The R Project SVN – R – Diff – /trunk/src/gnuwin32/malloc.c

-/* 27/03/2000 win32-api needs this for ANSI compliance */
-#define NONAMELESSUNION
-/* ---------- To make a malloc.h, start cutting here ------------ */
 /*
-  A version of malloc/free/realloc written by Doug Lea and released to the
+  This is a version (aka dlmalloc) of malloc/free/realloc written by
+  Doug Lea and released to the public domain.  Use, modify, and
+  redistribute this code without permission or acknowledgement in any
-  public domain.  Send questions/comments/complaints/performance data
+  way you wish.  Send questions, comments, complaints, performance
-  to dl@cs.oswego.edu
+  data, etc to dl@cs.oswego.edu
-* VERSION 2.6.6  Sun Mar  5 19:10:03 2000  Doug Lea  (dl at gee)
+* VERSION 2.7.2 Sat Aug 17 09:07:30 2002  Doug Lea  (dl at gee)
    Note: There may be an updated version of this malloc obtainable at
-           ftp://g.oswego.edu/pub/misc/malloc.c
+           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 (-O), and link it into another program. All
+  of the compile-time options default to reasonable values for use on
+  most unix platforms. Compile -DWIN32 for reasonable defaults on windows.
+  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-2.7.1.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.
 * Why use this malloc?
   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 a high-level description, see
+  allocator for malloc-intensive programs.
-     http://g.oswego.edu/dl/html/malloc.html
-* Synopsis of public routines
+  The main properties of the algorithms are:
-  (Much fuller descriptions are contained in the program documentation below.)
+  * For large (>= 512 bytes) requests, it is a pure best-fit allocator,
-  malloc(size_t n);
-     Return a pointer to a newly allocated chunk of at least n bytes, or null
+    with ties normally decided via FIFO (i.e. least recently used).
-     if no space is available.
+  * For small (<= 64 bytes by default) requests, it is a caching
-  free(Void_t* p);
-     Release the chunk of memory pointed to by p, or no effect if p is null.
+    allocator, that maintains pools of quickly recycled chunks.
-  realloc(Void_t* p, size_t n);
+  * In between, and for combinations of large and small requests, it does
-     Return a pointer to a chunk of size n that contains the same data
+    the best it can trying to meet both goals at once.
-     as does chunk p up to the minimum of (n, p's size) bytes, or null
+  * For very large requests (>= 128KB by default), it relies on system
-     if no space is available. The returned pointer may or may not be
+    memory mapping facilities, if supported.
-     the same as p. If p is null, equivalent to malloc.  Unless the
+  For a longer but slightly out of date high-level description, see
-     #define REALLOC_ZERO_BYTES_FREES below is set, realloc with a
+     http://gee.cs.oswego.edu/dl/html/malloc.html
-     size argument of zero (re)allocates a minimum-sized chunk.
+  You may already by default be using a C library containing a malloc
-  memalign(size_t alignment, size_t n);
+  that is  based on some version of this malloc (for example in
-     Return a pointer to a newly allocated chunk of n bytes, aligned
+  linux). You might still want to use the one in this file in order to
-     in accord with the alignment argument, which must be a power of
+  customize settings or to avoid overheads associated with library
-     two.
+  versions.
+* Contents, described in more detail in "description of public routines" below.
+  Standard (ANSI/SVID/...)  functions:
-  valloc(size_t n);
+    malloc(size_t n);
-     Equivalent to memalign(pagesize, n), where pagesize is the page
+    calloc(size_t n_elements, size_t element_size);
+    free(Void_t* p);
-     size of the system (or as near to this as can be figured out from
+    realloc(Void_t* p, size_t n);
-     all the includes/defines below.)
+    memalign(size_t alignment, size_t n);
-  pvalloc(size_t n);
+    valloc(size_t n);
+    mallinfo()
-     Equivalent to valloc(minimum-page-that-holds(n)), that is,
+    mallopt(int parameter_number, int parameter_value)
-     round up n to nearest pagesize.
+  Additional functions:
-  calloc(size_t unit, size_t quantity);
+    independent_calloc(size_t n_elements, size_t size, Void_t* chunks[]);
-     Returns a pointer to quantity * unit bytes, with all locations
+    independent_comalloc(size_t n_elements, size_t sizes[], Void_t* chunks[]);
-     set to zero.
+    pvalloc(size_t n);
-  cfree(Void_t* p);
+    cfree(Void_t* p);
-     Equivalent to free(p).
-  malloc_trim(size_t pad);
+    malloc_trim(size_t pad);
-     Release all but pad bytes of freed top-most memory back
-     to the system. Return 1 if successful, else 0.
-  malloc_usable_size(Void_t* p);
+    malloc_usable_size(Void_t* p);
-     Report the number usable allocated bytes associated with allocated
-     chunk p. This may or may not report more bytes than were requested,
-     due to alignment and minimum size constraints.
-  malloc_stats();
+    malloc_stats();
-     Prints brief summary statistics on stderr.
-  mallinfo()
-     Returns (by copy) a struct containing various summary statistics.
-  mallopt(int parameter_number, int parameter_value)
-     Changes one of the tunable parameters described below. Returns
-if successful in changing the parameter, else 0.
 * Vital statistics:
-  Alignment:                            8-byte
-byte alignment is currently hardwired into the design.  This
-       seems to suffice for all current machines and C compilers.
-  Assumed pointer representation:       4 or 8 bytes
+  Supported pointer representation:       4 or 8 bytes
-       Code for 8-byte pointers is untested by me but has worked
-       reliably by Wolfram Gloger, who contributed most of the
-       changes supporting this.
-  Assumed size_t  representation:       4 or 8 bytes
+  Supported size_t  representation:       4 or 8 bytes
        Note that size_t is allowed to be 4 bytes even if pointers are 8.
+       You can adjust this by defining INTERNAL_SIZE_T
+  Alignment:                              2 * sizeof(size_t) (default)
+       (i.e., 8 byte alignment with 4byte size_t). This suffices for
+       nearly all current machines and C compilers. However, you can
+       define MALLOC_ALIGNMENT to be wider than this if necessary.
-  Minimum overhead per allocated chunk: 4 or 8 bytes
+  Minimum overhead per allocated chunk:   4 or 8 bytes
-       Each malloced chunk has a hidden overhead of 4 bytes holding size
+       Each malloced chunk has a hidden word of overhead holding size
        and status information.
   Minimum allocated size: 4-byte ptrs:  16 bytes    (including 4 overhead)
 -byte ptrs:  24/32 bytes (including, 4/8 overhead)
        When a chunk is freed, 12 (for 4byte ptrs) or 20 (for 8 byte
        ptrs but 4 byte size) or 24 (for 8/8) additional bytes are
-       needed; 4 (8) for a trailing size field
+       needed; 4 (8) for a trailing size field and 8 (16) bytes for
-       and 8 (16) bytes for free list pointers. Thus, the minimum
+       free list pointers. Thus, the minimum allocatable size is
-       allocatable size is 16/24/32 bytes.
+/24/32 bytes.
        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 2 *
+       sizeof(size_t) bytes plus the remainder from a system page (the
+       minimal mmap unit); typically 4096 or 8192 bytes.
-  Maximum allocated size: 4-byte size_t: 2^31 -  8 bytes
+  Maximum allocated size:  4-byte size_t: 2^32 minus about two pages
--byte size_t: 2^63 - 16 bytes
+-byte size_t: 2^64 minus about two pages
-       It is assumed that (possibly signed) size_t bit values suffice to
+       It is assumed that (possibly signed) size_t values suffice to
        represent chunk sizes. `Possibly signed' is due to the fact
        that `size_t' may be defined on a system as either a signed or
-       an unsigned type. To be conservative, values that would appear
+       an unsigned type. The ISO C standard says that it must be
+       unsigned, but a few systems are known not to adhere to this.
+       Additionally, even when size_t is unsigned, sbrk (which is by
+       default used to obtain memory from system) accepts signed
+       arguments, and may not be able to handle size_t-wide arguments
-       as negative numbers are avoided.
+       with negative sign bit.  Generally, values that would
+       appear as negative after accounting for overhead and alignment
+       are supported only via mmap(), which does not have this
+       limitation.
-       Requests for sizes with a negative sign bit when the request
+       Requests for sizes outside the allowed range will perform an optional
+       failure action and then return null. (Requests may also
-       size is treaded as a long will return null.
+       also fail because a system is out of memory.)
-  Maximum overhead wastage per allocated chunk: normally 15 bytes
+  Thread-safety: NOT thread-safe unless USE_MALLOC_LOCK defined
-       Alignnment demands, plus the minimum allocatable size restriction
+       When USE_MALLOC_LOCK is defined, wrappers are created to
-       make the normal worst-case wastage 15 bytes (i.e., up to 15
+       surround every public call with either a pthread mutex or
-       more bytes will be allocated than were requested in malloc), with
+       a win32 spinlock (depending on WIN32). This is not
-       two exceptions:
+       especially fast, and can be a major bottleneck.
-. Because requests for zero bytes allocate non-zero space,
+       It is designed only to provide minimal protection
-            the worst case wastage for a request of zero bytes is 24 bytes.
+       in concurrent environments, and to provide a basis for
-. For requests >= mmap_threshold that are serviced via
+       extensions.  If you are using malloc in a concurrent program,
-            mmap(), the worst case wastage is 8 bytes plus the remainder
+       you would be far better off obtaining ptmalloc, which is
-            from a system page (the minimal mmap unit); typically 4096 bytes.
+       derived from a version of this malloc, and is well-tuned for
+       concurrent programs. (See http://www.malloc.de) Note that
-* Limitations
+       even when USE_MALLOC_LOCK is defined, you can can guarantee
-    Here are some features that are NOT currently supported
+       full thread-safety only if no threads acquire memory through
+       direct calls to MORECORE or other system-level allocators.
-    * No user-definable hooks for callbacks and the like.
+  Compliance: I believe it is compliant with the 1997 Single Unix Specification
-    * No automated mechanism for fully checking that all accesses
+       (See http://www.opennc.org). Also SVID/XPG, ANSI C, and probably
-      to malloced memory stay within their bounds.
-    * No support for compaction.
+       others as well.
 * Synopsis of compile-time options:
     People have reported using previous versions of this malloc on all
     versions of Unix, sometimes by tweaking some of the defines
     below. It has been tested most extensively on Solaris and
     Linux. It is also reported to work on WIN32 platforms.
-    People have also reported adapting this malloc for use in
+    People also report using it in stand-alone embedded systems.
-    stand-alone embedded systems.
-    The implementation is in straight, hand-tuned ANSI C.  Among other
+    The implementation is in straight, hand-tuned ANSI C.  It is not
-    consequences, it uses a lot of macros.  Because of this, to be at
+    at all modular. (Sorry!)  It uses a lot of macros.  To be at all
-    all usable, this code should be compiled using an optimizing compiler
+    usable, this code should be compiled using an optimizing compiler
-    (for example gcc -O2) that can simplify expressions and control
+    (for example gcc -O3) that can simplify expressions and control
+    paths. (FAQ: some macros import variables as arguments rather than
+    declare locals because people reported that some debuggers
-    paths.
+    otherwise get confused.)
-  __STD_C                  (default: derived from C compiler defines)
+    OPTION                     DEFAULT VALUE
-     Nonzero if using ANSI-standard C compiler, a C++ compiler, or
-     a C compiler sufficiently close to ANSI to get away with it.
+    Compilation Environment options:
-  DEBUG                    (default: NOT defined)
+    __STD_C                    derived from C compiler defines
-     Define to enable debugging. Adds fairly extensive assertion-based
-     checking to help track down memory errors, but noticeably slows down
-     execution.
-  REALLOC_ZERO_BYTES_FREES (default: NOT defined)
+    WIN32                      NOT defined
-     Define this if you think that realloc(p, 0) should be equivalent
-     to free(p). Otherwise, since malloc returns a unique pointer for
-     malloc(0), so does realloc(p, 0).
-  HAVE_MEMCPY               (default: defined)
+    HAVE_MEMCPY                defined
-     Define if you are not otherwise using ANSI STD C, but still
-     have memcpy and memset in your C library and want to use them.
-     Otherwise, simple internal versions are supplied.
-  USE_MEMCPY               (default: 1 if HAVE_MEMCPY is defined, 0 otherwise)
+    USE_MEMCPY                 1 if HAVE_MEMCPY is defined
-     Define as 1 if you want the C library versions of memset and
-     memcpy called in realloc and calloc (otherwise macro versions are used).
-     At least on some platforms, the simple macro versions usually
-     outperform libc versions.
-  HAVE_MMAP                 (default: defined as 1)
+    HAVE_MMAP                  defined as 1
-     Define to non-zero to optionally make malloc() use mmap() to
-     allocate very large blocks.
+    MMAP_CLEARS                1
-  HAVE_MREMAP                 (default: defined as 0 unless Linux libc set)
+    HAVE_MREMAP                0 unless linux defined
-     Define to non-zero to optionally make realloc() use mremap() to
-     reallocate very large blocks.
-  malloc_getpagesize        (default: derived from system #includes)
+    malloc_getpagesize         derived from system #includes, or 4096 if not
-     Either a constant or routine call returning the system page size.
-  HAVE_USR_INCLUDE_MALLOC_H (default: NOT defined)
+    HAVE_USR_INCLUDE_MALLOC_H  NOT defined
-     Optionally define if you are on a system with a /usr/include/malloc.h
+    LACKS_UNISTD_H             NOT defined unless WIN32
-     that declares struct mallinfo. It is not at all necessary to
+    LACKS_SYS_PARAM_H          NOT defined unless WIN32
-     define this even if you do, but will ensure consistency.
+    LACKS_SYS_MMAN_H           NOT defined unless WIN32
-  INTERNAL_SIZE_T           (default: size_t)
+    LACKS_FCNTL_H              NOT defined
-     Define to a 32-bit type (probably `unsigned int') if you are on a
+    Changing default word sizes:
--bit machine, yet do not want or need to allow malloc requests of
+    INTERNAL_SIZE_T            size_t
-     greater than 2^31 to be handled. This saves space, especially for
+    MALLOC_ALIGNMENT           2 * sizeof(INTERNAL_SIZE_T)
-     very small chunks.
+    PTR_UINT                   unsigned long
-  INTERNAL_LINUX_C_LIB      (default: NOT defined)
+    CHUNK_SIZE_T               unsigned long
-     Defined only when compiled as part of Linux libc.
+    Configuration and functionality options:
-     Also note that there is some odd internal name-mangling via defines
+    USE_DL_PREFIX              NOT defined
-     (for example, internally, `malloc' is named `mALLOc') needed
+    USE_PUBLIC_MALLOC_WRAPPERS NOT defined
-     when compiling in this case. These look funny but don't otherwise
+    USE_MALLOC_LOCK            NOT defined
-     affect anything.
-  WIN32                     (default: undefined)
+    DEBUG                      NOT defined
-     Define this on MS win (95, nt) platforms to compile in sbrk emulation.
+    REALLOC_ZERO_BYTES_FREES   NOT defined
-  LACKS_UNISTD_H            (default: undefined if not WIN32)
+    MALLOC_FAILURE_ACTION      errno = ENOMEM, if __STD_C defined, else no-op
-     Define this if your system does not have a <unistd.h>.
+    TRIM_FASTBINS              0
-  LACKS_SYS_PARAM_H         (default: undefined if not WIN32)
+    FIRST_SORTED_BIN_SIZE      512
-     Define this if your system does not have a <sys/param.h>.
+    Options for customizing MORECORE:
-  MORECORE                  (default: sbrk)
+    MORECORE                   sbrk
-     The name of the routine to call to obtain more memory from the system.
-  MORECORE_FAILURE          (default: -1)
+    MORECORE_CONTIGUOUS        1
-     The value returned upon failure of MORECORE.
+    MORECORE_CANNOT_TRIM       NOT defined
-  MORECORE_CLEARS           (default 1)
+    MMAP_AS_MORECORE_SIZE      (1024 * 1024)
-     True (1) if the routine mapped to MORECORE zeroes out memory (which
+    Tuning options that are also dynamically changeable via mallopt:
-     holds for sbrk).
+    DEFAULT_MXFAST             64
-  DEFAULT_TRIM_THRESHOLD
+    DEFAULT_TRIM_THRESHOLD     256 * 1024
-  DEFAULT_TOP_PAD
+    DEFAULT_TOP_PAD            0
-  DEFAULT_MMAP_THRESHOLD
+    DEFAULT_MMAP_THRESHOLD     256 * 1024
-  DEFAULT_MMAP_MAX
+    DEFAULT_MMAP_MAX           65536
-     Default values of tunable parameters (described in detail below)
-     controlling interaction with host system routines (sbrk, mmap, etc).
-     These values may also be changed dynamically via mallopt(). The
-     preset defaults are those that give best performance for typical
-     programs/systems.
-  USE_DL_PREFIX             (default: undefined)
-     Prefix all public routines with the string 'dl'.  Useful to
-     quickly avoid procedure declaration conflicts and linker symbol
-     conflicts with existing memory allocation routines.
+    There are several other #defined constants and macros that you
+    probably don't want to touch unless you are extending or adapting malloc.
+*/
+/*
+  WIN32 sets up defaults for MS environment and compilers.
+  Otherwise defaults are for unix.
 */
+/* #define WIN32 */
+#ifdef WIN32
+#define WIN32_LEAN_AND_MEAN
+#include <windows.h>
+/* Win32 doesn't supply or need the following headers */
+/*#define LACKS_UNISTD_H*/
+#define LACKS_SYS_PARAM_H
+#define LACKS_SYS_MMAN_H
+/* Use the supplied emulation of sbrk */
+#define MORECORE sbrk
+#define MORECORE_CONTIGUOUS 0
+#define MORECORE_FAILURE    ((void*)(-1))
+/* Use the supplied emulation of mmap and munmap */
+#define HAVE_MMAP 1
+#define MUNMAP_FAILURE  (-1)
+#define MMAP_CLEARS 1
+/* These values don't really matter in windows mmap emulation */
+#define MAP_PRIVATE 1
+#define MAP_ANONYMOUS 2
+#define PROT_READ 1
+#define PROT_WRITE 2
+/* Emulation functions defined at the end of this file */
+/* If USE_MALLOC_LOCK, use supplied critical-section-based lock functions */
+#ifdef USE_MALLOC_LOCK
-/* Preliminaries */
+static int slwait(int *sl);
+static int slrelease(int *sl);
+#endif
+static long getpagesize(void);
+static long getregionsize(void);
+static void *sbrk(long size);
+static void *mmap(void *ptr, long size, long prot, long type, long handle, long arg);
+static long munmap(void *ptr, long size);
+static void vminfo (unsigned long*free, unsigned long*reserved, unsigned long*committed);
+/* Commented our for R incompatibility
+static int cpuinfo (int whole, unsigned long*kernel, unsigned long*user);
+*/
+#endif
+/*
+  __STD_C should be nonzero if using ANSI-standard C compiler, a C++
+  compiler, or a C compiler sufficiently close to ANSI to get away
+  with it.
+*/
 #ifndef __STD_C
-#ifdef __STDC__
-#define __STD_C     1
+#if defined(__STDC__) || defined(_cplusplus)
-#else
-#if __cplusplus
 #define __STD_C     1
 #else
 #define __STD_C     0
-#endif /*__cplusplus*/
-#endif /*__STDC__*/
+#endif
 #endif /*__STD_C*/
+/*
+  Void_t* is the pointer type that malloc should say it returns
+*/
 #ifndef Void_t
 #if (__STD_C || defined(WIN32))
 #define Void_t      void
 #else
 #define Void_t      char
 #ifdef __cplusplus
 extern "C" {
 #endif
-#include <stdio.h>    /* needed for malloc_stats */
+/* define LACKS_UNISTD_H if your system does not have a <unistd.h>. */
-#include <stdlib.h>		/* for exit */
+/* #define  LACKS_UNISTD_H */
+#ifndef LACKS_UNISTD_H
+#include <unistd.h>
+#endif
-/*
+/* define LACKS_SYS_PARAM_H if your system does not have a <sys/param.h>. */
-  Compile-time options
+/* #define  LACKS_SYS_PARAM_H */
-*/
+#include <stdio.h>    /* needed for malloc_stats */
+#include <errno.h>    /* needed for optional MALLOC_FAILURE_ACTION */
 /*
-    Debugging:
+  Debugging:
-    Because freed chunks may be overwritten with link fields, this
+  Because freed chunks may be overwritten with bookkeeping fields, this
-    malloc will often die when freed memory is overwritten by user
+  malloc will often die when freed memory is overwritten by user
-    programs.  This can be very effective (albeit in an annoying way)
+  programs.  This can be very effective (albeit in an annoying way)
-    in helping track down dangling pointers.
+  in helping track down dangling pointers.
-    If you compile with -DDEBUG, a number of assertion checks are
+  If you compile with -DDEBUG, a number of assertion checks are
-    enabled that will catch more memory errors. You probably won't be
+  enabled that will catch more memory errors. You probably won't be
-    able to make much sense of the actual assertion errors, but they
+  able to make much sense of the actual assertion errors, but they
-    should help you locate incorrectly overwritten memory.  The
+  should help you locate incorrectly overwritten memory.  The
-    checking is fairly extensive, and will slow down execution
+  checking is fairly extensive, and will slow down execution
-    noticeably. Calling malloc_stats or mallinfo with DEBUG set will
+  noticeably. Calling malloc_stats or mallinfo with DEBUG set will
-    attempt to check every non-mmapped allocated and free chunk in the
+  attempt to check every non-mmapped allocated and free chunk in the
-    course of computing the summmaries. (By nature, mmapped regions
+  course of computing the summmaries. (By nature, mmapped regions
-    cannot be checked very much automatically.)
+  cannot be checked very much automatically.)
-    Setting DEBUG may also be helpful if you are trying to modify
+  Setting DEBUG may also be helpful if you are trying to modify
-    this code. The assertions in the check routines spell out in more
+  this code. The assertions in the check routines spell out in more
-    detail the assumptions and invariants underlying the algorithms.
+  detail the assumptions and invariants underlying the algorithms.
+  Setting DEBUG does NOT provide an automated mechanism for checking
+  that all accesses to malloced memory stay within their
+  bounds. However, there are several add-ons and adaptations of this
+  or other mallocs available that do this.
 */
+/* #define DEBUG 1 */
 #if DEBUG
 #include <assert.h>
 #else
 #define assert(x) ((void)0)
 #endif
+/*
+  The unsigned integer type used for comparing any two chunk sizes.
+  This should be at least as wide as size_t, but should not be signed.
+*/
+#ifndef CHUNK_SIZE_T
+#define CHUNK_SIZE_T unsigned long
+#endif
+/*
+  The unsigned integer type used to hold addresses when they are are
+  manipulated as integers. Except that it is not defined on all
+  systems, intptr_t would suffice.
+*/
+#ifndef PTR_UINT
+#define PTR_UINT unsigned long
+#endif
 /*
   INTERNAL_SIZE_T is the word-size used for internal bookkeeping
+  of chunk sizes.
+  The default version is the same as size_t.
+  While not strictly necessary, it is best to define this as an
+  unsigned type, even if size_t is a signed type. This may avoid some
+  artificial size limitations on some systems.
-  of chunk sizes. On a 64-bit machine, you can reduce malloc
+  On a 64-bit machine, you may be able to reduce malloc overhead by
-  overhead by defining INTERNAL_SIZE_T to be a 32 bit `unsigned int'
+  defining INTERNAL_SIZE_T to be a 32 bit `unsigned int' at the
-  at the expense of not being able to handle requests greater than
+  expense of not being able to handle more than 2^32 of malloced
-^31. This limitation is hardly ever a concern; you are encouraged
+  space. If this limitation is acceptable, you are encouraged to set
+  this unless you are on a platform requiring 16byte alignments. In
+  this case the alignment requirements turn out to negate any
+  potential advantages of decreasing size_t word size.
+  Implementors: Beware of the possible combinations of:
+     - INTERNAL_SIZE_T might be signed or unsigned, might be 32 or 64 bits,
+       and might be the same width as int or as long
+     - size_t might have different width and signedness as INTERNAL_SIZE_T
-  to set this. However, the default version is the same as size_t.
+     - int and long might be 32 or 64 bits, and might be the same width
+  To deal with this, most comparisons and difference computations
+  among INTERNAL_SIZE_Ts should cast them to CHUNK_SIZE_T, being
+  aware of the fact that casting an unsigned int to a wider long does
+  not sign-extend. (This also makes checking for negative numbers
+  awkward.) Some of these casts result in harmless compiler warnings
+  on some systems.
 */
 #ifndef INTERNAL_SIZE_T
 #define INTERNAL_SIZE_T size_t
 #endif
+/* The corresponding word size */
+#define SIZE_SZ                (sizeof(INTERNAL_SIZE_T))
+/*
+  MALLOC_ALIGNMENT is the minimum alignment for malloc'ed chunks.
+  It must be a power of two at least 2 * SIZE_SZ, 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.
+*/
+#ifndef MALLOC_ALIGNMENT
+#define MALLOC_ALIGNMENT       (2 * SIZE_SZ)
+#endif
+/* The corresponding bit mask value */
+#define MALLOC_ALIGN_MASK      (MALLOC_ALIGNMENT - 1)
 /*
   REALLOC_ZERO_BYTES_FREES 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).
 */
 /*   #define REALLOC_ZERO_BYTES_FREES */
+/*
+  TRIM_FASTBINS controls whether free() of a very small chunk can
+  immediately lead to trimming. Setting to true (1) can reduce memory
+  footprint, but will almost always slow down programs that use a lot
+  of small chunks.
+  Define this only if you are willing to give up some speed to more
+  aggressively reduce system-level memory footprint when releasing
+  memory in programs that use many small chunks.  You can get
+  essentially the same effect by setting MXFAST to 0, but this can
+  lead to even greater slowdowns in programs using many small chunks.
+  TRIM_FASTBINS is an in-between compile-time option, that disables
+  only those chunks bordering topmost memory from being placed in
+  fastbins.
+*/
+#ifndef TRIM_FASTBINS
+#define TRIM_FASTBINS  0
+#endif
 /*
+  USE_DL_PREFIX will prefix all public routines with the string 'dl'.
-  WIN32 causes an emulation of sbrk to be compiled in
+  This is necessary when you only want to use this malloc in one part
-  mmap-based options are not currently supported in WIN32.
+  of a program, using your regular system malloc elsewhere.
 */
-/* #define WIN32 */
+/* #define USE_DL_PREFIX */
-#ifdef WIN32
-#define MORECORE wsbrk
-#define HAVE_MMAP 0
-#define LACKS_UNISTD_H
-#define LACKS_SYS_PARAM_H
 /*
-  Include 'windows.h' to get the necessary declarations for the
+  USE_MALLOC_LOCK causes wrapper functions to surround each
-  Microsoft Visual C++ data structures and routines used in the 'sbrk'
+  callable routine with pthread mutex lock/unlock.
-  emulation.
-  Define WIN32_LEAN_AND_MEAN so that only the essential Microsoft
+  USE_MALLOC_LOCK forces USE_PUBLIC_MALLOC_WRAPPERS to be defined
-  Visual C++ header files are included.
 */
-#define WIN32_LEAN_AND_MEAN
+/* #define USE_MALLOC_LOCK */
+/*
+  If USE_PUBLIC_MALLOC_WRAPPERS is defined, every public routine is
+  actually a wrapper function that first calls MALLOC_PREACTION, then
+  calls the internal routine, and follows it with
+  MALLOC_POSTACTION. This is needed for locking, but you can also use
+  this, without USE_MALLOC_LOCK, for purposes of interception,
+  instrumentation, etc. It is a sad fact that using wrappers often
+  noticeably degrades performance of malloc-intensive programs.
+*/
-#include <windows.h>
+#ifdef USE_MALLOC_LOCK
+#define USE_PUBLIC_MALLOC_WRAPPERS
+#else
+/* #define USE_PUBLIC_MALLOC_WRAPPERS */
 #endif
+/*
+   Two-phase name translation.
+   All of the actual routines are given mangled names.
+   When wrappers are used, they become the public callable versions.
+   When DL_PREFIX is used, the callable names are prefixed.
+*/
+#ifndef USE_PUBLIC_MALLOC_WRAPPERS
+#define cALLOc      public_cALLOc
+#define fREe        public_fREe
+#define cFREe       public_cFREe
+#define mALLOc      public_mALLOc
+#define mEMALIGn    public_mEMALIGn
+#define rEALLOc     public_rEALLOc
+#define vALLOc      public_vALLOc
+#define pVALLOc     public_pVALLOc
+#define mALLINFo    public_mALLINFo
+#define mALLOPt     public_mALLOPt
+#define mTRIm       public_mTRIm
+#define mSTATs      public_mSTATs
+#define mUSABLe     public_mUSABLe
+#define iCALLOc     public_iCALLOc
+#define iCOMALLOc   public_iCOMALLOc
+#endif
+#ifdef USE_DL_PREFIX
+#define public_cALLOc    dlcalloc
+#define public_fREe      dlfree
+#define public_cFREe     dlcfree
+#define public_mALLOc    dlmalloc
+#define public_mEMALIGn  dlmemalign
+#define public_rEALLOc   dlrealloc
+#define public_vALLOc    dlvalloc
+#define public_pVALLOc   dlpvalloc
+#define public_mALLINFo  dlmallinfo
+#define public_mALLOPt   dlmallopt
+#define public_mTRIm     dlmalloc_trim
+#define public_mSTATs    dlmalloc_stats
+#define public_mUSABLe   dlmalloc_usable_size
+#define public_iCALLOc   dlindependent_calloc
+#define public_iCOMALLOc dlindependent_comalloc
+#else /* USE_DL_PREFIX */
+#define public_cALLOc    calloc
+#define public_fREe      free
+#define public_cFREe     cfree
+#define public_mALLOc    malloc
+#define public_mEMALIGn  memalign
+#define public_rEALLOc   realloc
+#define public_vALLOc    valloc
+#define public_pVALLOc   pvalloc
+#define public_mALLINFo  mallinfo
+#define public_mALLOPt   mallopt
+#define public_mTRIm     malloc_trim
+#define public_mSTATs    malloc_stats
+#define public_mUSABLe   malloc_usable_size
+#define public_iCALLOc   independent_calloc
+#define public_iCOMALLOc independent_comalloc
+#endif /* USE_DL_PREFIX */
 /*
   HAVE_MEMCPY should be defined if you are not otherwise using
   ANSI STD C, but still have memcpy and memset in your C library
   and want to use them in calloc and realloc. Otherwise simple
-  macro versions are defined here.
+  macro versions are defined below.
   USE_MEMCPY should be defined as 1 if you actually want to
   have memset and memcpy called. People report that the macro
-  versions are often enough faster than libc versions on many
+  versions are faster than libc versions on some systems.
-  systems that it is better to use them.
+  Even if USE_MEMCPY is set to 1, loops to copy/clear small chunks
+  (of <= 36 bytes) are manually unrolled in realloc and calloc.
 */
 #define HAVE_MEMCPY
 #ifndef USE_MEMCPY
 #else
 #define USE_MEMCPY 0
 #endif
 #endif
 #if (__STD_C || defined(HAVE_MEMCPY))
+#ifdef WIN32
+/* On Win32 memset and memcpy are already declared in windows.h */
+#else
 #if __STD_C
 void* memset(void*, int, size_t);
 void* memcpy(void*, const void*, size_t);
 #else
-#ifdef WIN32
-/* On Win32 platforms, 'memset()' and 'memcpy()' are already declared in
-   'windows.h' */
-#else
 Void_t* memset();
 Void_t* memcpy();
 #endif
 #endif
 #endif
+/*
+  MALLOC_FAILURE_ACTION is the action to take before "return 0" when
+  malloc fails to be able to return memory, either because memory is
+  exhausted or because of illegal arguments.
-#if USE_MEMCPY
+  By default, sets errno if running on STD_C platform, else does nothing.
+*/
-/* The following macros are only invoked with (2n+1)-multiples of
+#ifndef MALLOC_FAILURE_ACTION
+#if __STD_C
-   INTERNAL_SIZE_T units, with a positive integer n. This is exploited
+#define MALLOC_FAILURE_ACTION \
-   for fast inline execution when n is small. */
+   errno = ENOMEM;
+#else
-#define MALLOC_ZERO(charp, nbytes)                                            \
+#define MALLOC_FAILURE_ACTION
-do {                                                                          \
-  INTERNAL_SIZE_T mzsz = (nbytes);                                            \
-  if(mzsz <= 9*sizeof(mzsz)) {                                                \
-    INTERNAL_SIZE_T* mz = (INTERNAL_SIZE_T*) (charp);                         \
-    if(mzsz >= 5*sizeof(mzsz)) {     *mz++ = 0;                               \
-                                     *mz++ = 0;                               \
-      if(mzsz >= 7*sizeof(mzsz)) {   *mz++ = 0;                               \
-                                     *mz++ = 0;                               \
-        if(mzsz >= 9*sizeof(mzsz)) { *mz++ = 0;                               \
-                                     *mz++ = 0; }}}                           \
-                                     *mz++ = 0;                               \
-                                     *mz++ = 0;                               \
-                                     *mz   = 0;                               \
-  } else memset((charp), 0, mzsz);                                            \
-} while(0)
+#endif
+#endif
+/*
-#define MALLOC_COPY(dest,src,nbytes)                                          \
+  MORECORE-related declarations. By default, rely on sbrk
-do {                                                                          \
-  INTERNAL_SIZE_T mcsz = (nbytes);                                            \
-  if(mcsz <= 9*sizeof(mcsz)) {                                                \
-    INTERNAL_SIZE_T* mcsrc = (INTERNAL_SIZE_T*) (src);                        \
-    INTERNAL_SIZE_T* mcdst = (INTERNAL_SIZE_T*) (dest);                       \
-    if(mcsz >= 5*sizeof(mcsz)) {     *mcdst++ = *mcsrc++;                     \
-                                     *mcdst++ = *mcsrc++;                     \
-      if(mcsz >= 7*sizeof(mcsz)) {   *mcdst++ = *mcsrc++;                     \
-                                     *mcdst++ = *mcsrc++;                     \
-        if(mcsz >= 9*sizeof(mcsz)) { *mcdst++ = *mcsrc++;                     \
-                                     *mcdst++ = *mcsrc++; }}}                 \
-                                     *mcdst++ = *mcsrc++;                     \
-                                     *mcdst++ = *mcsrc++;                     \
-                                     *mcdst   = *mcsrc  ;                     \
-  } else memcpy(dest, src, mcsz);                                             \
-} while(0)
+*/
-#else /* !USE_MEMCPY */
+#ifdef LACKS_UNISTD_H
-/* Use Duff's device for good zeroing/copying performance. */
+#if !defined(__FreeBSD__) && !defined(__OpenBSD__) && !defined(__NetBSD__)
+#if __STD_C
+extern Void_t*     sbrk(ptrdiff_t);
+#else
+extern Void_t*     sbrk();
+#endif
+#endif
+#endif
+/*
-#define MALLOC_ZERO(charp, nbytes)                                            \
+  MORECORE is the name of the routine to call to obtain more memory
-do {                                                                          \
-  INTERNAL_SIZE_T* mzp = (INTERNAL_SIZE_T*)(charp);                           \
-  long mctmp = (nbytes)/sizeof(INTERNAL_SIZE_T), mcn;                         \
-  if (mctmp < 8) mcn = 0; else { mcn = (mctmp-1)/8; mctmp %= 8; }             \
+  from the system.  See below for general guidance on writing
-  switch (mctmp) {                                                            \
+  alternative MORECORE functions, as well as a version for WIN32 and a
-    case 0: for(;;) { *mzp++ = 0;                                             \
-    case 7:           *mzp++ = 0;                                             \
-    case 6:           *mzp++ = 0;                                             \
-    case 5:           *mzp++ = 0;                                             \
-    case 4:           *mzp++ = 0;                                             \
-    case 3:           *mzp++ = 0;                                             \
-    case 2:           *mzp++ = 0;                                             \
-    case 1:           *mzp++ = 0; if(mcn <= 0) break; mcn--; }                \
+  sample version for pre-OSX macos.
-  }                                                                           \
-} while(0)
+*/
+#ifndef MORECORE
-#define MALLOC_COPY(dest,src,nbytes)                                          \
+#define MORECORE sbrk
-do {                                                                          \
+#endif
+/*
-  INTERNAL_SIZE_T* mcsrc = (INTERNAL_SIZE_T*) src;                            \
+  MORECORE_FAILURE is the value returned upon failure of MORECORE
-  INTERNAL_SIZE_T* mcdst = (INTERNAL_SIZE_T*) dest;                           \
+  as well as mmap. Since it cannot be an otherwise valid memory address,
-  long mctmp = (nbytes)/sizeof(INTERNAL_SIZE_T), mcn;                         \
+  and must reflect values of standard sys calls, you probably ought not
-  if (mctmp < 8) mcn = 0; else { mcn = (mctmp-1)/8; mctmp %= 8; }             \
+  try to redefine it.
+*/
-  switch (mctmp) {                                                            \
+#ifndef MORECORE_FAILURE
-    case 0: for(;;) { *mcdst++ = *mcsrc++;                                    \
+#define MORECORE_FAILURE (-1)
+#endif
+/*
-    case 7:           *mcdst++ = *mcsrc++;                                    \
+  If MORECORE_CONTIGUOUS is true, take advantage of fact that
-    case 6:           *mcdst++ = *mcsrc++;                                    \
+  consecutive calls to MORECORE with positive arguments always return
-    case 5:           *mcdst++ = *mcsrc++;                                    \
+  contiguous increasing addresses.  This is true of unix sbrk.  Even
-    case 4:           *mcdst++ = *mcsrc++;                                    \
+  if not defined, when regions happen to be contiguous, malloc will
-    case 3:           *mcdst++ = *mcsrc++;                                    \
+  permit allocations spanning regions obtained from different
-    case 2:           *mcdst++ = *mcsrc++;                                    \
+  calls. But defining this when applicable enables some stronger
-    case 1:           *mcdst++ = *mcsrc++; if(mcn <= 0) break; mcn--; }       \
+  consistency checks and space efficiencies.
-  }                                                                           \
-} while(0)
+*/
+#ifndef MORECORE_CONTIGUOUS
+#define MORECORE_CONTIGUOUS 1
 #endif
+/*
+  Define MORECORE_CANNOT_TRIM if your version of 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.
+*/
+#define MORECORE_CANNOT_TRIM
 /*
-  Define HAVE_MMAP to optionally make malloc() use mmap() to
+  Define HAVE_MMAP as true to optionally make malloc() use mmap() to
   allocate very large blocks.  These will be returned to the
-  operating system immediately after a free().
+  operating system immediately after a free(). Also, if mmap
+  is available, it is used as a backup strategy in cases where
+  MORECORE fails to provide space from system.
+  This malloc is best tuned to work with mmap for large requests.
+  If you do not have mmap, operations involving very large chunks (1MB
+  or so) may be slower than you'd like.
 */
 #ifndef HAVE_MMAP
 #define HAVE_MMAP 1
 #endif
+#if HAVE_MMAP
+/*
+   Standard unix mmap using /dev/zero clears memory so calloc doesn't
+   need to.
+*/
+#ifndef MMAP_CLEARS
+#define MMAP_CLEARS 1
+#endif
+#else /* no mmap */
+#ifndef MMAP_CLEARS
+#define MMAP_CLEARS 0
+#endif
+#endif
+/*
+   MMAP_AS_MORECORE_SIZE is the minimum mmap size argument to use if
+   sbrk fails, and mmap is used as a backup (which is done only if
+   HAVE_MMAP).  The value must be a multiple of page size.  This
+   backup strategy generally applies only when systems have "holes" in
+   address space, so sbrk cannot perform contiguous expansion, but
+   there is still space available on system.  On systems for which
+   this is known to be useful (i.e. most linux kernels), this occurs
+   only when programs allocate huge amounts of memory.  Between this,
+   and the fact that mmap regions tend to be limited, the size should
+   be large, to avoid too many mmap calls and thus avoid running out
+   of kernel resources.
+*/
+#ifndef MMAP_AS_MORECORE_SIZE
+#define MMAP_AS_MORECORE_SIZE (1024 * 1024)
+#endif
 /*
   Define HAVE_MREMAP to make realloc() use mremap() to re-allocate
   large blocks.  This is currently only possible on Linux with
   kernel versions newer than 1.3.77.
 */
 #ifndef HAVE_MREMAP
-#ifdef INTERNAL_LINUX_C_LIB
+#ifdef linux
 #define HAVE_MREMAP 1
 #else
 #define HAVE_MREMAP 0
 #endif
-#endif
-#if HAVE_MMAP
-#include <unistd.h>
-#include <fcntl.h>
-#include <sys/mman.h>
-#if !defined(MAP_ANONYMOUS) && defined(MAP_ANON)
-#define MAP_ANONYMOUS MAP_ANON
-#endif
 #endif /* HAVE_MMAP */
-/*
-  Access to system page size. To the extent possible, this malloc
-  manages memory from the system in page-size units.
+/*
+  The system page size. To the extent possible, this malloc manages
+  memory from the system in page-size units.  Note that this value is
+  cached during initialization into a field of malloc_state. So even
+  if malloc_getpagesize is a function, it is only called once.
-  The following mechanics for getpagesize were adapted from
+  The following mechanics for getpagesize were adapted from bsd/gnu
+  getpagesize.h. If none of the system-probes here apply, a value of
+is used, which should be OK: If they don't apply, then using
-  bsd/gnu getpagesize.h
+  the actual value probably doesn't impact performance.
 */
+#ifndef malloc_getpagesize
 #ifndef LACKS_UNISTD_H
 #  include <unistd.h>
 #endif
-#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
+#      ifdef WIN32 /* use supplied emulation of getpagesize */
-#        define malloc_getpagesize (4096) /* TBD: Use 'GetSystemInfo' instead */
+#        define malloc_getpagesize getpagesize()
 #      else
 #        ifndef LACKS_SYS_PARAM_H
 #          include <sys/param.h>
 #        endif
 #        ifdef EXEC_PAGESIZE
 #            ifdef NBPC
 #              define malloc_getpagesize NBPC
 #            else
 #              ifdef PAGESIZE
 #                define malloc_getpagesize PAGESIZE
-#              else
+#              else /* just guess */
-#                define malloc_getpagesize (4096) /* just guess */
+#                define malloc_getpagesize (4096)
 #              endif
 #            endif
 #          endif
 #        endif
 #      endif
 #    endif
 #  endif
 #endif
 /*
   This version of malloc supports the standard SVID/XPG mallinfo
-  routine that returns a struct containing the same kind of
+  routine that returns a struct containing usage properties and
-  information you can get from malloc_stats. It should work on
-  any SVID/XPG compliant system that has a /usr/include/malloc.h
+  statistics. It should work on any SVID/XPG compliant system that has
-  defining struct mallinfo. (If you'd like to install such a thing
+  a /usr/include/malloc.h defining struct mallinfo. (If you'd like to
-  yourself, cut out the preliminary declarations as described above
+  install such a thing yourself, cut out the preliminary declarations
-  and below and save them in a malloc.h file. But there's no
+  as described above and below and save them in a malloc.h file. But
-  compelling reason to bother to do this.)
+  there's no compelling reason to bother to do this.)
   The main declaration needed is the mallinfo struct that is returned
   (by-copy) by mallinfo().  The SVID/XPG malloinfo struct contains a
-  bunch of fields, most of which are not even meaningful in this
+  bunch of fields that are not even meaningful in this version of
-  version of malloc. Some of these fields are are instead filled by
+  malloc.  These fields are are instead filled by mallinfo() with
-  mallinfo() with other numbers that might possibly be of interest.
+  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 an SVID2/XPG2 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
-  mallinfo() to work.
+  HAVE_USR_INCLUDE_MALLOC_H.
 */
 /* #define HAVE_USR_INCLUDE_MALLOC_H */
-#if HAVE_USR_INCLUDE_MALLOC_H
+#ifdef HAVE_USR_INCLUDE_MALLOC_H
 #include "/usr/include/malloc.h"
 #else
 /* SVID2/XPG mallinfo structure */
 struct mallinfo {
-  int arena;    /* total space allocated from system */
+  int arena;    /* non-mmapped space allocated from system */
-  int ordblks;  /* number of non-inuse chunks */
+  int ordblks;  /* number of free chunks */
-  int smblks;   /* unused -- always zero */
+  int smblks;   /* number of fastbin blocks */
   int hblks;    /* number of mmapped regions */
-  int hblkhd;   /* total space in mmapped regions */
+  int hblkhd;   /* space in mmapped regions */
-  int usmblks;  /* unused -- always zero */
+  int usmblks;  /* maximum total allocated space */
-  int fsmblks;  /* unused -- always zero */
+  int fsmblks;  /* space available in freed fastbin blocks */
   int uordblks; /* total allocated space */
-  int fordblks; /* total non-inuse space */
+  int fordblks; /* total free space */
   int keepcost; /* top-most, releasable (via malloc_trim) space */
 };
+/*
+  SVID/XPG defines four standard parameter numbers for mallopt,
+  normally defined in malloc.h.  Only one of these (M_MXFAST) is used
+  in this malloc. The others (M_NLBLKS, M_GRAIN, M_KEEP) don't apply,
+  so setting them has no effect. But this malloc also supports other
-/* SVID2/XPG mallopt options */
+  options in mallopt described below.
+*/
+#endif
-#define M_MXFAST  1    /* UNUSED in this malloc */
-#define M_NLBLKS  2    /* UNUSED in this malloc */
-#define M_GRAIN   3    /* UNUSED in this malloc */
+/* R Specific declarations here */
-#define M_KEEP    4    /* UNUSED in this malloc */
+/* The maximum via either sbrk or mmap */
-#endif
+unsigned long max_total_mem = 0;
-/* mallopt options that actually do something */
+extern unsigned int R_max_memory;
-#define M_TRIM_THRESHOLD    -1
+/* reserve 256MB to ensure large contiguous space */
-#define M_TOP_PAD           -2
+#define RESERVED_SIZE (256*1024*1024)
-#define M_MMAP_THRESHOLD    -3
+unsigned int R_reserved_size = RESERVED_SIZE;
-#define M_MMAP_MAX          -4
+/* ---------- description of public routines ------------ */
+/*
+  malloc(size_t n)
+  Returns a pointer to a newly allocated chunk of at least n bytes, or null
+  if no space is available. Additionally, on failure, errno is
+  set to ENOMEM on ANSI C systems.
+  If n is zero, malloc returns a minumum-sized chunk. (The minimum
+  size is 16 bytes on most 32bit systems, and 24 or 32 bytes on 64bit
+  systems.)  On most systems, size_t is an unsigned type, so calls
+  with negative arguments 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.
+*/
+#if __STD_C
+Void_t*  public_mALLOc(size_t);
+#else
+Void_t*  public_mALLOc();
+#endif
+/*
+  free(Void_t* 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. It can have arbitrary (i.e., bad!)
+  effects if p has already been freed.
+  Unless disabled (using mallopt), freeing very large spaces will
+  when possible, automatically trigger operations that give
+  back unused memory to the system, thus reducing program footprint.
+*/
+#if __STD_C
-#ifndef DEFAULT_TRIM_THRESHOLD
+void     public_fREe(Void_t*);
+#else
-#define DEFAULT_TRIM_THRESHOLD (128 * 1024)
+void     public_fREe();
 #endif
 /*
+  calloc(size_t n_elements, size_t element_size);
-    M_TRIM_THRESHOLD is the maximum amount of unused top-most memory
+  Returns a pointer to n_elements * element_size bytes, with all locations
+  set to zero.
+*/
+#if __STD_C
-      to keep before releasing via malloc_trim in free().
+Void_t*  public_cALLOc(size_t, size_t);
+#else
+Void_t*  public_cALLOc();
+#endif
-      Automatic trimming is mainly useful in long-lived programs.
-      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.
+/*
-      The trim threshold and the mmap control parameters (see below)
-      can be traded off with one another. Trimming and mmapping are
-      two different ways of releasing unused memory back to the
-      system. Between these two, it is often possible to keep
-      system-level demands of a long-lived program down to a bare
-      minimum. For example, in one test suite of sessions measuring
-      the XF86 X server on Linux, using a trim threshold of 128K and a
-      mmap threshold of 192K led to near-minimal long term resource
-      consumption.
+  realloc(Void_t* p, size_t n)
-      If you are using this malloc in a long-lived program, it should
-      pay to experiment with these values.  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's
-      worth it to tune for trimming rather tham memory mapping when a
-      program undergoes phases where several large chunks are
-      allocated and released in ways that can reuse each other's
+  Returns a pointer to a chunk of size n that contains the same data
-      storage, perhaps mixed with phases where there are no such
+  as does chunk p up to the minimum of (n, p's size) bytes, or null
-      chunks at all.  And in well-behaved long-lived programs,
-      controlling release of large blocks via trimming versus mapping
-      is usually faster.
+  if no space is available.
-      However, in most programs, these parameters serve mainly as
-      protection against the system-level effects of carrying around
-      massive amounts of unneeded memory. Since frequent calls to
-      sbrk, mmap, and munmap otherwise degrade performance, the default
-      parameters are set to relatively high values that serve only as
-      safeguards.
-      The default trim value is high enough to cause trimming only in
-      fairly extreme (by current memory consumption standards) cases.
-      It must be greater than page size to have any useful effect.  To
-      disable trimming completely, you can set to (unsigned long)(-1);
+  The returned pointer may or may not be the same as p. The algorithm
+  prefers extending p 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.  Unless the #define
+  REALLOC_ZERO_BYTES_FREES is set, realloc with a size argument of
+  zero (re)allocates a minimum-sized chunk.
+  Large chunks that were internally obtained via mmap will always
+  be reallocated using malloc-copy-free sequences unless
+  the system supports MREMAP (currently only linux).
+  The old unix realloc convention of allowing the last-free'd chunk
+  to be used as an argument to realloc is not supported.
+*/
-#ifndef DEFAULT_TOP_PAD
+#if __STD_C
-#define DEFAULT_TOP_PAD        (0)
+Void_t*  public_rEALLOc(Void_t*, size_t);
+#else
+Void_t*  public_rEALLOc();
 #endif
 /*
-    M_TOP_PAD is the amount of extra `padding' space to allocate or
+  memalign(size_t alignment, size_t n);
-      retain whenever sbrk is called. It is used in two ways internally:
+  Returns a pointer to a newly allocated chunk of n bytes, aligned
+  in accord with the alignment argument.
-      * When sbrk is called to extend the top of the arena to satisfy
+  The alignment argument should be a power of two. If the argument is
-        a new malloc request, this much padding is added to the sbrk
-        request.
-      * When malloc_trim is called automatically from free(),
-        it is used as the `pad' argument.
-      In both cases, the actual amount of padding is rounded
+  not a power of two, the nearest greater power is used.
-      so that the end of the arena is always a system page boundary.
-      The main reason for using padding is to avoid calling sbrk so
+-byte alignment is guaranteed by normal malloc calls, so don't
-      often. Having even a small pad greatly reduces the likelihood
-      that nearly every malloc request during program start-up (or
-      after trimming) will invoke sbrk, which needlessly wastes
-      time.
-      Automatic rounding-up to page-size units is normally sufficient
-      to avoid measurable overhead, so the default is 0.  However, in
-      systems where sbrk is relatively slow, it can pay to increase
-      this value, at the expense of carrying around more memory than
+  bother calling memalign with an argument of 8 or less.
-      the program needs.
+  Overreliance on memalign is a sure way to fragment space.
 */
+#if __STD_C
+Void_t*  public_mEMALIGn(size_t, size_t);
+#else
-#ifndef DEFAULT_MMAP_THRESHOLD
+Void_t*  public_mEMALIGn();
-#define DEFAULT_MMAP_THRESHOLD (128 * 1024)
 #endif
 /*
+  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.
+*/
+#if __STD_C
+Void_t*  public_vALLOc(size_t);
+#else
+Void_t*  public_vALLOc();
+#endif
-    M_MMAP_THRESHOLD is the request size threshold for using mmap()
-      to 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).
+/*
+  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
-      Segregating space in this way has the benefit that mmapped space
+  corresponding parameter to the argument value if it can (i.e., so
+  long as the value is meaningful), and returns 1 if successful else
-      can ALWAYS be individually released back to the system, which
+.  SVID/XPG/ANSI defines four standard param numbers for mallopt,
-      helps keep the system level memory demands of a long-lived
+  normally defined in malloc.h.  Only one of these (M_MXFAST) is used
-      program low. Mapped memory can never become `locked' between
+  in this malloc. The others (M_NLBLKS, M_GRAIN, M_KEEP) don't apply,
-      other chunks, as can happen with normally allocated chunks, which
+  so setting them has no effect. But this malloc also supports four
+  other options in mallopt. See below for details.  Briefly, supported
+  parameters are as follows (listed defaults are for "typical"
+  configurations).
-      menas that even trimming via malloc_trim would not release them.
+  Symbol            param #   default    allowed param values
+  M_MXFAST          1         64         0-80  (0 disables fastbins)
+  M_TRIM_THRESHOLD -1         256*1024   any   (-1U disables trimming)
+  M_TOP_PAD        -2         0          any
+  M_MMAP_THRESHOLD -3         256*1024   any   (or 0 if no MMAP support)
+  M_MMAP_MAX       -4         65536      any   (0 disables use of mmap)
+*/
+#if __STD_C
+int      public_mALLOPt(int, int);
+#else
+int      public_mALLOPt();
+#endif
-      However, it has the disadvantages that:
-. The space cannot be reclaimed, consolidated, and then
-            used to service later requests, as happens with normal chunks.
-. It can lead to more wastage because of mmap page alignment
+/*
-            requirements
+  mallinfo()
-. It causes malloc performance to be more dependent on host
-            system memory management support routines which may vary in
-            implementation quality and may impose arbitrary
-            limitations. Generally, servicing a request via normal
-            malloc steps is faster than going through a system's mmap.
+  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:    the number of fastbin blocks (i.e., small chunks that
-      All together, these considerations should lead you to use mmap
+               have been freed but not use resused or consolidated)
+  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:   total bytes held in fastbin blocks
+  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
-      only for relatively large requests.
+               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.
+*/
+#if __STD_C
+struct mallinfo public_mALLINFo(void);
+#else
+struct mallinfo public_mALLINFo();
+#endif
+/*
+  independent_calloc(size_t n_elements, size_t element_size, Void_t* 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;
+  }
 */
+#if __STD_C
+Void_t** public_iCALLOc(size_t, size_t, Void_t**);
+#else
+Void_t** public_iCALLOc();
+#endif
+/*
+  independent_comalloc(size_t n_elements, size_t sizes[], Void_t* 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
-#ifndef DEFAULT_MMAP_MAX
+  might be available for some of the elements.
+*/
-#if HAVE_MMAP
+#if __STD_C
-#define DEFAULT_MMAP_MAX       (64)
+Void_t** public_iCOMALLOc(size_t, size_t*, Void_t**);
 #else
-#define DEFAULT_MMAP_MAX       (0)
+Void_t** public_iCOMALLOc();
 #endif
+/*
+  pvalloc(size_t n);
+  Equivalent to valloc(minimum-page-that-holds(n)), that is,
+  round up n to nearest pagesize.
+ */
+#if __STD_C
+Void_t*  public_pVALLOc(size_t);
+#else
+Void_t*  public_pVALLOc();
 #endif
 /*
+  cfree(Void_t* p);
+  Equivalent to free(p).
-    M_MMAP_MAX is the maximum number of requests to simultaneously
+  cfree is needed/defined on some systems that pair it with calloc,
-      service using mmap. This parameter exists because:
+  for odd historical reasons (such as: cfree is used in example
+  code in the first edition of K&R).
+*/
+#if __STD_C
+void     public_cFREe(Void_t*);
+#else
+void     public_cFREe();
+#endif
-. Some systems have a limited number of internal tables for
+/*
-            use by mmap.
+  malloc_trim(size_t pad);
-. In most systems, overreliance on mmap can degrade overall
-            performance.
-. If a program allocates many large regions, it is probably
-            better off using normal sbrk-based allocation routines that
-            can reclaim and reallocate normal heap memory. Using a
-            small value allows transition into this mode after the
-            first few allocations.
+  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. 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
-      Setting to 0 disables all use of mmap.  If HAVE_MMAP is not set,
+  trailing space to leave untrimmed. If this argument is zero,
+  only the minimum amount of memory to maintain internal data
-      the default value is 0, and attempts to set it to non-zero values
+  structures will be left (one page or less). Non-zero arguments
+  can be supplied to maintain enough trailing space to service
+  future expected allocations without having to re-obtain memory
-      in mallopt will fail.
+  from the system.
+  Malloc_trim returns 1 if it actually released any memory, else 0.
+  On systems that do not support "negative sbrks", it will always
+  rreturn 0.
 */
+#if __STD_C
+int      public_mTRIm(size_t);
+#else
+int      public_mTRIm();
+#endif
+/*
+  malloc_usable_size(Void_t* 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);
+*/
+#if __STD_C
+size_t   public_mUSABLe(Void_t*);
+#else
+size_t   public_mUSABLe();
+#endif
 /*
+  malloc_stats();
-    USE_DL_PREFIX will prefix all public routines with the string 'dl'.
+  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
-      Useful to quickly avoid procedure declaration conflicts and linker
+  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
-      symbol conflicts with existing memory allocation routines.
+  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.
 */
+#if __STD_C
+void     public_mSTATs();
+#else
+void     public_mSTATs();
+#endif
-/* #define USE_DL_PREFIX */
+/* mallopt tuning options */
+/*
+  M_MXFAST is the maximum request size used for "fastbins", special bins
+  that hold returned chunks without consolidating their spaces. This
+  enables future requests for chunks of the same size to be handled
+  very quickly, but can increase fragmentation, and thus increase the
+  overall memory footprint of a program.
+  This malloc manages fastbins very conservatively yet still
+  efficiently, so fragmentation is rarely a problem for values less
+  than or equal to the default.  The maximum supported value of MXFAST
+  is 80. You wouldn't want it any higher than this anyway.  Fastbins
+  are designed especially for use with many small structs, objects or
+  strings -- the default handles structs/objects/arrays with sizes up
+  to 16 4byte fields, or small strings representing words, tokens,
+  etc. Using fastbins for larger objects normally worsens
+  fragmentation without improving speed.
+  M_MXFAST is set in REQUEST size units. It is internally used in
+  chunksize units, which adds padding and alignment.  You can reduce
+  M_MXFAST to 0 to disable all use of fastbins.  This causes the malloc
+  algorithm to be a closer approximation of fifo-best-fit in all cases,
+  not just for larger requests, but will generally cause it to be
+  slower.
+*/
+/* M_MXFAST is a standard SVID/XPG tuning option, usually listed in malloc.h */
+#ifndef M_MXFAST
+#define M_MXFAST            1
+#endif
+#ifndef DEFAULT_MXFAST
+#define DEFAULT_MXFAST     64
+#endif
+/*
+  M_TRIM_THRESHOLD is 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.
+  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.
+  The trim threshold and the mmap control parameters (see below)
+  can be traded off with one another. Trimming and mmapping are
+  two different ways of releasing unused memory back to the
+  system. Between these two, it is often possible to keep
+  system-level demands of a long-lived program down to a bare
+  minimum. For example, in one test suite of sessions measuring
+  the XF86 X server on Linux, using a trim threshold of 128K and a
+  mmap threshold of 192K led to near-minimal long term resource
+  consumption.
+  If you are using this malloc in a long-lived program, it should
+  pay to experiment with these values.  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's
+  worth it to tune for trimming rather tham memory mapping 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.  And in well-behaved long-lived programs,
+  controlling release of large blocks via trimming versus mapping
+  is usually faster.
+  However, in most programs, these parameters serve mainly as
+  protection against the system-level effects of carrying around
+  massive amounts of unneeded memory. Since frequent calls to
+  sbrk, mmap, and munmap otherwise degrade performance, the default
+  parameters are set to relatively high values that serve only as
+  safeguards.
+  The trim value must be greater than page size to have any useful
+  effect.  To disable trimming completely, you can set to
+  (unsigned long)(-1)
+  Trim settings interact with fastbin (MXFAST) settings: Unless
+  TRIM_FASTBINS is defined, automatic trimming never takes place upon
+  freeing a chunk with size less than or equal to MXFAST. Trimming is
+  instead delayed until subsequent freeing of larger chunks. However,
+  you can still force an attempted trim by calling malloc_trim.
+  Also, trimming is not generally possible in cases where
+  the main arena is obtained via mmap.
+  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.
+*/
+#define M_TRIM_THRESHOLD       -1
+#ifndef DEFAULT_TRIM_THRESHOLD
+#define DEFAULT_TRIM_THRESHOLD (256 * 1024)
+#endif
 /*
+  M_TOP_PAD is the amount of extra `padding' space to allocate or
+  retain whenever sbrk is called. It is used in two ways internally:
-  Special defines for linux libc
+  * When sbrk is called to extend the top of the arena to satisfy
+  a new malloc request, this much padding is added to the sbrk
+  request.
-  Except when compiled using these special defines for Linux libc
-  using weak aliases, this malloc is NOT designed to work in
-  multithreaded applications.  No semaphores or other concurrency
-  control are provided to ensure that multiple malloc or free calls
+  * When malloc_trim is called automatically from free(),
-  don't run at the same time, which could be disasterous. A single
-  semaphore could be used across malloc, realloc, and free (which is
-  essentially the effect of the linux weak alias approach). It would
-  be hard to obtain finer granularity.
+  it is used as the `pad' argument.
-*/
+  In both cases, the actual amount of padding is rounded
+  so that the end of the arena is always a system page boundary.
+  The main reason for using padding is to avoid calling sbrk so
+  often. Having even a small pad greatly reduces the likelihood
+  that nearly every malloc request during program start-up (or
+  after trimming) will invoke sbrk, which needlessly wastes
+  time.
+  Automatic rounding-up to page-size units is normally sufficient
+  to avoid measurable overhead, so the default is 0.  However, in
+  systems where sbrk is relatively slow, it can pay to increase
+  this value, at the expense of carrying around more memory than
-#ifdef INTERNAL_LINUX_C_LIB
+  the program needs.
+*/
-#if __STD_C
+#define M_TOP_PAD              -2
-Void_t * __default_morecore_init (ptrdiff_t);
+#ifndef DEFAULT_TOP_PAD
-Void_t *(*__morecore)(ptrdiff_t) = __default_morecore_init;
+#define DEFAULT_TOP_PAD        (0)
+#endif
+/*
+  M_MMAP_THRESHOLD is the request size threshold for using mmap()
+  to 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.
+. Mapped memory can never become `locked' between
+      other chunks, as can happen with normally allocated chunks, which
+      means that even trimming via malloc_trim would not release them.
+. On some systems with "holes" in address spaces, mmap can obtain
+      memory that sbrk cannot.
+  However, it has the disadvantages that:
+. The space cannot be reclaimed, consolidated, and then
+      used to service later requests, as happens with normal chunks.
+. It can lead to more wastage because of mmap page alignment
+      requirements
+. It causes malloc performance to be more dependent on host
+      system memory management support routines which may vary in
+      implementation quality and may impose arbitrary
+      limitations. Generally, servicing a request via normal
+      malloc steps is faster than going through a system's mmap.
+  The advantages of mmap nearly always outweigh disadvantages for
+  "large" chunks, but the value of "large" varies across systems.  The
+  default is an empirically derived value that works well in most
-#else
+  systems.
+*/
-Void_t * __default_morecore_init ();
+#define M_MMAP_THRESHOLD      -3
-Void_t *(*__morecore)() = __default_morecore_init;
+#ifndef DEFAULT_MMAP_THRESHOLD
+#define DEFAULT_MMAP_THRESHOLD (256 * 1024)
 #endif
+/*
+  M_MMAP_MAX is the maximum number of requests to simultaneously
-#define MORECORE (*__morecore)
+  service using mmap. This parameter exists because
+. Some systems have a limited number of internal tables for
+  use by mmap, and using more than a few of them may degrade
-#define MORECORE_FAILURE 0
+  performance.
+  The default is set to a value that serves only as a safeguard.
+  Setting to 0 disables use of mmap for servicing large requests.  If
+  HAVE_MMAP is not set, the default value is 0, and attempts to set it
-#define MORECORE_CLEARS 1
+  to non-zero values in mallopt will fail.
+*/
-#else /* INTERNAL_LINUX_C_LIB */
+#define M_MMAP_MAX             -4
+#ifndef DEFAULT_MMAP_MAX
-#if __STD_C
+#if HAVE_MMAP
-extern Void_t*     sbrk(ptrdiff_t);
+#define DEFAULT_MMAP_MAX       (65536)
 #else
-extern Void_t*     sbrk();
+#define DEFAULT_MMAP_MAX       (0)
 #endif
-#ifndef MORECORE
-#define MORECORE sbrk
 #endif
-#ifndef MORECORE_FAILURE
+#ifdef __cplusplus
-#define MORECORE_FAILURE -1
+};  /* end of extern "C" */
 #endif
+/*
+  ========================================================================
+  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" */
+/* --------------------- public wrappers ---------------------- */
-#ifndef MORECORE_CLEARS
+#ifdef USE_PUBLIC_MALLOC_WRAPPERS
+/* Declare all routines as internal */
+#if __STD_C
+static Void_t*  mALLOc(size_t);
+static void     fREe(Void_t*);
+static Void_t*  rEALLOc(Void_t*, size_t);
+static Void_t*  mEMALIGn(size_t, size_t);
+static Void_t*  vALLOc(size_t);
+static Void_t*  pVALLOc(size_t);
+static Void_t*  cALLOc(size_t, size_t);
+static Void_t** iCALLOc(size_t, size_t, Void_t**);
+static Void_t** iCOMALLOc(size_t, size_t*, Void_t**);
+static void     cFREe(Void_t*);
+static int      mTRIm(size_t);
+static size_t   mUSABLe(Void_t*);
+static void     mSTATs();
+static int      mALLOPt(int, int);
+static struct mallinfo mALLINFo(void);
+#else
+static Void_t*  mALLOc();
+static void     fREe();
+static Void_t*  rEALLOc();
+static Void_t*  mEMALIGn();
+static Void_t*  vALLOc();
+static Void_t*  pVALLOc();
+static Void_t*  cALLOc();
-#define MORECORE_CLEARS 1
+static Void_t** iCALLOc();
+static Void_t** iCOMALLOc();
+static void     cFREe();
+static int      mTRIm();
+static size_t   mUSABLe();
+static void     mSTATs();
+static int      mALLOPt();
+static struct mallinfo mALLINFo();
 #endif
+/*
+  MALLOC_PREACTION and MALLOC_POSTACTION should be
+  defined to return 0 on success, and nonzero on failure.
+  The return value of MALLOC_POSTACTION is currently ignored
+  in wrapper functions since there is no reasonable default
+  action to take on failure.
+*/
-#endif /* INTERNAL_LINUX_C_LIB */
+#ifdef USE_MALLOC_LOCK
+#ifdef WIN32
-#if defined(INTERNAL_LINUX_C_LIB) && defined(__ELF__)
-#define cALLOc		__libc_calloc
-#define fREe		__libc_free
-#define mALLOc		__libc_malloc
+static int mALLOC_MUTEx;
-#define mEMALIGn	__libc_memalign
-#define rEALLOc		__libc_realloc
-#define vALLOc		__libc_valloc
-#define pvALLOc		__libc_pvalloc
-#define mALLINFo	__libc_mallinfo
+#define MALLOC_PREACTION   slwait(&mALLOC_MUTEx)
-#define mALLOPt		__libc_mallopt
+#define MALLOC_POSTACTION  slrelease(&mALLOC_MUTEx)
-#pragma weak calloc = __libc_calloc
-#pragma weak free = __libc_free
-#pragma weak cfree = __libc_free
-#pragma weak malloc = __libc_malloc
-#pragma weak memalign = __libc_memalign
-#pragma weak realloc = __libc_realloc
-#pragma weak valloc = __libc_valloc
-#pragma weak pvalloc = __libc_pvalloc
-#pragma weak mallinfo = __libc_mallinfo
-#pragma weak mallopt = __libc_mallopt
 #else
-#ifdef USE_DL_PREFIX
-#define cALLOc		dlcalloc
-#define fREe		dlfree
-#define mALLOc		dlmalloc
-#define mEMALIGn	dlmemalign
-#define rEALLOc		dlrealloc
-#define vALLOc		dlvalloc
-#define pvALLOc		dlpvalloc
-#define mALLINFo	dlmallinfo
-#define mALLOPt		dlmallopt
-#else /* USE_DL_PREFIX */
-#define cALLOc		calloc
-#define fREe		free
-#define mALLOc		malloc
-#define mEMALIGn	memalign
-#define rEALLOc		realloc
+#include <pthread.h>
-#define vALLOc		valloc
-#define pvALLOc		pvalloc
-#define mALLINFo	mallinfo
-#define mALLOPt		mallopt
-#endif /* USE_DL_PREFIX */
-#endif
+static pthread_mutex_t mALLOC_MUTEx = PTHREAD_MUTEX_INITIALIZER;
+#define MALLOC_PREACTION   pthread_mutex_lock(&mALLOC_MUTEx)
-/* Public routines */
+#define MALLOC_POSTACTION  pthread_mutex_unlock(&mALLOC_MUTEx)
-#if __STD_C
+#endif /* USE_MALLOC_LOCK */
-Void_t* mALLOc(size_t);
-void    fREe(Void_t*);
-Void_t* rEALLOc(Void_t*, size_t);
-Void_t* mEMALIGn(size_t, size_t);
-Void_t* vALLOc(size_t);
-Void_t* pvALLOc(size_t);
-Void_t* cALLOc(size_t, size_t);
-void    cfree(Void_t*);
-int     malloc_trim(size_t);
-size_t  malloc_usable_size(Void_t*);
-void    malloc_stats();
-int     mALLOPt(int, int);
-struct mallinfo mALLINFo(void);
 #else
-Void_t* mALLOc();
-void    fREe();
-Void_t* rEALLOc();
-Void_t* mEMALIGn();
-Void_t* vALLOc();
-Void_t* pvALLOc();
-Void_t* cALLOc();
-void    cfree();
-int     malloc_trim();
-size_t  malloc_usable_size();
-void    malloc_stats();
-int     mALLOPt();
-struct mallinfo mALLINFo();
-#endif
+/* Substitute anything you like for these */
-#ifdef __cplusplus
+#define MALLOC_PREACTION   (0)
-};  /* end of extern "C" */
+#define MALLOC_POSTACTION  (0)
-#endif
-/* ---------- To make a malloc.h, end cutting here ------------ */
+#endif
+Void_t* public_mALLOc(size_t bytes) {
+  Void_t* m;
+  if (MALLOC_PREACTION != 0) {
+    return 0;
+  }
+  m = mALLOc(bytes);
+  if (MALLOC_POSTACTION != 0) {
+  }
+  return m;
+}
+void public_fREe(Void_t* m) {
+  if (MALLOC_PREACTION != 0) {
+    return;
-/*
+  }
-  Emulation of sbrk for WIN32
+  fREe(m);
-  All code within the ifdef WIN32 is untested by me.
+  if (MALLOC_POSTACTION != 0) {
+  }
+}
-  Thanks to Martin Fong and others for supplying this.
+Void_t* public_rEALLOc(Void_t* m, size_t bytes) {
+  if (MALLOC_PREACTION != 0) {
+    return 0;
+  }
+  m = rEALLOc(m, bytes);
+  if (MALLOC_POSTACTION != 0) {
+  }
+  return m;
-*/
+}
+Void_t* public_mEMALIGn(size_t alignment, size_t bytes) {
+  Void_t* m;
+  if (MALLOC_PREACTION != 0) {
+    return 0;
-/*
+  }
-  Quite heavily modified to make compile and remove redundant code,
-  and to coalesce adjacent blocks.  This version does not always
+  m = mEMALIGn(alignment, bytes);
-  allocate contiguous space, and it cannot return space from other
+  if (MALLOC_POSTACTION != 0) {
+  }
-  than the current allocation block.
+  return m;
+}
+Void_t* public_vALLOc(size_t bytes) {
+  Void_t* m;
+  if (MALLOC_PREACTION != 0) {
-  BDR 2000-8-20.
+    return 0;
+  }
+  m = vALLOc(bytes);
+  if (MALLOC_POSTACTION != 0) {
- */
+  }
+  return m;
+}
+Void_t* public_pVALLOc(size_t bytes) {
+  Void_t* m;
+  if (MALLOC_PREACTION != 0) {
-#ifdef WIN32
+    return 0;
+  }
-#include <windows.h>
+  m = pVALLOc(bytes);
-extern void R_Suicide(char *);
+  if (MALLOC_POSTACTION != 0) {
+  }
-static int PageSize=0;
+  return m;
+}
-#define AlignPage(add) (((add) + (PageSize-1)) & ~(PageSize-1))
+Void_t* public_cALLOc(size_t n, size_t elem_size) {
+  Void_t* m;
+  if (MALLOC_PREACTION != 0) {
+    return 0;
+  }
+  m = cALLOc(n, elem_size);
-#define AlignPage64K(add) (((add) + (0x10000 - 1)) & ~(0x10000 - 1))
+  if (MALLOC_POSTACTION != 0) {
+  }
+  return m;
+}
-/* reserve 256MB to ensure large contiguous space */
-#define RESERVED_SIZE (256*1024*1024)
-#define NEXT_SIZE (8*1024*1024)
-/* On most Windows systems, the bottom 2Gb of address space is
-   user space (on a few it is 3Gb) */
-#define TOP_MEMORY ((unsigned long)2*1024*1024*1024)
-static unsigned int gNextAddress = 0;
-static unsigned int gAddressBase = 0;
-static unsigned int gReservedSize = 0;
-static unsigned int totalAllocated = 0;
-extern unsigned int R_max_memory;
-unsigned int R_reserved_size = RESERVED_SIZE;
+Void_t** public_iCALLOc(size_t n, size_t elem_size, Void_t** chunks) {
-static
+  Void_t** m;
+  if (MALLOC_PREACTION != 0) {
-int getpagesize(void)
+    return 0;
-{
+  }
-    SYSTEM_INFO si;
+  m = iCALLOc(n, elem_size, chunks);
-    GetSystemInfo(&si);
+  if (MALLOC_POSTACTION != 0) {
+  }
-    return (int)si.dwPageSize;
+  return m;
+}
-static
-void *findRegion (void *start_address, unsigned long size)
+Void_t** public_iCOMALLOc(size_t n, size_t sizes[], Void_t** chunks) {
-{
+  Void_t** m;
-    MEMORY_BASIC_INFORMATION info;
+  if (MALLOC_PREACTION != 0) {
-    while ((unsigned long)start_address < TOP_MEMORY)
+    return 0;
-    {
+  }
-	VirtualQuery (start_address, &info, sizeof (info));
+  m = iCOMALLOc(n, sizes, chunks);
-	if (info.State == MEM_FREE && info.RegionSize >= size)
-	    return start_address;
+  if (MALLOC_POSTACTION != 0) {
-	else {
+  }
-/*	    Requested region is not available or not big enough
-	    so see if the next region is available.  Set 'start_address'
-	    to the next region and call 'VirtualQuery()' again. */
+  return m;
+}
-	    start_address = (char*)info.BaseAddress + info.RegionSize;
-/*	    Make sure we start looking for the next region
-	    on the *next* 64K boundary.  Otherwise, even if
-	    the new region is free according to
-	    'VirtualQuery()', the subsequent call to
-	    'VirtualAlloc()' (which follows the call to
-	    this routine in 'wsbrk()') will round *down*
-	    the requested address to a 64K boundary which
-	    we already know is an address in the
-	    unavailable region.  Thus, the subsequent call
-	    to 'VirtualAlloc()' will fail and bring us back
-	    here, causing us to go into an infinite loop. */
-	    start_address =
+void public_cFREe(Void_t* m) {
-		(void *) AlignPage64K((unsigned long) start_address);
+  if (MALLOC_PREACTION != 0) {
+    return;
-	}
+  }
+  cFREe(m);
+  if (MALLOC_POSTACTION != 0) {
-    }
+  }
-    return NULL;
+}
+int public_mTRIm(size_t s) {
-extern int R_Is_Running;
+  int result;
+  if (MALLOC_PREACTION != 0) {
+    return 0;
+  }
+  result = mTRIm(s);
-void Rf_warning(const char *, ...);
+  if (MALLOC_POSTACTION != 0) {
+  }
+  return result;
+}
-/* BDR 2000-10-28: add loop count as this infinitely looped */
+size_t public_mUSABLe(Void_t* m) {
+  size_t result;
-void *wsbrk (long size)
+  if (MALLOC_PREACTION != 0) {
+    return 0;
-{
+  }
-    void *tmp;
+  result = mUSABLe(m);
-    int count = 0;
+  if (MALLOC_POSTACTION != 0) {
+  }
-    if (PageSize == 0) PageSize = getpagesize();
+  return result;
+}
-    if (size > 0) {
+void public_mSTATs() {
-	/* will this take us over the limit? */
+  if (MALLOC_PREACTION != 0) {
-	if (totalAllocated + size > R_max_memory) {
+    return;
+  }
-	    if(R_Is_Running)
+  mSTATs();
-		Rf_warning("Reached total allocation of %dMb: see help(memory.size)", R_max_memory/1048576);
-	    return (void*)-1;
+  if (MALLOC_POSTACTION != 0) {
+  }
-	}
+}
-	/* first check if request fits in reserved space, and if not
-	   try to reserve the address space (never unreserved) */
-	if (gAddressBase == 0) {
-	    gReservedSize = max (R_reserved_size, AlignPage64K(size));
-	    gNextAddress = gAddressBase =
-		(unsigned int)VirtualAlloc (NULL, gReservedSize,
-					    MEM_RESERVE, PAGE_NOACCESS);
+struct mallinfo public_mALLINFo() {
-	    if(!gAddressBase) {
+  struct mallinfo m;
-		/* can't R_Suicide here
+  if (MALLOC_PREACTION != 0) {
-		   R_Suicide("unable to reserve initial space in wsbrk"); */
+    struct mallinfo nm = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
-		MessageBox(0, "Cannot reserve memory:\nterminating",
-			   "R fatal error",
+    return nm;
-			   MB_TASKMODAL | MB_ICONSTOP | MB_OK);
-		exit(3);
-	    }
+  }
-	} else if (AlignPage (gNextAddress + size) > (gAddressBase +
-						      gReservedSize)) {
-	    long new_size = max (NEXT_SIZE, AlignPage64K(size));
-	    void *new_address = (void*)(gAddressBase+gReservedSize);
-	    do {
+  m = mALLINFo();
-		/* if we failed twice, try somewhat desparately */
-		if (count >= 2)
+  if (MALLOC_POSTACTION != 0) {
-		    new_address = (void *)((unsigned int) new_address + new_size + NEXT_SIZE);
-		new_address = findRegion (new_address, new_size);
-		if (new_address == 0) return (void *) -1;
-		gNextAddress =
-		    (unsigned int)VirtualAlloc (new_address, new_size,
-						MEM_RESERVE, PAGE_NOACCESS);
-				/* repeat in case of race condition
-				  The region that we found has been grabbed
-				  by another thread */
-	    } while (gNextAddress == 0 && count++ < 10);
-	    if (gNextAddress == 0) return (void *) -1;
-	    if (new_address != (void*)gNextAddress)
-		R_Suicide("internal error in wsbrk");
+  }
-	    if (gNextAddress == gAddressBase + gReservedSize) {
-		/* allocated a contiguous block */
-		gReservedSize += new_size;
-	    } else {
+  return m;
-		gAddressBase = gNextAddress;
-		gReservedSize = new_size;
-	    }
-	}
+}
-	/* Now actually allocate, if not already in an allocated page */
-	if ((size + gNextAddress) > AlignPage(gNextAddress)) {
+int public_mALLOPt(int p, int v) {
-	    void* res;
+  int result;
-	    res = VirtualAlloc ((void*)AlignPage(gNextAddress),
-				(size + gNextAddress -
-				 AlignPage(gNextAddress)),
+  if (MALLOC_PREACTION != 0) {
-				MEM_COMMIT, PAGE_READWRITE);
-	    if (res == 0)
+    return 0;
-		return (void*)-1;
-	}
+  }
-	tmp = (void*)gNextAddress;
-	gNextAddress = (unsigned int)tmp + size;
-	totalAllocated += size + gNextAddress - AlignPage(gNextAddress);
-	return tmp;
+  result = mALLOPt(p, v);
-    } else if (size < 0) {
+  if (MALLOC_POSTACTION != 0) {
-	unsigned int alignedGoal = AlignPage(gNextAddress + size);
-	/* Trim by decommiting the virtual memory */
-	if (alignedGoal >= gAddressBase)
-	{
-	    VirtualFree ((void*)alignedGoal, gNextAddress - alignedGoal,
-			 MEM_DECOMMIT);
-	    totalAllocated -= gNextAddress - alignedGoal;
-	    gNextAddress = gNextAddress + size;
-	    return (void*)gNextAddress;
-	}
+  }
-	else /* requested release of more than we have in this block */
-	{
-	    VirtualFree ((void*)gAddressBase, gNextAddress - gAddressBase,
-			 MEM_DECOMMIT);
-	    totalAllocated -= gNextAddress - gAddressBase;
-	    gNextAddress = gAddressBase;
-	    return (void*)-1;
+  return result;
-	}
-    } else { /* size = 0 */
-	return (void*)gNextAddress;
-    }
+}
-#endif /* WIN32 */
+#endif
+/* ------------- Optional versions of memcopy ---------------- */
+#if USE_MEMCPY
+/*
+  Note: memcpy is ONLY invoked with non-overlapping regions,
+  so the (usually slower) memmove is not needed.
+*/
+#define MALLOC_COPY(dest, src, nbytes)  memcpy(dest, src, nbytes)
+#define MALLOC_ZERO(dest, nbytes)       memset(dest, 0,   nbytes)
+#else /* !USE_MEMCPY */
+/* Use Duff's device for good zeroing/copying performance. */
+#define MALLOC_ZERO(charp, nbytes)                                            \
+do {                                                                          \
+  INTERNAL_SIZE_T* mzp = (INTERNAL_SIZE_T*)(charp);                           \
+  CHUNK_SIZE_T  mctmp = (nbytes)/sizeof(INTERNAL_SIZE_T);                     \
+  long mcn;                                                                   \
+  if (mctmp < 8) mcn = 0; else { mcn = (mctmp-1)/8; mctmp %= 8; }             \
+  switch (mctmp) {                                                            \
+    case 0: for(;;) { *mzp++ = 0;                                             \
+    case 7:           *mzp++ = 0;                                             \
+    case 6:           *mzp++ = 0;                                             \
+    case 5:           *mzp++ = 0;                                             \
+    case 4:           *mzp++ = 0;                                             \
+    case 3:           *mzp++ = 0;                                             \
+    case 2:           *mzp++ = 0;                                             \
+    case 1:           *mzp++ = 0; if(mcn <= 0) break; mcn--; }                \
+  }                                                                           \
+} while(0)
+#define MALLOC_COPY(dest,src,nbytes)                                          \
+do {                                                                          \
+  INTERNAL_SIZE_T* mcsrc = (INTERNAL_SIZE_T*) src;                            \
+  INTERNAL_SIZE_T* mcdst = (INTERNAL_SIZE_T*) dest;                           \
+  CHUNK_SIZE_T  mctmp = (nbytes)/sizeof(INTERNAL_SIZE_T);                     \
+  long mcn;                                                                   \
+  if (mctmp < 8) mcn = 0; else { mcn = (mctmp-1)/8; mctmp %= 8; }             \
+  switch (mctmp) {                                                            \
+    case 0: for(;;) { *mcdst++ = *mcsrc++;                                    \
+    case 7:           *mcdst++ = *mcsrc++;                                    \
+    case 6:           *mcdst++ = *mcsrc++;                                    \
+    case 5:           *mcdst++ = *mcsrc++;                                    \
+    case 4:           *mcdst++ = *mcsrc++;                                    \
+    case 3:           *mcdst++ = *mcsrc++;                                    \
+    case 2:           *mcdst++ = *mcsrc++;                                    \
+    case 1:           *mcdst++ = *mcsrc++; if(mcn <= 0) break; mcn--; }       \
+  }                                                                           \
+} while(0)
+#endif
+/* ------------------ MMAP support ------------------  */
+#if HAVE_MMAP
+#ifndef LACKS_FCNTL_H
+#include <fcntl.h>
+#endif
+#ifndef LACKS_SYS_MMAN_H
+#include <sys/mman.h>
+#endif
+#if !defined(MAP_ANONYMOUS) && defined(MAP_ANON)
+#define MAP_ANONYMOUS MAP_ANON
+#endif
+/*
+   Nearly all versions of mmap support MAP_ANONYMOUS,
+   so the following is unlikely to be needed, but is
+   supplied just in case.
+*/
+#ifndef MAP_ANONYMOUS
+static int dev_zero_fd = -1; /* Cached file descriptor for /dev/zero. */
+#define MMAP(addr, size, prot, flags) ((dev_zero_fd < 0) ? \
+ (dev_zero_fd = open("/dev/zero", O_RDWR), \
+  mmap((addr), (size), (prot), (flags), dev_zero_fd, 0)) : \
+   mmap((addr), (size), (prot), (flags), dev_zero_fd, 0))
+#else
+#define MMAP(addr, size, prot, flags) \
+ (mmap((addr), (size), (prot), (flags)|MAP_ANONYMOUS, -1, 0))
+#endif
+#endif /* HAVE_MMAP */
 /*
-  Type declarations
+  -----------------------  Chunk representations -----------------------
 */
+/*
+  This struct declaration is misleading (but accurate and necessary).
+  It declares a "view" into memory allowing access to necessary
+  fields at known offsets from a given base. See explanation below.
+*/
-struct malloc_chunk
+struct malloc_chunk {
-{
-  INTERNAL_SIZE_T prev_size; /* Size of previous chunk (if free). */
+  INTERNAL_SIZE_T      prev_size;  /* Size of previous chunk (if free).  */
-  INTERNAL_SIZE_T size;      /* Size in bytes, including overhead. */
+  INTERNAL_SIZE_T      size;       /* Size in bytes, including overhead. */
-  struct malloc_chunk* fd;   /* double links -- used only if free. */
+  struct malloc_chunk* fd;         /* double links -- used only if free. */
   struct malloc_chunk* bk;
 };
 typedef struct malloc_chunk* mchunkptr;
 /*
    malloc_chunk details:
     (The following includes lightly edited explanations by Colin Plumb.)
     Chunks of memory are maintained using a `boundary tag' method as
     the malloc code, but "mem" is the pointer that is returned to the
     user.  "Nextchunk" is the beginning of the next contiguous chunk.
     Chunks always begin on even word boundries, so the mem portion
     (which is returned to the user) is also on an even word boundary, and
-    thus double-word aligned.
+    thus at least double-word aligned.
     Free chunks are stored in circular doubly-linked lists, and look like this:
     chunk-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
             |             Size of previous chunk                            |
     The P (PREV_INUSE) 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,
+    The very first chunk allocated always has this bit set,
-    preventing access to non-existent (or non-owned) memory.)
+    preventing access to non-existent (or non-owned) memory. If
+    prev_inuse 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.
     Note that the `foot' of the current chunk is actually represented
-    as the prev_size of the NEXT chunk. (This makes it easier to
+    as the prev_size of the NEXT chunk. This makes it easier to
+    deal with alignments etc but can be very confusing when trying
-    deal with alignments etc).
+    to extend or adapt this code.
     The two exceptions to all this are
-. The special chunk `top', which doesn't bother using the
+. The special chunk `top' doesn't bother using the
-        trailing size field since there is no
+        trailing size field since there is no next contiguous chunk
-        next contiguous chunk that would have to index off it. (After
+        that would have to index off it. After initialization, `top'
-        initialization, `top' is forced to always exist.  If it would
+        is forced to always exist.  If it would become less than
-        become less than MINSIZE bytes long, it is replenished via
+        MINSIZE bytes long, it is replenished.
-        malloc_extend_top.)
 . Chunks allocated via mmap, which have the second-lowest-order
         bit (IS_MMAPPED) set in their size fields.  Because they are
-        never merged or traversed from any other chunk, they have no
+        allocated one-by-one, each must contain its own trailing size field.
-        foot size or inuse information.
-    Available chunks are kept in any of several places (all declared below):
-    * `av': An array of chunks serving as bin headers for consolidated
-       chunks. Each bin is doubly linked.  The bins are approximately
-       proportionally (log) spaced.  There are a lot of these bins
-       (128). This may look excessive, but works very well in
-       practice.  All procedures maintain the invariant that no
-       consolidated chunk physically borders another one. Chunks in
-       bins are kept in size order, with ties going to the
-       approximately least recently used chunk.
+*/
-       The chunks in each bin are maintained in decreasing sorted order by
-       size.  This is irrelevant for the small bins, which all contain
-       the same-sized chunks, but facilitates best-fit allocation for
-       larger chunks. (These lists are just sequential. Keeping them in
-       order almost never requires enough traversal to warrant using
-       fancier ordered data structures.)  Chunks of the same size are
-       linked with the most recently freed at the front, and allocations
-       are taken from the back.  This results in LRU or FIFO allocation
-       order, which tends to give each chunk an equal opportunity to be
-       consolidated with adjacent freed chunks, resulting in larger free
-       chunks and less fragmentation.
-    * `top': The top-most available chunk (i.e., the one bordering the
-       end of available memory) is treated specially. It 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).
-    * `last_remainder': A bin holding only the remainder of the
-       most recently split (non-top) chunk. This bin is checked
-       before other non-fitting chunks, so as to provide better
-       locality for runs of sequentially allocated chunks.
-    *  Implicitly, through the host system's memory mapping tables.
-       If supported, requests greater than a threshold are usually
-       serviced via calls to mmap, and then later released via munmap.
+/*
+  ---------- Size and alignment checks and conversions ----------
 */
+/* conversion from malloc headers to user pointers, and back */
+#define chunk2mem(p)   ((Void_t*)((char*)(p) + 2*SIZE_SZ))
+#define mem2chunk(mem) ((mchunkptr)((char*)(mem) - 2*SIZE_SZ))
+/* The smallest possible chunk */
+#define MIN_CHUNK_SIZE        (sizeof(struct malloc_chunk))
+/* The smallest size we can malloc is an aligned minimal chunk */
-/*  sizes, alignments */
+#define MINSIZE  \
+  (CHUNK_SIZE_T)(((MIN_CHUNK_SIZE+MALLOC_ALIGN_MASK) & ~MALLOC_ALIGN_MASK))
-#define SIZE_SZ                (sizeof(INTERNAL_SIZE_T))
-#define MALLOC_ALIGNMENT       (SIZE_SZ + SIZE_SZ)
-#define MALLOC_ALIGN_MASK      (MALLOC_ALIGNMENT - 1)
-#define MINSIZE                (sizeof(struct malloc_chunk))
+/* Check if m has acceptable alignment */
-/* conversion from malloc headers to user pointers, and back */
+#define aligned_OK(m)  (((PTR_UINT)((m)) & (MALLOC_ALIGN_MASK)) == 0)
-#define chunk2mem(p)   ((Void_t*)((char*)(p) + 2*SIZE_SZ))
-#define mem2chunk(mem) ((mchunkptr)((char*)(mem) - 2*SIZE_SZ))
+/*
-/* pad request bytes into a usable size */
+   Check if a request is so large that it would wrap around zero when
+   padded and aligned. To simplify some other code, the bound is made
+   low enough so that adding MINSIZE will also not wrap around sero.
+*/
-#define request2size(req) \
+#define REQUEST_OUT_OF_RANGE(req)                                 \
- (((long)((req) + (SIZE_SZ + MALLOC_ALIGN_MASK)) < \
+  ((CHUNK_SIZE_T)(req) >=                                        \
-  (long)(MINSIZE + MALLOC_ALIGN_MASK)) ? MINSIZE : \
+   (CHUNK_SIZE_T)(INTERNAL_SIZE_T)(-2 * MINSIZE))
-   (((req) + (SIZE_SZ + MALLOC_ALIGN_MASK)) & ~(MALLOC_ALIGN_MASK)))
-/* Check if m has acceptable alignment */
+/* pad request bytes into a usable size -- internal version */
+#define request2size(req)                                         \
+  (((req) + SIZE_SZ + MALLOC_ALIGN_MASK < MINSIZE)  ?             \
+   MINSIZE :                                                      \
-#define aligned_OK(m)    (((unsigned long)((m)) & (MALLOC_ALIGN_MASK)) == 0)
+   ((req) + SIZE_SZ + MALLOC_ALIGN_MASK) & ~MALLOC_ALIGN_MASK)
+/*  Same, except also perform argument check */
+#define checked_request2size(req, sz)                             \
+  if (REQUEST_OUT_OF_RANGE(req)) {                                \
+    MALLOC_FAILURE_ACTION;                                        \
+    return 0;                                                     \
+  }                                                               \
+  (sz) = request2size(req);
 /*
-  Physical chunk operations
+  --------------- Physical chunk operations ---------------
 */
 /* size field is or'ed with PREV_INUSE when previous adjacent chunk in use */
 #define PREV_INUSE 0x1
-/* size field is or'ed with IS_MMAPPED if the chunk was obtained with mmap() */
+/* extract inuse bit of previous chunk */
+#define prev_inuse(p)       ((p)->size & PREV_INUSE)
+/* size field is or'ed with IS_MMAPPED if the chunk was obtained with mmap() */
 #define IS_MMAPPED 0x2
+/* check for mmap()'ed chunk */
+#define chunk_is_mmapped(p) ((p)->size & IS_MMAPPED)
+/*
-/* Bits to mask off when extracting size */
+  Bits to mask off when extracting size
+  Note: IS_MMAPPED is intentionally not masked off from size field in
+  macros for which mmapped chunks should never be seen. This should
+  cause helpful core dumps to occur if it is tried by accident by
+  people extending or adapting this malloc.
+*/
 #define SIZE_BITS (PREV_INUSE|IS_MMAPPED)
+/* Get size, ignoring use bits */
+#define chunksize(p)         ((p)->size & ~(SIZE_BITS))
-/* Ptr to next physical malloc_chunk. */
+/* Ptr to next physical malloc_chunk. */
 #define next_chunk(p) ((mchunkptr)( ((char*)(p)) + ((p)->size & ~PREV_INUSE) ))
 /* Ptr to previous physical malloc_chunk */
-#define prev_chunk(p)\
-   ((mchunkptr)( ((char*)(p)) - ((p)->prev_size) ))
+#define prev_chunk(p) ((mchunkptr)( ((char*)(p)) - ((p)->prev_size) ))
 /* Treat space at ptr + offset as a chunk */
 #define chunk_at_offset(p, s)  ((mchunkptr)(((char*)(p)) + (s)))
-/*
-  Dealing with use bits
-*/
 /* extract p's inuse bit */
 #define inuse(p)\
 ((((mchunkptr)(((char*)(p))+((p)->size & ~PREV_INUSE)))->size) & PREV_INUSE)
-/* extract inuse bit of previous chunk */
-#define prev_inuse(p)  ((p)->size & PREV_INUSE)
-/* check for mmap()'ed chunk */
-#define chunk_is_mmapped(p) ((p)->size & IS_MMAPPED)
-/* set/clear chunk as in use without otherwise disturbing */
+/* set/clear chunk as being inuse without otherwise disturbing */
 #define set_inuse(p)\
 ((mchunkptr)(((char*)(p)) + ((p)->size & ~PREV_INUSE)))->size |= PREV_INUSE
 #define clear_inuse(p)\
 ((mchunkptr)(((char*)(p)) + ((p)->size & ~PREV_INUSE)))->size &= ~(PREV_INUSE)
-/* check/set/clear inuse bits in known places */
+/* check/set/clear inuse bits in known places */
 #define inuse_bit_at_offset(p, s)\
  (((mchunkptr)(((char*)(p)) + (s)))->size & PREV_INUSE)
 #define set_inuse_bit_at_offset(p, s)\
  (((mchunkptr)(((char*)(p)) + (s)))->size |= PREV_INUSE)
 #define clear_inuse_bit_at_offset(p, s)\
  (((mchunkptr)(((char*)(p)) + (s)))->size &= ~(PREV_INUSE))
-/*
-  Dealing with size fields
+/* Set size at head, without disturbing its use bit */
-*/
-/* Get size, ignoring use bits */
+#define set_head_size(p, s)  ((p)->size = (((p)->size & PREV_INUSE) | (s)))
+/* Set size/use field */
-#define chunksize(p)          ((p)->size & ~(SIZE_BITS))
+#define set_head(p, s)       ((p)->size = (s))
-/* Set size at head, without disturbing its use bit */
+/* Set size at footer (only when chunk is not in use) */
+#define set_foot(p, s)       (((mchunkptr)((char*)(p) + (s)))->prev_size = (s))
-#define set_head_size(p, s)   ((p)->size = (((p)->size & PREV_INUSE) | (s)))
+/*
-/* Set size/use ignoring previous bits in header */
+  -------------------- Internal data structures --------------------
+   All internal state is held in an instance of malloc_state defined
+   below. There are no other static variables, except in two optional
+   cases:
+   * If USE_MALLOC_LOCK is defined, the mALLOC_MUTEx declared above.
-#define set_head(p, s)        ((p)->size = (s))
+   * If HAVE_MMAP is true, but mmap doesn't support
+     MAP_ANONYMOUS, a dummy file descriptor for mmap.
+   Beware of lots of tricks that minimize the total bookkeeping space
+   requirements. The result is a little over 1K bytes (for 4byte
+   pointers and size_t.)
+*/
+/*
+  Bins
+    An array of bin headers for free chunks. Each bin is doubly
+    linked.  The bins are approximately proportionally (log) spaced.
+    There are a lot of these bins (128). This may look excessive, but
+    works very well in practice.  Most bins hold sizes that are
+    unusual as malloc request sizes, but are more usual for fragments
+    and consolidated sets of chunks, which is what these bins hold, so
+    they can be found quickly.  All procedures maintain the invariant
+    that no consolidated chunk physically borders another one, so each
+    chunk in a list is known to be preceeded and followed by either
+    inuse chunks or the ends of memory.
+    Chunks in bins are kept in size order, with ties going to the
-/* Set size at footer (only when chunk is not in use) */
+    approximately least recently used chunk. Ordering isn't needed
+    for the small bins, which all contain the same-sized chunks, but
+    facilitates best-fit allocation for larger chunks. These lists
+    are just sequential. Keeping them in order almost never requires
+    enough traversal to warrant using fancier ordered data
+    structures.
+    Chunks of the same size are linked with the most
+    recently freed at the front, and allocations are taken from the
+    back.  This results in LRU (FIFO) allocation order, which tends
+    to give each chunk an equal opportunity to be consolidated with
+    adjacent freed chunks, resulting in larger free chunks and less
+    fragmentation.
+    To simplify use in double-linked lists, each bin header acts
+    as a malloc_chunk. This avoids special-casing for headers.
+    But to conserve space and improve locality, we allocate
+    only the fd/bk pointers of bins, and then use repositioning tricks
+    to treat these as the fields of a malloc_chunk*.
+*/
-#define set_foot(p, s)   (((mchunkptr)((char*)(p) + (s)))->prev_size = (s))
+typedef struct malloc_chunk* mbinptr;
+/* addressing -- note that bin_at(0) does not exist */
+#define bin_at(m, i) ((mbinptr)((char*)&((m)->bins[(i)<<1]) - (SIZE_SZ<<1)))
+/* analog of ++bin */
+#define next_bin(b)  ((mbinptr)((char*)(b) + (sizeof(mchunkptr)<<1)))
+/* Reminders about list directionality within bins */
+#define first(b)     ((b)->fd)
+#define last(b)      ((b)->bk)
+/* Take a chunk off a bin list */
+#define unlink(P, BK, FD) {                                            \
+  FD = P->fd;                                                          \
+  BK = P->bk;                                                          \
+  FD->bk = BK;                                                         \
+  BK->fd = FD;                                                         \
+}
 /*
-   Bins
+  Indexing
-    The bins, `av_' are an array of pairs of pointers serving as the
-    heads of (initially empty) doubly-linked lists of chunks, laid out
-    in a way so that each pair can be treated as if it were in a
-    malloc_chunk. (This way, the fd/bk offsets for linking bin heads
-    and chunks are the same).
     Bins for sizes < 512 bytes contain chunks of all the same size, spaced
-bytes apart. Larger bins are approximately logarithmically
+bytes apart. Larger bins are approximately logarithmically spaced:
-    spaced. (See the table below.) The `av_' array is never mentioned
-    directly in the code, but instead via bin access macros.
-    Bin layout:
 bins of size       8
 bins of size      64
 bins of size     512
 bins of size    4096
 bins of size   32768
 bins of size  262144
 bin  of size what's left
-    There is actually a little bit of slop in the numbers in bin_index
+    The bins top out around 1MB because we expect to service large
+    requests via mmap.
+*/
+#define NBINS              96
+#define NSMALLBINS         32
+#define SMALLBIN_WIDTH      8
+#define MIN_LARGE_SIZE    256
+#define in_smallbin_range(sz)  \
-    for the sake of speed. This makes no difference elsewhere.
+  ((CHUNK_SIZE_T)(sz) < (CHUNK_SIZE_T)MIN_LARGE_SIZE)
-    The special chunks `top' and `last_remainder' get their own bins,
-    (this is implemented via yet more trickery with the av_ array),
+#define smallbin_index(sz)     (((unsigned)(sz)) >> 3)
-    although `top' is never properly linked to its bin since it is
-    always handled specially.
+/*
+  Compute index for size. We expect this to be inlined when
+  compiled with optimization, else not, which works out well.
 */
+static int largebin_index(unsigned int sz) {
+  unsigned int  x = sz >> SMALLBIN_WIDTH;
+  unsigned int m;            /* bit position of highest set bit of m */
-#define NAV             128   /* number of bins */
+  if (x >= 0x10000) return NBINS-1;
+  /* On intel, use BSRL instruction to find highest bit */
+#if defined(__GNUC__) && defined(i386)
+  __asm__("bsrl %1,%0\n\t"
+          : "=r" (m)
+          : "g"  (x));
+#else
+  {
+    /*
+      Based on branch-free nlz algorithm in chapter 5 of Henry
-typedef struct malloc_chunk* mbinptr;
+      S. Warren Jr's book "Hacker's Delight".
+    */
+    unsigned int n = ((x - 0x100) >> 16) & 8;
+    x <<= n;
+    m = ((x - 0x1000) >> 16) & 4;
+    n += m;
+    x <<= m;
+    m = ((x - 0x4000) >> 16) & 2;
+    n += m;
+    x = (x << m) >> 14;
+    m = 13 - n + (x & ~(x>>1));
+  }
+#endif
+  /* Use next 2 bits to create finer-granularity bins */
+  return NSMALLBINS + (m << 2) + ((sz >> (m + 6)) & 3);
+}
+#define bin_index(sz) \
+ ((in_smallbin_range(sz)) ? smallbin_index(sz) : largebin_index(sz))
+/*
+  FIRST_SORTED_BIN_SIZE is the chunk size corresponding to the
+  first bin that is maintained in sorted order. This must
+  be the smallest size corresponding to a given bin.
+  Normally, this should be MIN_LARGE_SIZE. But you can weaken
+  best fit guarantees to sometimes speed up malloc by increasing value.
-/* access macros */
+  Doing this means that malloc may choose a chunk that is
+  non-best-fitting by up to the width of the bin.
+  Some useful cutoff values:
+- all bins sorted
-#define bin_at(i)      ((mbinptr)((char*)&(av_[2*(i) + 2]) - 2*SIZE_SZ))
+- leaves bins <=     64 bytes wide unsorted
-#define next_bin(b)    ((mbinptr)((char*)(b) + 2 * sizeof(mbinptr)))
+- leaves bins <=    512 bytes wide unsorted
-#define prev_bin(b)    ((mbinptr)((char*)(b) - 2 * sizeof(mbinptr)))
+- leaves bins <=   4096 bytes wide unsorted
+- leaves bins <=  32768 bytes wide unsorted
+       -1 - no bins sorted (not recommended!)
+*/
+#define FIRST_SORTED_BIN_SIZE MIN_LARGE_SIZE
+/* #define FIRST_SORTED_BIN_SIZE 65536 */
 /*
+  Unsorted chunks
+    All remainders from chunk splits, as well as all returned chunks,
-   The first 2 bins are never indexed. The corresponding av_ cells are instead
+    are first placed in the "unsorted" bin. They are then placed
-   used for bookkeeping. This is not to save space, but to simplify
+    in regular bins after malloc gives them ONE chance to be used before
-   indexing, maintain locality, and avoid some initialization tests.
+    binning. So, basically, the unsorted_chunks list acts as a queue,
+    with chunks being placed on it in free (and malloc_consolidate),
+    and taken off (to be either used or placed in bins) in malloc.
 */
-#define top            (bin_at(0)->fd)   /* The topmost chunk */
+/* The otherwise unindexable 1-bin is used to hold unsorted chunks. */
-#define last_remainder (bin_at(1))       /* remainder from last split */
+#define unsorted_chunks(M)          (bin_at(M, 1))
+/*
+  Top
+    The top-most available chunk (i.e., the one bordering the end of
+    available memory) is treated specially. It 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).  Because top initially
+    points to its own bin with initial zero size, thus forcing
+    extension on the first malloc request, we avoid having any special
+    code in malloc to check whether it even exists yet. But we still
+    need to do so when getting memory from system, so we make
+    initial_top treat the bin as a legal but unusable chunk during the
+    interval between initialization and the first call to
+    sYSMALLOc. (This is somewhat delicate, since it relies on
+    the 2 preceding words to be zero during this interval as well.)
+*/
+/* Conveniently, the unsorted bin can be used as dummy top on first call */
+#define initial_top(M)              (unsorted_chunks(M))
 /*
+  Binmap
+    To help compensate for the large number of bins, a one-level index
-   Because top initially points to its own bin with initial
+    structure is used for bin-by-bin searching.  `binmap' is a
+    bitvector recording whether bins are definitely empty so they can
-   zero size, thus forcing extension on the first malloc request,
+    be skipped over during during traversals.  The bits are NOT always
-   we avoid having any special code in malloc to check whether
+    cleared as soon as bins are empty, but instead only
-   it even exists yet. But we still need to in malloc_extend_top.
+    when they are noticed to be empty during traversal in malloc.
 */
+/* Conservatively use 32 bits per map word, even if on 64bit system */
+#define BINMAPSHIFT      5
+#define BITSPERMAP       (1U << BINMAPSHIFT)
-#define initial_top    ((mchunkptr)(bin_at(0)))
+#define BINMAPSIZE       (NBINS / BITSPERMAP)
-/* Helper macro to initialize bins */
+#define idx2block(i)     ((i) >> BINMAPSHIFT)
+#define idx2bit(i)       ((1U << ((i) & ((1U << BINMAPSHIFT)-1))))
+#define mark_bin(m,i)    ((m)->binmap[idx2block(i)] |=  idx2bit(i))
+#define unmark_bin(m,i)  ((m)->binmap[idx2block(i)] &= ~(idx2bit(i)))
-#define IAV(i)  bin_at(i), bin_at(i)
+#define get_binmap(m,i)  ((m)->binmap[idx2block(i)] &   idx2bit(i))
-static mbinptr av_[NAV * 2 + 2] = {
-, 0,
- IAV(0),   IAV(1),   IAV(2),   IAV(3),   IAV(4),   IAV(5),   IAV(6),   IAV(7),
- IAV(8),   IAV(9),   IAV(10),  IAV(11),  IAV(12),  IAV(13),  IAV(14),  IAV(15),
- IAV(16),  IAV(17),  IAV(18),  IAV(19),  IAV(20),  IAV(21),  IAV(22),  IAV(23),
- IAV(24),  IAV(25),  IAV(26),  IAV(27),  IAV(28),  IAV(29),  IAV(30),  IAV(31),
- IAV(32),  IAV(33),  IAV(34),  IAV(35),  IAV(36),  IAV(37),  IAV(38),  IAV(39),
- IAV(40),  IAV(41),  IAV(42),  IAV(43),  IAV(44),  IAV(45),  IAV(46),  IAV(47),
- IAV(48),  IAV(49),  IAV(50),  IAV(51),  IAV(52),  IAV(53),  IAV(54),  IAV(55),
- IAV(56),  IAV(57),  IAV(58),  IAV(59),  IAV(60),  IAV(61),  IAV(62),  IAV(63),
- IAV(64),  IAV(65),  IAV(66),  IAV(67),  IAV(68),  IAV(69),  IAV(70),  IAV(71),
- IAV(72),  IAV(73),  IAV(74),  IAV(75),  IAV(76),  IAV(77),  IAV(78),  IAV(79),
- IAV(80),  IAV(81),  IAV(82),  IAV(83),  IAV(84),  IAV(85),  IAV(86),  IAV(87),
- IAV(88),  IAV(89),  IAV(90),  IAV(91),  IAV(92),  IAV(93),  IAV(94),  IAV(95),
- IAV(96),  IAV(97),  IAV(98),  IAV(99),  IAV(100), IAV(101), IAV(102), IAV(103),
- IAV(104), IAV(105), IAV(106), IAV(107), IAV(108), IAV(109), IAV(110), IAV(111),
- IAV(112), IAV(113), IAV(114), IAV(115), IAV(116), IAV(117), IAV(118), IAV(119),
- IAV(120), IAV(121), IAV(122), IAV(123), IAV(124), IAV(125), IAV(126), IAV(127)
-};
+/*
+  Fastbins
+    An array of lists holding recently freed small chunks.  Fastbins
+    are not doubly linked.  It is faster to single-link them, and
+    since chunks are never removed from the middles of these lists,
+    double linking is not necessary. Also, unlike regular bins, they
+    are not even processed in FIFO order (they use faster LIFO) since
+    ordering doesn't much matter in the transient contexts in which
+    fastbins are normally used.
+    Chunks in fastbins keep their inuse bit set, so they cannot
+    be consolidated with other free chunks. malloc_consolidate
+    releases all chunks in fastbins and consolidates them with
-/* field-extraction macros */
+    other free chunks.
+*/
-#define first(b) ((b)->fd)
+typedef struct malloc_chunk* mfastbinptr;
+/* offset 2 to use otherwise unindexable first 2 bins */
+#define fastbin_index(sz)        ((((unsigned int)(sz)) >> 3) - 2)
+/* The maximum fastbin request size we support */
-#define last(b)  ((b)->bk)
+#define MAX_FAST_SIZE     80
+#define NFASTBINS  (fastbin_index(request2size(MAX_FAST_SIZE))+1)
 /*
+  FASTBIN_CONSOLIDATION_THRESHOLD is the size of a chunk in free()
+  that triggers automatic consolidation of possibly-surrounding
+  fastbin chunks. This is a heuristic, so the exact value should not
+  matter too much. It is defined at half the default trim threshold as a
+  compromise heuristic to only attempt consolidation if it is likely
+  to lead to trimming. However, it is not dynamically tunable, since
+  consolidation reduces fragmentation surrounding loarge chunks even
-  Indexing into bins
+  if trimming is not used.
 */
-#define bin_index(sz)                                                          \
+#define FASTBIN_CONSOLIDATION_THRESHOLD  \
-(((((unsigned long)(sz)) >> 9) ==    0) ?       (((unsigned long)(sz)) >>  3): \
- ((((unsigned long)(sz)) >> 9) <=    4) ?  56 + (((unsigned long)(sz)) >>  6): \
- ((((unsigned long)(sz)) >> 9) <=   20) ?  91 + (((unsigned long)(sz)) >>  9): \
- ((((unsigned long)(sz)) >> 9) <=   84) ? 110 + (((unsigned long)(sz)) >> 12): \
- ((((unsigned long)(sz)) >> 9) <=  340) ? 119 + (((unsigned long)(sz)) >> 15): \
- ((((unsigned long)(sz)) >> 9) <= 1364) ? 124 + (((unsigned long)(sz)) >> 18): \
+  ((unsigned long)(DEFAULT_TRIM_THRESHOLD) >> 1)
-)
 /*
-  bins for chunks < 512 are all spaced 8 bytes apart, and hold
+  Since the lowest 2 bits in max_fast don't matter in size comparisons,
-  identically sized chunks. This is exploited in malloc.
+  they are used as flags.
 */
+/*
-#define MAX_SMALLBIN         63
+  ANYCHUNKS_BIT held in max_fast indicates that there may be any
-#define MAX_SMALLBIN_SIZE   512
+  freed chunks at all. It is set true when entering a chunk into any
-#define SMALLBIN_WIDTH        8
+  bin.
+*/
+#define ANYCHUNKS_BIT        (1U)
-#define smallbin_index(sz)  (((unsigned long)(sz)) >> 3)
+#define have_anychunks(M)     (((M)->max_fast &  ANYCHUNKS_BIT))
+#define set_anychunks(M)      ((M)->max_fast |=  ANYCHUNKS_BIT)
+#define clear_anychunks(M)    ((M)->max_fast &= ~ANYCHUNKS_BIT)
 /*
+  FASTCHUNKS_BIT held in max_fast indicates that there are probably
-   Requests are `small' if both the corresponding and the next bin are small
+  some fastbin chunks. It is set true on entering a chunk into any
+  fastbin, and cleared only in malloc_consolidate.
+*/
+#define FASTCHUNKS_BIT        (2U)
+#define have_fastchunks(M)   (((M)->max_fast &  FASTCHUNKS_BIT))
+#define set_fastchunks(M)    ((M)->max_fast |=  (FASTCHUNKS_BIT|ANYCHUNKS_BIT))
+#define clear_fastchunks(M)  ((M)->max_fast &= ~(FASTCHUNKS_BIT))
+/*
+   Set value of max_fast.
+   Use impossibly small value if 0.
 */
+#define set_max_fast(M, s) \
-#define is_small_request(nb) (nb < MAX_SMALLBIN_SIZE - SMALLBIN_WIDTH)
+  (M)->max_fast = (((s) == 0)? SMALLBIN_WIDTH: request2size(s)) | \
+  ((M)->max_fast &  (FASTCHUNKS_BIT|ANYCHUNKS_BIT))
+#define get_max_fast(M) \
+  ((M)->max_fast & ~(FASTCHUNKS_BIT | ANYCHUNKS_BIT))
 /*
-    To help compensate for the large number of bins, a one-level index
+  morecore_properties is a status word holding dynamically discovered
-    structure is used for bin-by-bin searching.  `binblocks' is a
-    one-word bitvector recording whether groups of BINBLOCKWIDTH bins
-    have any (possibly) non-empty bins, so they can be skipped over
-    all at once during during traversals. The bits are NOT always
+  or controlled properties of the morecore function
-    cleared as soon as all bins in a block are empty, but instead only
-    when all are noticed to be empty during traversal in malloc.
 */
-#define BINBLOCKWIDTH     4   /* bins per block */
+#define MORECORE_CONTIGUOUS_BIT  (1U)
+#define contiguous(M) \
+        (((M)->morecore_properties &  MORECORE_CONTIGUOUS_BIT))
+#define noncontiguous(M) \
-#define binblocks      (bin_at(0)->size) /* bitvector of nonempty blocks */
+        (((M)->morecore_properties &  MORECORE_CONTIGUOUS_BIT) == 0)
+#define set_contiguous(M) \
+        ((M)->morecore_properties |=  MORECORE_CONTIGUOUS_BIT)
+#define set_noncontiguous(M) \
+        ((M)->morecore_properties &= ~MORECORE_CONTIGUOUS_BIT)
-/* bin<->block macros */
+/*
-#define idx2binblock(ix)    ((unsigned)1 << (ix / BINBLOCKWIDTH))
+   ----------- Internal state representation and initialization -----------
-#define mark_binblock(ii)   (binblocks |= idx2binblock(ii))
-#define clear_binblock(ii)  (binblocks &= ~(idx2binblock(ii)))
+*/
+struct malloc_state {
+  /* The maximum chunk size to be eligible for fastbin */
+  INTERNAL_SIZE_T  max_fast;   /* low 2 bits used as flags */
+  /* Fastbins */
+  mfastbinptr      fastbins[NFASTBINS];
-/*  Other static bookkeeping data */
+  /* Base of the topmost chunk -- not otherwise kept in a bin */
+  mchunkptr        top;
-/* variables holding tunable values */
+  /* The remainder from the most recent split of a small request */
+  mchunkptr        last_remainder;
-static unsigned long trim_threshold   = DEFAULT_TRIM_THRESHOLD;
-static unsigned long top_pad          = DEFAULT_TOP_PAD;
+  /* Normal bins packed as described above */
-static unsigned int  n_mmaps_max      = DEFAULT_MMAP_MAX;
+  mchunkptr        bins[NBINS * 2];
-static unsigned long mmap_threshold   = DEFAULT_MMAP_THRESHOLD;
-/* The first value returned from sbrk */
+  /* Bitmap of bins. Trailing zero map handles cases of largest binned size */
-static char* sbrk_base = (char*)(-1);
+  unsigned int     binmap[BINMAPSIZE+1];
-/* The maximum memory obtained from system via sbrk */
+  /* Tunable parameters */
+  CHUNK_SIZE_T     trim_threshold;
+  INTERNAL_SIZE_T  top_pad;
-static unsigned long max_sbrked_mem = 0;
+  INTERNAL_SIZE_T  mmap_threshold;
-/* The maximum via either sbrk or mmap */
+  /* Memory map support */
+  int              n_mmaps;
+  int              n_mmaps_max;
-unsigned long max_total_mem = 0;
+  int              max_n_mmaps;
-/* internal working copy of mallinfo */
+  /* Cache malloc_getpagesize */
-static struct mallinfo current_mallinfo = {  0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
+  unsigned int     pagesize;
-/* The total memory obtained from system via sbrk */
+  /* Track properties of MORECORE */
-#define sbrked_mem  (current_mallinfo.arena)
+  unsigned int     morecore_properties;
-/* Tracking mmaps */
+  /* Statistics */
+  INTERNAL_SIZE_T  mmapped_mem;
+  INTERNAL_SIZE_T  sbrked_mem;
+  INTERNAL_SIZE_T  max_sbrked_mem;
+  INTERNAL_SIZE_T  max_mmapped_mem;
+  INTERNAL_SIZE_T  max_total_mem;
+};
-static unsigned int n_mmaps = 0;
-static unsigned long mmapped_mem = 0;
+typedef struct malloc_state *mstate;
-#ifndef WIN32
-static unsigned int max_n_mmaps = 0;
-static unsigned long max_mmapped_mem = 0;
-#endif
+/*
+   There is exactly one instance of this struct in this malloc.
+   If you are adapting this malloc in a way that does NOT use a static
+   malloc_state, you MUST explicitly zero-fill it before using. This
+   malloc relies on the property that malloc_state is initialized to
+   all zeroes (as is true of C statics).
+*/
+static struct malloc_state av_;  /* never directly referenced */
 /*
-  Debugging support
+   All uses of av_ are via get_malloc_state().
+   At most one "call" to get_malloc_state is made per invocation of
+   the public versions of malloc and free, but other routines
+   that in turn invoke malloc and/or free may call more then once.
+   Also, it is called in check* routines if DEBUG is set.
 */
+#define get_malloc_state() (&(av_))
+/*
+  Initialize a malloc_state struct.
+  This is called only from within malloc_consolidate, which needs
+  be called in the same contexts anyway.  It is never called directly
+  outside of malloc_consolidate because some optimizing compilers try
+  to inline it at all call points, which turns out not to be an
+  optimization at all. (Inlining it in malloc_consolidate is fine though.)
+*/
-#if DEBUG
+#if __STD_C
+static void malloc_init_state(mstate av)
+#else
+static void malloc_init_state(av) mstate av;
+#endif
+{
+  int     i;
+  mbinptr bin;
+  /* Establish circular links for normal bins */
+  for (i = 1; i < NBINS; ++i) {
+    bin = bin_at(av,i);
+    bin->fd = bin->bk = bin;
+  }
+  av->top_pad        = DEFAULT_TOP_PAD;
+  av->n_mmaps_max    = DEFAULT_MMAP_MAX;
+  av->mmap_threshold = DEFAULT_MMAP_THRESHOLD;
+  av->trim_threshold = DEFAULT_TRIM_THRESHOLD;
+#if MORECORE_CONTIGUOUS
+  set_contiguous(av);
+#else
+  set_noncontiguous(av);
+#endif
+  set_max_fast(av, DEFAULT_MXFAST);
+  av->top            = initial_top(av);
+  av->pagesize       = malloc_getpagesize;
+}
+/*
+   Other internal utilities operating on mstates
+*/
+#if __STD_C
+static Void_t*  sYSMALLOc(INTERNAL_SIZE_T, mstate);
+#ifndef MORECORE_CANNOT_TRIM
+static int      sYSTRIm(size_t, mstate);
+#endif
+static void     malloc_consolidate(mstate);
+static Void_t** iALLOc(size_t, size_t*, int, Void_t**);
+#else
+static Void_t*  sYSMALLOc();
+static int      sYSTRIm();
+static void     malloc_consolidate();
+static Void_t** iALLOc();
+#endif
 /*
+  Debugging support
   These routines make a number of assertions about the states
   of data structures that should be true at all times. If any
   are not true, it's very likely that a user program has somehow
   trashed memory. (It's also possible that there is a coding error
   in malloc. In which case, please report it!)
 */
+#if ! DEBUG
+#define check_chunk(P)
+#define check_free_chunk(P)
+#define check_inuse_chunk(P)
+#define check_remalloced_chunk(P,N)
+#define check_malloced_chunk(P,N)
+#define check_malloc_state()
+#else
+#define check_chunk(P)              do_check_chunk(P)
+#define check_free_chunk(P)         do_check_free_chunk(P)
+#define check_inuse_chunk(P)        do_check_inuse_chunk(P)
+#define check_remalloced_chunk(P,N) do_check_remalloced_chunk(P,N)
+#define check_malloced_chunk(P,N)   do_check_malloced_chunk(P,N)
+#define check_malloc_state()        do_check_malloc_state()
+/*
+  Properties of all chunks
+*/
 #if __STD_C
 static void do_check_chunk(mchunkptr p)
 #else
 static void do_check_chunk(p) mchunkptr p;
 #endif
+{
+  mstate av = get_malloc_state();
-  INTERNAL_SIZE_T sz = p->size & ~PREV_INUSE;
+  CHUNK_SIZE_T  sz = chunksize(p);
+  /* min and max possible addresses assuming contiguous allocation */
+  char* max_address = (char*)(av->top) + chunksize(av->top);
+  char* min_address = max_address - av->sbrked_mem;
-  /* No checkable chunk is mmapped */
-  assert(!chunk_is_mmapped(p));
+  if (!chunk_is_mmapped(p)) {
-  /* Check for legal address ... */
+    /* Has legal address ... */
+    if (p != av->top) {
+      if (contiguous(av)) {
-  assert((char*)p >= sbrk_base);
+        assert(((char*)p) >= min_address);
+        assert(((char*)p + sz) <= ((char*)(av->top)));
+      }
+    }
-  if (p != top)
+    else {
+      /* top size is always at least MINSIZE */
-    assert((char*)p + sz <= (char*)top);
+      assert((CHUNK_SIZE_T)(sz) >= MINSIZE);
+      /* top predecessor always marked inuse */
+      assert(prev_inuse(p));
+    }
+  }
-  else
+  else {
+#if HAVE_MMAP
+    /* address is outside main heap  */
+    if (contiguous(av) && av->top != initial_top(av)) {
-    assert((char*)p + sz <= sbrk_base + sbrked_mem);
+      assert(((char*)p) < min_address || ((char*)p) > max_address);
+    }
+    /* chunk is page-aligned */
+    assert(((p->prev_size + sz) & (av->pagesize-1)) == 0);
+    /* mem is aligned */
+    assert(aligned_OK(chunk2mem(p)));
+#else
+    /* force an appropriate assert violation if debug set */
+    assert(!chunk_is_mmapped(p));
+#endif
+  }
+}
+/*
+  Properties of free chunks
+*/
 #if __STD_C
 static void do_check_free_chunk(mchunkptr p)
 #else
 static void do_check_free_chunk(p) mchunkptr p;
 #endif
+{
+  mstate av = get_malloc_state();
   INTERNAL_SIZE_T sz = p->size & ~PREV_INUSE;
   mchunkptr next = chunk_at_offset(p, sz);
   do_check_chunk(p);
-  /* Check whether it claims to be free ... */
+  /* Chunk must claim to be free ... */
   assert(!inuse(p));
+  assert (!chunk_is_mmapped(p));
   /* Unless a special marker, must have OK fields */
-  if ((long)sz >= (long)MINSIZE)
+  if ((CHUNK_SIZE_T)(sz) >= MINSIZE)
+  {
     assert((sz & MALLOC_ALIGN_MASK) == 0);
     assert(aligned_OK(chunk2mem(p)));
     /* ... matching footer field */
     assert(next->prev_size == sz);
     /* ... and is fully consolidated */
     assert(prev_inuse(p));
-    assert (next == top || inuse(next));
+    assert (next == av->top || inuse(next));
     /* ... and has minimally sane links */
     assert(p->fd->bk == p);
     assert(p->bk->fd == p);
+  }
   else /* markers are always of size SIZE_SZ */
     assert(sz == SIZE_SZ);
+}
+/*
+  Properties of inuse chunks
+*/
 #if __STD_C
 static void do_check_inuse_chunk(mchunkptr p)
 #else
 static void do_check_inuse_chunk(p) mchunkptr p;
 #endif
+{
+  mstate av = get_malloc_state();
-  mchunkptr next = next_chunk(p);
+  mchunkptr next;
   do_check_chunk(p);
+  if (chunk_is_mmapped(p))
+    return; /* mmapped chunks have no next/prev */
   /* Check whether it claims to be in use ... */
   assert(inuse(p));
+  next = next_chunk(p);
   /* ... and is surrounded by OK chunks.
     Since more things can be checked with free chunks than inuse ones,
     if an inuse chunk borders them and debug is on, it's worth doing them.
   */
-  if (!prev_inuse(p))
+  if (!prev_inuse(p))  {
-  {
+    /* Note that we cannot even look at prev unless it is not inuse */
     mchunkptr prv = prev_chunk(p);
     assert(next_chunk(prv) == p);
     do_check_free_chunk(prv);
+  }
-  if (next == top)
+  if (next == av->top) {
-  {
     assert(prev_inuse(next));
     assert(chunksize(next) >= MINSIZE);
+  }
   else if (!inuse(next))
     do_check_free_chunk(next);
+}
+/*
+  Properties of chunks recycled from fastbins
+*/
 #if __STD_C
-static void do_check_malloced_chunk(mchunkptr p, INTERNAL_SIZE_T s)
+static void do_check_remalloced_chunk(mchunkptr p, INTERNAL_SIZE_T s)
 #else
-static void do_check_malloced_chunk(p, s) mchunkptr p; INTERNAL_SIZE_T s;
+static void do_check_remalloced_chunk(p, s) mchunkptr p; INTERNAL_SIZE_T s;
 #endif
+{
   INTERNAL_SIZE_T sz = p->size & ~PREV_INUSE;
-  long room = sz - s;
   do_check_inuse_chunk(p);
   /* Legal size ... */
-  assert((long)sz >= (long)MINSIZE);
   assert((sz & MALLOC_ALIGN_MASK) == 0);
-  assert(room >= 0);
-  assert(room < (long)MINSIZE);
+  assert((CHUNK_SIZE_T)(sz) >= MINSIZE);
   /* ... and alignment */
   assert(aligned_OK(chunk2mem(p)));
-  /* ... and was allocated at front of an available chunk */
+  /* chunk is less than MINSIZE more than request */
-  assert(prev_inuse(p));
+  assert((long)(sz) - (long)(s) >= 0);
+  assert((long)(sz) - (long)(s + MINSIZE) < 0);
+}
+/*
+  Properties of nonrecycled chunks at the point they are malloced
+*/
-#define check_free_chunk(P)  do_check_free_chunk(P)
-#define check_inuse_chunk(P) do_check_inuse_chunk(P)
-#define check_chunk(P) do_check_chunk(P)
+#if __STD_C
-#define check_malloced_chunk(P,N) do_check_malloced_chunk(P,N)
+static void do_check_malloced_chunk(mchunkptr p, INTERNAL_SIZE_T s)
 #else
-#define check_free_chunk(P)
-#define check_inuse_chunk(P)
-#define check_chunk(P)
-#define check_malloced_chunk(P,N)
+static void do_check_malloced_chunk(p, s) mchunkptr p; INTERNAL_SIZE_T s;
 #endif
+{
+  /* same as recycled case ... */
+  do_check_remalloced_chunk(p, s);
+  /*
+    ... plus,  must obey implementation invariant that prev_inuse is
+    always true of any allocated chunk; i.e., that 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.  This does not necessarily hold however for chunks
+    recycled via fastbins.
+  */
-/*
-  Macro-based internal utilities
+  assert(prev_inuse(p));
-*/
+}
 /*
-  Linking chunks in bin lists.
+  Properties of malloc_state.
-  Call these only with variables, not arbitrary expressions, as arguments.
-*/
-/*
-  Place chunk p of size s in its bin, in size order,
-  putting it ahead of others of same size.
-*/
+  This may be useful for debugging malloc, as well as detecting user
+  programmer errors that somehow write into malloc_state.
-#define frontlink(P, S, IDX, BK, FD)                                          \
-{                                                                             \
-  if (S < MAX_SMALLBIN_SIZE)                                                  \
-  {                                                                           \
-    IDX = smallbin_index(S);                                                  \
+  If you are extending or experimenting with this malloc, you can
-    mark_binblock(IDX);                                                       \
+  probably figure out how to hack this routine to print out or
-    BK = bin_at(IDX);                                                         \
-    FD = BK->fd;                                                              \
-    P->bk = BK;                                                               \
-    P->fd = FD;                                                               \
-    FD->bk = BK->fd = P;                                                      \
-  }                                                                           \
-  else                                                                        \
-  {                                                                           \
-    IDX = bin_index(S);                                                       \
-    BK = bin_at(IDX);                                                         \
-    FD = BK->fd;                                                              \
-    if (FD == BK) mark_binblock(IDX);                                         \
-    else                                                                      \
-    {                                                                         \
-      while (FD != BK && S < chunksize(FD)) FD = FD->fd;                      \
+  display chunk addresses, sizes, bins, and other instrumentation.
-      BK = FD->bk;                                                            \
-    }                                                                         \
-    P->bk = BK;                                                               \
-    P->fd = FD;                                                               \
-    FD->bk = BK->fd = P;                                                      \
-  }                                                                           \
-}
+*/
-/* take a chunk off a list */
-#define unlink(P, BK, FD)                                                     \
+static void do_check_malloc_state()
+{
-{                                                                             \
+  mstate av = get_malloc_state();
+  int i;
+  mchunkptr p;
+  mchunkptr q;
+  mbinptr b;
-  BK = P->bk;                                                                 \
+  unsigned int binbit;
+  int empty;
-  FD = P->fd;                                                                 \
+  unsigned int idx;
-  FD->bk = BK;                                                                \
+  INTERNAL_SIZE_T size;
-  BK->fd = FD;                                                                \
+  CHUNK_SIZE_T  total = 0;
-}                                                                             \
+  int max_fast_bin;
-/* Place p as the last remainder */
+  /* internal size_t must be no wider than pointer type */
+  assert(sizeof(INTERNAL_SIZE_T) <= sizeof(char*));
-#define link_last_remainder(P)                                                \
-{                                                                             \
-  last_remainder->fd = last_remainder->bk =  P;                               \
+  /* alignment is a power of 2 */
-  P->fd = P->bk = last_remainder;                                             \
+  assert((MALLOC_ALIGNMENT & (MALLOC_ALIGNMENT-1)) == 0);
-}
-/* Clear the last_remainder bin */
+  /* cannot run remaining checks until fully initialized */
+  if (av->top == 0 || av->top == initial_top(av))
+    return;
-#define clear_last_remainder \
+  /* pagesize is a power of 2 */
-  (last_remainder->fd = last_remainder->bk = last_remainder)
+  assert((av->pagesize & (av->pagesize-1)) == 0);
+  /* properties of fastbins */
+  /* max_fast is in allowed range */
+  assert(get_max_fast(av) <= request2size(MAX_FAST_SIZE));
+  max_fast_bin = fastbin_index(av->max_fast);
+  for (i = 0; i < NFASTBINS; ++i) {
+    p = av->fastbins[i];
+    /* all bins past max_fast are empty */
+    if (i > max_fast_bin)
+      assert(p == 0);
+    while (p != 0) {
+      /* each chunk claims to be inuse */
+      do_check_inuse_chunk(p);
+      total += chunksize(p);
+      /* chunk belongs in this bin */
+      assert(fastbin_index(chunksize(p)) == i);
+      p = p->fd;
+    }
+  }
+  if (total != 0)
+    assert(have_fastchunks(av));
+  else if (!have_fastchunks(av))
+    assert(total == 0);
+  /* check normal bins */
+  for (i = 1; i < NBINS; ++i) {
+    b = bin_at(av,i);
+    /* binmap is accurate (except for bin 1 == unsorted_chunks) */
+    if (i >= 2) {
+      binbit = get_binmap(av,i);
+      empty = last(b) == b;
+      if (!binbit)
+        assert(empty);
+      else if (!empty)
+        assert(binbit);
+    }
+    for (p = last(b); p != b; p = p->bk) {
-/* Routines dealing with mmap(). */
+      /* each chunk claims to be free */
+      do_check_free_chunk(p);
+      size = chunksize(p);
+      total += size;
+      if (i >= 2) {
+        /* chunk belongs in bin */
+        idx = bin_index(size);
+        assert(idx == i);
+        /* lists are sorted */
+        if ((CHUNK_SIZE_T) size >= (CHUNK_SIZE_T)(FIRST_SORTED_BIN_SIZE)) {
+          assert(p->bk == b ||
+                 (CHUNK_SIZE_T)chunksize(p->bk) >=
+                 (CHUNK_SIZE_T)chunksize(p));
+        }
+      }
+      /* chunk is followed by a legal chain of inuse chunks */
+      for (q = next_chunk(p);
+           (q != av->top && inuse(q) &&
+             (CHUNK_SIZE_T)(chunksize(q)) >= MINSIZE);
+           q = next_chunk(q))
+        do_check_inuse_chunk(q);
+    }
+  }
-#if HAVE_MMAP
+  /* top chunk is OK */
+  check_chunk(av->top);
-#if __STD_C
-static mchunkptr mmap_chunk(size_t size)
-#else
-static mchunkptr mmap_chunk(size) size_t size;
-#endif
-{
-  size_t page_mask = malloc_getpagesize - 1;
+  /* sanity checks for statistics */
-  mchunkptr p;
+  assert(total <= (CHUNK_SIZE_T)(av->max_total_mem));
-#ifndef MAP_ANONYMOUS
+  assert(av->n_mmaps >= 0);
-  static int fd = -1;
+  assert(av->n_mmaps <= av->max_n_mmaps);
-#endif
+  assert((CHUNK_SIZE_T)(av->sbrked_mem) <=
-  if(n_mmaps >= n_mmaps_max) return 0; /* too many regions */
+         (CHUNK_SIZE_T)(av->max_sbrked_mem));
-  /* For mmapped chunks, the overhead is one SIZE_SZ unit larger, because
+  assert((CHUNK_SIZE_T)(av->mmapped_mem) <=
-   * there is no following chunk whose prev_size field could be used.
-   */
-  size = (size + SIZE_SZ + page_mask) & ~page_mask;
+         (CHUNK_SIZE_T)(av->max_mmapped_mem));
-#ifdef MAP_ANONYMOUS
-  p = (mchunkptr)mmap(0, size, PROT_READ|PROT_WRITE,
+  assert((CHUNK_SIZE_T)(av->max_total_mem) >=
-		      MAP_PRIVATE|MAP_ANONYMOUS, -1, 0);
+         (CHUNK_SIZE_T)(av->mmapped_mem) + (CHUNK_SIZE_T)(av->sbrked_mem));
-#else /* !MAP_ANONYMOUS */
-  if (fd < 0)
-  {
-    fd = open("/dev/zero", O_RDWR);
-    if(fd < 0) return 0;
-  }
+}
-  p = (mchunkptr)mmap(0, size, PROT_READ|PROT_WRITE, MAP_PRIVATE, fd, 0);
 #endif
-  if(p == (mchunkptr)-1) return 0;
-  n_mmaps++;
-  if (n_mmaps > max_n_mmaps) max_n_mmaps = n_mmaps;
+/* ----------- Routines dealing with system allocation -------------- */
-  /* We demand that eight bytes into a page must be 8-byte aligned. */
-  assert(aligned_OK(chunk2mem(p)));
+/*
-  /* The offset to the start of the mmapped region is stored
+  sysmalloc handles malloc cases requiring more memory from the system.
-   * in the prev_size field of the chunk; normally it is zero,
+  On entry, it is assumed that av->top does not have enough
-   * but that can be changed in memalign().
+  space to service request for nb bytes, thus requiring that av->top
-   */
-  p->prev_size = 0;
-  set_head(p, size|IS_MMAPPED);
-  mmapped_mem += size;
+  be extended or replaced.
-  if ((unsigned long)mmapped_mem > (unsigned long)max_mmapped_mem)
-    max_mmapped_mem = mmapped_mem;
-  if ((unsigned long)(mmapped_mem + sbrked_mem) > (unsigned long)max_total_mem)
-    max_total_mem = mmapped_mem + sbrked_mem;
-  return p;
-}
+*/
 #if __STD_C
-static void munmap_chunk(mchunkptr p)
+static Void_t* sYSMALLOc(INTERNAL_SIZE_T nb, mstate av)
 #else
-static void munmap_chunk(p) mchunkptr p;
+static Void_t* sYSMALLOc(nb, av) INTERNAL_SIZE_T nb; mstate av;
 #endif
+{
-  INTERNAL_SIZE_T size = chunksize(p);
-  int ret;
-  assert (chunk_is_mmapped(p));
-  assert(! ((char*)p >= sbrk_base && (char*)p < sbrk_base + sbrked_mem));
+  mchunkptr       old_top;        /* incoming value of av->top */
-  assert((n_mmaps > 0));
+  INTERNAL_SIZE_T old_size;       /* its size */
-  assert(((p->prev_size + size) & (malloc_getpagesize-1)) == 0);
+  char*           old_end;        /* its end address */
-  n_mmaps--;
+  long            size;           /* arg to first MORECORE or mmap call */
-  mmapped_mem -= (size + p->prev_size);
+  char*           brk;            /* return value from MORECORE */
-  ret = munmap((char *)p - p->prev_size, size + p->prev_size);
+  long            correction;     /* arg to 2nd MORECORE call */
+  char*           snd_brk;        /* 2nd return val */
-  /* munmap returns non-zero on failure */
+  INTERNAL_SIZE_T front_misalign; /* unusable bytes at front of new space */
+  INTERNAL_SIZE_T end_misalign;   /* partial page left at end of new space */
-  assert(ret == 0);
+  char*           aligned_brk;    /* aligned offset into brk */
-}
+  mchunkptr       p;              /* the allocated/returned chunk */
+  mchunkptr       remainder;      /* remainder from allocation */
-#if HAVE_MREMAP
+  CHUNK_SIZE_T    remainder_size; /* its size */
-#if __STD_C
-static mchunkptr mremap_chunk(mchunkptr p, size_t new_size)
+  CHUNK_SIZE_T    sum;            /* for updating stats */
-#else
-static mchunkptr mremap_chunk(p, new_size) mchunkptr p; size_t new_size;
-#endif
-{
-  size_t page_mask = malloc_getpagesize - 1;
-  INTERNAL_SIZE_T offset = p->prev_size;
-  INTERNAL_SIZE_T size = chunksize(p);
-  char *cp;
-  assert (chunk_is_mmapped(p));
-  assert(! ((char*)p >= sbrk_base && (char*)p < sbrk_base + sbrked_mem));
-  assert((n_mmaps > 0));
-  assert(((size + offset) & (malloc_getpagesize-1)) == 0);
+  size_t          pagemask  = av->pagesize - 1;
+  /*
-  /* Note the extra SIZE_SZ overhead as in mmap_chunk(). */
+    If there is space available in fastbins, consolidate and retry
+    malloc from scratch rather than getting memory from system.  This
-  new_size = (new_size + offset + SIZE_SZ + page_mask) & ~page_mask;
+    can occur only if nb is in smallbin range so we didn't consolidate
-  cp = (char *)mremap((char *)p - offset, size + offset, new_size, 1);
+    upon entry to malloc. It is much easier to handle this case here
+    than in malloc proper.
+  */
-  if (cp == (char *)-1) return 0;
+  if (have_fastchunks(av)) {
+    assert(in_smallbin_range(nb));
+    malloc_consolidate(av);
+    return mALLOc(nb - MALLOC_ALIGN_MASK);
+  }
-  p = (mchunkptr)(cp + offset);
-  assert(aligned_OK(chunk2mem(p)));
+#if HAVE_MMAP
+  /*
+    If have mmap, and the request size meets the mmap threshold, and
-  assert((p->prev_size == offset));
+    the system supports mmap, and there are few enough currently
-  set_head(p, (new_size - offset)|IS_MMAPPED);
+    allocated mmapped regions, try to directly map this request
+    rather than expanding top.
+  */
-  mmapped_mem -= size + offset;
-  mmapped_mem += new_size;
-  if ((unsigned long)mmapped_mem > (unsigned long)max_mmapped_mem)
+  if ((CHUNK_SIZE_T)(nb) >= (CHUNK_SIZE_T)(av->mmap_threshold) &&
-    max_mmapped_mem = mmapped_mem;
+      (av->n_mmaps < av->n_mmaps_max)) {
-  if ((unsigned long)(mmapped_mem + sbrked_mem) > (unsigned long)max_total_mem)
-    max_total_mem = mmapped_mem + sbrked_mem;
-  return p;
-}
-#endif /* HAVE_MREMAP */
+    char* mm;             /* return value from mmap call*/
+    /*
+      Round up size to nearest page.  For mmapped chunks, the overhead
+      is one SIZE_SZ unit larger than for normal chunks, because there
+      is no following chunk whose prev_size field could be used.
-#endif /* HAVE_MMAP */
+    */
+    size = (nb + SIZE_SZ + MALLOC_ALIGN_MASK + pagemask) & ~pagemask;
+    /* Don't try if size wraps around 0 */
+    if ((CHUNK_SIZE_T)(size) > (CHUNK_SIZE_T)(nb)) {
+      mm = (char*)(MMAP(0, size, PROT_READ|PROT_WRITE, MAP_PRIVATE));
+      if (mm != (char*)(MORECORE_FAILURE)) {
+        /*
+          The offset to the start of the mmapped region is stored
+          in the prev_size field of the chunk. This allows us to adjust
+          returned start address to meet alignment requirements here
+          and in memalign(), and still be able to compute proper
+          address argument for later munmap in free() and realloc().
+        */
+        front_misalign = (INTERNAL_SIZE_T)chunk2mem(mm) & MALLOC_ALIGN_MASK;
+        if (front_misalign > 0) {
+          correction = MALLOC_ALIGNMENT - front_misalign;
+          p = (mchunkptr)(mm + correction);
+          p->prev_size = correction;
+          set_head(p, (size - correction) |IS_MMAPPED);
+        }
+        else {
+          p = (mchunkptr)mm;
+          p->prev_size = 0;
+          set_head(p, size|IS_MMAPPED);
+        }
+        /* update statistics */
+        if (++av->n_mmaps > av->max_n_mmaps)
+          av->max_n_mmaps = av->n_mmaps;
+        sum = av->mmapped_mem += size;
+        if (sum > (CHUNK_SIZE_T)(av->max_mmapped_mem))
+          av->max_mmapped_mem = sum;
+        sum += av->sbrked_mem;
+        if (sum > (CHUNK_SIZE_T)(av->max_total_mem))
+          av->max_total_mem = sum;
+		/* R Memory Statistics */
+		max_total_mem = av->max_total_mem;
+        check_chunk(p);
+        return chunk2mem(p);
+      }
+    }
+  }
+#endif
-/*
-  Extend the top-most chunk by obtaining memory from system.
+  /* Record incoming configuration of top */
-  Main interface to sbrk (but see also malloc_trim).
-*/
-#if __STD_C
-static void malloc_extend_top(INTERNAL_SIZE_T nb)
+  old_top  = av->top;
-#else
-static void malloc_extend_top(nb) INTERNAL_SIZE_T nb;
+  old_size = chunksize(old_top);
-#endif
-{
-  char*     brk;                  /* return value from sbrk */
+  old_end  = (char*)(chunk_at_offset(old_top, old_size));
-  INTERNAL_SIZE_T front_misalign; /* unusable bytes at front of sbrked space */
-  INTERNAL_SIZE_T correction;     /* bytes for 2nd sbrk call */
-  char*     new_brk;              /* return of 2nd sbrk call */
-  INTERNAL_SIZE_T top_size;       /* new size of top chunk */
-  mchunkptr old_top     = top;  /* Record state of old top */
-  INTERNAL_SIZE_T old_top_size = chunksize(old_top);
-  char*     old_end      = (char*)(chunk_at_offset(old_top, old_top_size));
+  brk = snd_brk = (char*)(MORECORE_FAILURE);
+  /*
-  /* Pad request with top_pad plus minimal overhead */
+     If not the first time through, we require old_size to be
+     at least MINSIZE and to have prev_inuse set.
+  */
+  assert((old_top == initial_top(av) && old_size == 0) ||
-  INTERNAL_SIZE_T    sbrk_size     = nb + top_pad + MINSIZE;
+         ((CHUNK_SIZE_T) (old_size) >= MINSIZE &&
-  unsigned long pagesz    = malloc_getpagesize;
+          prev_inuse(old_top)));
-  /* If not the first time through, round to preserve page boundary */
-  /* Otherwise, we need to correct to a page size below anyway. */
+  /* Precondition: not enough current space to satisfy nb request */
-  /* (We also correct below if an intervening foreign sbrk call.) */
+  assert((CHUNK_SIZE_T)(old_size) < (CHUNK_SIZE_T)(nb + MINSIZE));
-  if (sbrk_base != (char*)(-1))
+  /* Precondition: all fastbins are consolidated */
-    sbrk_size = (sbrk_size + (pagesz - 1)) & ~(pagesz - 1);
+  assert(!have_fastchunks(av));
-  brk = (char*)(MORECORE (sbrk_size));
-  /* Fail if sbrk failed or if a foreign sbrk call killed our space */
+  /* Request enough space for nb + pad + overhead */
-  if (brk == (char*)(MORECORE_FAILURE) ||
-      (brk < old_end && old_top != initial_top))
-    return;
-  sbrked_mem += sbrk_size;
+  size = nb + av->top_pad + MINSIZE;
-  if (brk == old_end) /* can just add bytes to current top */
-  {
-    top_size = sbrk_size + old_top_size;
-    set_head(top, top_size | PREV_INUSE);
-  }
+  /*
-  else
-  {
-    if (sbrk_base == (char*)(-1))  /* First time through. Record base */
+    If contiguous, we can subtract out existing space that we hope to
-      sbrk_base = brk;
-    else  /* Someone else called sbrk().  Count those bytes as sbrked_mem. */
+    combine with new space. We add it back later only if
-      sbrked_mem += brk - (char*)old_end;
-    /* Guarantee alignment of first new chunk made from this space */
+    we don't actually get contiguous space.
-    front_misalign = (unsigned long)chunk2mem(brk) & MALLOC_ALIGN_MASK;
-    if (front_misalign > 0)
-    {
-      correction = (MALLOC_ALIGNMENT) - front_misalign;
-      brk += correction;
-    }
+  */
-    else
-      correction = 0;
+  if (contiguous(av))
-    /* Guarantee the next brk will be at a page boundary */
+    size -= old_size;
+  /*
+    Round to a multiple of page size.
+    If MORECORE is not contiguous, this ensures that we only call it
+    with whole-page arguments.  And if MORECORE is contiguous and
-    correction += ((((unsigned long)(brk + sbrk_size))+(pagesz-1)) &
+    this is not first time through, this preserves page-alignment of
-                   ~(pagesz - 1)) - ((unsigned long)(brk + sbrk_size));
+    previous calls. Otherwise, we correct to page-align below.
+  */
-    /* Allocate correction */
-    new_brk = (char*)(MORECORE (correction));
+  size = (size + pagemask) & ~pagemask;
-    if (new_brk == (char*)(MORECORE_FAILURE)) return;
+  /*
+    Don't try to call MORECORE if argument is so big as to appear
+    negative. Note that since mmap takes size_t arg, it may succeed
-    sbrked_mem += correction;
+    below even if we cannot call MORECORE.
+  */
-    top = (mchunkptr)brk;
+  if (size > 0)
-    top_size = new_brk - brk + correction;
-    set_head(top, top_size | PREV_INUSE);
+    brk = (char*)(MORECORE(size));
+  /*
+    If have mmap, try using it as a backup when MORECORE fails or
+    cannot be used. This is worth doing on systems that have "holes" in
+    address space, so sbrk cannot extend to give contiguous space, but
+    space is available elsewhere.  Note that we ignore mmap max count
+    and threshold limits, since the space will not be used as a
-    if (old_top != initial_top)
+    segregated mmap region.
-    {
+  */
-      /* There must have been an intervening foreign sbrk call. */
+#if HAVE_MMAP
-      /* A double fencepost is necessary to prevent consolidation */
+  if (brk == (char*)(MORECORE_FAILURE)) {
-      /* If not enough space to do this, then user did something very wrong */
+    /* Cannot merge with old top, so add its size back in */
+    if (contiguous(av))
+      size = (size + old_size + pagemask) & ~pagemask;
+    /* If we are relying on mmap as backup, then use larger units */
+    if ((CHUNK_SIZE_T)(size) < (CHUNK_SIZE_T)(MMAP_AS_MORECORE_SIZE))
-      if (old_top_size < MINSIZE)
+      size = MMAP_AS_MORECORE_SIZE;
+    /* Don't try if size wraps around 0 */
+    if ((CHUNK_SIZE_T)(size) > (CHUNK_SIZE_T)(nb)) {
+      brk = (char*)(MMAP(0, size, PROT_READ|PROT_WRITE, MAP_PRIVATE));
-      {
+      if (brk != (char*)(MORECORE_FAILURE)) {
-        set_head(top, PREV_INUSE); /* will force null return from malloc */
+        /* We do not need, and cannot use, another sbrk call to find end */
+        snd_brk = brk + size;
-        return;
+        /*
+           Record that we no longer have a contiguous sbrk region.
+           After the first time mmap is used as backup, we do not
+           ever rely on contiguous space since this could incorrectly
+           bridge regions.
+        */
+        set_noncontiguous(av);
+      }
-      /* Also keep size a multiple of MALLOC_ALIGNMENT */
-      old_top_size = (old_top_size - 3*SIZE_SZ) & ~MALLOC_ALIGN_MASK;
-      set_head_size(old_top, old_top_size);
-      chunk_at_offset(old_top, old_top_size          )->size =
-        SIZE_SZ|PREV_INUSE;
-      chunk_at_offset(old_top, old_top_size + SIZE_SZ)->size =
-        SIZE_SZ|PREV_INUSE;
-      /* If possible, release the rest. */
-      if (old_top_size >= MINSIZE)
-        fREe(chunk2mem(old_top));
+    }
+  }
+#endif
-  if ((unsigned long)sbrked_mem > (unsigned long)max_sbrked_mem)
+  if (brk != (char*)(MORECORE_FAILURE)) {
-    max_sbrked_mem = sbrked_mem;
+    av->sbrked_mem += size;
-  if ((unsigned long)(mmapped_mem + sbrked_mem) > (unsigned long)max_total_mem)
-    max_total_mem = mmapped_mem + sbrked_mem;
-  /* We always land on a page boundary */
-  assert(((unsigned long)((char*)top + top_size) & (pagesz - 1)) == 0);
-}
+    /*
+      If MORECORE extends previous space, we can likewise extend top size.
+    */
+    if (brk == old_end && snd_brk == (char*)(MORECORE_FAILURE)) {
-/* Main public routines */
+      set_head(old_top, (size + old_size) | PREV_INUSE);
+    }
+    /*
+      Otherwise, make adjustments:
+      * If the first time through or noncontiguous, we need to call sbrk
+        just to find out where the end of memory lies.
+      * We need to ensure that all returned chunks from malloc will meet
+        MALLOC_ALIGNMENT
+      * If there was an intervening foreign sbrk, we need to adjust sbrk
+        request size to account for fact that we will not be able to
+        combine new space with existing space in old_top.
+      * Almost all systems internally allocate whole pages at a time, in
+        which case we might as well use the whole last page of request.
+        So we allocate enough more memory to hit a page boundary now,
+        which in turn causes future contiguous calls to page-align.
+    */
+    else {
+      front_misalign = 0;
+      end_misalign = 0;
+      correction = 0;
+      aligned_brk = brk;
+      /*
+        If MORECORE returns an address lower than we have seen before,
+        we know it isn't really contiguous.  This and some subsequent
+        checks help cope with non-conforming MORECORE functions and
+        the presence of "foreign" calls to MORECORE from outside of
+        malloc or by other threads.  We cannot guarantee to detect
+        these in all cases, but cope with the ones we do detect.
+      */
+      if (contiguous(av) && old_size != 0 && brk < old_end) {
+        set_noncontiguous(av);
+      }
+      /* handle contiguous cases */
+      if (contiguous(av)) {
-/*
+        /*
+           We can tolerate forward non-contiguities here (usually due
+           to foreign calls) but treat them as part of our space for
-  Malloc Algorthim:
+           stats reporting.
+        */
+        if (old_size != 0)
+          av->sbrked_mem += brk - old_end;
+        /* Guarantee alignment of first new chunk made from this space */
+        front_misalign = (INTERNAL_SIZE_T)chunk2mem(brk) & MALLOC_ALIGN_MASK;
+        if (front_misalign > 0) {
+          /*
+            Skip over some bytes to arrive at an aligned position.
+            We don't need to specially mark these wasted front bytes.
+            They will never be accessed anyway because
+            prev_inuse of av->top (and any chunk created from its start)
+            is always true after initialization.
+          */
+          correction = MALLOC_ALIGNMENT - front_misalign;
+          aligned_brk += correction;
+        }
+        /*
-    The requested size is first converted into a usable form, `nb'.
+          If this isn't adjacent to existing space, then we will not
-    This currently means to add 4 bytes overhead plus possibly more to
+          be able to merge with old_top space, so must add to 2nd request.
+        */
+        correction += old_size;
-    obtain 8-byte alignment and/or to obtain a size of at least
+        /* Extend the end address to hit a page boundary */
+        end_misalign = (INTERNAL_SIZE_T)(brk + size + correction);
+        correction += ((end_misalign + pagemask) & ~pagemask) - end_misalign;
+        assert(correction >= 0);
+        snd_brk = (char*)(MORECORE(correction));
+        if (snd_brk == (char*)(MORECORE_FAILURE)) {
+          /*
+            If can't allocate correction, try to at least find out current
-    MINSIZE (currently 16 bytes), the smallest allocatable size.
+            brk.  It might be enough to proceed without failing.
+          */
+          correction = 0;
+          snd_brk = (char*)(MORECORE(0));
+        }
+        else if (snd_brk < brk) {
+          /*
+            If the second call gives noncontiguous space even though
+            it says it won't, the only course of action is to ignore
+            results of second call, and conservatively estimate where
-    (All fits are considered `exact' if they are within MINSIZE bytes.)
+            the first call left us. Also set noncontiguous, so this
+            won't happen again, leaving at most one hole.
+            Note that this check is intrinsically incomplete.  Because
+            MORECORE is allowed to give more space than we ask for,
+            there is no reliable way to detect a noncontiguity
+            producing a forward gap for the second call.
+          */
+          snd_brk = brk + size;
+          correction = 0;
+          set_noncontiguous(av);
+        }
+      }
+      /* handle non-contiguous cases */
+      else {
+        /* MORECORE/mmap must correctly align */
+        assert(aligned_OK(chunk2mem(brk)));
+        /* Find out current end of memory */
+        if (snd_brk == (char*)(MORECORE_FAILURE)) {
+          snd_brk = (char*)(MORECORE(0));
+          av->sbrked_mem += snd_brk - brk - size;
+        }
+      }
+      /* Adjust top based on results of second sbrk */
+      if (snd_brk != (char*)(MORECORE_FAILURE)) {
+        av->top = (mchunkptr)aligned_brk;
+        set_head(av->top, (snd_brk - aligned_brk + correction) | PREV_INUSE);
+        av->sbrked_mem += correction;
+        /*
-    From there, the first successful of the following steps is taken:
+          If not the first time through, we either have a
+          gap due to foreign sbrk or a non-contiguous region.  Insert a
+          double fencepost at old_top to prevent consolidation with space
+          we don't own. These fenceposts are artificial chunks that are
+          marked as inuse and are in any case too small to use.  We need
+          two to make sizes and alignments work out.
+        */
+        if (old_size != 0) {
+          /*
+             Shrink old_top to insert fenceposts, keeping size a
+             multiple of MALLOC_ALIGNMENT. We know there is at least
+             enough space in old_top to do this.
+          */
+          old_size = (old_size - 3*SIZE_SZ) & ~MALLOC_ALIGN_MASK;
+          set_head(old_top, old_size | PREV_INUSE);
+          /*
+            Note that the following assignments completely overwrite
+            old_top when old_size was previously MINSIZE.  This is
+            intentional. We need the fencepost, even if old_top otherwise gets
+            lost.
+          */
+          chunk_at_offset(old_top, old_size          )->size =
+            SIZE_SZ|PREV_INUSE;
+          chunk_at_offset(old_top, old_size + SIZE_SZ)->size =
+            SIZE_SZ|PREV_INUSE;
+          /*
+             If possible, release the rest, suppressing trimming.
+          */
+          if (old_size >= MINSIZE) {
+            INTERNAL_SIZE_T tt = av->trim_threshold;
+            av->trim_threshold = (INTERNAL_SIZE_T)(-1);
+            fREe(chunk2mem(old_top));
+            av->trim_threshold = tt;
+          }
+        }
+      }
+    }
+    /* Update statistics */
+    sum = av->sbrked_mem;
+    if (sum > (CHUNK_SIZE_T)(av->max_sbrked_mem))
+      av->max_sbrked_mem = sum;
+    sum += av->mmapped_mem;
+    if (sum > (CHUNK_SIZE_T)(av->max_total_mem))
+      av->max_total_mem = sum;
+	/* R Memory Statistics */
+	max_total_mem = av->max_total_mem;
+    check_malloc_state();
+    /* finally, do the allocation */
+    p = av->top;
+    size = chunksize(p);
+    /* check that one of the above allocation paths succeeded */
+    if ((CHUNK_SIZE_T)(size) >= (CHUNK_SIZE_T)(nb + MINSIZE)) {
+      remainder_size = size - nb;
+      remainder = chunk_at_offset(p, nb);
+      av->top = remainder;
+      set_head(p, nb | PREV_INUSE);
+      set_head(remainder, remainder_size | PREV_INUSE);
+      check_malloced_chunk(p, nb);
+      return chunk2mem(p);
+    }
-. The bin corresponding to the request size is scanned, and if
-         a chunk of exactly the right size is found, it is taken.
+  }
-. The most recently remaindered chunk is used if it is big
-         enough.  This is a form of (roving) first fit, used only in
-         the absence of exact fits. Runs of consecutive requests use
-         the remainder of the chunk used for the previous such request
-         whenever possible. This limited use of a first-fit style
+  /* catch all failure paths */
-         allocation strategy tends to give contiguous chunks
+  MALLOC_FAILURE_ACTION;
-         coextensive lifetimes, which improves locality and can reduce
-         fragmentation in the long run.
+  return 0;
+}
-. Other bins are scanned in increasing size order, using a
-         chunk big enough to fulfill the request, and splitting off
-         any remainder.  This search is strictly by best-fit; i.e.,
-         the smallest (with ties going to approximately the least
-         recently used) chunk that fits is selected.
-. If large enough, the chunk bordering the end of memory
-         (`top') is split off. (This use of `top' is in accord with
-         the best-fit search rule.  In effect, `top' is treated as
-         larger (and thus less well fitting) than any other available
-         chunk since it can be extended to be as large as necessary
-         (up to system limitations).
-. If the request size meets the mmap threshold and the
-         system supports mmap, and there are few enough currently
-         allocated mmapped regions, and a call to mmap succeeds,
-         the request is allocated via direct memory mapping.
-. Otherwise, the top of memory is extended by
+#ifndef MORECORE_CANNOT_TRIM
+/*
-         obtaining more space from the system (normally using sbrk,
+  sYSTRIm is an inverse of sorts to sYSMALLOc.  It gives memory back
-         but definable to anything else via the MORECORE macro).
+  to the system (via negative arguments to sbrk) if there is unused
-         Memory is gathered from the system (in system page-sized
+  memory at the `high' end of the malloc pool. It is called
-         units) in a way that allows chunks obtained across different
+  automatically by free() when top space exceeds the trim
-         sbrk calls to be consolidated, but does not require
+  threshold. It is also called by the public malloc_trim routine.  It
-         contiguous memory. Thus, it should be safe to intersperse
+  returns 1 if it actually released any memory, else 0.
-         mallocs with other sbrk calls.
+*/
+#if __STD_C
+static int sYSTRIm(size_t pad, mstate av)
+#else
+static int sYSTRIm(pad, av) size_t pad; mstate av;
+#endif
+{
+  long  top_size;        /* Amount of top-most memory */
+  long  extra;           /* Amount to release */
+  long  released;        /* Amount actually released */
+  char* current_brk;     /* address returned by pre-check sbrk call */
+  char* new_brk;         /* address returned by post-check sbrk call */
+  size_t pagesz;
+  pagesz = av->pagesize;
+  top_size = chunksize(av->top);
-      All allocations are made from the the `lowest' part of any found
+  /* Release in pagesize units, keeping at least one page */
+  extra = ((top_size - pad - MINSIZE + (pagesz-1)) / pagesz - 1) * pagesz;
+  if (extra > 0) {
+    /*
-      chunk. (The implementation invariant is that prev_inuse is
+      Only proceed if end of memory is where we last set it.
+      This avoids problems if there were foreign sbrk calls.
+    */
+    current_brk = (char*)(MORECORE(0));
+    if (current_brk == (char*)(av->top) + top_size) {
+      /*
+        Attempt to release memory. We ignore MORECORE return value,
-      always true of any allocated chunk; i.e., that each allocated
+        and instead call again to find out where new end of memory is.
-      chunk borders either a previously allocated and still in-use chunk,
+        This avoids problems if first call releases less than we asked,
+        of if failure somehow altered brk value. (We could still
+        encounter problems if it altered brk in some very bad way,
+        but the only thing we can do is adjust anyway, which will cause
-      or the base of its memory arena.)
+        some downstream failure.)
+      */
+      MORECORE(-extra);
+      new_brk = (char*)(MORECORE(0));
+      if (new_brk != (char*)MORECORE_FAILURE) {
+        released = (long)(current_brk - new_brk);
+        if (released != 0) {
+          /* Success. Adjust top. */
+          av->sbrked_mem -= released;
+          set_head(av->top, (top_size - released) | PREV_INUSE);
+          check_malloc_state();
+          return 1;
+        }
+      }
+    }
+  }
+  return 0;
+}
+#endif
+/*
+  ------------------------------ malloc ------------------------------
 */
 #if __STD_C
 Void_t* mALLOc(size_t bytes)
 #else
-Void_t* mALLOc(bytes) size_t bytes;
+  Void_t* mALLOc(bytes) size_t bytes;
 #endif
+{
-  mchunkptr victim;                  /* inspected/selected chunk */
-  INTERNAL_SIZE_T victim_size;       /* its size */
-  int       idx;                     /* index for bin traversal */
-  mbinptr   bin;                     /* associated bin */
-  mchunkptr remainder;               /* remainder from a split */
-  long      remainder_size;          /* its size */
-  int       remainder_index;         /* its bin index */
-  unsigned long block;               /* block traverser bit */
-  int       startidx;                /* first bin of a traversed block */
-  mchunkptr fwd;                     /* misc temp for linking */
-  mchunkptr bck;                     /* misc temp for linking */
-  mbinptr q;                         /* misc temp */
+  mstate av = get_malloc_state();
-  INTERNAL_SIZE_T nb;
+  INTERNAL_SIZE_T nb;               /* normalized request size */
+  unsigned int    idx;              /* associated bin index */
+  mbinptr         bin;              /* associated bin */
+  mfastbinptr*    fb;               /* associated fastbin */
+  mchunkptr       victim;           /* inspected/selected chunk */
+  INTERNAL_SIZE_T size;             /* its size */
+  int             victim_index;     /* its bin index */
+  mchunkptr       remainder;        /* remainder from a split */
-  if ((long)bytes < 0) return 0;
+  CHUNK_SIZE_T    remainder_size;   /* its size */
+  unsigned int    block;            /* bit map traverser */
+  unsigned int    bit;              /* bit map traverser */
+  unsigned int    map;              /* current word of binmap */
+  mchunkptr       fwd;              /* misc temp for linking */
-  nb = request2size(bytes);  /* padded request size; */
+  mchunkptr       bck;              /* misc temp for linking */
+  /*
+    Convert request size to internal form by adding SIZE_SZ bytes
+    overhead plus possibly more to obtain necessary alignment and/or
+    to obtain a size of at least MINSIZE, the smallest allocatable
+    size. Also, checked_request2size traps (returning 0) request sizes
-  /* Check for exact match in a bin */
+    that are so large that they wrap around zero when padded and
+    aligned.
+  */
-  if (is_small_request(nb))  /* Faster version for small requests */
-  {
-    idx = smallbin_index(nb);
+  checked_request2size(bytes, nb);
+  /*
+    Bypass search if no frees yet
+   */
+  if (!have_anychunks(av)) {
-    /* No traversal or size check necessary for small bins.  */
+    if (av->max_fast == 0) /* initialization check */
+      malloc_consolidate(av);
+    goto use_top;
+  }
-    q = bin_at(idx);
+  /*
-    victim = last(q);
+    If the size qualifies as a fastbin, first check corresponding bin.
+  */
-    /* Also scan the next one, since it would have a remainder < MINSIZE */
+  if ((CHUNK_SIZE_T)(nb) <= (CHUNK_SIZE_T)(av->max_fast)) {
-    if (victim == q)
-    {
-      q = next_bin(q);
-      victim = last(q);
+    fb = &(av->fastbins[(fastbin_index(nb))]);
-    }
-    if (victim != q)
+    if ( (victim = *fb) != 0) {
-    {
-      victim_size = chunksize(victim);
-      unlink(victim, bck, fwd);
+      *fb = victim->fd;
-      set_inuse_bit_at_offset(victim, victim_size);
-      check_malloced_chunk(victim, nb);
+      check_remalloced_chunk(victim, nb);
       return chunk2mem(victim);
+    }
-    idx += 2; /* Set for bin scan below. We've already scanned 2 bins. */
+  }
-  else
-  {
-    idx = bin_index(nb);
-    bin = bin_at(idx);
-    for (victim = last(bin); victim != bin; victim = victim->bk)
-    {
+  /*
+    If a small request, check regular bin.  Since these "smallbins"
-      victim_size = chunksize(victim);
+    hold one size each, no searching within bins is necessary.
+    (For a large request, we need to wait until unsorted chunks are
+    processed to find best fit. But for small ones, fits are exact
-      remainder_size = victim_size - nb;
+    anyway, so we can check now, which is faster.)
+  */
-      if (remainder_size >= (long)MINSIZE) /* too big */
+  if (in_smallbin_range(nb)) {
-      {
-        --idx; /* adjust to rescan below after checking last remainder */
+    idx = smallbin_index(nb);
-        break;
+    bin = bin_at(av,idx);
-      }
-      else if (remainder_size >= 0) /* exact fit */
+    if ( (victim = last(bin)) != bin) {
-      {
-        unlink(victim, bck, fwd);
+      bck = victim->bk;
-        set_inuse_bit_at_offset(victim, victim_size);
+      set_inuse_bit_at_offset(victim, nb);
+      bin->bk = bck;
+      bck->fd = bin;
-        check_malloced_chunk(victim, nb);
+      check_malloced_chunk(victim, nb);
-        return chunk2mem(victim);
+      return chunk2mem(victim);
-      }
+    }
-    ++idx;
+  }
+  /*
+     If this is a large request, consolidate fastbins before continuing.
+     While it might look excessive to kill all fastbins before
-  /* Try to use the last split-off remainder */
+     even seeing if there is space available, this avoids
+     fragmentation problems normally associated with fastbins.
+     Also, in practice, programs tend to have runs of either small or
+     large requests, but less often mixtures, so consolidation is not
+     invoked all that often in most programs. And the programs that
+     it is called frequently in otherwise tend to fragment.
+  */
-  if ( (victim = last_remainder->fd) != last_remainder)
-  {
+  else {
+    idx = largebin_index(nb);
-    victim_size = chunksize(victim);
+    if (have_fastchunks(av))
-    remainder_size = victim_size - nb;
+      malloc_consolidate(av);
+  }
+  /*
+    Process recently freed or remaindered chunks, taking one only if
+    it is exact fit, or, if this a small request, the chunk is remainder from
+    the most recent non-exact fit.  Place other traversed chunks in
+    bins.  Note that this step is the only place in any routine where
+    chunks are placed in bins.
+  */
+  while ( (victim = unsorted_chunks(av)->bk) != unsorted_chunks(av)) {
+    bck = victim->bk;
+    size = chunksize(victim);
+    /*
+       If a small request, try to use last remainder if it is the
+       only chunk in unsorted bin.  This helps promote locality for
-    if (remainder_size >= (long)MINSIZE) /* re-split */
+       runs of consecutive small requests. This is the only
+       exception to best-fit, and applies only when there is
+       no exact fit for a small chunk.
+    */
-    {
+    if (in_smallbin_range(nb) &&
+        bck == unsorted_chunks(av) &&
+        victim == av->last_remainder &&
+        (CHUNK_SIZE_T)(size) > (CHUNK_SIZE_T)(nb + MINSIZE)) {
+      /* split and reattach remainder */
+      remainder_size = size - nb;
       remainder = chunk_at_offset(victim, nb);
+      unsorted_chunks(av)->bk = unsorted_chunks(av)->fd = remainder;
+      av->last_remainder = remainder;
+      remainder->bk = remainder->fd = unsorted_chunks(av);
       set_head(victim, nb | PREV_INUSE);
-      link_last_remainder(remainder);
       set_head(remainder, remainder_size | PREV_INUSE);
       set_foot(remainder, remainder_size);
       check_malloced_chunk(victim, nb);
       return chunk2mem(victim);
+    }
+    /* remove from unsorted list */
+    unsorted_chunks(av)->bk = bck;
-    clear_last_remainder;
+    bck->fd = unsorted_chunks(av);
-    if (remainder_size >= 0)  /* exhaust */
+    /* Take now instead of binning if exact fit */
-    {
+    if (size == nb) {
-      set_inuse_bit_at_offset(victim, victim_size);
+      set_inuse_bit_at_offset(victim, size);
       check_malloced_chunk(victim, nb);
       return chunk2mem(victim);
+    }
-    /* Else place in bin */
+    /* place chunk in bin */
+    if (in_smallbin_range(size)) {
-    frontlink(victim, victim_size, remainder_index, bck, fwd);
+      victim_index = smallbin_index(size);
+      bck = bin_at(av, victim_index);
+      fwd = bck->fd;
+    }
+    else {
+      victim_index = largebin_index(size);
+      bck = bin_at(av, victim_index);
+      fwd = bck->fd;
+      if (fwd != bck) {
+        /* if smaller than smallest, place first */
+        if ((CHUNK_SIZE_T)(size) < (CHUNK_SIZE_T)(bck->bk->size)) {
+          fwd = bck;
+          bck = bck->bk;
+        }
+        else if ((CHUNK_SIZE_T)(size) >=
+                 (CHUNK_SIZE_T)(FIRST_SORTED_BIN_SIZE)) {
+          /* maintain large bins in sorted order */
+          size |= PREV_INUSE; /* Or with inuse bit to speed comparisons */
+          while ((CHUNK_SIZE_T)(size) < (CHUNK_SIZE_T)(fwd->size))
+            fwd = fwd->fd;
+          bck = fwd->bk;
+        }
+      }
+    }
+    mark_bin(av, victim_index);
+    victim->bk = bck;
+    victim->fd = fwd;
+    fwd->bk = victim;
+    bck->fd = victim;
+  }
   /*
-     If there are any possibly nonempty big-enough blocks,
+    If a large request, scan through the chunks of current bin to
-     search for best fitting chunk by scanning bins in blockwidth units.
+    find one that fits.  (This will be the smallest that fits unless
+    FIRST_SORTED_BIN_SIZE has been changed from default.)  This is
+    the only step where an unbounded number of chunks might be
+    scanned without doing anything useful with them. However the
+    lists tend to be short.
   */
-  if ( (block = idx2binblock(idx)) <= binblocks)
-  {
-    /* Get to the first marked block */
-    if ( (block & binblocks) == 0)
+  if (!in_smallbin_range(nb)) {
-    {
-      /* force to an even block boundary */
-      idx = (idx & ~(BINBLOCKWIDTH - 1)) + BINBLOCKWIDTH;
-      block <<= 1;
-      while ((block & binblocks) == 0)
-      {
-        idx += BINBLOCKWIDTH;
-        block <<= 1;
-      }
-    }
-    /* For each possibly nonempty block ... */
-    for (;;)
-    {
-      startidx = idx;          /* (track incomplete blocks) */
-      q = bin = bin_at(idx);
+    bin = bin_at(av, idx);
-      /* For each bin in this block ... */
-      do
-      {
-        /* Find and use first big enough chunk ... */
-        for (victim = last(bin); victim != bin; victim = victim->bk)
+    for (victim = last(bin); victim != bin; victim = victim->bk) {
-        {
-          victim_size = chunksize(victim);
+      size = chunksize(victim);
-          remainder_size = victim_size - nb;
-          if (remainder_size >= (long)MINSIZE) /* split */
-          {
-            remainder = chunk_at_offset(victim, nb);
+      if ((CHUNK_SIZE_T)(size) >= (CHUNK_SIZE_T)(nb)) {
-            set_head(victim, nb | PREV_INUSE);
+        remainder_size = size - nb;
-            unlink(victim, bck, fwd);
+        unlink(victim, bck, fwd);
-            link_last_remainder(remainder);
-            set_head(remainder, remainder_size | PREV_INUSE);
-            set_foot(remainder, remainder_size);
-            check_malloced_chunk(victim, nb);
-            return chunk2mem(victim);
-          }
-          else if (remainder_size >= 0)  /* take */
-          {
+        /* Exhaust */
-            set_inuse_bit_at_offset(victim, victim_size);
+        if (remainder_size < MINSIZE)  {
-            unlink(victim, bck, fwd);
+          set_inuse_bit_at_offset(victim, size);
-            check_malloced_chunk(victim, nb);
+          check_malloced_chunk(victim, nb);
-            return chunk2mem(victim);
+          return chunk2mem(victim);
-          }
-        }
-       bin = next_bin(bin);
-      } while ((++idx & (BINBLOCKWIDTH - 1)) != 0);
-      /* Clear out the block bit. */
-      do   /* Possibly backtrack to try to clear a partial block */
-      {
-        if ((startidx & (BINBLOCKWIDTH - 1)) == 0)
-        {
-          binblocks &= ~block;
-          break;
+        }
-        --startidx;
+        /* Split */
-       q = prev_bin(q);
+        else {
+          remainder = chunk_at_offset(victim, nb);
+          unsorted_chunks(av)->bk = unsorted_chunks(av)->fd = remainder;
+          remainder->bk = remainder->fd = unsorted_chunks(av);
-      } while (first(q) == q);
+          set_head(victim, nb | PREV_INUSE);
+          set_head(remainder, remainder_size | PREV_INUSE);
-      /* Get to the next possibly nonempty block */
+          set_foot(remainder, remainder_size);
+          check_malloced_chunk(victim, nb);
+          return chunk2mem(victim);
+        }
+      }
+    }
+  }
+  /*
+    Search for a chunk by scanning bins, starting with next largest
+    bin. This search is strictly by best-fit; i.e., the smallest
+    (with ties going to approximately the least recently used) chunk
+    that fits is selected.
+    The bitmap avoids needing to check that most blocks are nonempty.
+  */
+  ++idx;
+  bin = bin_at(av,idx);
+  block = idx2block(idx);
+  map = av->binmap[block];
+  bit = idx2bit(idx);
+  for (;;) {
-      if ( (block <<= 1) <= binblocks && (block != 0) )
+    /* Skip rest of block if there are no more set bits in this block.  */
+    if (bit > map || bit == 0) {
-      {
+      do {
+        if (++block >= BINMAPSIZE)  /* out of bins */
+          goto use_top;
-        while ((block & binblocks) == 0)
+      } while ( (map = av->binmap[block]) == 0);
-        {
+      bin = bin_at(av, (block << BINMAPSHIFT));
+      bit = 1;
+    }
+    /* Advance to bin with set bit. There must be one. */
+    while ((bit & map) == 0) {
+      bin = next_bin(bin);
+      bit <<= 1;
+      assert(bit != 0);
+    }
+    /* Inspect the bin. It is likely to be non-empty */
+    victim = last(bin);
+    /*  If a false alarm (empty bin), clear the bit. */
+    if (victim == bin) {
+      av->binmap[block] = map &= ~bit; /* Write through */
-          idx += BINBLOCKWIDTH;
+      bin = next_bin(bin);
-          block <<= 1;
+      bit <<= 1;
+    }
+    else {
+      size = chunksize(victim);
+      /*  We know the first chunk in this bin is big enough to use. */
+      assert((CHUNK_SIZE_T)(size) >= (CHUNK_SIZE_T)(nb));
+      remainder_size = size - nb;
+      /* unlink */
+      bck = victim->bk;
+      bin->bk = bck;
+      bck->fd = bin;
+      /* Exhaust */
+      if (remainder_size < MINSIZE) {
+        set_inuse_bit_at_offset(victim, size);
+        check_malloced_chunk(victim, nb);
+        return chunk2mem(victim);
+      }
+      /* Split */
+      else {
+        remainder = chunk_at_offset(victim, nb);
+        unsorted_chunks(av)->bk = unsorted_chunks(av)->fd = remainder;
+        remainder->bk = remainder->fd = unsorted_chunks(av);
+        /* advertise as last remainder */
+        if (in_smallbin_range(nb))
+          av->last_remainder = remainder;
-        }
+        set_head(victim, nb | PREV_INUSE);
+        set_head(remainder, remainder_size | PREV_INUSE);
+        set_foot(remainder, remainder_size);
+        check_malloced_chunk(victim, nb);
+        return chunk2mem(victim);
+      }
-      else
-        break;
+    }
+  }
+  use_top:
+  /*
+    If large enough, split off the chunk bordering the end of memory
+    (held in av->top). Note that this is in accord with the best-fit
+    search rule.  In effect, av->top is treated as larger (and thus
+    less well fitting) than any other available chunk since it can
+    be extended to be as large as necessary (up to system
-  /* Try to use top chunk */
+    limitations).
-  /* Require that there be a remainder, ensuring top always exists  */
+    We require that av->top always exists (i.e., has size >=
+    MINSIZE) after initialization, so if it would otherwise be
+    exhuasted by current request, it is replenished. (The main
-  if ( (remainder_size = chunksize(top) - nb) < (long)MINSIZE)
+    reason for ensuring it exists is that we may need MINSIZE space
+    to put in fenceposts in sysmalloc.)
-  {
+  */
-#if HAVE_MMAP
+  victim = av->top;
-    /* If big and would otherwise need to extend, try to use mmap instead */
+  size = chunksize(victim);
-    if ((unsigned long)nb >= (unsigned long)mmap_threshold &&
+  if ((CHUNK_SIZE_T)(size) >= (CHUNK_SIZE_T)(nb + MINSIZE)) {
-        (victim = mmap_chunk(nb)) != 0)
+    remainder_size = size - nb;
-      return chunk2mem(victim);
+    remainder = chunk_at_offset(victim, nb);
-#endif
-    /* Try to extend */
+    av->top = remainder;
-    malloc_extend_top(nb);
+    set_head(victim, nb | PREV_INUSE);
-    if ( (remainder_size = chunksize(top) - nb) < (long)MINSIZE)
+    set_head(remainder, remainder_size | PREV_INUSE);
+    check_malloced_chunk(victim, nb);
-      return 0; /* propagate failure */
+    return chunk2mem(victim);
+  }
-  victim = top;
+  /*
-  set_head(victim, nb | PREV_INUSE);
-  top = chunk_at_offset(victim, nb);
-  set_head(top, remainder_size | PREV_INUSE);
+     If no space in top, relay to handle system-dependent cases
-  check_malloced_chunk(victim, nb);
+  */
-  return chunk2mem(victim);
+  return sYSMALLOc(nb, av);
+}
 /*
-  free() algorithm :
-    cases:
-. free(0) has no effect.
-. If the chunk was allocated via mmap, it is release via munmap().
-. If a returned chunk borders the current high end of memory,
-          it is consolidated into the top, and if the total unused
-          topmost memory exceeds the trim threshold, malloc_trim is
+  ------------------------------ free ------------------------------
-          called.
-. Other chunks are consolidated as they arrive, and
-          placed in corresponding bins. (This includes the case of
-          consolidating with the current `last_remainder').
 */
 #if __STD_C
 void fREe(Void_t* mem)
 #else
 void fREe(mem) Void_t* mem;
 #endif
+{
-  mchunkptr p;         /* chunk corresponding to mem */
-  INTERNAL_SIZE_T hd;  /* its head field */
-  INTERNAL_SIZE_T sz;  /* its size */
-  int       idx;       /* its bin index */
-  mchunkptr next;      /* next contiguous chunk */
-  INTERNAL_SIZE_T nextsz; /* its size */
+  mstate av = get_malloc_state();
-  INTERNAL_SIZE_T prevsz; /* size of previous contiguous chunk */
-  mchunkptr bck;       /* misc temp for linking */
-  mchunkptr fwd;       /* misc temp for linking */
-  int       islr;      /* track whether merging with last_remainder */
-  if (mem == 0)                              /* free(0) has no effect */
+  mchunkptr       p;           /* chunk corresponding to mem */
-    return;
+  INTERNAL_SIZE_T size;        /* its size */
+  mfastbinptr*    fb;          /* associated fastbin */
-  p = mem2chunk(mem);
+  mchunkptr       nextchunk;   /* next contiguous chunk */
-  hd = p->size;
+  INTERNAL_SIZE_T nextsize;    /* its size */
+  int             nextinuse;   /* true if nextchunk is used */
+  INTERNAL_SIZE_T prevsize;    /* size of previous contiguous chunk */
+  mchunkptr       bck;         /* misc temp for linking */
+  mchunkptr       fwd;         /* misc temp for linking */
-#if HAVE_MMAP
-  if (hd & IS_MMAPPED)                       /* release mmapped memory. */
+  /* free(0) has no effect */
-  {
+  if (mem != 0) {
-    munmap_chunk(p);
+    p = mem2chunk(mem);
-    return;
+    size = chunksize(p);
-  }
-#endif
-  check_inuse_chunk(p);
+    check_inuse_chunk(p);
-  sz = hd & ~PREV_INUSE;
+    /*
-  next = chunk_at_offset(p, sz);
+      If eligible, place chunk on a fastbin so it can be found
-  nextsz = chunksize(next);
+      and used quickly in malloc.
+    */
-  if (next == top)                            /* merge with top */
+    if ((CHUNK_SIZE_T)(size) <= (CHUNK_SIZE_T)(av->max_fast)
-  {
-    sz += nextsz;
+#if TRIM_FASTBINS
+        /*
-    if (!(hd & PREV_INUSE))                    /* consolidate backward */
+           If TRIM_FASTBINS set, don't place chunks
+           bordering top into fastbins
+        */
+        && (chunk_at_offset(p, size) != av->top)
+#endif
-    {
+        ) {
-      prevsz = p->prev_size;
+      set_fastchunks(av);
-      p = chunk_at_offset(p, -((long) prevsz));
+      fb = &(av->fastbins[fastbin_index(size)]);
-      sz += prevsz;
+      p->fd = *fb;
-      unlink(p, bck, fwd);
+      *fb = p;
+    }
-    set_head(p, sz | PREV_INUSE);
-    top = p;
+    /*
-    if ((unsigned long)(sz) >= (unsigned long)trim_threshold)
+       Consolidate other non-mmapped chunks as they arrive.
-      malloc_trim(top_pad);
-    return;
-  }
+    */
-  set_head(next, nextsz);                    /* clear inuse bit */
+    else if (!chunk_is_mmapped(p)) {
-  islr = 0;
+      set_anychunks(av);
-  if (!(hd & PREV_INUSE))                    /* consolidate backward */
-  {
-    prevsz = p->prev_size;
-    p = chunk_at_offset(p, -((long) prevsz));
+      nextchunk = chunk_at_offset(p, size);
-    sz += prevsz;
+      nextsize = chunksize(nextchunk);
-    if (p->fd == last_remainder)             /* keep as last_remainder */
+      /* consolidate backward */
+      if (!prev_inuse(p)) {
-      islr = 1;
+        prevsize = p->prev_size;
-    else
+        size += prevsize;
+        p = chunk_at_offset(p, -((long) prevsize));
-      unlink(p, bck, fwd);
+        unlink(p, bck, fwd);
-  }
+      }
-  if (!(inuse_bit_at_offset(next, nextsz)))   /* consolidate forward */
-  {
-    sz += nextsz;
-    if (!islr && next->fd == last_remainder)  /* re-insert last_remainder */
+      if (nextchunk != av->top) {
-    {
-      islr = 1;
+        /* get and clear inuse bit */
+        nextinuse = inuse_bit_at_offset(nextchunk, nextsize);
-      link_last_remainder(p);
+        set_head(nextchunk, nextsize);
-    }
+        /* consolidate forward */
-    else
+        if (!nextinuse) {
-      unlink(next, bck, fwd);
+          unlink(nextchunk, bck, fwd);
+          size += nextsize;
-  }
+        }
+        /*
+          Place the chunk in unsorted chunk list. Chunks are
+          not placed into regular bins until after they have
+          been given one chance to be used in malloc.
+        */
+        bck = unsorted_chunks(av);
+        fwd = bck->fd;
+        p->bk = bck;
+        p->fd = fwd;
+        bck->fd = p;
+        fwd->bk = p;
+        set_head(p, size | PREV_INUSE);
+        set_foot(p, size);
+        check_free_chunk(p);
+      }
+      /*
+         If the chunk borders the current high end of memory,
-  set_head(p, sz | PREV_INUSE);
+         consolidate into top
-  set_foot(p, sz);
+      */
-  if (!islr)
+      else {
+        size += nextsize;
-    frontlink(p, sz, idx, bck, fwd);
+        set_head(p, size | PREV_INUSE);
+        av->top = p;
+        check_chunk(p);
-}
+      }
+      /*
+        If freeing a large space, consolidate possibly-surrounding
+        chunks. Then, if the total unused topmost memory exceeds trim
+        threshold, ask malloc_trim to reduce top.
+        Unless max_fast is 0, we don't know if there are fastbins
+        bordering top, so we cannot tell for sure whether threshold
+        has been reached unless fastbins are consolidated.  But we
+        don't want to consolidate on each free.  As a compromise,
+        consolidation is performed if FASTBIN_CONSOLIDATION_THRESHOLD
+        is reached.
+      */
+      if ((CHUNK_SIZE_T)(size) >= FASTBIN_CONSOLIDATION_THRESHOLD) {
+        if (have_fastchunks(av))
+          malloc_consolidate(av);
+#ifndef MORECORE_CANNOT_TRIM
+        if ((CHUNK_SIZE_T)(chunksize(av->top)) >=
+            (CHUNK_SIZE_T)(av->trim_threshold))
+          sYSTRIm(av->top_pad, av);
+#endif
+      }
+    }
+    /*
+      If the chunk was allocated via mmap, release via munmap()
+      Note that if HAVE_MMAP is false but chunk_is_mmapped is
+      true, then user must have overwritten memory. There's nothing
+      we can do to catch this error unless DEBUG is set, in which case
+      check_inuse_chunk (above) will have triggered error.
+    */
+    else {
+#if HAVE_MMAP
+      int ret;
+      INTERNAL_SIZE_T offset = p->prev_size;
+      av->n_mmaps--;
+      av->mmapped_mem -= (size + offset);
+      ret = munmap((char*)p - offset, size + offset);
+      /* munmap returns non-zero on failure */
+      assert(ret == 0);
+#endif
+    }
+  }
+}
 /*
+  ------------------------- malloc_consolidate -------------------------
-  Realloc algorithm:
+  malloc_consolidate is a specialized version of free() that tears
-    Chunks that were obtained via mmap cannot be extended or shrunk
+  down chunks held in fastbins.  Free itself cannot be used for this
-    unless HAVE_MREMAP is defined, in which case mremap is used.
+  purpose since, among other things, it might place chunks back onto
-    Otherwise, if their reallocation is for additional space, they are
+  fastbins.  So, instead, we need to use a minor variant of the same
-    copied.  If for less, they are just left alone.
+  code.
-    Otherwise, if the reallocation is for additional space, and the
+  Also, because this routine needs to be called the first time through
-    chunk can be extended, it is, else a malloc-copy-free sequence is
+  malloc anyway, it turns out to be the perfect place to trigger
-    taken.  There are several different ways that a chunk could be
-    extended. All are tried:
+  initialization code.
+*/
+#if __STD_C
+static void malloc_consolidate(mstate av)
+#else
+static void malloc_consolidate(av) mstate av;
+#endif
+{
+  mfastbinptr*    fb;                 /* current fastbin being consolidated */
+  mfastbinptr*    maxfb;              /* last fastbin (for loop control) */
+  mchunkptr       p;                  /* current chunk being consolidated */
-       * Extending forward into following adjacent free chunk.
+  mchunkptr       nextp;              /* next chunk to consolidate */
-       * Shifting backwards, joining preceding adjacent space
+  mchunkptr       unsorted_bin;       /* bin header */
-       * Both shifting backwards and extending forward.
+  mchunkptr       first_unsorted;     /* chunk to link to */
-       * Extending into newly sbrked space
+  /* These have same use as in free() */
+  mchunkptr       nextchunk;
+  INTERNAL_SIZE_T size;
+  INTERNAL_SIZE_T nextsize;
+  INTERNAL_SIZE_T prevsize;
+  int             nextinuse;
+  mchunkptr       bck;
+  mchunkptr       fwd;
+  /*
-    Unless the #define REALLOC_ZERO_BYTES_FREES is set, realloc with a
+    If max_fast is 0, we know that av hasn't
-    size argument of zero (re)allocates a minimum-sized chunk.
+    yet been initialized, in which case do so below
+  */
-    If the reallocation is for less space, and the new request is for
+  if (av->max_fast != 0) {
-    a `small' (<512 bytes) size, then the newly unused space is lopped
-    off and freed.
+    clear_fastchunks(av);
-    The old unix realloc convention of allowing the last-free'd chunk
-    to be used as an argument to realloc is no longer supported.
-    I don't know of any programs still relying on this feature,
-    and allowing it would also allow too many other incorrect
-    usages of realloc to be sensible.
+    unsorted_bin = unsorted_chunks(av);
+    /*
+      Remove each chunk from fast bin and consolidate it, placing it
+      then in unsorted bin. Among other reasons for doing this,
+      placing in unsorted bin avoids needing to calculate actual bins
+      until malloc is sure that chunks aren't immediately going to be
+      reused anyway.
+    */
+    maxfb = &(av->fastbins[fastbin_index(av->max_fast)]);
+    fb = &(av->fastbins[0]);
+    do {
+      if ( (p = *fb) != 0) {
+        *fb = 0;
+        do {
+          check_inuse_chunk(p);
+          nextp = p->fd;
+          /* Slightly streamlined version of consolidation code in free() */
+          size = p->size & ~PREV_INUSE;
+          nextchunk = chunk_at_offset(p, size);
+          nextsize = chunksize(nextchunk);
+          if (!prev_inuse(p)) {
+            prevsize = p->prev_size;
+            size += prevsize;
+            p = chunk_at_offset(p, -((long) prevsize));
+            unlink(p, bck, fwd);
+          }
+          if (nextchunk != av->top) {
+            nextinuse = inuse_bit_at_offset(nextchunk, nextsize);
+            set_head(nextchunk, nextsize);
+            if (!nextinuse) {
+              size += nextsize;
+              unlink(nextchunk, bck, fwd);
+            }
+            first_unsorted = unsorted_bin->fd;
+            unsorted_bin->fd = p;
+            first_unsorted->bk = p;
+            set_head(p, size | PREV_INUSE);
+            p->bk = unsorted_bin;
+            p->fd = first_unsorted;
+            set_foot(p, size);
+          }
+          else {
+            size += nextsize;
+            set_head(p, size | PREV_INUSE);
+            av->top = p;
+          }
+        } while ( (p = nextp) != 0);
+      }
+    } while (fb++ != maxfb);
+  }
+  else {
+    malloc_init_state(av);
+    check_malloc_state();
+  }
+}
+/*
+  ------------------------------ realloc ------------------------------
 */
 #if __STD_C
 Void_t* rEALLOc(Void_t* oldmem, size_t bytes)
 #else
 Void_t* rEALLOc(oldmem, bytes) Void_t* oldmem; size_t bytes;
 #endif
+{
-  INTERNAL_SIZE_T    nb;      /* padded request size */
+  mstate av = get_malloc_state();
-  mchunkptr oldp;             /* chunk corresponding to oldmem */
-  INTERNAL_SIZE_T    oldsize; /* its size */
+  INTERNAL_SIZE_T  nb;              /* padded request size */
-  mchunkptr newp;             /* chunk to return */
+  mchunkptr        oldp;            /* chunk corresponding to oldmem */
-  INTERNAL_SIZE_T    newsize; /* its size */
+  INTERNAL_SIZE_T  oldsize;         /* its size */
-  Void_t*   newmem;           /* corresponding user mem */
-  mchunkptr next;             /* next contiguous chunk after oldp */
+  mchunkptr        newp;            /* chunk to return */
-  INTERNAL_SIZE_T  nextsize;  /* its size */
+  INTERNAL_SIZE_T  newsize;         /* its size */
+  Void_t*          newmem;          /* corresponding user mem */
-  mchunkptr prev;             /* previous contiguous chunk before oldp */
+  mchunkptr        next;            /* next contiguous chunk after oldp */
-  INTERNAL_SIZE_T  prevsize;  /* its size */
-  mchunkptr remainder;        /* holds split off extra space from newp */
+  mchunkptr        remainder;       /* extra space at end of newp */
-  INTERNAL_SIZE_T  remainder_size;   /* its size */
+  CHUNK_SIZE_T     remainder_size;  /* its size */
+  mchunkptr        bck;             /* misc temp for linking */
+  mchunkptr        fwd;             /* misc temp for linking */
+  CHUNK_SIZE_T     copysize;        /* bytes to copy */
+  unsigned int     ncopies;         /* INTERNAL_SIZE_T words to copy */
+  INTERNAL_SIZE_T* s;               /* copy source */
+  INTERNAL_SIZE_T* d;               /* copy destination */
-  mchunkptr bck;              /* misc temp for linking */
-  mchunkptr fwd;              /* misc temp for linking */
 #ifdef REALLOC_ZERO_BYTES_FREES
-  if (bytes == 0) { fREe(oldmem); return 0; }
+  if (bytes == 0) {
+    fREe(oldmem);
+    return 0;
+  }
 #endif
-  if ((long)bytes < 0) return 0;
   /* realloc of null is supposed to be same as malloc */
   if (oldmem == 0) return mALLOc(bytes);
-  newp    = oldp    = mem2chunk(oldmem);
-  newsize = oldsize = chunksize(oldp);
+  checked_request2size(bytes, nb);
-  nb = request2size(bytes);
-#if HAVE_MMAP
-  if (chunk_is_mmapped(oldp))
-  {
-#if HAVE_MREMAP
-    newp = mremap_chunk(oldp, nb);
+  oldp    = mem2chunk(oldmem);
-    if(newp) return chunk2mem(newp);
-#endif
-    /* Note the extra SIZE_SZ overhead. */
-    if(oldsize - SIZE_SZ >= nb) return oldmem; /* do nothing */
-    /* Must alloc, copy, free. */
-    newmem = mALLOc(bytes);
-    if (newmem == 0) return 0; /* propagate failure */
-    MALLOC_COPY(newmem, oldmem, oldsize - 2*SIZE_SZ);
-    munmap_chunk(oldp);
+  oldsize = chunksize(oldp);
-    return newmem;
-  }
-#endif
   check_inuse_chunk(oldp);
-  if ((long)(oldsize) < (long)(nb))
+  if (!chunk_is_mmapped(oldp)) {
-  {
+    if ((CHUNK_SIZE_T)(oldsize) >= (CHUNK_SIZE_T)(nb)) {
-    /* Try expanding forward */
+      /* already big enough; split below */
+      newp = oldp;
+      newsize = oldsize;
+    }
-    next = chunk_at_offset(oldp, oldsize);
-    if (next == top || !inuse(next))
-    {
+    else {
-      nextsize = chunksize(next);
+      next = chunk_at_offset(oldp, oldsize);
-      /* Forward into top only if a remainder */
+      /* Try to expand forward into top */
-      if (next == top)
+      if (next == av->top &&
-      {
-        if ((long)(nextsize + newsize) >= (long)(nb + MINSIZE))
+          (CHUNK_SIZE_T)(newsize = oldsize + chunksize(next)) >=
-        {
+          (CHUNK_SIZE_T)(nb + MINSIZE)) {
-          newsize += nextsize;
+        set_head_size(oldp, nb);
-          top = chunk_at_offset(oldp, nb);
+        av->top = chunk_at_offset(oldp, nb);
-          set_head(top, (newsize - nb) | PREV_INUSE);
+        set_head(av->top, (newsize - nb) | PREV_INUSE);
-          set_head_size(oldp, nb);
-          return chunk2mem(oldp);
+        return chunk2mem(oldp);
-        }
+      }
+      /* Try to expand forward into next chunk;  split off remainder below */
+      else if (next != av->top &&
-      /* Forward into next chunk */
+               !inuse(next) &&
-      else if (((long)(nextsize + newsize) >= (long)(nb)))
+               (CHUNK_SIZE_T)(newsize = oldsize + chunksize(next)) >=
+               (CHUNK_SIZE_T)(nb)) {
-      {
+        newp = oldp;
         unlink(next, bck, fwd);
-        newsize  += nextsize;
-        goto split;
+      }
-    }
-    else
-    {
-      next = 0;
-      nextsize = 0;
-    }
-    /* Try shifting backwards. */
-    if (!prev_inuse(oldp))
-    {
-      prev = prev_chunk(oldp);
-      prevsize = chunksize(prev);
-      /* try forward + backward first to save a later consolidation */
+      /* allocate, copy, free */
+      else {
+        newmem = mALLOc(nb - MALLOC_ALIGN_MASK);
-      if (next != 0)
+        if (newmem == 0)
+          return 0; /* propagate failure */
-      {
-        /* into top */
+        newp = mem2chunk(newmem);
-        if (next == top)
+        newsize = chunksize(newp);
-        {
-          if ((long)(nextsize + prevsize + newsize) >= (long)(nb + MINSIZE))
-          {
+        /*
-            unlink(prev, bck, fwd);
+          Avoid copy if newp is next chunk after oldp.
-            newp = prev;
+        */
-            newsize += prevsize + nextsize;
-            newmem = chunk2mem(newp);
+        if (newp == next) {
-            MALLOC_COPY(newmem, oldmem, oldsize - SIZE_SZ);
-            top = chunk_at_offset(newp, nb);
-            set_head(top, (newsize - nb) | PREV_INUSE);
-            set_head_size(newp, nb);
+          newsize += oldsize;
-            return newmem;
+          newp = oldp;
-          }
+        }
+        else {
+          /*
+            Unroll copy of <= 36 bytes (72 if 8byte sizes)
+            We know that contents have an odd number of
+            INTERNAL_SIZE_T-sized words; minimally 3.
+          */
-        /* into next chunk */
+          copysize = oldsize - SIZE_SZ;
+          s = (INTERNAL_SIZE_T*)(oldmem);
+          d = (INTERNAL_SIZE_T*)(newmem);
-        else if (((long)(nextsize + prevsize + newsize) >= (long)(nb)))
+          ncopies = copysize / sizeof(INTERNAL_SIZE_T);
+          assert(ncopies >= 3);
-        {
-          unlink(next, bck, fwd);
+          if (ncopies > 9)
+            MALLOC_COPY(d, s, copysize);
+          else {
+            *(d+0) = *(s+0);
-          unlink(prev, bck, fwd);
+            *(d+1) = *(s+1);
-          newp = prev;
+            *(d+2) = *(s+2);
+            if (ncopies > 4) {
+              *(d+3) = *(s+3);
+              *(d+4) = *(s+4);
+              if (ncopies > 6) {
+                *(d+5) = *(s+5);
-          newsize += nextsize + prevsize;
+                *(d+6) = *(s+6);
+                if (ncopies > 8) {
-          newmem = chunk2mem(newp);
+                  *(d+7) = *(s+7);
-          MALLOC_COPY(newmem, oldmem, oldsize - SIZE_SZ);
+                  *(d+8) = *(s+8);
+                }
+              }
+            }
+          }
-          goto split;
+          fREe(oldmem);
+          check_inuse_chunk(newp);
+          return chunk2mem(newp);
+        }
+      }
-      /* backward only */
-      if (prev != 0 && (long)(prevsize + newsize) >= (long)nb)
-      {
-        unlink(prev, bck, fwd);
-        newp = prev;
-        newsize += prevsize;
-        newmem = chunk2mem(newp);
-        MALLOC_COPY(newmem, oldmem, oldsize - SIZE_SZ);
-        goto split;
-      }
+    }
-    /* Must allocate */
+    /* If possible, free extra space in old or extended chunk */
-    newmem = mALLOc (bytes);
-    if (newmem == 0)  /* propagate failure */
+    assert((CHUNK_SIZE_T)(newsize) >= (CHUNK_SIZE_T)(nb));
-      return 0;
-    /* Avoid copy if newp is next chunk after oldp. */
+    remainder_size = newsize - nb;
-    /* (This can only happen when new chunk is sbrk'ed.) */
+    if (remainder_size < MINSIZE) { /* not enough extra to split off */
+      set_head_size(newp, newsize);
-    if ( (newp = mem2chunk(newmem)) == next_chunk(oldp))
+      set_inuse_bit_at_offset(newp, newsize);
-    {
+    }
+    else { /* split remainder */
-      newsize += chunksize(newp);
+      remainder = chunk_at_offset(newp, nb);
-      newp = oldp;
+      set_head_size(newp, nb);
+      set_head(remainder, remainder_size | PREV_INUSE);
+      /* Mark remainder as inuse so free() won't complain */
+      set_inuse_bit_at_offset(remainder, remainder_size);
-      goto split;
+      fREe(chunk2mem(remainder));
+    }
-    /* Otherwise copy, free, and exit */
-    MALLOC_COPY(newmem, oldmem, oldsize - SIZE_SZ);
-    fREe(oldmem);
+    check_inuse_chunk(newp);
-    return newmem;
+    return chunk2mem(newp);
+  }
+  /*
+    Handle mmap cases
+  */
- split:  /* split off extra room in old or expanded chunk */
-  if (newsize - nb >= MINSIZE) /* split off remainder */
-  {
-    remainder = chunk_at_offset(newp, nb);
-    remainder_size = newsize - nb;
-    set_head_size(newp, nb);
-    set_head(remainder, remainder_size | PREV_INUSE);
-    set_inuse_bit_at_offset(remainder, remainder_size);
-    fREe(chunk2mem(remainder)); /* let free() deal with it */
-  }
-  else
+  else {
-  {
-    set_head_size(newp, newsize);
-    set_inuse_bit_at_offset(newp, newsize);
-  }
+#if HAVE_MMAP
+#if HAVE_MREMAP
+    INTERNAL_SIZE_T offset = oldp->prev_size;
+    size_t pagemask = av->pagesize - 1;
+    char *cp;
+    CHUNK_SIZE_T  sum;
+    /* Note the extra SIZE_SZ overhead */
+    newsize = (nb + offset + SIZE_SZ + pagemask) & ~pagemask;
+    /* don't need to remap if still within same page */
+    if (oldsize == newsize - offset)
+      return oldmem;
+    cp = (char*)mremap((char*)oldp - offset, oldsize + offset, newsize, 1);
+    if (cp != (char*)MORECORE_FAILURE) {
+      newp = (mchunkptr)(cp + offset);
+      set_head(newp, (newsize - offset)|IS_MMAPPED);
-  check_inuse_chunk(newp);
+      assert(aligned_OK(chunk2mem(newp)));
+      assert((newp->prev_size == offset));
+      /* update statistics */
+      sum = av->mmapped_mem += newsize - oldsize;
+      if (sum > (CHUNK_SIZE_T)(av->max_mmapped_mem))
+        av->max_mmapped_mem = sum;
+      sum += av->sbrked_mem;
+      if (sum > (CHUNK_SIZE_T)(av->max_total_mem))
+        av->max_total_mem = sum;
+	  /* R Memory Statistics */
+ 	  max_total_mem = av->max_total_mem;
-  return chunk2mem(newp);
+      return chunk2mem(newp);
-}
+    }
+#endif
+    /* Note the extra SIZE_SZ overhead. */
+    if ((CHUNK_SIZE_T)(oldsize) >= (CHUNK_SIZE_T)(nb + SIZE_SZ))
+      newmem = oldmem; /* do nothing */
+    else {
+      /* Must alloc, copy, free. */
+      newmem = mALLOc(nb - MALLOC_ALIGN_MASK);
+      if (newmem != 0) {
+        MALLOC_COPY(newmem, oldmem, oldsize - 2*SIZE_SZ);
+        fREe(oldmem);
+      }
+    }
+    return newmem;
+#else
+    /* If !HAVE_MMAP, but chunk_is_mmapped, user must have overwritten mem */
+    check_malloc_state();
+    MALLOC_FAILURE_ACTION;
+    return 0;
+#endif
+  }
+}
 /*
-  memalign algorithm:
-    memalign requests more than enough space from malloc, finds a spot
-    within that chunk that meets the alignment request, and then
-    possibly frees the leading and trailing space.
-    The alignment argument must be a power of two. This property is not
-    checked by memalign, so misuse may result in random runtime errors.
--byte alignment is guaranteed by normal malloc calls, so don't
-    bother calling memalign with an argument of 8 or less.
+  ------------------------------ memalign ------------------------------
-    Overreliance on memalign is a sure way to fragment space.
 */
 #if __STD_C
 Void_t* mEMALIGn(size_t alignment, size_t bytes)
 #else
 Void_t* mEMALIGn(alignment, bytes) size_t alignment; size_t bytes;
 #endif
+{
-  INTERNAL_SIZE_T    nb;      /* padded  request size */
+  INTERNAL_SIZE_T nb;             /* padded  request size */
-  char*     m;                /* memory returned by malloc call */
+  char*           m;              /* memory returned by malloc call */
-  mchunkptr p;                /* corresponding chunk */
+  mchunkptr       p;              /* corresponding chunk */
-  char*     brk;              /* alignment point within p */
+  char*           brk;            /* alignment point within p */
-  mchunkptr newp;             /* chunk to return */
+  mchunkptr       newp;           /* chunk to return */
-  INTERNAL_SIZE_T  newsize;   /* its size */
+  INTERNAL_SIZE_T newsize;        /* its size */
-  INTERNAL_SIZE_T  leadsize;  /* leading space befor alignment point */
+  INTERNAL_SIZE_T leadsize;       /* leading space before alignment point */
-  mchunkptr remainder;        /* spare room at end to split off */
+  mchunkptr       remainder;      /* spare room at end to split off */
-  long      remainder_size;   /* its size */
+  CHUNK_SIZE_T    remainder_size; /* its size */
-  if ((long)bytes < 0) return 0;
+  INTERNAL_SIZE_T size;
   /* If need less alignment than we give anyway, just relay to malloc */
   if (alignment <= MALLOC_ALIGNMENT) return mALLOc(bytes);
   /* Otherwise, ensure that it is at least a minimum chunk size */
   if (alignment <  MINSIZE) alignment = MINSIZE;
+  /* Make sure alignment is power of 2 (in case MINSIZE is not).  */
+  if ((alignment & (alignment - 1)) != 0) {
+    size_t a = MALLOC_ALIGNMENT * 2;
+    while ((CHUNK_SIZE_T)a < (CHUNK_SIZE_T)alignment) a <<= 1;
+    alignment = a;
+  }
+  checked_request2size(bytes, nb);
+  /*
+    Strategy: find a spot within that chunk that meets the alignment
+    request, and then possibly free the leading and trailing space.
+  */
   /* Call malloc with worst case padding to hit alignment. */
-  nb = request2size(bytes);
   m  = (char*)(mALLOc(nb + alignment + MINSIZE));
   if (m == 0) return 0; /* propagate failure */
   p = mem2chunk(m);
-  if ((((unsigned long)(m)) % alignment) == 0) /* aligned */
+  if ((((PTR_UINT)(m)) % alignment) != 0) { /* misaligned */
-  {
-#if HAVE_MMAP
-    if(chunk_is_mmapped(p))
-      return chunk2mem(p); /* nothing more to do */
-#endif
-  }
-  else /* misaligned */
-  {
     /*
-      Find an aligned spot inside chunk.
+      Find an aligned spot inside chunk.  Since we need to give back
-      Since we need to give back leading space in a chunk of at
-      least MINSIZE, if the first calculation places us at
+      leading space in a chunk of at least MINSIZE, if the first
-      a spot with less than MINSIZE leader, we can move to the
+      calculation places us at a spot with less than MINSIZE leader,
-      next aligned spot -- we've allocated enough total room so that
+      we can move to the next aligned spot -- we've allocated enough
-      this is always possible.
+      total room so that this is always possible.
     */
-    brk = (char*)mem2chunk(((unsigned long)(m + alignment - 1)) & -((signed) alignment));
+    brk = (char*)mem2chunk((PTR_UINT)(((PTR_UINT)(m + alignment - 1)) &
+                           -((signed long) alignment)));
-    if ((long)(brk - (char*)(p)) < MINSIZE) brk = brk + alignment;
+    if ((CHUNK_SIZE_T)(brk - (char*)(p)) < MINSIZE)
+      brk += alignment;
     newp = (mchunkptr)brk;
     leadsize = brk - (char*)(p);
     newsize = chunksize(p) - leadsize;
-#if HAVE_MMAP
+    /* For mmapped chunks, just adjust offset */
-    if(chunk_is_mmapped(p))
+    if (chunk_is_mmapped(p)) {
-    {
       newp->prev_size = p->prev_size + leadsize;
       set_head(newp, newsize|IS_MMAPPED);
       return chunk2mem(newp);
+    }
-#endif
-    /* give back leader, use the rest */
+    /* Otherwise, give back leader, use the rest */
     set_head(newp, newsize | PREV_INUSE);
     set_inuse_bit_at_offset(newp, newsize);
     set_head_size(p, leadsize);
     fREe(chunk2mem(p));
     p = newp;
+    assert (newsize >= nb &&
-    assert (newsize >= nb && (((unsigned long)(chunk2mem(p))) % alignment) == 0);
+            (((PTR_UINT)(chunk2mem(p))) % alignment) == 0);
+  }
   /* Also give back spare room at the end */
+  if (!chunk_is_mmapped(p)) {
-  remainder_size = chunksize(p) - nb;
+    size = chunksize(p);
+    if ((CHUNK_SIZE_T)(size) > (CHUNK_SIZE_T)(nb + MINSIZE)) {
-  if (remainder_size >= (long)MINSIZE)
+      remainder_size = size - nb;
-  {
-    remainder = chunk_at_offset(p, nb);
+      remainder = chunk_at_offset(p, nb);
-    set_head(remainder, remainder_size | PREV_INUSE);
+      set_head(remainder, remainder_size | PREV_INUSE);
-    set_head_size(p, nb);
+      set_head_size(p, nb);
-    fREe(chunk2mem(remainder));
+      fREe(chunk2mem(remainder));
+    }
+  }
   check_inuse_chunk(p);
   return chunk2mem(p);
+}
 /*
-    valloc just invokes memalign with alignment argument equal
+  ------------------------------ calloc ------------------------------
-    to the page size of the system (or as near to this as can
-    be figured out from all the includes/defines above.)
 */
 #if __STD_C
-Void_t* vALLOc(size_t bytes)
+Void_t* cALLOc(size_t n_elements, size_t elem_size)
 #else
-Void_t* vALLOc(bytes) size_t bytes;
+Void_t* cALLOc(n_elements, elem_size) size_t n_elements; size_t elem_size;
 #endif
+{
+  mchunkptr p;
+  CHUNK_SIZE_T  clearsize;
+  CHUNK_SIZE_T  nclears;
+  INTERNAL_SIZE_T* d;
-  return mEMALIGn (malloc_getpagesize, bytes);
+  Void_t* mem = mALLOc(n_elements * elem_size);
+  if (mem != 0) {
+    p = mem2chunk(mem);
+    if (!chunk_is_mmapped(p))
+    {
+      /*
+        Unroll clear of <= 36 bytes (72 if 8byte sizes)
+        We know that contents have an odd number of
+        INTERNAL_SIZE_T-sized words; minimally 3.
+      */
+      d = (INTERNAL_SIZE_T*)mem;
+      clearsize = chunksize(p) - SIZE_SZ;
+      nclears = clearsize / sizeof(INTERNAL_SIZE_T);
+      assert(nclears >= 3);
+      if (nclears > 9)
+        MALLOC_ZERO(d, clearsize);
+      else {
+        *(d+0) = 0;
+        *(d+1) = 0;
+        *(d+2) = 0;
+        if (nclears > 4) {
+          *(d+3) = 0;
+          *(d+4) = 0;
+          if (nclears > 6) {
+            *(d+5) = 0;
+            *(d+6) = 0;
+            if (nclears > 8) {
+              *(d+7) = 0;
+              *(d+8) = 0;
+            }
+          }
+        }
+      }
+    }
+#if ! MMAP_CLEARS
+    else
+    {
+      d = (INTERNAL_SIZE_T*)mem;
+      /*
+        Note the additional SIZE_SZ
+      */
+      clearsize = chunksize(p) - 2*SIZE_SZ;
+      MALLOC_ZERO(d, clearsize);
+    }
+#endif
+  }
+  return mem;
+}
 /*
-  pvalloc just invokes valloc for the nearest pagesize
+  ------------------------------ cfree ------------------------------
-  that will accommodate request
 */
 #if __STD_C
-Void_t* pvALLOc(size_t bytes)
+void cFREe(Void_t *mem)
 #else
-Void_t* pvALLOc(bytes) size_t bytes;
+void cFREe(mem) Void_t *mem;
 #endif
+{
-  size_t pagesize = malloc_getpagesize;
+  fREe(mem);
-  return mEMALIGn (pagesize, (bytes + pagesize - 1) & ~(pagesize - 1));
+}
 /*
+  ------------------------- independent_calloc -------------------------
+*/
+#if __STD_C
+Void_t** iCALLOc(size_t n_elements, size_t elem_size, Void_t* chunks[])
+#else
+Void_t** iCALLOc(n_elements, elem_size, chunks) size_t n_elements; size_t elem_size; Void_t* chunks[];
+#endif
+{
-  calloc calls malloc, then zeroes out the allocated chunk.
+  size_t sz = elem_size; /* serves as 1-element array */
+  /* opts arg of 3 means all elements are same size, and should be cleared */
+  return iALLOc(n_elements, &sz, 3, chunks);
+}
+/*
+  ------------------------- independent_comalloc -------------------------
 */
 #if __STD_C
-Void_t* cALLOc(size_t n, size_t elem_size)
+Void_t** iCOMALLOc(size_t n_elements, size_t sizes[], Void_t* chunks[])
 #else
-Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
+Void_t** iCOMALLOc(n_elements, sizes, chunks) size_t n_elements; size_t sizes[]; Void_t* chunks[];
 #endif
+{
-  mchunkptr p;
-  INTERNAL_SIZE_T csz;
+  return iALLOc(n_elements, sizes, 0, chunks);
+}
+/*
+  ------------------------------ ialloc ------------------------------
+  ialloc provides common support for independent_X routines, handling all of
-  INTERNAL_SIZE_T sz = n * elem_size;
+  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
+*/
+#if __STD_C
-  /* check if expand_top called, in which case don't need to clear */
+static Void_t** iALLOc(size_t n_elements,
-#if MORECORE_CLEARS
+                       size_t* sizes,
-  mchunkptr oldtop = top;
+                       int opts,
-  INTERNAL_SIZE_T oldtopsize = chunksize(top);
+                       Void_t* chunks[])
+#else
+static Void_t** iALLOc(n_elements, sizes, opts, chunks) size_t n_elements; size_t* sizes; int opts; Void_t* chunks[];
 #endif
+{
+  mstate av = get_malloc_state();
+  INTERNAL_SIZE_T element_size;   /* chunksize of each element, if all same */
+  INTERNAL_SIZE_T contents_size;  /* total size of elements */
+  INTERNAL_SIZE_T array_size;     /* request size of pointer array */
+  Void_t*         mem;            /* malloced aggregate space */
+  mchunkptr       p;              /* corresponding chunk */
+  INTERNAL_SIZE_T remainder_size; /* remaining bytes while splitting */
+  Void_t**        marray;         /* either "chunks" or malloced ptr array */
+  mchunkptr       array_chunk;    /* chunk for malloced ptr array */
+  int             mmx;            /* to disable mmap */
+  INTERNAL_SIZE_T size;
+  size_t          i;
+  /* Ensure initialization */
+  if (av->max_fast == 0) malloc_consolidate(av);
+  /* 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)
-  Void_t* mem = mALLOc (sz);
+      return (Void_t**) mALLOc(0);
+    marray = 0;
+    array_size = request2size(n_elements * (sizeof(Void_t*)));
+  }
+  /* 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;
-  if ((long)n < 0) return 0;
+    contents_size = 0;
+    for (i = 0; i != n_elements; ++i)
+      contents_size += request2size(sizes[i]);
+  }
+  /* subtract out alignment bytes from total to minimize overallocation */
+  size = contents_size + array_size - MALLOC_ALIGN_MASK;
+  /*
+     Allocate the aggregate chunk.
+     But first disable mmap so malloc won't use it, since
+     we would not be able to later free/realloc space internal
+     to a segregated mmap region.
+ */
+  mmx = av->n_mmaps_max;   /* disable mmap */
+  av->n_mmaps_max = 0;
+  mem = mALLOc(size);
+  av->n_mmaps_max = mmx;   /* reset mmap */
   if (mem == 0)
     return 0;
-  else
-  {
-    p = mem2chunk(mem);
-    /* Two optional cases in which clearing not necessary */
+  p = mem2chunk(mem);
+  assert(!chunk_is_mmapped(p));
+  remainder_size = chunksize(p);
-#if HAVE_MMAP
+  if (opts & 0x2) {       /* optionally clear the elements */
-    if (chunk_is_mmapped(p)) return mem;
+    MALLOC_ZERO(mem, remainder_size - SIZE_SZ - array_size);
-#endif
+  }
+  /* If not provided, allocate the pointer array as final part of chunk */
+  if (marray == 0) {
+    array_chunk = chunk_at_offset(p, contents_size);
+    marray = (Void_t**) (chunk2mem(array_chunk));
+    set_head(array_chunk, (remainder_size - contents_size) | PREV_INUSE);
-    csz = chunksize(p);
+    remainder_size = contents_size;
+  }
+  /* split out elements */
-#if MORECORE_CLEARS
+  for (i = 0; ; ++i) {
+    marray[i] = chunk2mem(p);
+    if (i != n_elements-1) {
+      if (element_size != 0)
-    if (p == oldtop && csz > oldtopsize)
+        size = element_size;
-    {
+      else
-      /* clear only the bytes from non-freshly-sbrked memory */
+        size = request2size(sizes[i]);
-      csz = oldtopsize;
+      remainder_size -= size;
+      set_head(p, size | PREV_INUSE);
+      p = chunk_at_offset(p, size);
+    }
+    else { /* the final element absorbs any overallocation slop */
+      set_head(p, remainder_size | PREV_INUSE);
+      break;
-#endif
+    }
+  }
+#if DEBUG
+  if (marray != chunks) {
+    /* final element must have exactly exhausted chunk */
+    if (element_size != 0)
-    MALLOC_ZERO(mem, csz - SIZE_SZ);
+      assert(remainder_size == element_size);
-    return mem;
+    else
+      assert(remainder_size == request2size(sizes[i]));
+    check_inuse_chunk(mem2chunk(marray));
+  }
-}
+  for (i = 0; i != n_elements; ++i)
+    check_inuse_chunk(mem2chunk(marray[i]));
+#endif
+  return marray;
-/*
+}
-  cfree just calls free. It is needed/defined on some systems
-  that pair it with calloc, presumably for odd historical reasons.
+/*
+  ------------------------------ valloc ------------------------------
 */
-#if !defined(INTERNAL_LINUX_C_LIB) || !defined(__ELF__)
 #if __STD_C
-void cfree(Void_t *mem)
+Void_t* vALLOc(size_t bytes)
 #else
-void cfree(mem) Void_t *mem;
+Void_t* vALLOc(bytes) size_t bytes;
 #endif
+{
+  /* Ensure initialization */
-  fREe(mem);
+  mstate av = get_malloc_state();
+  if (av->max_fast == 0) malloc_consolidate(av);
+  return mEMALIGn(av->pagesize, bytes);
+}
-#endif
 /*
+  ------------------------------ pvalloc ------------------------------
+*/
-    Malloc_trim gives memory back to the system (via negative
-    arguments to sbrk) if there is unused memory at the `high' end of
-    the malloc pool. 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,
+#if __STD_C
-    only the minimum amount of memory to maintain internal data
+Void_t* pVALLOc(size_t bytes)
+#else
-    structures will be left (one page or less). Non-zero arguments
+Void_t* pVALLOc(bytes) size_t bytes;
+#endif
+{
-    can be supplied to maintain enough trailing space to service
+  mstate av = get_malloc_state();
-    future expected allocations without having to re-obtain memory
-    from the system.
+  size_t pagesz;
+  /* Ensure initialization */
+  if (av->max_fast == 0) malloc_consolidate(av);
+  pagesz = av->pagesize;
-    Malloc_trim returns 1 if it actually released any memory, else 0.
+  return mEMALIGn(pagesz, (bytes + pagesz - 1) & ~(pagesz - 1));
+}
+/*
+  ------------------------------ malloc_trim ------------------------------
 */
 #if __STD_C
-int malloc_trim(size_t pad)
+int mTRIm(size_t pad)
 #else
-int malloc_trim(pad) size_t pad;
+int mTRIm(pad) size_t pad;
 #endif
+{
-  long  top_size;        /* Amount of top-most memory */
+  mstate av = get_malloc_state();
-  long  extra;           /* Amount to release */
+  /* Ensure initialization/consolidation */
-  char* current_brk;     /* address returned by pre-check sbrk call */
-  char* new_brk;         /* address returned by negative sbrk call */
-  unsigned long pagesz = malloc_getpagesize;
+  malloc_consolidate(av);
-  top_size = chunksize(top);
+#ifndef MORECORE_CANNOT_TRIM
-  extra = ((top_size - pad - MINSIZE + (pagesz-1)) / pagesz - 1) * pagesz;
-  if (extra < (long)pagesz)  /* Not enough memory to release */
-    return 0;
+  return sYSTRIm(pad, av);
-  else
+#else
-  {
-    /* Test to make sure no one else called sbrk */
-    current_brk = (char*)(MORECORE (0));
-    if (current_brk != (char*)(top) + top_size)
-      return 0;     /* Apparently we don't own memory; must fail */
-    else
-    {
-      new_brk = (char*)(MORECORE (-extra));
-      if (new_brk == (char*)(MORECORE_FAILURE)) /* sbrk failed? */
-      {
-        /* Try to figure out what we have */
-        current_brk = (char*)(MORECORE (0));
-        top_size = current_brk - (char*)top;
-        if (top_size >= (long)MINSIZE) /* if not, we are very very dead! */
-        {
-          sbrked_mem = current_brk - sbrk_base;
-          set_head(top, top_size | PREV_INUSE);
-        }
-        check_chunk(top);
-        return 0;
+  return 0;
-      }
-      else
-      {
-        /* Success. Adjust top accordingly. */
-        set_head(top, (top_size - extra) | PREV_INUSE);
-        sbrked_mem -= extra;
-        check_chunk(top);
-        return 1;
-      }
-    }
+#endif
-  }
+}
 /*
-  malloc_usable_size:
-    This routine tells you how many bytes you can actually use in an
-    allocated chunk, which may be more than you requested (although
-    often not). You can use this many bytes without worrying about
-    overwriting other allocated objects. Not a particularly great
+  ------------------------- malloc_usable_size -------------------------
-    programming practice, but still sometimes useful.
 */
 #if __STD_C
-size_t malloc_usable_size(Void_t* mem)
+size_t mUSABLe(Void_t* mem)
 #else
-size_t malloc_usable_size(mem) Void_t* mem;
+size_t mUSABLe(mem) Void_t* mem;
 #endif
+{
   mchunkptr p;
-  if (mem == 0)
+  if (mem != 0) {
-    return 0;
-  else
-  {
     p = mem2chunk(mem);
-    if(!chunk_is_mmapped(p))
+    if (chunk_is_mmapped(p))
-    {
-      if (!inuse(p)) return 0;
+      return chunksize(p) - 2*SIZE_SZ;
-      check_inuse_chunk(p);
+    else if (inuse(p))
       return chunksize(p) - SIZE_SZ;
-    }
-    return chunksize(p) - 2*SIZE_SZ;
+  }
+  return 0;
+}
+/*
+  ------------------------------ mallinfo ------------------------------
+*/
-/* Utility to update current_mallinfo for malloc_stats and mallinfo() */
-static void malloc_update_mallinfo()
+struct mallinfo mALLINFo()
+{
+  mstate av = get_malloc_state();
+  struct mallinfo mi;
   int i;
   mbinptr b;
   mchunkptr p;
+  INTERNAL_SIZE_T avail;
+  INTERNAL_SIZE_T fastavail;
-#if DEBUG
+  int nblocks;
+  int nfastblocks;
+  /* Ensure initialization */
+  if (av->top == 0)  malloc_consolidate(av);
+  check_malloc_state();
+  /* Account for top */
+  avail = chunksize(av->top);
+  nblocks = 1;  /* top always exists */
+  /* traverse fastbins */
+  nfastblocks = 0;
-  mchunkptr q;
+  fastavail = 0;
+  for (i = 0; i < NFASTBINS; ++i) {
+    for (p = av->fastbins[i]; p != 0; p = p->fd) {
+      ++nfastblocks;
+      fastavail += chunksize(p);
-#endif
+    }
+  }
-  INTERNAL_SIZE_T avail = chunksize(top);
+  avail += fastavail;
-  int   navail = ((long)(avail) >= (long)MINSIZE)? 1 : 0;
+  /* traverse regular bins */
-  for (i = 1; i < NAV; ++i)
+  for (i = 1; i < NBINS; ++i) {
-  {
-    b = bin_at(i);
+    b = bin_at(av, i);
-    for (p = last(b); p != b; p = p->bk)
+    for (p = last(b); p != b; p = p->bk) {
-    {
-#if DEBUG
-      check_free_chunk(p);
+      ++nblocks;
-      for (q = next_chunk(p);
-           q < top && inuse(q) && (long)(chunksize(q)) >= (long)MINSIZE;
-           q = next_chunk(q))
-        check_inuse_chunk(q);
-#endif
       avail += chunksize(p);
-      navail++;
+    }
+  }
+  mi.smblks = nfastblocks;
+  mi.ordblks = nblocks;
-  current_mallinfo.ordblks = navail;
+  mi.fordblks = avail;
-  current_mallinfo.uordblks = sbrked_mem - avail;
+  mi.uordblks = av->sbrked_mem - avail;
-  current_mallinfo.fordblks = avail;
+  mi.arena = av->sbrked_mem;
-  current_mallinfo.hblks = n_mmaps;
+  mi.hblks = av->n_mmaps;
-  current_mallinfo.hblkhd = mmapped_mem;
+  mi.hblkhd = av->mmapped_mem;
+  mi.fsmblks = fastavail;
-  current_mallinfo.keepcost = chunksize(top);
+  mi.keepcost = chunksize(av->top);
+  mi.usmblks = av->max_total_mem;
-}
-/*
-  malloc_stats:
+#ifdef TRACE
+#ifdef WIN32
+  {
-    Prints on stderr the amount of space obtain from the system (both
+    CHUNK_SIZE_T  free, reserved, committed;
-    via sbrk and mmap), the maximum amount (which may be more than
+    vminfo (&free, &reserved, &committed);
-    current if malloc_trim and/or munmap got called), the maximum
+    printf("free bytes       = %10lu\n",
-    number of simultaneous mmap regions used, and the current number
+            free);
-    of bytes allocated via malloc (or realloc, etc) but not yet
+    printf("reserved bytes   = %10lu\n",
-    freed. (Note that this is the number of bytes allocated, not the
+            reserved);
-    number requested. It will be larger than the number requested
+    printf("committed bytes  = %10lu\n",
-    because of alignment and bookkeeping overhead.)
+            committed);
+  }
+#endif
-*/
-void malloc_stats()
-{
-  malloc_update_mallinfo();
-  fprintf(stderr, "max system bytes = %10u\n",
+  printf("max system bytes = %10lu\n",
-          (unsigned int)(max_total_mem));
+          (CHUNK_SIZE_T)(mi.usmblks));
-  fprintf(stderr, "system bytes     = %10u\n",
+  printf("system bytes     = %10lu\n",
-          (unsigned int)(sbrked_mem + mmapped_mem));
+          (CHUNK_SIZE_T)(mi.arena + mi.hblkhd));
-  fprintf(stderr, "in use bytes     = %10u\n",
+  printf("in use bytes     = %10lu\n",
-          (unsigned int)(current_mallinfo.uordblks + mmapped_mem));
+          (CHUNK_SIZE_T)(mi.uordblks + mi.hblkhd));
-#if HAVE_MMAP
-  fprintf(stderr, "max mmap regions = %10u\n",
-          (unsigned int)max_n_mmaps);
 #endif
+  return mi;
+}
 /*
-  mallinfo returns a copy of updated current mallinfo.
+  ------------------------------ malloc_stats ------------------------------
 */
-struct mallinfo mALLINFo()
+void mSTATs()
+{
-  malloc_update_mallinfo();
+  struct mallinfo mi = mALLINFo();
-  return current_mallinfo;
-}
+#ifdef WIN32
+  {
+    CHUNK_SIZE_T  free, reserved, committed;
+    vminfo (&free, &reserved, &committed);
+    fprintf(stderr, "free bytes       = %10lu\n",
+            free);
+    fprintf(stderr, "reserved bytes   = %10lu\n",
+            reserved);
+    fprintf(stderr, "committed bytes  = %10lu\n",
+            committed);
+  }
+#endif
-/*
+  fprintf(stderr, "max system bytes = %10lu\n",
-  mallopt:
+          (CHUNK_SIZE_T)(mi.usmblks));
+  fprintf(stderr, "system bytes     = %10lu\n",
+          (CHUNK_SIZE_T)(mi.arena + mi.hblkhd));
+  fprintf(stderr, "in use bytes     = %10lu\n",
+          (CHUNK_SIZE_T)(mi.uordblks + mi.hblkhd));
+#ifdef WIN32
-    mallopt is the general SVID/XPG interface to tunable parameters.
+/* Commented out for R Build Compatibility
+  {
+    CHUNK_SIZE_T  kernel, user;
-    The format is to provide a (parameter-number, parameter-value) pair.
+    if (cpuinfo (TRUE, &kernel, &user)) {
-    mallopt then sets the corresponding parameter to the argument
+      fprintf(stderr, "kernel ms        = %10lu\n",
+              kernel);
-    value if it can (i.e., so long as the value is meaningful),
+      fprintf(stderr, "user ms          = %10lu\n",
-    and returns 1 if successful else 0.
+              user);
+    }
+  }
+ */
+#endif
+}
-    See descriptions of tunable parameters above.
+/*
+  ------------------------------ mallopt ------------------------------
 */
 #if __STD_C
 int mALLOPt(int param_number, int value)
 #else
 int mALLOPt(param_number, value) int param_number; int value;
 #endif
+{
-  switch(param_number)
+  mstate av = get_malloc_state();
-  {
-    case M_TRIM_THRESHOLD:
-      trim_threshold = value; return 1;
+  /* Ensure initialization/consolidation */
-    case M_TOP_PAD:
+  malloc_consolidate(av);
-      top_pad = value; return 1;
+  switch(param_number) {
-    case M_MMAP_THRESHOLD:
+  case M_MXFAST:
-      mmap_threshold = value; return 1;
+    if (value >= 0 && value <= MAX_FAST_SIZE) {
-    case M_MMAP_MAX:
+      set_max_fast(av, value);
-#if HAVE_MMAP
+      return 1;
-      n_mmaps_max = value; return 1;
+    }
-#else
+    else
-      if (value != 0) return 0; else  n_mmaps_max = value; return 1;
-#endif
+      return 0;
+  case M_TRIM_THRESHOLD:
+    av->trim_threshold = value;
-    default:
+    return 1;
+  case M_TOP_PAD:
+    av->top_pad = value;
+    return 1;
+  case M_MMAP_THRESHOLD:
+    av->mmap_threshold = value;
+    return 1;
+  case M_MMAP_MAX:
+#if !HAVE_MMAP
+    if (value != 0)
       return 0;
+#endif
+    av->n_mmaps_max = value;
+    return 1;
+  default:
+    return 0;
+  }
+}
+/*
+  -------------------- Alternative MORECORE functions --------------------
+*/
 /*
+  General Requirements for MORECORE.
+  The MORECORE function must have the following properties:
+  If MORECORE_CONTIGUOUS is false:
+    * MORECORE must allocate in multiples of pagesize. It will
+      only be called with arguments that are multiples of pagesize.
+    * MORECORE(0) must return an address that is at least
+      MALLOC_ALIGNMENT aligned. (Page-aligning always suffices.)
+  else (i.e. If MORECORE_CONTIGUOUS is true):
+    * Consecutive calls to MORECORE with positive arguments
+      return increasing addresses, indicating that space has been
+      contiguously extended.
+    * MORECORE need not allocate in multiples of pagesize.
+      Calls to MORECORE need not have args of multiples of pagesize.
+    * MORECORE need not page-align.
+  In either case:
+    * MORECORE may allocate more memory than requested. (Or even less,
+      but this will generally 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. This malloc does NOT call MORECORE(0)
+      until at least one call with positive arguments is made, so
+      the initial value returned is not important.
+    * 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 MORECORE_FAILURE 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,
+  There is some variation across systems about the type of the
+  argument to sbrk/MORECORE. If size_t is unsigned, then it cannot
+  actually be size_t, because sbrk supports negative args, so it is
+  normally the signed type of the same width as size_t (sometimes
+  declared as "intptr_t", and sometimes "ptrdiff_t").  It doesn't much
+  matter though. Internally, we use "long" as arguments, which should
+  work across all reasonable possibilities.
+  Additionally, if MORECORE ever returns failure for a positive
+  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 map noncontiguous space.
+  If you'd like mmap to ALWAYS be used, you can define MORECORE to be
+  a function that always returns MORECORE_FAILURE.
+  Malloc only has limited ability to detect failures of MORECORE
+  to supply contiguous space when it says it can. In particular,
+  multithreaded programs that do not use locks may result in
+  rece conditions across calls to MORECORE that result in gaps
+  that cannot be detected as such, and subsequent corruption.
+  If you are using this malloc with something other than sbrk (or its
+  emulation) to supply memory regions, you probably want to set
+  MORECORE_CONTIGUOUS as false.  As an example, 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
+      #define MORECORE_CONTIGUOUS 0
+  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 * 1024)
+  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 *) MORECORE_FAILURE;
+      }
+      // save ptrs so they can be freed during cleanup
+      our_os_pools[next_os_pool] = ptr;
+      next_os_pool++;
+      ptr = (void *) ((((CHUNK_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 *) MORECORE_FAILURE;
+    }
+    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;
+      }
+  }
+*/
+/*
+  --------------------------------------------------------------
+  Emulation of sbrk for win32.
+  Donated by J. Walter <Walter@GeNeSys-e.de>.
+  For additional information about this code, and malloc on Win32, see
+     http://www.genesys-e.de/jwalter/
+*/
+#ifdef WIN32
+#ifdef _DEBUG
+/* #define TRACE */
+/* #define TRACESB */
+#endif
+/* #define TRACE */
+/* #define TRACESB */
+/* Support for USE_MALLOC_LOCK */
+#ifdef USE_MALLOC_LOCK
+/* Wait for spin lock */
+static int slwait (int *sl) {
+    while (InterlockedCompareExchange ((void **) sl, (void *) 1, (void *) 0) != 0)
+	    Sleep (0);
+    return 0;
+}
+/* Release spin lock */
+static int slrelease (int *sl) {
+    InterlockedExchange (sl, 0);
+    return 0;
+}
+#ifdef NEEDED
+/* Spin lock for emulation code */
+static int g_sl;
+#endif
+#endif /* USE_MALLOC_LOCK */
+/* getpagesize for windows */
+static long getpagesize (void) {
+    static long g_pagesize = 0;
+    if (! g_pagesize) {
+        SYSTEM_INFO system_info;
+        GetSystemInfo (&system_info);
+        g_pagesize = system_info.dwPageSize;
+    }
+    return g_pagesize;
+}
+static long getregionsize (void) {
+    static long g_regionsize = 0;
+    if (! g_regionsize) {
+        SYSTEM_INFO system_info;
+        GetSystemInfo (&system_info);
+        g_regionsize = system_info.dwAllocationGranularity;
+    }
+    return g_regionsize;
+}
+/* A region list entry */
+typedef struct _region_list_entry {
+    void *top_allocated;
+    void *top_committed;
+    void *top_reserved;
+    long reserve_size;
+    struct _region_list_entry *previous;
+} region_list_entry;
+/* Allocate and link a region entry in the region list */
+static int region_list_append (region_list_entry **last, void *base_reserved, long reserve_size) {
+#ifdef TRACESB
+	printf ("region_list_append %p %p %lu\n", last, base_reserved, reserve_size);
+#endif
+#ifdef DEBUG
+	assert(HeapValidate(GetProcessHeap(), 0, NULL));
+#endif
+    region_list_entry *next = HeapAlloc (GetProcessHeap (), 0, sizeof (region_list_entry));
+    if (! next)
+	{
+#ifdef TRACESB
+	    printf ("region_list_append HeapAlloc failed\n");
+#endif
+        return FALSE;
+	}
+    next->top_allocated = (char *) base_reserved;
+    next->top_committed = (char *) base_reserved;
+    next->top_reserved = (char *) base_reserved + reserve_size;
+    next->reserve_size = reserve_size;
+    next->previous = *last;
+    *last = next;
+    return TRUE;
+}
+/* Free and unlink the last region entry from the region list */
+static int region_list_remove (region_list_entry **last) {
+    region_list_entry *previous = (*last)->previous;
+#ifdef TRACESB
+	printf ("region_list_remove %p %p\n", last, previous);
+#endif
+#ifdef DEBUG
+	assert(HeapValidate(GetProcessHeap(), 0, NULL));
+#endif
+    if (! HeapFree (GetProcessHeap (), sizeof (region_list_entry), *last)) {
+#ifdef TRACESB
+	    printf ("region_list_remove HeapFree failed\n");
+#endif
+        return FALSE;
+	}
+    *last = previous;
+    return TRUE;
+}
+#define CEIL(size,to)	(((size)+(to)-1)&~((to)-1))
+#define FLOOR(size,to)	((size)&~((to)-1))
+#define SBRK_SCALE  0
+/* #define SBRK_SCALE  1 */
+/* #define SBRK_SCALE  2 */
+/* #define SBRK_SCALE  4  */
+/* sbrk for windows */
+static void *sbrk (long size) {
+    static long g_pagesize, g_my_pagesize;
+    static long g_regionsize, g_my_regionsize;
+    static region_list_entry *g_last;
+    void *result = (void *) MORECORE_FAILURE;
+#ifdef TRACESB
+    printf ("sbrk %ld\n", size);
+#endif
+#if defined (USE_MALLOC_LOCK) && defined (NEEDED)
+    /* Wait for spin lock */
+    slwait (&g_sl);
+#endif
+    /* First time initialization */
+    if (! g_pagesize) {
+        g_pagesize = getpagesize ();
+        g_my_pagesize = g_pagesize << SBRK_SCALE;
+    }
+    if (! g_regionsize) {
+        g_regionsize = getregionsize ();
+        g_my_regionsize = g_regionsize << SBRK_SCALE;
+    }
+    if (! g_last) {
+        if (! region_list_append (&g_last, 0, 0))
+           goto sbrk_exit;
+    }
+    /* Assert invariants */
+    assert (g_last);
+    assert ((char *) g_last->top_reserved - g_last->reserve_size <= (char *) g_last->top_allocated &&
+            g_last->top_allocated <= g_last->top_committed);
+    assert ((char *) g_last->top_reserved - g_last->reserve_size <= (char *) g_last->top_committed &&
+            g_last->top_committed <= g_last->top_reserved &&
+            (unsigned) g_last->top_committed % g_pagesize == 0);
+    assert ((unsigned) g_last->top_reserved % g_regionsize == 0);
+    assert ((unsigned) g_last->reserve_size % g_regionsize == 0);
+    /* Allocation requested? */
+    if (size >= 0) {
+        /* Allocation size is the requested size */
+        long allocate_size = size;
+        /* Compute the size to commit */
+        long to_commit = (char *) g_last->top_allocated + allocate_size - (char *) g_last->top_committed;
+        /* Do we reach the commit limit? */
+        if (to_commit > 0) {
+            /* Round size to commit */
+            long commit_size = CEIL (to_commit, g_my_pagesize);
+            /* Compute the size to reserve */
+            long to_reserve = (char *) g_last->top_committed + commit_size - (char *) g_last->top_reserved;
+            /* Do we reach the reserve limit? */
+            if (to_reserve > 0) {
+                /* Compute the remaining size to commit in the current region */
+                long remaining_commit_size = (char *) g_last->top_reserved - (char *) g_last->top_committed;
+                if (remaining_commit_size > 0) {
+                    /* Assert preconditions */
+                    assert ((unsigned) g_last->top_committed % g_pagesize == 0);
+                    assert (0 < remaining_commit_size && remaining_commit_size % g_pagesize == 0); {
+                        /* Commit this */
+                        void *base_committed = VirtualAlloc (g_last->top_committed, remaining_commit_size,
+							                                 MEM_COMMIT, PAGE_READWRITE);
+                        /* Check returned pointer for consistency */
+                        if (base_committed != g_last->top_committed)
+                            goto sbrk_exit;
+                        /* Assert postconditions */
+                        assert ((unsigned) base_committed % g_pagesize == 0);
+#ifdef TRACESB
+                        printf ("Commit %p %ld\n", base_committed, remaining_commit_size);
+#endif
+                        /* Adjust the regions commit top */
+                        g_last->top_committed = (char *) base_committed + remaining_commit_size;
+                    }
+                } {
+                    /* Now we are going to search and reserve. */
+                    int contiguous = -1;
+                    int found = FALSE;
+                    MEMORY_BASIC_INFORMATION memory_info;
+                    void *base_reserved;
+                    long reserve_size;
+                    do {
+                        /* Assume contiguous memory */
+                        contiguous = TRUE;
+                        /* Round size to reserve */
+                        reserve_size = CEIL (to_reserve, g_my_regionsize);
+                        /* Start with the current region's top */
+                        memory_info.BaseAddress = g_last->top_reserved;
+                        /* Assert preconditions */
+                        assert ((unsigned) memory_info.BaseAddress % g_pagesize == 0);
+                        assert (0 < reserve_size && reserve_size % g_regionsize == 0);
+                        while (VirtualQuery (memory_info.BaseAddress, &memory_info, sizeof (memory_info))) {
+                            /* Assert postconditions */
+                            assert ((unsigned) memory_info.BaseAddress % g_pagesize == 0);
+#ifdef TRACESB
+                            printf ("Query %p %ld %s\n", memory_info.BaseAddress, memory_info.RegionSize,
+                                    memory_info.State == MEM_FREE ? "FREE":
+                                    (memory_info.State == MEM_RESERVE ? "RESERVED":
+                                     (memory_info.State == MEM_COMMIT ? "COMMITTED": "?")));
+#endif
+                            /* Region is free, well aligned and big enough: we are done */
+                            if (memory_info.State == MEM_FREE &&
+                                (unsigned) memory_info.BaseAddress % g_regionsize == 0 &&
+                                memory_info.RegionSize >= (unsigned) reserve_size) {
+                                found = TRUE;
+                                break;
+                            }
+                            /* From now on we can't get contiguous memory! */
+                            contiguous = FALSE;
+                            /* Recompute size to reserve */
+                            reserve_size = CEIL (allocate_size, g_my_regionsize);
+                            memory_info.BaseAddress = (char *) memory_info.BaseAddress + memory_info.RegionSize;
+                            /* Assert preconditions */
+                            assert ((unsigned) memory_info.BaseAddress % g_pagesize == 0);
+                            assert (0 < reserve_size && reserve_size % g_regionsize == 0);
+                        }
+                        /* Search failed? */
+                        if (! found)
+                            goto sbrk_exit;
+                        /* Assert preconditions */
+                        assert ((unsigned) memory_info.BaseAddress % g_regionsize == 0);
+                        assert (0 < reserve_size && reserve_size % g_regionsize == 0);
+                        /* Try to reserve this */
+                        base_reserved = VirtualAlloc (memory_info.BaseAddress, reserve_size,
+					                                  MEM_RESERVE, PAGE_NOACCESS);
+                        if (! base_reserved) {
+                            int rc = GetLastError ();
+                            if (rc != ERROR_INVALID_ADDRESS)
+                                goto sbrk_exit;
+                        }
+                        /* A null pointer signals (hopefully) a race condition with another thread. */
+                        /* In this case, we try again. */
+                    } while (! base_reserved);
+                    /* Check returned pointer for consistency */
+                    if (memory_info.BaseAddress && base_reserved != memory_info.BaseAddress)
+                        goto sbrk_exit;
+                    /* Assert postconditions */
+                    assert ((unsigned) base_reserved % g_regionsize == 0);
+#ifdef TRACESB
+                    printf ("Reserve %p %ld\n", base_reserved, reserve_size);
+#endif
+                    /* Did we get contiguous memory? */
+                    if (contiguous) {
+                        long start_size = (char *) g_last->top_committed - (char *) g_last->top_allocated;
+                        /* Adjust allocation size */
+                        allocate_size -= start_size;
+                        /* Adjust the regions allocation top */
+                        g_last->top_allocated = g_last->top_committed;
+                        /* Recompute the size to commit */
+                        to_commit = (char *) g_last->top_allocated + allocate_size - (char *) g_last->top_committed;
+                        /* Round size to commit */
+                        commit_size = CEIL (to_commit, g_my_pagesize);
+                    }
+                    /* Append the new region to the list */
+                    if (! region_list_append (&g_last, base_reserved, reserve_size))
+                        goto sbrk_exit;
+                    /* Didn't we get contiguous memory? */
+                    if (! contiguous) {
+                        /* Recompute the size to commit */
+                        to_commit = (char *) g_last->top_allocated + allocate_size - (char *) g_last->top_committed;
+                        /* Round size to commit */
+                        commit_size = CEIL (to_commit, g_my_pagesize);
+                    }
+                }
+            }
+            /* Assert preconditions */
+            assert ((unsigned) g_last->top_committed % g_pagesize == 0);
+            assert (0 < commit_size && commit_size % g_pagesize == 0); {
+                /* Commit this */
+                void *base_committed = VirtualAlloc (g_last->top_committed, commit_size,
+				    			                     MEM_COMMIT, PAGE_READWRITE);
+                /* Check returned pointer for consistency */
+                if (base_committed != g_last->top_committed)
+                    goto sbrk_exit;
+                /* Assert postconditions */
+                assert ((unsigned) base_committed % g_pagesize == 0);
+#ifdef TRACESB
+                printf ("Commit %p %ld\n", base_committed, commit_size);
+#endif
+                /* Adjust the regions commit top */
+                g_last->top_committed = (char *) base_committed + commit_size;
+            }
+        }
+        /* Adjust the regions allocation top */
+        g_last->top_allocated = (char *) g_last->top_allocated + allocate_size;
+        result = (char *) g_last->top_allocated - size;
+    /* Deallocation requested? */
+    } else if (size < 0) {
+        long deallocate_size = - size;
+        /* As long as we have a region to release */
+        while ((char *) g_last->top_allocated - deallocate_size < (char *) g_last->top_reserved - g_last->reserve_size) {
+            /* Get the size to release */
+            long release_size = g_last->reserve_size;
+            /* Get the base address */
+            void *base_reserved = (char *) g_last->top_reserved - release_size;
+            /* Assert preconditions */
+            assert ((unsigned) base_reserved % g_regionsize == 0);
+            assert (0 < release_size && release_size % g_regionsize == 0); {
+                /* Release this */
+                int rc = VirtualFree (base_reserved, 0,
+                                      MEM_RELEASE);
+                /* Check returned code for consistency */
+                if (! rc)
+                    goto sbrk_exit;
+#ifdef TRACESB
+                printf ("Release %p %ld\n", base_reserved, release_size);
+#endif
+            }
+            /* Adjust deallocation size */
+            deallocate_size -= (char *) g_last->top_allocated - (char *) base_reserved;
+            /* Remove the old region from the list */
+            if (! region_list_remove (&g_last))
+                goto sbrk_exit;
+        } {
+            /* Compute the size to decommit */
+            long to_decommit = (char *) g_last->top_committed - ((char *) g_last->top_allocated - deallocate_size);
+            if (to_decommit >= g_my_pagesize) {
+                /* Compute the size to decommit */
+                long decommit_size = FLOOR (to_decommit, g_my_pagesize);
+                /*  Compute the base address */
+                void *base_committed = (char *) g_last->top_committed - decommit_size;
+                /* Assert preconditions */
+                assert ((unsigned) base_committed % g_pagesize == 0);
+                assert (0 < decommit_size && decommit_size % g_pagesize == 0); {
+                    /* Decommit this */
+                    int rc = VirtualFree ((char *) base_committed, decommit_size,
+                                          MEM_DECOMMIT);
+                    /* Check returned code for consistency */
+                    if (! rc)
+                        goto sbrk_exit;
+#ifdef TRACESB
+                    printf ("Decommit %p %ld\n", base_committed, decommit_size);
+#endif
+                }
+                /* Adjust deallocation size and regions commit and allocate top */
+                deallocate_size -= (char *) g_last->top_allocated - (char *) base_committed;
+                g_last->top_committed = base_committed;
+                g_last->top_allocated = base_committed;
+            }
+        }
+        /* Adjust regions allocate top */
+        g_last->top_allocated = (char *) g_last->top_allocated - deallocate_size;
+        /* Check for underflow */
+        if ((char *) g_last->top_reserved - g_last->reserve_size > (char *) g_last->top_allocated ||
+            g_last->top_allocated > g_last->top_committed) {
+            /* Adjust regions allocate top */
+            g_last->top_allocated = (char *) g_last->top_reserved - g_last->reserve_size;
+            goto sbrk_exit;
+        }
+        result = g_last->top_allocated;
+    }
+    /* Assert invariants */
+    assert (g_last);
+    assert ((char *) g_last->top_reserved - g_last->reserve_size <= (char *) g_last->top_allocated &&
+            g_last->top_allocated <= g_last->top_committed);
+    assert ((char *) g_last->top_reserved - g_last->reserve_size <= (char *) g_last->top_committed &&
+            g_last->top_committed <= g_last->top_reserved &&
+            (unsigned) g_last->top_committed % g_pagesize == 0);
+    assert ((unsigned) g_last->top_reserved % g_regionsize == 0);
+    assert ((unsigned) g_last->reserve_size % g_regionsize == 0);
+sbrk_exit:
+#if defined (USE_MALLOC_LOCK) && defined (NEEDED)
+    /* Release spin lock */
+    slrelease (&g_sl);
+#endif
+    return result;
+}
+/* mmap for windows */
+static void *mmap (void *ptr, long size, long prot, long type, long handle, long arg) {
+    static long g_pagesize;
+    static long g_regionsize;
+#ifdef TRACESB
+    printf ("mmap %ld\n", size);
+#endif
+#if defined (USE_MALLOC_LOCK) && defined (NEEDED)
+    /* Wait for spin lock */
+    slwait (&g_sl);
+#endif
+    /* First time initialization */
+    if (! g_pagesize)
+        g_pagesize = getpagesize ();
+    if (! g_regionsize)
+        g_regionsize = getregionsize ();
+    /* Assert preconditions */
+    assert ((unsigned) ptr % g_regionsize == 0);
+    assert (size % g_pagesize == 0);
+    /* Allocate this */
+    ptr = VirtualAlloc (ptr, size,
+					    MEM_RESERVE | MEM_COMMIT | MEM_TOP_DOWN, PAGE_READWRITE);
+    if (! ptr) {
+        ptr = (void *) MORECORE_FAILURE;
+        goto mmap_exit;
+    }
+    /* Assert postconditions */
+    assert ((unsigned) ptr % g_regionsize == 0);
+#ifdef TRACESB
+    printf ("Commit %p %ld\n", ptr, size);
+#endif
+mmap_exit:
+#if defined (USE_MALLOC_LOCK) && defined (NEEDED)
+    /* Release spin lock */
+    slrelease (&g_sl);
+#endif
+    return ptr;
+}
+/* munmap for windows */
+static long munmap (void *ptr, long size) {
+    static long g_pagesize;
+    static long g_regionsize;
+    int rc = MUNMAP_FAILURE;
+#ifdef TRACESB
+    printf ("munmap %p %ld\n", ptr, size);
+#endif
+#if defined (USE_MALLOC_LOCK) && defined (NEEDED)
+    /* Wait for spin lock */
+    slwait (&g_sl);
+#endif
+    /* First time initialization */
+    if (! g_pagesize)
+        g_pagesize = getpagesize ();
+    if (! g_regionsize)
+        g_regionsize = getregionsize ();
+    /* Assert preconditions */
+    assert ((unsigned) ptr % g_regionsize == 0);
+    assert (size % g_pagesize == 0);
+    /* Free this */
+    if (! VirtualFree (ptr, 0,
+                       MEM_RELEASE))
+        goto munmap_exit;
+    rc = 0;
+#ifdef TRACESB
+    printf ("Release %p %ld\n", ptr, size);
+#endif
+munmap_exit:
+#if defined (USE_MALLOC_LOCK) && defined (NEEDED)
+    /* Release spin lock */
+    slrelease (&g_sl);
+#endif
+    return rc;
+}
+static void vminfo (CHUNK_SIZE_T  *free, CHUNK_SIZE_T  *reserved, CHUNK_SIZE_T  *committed) {
+    MEMORY_BASIC_INFORMATION memory_info;
+    memory_info.BaseAddress = 0;
+    *free = *reserved = *committed = 0;
+    while (VirtualQuery (memory_info.BaseAddress, &memory_info, sizeof (memory_info))) {
+        switch (memory_info.State) {
+        case MEM_FREE:
+            *free += memory_info.RegionSize;
+            break;
+        case MEM_RESERVE:
+            *reserved += memory_info.RegionSize;
+            break;
+        case MEM_COMMIT:
+            *committed += memory_info.RegionSize;
+            break;
+        }
+        memory_info.BaseAddress = (char *) memory_info.BaseAddress + memory_info.RegionSize;
+    }
+}
+/* Commented out for R Build Compatibility
+static int cpuinfo (int whole, CHUNK_SIZE_T  *kernel, CHUNK_SIZE_T  *user) {
+    if (whole) {
+        __int64 creation64, exit64, kernel64, user64;
+        int rc = GetProcessTimes (GetCurrentProcess (),
+                                  (FILETIME *) &creation64,
+                                  (FILETIME *) &exit64,
+                                  (FILETIME *) &kernel64,
+                                  (FILETIME *) &user64);
+        if (! rc) {
+            *kernel = 0;
+            *user = 0;
+            return FALSE;
+        }
+        *kernel = (CHUNK_SIZE_T) (kernel64 / 10000);
+        *user = (CHUNK_SIZE_T) (user64 / 10000);
+        return TRUE;
+    } else {
+        __int64 creation64, exit64, kernel64, user64;
+        int rc = GetThreadTimes (GetCurrentThread (),
+                                 (FILETIME *) &creation64,
+                                 (FILETIME *) &exit64,
+                                 (FILETIME *) &kernel64,
+                                 (FILETIME *) &user64);
+        if (! rc) {
+            *kernel = 0;
+            *user = 0;
+            return FALSE;
+        }
+        *kernel = (CHUNK_SIZE_T) (kernel64 / 10000);
+        *user = (CHUNK_SIZE_T) (user64 / 10000);
+        return TRUE;
+    }
+}
+*/
+#endif /* WIN32 */
+/* ------------------------------------------------------------
 History:
+    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@yahoo.com>
+      * 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 up header file inclusion for 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
+           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)
       * Based loosely on libg++-1.2X malloc. (It retains some of the overall
          structure of old version,  but most details differ.)
 */
+/* Need strdup here so we don't get calls to the wrong heap function inside the library */
 char * strdup (const char *str)
+{
   char *newstr;
   newstr = (char *) malloc (strlen (str) + 1);

The R Project SVN R

(root)/trunk/src/gnuwin32/malloc.c @ 28362 – Rev 20730 → 26855