEXT_FULL;
+
+ // Use RtlCaptureContext() if we have it as it saves an exception throw
+ if((VOID (WINAPI *)(PCONTEXT)) -1==RtlCaptureContextAddr)
+ RtlCaptureContextAddr=(VOID (WINAPI *)(PCONTEXT)) GetProcAddress(GetModuleHandle(L"kernel32"), "RtlCaptureContext");
+ if(RtlCaptureContextAddr)
+ RtlCaptureContextAddr(&ct);
+ else
+ { // This is nasty, but it works
+ __try
+ {
+ int *foo=0;
+ *foo=78;
+ }
+ __except (ExceptionFilter(GetExceptionCode(), GetExceptionInformation(), &ct))
+ {
+ }
+ }
+
+ sf.AddrPC.Mode=sf.AddrStack.Mode=sf.AddrFrame.Mode=AddrModeFlat;
+#if !(defined(_M_AMD64) || defined(_M_X64))
+ sf.AddrPC.Offset =ct.Eip;
+ sf.AddrStack.Offset=ct.Esp;
+ sf.AddrFrame.Offset=ct.Ebp;
+#else
+ sf.AddrPC.Offset =ct.Rip;
+ sf.AddrStack.Offset=ct.Rsp;
+ sf.AddrFrame.Offset=ct.Rbp; // maybe Rdi?
+#endif
+ for(i2=0; i2<NEDMALLOC_STACKBACKTRACEDEPTH; i2++)
+ {
+ IMAGEHLP_MODULE64 ihm={ sizeof(IMAGEHLP_MODULE64) };
+ char temp[MAX_PATH+sizeof(IMAGEHLP_SYMBOL64)];
+ IMAGEHLP_SYMBOL64 *ihs;
+ IMAGEHLP_LINE64 ihl={ sizeof(IMAGEHLP_LINE64) };
+ DWORD64 offset;
+ if(!StackWalk64(
+#if !(defined(_M_AMD64) || defined(_M_X64))
+ IMAGE_FILE_MACHINE_I386,
+#else
+ IMAGE_FILE_MACHINE_AMD64,
+#endif
+ sym_myprocess, mythread, &sf, &ct, NULL, SymFunctionTableAccess64, GetModBase, NULL))
+ break;
+ if(0==sf.AddrPC.Offset)
+ break;
+ i=i2;
+ if(i) // Skip first entry relating to this function
+ {
+ DWORD lineoffset=0;
+ p->stack[i-1].pc=(void *)(size_t) sf.AddrPC.Offset;
+ if(SymGetModuleInfo64(sym_myprocess, sf.AddrPC.Offset, &ihm))
+ {
+ char *leaf;
+ leaf=strrchr(ihm.ImageName, '\\');
+ if(!leaf) leaf=ihm.ImageName-1;
+ COPY_STRING(p->stack[i-1].module, leaf+1, sizeof(p->stack[i-1].module));
+ }
+ else strcpy(p->stack[i-1].module, "<unknown>");
+ //fxmessage("WARNING: SymGetModuleInfo64() returned error code %d\n", GetLastError());
+ memset(temp, 0, MAX_PATH+sizeof(IMAGEHLP_SYMBOL64));
+ ihs=(IMAGEHLP_SYMBOL64 *) temp;
+ ihs->SizeOfStruct=sizeof(IMAGEHLP_SYMBOL64);
+ ihs->Address=sf.AddrPC.Offset;
+ ihs->MaxNameLength=MAX_PATH;
+
+ if(SymGetSymFromAddr64(sym_myprocess, sf.AddrPC.Offset, &offset, ihs))
+ {
+ COPY_STRING(p->stack[i-1].functname, ihs->Name, sizeof(p->stack[i-1].functname));
+ if(strlen(p->stack[i-1].functname)<sizeof(p->stack[i-1].functname)-8)
+ {
+ sprintf(strchr(p->stack[i-1].functname, 0), " +0x%x", offset);
+ }
+ }
+ else
+ strcpy(p->stack[i-1].functname, "<unknown>");
+ if(SymGetLineFromAddr64(sym_myprocess, sf.AddrPC.Offset, &lineoffset, &ihl))
+ {
+ char *leaf;
+ p->stack[i-1].lineno=ihl.LineNumber;
+
+ leaf=strrchr(ihl.FileName, '\\');
+ if(!leaf) leaf=ihl.FileName-1;
+ COPY_STRING(p->stack[i-1].file, leaf+1, sizeof(p->stack[i-1].file));
+ }
+ else
+ strcpy(p->stack[i-1].file, "<unknown>");
+ }
+ }
+}
+#pragma optimize("g", on)
+#else
+static void DoStackWalk(logentry *p) THROWSPEC
+{
+ void *backtr[NEDMALLOC_STACKBACKTRACEDEPTH];
+ size_t size;
+ char **strings;
+ size_t i2;
+ size=backtrace(backtr, NEDMALLOC_STACKBACKTRACEDEPTH);
+ strings=backtrace_symbols(backtr, size);
+ for(i2=0; i2<size; i2++)
+ { // Format can be <file path>(<mangled symbol>+0x<offset>) [<pc>]
+ // or can be <file path> [<pc>]
+ int start=0, end=strlen(strings[i2]), which=0, idx;
+ for(idx=0; idx<end; idx++)
+ {
+ if(0==which && (' '==strings[i2][idx] || '('==strings[i2][idx]))
+ {
+ int len=FXMIN(idx-start, (int) sizeof(p->stack[i2].file));
+ memcpy(p->stack[i2].file, strings[i2]+start, len);
+ p->stack[i2].file[len]=0;
+ which=(' '==strings[i2][idx]) ? 2 : 1;
+ start=idx+1;
+ }
+ else if(1==which && ')'==strings[i2][idx])
+ {
+ FXString functname(strings[i2]+start, idx-start);
+ FXint offset=functname.rfind("+0x");
+ FXString rawsymbol(functname.left(offset));
+ FXString symbol(rawsymbol.length() ? fxdemanglesymbol(rawsymbol, false) : rawsymbol);
+ symbol.append(functname.mid(offset));
+ int len=FXMIN(symbol.length(), (int) sizeof(p->stack[i2].functname));
+ memcpy(p->stack[i2].functname, symbol.text(), len);
+ p->stack[i2].functname[len]=0;
+ which=2;
+ }
+ else if(2==which && '['==strings[i2][idx])
+ {
+ start=idx+1;
+ which=3;
+ }
+ else if(3==which && ']'==strings[i2][idx])
+ {
+ FXString v(strings[i2]+start+2, idx-start-2);
+ p->stack[i2].pc=(void *)(FXuval)v.toULong(0, 16);
+ }
+ }
+ }
+ free(strings);
+}
+#endif
+#endif
+#endif
+static FORCEINLINE logentry *LogOperation(threadcache *tc, nedpool *np, LogEntryType type, int mspace, size_t size, void *mem, size_t alignment, unsigned flags, void *returned) THROWSPEC
+{
+#if ENABLE_LOGGING
+ if(tc->logentries && NEDMALLOC_TESTLOGENTRY(tc, np, type, mspace, size, mem, alignment, flags, returned))
+ {
+ logentry *le;
+ if(tc->logentriesptr==tc->logentriesend)
+ {
+ mchunkptr cp=mem2chunk(tc->logentries);
+ size_t logentrieslen=chunksize(cp)-overhead_for(cp);
+ le=(logentry *) CallRealloc(0, tc->logentries, 0, logentrieslen, (logentrieslen*3)/2, 0, M2_ZERO_MEMORY|M2_ALWAYS_MMAP|M2_RESERVE_MULT(8));
+ if(!le) return 0;
+ tc->logentriesptr=le+(tc->logentriesptr-tc->logentries);
+ tc->logentries=le;
+ cp=mem2chunk(tc->logentries);
+ logentrieslen=(chunksize(cp)-overhead_for(cp))/sizeof(logentry);
+ tc->logentriesend=tc->logentries+logentrieslen;
+ }
+ le=tc->logentriesptr++;
+ assert(le+1<=tc->logentriesend);
+ le->timestamp=GetTimestamp();
+ le->np=np;
+ le->type=type;
+ le->mspace=mspace;
+ le->size=size;
+ le->mem=mem;
+ le->alignment=alignment;
+ le->flags=flags;
+ le->returned=returned;
+#if NEDMALLOC_STACKBACKTRACEDEPTH
+ DoStackWalk(le);
+#endif
+ return le;
+ }
+#endif
+ return 0;
+}
+
+static FORCEINLINE NEDMALLOCNOALIASATTR unsigned int size2binidx(size_t _size) THROWSPEC
+{ /* 8=1000 16=10000 20=10100 24=11000 32=100000 48=110000 4096=1000000000000 */
+ unsigned int topbit, size=(unsigned int)(_size>>4);
+ /* 16=1 20=1 24=1 32=10 48=11 64=100 96=110 128=1000 4096=100000000 */
+
+#if defined(__GNUC__)
+ topbit = sizeof(size)*__CHAR_BIT__ - 1 - __builtin_clz(size);
+#elif defined(_MSC_VER) && _MSC_VER>=1300
+ {
+ unsigned long bsrTopBit;
+
+ _BitScanReverse(&bsrTopBit, size);
+
+ topbit = bsrTopBit;
+ }
+#else
+#if 0
+ union {
+ unsigned asInt[2];
+ double asDouble;
+ };
+ int n;
+
+ asDouble = (double)size + 0.5;
+ topbit = (asInt[!FOX_BIGENDIAN] >> 20) - 1023;
+#else
+ {
+ unsigned int x=size;
+ x = x | (x >> 1);
+ x = x | (x >> 2);
+ x = x | (x >> 4);
+ x = x | (x >> 8);
+ x = x | (x >>16);
+ x = ~x;
+ x = x - ((x >> 1) & 0x55555555);
+ x = (x & 0x33333333) + ((x >> 2) & 0x33333333);
+ x = (x + (x >> 4)) & 0x0F0F0F0F;
+ x = x + (x << 8);
+ x = x + (x << 16);
+ topbit=31 - (x >> 24);
+ }
+#endif
+#endif
+ return topbit;
+}
+
+
+#ifdef FULLSANITYCHECKS
+static void tcsanitycheck(threadcacheblk *RESTRICT *RESTRICT ptr) THROWSPEC
+{
+ assert((ptr[0] && ptr[1]) || (!ptr[0] && !ptr[1]));
+ if(ptr[0] && ptr[1])
+ {
+ assert(ptr[0]->isforeign || nedblkmstate(ptr[0]));
+ assert(ptr[1]->isforeign || nedblkmstate(ptr[1]));
+ assert(nedblksize(0, ptr[0], 0)>=sizeof(threadcacheblk));
+ assert(nedblksize(0, ptr[1], 0)>=sizeof(threadcacheblk));
+ assert(*(unsigned int *) "NEDN"==ptr[0]->magic);
+ assert(*(unsigned int *) "NEDN"==ptr[1]->magic);
+ assert(!ptr[0]->prev);
+ assert(!ptr[1]->next);
+ if(ptr[0]==ptr[1])
+ {
+ assert(!ptr[0]->next);
+ assert(!ptr[1]->prev);
+ }
+ }
+}
+static void tcfullsanitycheck(threadcache *tc) THROWSPEC
+{
+ threadcacheblk *RESTRICT *RESTRICT tcbptr=tc->bins;
+ int n;
+ for(n=0; n<=THREADCACHEMAXBINS; n++, tcbptr+=2)
+ {
+ threadcacheblk *RESTRICT b, *RESTRICT ob=0;
+ tcsanitycheck(tcbptr);
+ for(b=tcbptr[0]; b; ob=b, b=b->next)
+ {
+ assert(b->isforeign || nedblkmstate(b));
+ assert(nedblksize(0, b, 0)>=sizeof(threadcacheblk));
+ assert(*(unsigned int *) "NEDN"==b->magic);
+ assert(!ob || ob->next==b);
+ assert(!ob || b->prev==ob);
+ }
+ }
+}
+#endif
+
+static NOINLINE int InitPool(nedpool *RESTRICT p, size_t capacity, int threads) THROWSPEC;
+static NOINLINE void RemoveCacheEntries(nedpool *RESTRICT p, threadcache *RESTRICT tc, unsigned int age) THROWSPEC
+{
+#ifdef FULLSANITYCHECKS
+ tcfullsanitycheck(tc);
+#endif
+ if(tc->freeInCache)
+ {
+ threadcacheblk *RESTRICT *RESTRICT tcbptr=tc->bins;
+ int n;
+ for(n=0; n<=THREADCACHEMAXBINS; n++, tcbptr+=2)
+ {
+ threadcacheblk *RESTRICT *RESTRICT tcb=tcbptr+1; /* come from oldest end of list */
+ /*tcsanitycheck(tcbptr);*/
+ for(; *tcb && tc->frees-(*tcb)->lastUsed>=age; )
+ {
+ threadcacheblk *RESTRICT f=*tcb;
+ size_t blksize=f->size; /*nedblksize(f);*/
+ assert(blksize<=nedblksize(0, f, 0));
+ assert(blksize);
+#ifdef FULLSANITYCHECKS
+ assert(*(unsigned int *) "NEDN"==(*tcb)->magic);
+#endif
+ *tcb=(*tcb)->prev;
+ if(*tcb)
+ (*tcb)->next=0;
+ else
+ *tcbptr=0;
+ tc->freeInCache-=blksize;
+ assert((long) tc->freeInCache>=0);
+ CallFree(0, f, f->isforeign);
+ /*tcsanitycheck(tcbptr);*/
+ LogOperation(tc, p, LOGENTRY_THREADCACHE_CLEAN, age, blksize, f, 0, 0, 0);
+ }
+ }
+ }
+#ifdef FULLSANITYCHECKS
+ tcfullsanitycheck(tc);
+#endif
+}
+size_t nedflushlogs(nedpool *p, char *filepath) THROWSPEC
+{
+ size_t count=0;
+ if(!p)
+ {
+ p=&syspool;
+ if(!syspool.threads) InitPool(&syspool, 0, -1);
+ }
+ if(p->caches)
+ {
+ threadcache *tc;
+ int n;
+#if ENABLE_LOGGING
+ int haslogentries=0;
+#endif
+ for(n=0; n<THREADCACHEMAXCACHES; n++)
+ {
+ if((tc=p->caches[n]))
+ {
+ count+=tc->freeInCache;
+ tc->frees++;
+ RemoveCacheEntries(p, tc, 0);
+ assert(!tc->freeInCache);
+#if ENABLE_LOGGING
+ haslogentries|=!!tc->logentries;
+#endif
+ }
+ }
+#if ENABLE_LOGGING
+ if(haslogentries)
+ {
+ char buffer[MAX_PATH]=NEDMALLOC_LOGFILE;
+ FILE *oh;
+ fpos_t pos1, pos2;
+ if(!filepath) filepath=buffer;
+ oh=fopen(filepath, "r+");
+ while(!oh)
+ {
+ char *bptr;
+ if((oh=fopen(filepath, "w"))) break;
+ if(ENOSPC==errno) break;
+ bptr=strrchr(filepath, '.');
+ if(bptr-filepath>=MAX_PATH-6) abort();
+ memcpy(bptr, "!.csv", 6);
+ }
+ if(oh)
+ {
+ fgetpos(oh, &pos1);
+ fseek(oh, 0, SEEK_END);
+ fgetpos(oh, &pos2);
+ if(pos1==pos2)
+ fprintf(oh, "Timestamp, Pool, Operation, MSpace, Size, Block, Alignment, Flags, Returned,\"Stack Backtrace\"\n");
+ for(n=0; n<THREADCACHEMAXCACHES; n++)
+ {
+ if((tc=p->caches[n]) && tc->logentries)
+ {
+ logentry *le;
+ for(le=tc->logentries; le<tc->logentriesptr; le++)
+ {
+ const char *LogEntryTypeString=LogEntryTypeStrings[size2binidx(((size_t)le->type)<<4)];
+ char stackbacktrace[16384]="?";
+#if NEDMALLOC_STACKBACKTRACEDEPTH
+ char *sbtp=stackbacktrace;
+ int i;
+ for(i=0; i<NEDMALLOC_STACKBACKTRACEDEPTH && le->stack[i].pc; i++)
+ {
+ sbtp+=sprintf(sbtp, "0x%p:%s:%s (%s:%u),",
+ le->stack[i].pc, le->stack[i].module, le->stack[i].functname, le->stack[i].file, le->stack[i].lineno);
+ if(sbtp>=stackbacktrace+sizeof(stackbacktrace)) abort();
+ }
+ if(NEDMALLOC_STACKBACKTRACEDEPTH==i)
+ strcpy(sbtp, "<backtrace may continue ...>");
+ else
+ strcpy(sbtp, "<backtrace ends>");
+ if(strchr(sbtp, 0)>=stackbacktrace+sizeof(stackbacktrace)) abort();
+#endif
+ fprintf(oh, "%llu, 0x%p, %s, %d, %Iu, 0x%p, %Iu, 0x%x, 0x%p,\"%s\"\n",
+ le->timestamp, le->np, LogEntryTypeString, le->mspace, le->size, le->mem, le->alignment, le->flags, le->returned, stackbacktrace);
+ }
+ CallFree(0, tc->logentries, 0);
+ tc->logentries=tc->logentriesptr=tc->logentriesend=0;
+ }
+ }
+ fclose(oh);
+ }
+ }
+#endif
+ }
+ return count;
+}
+static void DestroyCaches(nedpool *RESTRICT p) THROWSPEC
+{
+ if(p->caches)
+ {
+ threadcache *tc;
+ int n;
+ nedflushlogs(p, 0);
+ for(n=0; n<THREADCACHEMAXCACHES; n++)
+ {
+ if((tc=p->caches[n]))
+ {
+ tc->mymspace=-1;
+ tc->threadid=0;
+ CallFree(0, tc, 0);
+ p->caches[n]=0;
+ }
+ }
+ }
+}
+
+static NOINLINE threadcache *AllocCache(nedpool *RESTRICT p) THROWSPEC
+{
+ threadcache *tc=0;
+ int n, end;
+#if USE_LOCKS
+ ACQUIRE_LOCK(&p->mutex);
+#endif
+ for(n=0; n<THREADCACHEMAXCACHES && p->caches[n]; n++);
+ if(THREADCACHEMAXCACHES==n)
+ { /* List exhausted, so disable for this thread */
+#if USE_LOCKS
+ RELEASE_LOCK(&p->mutex);
+#endif
+ return 0;
+ }
+ tc=p->caches[n]=(threadcache *) CallMalloc(p->m[0], sizeof(threadcache), 0, M2_ZERO_MEMORY);
+ if(!tc)
+ {
+#if USE_LOCKS
+ RELEASE_LOCK(&p->mutex);
+#endif
+ return 0;
+ }
+#ifdef FULLSANITYCHECKS
+ tc->magic1=*(unsigned int *)"NEDMALC1";
+ tc->magic2=*(unsigned int *)"NEDMALC2";
+#endif
+ tc->threadid=
+#if USE_LOCKS
+ (long)(size_t)CURRENT_THREAD;
+#else
+ 1;
+#endif
+ for(end=1; p->m[end]; end++);
+ tc->mymspace=abs(tc->threadid) % end;
+#if ENABLE_LOGGING
+ {
+ mchunkptr cp;
+ size_t logentrieslen=2048/sizeof(logentry); /* One page */
+ tc->logentries=tc->logentriesptr=(logentry *) CallMalloc(p->m[0], logentrieslen*sizeof(logentry), 0, M2_ZERO_MEMORY|M2_ALWAYS_MMAP|M2_RESERVE_MULT(8));
+ if(!tc->logentries)
+ {
+#if USE_LOCKS
+ RELEASE_LOCK(&p->mutex);
+#endif
+ return 0;
+ }
+ cp=mem2chunk(tc->logentries);
+ logentrieslen=(chunksize(cp)-overhead_for(cp))/sizeof(logentry);
+ tc->logentriesend=tc->logentries+logentrieslen;
+ }
+#endif
+#if USE_LOCKS
+ RELEASE_LOCK(&p->mutex);
+#endif
+ if(TLSSET(p->mycache, (void *)(size_t)(n+1))) abort();
+ return tc;
+}
+
+static void *threadcache_malloc(nedpool *RESTRICT p, threadcache *RESTRICT tc, size_t *RESTRICT _size) THROWSPEC
+{
+ void *RESTRICT ret=0;
+ size_t size=*_size, blksize=0;
+ unsigned int bestsize;
+ unsigned int idx=size2binidx(size);
+ threadcacheblk *RESTRICT blk, *RESTRICT *RESTRICT binsptr;
+#ifdef FULLSANITYCHECKS
+ tcfullsanitycheck(tc);
+#endif
+ /* Calculate best fit bin size */
+ bestsize=1<<(idx+4);
+#if 0
+ /* Finer grained bin fit */
+ idx<<=1;
+ if(size>bestsize)
+ {
+ idx++;
+ bestsize+=bestsize>>1;
+ }
+ if(size>bestsize)
+ {
+ idx++;
+ bestsize=1<<(4+(idx>>1));
+ }
+#else
+ if(size>bestsize)
+ {
+ idx++;
+ bestsize<<=1;
+ }
+#endif
+ assert(bestsize>=size);
+ if(size<bestsize) size=bestsize;
+ assert(size<=THREADCACHEMAX);
+ assert(idx<=THREADCACHEMAXBINS);
+ binsptr=&tc->bins[idx*2];
+ /* Try to match close, but move up a bin if necessary */
+ blk=*binsptr;
+ if(!blk || blk->size<size)
+ { /* Bump it up a bin */
+ if(idx<THREADCACHEMAXBINS)
+ {
+ idx++;
+ binsptr+=2;
+ blk=*binsptr;
+ }
+ }
+ if(blk)
+ {
+ blksize=blk->size; /*nedblksize(blk);*/
+ assert(nedblksize(0, blk, 0)>=blksize);
+ assert(blksize>=size);
+ if(blk->next)
+ blk->next->prev=0;
+ *binsptr=blk->next;
+ if(!*binsptr)
+ binsptr[1]=0;
+#ifdef FULLSANITYCHECKS
+ blk->magic=0;
+#endif
+ assert(binsptr[0]!=blk && binsptr[1]!=blk);
+ assert(nedblksize(0, blk, 0)>=sizeof(threadcacheblk) && nedblksize(0, blk, 0)<=THREADCACHEMAX+CHUNK_OVERHEAD);
+ /*printf("malloc: %p, %p, %p, %lu\n", p, tc, blk, (long) _size);*/
+ ret=(void *) blk;
+ }
+ ++tc->mallocs;
+ if(ret)
+ {
+ assert(blksize>=size);
+ ++tc->successes;
+ tc->freeInCache-=blksize;
+ assert((long) tc->freeInCache>=0);
+ }
+#if defined(DEBUG) && 0
+ if(!(tc->mallocs & 0xfff))
+ {
+ printf("*** threadcache=%u, mallocs=%u (%f), free=%u (%f), freeInCache=%u\n", (unsigned int) tc->threadid, tc->mallocs,
+ (float) tc->successes/tc->mallocs, tc->frees, (float) tc->successes/tc->frees, (unsigned int) tc->freeInCache);
+ }
+#endif
+#ifdef FULLSANITYCHECKS
+ tcfullsanitycheck(tc);
+#endif
+ *_size=size;
+ return ret;
+}
+static NOINLINE void ReleaseFreeInCache(nedpool *RESTRICT p, threadcache *RESTRICT tc, int mymspace) THROWSPEC
+{
+ unsigned int age=THREADCACHEMAXFREESPACE/8192;
+#if USE_LOCKS
+ /*ACQUIRE_LOCK(&p->m[mymspace]->mutex);*/
+#endif
+ while(age && tc->freeInCache>=THREADCACHEMAXFREESPACE)
+ {
+ RemoveCacheEntries(p, tc, age);
+ /*printf("*** Removing cache entries older than %u (%u)\n", age, (unsigned int) tc->freeInCache);*/
+ age>>=1;
+ }
+#if USE_LOCKS
+ /*RELEASE_LOCK(&p->m[mymspace]->mutex);*/
+#endif
+}
+static void threadcache_free(nedpool *RESTRICT p, threadcache *RESTRICT tc, int mymspace, void *RESTRICT mem, size_t size, int isforeign) THROWSPEC
+{
+ unsigned int bestsize;
+ unsigned int idx=size2binidx(size);
+ threadcacheblk *RESTRICT *RESTRICT binsptr, *RESTRICT tck=(threadcacheblk *RESTRICT) mem;
+ assert(size>=sizeof(threadcacheblk) && size<=THREADCACHEMAX+CHUNK_OVERHEAD);
+#ifdef DEBUG
+ /* Make sure this is a valid memory block */
+ assert(nedblksize(0, mem, 0));
+#endif
+#ifdef FULLSANITYCHECKS
+ tcfullsanitycheck(tc);
+#endif
+ /* Calculate best fit bin size */
+ bestsize=1<<(idx+4);
+#if 0
+ /* Finer grained bin fit */
+ idx<<=1;
+ if(size>bestsize)
+ {
+ unsigned int biggerbestsize=bestsize+bestsize<<1;
+ if(size>=biggerbestsize)
+ {
+ idx++;
+ bestsize=biggerbestsize;
+ }
+ }
+#endif
+ if(bestsize!=size) /* dlmalloc can round up, so we round down to preserve indexing */
+ size=bestsize;
+ binsptr=&tc->bins[idx*2];
+ assert(idx<=THREADCACHEMAXBINS);
+ if(tck==*binsptr)
+ {
+ fprintf(stderr, "nedmalloc: Attempt to free already freed memory block %p - aborting!\n", tck);
+ abort();
+ }
+#ifdef FULLSANITYCHECKS
+ tck->magic=*(unsigned int *) "NEDN";
+#endif
+ tck->isforeign=isforeign;
+ tck->lastUsed=++tc->frees;
+ tck->size=(unsigned int) size;
+ tck->next=*binsptr;
+ tck->prev=0;
+ if(tck->next)
+ tck->next->prev=tck;
+ else
+ binsptr[1]=tck;
+ assert(!*binsptr || (*binsptr)->size==tck->size);
+ *binsptr=tck;
+ assert(tck==tc->bins[idx*2]);
+ assert(tc->bins[idx*2+1]==tck || binsptr[0]->next->prev==tck);
+ /*printf("free: %p, %p, %p, %lu\n", p, tc, mem, (long) size);*/
+ tc->freeInCache+=size;
+#ifdef FULLSANITYCHECKS
+ tcfullsanitycheck(tc);
+#endif
+#if 1
+ if(tc->freeInCache>=THREADCACHEMAXFREESPACE)
+ ReleaseFreeInCache(p, tc, mymspace);
+#endif
+}
+
+
+
+
+static NOINLINE int InitPool(nedpool *RESTRICT p, size_t capacity, int threads) THROWSPEC
+{ /* threads is -1 for system pool */
+ ensure_initialization();
+ ACQUIRE_MALLOC_GLOBAL_LOCK();
+ if(p->threads) goto done;
+#if USE_LOCKS
+ if(INITIAL_LOCK(&p->mutex)) goto err;
+#endif
+ if(TLSALLOC(&p->mycache)) goto err;
+#if USE_ALLOCATOR==0
+ p->m[0]=(mstate) mspacecounter++;
+#elif USE_ALLOCATOR==1
+ if(!(p->m[0]=(mstate) create_mspace(capacity, 1))) goto err;
+#ifdef HAVE_VALGRIND
+ VALGRIND_CREATE_MEMPOOL(p->m[0], 0, 1);
+#endif
+ p->m[0]->extp=p;
+#endif
+ p->threads=(threads>MAXTHREADSINPOOL) ? MAXTHREADSINPOOL : (threads<=0) ? DEFAULTMAXTHREADSINPOOL : threads;
+done:
+ RELEASE_MALLOC_GLOBAL_LOCK();
+ return 1;
+err:
+ if(threads<0)
+ abort(); /* If you can't allocate for system pool, we're screwed */
+ DestroyCaches(p);
+ if(p->m[0])
+ {
+#if USE_ALLOCATOR==1
+#ifdef HAVE_VALGRIND
+ VALGRIND_DESTROY_MEMPOOL(p->m[0]);
+#endif
+ destroy_mspace(p->m[0]);
+#endif
+ p->m[0]=0;
+ }
+ if(p->mycache)
+ {
+ if(TLSFREE(p->mycache)) abort();
+ p->mycache=0;
+ }
+ RELEASE_MALLOC_GLOBAL_LOCK();
+ return 0;
+}
+static NOINLINE mstate FindMSpace(nedpool *RESTRICT p, threadcache *RESTRICT tc, int *RESTRICT lastUsed, size_t size) THROWSPEC
+{ /* Gets called when thread's last used mspace is in use. The strategy
+ is to run through the list of all available mspaces looking for an
+ unlocked one and if we fail, we create a new one so long as we don't
+ exceed p->threads */
+ int n, end;
+ n=end=*lastUsed+1;
+#if USE_LOCKS
+ for(; p->m[n]; end=++n)
+ {
+ if(TRY_LOCK(&p->m[n]->mutex)) goto found;
+ }
+ for(n=0; n<*lastUsed && p->m[n]; n++)
+ {
+ if(TRY_LOCK(&p->m[n]->mutex)) goto found;
+ }
+ if(end<p->threads)
+ {
+ mstate temp;
+#if USE_ALLOCATOR==0
+ temp=(mstate) mspacecounter++;
+#elif USE_ALLOCATOR==1
+ if(!(temp=(mstate) create_mspace(size, 1)))
+ goto badexit;
+#endif
+ /* Now we're ready to modify the lists, we lock */
+ ACQUIRE_LOCK(&p->mutex);
+ while(p->m[end] && end<p->threads)
+ end++;
+ if(end>=p->threads)
+ { /* Drat, must destroy it now */
+ RELEASE_LOCK(&p->mutex);
+#if USE_ALLOCATOR==1
+ destroy_mspace((mstate) temp);
+#endif
+ goto badexit;
+ }
+ /* We really want to make sure this goes into memory now but we
+ have to be careful of breaking aliasing rules, so write it twice */
+ {
+ volatile struct malloc_state **_m=(volatile struct malloc_state **) &p->m[end];
+ *_m=(p->m[end]=temp);
+ }
+#if USE_ALLOCATOR==1 && defined(HAVE_VALGRIND)
+ VALGRIND_CREATE_MEMPOOL(temp, 0, 1);
+#endif
+ ACQUIRE_LOCK(&p->m[end]->mutex);
+ /*printf("Created mspace idx %d\n", end);*/
+ RELEASE_LOCK(&p->mutex);
+ n=end;
+ goto found;
+ }
+ /* Let it lock on the last one it used */
+badexit:
+ ACQUIRE_LOCK(&p->m[*lastUsed]->mutex);
+ return p->m[*lastUsed];
+#endif
+found:
+ *lastUsed=n;
+ if(tc)
+ tc->mymspace=n;
+ else
+ {
+ if(TLSSET(p->mycache, (void *)(size_t)(-(n+1)))) abort();
+ }
+ return p->m[n];
+}
+
+typedef struct PoolList_t
+{
+ size_t size; /* Size of list */
+ size_t length; /* Actual entries in list */
+#ifdef DEBUG
+ nedpool *list[1]; /* Force testing of list expansion */
+#else
+ nedpool *list[16];
+#endif
+} PoolList;
+#if USE_LOCKS
+static MLOCK_T poollistlock;
+#endif
+static PoolList *poollist;
+NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR nedpool *nedcreatepool(size_t capacity, int threads) THROWSPEC
+{
+ nedpool *ret=0;
+ if(!poollist)
+ {
+ PoolList *newpoollist=0;
+ if(!(newpoollist=(PoolList *) nedpcalloc(0, 1, sizeof(PoolList)+sizeof(nedpool *)))) return 0;
+#if USE_LOCKS
+ INITIAL_LOCK(&poollistlock);
+ ACQUIRE_LOCK(&poollistlock);
+#endif
+ poollist=newpoollist;
+ poollist->size=sizeof(poollist->list)/sizeof(nedpool *);
+ }
+#if USE_LOCKS
+ else
+ ACQUIRE_LOCK(&poollistlock);
+#endif
+ if(poollist->length==poollist->size)
+ {
+ PoolList *newpoollist=0;
+ size_t newsize=0, toclearsize;
+ newsize=sizeof(PoolList)+(poollist->size+1)*sizeof(nedpool *);
+ if(!(newpoollist=(PoolList *) nedprealloc(0, poollist, newsize))) goto badexit;
+ poollist=newpoollist;
+ toclearsize=newsize-sizeof(PoolList)-(poollist->size)*sizeof(nedpool *);
+ memset(&poollist->list[poollist->size], 0, toclearsize);
+ poollist->size+=toclearsize/sizeof(nedpool *);
+ assert(poollist->size>poollist->length);
+ }
+ if(!(ret=(nedpool *) nedpcalloc(0, 1, sizeof(nedpool)))) goto badexit;
+ if(!InitPool(ret, capacity, threads))
+ {
+ nedpfree(0, ret);
+ goto badexit;
+ }
+ poollist->list[poollist->length++]=ret;
+badexit:
+ {
+#if USE_LOCKS
+ RELEASE_LOCK(&poollistlock);
+#endif
+ }
+ return ret;
+}
+void neddestroypool(nedpool *p) THROWSPEC
+{
+ unsigned int n;
+#if USE_LOCKS
+ ACQUIRE_LOCK(&p->mutex);
+#endif
+ DestroyCaches(p);
+ for(n=0; p->m[n]; n++)
+ {
+#if USE_ALLOCATOR==1
+#ifdef HAVE_VALGRIND
+ VALGRIND_DESTROY_MEMPOOL(p->m[n]);
+#endif
+ destroy_mspace(p->m[n]);
+#endif
+ p->m[n]=0;
+ }
+#if USE_LOCKS
+ RELEASE_LOCK(&p->mutex);
+ DESTROY_LOCK(&p->mutex);
+#endif
+ if(TLSFREE(p->mycache)) abort();
+ nedpfree(0, p);
+#if USE_LOCKS
+ ACQUIRE_LOCK(&poollistlock);
+#endif
+ assert(poollist);
+ for(n=0; n<poollist->length && poollist->list[n]!=p; n++)
+ /* empty */;
+ assert(n!=poollist->length);
+ memmove(&poollist->list[n], &poollist->list[n+1], (size_t)&poollist->list[poollist->length]-(size_t)&poollist->list[n]);
+ if(!--poollist->length)
+ {
+ assert(!poollist->list[0]);
+ nedpfree(0, poollist);
+ poollist=0;
+ }
+#if USE_LOCKS
+ RELEASE_LOCK(&poollistlock);
+#endif
+}
+void neddestroysyspool() THROWSPEC
+{
+ nedpool *p=&syspool;
+ int n;
+#if USE_LOCKS
+ ACQUIRE_LOCK(&p->mutex);
+#endif
+ DestroyCaches(p);
+ for(n=0; p->m[n]; n++)
+ {
+#if USE_ALLOCATOR==1
+#ifdef HAVE_VALGRIND
+ VALGRIND_DESTROY_MEMPOOL(p->m[n]);
+#endif
+ destroy_mspace(p->m[n]);
+#endif
+ p->m[n]=0;
+ }
+ /* Render syspool unusable */
+ for(n=0; n<THREADCACHEMAXCACHES; n++)
+ p->caches[n]=(threadcache *)(size_t)(sizeof(size_t)>4 ? 0xdeadbeefdeadbeefULL : 0xdeadbeefUL);
+ for(n=0; n<MAXTHREADSINPOOL+1; n++)
+ p->m[n]=(mstate)(size_t)(sizeof(size_t)>4 ? 0xdeadbeefdeadbeefULL : 0xdeadbeefUL);
+ if(TLSFREE(p->mycache)) abort();
+#if USE_LOCKS
+ RELEASE_LOCK(&p->mutex);
+ DESTROY_LOCK(&p->mutex);
+#endif
+}
+nedpool **nedpoollist() THROWSPEC
+{
+ nedpool **ret=0;
+ if(poollist)
+ {
+#if USE_LOCKS
+ ACQUIRE_LOCK(&poollistlock);
+#endif
+ if(!(ret=(nedpool **) nedmalloc((poollist->length+1)*sizeof(nedpool *)))) goto badexit;
+ memcpy(ret, poollist->list, (poollist->length+1)*sizeof(nedpool *));
+badexit:
+ {
+#if USE_LOCKS
+ RELEASE_LOCK(&poollistlock);
+#endif
+ }
+ }
+ return ret;
+}
+
+void nedpsetvalue(nedpool *p, void *v) THROWSPEC
+{
+ if(!p) { p=&syspool; if(!syspool.threads) InitPool(&syspool, 0, -1); }
+ p->uservalue=v;
+}
+void *nedgetvalue(nedpool **p, void *mem) THROWSPEC
+{
+ nedpool *np=0;
+ mstate fm=nedblkmstate(mem);
+ if(!fm || !fm->extp) return 0;
+ np=(nedpool *) fm->extp;
+ if(p) *p=np;
+ return np->uservalue;
+}
+
+void nedtrimthreadcache(nedpool *p, int disable) THROWSPEC
+{
+ int mycache;
+ if(!p)
+ {
+ p=&syspool;
+ if(!syspool.threads) InitPool(&syspool, 0, -1);
+ }
+ mycache=(int)(size_t) TLSGET(p->mycache);
+ if(!mycache)
+ { /* Set to mspace 0 */
+ if(disable && TLSSET(p->mycache, (void *)(size_t)-1)) abort();
+ }
+ else if(mycache>0)
+ { /* Set to last used mspace */
+ threadcache *tc=p->caches[mycache-1];
+#if defined(DEBUG)
+ printf("Threadcache utilisation: %lf%% in cache with %lf%% lost to other threads\n",
+ 100.0*tc->successes/tc->mallocs, 100.0*((double) tc->mallocs-tc->frees)/tc->mallocs);
+#endif
+ if(disable && TLSSET(p->mycache, (void *)(size_t)(-tc->mymspace))) abort();
+ tc->frees++;
+ RemoveCacheEntries(p, tc, 0);
+ assert(!tc->freeInCache);
+ if(disable)
+ {
+ tc->mymspace=-1;
+ tc->threadid=0;
+ CallFree(0, p->caches[mycache-1], 0);
+ p->caches[mycache-1]=0;
+ }
+ }
+}
+void neddisablethreadcache(nedpool *p) THROWSPEC
+{
+ nedtrimthreadcache(p, 1);
+}
+
+#if USE_LOCKS && USE_ALLOCATOR==1
+#define GETMSPACE(m,p,tc,ms,s,action) \
+ do \
+ { \
+ mstate m = GetMSpace((p),(tc),(ms),(s)); \
+ action; \
+ RELEASE_LOCK(&m->mutex); \
+ } while (0)
+#else
+#define GETMSPACE(m,p,tc,ms,s,action) \
+ do \
+ { \
+ mstate m = GetMSpace((p),(tc),(ms),(s)); \
+ action; \
+ } while (0)
+#endif
+
+static FORCEINLINE mstate GetMSpace(nedpool *RESTRICT p, threadcache *RESTRICT tc, int mymspace, size_t size) THROWSPEC
+{ /* Returns a locked and ready for use mspace */
+ mstate m=p->m[mymspace];
+ assert(m);
+#if USE_LOCKS && USE_ALLOCATOR==1
+ if(!TRY_LOCK(&p->m[mymspace]->mutex)) m=FindMSpace(p, tc, &mymspace, size);
+ /*assert(IS_LOCKED(&p->m[mymspace]->mutex));*/
+#endif
+ return m;
+}
+static NOINLINE void GetThreadCache_cold1(nedpool *RESTRICT *RESTRICT p) THROWSPEC
+{
+ *p=&syspool;
+ if(!syspool.threads) InitPool(&syspool, 0, -1);
+}
+static NOINLINE void GetThreadCache_cold2(nedpool *RESTRICT *RESTRICT p, threadcache *RESTRICT *RESTRICT tc, int *RESTRICT mymspace, int mycache) THROWSPEC
+{
+ if(!mycache)
+ { /* Need to allocate a new cache */
+ *tc=AllocCache(*p);
+ if(!*tc)
+ { /* Disable */
+ if(TLSSET((*p)->mycache, (void *)(size_t)-1)) abort();
+ *mymspace=0;
+ }
+ else
+ *mymspace=(*tc)->mymspace;
+ }
+ else
+ { /* Cache disabled, but we do have an assigned thread pool */
+ *tc=0;
+ *mymspace=-mycache-1;
+ }
+}
+static FORCEINLINE void GetThreadCache(nedpool *RESTRICT *RESTRICT p, threadcache *RESTRICT *RESTRICT tc, int *RESTRICT mymspace, size_t *RESTRICT size) THROWSPEC
+{
+ int mycache;
+#if THREADCACHEMAX
+ if(size && *size<sizeof(threadcacheblk)) *size=sizeof(threadcacheblk);
+#endif
+ if(!*p)
+ GetThreadCache_cold1(p);
+ mycache=(int)(size_t) TLSGET((*p)->mycache);
+ if(mycache>0)
+ { /* Already have a cache */
+ *tc=(*p)->caches[mycache-1];
+ *mymspace=(*tc)->mymspace;
+ }
+ else GetThreadCache_cold2(p, tc, mymspace, mycache);
+ assert(*mymspace>=0);
+#if USE_LOCKS
+ assert(!(*tc) || (long)(size_t)CURRENT_THREAD==(*tc)->threadid);
+#endif
+#ifdef FULLSANITYCHECKS
+ if(*tc)
+ {
+ if(*(unsigned int *)"NEDMALC1"!=(*tc)->magic1 || *(unsigned int *)"NEDMALC2"!=(*tc)->magic2)
+ {
+ abort();
+ }
+ }
+#endif
+}
+
+NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedpmalloc2(nedpool *p, size_t size, size_t alignment, unsigned flags) THROWSPEC
+{
+ void *ret=0;
+ threadcache *tc;
+ int mymspace;
+ GetThreadCache(&p, &tc, &mymspace, &size);
+#if THREADCACHEMAX
+ if(alignment<=MALLOC_ALIGNMENT && !(flags & NM_FLAGS_MASK) && tc && size<=THREADCACHEMAX)
+ { /* Use the thread cache */
+ if((ret=threadcache_malloc(p, tc, &size)))
+ {
+ if((flags & M2_ZERO_MEMORY))
+ memset(ret, 0, size);
+ LogOperation(tc, p, LOGENTRY_THREADCACHE_MALLOC, mymspace, size, 0, alignment, flags, ret);
+ }
+ }
+#endif
+ if(!ret)
+ { /* Use this thread's mspace */
+ GETMSPACE(m, p, tc, mymspace, size,
+ ret=CallMalloc(m, size, alignment, flags));
+ if(ret)
+ LogOperation(tc, p, LOGENTRY_POOL_MALLOC, mymspace, size, 0, alignment, flags, ret);
+ }
+ LogOperation(tc, p, LOGENTRY_MALLOC, mymspace, size, 0, alignment, flags, ret);
+ return ret;
+}
+NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedprealloc2(nedpool *p, void *mem, size_t size, size_t alignment, unsigned flags) THROWSPEC
+{
+ void *ret=0;
+ threadcache *tc;
+ int mymspace, isforeign=1;
+ size_t memsize;
+ if(!mem) return nedpmalloc2(p, size, alignment, flags);
+#if REALLOC_ZERO_BYTES_FREES
+ if(!size)
+ {
+ nedpfree2(p, mem, flags);
+ return 0;
+ }
+#endif
+ memsize=nedblksize(&isforeign, mem, flags);
+ assert(memsize);
+ if(!memsize)
+ {
+ fprintf(stderr, "nedmalloc: nedprealloc() called with a block not created by nedmalloc!\n");
+ abort();
+ }
+ else if(size<=memsize && memsize-size<
+#ifdef DEBUG
+ 32
+#else
+ 1024
+#endif
+ ) /* If realloc size is within 1Kb smaller than existing, noop it */
+ return mem;
+ GetThreadCache(&p, &tc, &mymspace, &size);
+#if THREADCACHEMAX
+ if(alignment<=MALLOC_ALIGNMENT && !(flags & NM_FLAGS_MASK) && tc && size && size<=THREADCACHEMAX)
+ { /* Use the thread cache */
+ if((ret=threadcache_malloc(p, tc, &size)))
+ {
+ size_t tocopy=memsize<size ? memsize : size;
+ memcpy(ret, mem, tocopy);
+ if((flags & M2_ZERO_MEMORY) && size>memsize)
+ memset((void *)((size_t)ret+memsize), 0, size-memsize);
+ LogOperation(tc, p, LOGENTRY_THREADCACHE_MALLOC, mymspace, size, mem, alignment, flags, ret);
+ if(!isforeign && memsize>=sizeof(threadcacheblk) && memsize<=(THREADCACHEMAX+CHUNK_OVERHEAD))
+ {
+ threadcache_free(p, tc, mymspace, mem, memsize, isforeign);
+ LogOperation(tc, p, LOGENTRY_THREADCACHE_FREE, mymspace, memsize, mem, 0, 0, 0);
+ }
+ else
+ {
+ CallFree(0, mem, isforeign);
+ LogOperation(tc, p, LOGENTRY_POOL_FREE, mymspace, memsize, mem, 0, 0, 0);
+ }
+ }
+ }
+#endif
+ if(!ret)
+ { /* Reallocs always happen in the mspace they happened in, so skip
+ locking the preferred mspace for this thread */
+ ret=CallRealloc(p->m[mymspace], mem, isforeign, memsize, size, alignment, flags);
+ if(ret)
+ LogOperation(tc, p, LOGENTRY_POOL_REALLOC, mymspace, size, mem, alignment, flags, ret);
+ }
+ LogOperation(tc, p, LOGENTRY_REALLOC, mymspace, size, mem, alignment, flags, ret);
+ return ret;
+}
+NEDMALLOCNOALIASATTR void nedpfree2(nedpool *p, void *mem, unsigned flags) THROWSPEC
+{ /* Frees always happen in the mspace they happened in, so skip
+ locking the preferred mspace for this thread */
+ threadcache *tc;
+ int mymspace, isforeign=1;
+ size_t memsize;
+ if(!mem)
+ { /* If you tried this on FreeBSD you'd be sorry! */
+#ifdef DEBUG
+ fprintf(stderr, "nedmalloc: WARNING nedpfree() called with zero. This is not portable behaviour!\n");
+#endif
+ return;
+ }
+ memsize=nedblksize(&isforeign, mem, flags);
+ assert(memsize);
+ if(!memsize)
+ {
+ fprintf(stderr, "nedmalloc: nedpfree() called with a block not created by nedmalloc!\n");
+ abort();
+ }
+ GetThreadCache(&p, &tc, &mymspace, 0);
+#if THREADCACHEMAX
+ if(mem && tc && !isforeign && memsize>=sizeof(threadcacheblk) && memsize<=(THREADCACHEMAX+CHUNK_OVERHEAD))
+ {
+ threadcache_free(p, tc, mymspace, mem, memsize, isforeign);
+ LogOperation(tc, p, LOGENTRY_THREADCACHE_FREE, mymspace, memsize, mem, 0, 0, 0);
+ }
+ else
+#endif
+ {
+ CallFree(0, mem, isforeign);
+ LogOperation(tc, p, LOGENTRY_POOL_FREE, mymspace, memsize, mem, 0, 0, 0);
+ }
+ LogOperation(tc, p, LOGENTRY_FREE, mymspace, memsize, mem, 0, 0, 0);
+}
+NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedpmalloc(nedpool *p, size_t size) THROWSPEC
+{
+ unsigned flags=NEDMALLOC_FORCERESERVE(p, 0, size);
+ return nedpmalloc2(p, size, 0, flags);
+}
+NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedpcalloc(nedpool *p, size_t no, size_t size) THROWSPEC
+{
+ size_t bytes=no*size;
+ unsigned flags=NEDMALLOC_FORCERESERVE(p, 0, bytes);
+ /* Avoid multiplication overflow. */
+ if(size && no!=bytes/size) return 0;
+ return nedpmalloc2(p, bytes, 0, M2_ZERO_MEMORY|flags);
+}
+NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedprealloc(nedpool *p, void *mem, size_t size) THROWSPEC
+{
+ unsigned flags=NEDMALLOC_FORCERESERVE(p, mem, size);
+#if ENABLE_USERMODEPAGEALLOCATOR
+ /* If the user mode page allocator is turned on in a 32 bit process,
+ don't automatically reserve eight times the address space. */
+ if(8==sizeof(size_t) || !OSHavePhysicalPageSupport())
+#endif
+ { /* If he reallocs even once, it's probably wise to turn on address space reservation.
+ If the size is larger than mmap_threshold then it'll set the reserve. */
+ if(!(flags & M2_RESERVE_MASK)) flags=M2_RESERVE_MULT(8);
+ }
+ return nedprealloc2(p, mem, size, 0, flags);
+}
+NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedpmemalign(nedpool *p, size_t alignment, size_t bytes) THROWSPEC
+{
+ unsigned flags=NEDMALLOC_FORCERESERVE(p, 0, bytes);
+ return nedpmalloc2(p, bytes, alignment, flags);
+}
+NEDMALLOCNOALIASATTR void nedpfree(nedpool *p, void *mem) THROWSPEC
+{
+ nedpfree2(p, mem, 0);
+}
+
+struct nedmallinfo nedpmallinfo(nedpool *p) THROWSPEC
+{
+ int n;
+ struct nedmallinfo ret={0};
+ if(!p) { p=&syspool; if(!syspool.threads) InitPool(&syspool, 0, -1); }
+ for(n=0; p->m[n]; n++)
+ {
+#if USE_ALLOCATOR==1 && !NO_MALLINFO
+ struct mallinfo t=mspace_mallinfo(p->m[n]);
+ ret.arena+=t.arena;
+ ret.ordblks+=t.ordblks;
+ ret.hblkhd+=t.hblkhd;
+ ret.usmblks+=t.usmblks;
+ ret.uordblks+=t.uordblks;
+ ret.fordblks+=t.fordblks;
+ ret.keepcost+=t.keepcost;
+#endif
+ }
+ return ret;
+}
+int nedpmallopt(nedpool *p, int parno, int value) THROWSPEC
+{
+#if USE_ALLOCATOR==1
+ return mspace_mallopt(parno, value);
+#else
+ return 0;
+#endif
+}
+NEDMALLOCNOALIASATTR void* nedmalloc_internals(size_t *granularity, size_t *magic) THROWSPEC
+{
+#if USE_ALLOCATOR==1
+ if(granularity) *granularity=mparams.granularity;
+ if(magic) *magic=mparams.magic;
+ return (void *) &syspool;
+#else
+ if(granularity) *granularity=0;
+ if(magic) *magic=0;
+ return 0;
+#endif
+}
+int nedpmalloc_trim(nedpool *p, size_t pad) THROWSPEC
+{
+ int n, ret=0;
+ if(!p) { p=&syspool; if(!syspool.threads) InitPool(&syspool, 0, -1); }
+ for(n=0; p->m[n]; n++)
+ {
+#if USE_ALLOCATOR==1
+ ret+=mspace_trim(p->m[n], pad);
+#endif
+ }
+ return ret;
+}
+void nedpmalloc_stats(nedpool *p) THROWSPEC
+{
+ int n;
+ if(!p) { p=&syspool; if(!syspool.threads) InitPool(&syspool, 0, -1); }
+ for(n=0; p->m[n]; n++)
+ {
+#if USE_ALLOCATOR==1
+ mspace_malloc_stats(p->m[n]);
+#endif
+ }
+}
+size_t nedpmalloc_footprint(nedpool *p) THROWSPEC
+{
+ size_t ret=0;
+ int n;
+ if(!p) { p=&syspool; if(!syspool.threads) InitPool(&syspool, 0, -1); }
+ for(n=0; p->m[n]; n++)
+ {
+#if USE_ALLOCATOR==1
+ ret+=mspace_footprint(p->m[n]);
+#endif
+ }
+ return ret;
+}
+static FORCEINLINE NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void ** CallIndependentCalloc(void *RESTRICT m, size_t elemsno, size_t elemsize, void **chunks) THROWSPEC
+{
+ void **ret;
+#if USE_ALLOCATOR==0
+ ret=(void **) unsupported_operation("independent_calloc");
+#elif USE_ALLOCATOR==1
+ ret=mspace_independent_calloc(m, elemsno, elemsize, chunks);
+#ifdef HAVE_VALGRIND
+ if(ret)
+ {
+ size_t n;
+ for(n=0; n<elemsno; n++)
+ VALGRIND_MEMPOOL_ALLOC(m, ret[n], elemsize);
+ }
+#endif
+#endif
+ if(ret)
+ {
+ if(!leastusedaddress || (void *)((mstate) m)->least_addr<leastusedaddress) leastusedaddress=(void *)((mstate) m)->least_addr;
+ /*if(!largestusedblock || truesize>largestusedblock) largestusedblock=(truesize+mparams.page_size) & ~(mparams.page_size-1);*/
+ }
+ return ret;
+}
+NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void **nedpindependent_calloc(nedpool *p, size_t elemsno, size_t elemsize, void **chunks) THROWSPEC
+{
+ void **ret;
+ threadcache *tc;
+ int mymspace;
+ GetThreadCache(&p, &tc, &mymspace, &elemsize);
+ GETMSPACE(m, p, tc, mymspace, elemsno*elemsize,
+ ret=CallIndependentCalloc(m, elemsno, elemsize, chunks));
+#if ENABLE_LOGGING
+ if(ret && (ENABLE_LOGGING & LOGENTRY_POOL_MALLOC))
+ {
+ size_t n;
+ for(n=0; n<elemsno; n++)
+ {
+ LogOperation(tc, p, LOGENTRY_POOL_MALLOC, mymspace, elemsize, 0, 0, M2_ZERO_MEMORY, ret[n]);
+ }
+ }
+#endif
+ return ret;
+}
+static FORCEINLINE NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void **CallIndependentComalloc(void *RESTRICT m, size_t elems, size_t *RESTRICT sizes, void **chunks) THROWSPEC
+{
+ void **ret;
+#if USE_ALLOCATOR==0
+ ret=(void **) unsupported_operation("independent_comalloc");
+#elif USE_ALLOCATOR==1
+ ret=mspace_independent_comalloc(m, elems, sizes, chunks);
+#ifdef HAVE_VALGRIND
+ if(ret)
+ {
+ size_t n;
+ for(n=0; n<elems; n++)
+ VALGRIND_MEMPOOL_ALLOC(m, ret[n], sizes[n]);
+ }
+#endif
+#endif
+ if(ret)
+ {
+ if(!leastusedaddress || (void *)((mstate) m)->least_addr<leastusedaddress) leastusedaddress=(void *)((mstate) m)->least_addr;
+ /*if(!largestusedblock || truesize>largestusedblock) largestusedblock=(truesize+mparams.page_size) & ~(mparams.page_size-1);*/
+ }
+ return ret;
+}
+NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void **nedpindependent_comalloc(nedpool *p, size_t elems, size_t *sizes, void **chunks) THROWSPEC
+{
+ void **ret;
+ threadcache *tc;
+ int mymspace;
+ size_t i, *adjustedsizes=(size_t *) alloca(elems*sizeof(size_t));
+ if(!adjustedsizes) return 0;
+ for(i=0; i<elems; i++)
+ adjustedsizes[i]=sizes[i]<sizeof(threadcacheblk) ? sizeof(threadcacheblk) : sizes[i];
+ GetThreadCache(&p, &tc, &mymspace, 0);
+ GETMSPACE(m, p, tc, mymspace, 0,
+ ret=CallIndependentComalloc(m, elems, adjustedsizes, chunks));
+#if ENABLE_LOGGING
+ if(ret && (ENABLE_LOGGING & LOGENTRY_POOL_MALLOC))
+ {
+ size_t n;
+ for(n=0; n<elems; n++)
+ {
+ LogOperation(tc, p, LOGENTRY_POOL_MALLOC, mymspace, sizes[n], 0, 0, 0, ret[n]);
+ }
+ }
+#endif
+ return ret;
+}
+
+#if defined(__cplusplus)
+}
+#endif
+
+#ifdef _MSC_VER
+#pragma warning(pop)
+#endif
diff --git a/nedmalloc.h b/nedmalloc.h new file mode 100644 index 0000000..ba699be --- /dev/null +++ b/nedmalloc.h @@ -0,0 +1,1620 @@ +/* nedalloc, an alternative malloc implementation for multiple threads without +lock contention based on dlmalloc v2.8.4. (C) 2005-2010 Niall Douglas + +Boost Software License - Version 1.0 - August 17th, 2003 + +Permission is hereby granted, free of charge, to any person or organization +obtaining a copy of the software and accompanying documentation covered by +this license (the "Software") to use, reproduce, display, distribute, +execute, and transmit the Software, and to prepare derivative works of the +Software, and to permit third-parties to whom the Software is furnished to +do so, all subject to the following: + +The copyright notices in the Software and this entire statement, including +the above license grant, this restriction and the following disclaimer, +must be included in all copies of the Software, in whole or in part, and +all derivative works of the Software, unless such copies or derivative +works are solely in the form of machine-executable object code generated by +a source language processor. + +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE, TITLE AND NON-INFRINGEMENT. IN NO EVENT +SHALL THE COPYRIGHT HOLDERS OR ANYONE DISTRIBUTING THE SOFTWARE BE LIABLE +FOR ANY DAMAGES OR OTHER LIABILITY, WHETHER IN CONTRACT, TORT OR OTHERWISE, +ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER +DEALINGS IN THE SOFTWARE. +*/ + +#ifndef NEDMALLOC_H +#define NEDMALLOC_H + +/*! \file nedmalloc.h +\brief Defines the functionality provided by nedalloc. +*/ + +/*! \mainpage + +<a href="../../Readme.html">Please see the Readme.html</a> +*/ + +/*! \def NEDMALLOC_DEBUG +\brief Defines the assertion checking performed by nedalloc + +NEDMALLOC_DEBUG can be defined to cause DEBUG to be set differently for nedmalloc +than for the rest of the build. Remember to set NDEBUG to disable all assertion +checking too. +*/ + +/*! \def ENABLE_LARGE_PAGES +\brief Defines whether nedalloc uses large pages (>=2Mb) + +ENABLE_LARGE_PAGES enables support for requesting memory from the system in large +(typically >=2Mb) pages if the host OS supports this. These occupy just a single +TLB entry and can significantly improve performance in large working set applications. +*/ + +/*! \def ENABLE_FAST_HEAP_DETECTION +\brief Defines whether nedalloc takes platform specific shortcuts when detecting foreign blocks. + +ENABLE_FAST_HEAP_DETECTION enables special logic to detect blocks allocated +by the system heap. This avoids 1.5%-2% overhead when checking for non-nedmalloc +blocks, but it assumes that the NT and glibc heaps function in a very specific +fashion which may not hold true across OS upgrades. +*/ + +/*! \def HAVE_CPP0XRVALUEREFS +\ingroup C++ +\brief Enables rvalue references + +Define to enable the usage of rvalue references which enables move semantics and +other things. Automatically defined if __cplusplus indicates a C++0x compiler, +otherwise you'll need to set it yourself. +*/ + +/*! \def HAVE_CPP0XVARIADICTEMPLATES +\ingroup C++ +\brief Enables variadic templates + +Define to enable the usage of variadic templates which enables the use of arbitrary +numbers of policies and other useful things. Automatically defined if __cplusplus +indicates a C++0x compiler, otherwise you'll need to set it yourself. +*/ + +/*! \def HAVE_CPP0XSTATICASSERT +\ingroup C++ +\brief Enables static assertions + +Define to enable the usage of static assertions. Automatically defined if __cplusplus +indicates a C++0x compiler, otherwise you'll need to set it yourself. +*/ + +/*! \def HAVE_CPP0XTYPETRAITS +\ingroup C++ +\brief Enables type traits + +Define to enable the usage of <type_traits>. Automatically defined if __cplusplus +indicates a C++0x compiler, otherwise you'll need to set it yourself. +*/ + +#if __cplusplus > 199711L || defined(HAVE_CPP0X) /* Do we have C++0x? */ +#undef HAVE_CPP0XRVALUEREFS +#define HAVE_CPP0XRVALUEREFS 1 +#undef HAVE_CPP0XVARIADICTEMPLATES +#define HAVE_CPP0XVARIADICTEMPLATES 1 +#undef HAVE_CPP0XSTATICASSERT +#define HAVE_CPP0XSTATICASSERT 1 +#undef HAVE_CPP0XTYPETRAITS +#define HAVE_CPP0XTYPETRAITS 1 +#endif + +#include <stddef.h> /* for size_t */ + +/*! \def NEDMALLOCEXTSPEC +\brief Defines how nedalloc's API is to be made visible. + +NEDMALLOCEXTSPEC can be defined to be __declspec(dllexport) or +__attribute__ ((visibility("default"))) or whatever you like. It defaults +to extern unless NEDMALLOC_DLL_EXPORTS is set as it would be when building +nedmalloc.dll. + */ +#ifndef NEDMALLOCEXTSPEC + #ifdef NEDMALLOC_DLL_EXPORTS + #ifdef WIN32 + #define NEDMALLOCEXTSPEC extern __declspec(dllexport) + #elif defined(__GNUC__) + #define NEDMALLOCEXTSPEC extern __attribute__ ((visibility("default"))) + #endif + #ifndef ENABLE_TOLERANT_NEDMALLOC + #define ENABLE_TOLERANT_NEDMALLOC 1 + #endif + #else + #define NEDMALLOCEXTSPEC extern + #endif +#endif + +/*! \def NEDMALLOCDEPRECATED +\brief Defined to mark an API as deprecated */ +#ifndef NEDMALLOCDEPRECATED +#if defined(_MSC_VER) && !defined(__GCCXML__) + #define NEDMALLOCDEPRECATED __declspec(deprecated) +#elif defined(__GNUC__) && !defined(__GCCXML__) + #define NEDMALLOCDEPRECATED __attribute ((deprecated)) +#else +//! Marks a function as being deprecated + #define NEDMALLOCDEPRECATED +#endif +#endif + +/*! \def RESTRICT +\brief Defined to the restrict keyword or equivalent if available */ +#ifndef RESTRICT +#if __STDC_VERSION__ >= 199901L /* C99 or better */ + #define RESTRICT restrict +#else + #if defined(_MSC_VER) && _MSC_VER>=1400 + #define RESTRICT __restrict + #endif + #ifdef __GNUC__ + #define RESTRICT __restrict + #endif +#endif +#ifndef RESTRICT + #define RESTRICT +#endif +#endif + +#if defined(_MSC_VER) && _MSC_VER>=1400 + #define NEDMALLOCPTRATTR __declspec(restrict) + #define NEDMALLOCNOALIASATTR __declspec(noalias) +#endif +#ifdef __GNUC__ + #define NEDMALLOCPTRATTR __attribute__ ((malloc)) +#endif +/*! \def NEDMALLOCPTRATTR +\brief Defined to the specifier for a pointer which points to a memory block. Like NEDMALLOCNOALIASATTR, but sadly not identical. */ +#ifndef NEDMALLOCPTRATTR + #define NEDMALLOCPTRATTR +#endif +/*! \def NEDMALLOCNOALIASATTR +\brief Defined to the specifier for a pointer which does not alias any other variable. */ +#ifndef NEDMALLOCNOALIASATTR + #define NEDMALLOCNOALIASATTR +#endif + +/*! \def USE_MAGIC_HEADERS +\brief Defines whether nedalloc should use magic headers in foreign heap block detection + +USE_MAGIC_HEADERS causes nedalloc to allocate an extra three sizeof(size_t) +to each block. nedpfree() and nedprealloc() can then automagically know when +to free a system allocated block. Enabling this typically adds 20-50% to +application memory usage, and is mandatory if USE_ALLOCATOR is not 1. +*/ +#ifndef USE_MAGIC_HEADERS + #define USE_MAGIC_HEADERS 0 +#endif + +/*! \def USE_ALLOCATOR +\brief Defines the underlying allocator to use + +USE_ALLOCATOR can be one of these settings (it defaults to 1): + 0: System allocator (nedmalloc now simply acts as a threadcache) which is + very useful for testing with valgrind and Glowcode. + WARNING: Intended for DEBUG USE ONLY - not all functions work correctly. + 1: dlmalloc +*/ +#ifndef USE_ALLOCATOR + #define USE_ALLOCATOR 1 /* dlmalloc */ +#endif + +#if !USE_ALLOCATOR && !USE_MAGIC_HEADERS +#error If you are using the system allocator then you MUST use magic headers +#endif + +/*! \def REPLACE_SYSTEM_ALLOCATOR +\brief Defines whether to replace the system allocator (malloc(), free() et al) with nedalloc's implementation. + +REPLACE_SYSTEM_ALLOCATOR on POSIX causes nedalloc's functions to be called +malloc, free etc. instead of nedmalloc, nedfree etc. You may or may not want +this. On Windows it causes nedmalloc to patch all loaded DLLs and binaries +to replace usage of the system allocator. + +Always turns on ENABLE_TOLERANT_NEDMALLOC. +*/ +#ifdef REPLACE_SYSTEM_ALLOCATOR + #if USE_ALLOCATOR==0 + #error Cannot combine using the system allocator with replacing the system allocator + #endif + #ifndef ENABLE_TOLERANT_NEDMALLOC + #define ENABLE_TOLERANT_NEDMALLOC 1 + #endif + #ifndef WIN32 /* We have a dedicated patcher for Windows */ + #define nedmalloc malloc + #define nedmalloc2 malloc2 + #define nedcalloc calloc + #define nedrealloc realloc + #define nedrealloc2 realloc2 + #define nedfree free + #define nedfree2 free2 + #define nedmemalign memalign + #define nedmallinfo mallinfo + #define nedmallopt mallopt + #define nedmalloc_trim malloc_trim + #define nedmalloc_stats malloc_stats + #define nedmalloc_footprint malloc_footprint + #define nedindependent_calloc independent_calloc + #define nedindependent_comalloc independent_comalloc + #ifdef __GNUC__ + #define nedmemsize malloc_usable_size + #endif + #endif +#endif + +/*! \def ENABLE_TOLERANT_NEDMALLOC +\brief Defines whether nedalloc should check for blocks from the system allocator. + +ENABLE_TOLERANT_NEDMALLOC is automatically turned on if REPLACE_SYSTEM_ALLOCATOR +is set or the Windows DLL is being built. This causes nedmalloc to detect when a +system allocator block is passed to it and to handle it appropriately. Note that +without USE_MAGIC_HEADERS there is a very tiny chance that nedmalloc will segfault +on non-Windows builds (it uses Win32 SEH to trap segfaults on Windows and there +is no comparable system on POSIX). +*/ + +#if defined(__cplusplus) +extern "C" { +#endif +/*! \brief Returns information about a memory pool */ +struct nedmallinfo { + size_t arena; /*!< non-mmapped space allocated from system */ + size_t ordblks; /*!< number of free chunks */ + size_t smblks; /*!< always 0 */ + size_t hblks; /*!< always 0 */ + size_t hblkhd; /*!< space in mmapped regions */ + size_t usmblks; /*!< maximum total allocated space */ + size_t fsmblks; /*!< always 0 */ + size_t uordblks; /*!< total allocated space */ + size_t fordblks; /*!< total free space */ + size_t keepcost; /*!< releasable (via malloc_trim) space */ +}; +#if defined(__cplusplus) +} +#endif + +/*! \def NO_NED_NAMESPACE +\brief Defines the use of the nedalloc namespace for the C functions. + +NO_NED_NAMESPACE prevents the functions from being defined in the nedalloc +namespace when in C++ (uses the global C namespace instead). +*/ +/*! \def THROWSPEC +\brief Defined to throw() or noexcept(true) (as in, throws nothing) under C++, otherwise nothing. +*/ +#if defined(__cplusplus) + #if !defined(NO_NED_NAMESPACE) +namespace nedalloc { + #else +extern "C" { + #endif + #if __cplusplus > 199711L + #define THROWSPEC noexcept(true) + #else + #define THROWSPEC throw() + #endif +#else + #define THROWSPEC +#endif + +/* These are the global functions */ + +/*! \defgroup v2malloc The v2 malloc API + +\warning This API is being completely retired in v1.10 beta 2 and replaced with the API +being developed for inclusion into the C1X programming language standard + +For the v1.10 release which was generously sponsored by +<a href="http://www.ara.com/" target="_blank">Applied Research Associates (USA)</a>, +a new general purpose allocator API was designed which is intended to remedy many +of the long standing problems and inefficiencies introduced by the ISO C allocator +API. Internally nedalloc's implementations of nedmalloc(), nedcalloc(), nedmemalign() +and nedrealloc() call into this API: + +<ul> + <li><code>void* malloc2(size_t bytes, size_t alignment, unsigned flags)</code></li> + <li><code>void* realloc2(void* mem, size_t bytes, size_t alignment, unsigned + flags)</code></li> + <li><code>void free2(void* mem, unsigned flags)</code></li> +</ul> + +If nedmalloc.h is being included by C++ code, the alignment and flags parameters +default to zero which makes the new API identical to the old API (roll on the introduction +of default parameters to C!). The ability for realloc2() to take an alignment is +<em>particularly</em> useful for extending aligned vector arrays such as SSE/AVX +vector arrays. Hitherto SSE/AVX vector code had to jump through all sorts of unpleasant +hoops to maintain alignment :(. + +Note that using any of these flags other than M2_ZERO_MEMORY or any alignment +other than zero inhibits the threadcache. + +Currently MREMAP support is limited to Linux and Windows. Patches implementing +support for other platforms are welcome. + +On Linux the non portable mremap() kernel function is currently used, so in fact +the M2_RESERVE_* options are currently ignored. + +On Windows, there are two different MREMAP implementations which are chosen according +to whether a 32 bit or a 64 bit build is being performed. The 32 bit implementation +is based on Win32 file mappings where it reserves the address space within the Windows +VM system, so you can safely specify silly reservation quantities like 2Gb per block +and not exhaust local process address space. Note however that on x86 this costs +2Kb (1Kb if PAE is off) of kernel memory per Mb reserved, and as kernel memory has +a hard limit of 447Mb on x86 you will find the total address space reservable in +the system is limited. On x64, or if you define WIN32_DIRECT_USE_FILE_MAPPINGS=0 +on x86, a much faster implementation of using VirtualAlloc(MEM_RESERVE) to directly +reserve the address space is used. + +When using M2_RESERVE_* with realloc2(), the setting only takes effect when the +mmapped chunk has exceeded its reservation space and a new reservation space needs +to be created. +*/ + +#ifndef M2_FLAGS_DEFINED +#define M2_FLAGS_DEFINED + +/*! \def M2_ZERO_MEMORY +\ingroup v2malloc +\brief Sets the contents of the allocated block (or any increase in the allocated +block) to zero. + +Note that this zeroes only the increase from what dlmalloc thinks +the chunk's size is, so if you realloc2() a block which wasn't allocated using +malloc2() using this flag then you may have garbage just before the newly extended +space. + +\li <strong>Rationale:</strong> Memory returned by the system is guaranteed to +be zero on most platforms, and hence dlmalloc knows when it can skip zeroing +memory. This improves performance. +*/ +#define M2_ZERO_MEMORY (1<<0) + +/*! \def M2_PREVENT_MOVE +\ingroup v2malloc +\brief Cause realloc2() to attempt to extend a block in place, but to never move +it. + +\li <strong>Rationale:</strong> C++ makes almost no use of realloc(), even for +contiguous arrays such as std::vector<> because most C++ objects cannot be relocated +in memory without a copy or rvalue construction (though some clever STL implementations +specialise for Plain Old Data (POD) types, and use realloc() then and only then). +This flag allows C++ containers to speculatively try to extend in place, thus +improving performance <em>especially</em> for large allocations which will use +mmap(). +*/ +#define M2_PREVENT_MOVE (1<<1) + +/*! \def M2_ALWAYS_MMAP +\ingroup v2malloc +\brief Always allocate as though mmap_threshold were being exceeded. + +In the case of realloc2(), note that setting this bit will not necessarily mmap a chunk +which isn't already mmapped, but it will force a mmapped chunk if new memory +needs allocating. + +\li <strong>Rationale:</strong> If you know that an array you are allocating +is going to be repeatedly extended up into the hundred of kilobytes range, then +you can avoid the constant memory copying into larger blocks by specifying this +flag at the beginning along with one of the M2_RESERVE_* flags below. This can +<strong>greatly</strong> improve performance for large arrays. +*/ +#define M2_ALWAYS_MMAP (1<<2) +#define M2_RESERVED1 (1<<3) +#define M2_RESERVED2 (1<<4) +#define M2_RESERVED3 (1<<5) +#define M2_RESERVED4 (1<<6) +#define M2_RESERVED5 (1<<7) +#define M2_RESERVE_ISMULTIPLIER (1<<15) +/* 7 bits is given to the address reservation specifier. +This lets you set a multiplier (bit 15 set) or a 1<< shift value. +*/ +#define M2_RESERVE_MASK 0x00007f00 + +/*! \def M2_RESERVE_MULT(n) +\ingroup v2malloc +\brief Reserve n times as much address space such that mmapped realloc2(size <= +n * original size) avoids memory copying and hence is much faster. +*/ +#define M2_RESERVE_MULT(n) (M2_RESERVE_ISMULTIPLIER|(((n)<<8)&M2_RESERVE_MASK)) + +/*! \def M2_RESERVE_SHIFT(n) +\ingroup v2malloc +\brief Reserve (1<<n) bytes of address space such that mmapped realloc2(size <= +(1<<n)) avoids memory copying and hence is much faster. +*/ +#define M2_RESERVE_SHIFT(n) (((n)<<8)&M2_RESERVE_MASK) +#define M2_FLAGS_MASK 0x0000ffff +#define M2_CUSTOM_FLAGS_BEGIN (1<<16) +#define M2_CUSTOM_FLAGS_MASK 0xffff0000 + +/*! \def NM_SKIP_TOLERANCE_CHECKS +\ingroup v2malloc +\brief Causes nedmalloc to not inspect the block being passed to see if it belongs +to the system allocator. Can improve speed by up to 10%. +*/ +#define NM_SKIP_TOLERANCE_CHECKS (1<<31) +#endif /* M2_FLAGS_DEFINED */ + + +#if defined(__cplusplus) +/*! \brief Gets the usable size of an allocated block. + +Note this will always be bigger than what was +asked for due to rounding etc. Optionally returns 1 in isforeign if the block came from the +system allocator - note that there is a small (>0.01%) but real chance of segfault on non-Windows +systems when passing non-nedmalloc blocks if you don't use USE_MAGIC_HEADERS. +*/ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR size_t nedblksize(int *RESTRICT isforeign, void *RESTRICT mem, unsigned flags=0) THROWSPEC; +#else +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR size_t nedblksize(int *RESTRICT isforeign, void *RESTRICT mem, unsigned flags) THROWSPEC; +#endif +/*! \brief Identical to nedblksize() except without the isforeign */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR size_t nedmemsize(void *RESTRICT mem) THROWSPEC; + +/*! \brief Equivalent to nedpsetvalue((nedpool *) 0, v) */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR void nedsetvalue(void *v) THROWSPEC; + +/*! \brief Equivalent to nedpmalloc2((nedpool *) 0, size, 0, 0) */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedmalloc(size_t size) THROWSPEC; +/*! \brief Equivalent to nedpmalloc2((nedpool *) 0, no*size, 0, M2_ZERO_MEMORY) */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedcalloc(size_t no, size_t size) THROWSPEC; +/*! \brief Equivalent to nedprealloc2((nedpool *) 0, size, mem, size, 0, M2_RESERVE_MULT(8)) */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedrealloc(void *mem, size_t size) THROWSPEC; +/*! \brief Equivalent to nedpfree2((nedpool *) 0, mem, 0) */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR void nedfree(void *mem) THROWSPEC; +/*! \brief Equivalent to nedpmalloc2((nedpool *) 0, size, alignment, 0) */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedmemalign(size_t alignment, size_t bytes) THROWSPEC; + +#if defined(__cplusplus) +/*! \ingroup v2malloc +\brief Equivalent to nedpmalloc2((nedpool *) 0, size, alignment, flags) */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedmalloc2(size_t size, size_t alignment=0, unsigned flags=0) THROWSPEC; +/*! \ingroup v2malloc +\brief Equivalent to nedprealloc2((nedpool *) 0, mem, size, alignment, flags) */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedrealloc2(void *mem, size_t size, size_t alignment=0, unsigned flags=0) THROWSPEC; +/*! \ingroup v2malloc +\brief Equivalent to nedpfree2((nedpool *) 0, mem, flags) */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR void nedfree2(void *mem, unsigned flags=0) THROWSPEC; +#else +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedmalloc2(size_t size, size_t alignment, unsigned flags) THROWSPEC; +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedrealloc2(void *mem, size_t size, size_t alignment, unsigned flags) THROWSPEC; +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR void nedfree2(void *mem, unsigned flags) THROWSPEC; +#endif + +/*! \brief Equivalent to nedpmallinfo((nedpool *) 0) */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR struct nedmallinfo nedmallinfo(void) THROWSPEC; +/*! \brief Equivalent to nedpmallopt((nedpool *) 0, parno, value) */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR int nedmallopt(int parno, int value) THROWSPEC; +/*! \brief Returns the internal allocation granularity and the magic header XOR used for internal consistency checks. */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR void* nedmalloc_internals(size_t *granularity, size_t *magic) THROWSPEC; +/*! \brief Equivalent to nedpmalloc_trim((nedpool *) 0, pad) */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR int nedmalloc_trim(size_t pad) THROWSPEC; +/*! \brief Equivalent to nedpmalloc_stats((nedpool *) 0) */ +NEDMALLOCEXTSPEC void nedmalloc_stats(void) THROWSPEC; +/*! \brief Equivalent to nedpmalloc_footprint((nedpool *) 0) */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR size_t nedmalloc_footprint(void) THROWSPEC; +/*! \brief Equivalent to nedpindependent_calloc((nedpool *) 0, elemsno, elemsize, chunks) */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void **nedindependent_calloc(size_t elemsno, size_t elemsize, void **chunks) THROWSPEC; +/*! \brief Equivalent to nedpindependent_comalloc((nedpool *) 0, elems, sizes, chunks) */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void **nedindependent_comalloc(size_t elems, size_t *sizes, void **chunks) THROWSPEC; + +/*! \brief Destroys the system memory pool used by the functions above. + +Useful for when you have nedmalloc in a DLL you're about to unload. +If you call ANY nedmalloc functions after calling this you will +get a fatal exception! +*/ +NEDMALLOCEXTSPEC void neddestroysyspool() THROWSPEC; + +/*! \brief A nedpool type */ +struct nedpool_t; +/*! \brief A nedpool type */ +typedef struct nedpool_t nedpool; + +/*! \brief Creates a memory pool for use with the nedp* functions below. + +Capacity is how much to allocate immediately (if you know you'll be allocating a lot +of memory very soon) which you can leave at zero. Threads specifies how many threads +will *normally* be accessing the pool concurrently. Setting this to zero means it +extends on demand, but be careful of this as it can rapidly consume system resources +where bursts of concurrent threads use a pool at once. +*/ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR nedpool *nedcreatepool(size_t capacity, int threads) THROWSPEC; + +/*! \brief Destroys a memory pool previously created by nedcreatepool(). +*/ +NEDMALLOCEXTSPEC void neddestroypool(nedpool *p) THROWSPEC; + +/*! \brief Returns a zero terminated snapshot of threadpools existing at the time of call. + +Call nedfree() on the returned list when you are done. Returns zero if there is only the +system pool in existence. +*/ +NEDMALLOCEXTSPEC nedpool **nedpoollist() THROWSPEC; + +/*! \brief Sets a value to be associated with a pool. + +You can retrieve this value by passing any memory block allocated from that pool. +*/ +NEDMALLOCEXTSPEC void nedpsetvalue(nedpool *p, void *v) THROWSPEC; + +/*! \brief Gets a previously set value using nedpsetvalue() or zero if memory is unknown. + +Optionally can also retrieve pool. You can detect an unknown block by the return +being zero and *p being unmodifed. +*/ +NEDMALLOCEXTSPEC void *nedgetvalue(nedpool **p, void *mem) THROWSPEC; + +/*! \brief Trims the thread cache for the calling thread, returning any existing cache +data to the central pool. + +Remember to ALWAYS call with zero if you used the system pool. Setting disable to +non-zero replicates neddisablethreadcache(). +*/ +NEDMALLOCEXTSPEC void nedtrimthreadcache(nedpool *p, int disable) THROWSPEC; + +/*! \brief Disables the thread cache for the calling thread, returning any existing cache +data to the central pool. + +Remember to ALWAYS call with zero if you used the system pool. +*/ +NEDMALLOCEXTSPEC void neddisablethreadcache(nedpool *p) THROWSPEC; + +/*! \brief Releases all memory in all threadcaches in the pool, and writes all +accumulated memory operations to the log if enabled. + +You can pass zero for filepath to use the compiled default, or else a char[MAX_PATH] +containing the path you wish to use for the log file. The log file is always +appended to if it already exists. After writing the logs, the logging ability +is disabled for that pool. + +\warning Do NOT call this if the pool is in use - this call is NOT threadsafe. +*/ +NEDMALLOCEXTSPEC size_t nedflushlogs(nedpool *p, char *filepath) THROWSPEC; + + +/*! \brief Equivalent to nedpmalloc2(p, size, 0, 0) */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedpmalloc(nedpool *p, size_t size) THROWSPEC; +/*! \brief Equivalent to nedpmalloc2(p, no*size, 0, M2_ZERO_MEMORY) */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedpcalloc(nedpool *p, size_t no, size_t size) THROWSPEC; +/*! \brief Equivalent to nedprealloc2(p, mem, size, 0, M2_RESERVE_MULT(8)) */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedprealloc(nedpool *p, void *mem, size_t size) THROWSPEC; +/*! \brief Equivalent to nedpfree2(p, mem, 0) */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR void nedpfree(nedpool *p, void *mem) THROWSPEC; +/*! \brief Equivalent to nedpmalloc2(p, bytes, alignment, 0) */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedpmemalign(nedpool *p, size_t alignment, size_t bytes) THROWSPEC; +#if defined(__cplusplus) +/*! \ingroup v2malloc +\brief Allocates a block of memory sized \em size from pool \em p, aligned to \em alignment and according to the flags \em flags. +*/ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedpmalloc2(nedpool *p, size_t size, size_t alignment=0, unsigned flags=0) THROWSPEC; +/*! \ingroup v2malloc +\brief Resizes the block of memory at \em mem in pool \em p to size \em size, aligned to \em alignment and according to the flags \em flags. +*/ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedprealloc2(nedpool *p, void *mem, size_t size, size_t alignment=0, unsigned flags=0) THROWSPEC; +/*! \brief Frees the block \em mem from the pool \em p according to flags \em flags. */ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR void nedpfree2(nedpool *p, void *mem, unsigned flags=0) THROWSPEC; +#else +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedpmalloc2(nedpool *p, size_t size, size_t alignment, unsigned flags) THROWSPEC; +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void * nedprealloc2(nedpool *p, void *mem, size_t size, size_t alignment, unsigned flags) THROWSPEC; +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR void nedpfree2(nedpool *p, void *mem, unsigned flags) THROWSPEC; +#endif +/*! \brief Returns information about the memory pool */ +NEDMALLOCEXTSPEC struct nedmallinfo nedpmallinfo(nedpool *p) THROWSPEC; +/*! \brief Changes the operational parameters of the memory pool */ +NEDMALLOCEXTSPEC int nedpmallopt(nedpool *p, int parno, int value) THROWSPEC; +/*! \brief Tries to release as much free memory back to the system as possible, leaving \em pad remaining per threadpool. */ +NEDMALLOCEXTSPEC int nedpmalloc_trim(nedpool *p, size_t pad) THROWSPEC; +/*! \brief Prints some operational statistics to stdout. */ +NEDMALLOCEXTSPEC void nedpmalloc_stats(nedpool *p) THROWSPEC; +/*! \brief Returns how much memory is currently in use by the memory pool */ +NEDMALLOCEXTSPEC size_t nedpmalloc_footprint(nedpool *p) THROWSPEC; +/*! \brief Returns a series of guaranteed consecutive cleared memory allocations. + + independent_calloc is similar to calloc, but instead of returning a + single cleared space, it returns an array of pointers to n_elements + independent elements that can hold contents of size elem_size, each + of which starts out cleared, and can be independently freed, + realloc'ed etc. The elements are guaranteed to be adjacently + allocated (this is not guaranteed to occur with multiple callocs or + mallocs), which may also improve cache locality in some + applications. + + The "chunks" argument is optional (i.e., may be null, which is + probably the most typical usage). If it is null, the returned array + is itself dynamically allocated and should also be freed when it is + no longer needed. Otherwise, the chunks array must be of at least + n_elements in length. It is filled in with the pointers to the + chunks. + + In either case, independent_calloc returns this pointer array, or + null if the allocation failed. If n_elements is zero and "chunks" + is null, it returns a chunk representing an array with zero elements + (which should be freed if not wanted). + + Each element must be individually freed when it is no longer + needed. If you'd like to instead be able to free all at once, you + should instead use regular calloc and assign pointers into this + space to represent elements. (In this case though, you cannot + independently free elements.) + + independent_calloc simplifies and speeds up implementations of many + kinds of pools. It may also be useful when constructing large data + structures that initially have a fixed number of fixed-sized nodes, + but the number is not known at compile time, and some of the nodes + may later need to be freed. For example: + + struct Node { int item; struct Node* next; }; + + struct Node* build_list() { + struct Node** pool; + int n = read_number_of_nodes_needed(); + if (n <= 0) return 0; + pool = (struct Node**)(independent_calloc(n, sizeof(struct Node), 0); + if (pool == 0) die(); + // organize into a linked list... + struct Node* first = pool[0]; + for (i = 0; i < n-1; ++i) + pool[i]->next = pool[i+1]; + free(pool); // Can now free the array (or not, if it is needed later) + return first; + } +*/ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void **nedpindependent_calloc(nedpool *p, size_t elemsno, size_t elemsize, void **chunks) THROWSPEC; +/*! \brief Returns a series of guaranteed consecutive allocations. + + independent_comalloc allocates, all at once, a set of n_elements + chunks with sizes indicated in the "sizes" array. It returns + an array of pointers to these elements, each of which can be + independently freed, realloc'ed etc. The elements are guaranteed to + be adjacently allocated (this is not guaranteed to occur with + multiple callocs or mallocs), which may also improve cache locality + in some applications. + + The "chunks" argument is optional (i.e., may be null). If it is null + the returned array is itself dynamically allocated and should also + be freed when it is no longer needed. Otherwise, the chunks array + must be of at least n_elements in length. It is filled in with the + pointers to the chunks. + + In either case, independent_comalloc returns this pointer array, or + null if the allocation failed. If n_elements is zero and chunks is + null, it returns a chunk representing an array with zero elements + (which should be freed if not wanted). + + Each element must be individually freed when it is no longer + needed. If you'd like to instead be able to free all at once, you + should instead use a single regular malloc, and assign pointers at + particular offsets in the aggregate space. (In this case though, you + cannot independently free elements.) + + independent_comallac differs from independent_calloc in that each + element may have a different size, and also that it does not + automatically clear elements. + + independent_comalloc can be used to speed up allocation in cases + where several structs or objects must always be allocated at the + same time. For example: + + struct Head { ... } + struct Foot { ... } + + void send_message(char* msg) { + int msglen = strlen(msg); + size_t sizes[3] = { sizeof(struct Head), msglen, sizeof(struct Foot) }; + void* chunks[3]; + if (independent_comalloc(3, sizes, chunks) == 0) + die(); + struct Head* head = (struct Head*)(chunks[0]); + char* body = (char*)(chunks[1]); + struct Foot* foot = (struct Foot*)(chunks[2]); + // ... + } + + In general though, independent_comalloc is worth using only for + larger values of n_elements. For small values, you probably won't + detect enough difference from series of malloc calls to bother. + + Overuse of independent_comalloc can increase overall memory usage, + since it cannot reuse existing noncontiguous small chunks that + might be available for some of the elements. +*/ +NEDMALLOCEXTSPEC NEDMALLOCNOALIASATTR NEDMALLOCPTRATTR void **nedpindependent_comalloc(nedpool *p, size_t elems, size_t *sizes, void **chunks) THROWSPEC; + +#if defined(__cplusplus) +} /* namespace or extern "C" */ +#include <new> +#include <memory> +#ifdef HAVE_CPP0XTYPETRAITS +#include <type_traits> +#endif + +// Touch into existence for future platforms +namespace std { namespace tr1 { } } + +/*! \defgroup C++ C++ language support + +Thanks to the generous support of Applied Research Associates (USA), nedalloc has extensive +C++ language support which uses C++ metaprogramming techniques to provide a policy driven +STL container reimplementor. The metaprogramming silently overrides or replaces the STL implementation +on your system (MSVC and GCC are the two currently supported) to \b substantially improve +the performance of STL containers by making use of nedalloc's additional features. + +Sounds difficult to use? Not really. Simply do this: +\code +using namespace nedalloc; +typedef nedallocatorise<std::vector, unsigned int, + nedpolicy::typeIsPOD<true>::policy, + nedpolicy::mmap<>::policy, + nedpolicy::reserveN<26>::policy // 1<<26 = 64Mb. 10,000,000 * sizeof(unsigned int) = 38Mb. +>::value myvectortype; +myvectortype a; +for(int n=0; n<10000000; n++) + a.push_back(n); +\endcode + +The metaprogramming requires a new C++ compiler (> year 2008), and it will readily make use +of a C++0x compiler where it will use rvalue referencing, variadic templates, type traits and more. +Visual Studio 2008 or later is sufficent, as is GCC v4.4 or later. + +nedalloc's metaprogramming is designed to be extensible, so the rest of this page is intended for those +wishing to customise the metaprogramming. If you simply wish to know how to use the +nedalloc::nedallocator STL allocator or the nedalloc::nedallocatorise STL reimplementor, please refer +to test.cpp which gives several examples of usage. + +<h2>Extending the metaprogramming:</h2> +A nedallocator policy looks as follows: +\code +namespace nedpolicy { + template<size_t size, size_t alignment> struct sizedalign + { + template<class Base> class policy : public Base + { + template<class implementation> friend class nedallocatorI::baseimplementation; + protected: + size_t policy_alignment(size_t bytes) const + { + return (bytes < size) ? alignment : 0; + } + }; + }; +} +\endcode +The policy above implements a size based alignment, so if the block being allocated is +less than \em size then it causes \em alignment to be used, otherwise it does not align. +The sizedalign struct is merely a template parameter encapsulator used to capture +additional parameters, so the real policy is in fact the class policy held within in. +If you did not need to specify any additional parameters e.g. if you were defining +policy_nedpool(), then you would directly define a policy returning your nedpool and pass +it directly to nedallocator<>. + +The primary policy functions which are intended to be overridden are listed in +nedalloc::nedallocatorI::baseimplementation in nedmalloc.h and are prefixed by "policy_". +However, there is absolutely no reason why the meatier functions such as +nedalloc::nedallocatorI::baseimplementation::allocate() cannot be overriden, and indeed +some of the policies defined in nedmalloc.h do just that. + +Policy composition is handled by a dedicated recursive variadic template called +nedalloc::nedallocatorI::policycompositor. If you have \em really specialised needs, you +can partially specialise this class to make it do all sorts of interesting things - hence +its separation into its own class. +*/ + +/*! \brief The nedalloc namespace */ +namespace nedalloc { + +/*! \def NEDSTATIC_ASSERT(expr, msg) +\brief Generates a static assertion if (expr)==0 at compile time. + +Make SURE your message contains no spaces or anything else which would make it an invalid +variable name. +*/ +#ifndef HAVE_CPP0XSTATICASSERT +template<bool> struct StaticAssert; +template<> struct StaticAssert<true> +{ + StaticAssert() { } +}; +#define NEDSTATIC_ASSERT(expr, msg) \ + nedalloc::StaticAssert<(expr)!=0> ERROR_##msg +#else +#define NEDSTATIC_ASSERT(expr, msg) static_assert((expr)!=0, #msg ) +#endif + +/*! \brief The policy namespace in which all nedallocator policies live. */ +namespace nedpolicy { + /*! \class empty + \ingroup C++ + \brief An empty policy which does nothing. + */ + template<class Base> class empty : public Base + { + }; +} + +/*! \brief The implementation namespace where the internals live. */ +namespace nedallocatorI +{ + using namespace std; + using namespace tr1; + + /* Roll on variadic templates is all I can say! */ +#ifdef HAVE_CPP0XVARIADICTEMPLATES + template<class Impl, template<class> class... policies> class policycompositor; + template<class Impl, template<class> class A, template<class> class... policies> class policycompositor<Impl, A, policies...> + { + typedef policycompositor<Impl, policies...> temp; + public: + typedef A<typename temp::value> value; + }; +#else + template<class Impl, + template<class> class A=nedpolicy::empty, + template<class> class B=nedpolicy::empty, + template<class> class C=nedpolicy::empty, + template<class> class D=nedpolicy::empty, + template<class> class E=nedpolicy::empty, + template<class> class F=nedpolicy::empty, + template<class> class G=nedpolicy::empty, + template<class> class H=nedpolicy::empty, + template<class> class I=nedpolicy::empty, + template<class> class J=nedpolicy::empty, + template<class> class K=nedpolicy::empty, + template<class> class L=nedpolicy::empty, + template<class> class M=nedpolicy::empty, + template<class> class N=nedpolicy::empty, + template<class> class O=nedpolicy::empty + > class policycompositor + { + typedef policycompositor<Impl, B, C, D, E, F, G, H, I, J, K, L, M, N, O> temp; + public: + typedef A<typename temp::value> value; + }; +#endif + template<class Impl> class policycompositor<Impl> + { + public: + typedef Impl value; + }; +} + +template<typename T, +#ifdef HAVE_CPP0XVARIADICTEMPLATES + template<class> class... policies +#else + template<class> class policy1=nedpolicy::empty, + template<class> class policy2=nedpolicy::empty, + template<class> class policy3=nedpolicy::empty, + template<class> class policy4=nedpolicy::empty, + template<class> class policy5=nedpolicy::empty, + template<class> class policy6=nedpolicy::empty, + template<class> class policy7=nedpolicy::empty, + template<class> class policy8=nedpolicy::empty, + template<class> class policy9=nedpolicy::empty, + template<class> class policy10=nedpolicy::empty, + template<class> class policy11=nedpolicy::empty, + template<class> class policy12=nedpolicy::empty, + template<class> class policy13=nedpolicy::empty, + template<class> class policy14=nedpolicy::empty, + template<class> class policy15=nedpolicy::empty +#endif +> class nedallocator; + +namespace nedallocatorI +{ + /*! \brief The base implementation class */ + template<class implementation> class baseimplementation + { + //NEDSTATIC_ASSERT(false, Bad_policies_specified); + }; + /*! \brief The base implementation class */ + template<typename T, +#ifdef HAVE_CPP0XVARIADICTEMPLATES + template<class> class... policies +#else + template<class> class policy1, + template<class> class policy2, + template<class> class policy3, + template<class> class policy4, + template<class> class policy5, + template<class> class policy6, + template<class> class policy7, + template<class> class policy8, + template<class> class policy9, + template<class> class policy10, + template<class> class policy11, + template<class> class policy12, + template<class> class policy13, + template<class> class policy14, + template<class> class policy15 +#endif + > class baseimplementation<nedallocator<T, +#ifdef HAVE_CPP0XVARIADICTEMPLATES +policies... +#else + policy1, policy2, policy3, policy4, policy5, + policy6, policy7, policy8, policy9, policy10, + policy11, policy12, policy13, policy14, policy15 +#endif + > > + { + protected: + //! \brief The most derived nedallocator implementation type + typedef nedallocator<T, +#ifdef HAVE_CPP0XVARIADICTEMPLATES + policies... +#else + policy1, policy2, policy3, policy4, policy5, + policy6, policy7, policy8, policy9, policy10, + policy11, policy12, policy13, policy14, policy15 +#endif + > implementationType; + //! \brief Returns a this for the most derived nedallocator implementation type + implementationType *_this() { return static_cast<implementationType *>(this); } + //! \brief Returns a this for the most derived nedallocator implementation type + const implementationType *_this() const { return static_cast<const implementationType *>(this); } + //! \brief Specifies the nedpool to use. Defaults to zero (the system pool). + nedpool *policy_nedpool(size_t bytes) const + { + return 0; + } + //! \brief Specifies the granularity to use. Defaults to \em bytes (no granularity). + size_t policy_granularity(size_t bytes) const + { + return bytes; + } + //! \brief Specifies the alignment to use. Defaults to zero (no alignment). + size_t policy_alignment(size_t bytes) const + { + return 0; + } + //! \brief Specifies the flags to use. Defaults to zero (no flags). + unsigned policy_flags(size_t bytes) const + { + return 0; + } + //! \brief Specifies what to do when the allocation fails. Defaults to throwing std::bad_alloc. + void policy_throwbadalloc(size_t bytes) const + { + throw std::bad_alloc(); + } + //! \brief Specifies if the type is POD. Is std::is_trivially_copyable<T>::value on C++0x compilers, otherwise false. + static const bool policy_typeIsPOD= +#ifdef HAVE_CPP0XTYPETRAITS +#if defined(__GNUC__) && (__GNUC__ * 10000 + __GNUC_MINOR__ * 100 + __GNUC_PATCHLEVEL__) < 40900 + is_pod<T>::value; +#else + is_trivially_copyable<T>::value; +#endif +#else + false; +#endif + public: + typedef T *pointer; + typedef const T *const_pointer; + typedef T &reference; + typedef const T &const_reference; + typedef T value_type; + typedef size_t size_type; + typedef ptrdiff_t difference_type; + T *address(T &r) const { return &r; } + const T *address(const T &s) const { return &s; } + size_t max_size() const { return (static_cast<size_t>(0) - static_cast<size_t>(1)) / sizeof(T); } + bool operator!=(const baseimplementation &other) const { return !(*this == other); } + bool operator==(const baseimplementation &other) const { return true; } + + void construct(T *const p, const T &t) const { + void *const _p = static_cast<void *>(p); + new (_p) T(t); + } + void destroy(T *const p) const { + p->~T(); + } + baseimplementation() { } + baseimplementation(const baseimplementation &) { } +#ifdef HAVE_CPP0XRVALUEREFS + baseimplementation(baseimplementation &&) { } +#endif + template<typename U> struct rebind { + typedef nedallocator<U, +#ifdef HAVE_CPP0XVARIADICTEMPLATES + policies... +#else + policy1, policy2, policy3, policy4, policy5, + policy6, policy7, policy8, policy9, policy10, + policy11, policy12, policy13, policy14, policy15 +#endif + > other; + }; + template<typename U> baseimplementation(const nedallocator<U, +#ifdef HAVE_CPP0XVARIADICTEMPLATES + policies... +#else + policy1, policy2, policy3, policy4, policy5, + policy6, policy7, policy8, policy9, policy10, + policy11, policy12, policy13, policy14, policy15 +#endif + > &) { } + + T *allocate(const size_t n) const { + // Leave these spelled out to aid debugging + const size_t t_size = sizeof(T); + size_t size = _this()->policy_granularity(n*t_size); + nedpool *pool = _this()->policy_nedpool(size); + size_t alignment = _this()->policy_alignment(size); + unsigned flags = _this()->policy_flags(size); + void *ptr = nedpmalloc2(pool, size, alignment, flags); + if(!ptr) + _this()->policy_throwbadalloc(size); + return static_cast<T *>(ptr); + } + void deallocate(T *p, const size_t n) const { + nedpfree(0/*not needed*/, p); + } + template<typename U> T *allocate(const size_t n, const U * /* hint */) const { + return allocate(n); + } + private: + baseimplementation &operator=(const baseimplementation &); + }; + +} + +namespace nedpolicy +{ + /*! \class granulate + \ingroup C++ + \brief A policy setting the granularity of the allocated memory. + + Memory is sized according to (size+granularity-1) & ~(granularity-1). + In other words, granularity \b must be a power of two. + */ + template<size_t granularity> struct granulate + { + template<class Base> class policy : public Base + { + template<class implementation> friend class nedallocatorI::baseimplementation; + protected: + size_t policy_granularity(size_t bytes) const + { + return (bytes+granularity-1) & ~(granularity-1); + } + }; + }; + /*! \class align + \ingroup C++ + \brief A policy setting the alignment of the allocated memory. + */ + template<size_t alignment> struct align + { + template<class Base> class policy : public Base + { + template<class implementation> friend class nedallocatorI::baseimplementation; + protected: + size_t policy_alignment(size_t bytes) const + { + return alignment; + } + }; + }; + /*! \class zero + \ingroup C++ + \brief A policy causing the zeroing of the allocated memory. + */ + template<bool dozero=true> struct zero + { + template<class Base> class policy : public Base + { + template<class implementation> friend class nedallocatorI::baseimplementation; + protected: + unsigned policy_flags(size_t bytes) const + { + return dozero ? Base::policy_flags(bytes)|M2_ZERO_MEMORY : Base::policy_flags(bytes); + } + }; + }; + /*! \class preventmove + \ingroup C++ + \brief A policy preventing the moving of the allocated memory. + */ + template<bool doprevent=true> struct preventmove + { + template<class Base> class policy : public Base + { + template<class implementation> friend class nedallocatorI::baseimplementation; + protected: + unsigned policy_flags(size_t bytes) const + { + return doprevent ? Base::policy_flags(bytes)|M2_PREVENT_MOVE : Base::policy_flags(bytes); + } + }; + }; + /*! \class mmap + \ingroup C++ + \brief A policy causing the mmapping of the allocated memory. + */ + template<bool dommap=true> struct mmap + { + template<class Base> class policy : public Base + { + template<class implementation> friend class nedallocatorI::baseimplementation; + protected: + unsigned policy_flags(size_t bytes) const + { + return dommap ? Base::policy_flags(bytes)|M2_ALWAYS_MMAP : Base::policy_flags(bytes); + } + }; + }; + /*! \class reserveX + \ingroup C++ + \brief A policy causing the address reservation of X times the allocated memory. + */ + template<size_t X> struct reserveX + { + template<class Base> class policy : public Base + { + template<class implementation> friend class nedallocatorI::baseimplementation; + protected: + unsigned policy_flags(size_t bytes) const + { + return Base::policy_flags(bytes)|M2_RESERVE_MULT(X); + } + }; + }; + /*! \class reserveN + \ingroup C++ + \brief A policy causing the address reservation of (1<<N) bytes of memory. + */ + template<size_t N> struct reserveN + { + template<class Base> class policy : public Base + { + template<class implementation> friend class nedallocatorI::baseimplementation; + protected: + unsigned policy_flags(size_t bytes) const + { + return Base::policy_flags(bytes)|M2_RESERVE_SHIFT(N); + } + }; + }; + /*! \class badalloc + \ingroup C++ + \brief A policy specifying what to throw when an allocation failure occurs. + + A type specialisation exists for badalloc<void> which is equivalent to new(nothrow) + i.e. return zero and don't throw anything. + */ + template<typename T> struct badalloc + { + template<class Base> class policy : public Base + { + template<class implementation> friend class nedallocatorI::baseimplementation; + protected: + void policy_throwbadalloc(size_t bytes) const + { + throw T(); + } + }; + }; + template<> struct badalloc<void> + { + template<class Base> class policy : public Base + { + template<class implementation> friend class nedallocatorI::baseimplementation; + protected: + void policy_throwbadalloc(size_t bytes) const + { + } + }; + }; + /*! \class typeIsPOD + \ingroup C++ + \brief A policy forcing the treatment of the type as Plain Old Data (POD) + + On C++0x compilers, the <type_traits> is_trivially_copyable<type>::value is used by default. + When treated as POD, memcpy() is used instead + of copy construction and realloc() is permitted to move the memory contents when + resizing. + */ + template<bool ispod> struct typeIsPOD + { + template<class Base> class policy : public Base + { + template<class implementation> friend class nedallocatorI::baseimplementation; + protected: + static const bool policy_typeIsPOD=ispod; + }; + }; +} + +/*! \class nedallocator +\ingroup C++ +\brief A policy driven STL allocator which uses nedmalloc + +One of the lesser known features of STL container classes is their ability to take +an allocator implementation class, so where you had std::vector<Foo> you can now +have std::vector<Foo, nedalloc::nedallocator< std::vector<Foo> > such that +std::vector<> will now use nedalloc as the policy specifies. + +You <b>almost certainly</b> don't want to use this directly except in the naive +case. See nedalloc::nedallocatorise to see what I mean. +*/ +template<typename T, +#ifdef HAVE_CPP0XVARIADICTEMPLATES + template<class> class... policies +#else + template<class> class policy1, + template<class> class policy2, + template<class> class policy3, + template<class> class policy4, + template<class> class policy5, + template<class> class policy6, + template<class> class policy7, + template<class> class policy8, + template<class> class policy9, + template<class> class policy10, + template<class> class policy11, + template<class> class policy12, + template<class> class policy13, + template<class> class policy14, + template<class> class policy15 +#endif +> class nedallocator : public nedallocatorI::policycompositor< +#ifdef HAVE_CPP0XVARIADICTEMPLATES + nedallocatorI::baseimplementation<nedallocator<T, policies...> >, + policies... +#else + nedallocatorI::baseimplementation<nedallocator<T, + policy1, policy2, policy3, policy4, policy5, + policy6, policy7, policy8, policy9, policy10, + policy11, policy12, policy13, policy14, policy15 + > >, + policy1, policy2, policy3, policy4, policy5, + policy6, policy7, policy8, policy9, policy10, + policy11, policy12, policy13, policy14, policy15 +#endif +>::value +{ + typedef typename nedallocatorI::policycompositor< +#ifdef HAVE_CPP0XVARIADICTEMPLATES + nedallocatorI::baseimplementation<nedallocator<T, policies...> >, + policies... +#else + nedallocatorI::baseimplementation<nedallocator<T, + policy1, policy2, policy3, policy4, policy5, + policy6, policy7, policy8, policy9, policy10, + policy11, policy12, policy13, policy14, policy15 + > >, + policy1, policy2, policy3, policy4, policy5, + policy6, policy7, policy8, policy9, policy10, + policy11, policy12, policy13, policy14, policy15 +#endif + >::value Base; +public: + nedallocator() { } + nedallocator(const nedallocator &o) : Base(o) { } +#ifdef HAVE_CPP0XRVALUEREFS + nedallocator(nedallocator &&o) : Base(std::move(o)) { } +#endif + /* This templated constructor and rebind() are used by MSVC's secure iterator checker. + I think it's best to not copy state even though it may break policies which store data. */ + template<typename U> nedallocator(const nedallocator<U, +#ifdef HAVE_CPP0XVARIADICTEMPLATES + policies... +#else + policy1, policy2, policy3, policy4, policy5, + policy6, policy7, policy8, policy9, policy10, + policy11, policy12, policy13, policy14, policy15 +#endif + > &o) { } +#ifdef HAVE_CPP0XRVALUEREFS + template<typename U> nedallocator(nedallocator<U, +#ifdef HAVE_CPP0XVARIADICTEMPLATES + policies... +#else + policy1, policy2, policy3, policy4, policy5, + policy6, policy7, policy8, policy9, policy10, + policy11, policy12, policy13, policy14, policy15 +#endif + > &&o) { } +#endif + + template<typename U> struct rebind { + typedef nedallocator<U, +#ifdef HAVE_CPP0XVARIADICTEMPLATES + policies... +#else + policy1, policy2, policy3, policy4, policy5, + policy6, policy7, policy8, policy9, policy10, + policy11, policy12, policy13, policy14, policy15 +#endif + > other; + }; +}; + +namespace nedallocatorI { + // Holds a static allocator instance shared by anything allocating from allocator + template<class allocator> struct StaticAllocator + { + static allocator &get() + { + static allocator a; + return a; + } + }; + // RAII holder for a Newed object + template<typename T, class allocator> struct PtrHolder + { + T *mem; + PtrHolder(T *_mem) : mem(_mem) { } + ~PtrHolder() + { + if(mem) + { + allocator &a=nedallocatorI::StaticAllocator<allocator>::get(); + a.deallocate(mem, sizeof(T)); + mem=0; + } + } + T *release() { T *ret=mem; mem=0; return ret; } + T *operator *() { return mem; } + const T *operator *() const { return mem; } + }; +} +/*! \brief Allocates the memory for an instance of object \em T and constructs it. + +If an exception is thrown during construction, the memory is freed before +rethrowing the exception. + +Usage is very simple: +\code + SSEVectorType *foo1=New<SSEVectorType>(4, 5, 6, 7); +\endcode +*/ +#ifdef HAVE_CPP0XVARIADICTEMPLATES +template<typename T, class allocator=nedallocator<T>, typename... Parameters> inline T *New(const Parameters&... parameters) +#else +template<typename T, class allocator> inline T *New() +#endif +{ + allocator &a=nedallocatorI::StaticAllocator<allocator>::get(); + nedallocatorI::PtrHolder<T, allocator> ret(a.allocate(sizeof(T))); + if(*ret) + { +#ifdef HAVE_CPP0XVARIADICTEMPLATES + new((void *) *ret) T(parameters...); +#else + new((void *) *ret) T; +#endif + } + return ret.release(); +} +#ifndef HAVE_CPP0XVARIADICTEMPLATES +// Extremely annoying not to have default template arguments for functions pre-C++0x +template<typename T> inline T *New() +{ + return New<T, nedallocator<T> >(); +} +// Also, it's painful to replicate function overloads :( +#define NEDMALLOC_NEWIMPL \ +template<typename T, class allocator, NEDMALLOC_NEWIMPLTYPES> inline T *New(NEDMALLOC_NEWIMPLPARSDEFS) \ +{ \ + allocator &a=nedallocatorI::StaticAllocator<allocator>::get(); \ + nedallocatorI::PtrHolder<T, allocator> ret(a.allocate(sizeof(T))); \ + if(*ret) \ + { \ + new((void *) *ret) T(NEDMALLOC_NEWIMPLPARS); \ + } \ + return ret.release(); \ +} \ +template<typename T, NEDMALLOC_NEWIMPLTYPES> inline T *New(NEDMALLOC_NEWIMPLPARSDEFS)\ +{ \ + return New<T, nedallocator<T> >(NEDMALLOC_NEWIMPLPARS); \ +} +#define NEDMALLOC_NEWIMPLTYPES typename P1 +#define NEDMALLOC_NEWIMPLPARSDEFS const P1 &p1 +#define NEDMALLOC_NEWIMPLPARS p1 +NEDMALLOC_NEWIMPL +#undef NEDMALLOC_NEWIMPLTYPES +#undef NEDMALLOC_NEWIMPLPARSDEFS +#undef NEDMALLOC_NEWIMPLPARS + +#define NEDMALLOC_NEWIMPLTYPES typename P1, typename P2 +#define NEDMALLOC_NEWIMPLPARSDEFS const P1 &p1, const P2 &p2 +#define NEDMALLOC_NEWIMPLPARS p1, p2 +NEDMALLOC_NEWIMPL +#undef NEDMALLOC_NEWIMPLTYPES +#undef NEDMALLOC_NEWIMPLPARSDEFS +#undef NEDMALLOC_NEWIMPLPARS + +#define NEDMALLOC_NEWIMPLTYPES typename P1, typename P2, typename P3 +#define NEDMALLOC_NEWIMPLPARSDEFS const P1 &p1, const P2 &p2, const P3 &p3 +#define NEDMALLOC_NEWIMPLPARS p1, p2, p3 +NEDMALLOC_NEWIMPL +#undef NEDMALLOC_NEWIMPLTYPES +#undef NEDMALLOC_NEWIMPLPARSDEFS +#undef NEDMALLOC_NEWIMPLPARS + +#define NEDMALLOC_NEWIMPLTYPES typename P1, typename P2, typename P3, typename P4 +#define NEDMALLOC_NEWIMPLPARSDEFS const P1 &p1, const P2 &p2, const P3 &p3, const P4 &p4 +#define NEDMALLOC_NEWIMPLPARS p1, p2, p3, p4 +NEDMALLOC_NEWIMPL +#undef NEDMALLOC_NEWIMPLTYPES +#undef NEDMALLOC_NEWIMPLPARSDEFS +#undef NEDMALLOC_NEWIMPLPARS + +#define NEDMALLOC_NEWIMPLTYPES typename P1, typename P2, typename P3, typename P4, typename P5 +#define NEDMALLOC_NEWIMPLPARSDEFS const P1 &p1, const P2 &p2, const P3 &p3, const P4 &p4, const P5 &p5 +#define NEDMALLOC_NEWIMPLPARS p1, p2, p3, p4, p5 +NEDMALLOC_NEWIMPL +#undef NEDMALLOC_NEWIMPLTYPES +#undef NEDMALLOC_NEWIMPLPARSDEFS +#undef NEDMALLOC_NEWIMPLPARS + +#undef NEDMALLOC_NEWIMPL +#endif + +/*! \brief Destructs an instance of object T, and releases the memory used to store it. +*/ +template<class allocator, typename T> inline void Delete(const T *_obj) +{ + T *obj=const_cast<T *>(_obj); + allocator &a=nedallocatorI::StaticAllocator<allocator>::get(); + obj->~T(); + a.deallocate(obj, sizeof(T)); +} +template<typename T> inline void Delete(const T *obj) { Delete<nedallocator<T> >(obj); } + +/*! \class nedallocatorise +\ingroup C++ +\brief Reimplements a given STL container to make full and efficient usage of nedalloc +\param stlcontainer The STL container you wish to reimplement +\param T The type to be contained +\param policies... Any policies you want applied to the allocator + + +This is a clever bit of C++ metaprogramming if I do say so myself! What it does +is to specialise a STL container implementation to make full use of nedalloc's +advanced facilities, so for example if you do: +\code +using namespace nedalloc; +typedef nedallocatorise<std::vector, unsigned int, + nedpolicy::typeIsPOD<true>::policy, + nedpolicy::mmap<>::policy, + nedpolicy::reserveN<26>::policy // 1<<26 = 64Mb. 10,000,000 * sizeof(unsigned int) = 38Mb. +>::value myvectortype; +myvectortype a; +for(int n=0; n<10000000; n++) + a.push_back(n); +\endcode +What happens here is that nedallocatorise reimplements the parts of +std::vector which extend and shrink the actual memory allocation. +Because the typeIsPOD policy is specified, it means that realloc() +rather than realloc(M2_PREVENT_MOVE) can be used. Also, because the +mmap and the reserveN policies are specified, std::vector immediately +reserves 64Mb of address space and forces the immediate use of mmap(). +This allows you to push_back() a lot of data very, very quickly indeed. +You will also find that pop_back() actually reduces the allocation now +(most implementations don't bother ever releasing memory except when +reaching empty or when resize() is called). When mmapped, reserve() +is automatically held at a minimum of <page size>/sizeof(type) though +larger values are respected. + +test.cpp has a benchmark of the speed differences you may realise, plus +an example of usage. +*/ +template<template<typename, class> class stlcontainer, + typename T, +#ifdef HAVE_CPP0XVARIADICTEMPLATES + template<class> class... policies +#else + template<class> class policy1=nedpolicy::empty, + template<class> class policy2=nedpolicy::empty, + template<class> class policy3=nedpolicy::empty, + template<class> class policy4=nedpolicy::empty, + template<class> class policy5=nedpolicy::empty, + template<class> class policy6=nedpolicy::empty, + template<class> class policy7=nedpolicy::empty, + template<class> class policy8=nedpolicy::empty, + template<class> class policy9=nedpolicy::empty, + template<class> class policy10=nedpolicy::empty, + template<class> class policy11=nedpolicy::empty, + template<class> class policy12=nedpolicy::empty, + template<class> class policy13=nedpolicy::empty, + template<class> class policy14=nedpolicy::empty, + template<class> class policy15=nedpolicy::empty +#endif +> class nedallocatorise +{ +public: + //! The reimplemented STL container type + typedef stlcontainer<T, nedallocator<T, +#ifdef HAVE_CPP0XVARIADICTEMPLATES + policies... +#else + policy1, policy2, policy3, policy4, policy5, + policy6, policy7, policy8, policy9, policy10, + policy11, policy12, policy13, policy14, policy15 +#endif + > > value; +}; + +} /* namespace */ +#endif + +/* Some miscellaneous dlmalloc option documentation */ + +#ifdef DOXYGEN_IS_PARSING_ME +/* Just some false defines to keep doxygen happy */ + +#define NEDMALLOC_DEBUG DEBUG +#define ENABLE_LARGE_PAGES undef +#define ENABLE_FAST_HEAP_DETECTION undef +#define REPLACE_SYSTEM_ALLOCATOR undef +#define ENABLE_TOLERANT_NEDMALLOC undef +#define NO_NED_NAMESPACE undef + +/*! \def MALLOC_ALIGNMENT +\brief Defines what alignment normally returned blocks should use. Is 16 bytes on Mac OS X, otherwise 8 bytes. */ +#define MALLOC_ALIGNMENT 8 + +/*! \def USE_LOCKS +\brief Defines the threadsafety of nedalloc + +USE_LOCKS can be 2 if you want to define your own MLOCK_T, INITIAL_LOCK, +ACQUIRE_LOCK, RELEASE_LOCK, TRY_LOCK, IS_LOCKED and NULL_LOCK_INITIALIZER. +*/ +#define USE_LOCKS 1 + +/*! \def DEFAULT_GRANULARITY +\brief Defines the granularity in which to request or free system memory. +*/ +#define DEFAULT_GRANULARITY (2*1024*1024) + +/*! \def DEFAULT_TRIM_THRESHOLD +\brief Defines how much memory must be free before returning it to the system. +*/ +#define DEFAULT_TRIM_THRESHOLD (2*1024*1024) + +/*! \def DEFAULT_MMAP_THRESHOLD +\brief Defines the threshold above which mmap() is used to perform direct allocation. +*/ +#define DEFAULT_MMAP_THRESHOLD (256*1024) + +/*! \def MAX_RELEASE_CHECK_RATE +\brief Defines how many free() ops should occur before checking how much free memory there is. +*/ +#define MAX_RELEASE_CHECK_RATE 4095 + +/*! \def NEDMALLOC_FORCERESERVE +\brief Lets you force address space reservation in the \b standard malloc API + +Note that by default realloc() sets M2_RESERVE_MULT(8) when thunking to realloc2(), +so you probably don't need to override this +*/ +#define NEDMALLOC_FORCERESERVE(p, mem, size) 0 + +/*! \def NEDMALLOC_TESTLOGENTRY +\brief Used to determine whether a given memory operation should be logged. +*/ +#define NEDMALLOC_TESTLOGENTRY(tc, np, type, mspace, size, mem, alignment, flags, returned) ((type)&ENABLE_LOGGING) + +/*! \def NEDMALLOC_STACKBACKTRACEDEPTH +\brief Turns on stack backtracing in the logger. + +You almost certainly want to constrain what gets logged using NEDMALLOC_TESTLOGENTRY +if you turn this on as the sheer volume of data output can make execution very slow. +*/ +#define NEDMALLOC_STACKBACKTRACEDEPTH 0 + +#endif + +#endif |