about summary refs log tree commit diff
path: root/malloc/arena.c
blob: 9018a4f0e9f3b871a09a282cbfb0cd7eb83c7c56 (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
/* Malloc implementation for multiple threads without lock contention.
   Copyright (C) 2001, 2002, 2003, 2004 Free Software Foundation, Inc.
   This file is part of the GNU C Library.
   Contributed by Wolfram Gloger <wg@malloc.de>, 2001.

   The GNU C Library is free software; you can redistribute it and/or
   modify it under the terms of the GNU Lesser General Public License as
   published by the Free Software Foundation; either version 2.1 of the
   License, or (at your option) any later version.

   The GNU C Library is distributed in the hope that it will be useful,
   but WITHOUT ANY WARRANTY; without even the implied warranty of
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
   Lesser General Public License for more details.

   You should have received a copy of the GNU Lesser General Public
   License along with the GNU C Library; see the file COPYING.LIB.  If not,
   write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
   Boston, MA 02111-1307, USA.  */

/* $Id$ */

/* Compile-time constants.  */

#define HEAP_MIN_SIZE (32*1024)
#ifndef HEAP_MAX_SIZE
#define HEAP_MAX_SIZE (1024*1024) /* must be a power of two */
#endif

/* HEAP_MIN_SIZE and HEAP_MAX_SIZE limit the size of mmap()ed heaps
   that are dynamically created for multi-threaded programs.  The
   maximum size must be a power of two, for fast determination of
   which heap belongs to a chunk.  It should be much larger than the
   mmap threshold, so that requests with a size just below that
   threshold can be fulfilled without creating too many heaps.  */


#ifndef THREAD_STATS
#define THREAD_STATS 0
#endif

/* If THREAD_STATS is non-zero, some statistics on mutex locking are
   computed.  */

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

#define top(ar_ptr) ((ar_ptr)->top)

/* A heap is a single contiguous memory region holding (coalesceable)
   malloc_chunks.  It is allocated with mmap() and always starts at an
   address aligned to HEAP_MAX_SIZE.  Not used unless compiling with
   USE_ARENAS. */

typedef struct _heap_info {
  mstate ar_ptr; /* Arena for this heap. */
  struct _heap_info *prev; /* Previous heap. */
  size_t size;   /* Current size in bytes. */
  size_t pad;    /* Make sure the following data is properly aligned. */
} heap_info;

/* Thread specific data */

static tsd_key_t arena_key;
static mutex_t list_lock;

#if THREAD_STATS
static int stat_n_heaps;
#define THREAD_STAT(x) x
#else
#define THREAD_STAT(x) do ; while(0)
#endif

/* Mapped memory in non-main arenas (reliable only for NO_THREADS). */
static unsigned long arena_mem;

/* Already initialized? */
int __malloc_initialized = -1;

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

#if USE_ARENAS

/* arena_get() acquires an arena and locks the corresponding mutex.
   First, try the one last locked successfully by this thread.  (This
   is the common case and handled with a macro for speed.)  Then, loop
   once over the circularly linked list of arenas.  If no arena is
   readily available, create a new one.  In this latter case, `size'
   is just a hint as to how much memory will be required immediately
   in the new arena. */

#define arena_get(ptr, size) do { \
  Void_t *vptr = NULL; \
  ptr = (mstate)tsd_getspecific(arena_key, vptr); \
  if(ptr && !mutex_trylock(&ptr->mutex)) { \
    THREAD_STAT(++(ptr->stat_lock_direct)); \
  } else \
    ptr = arena_get2(ptr, (size)); \
} while(0)

/* find the heap and corresponding arena for a given ptr */

#define heap_for_ptr(ptr) \
 ((heap_info *)((unsigned long)(ptr) & ~(HEAP_MAX_SIZE-1)))
#define arena_for_chunk(ptr) \
 (chunk_non_main_arena(ptr) ? heap_for_ptr(ptr)->ar_ptr : &main_arena)

#else /* !USE_ARENAS */

/* There is only one arena, main_arena. */

#if THREAD_STATS
#define arena_get(ar_ptr, sz) do { \
  ar_ptr = &main_arena; \
  if(!mutex_trylock(&ar_ptr->mutex)) \
    ++(ar_ptr->stat_lock_direct); \
  else { \
    (void)mutex_lock(&ar_ptr->mutex); \
    ++(ar_ptr->stat_lock_wait); \
  } \
} while(0)
#else
#define arena_get(ar_ptr, sz) do { \
  ar_ptr = &main_arena; \
  (void)mutex_lock(&ar_ptr->mutex); \
} while(0)
#endif
#define arena_for_chunk(ptr) (&main_arena)

#endif /* USE_ARENAS */

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

#ifndef NO_THREADS

/* atfork support.  */

static __malloc_ptr_t (*save_malloc_hook) (size_t __size,
					   __const __malloc_ptr_t);
# if !defined _LIBC || !defined USE_TLS || (defined SHARED && !USE___THREAD)
static __malloc_ptr_t (*save_memalign_hook) (size_t __align, size_t __size,
					     __const __malloc_ptr_t);
# endif
static void           (*save_free_hook) (__malloc_ptr_t __ptr,
					 __const __malloc_ptr_t);
static Void_t*        save_arena;

/* Magic value for the thread-specific arena pointer when
   malloc_atfork() is in use.  */

#define ATFORK_ARENA_PTR ((Void_t*)-1)

/* The following hooks are used while the `atfork' handling mechanism
   is active. */

static Void_t*
malloc_atfork(size_t sz, const Void_t *caller)
{
  Void_t *vptr = NULL;
  Void_t *victim;

  tsd_getspecific(arena_key, vptr);
  if(vptr == ATFORK_ARENA_PTR) {
    /* We are the only thread that may allocate at all.  */
    if(save_malloc_hook != malloc_check) {
      return _int_malloc(&main_arena, sz);
    } else {
      if(top_check()<0)
        return 0;
      victim = _int_malloc(&main_arena, sz+1);
      return mem2mem_check(victim, sz);
    }
  } else {
    /* Suspend the thread until the `atfork' handlers have completed.
       By that time, the hooks will have been reset as well, so that
       mALLOc() can be used again. */
    (void)mutex_lock(&list_lock);
    (void)mutex_unlock(&list_lock);
    return public_mALLOc(sz);
  }
}

static void
free_atfork(Void_t* mem, const Void_t *caller)
{
  Void_t *vptr = NULL;
  mstate ar_ptr;
  mchunkptr p;                          /* chunk corresponding to mem */

  if (mem == 0)                              /* free(0) has no effect */
    return;

  p = mem2chunk(mem);         /* do not bother to replicate free_check here */

#if HAVE_MMAP
  if (chunk_is_mmapped(p))                       /* release mmapped memory. */
  {
    munmap_chunk(p);
    return;
  }
#endif

  ar_ptr = arena_for_chunk(p);
  tsd_getspecific(arena_key, vptr);
  if(vptr != ATFORK_ARENA_PTR)
    (void)mutex_lock(&ar_ptr->mutex);
  _int_free(ar_ptr, mem);
  if(vptr != ATFORK_ARENA_PTR)
    (void)mutex_unlock(&ar_ptr->mutex);
}

/* The following two functions are registered via thread_atfork() to
   make sure that the mutexes remain in a consistent state in the
   fork()ed version of a thread.  Also adapt the malloc and free hooks
   temporarily, because the `atfork' handler mechanism may use
   malloc/free internally (e.g. in LinuxThreads). */

static void
ptmalloc_lock_all (void)
{
  mstate ar_ptr;

  if(__malloc_initialized < 1)
    return;
  (void)mutex_lock(&list_lock);
  for(ar_ptr = &main_arena;;) {
    (void)mutex_lock(&ar_ptr->mutex);
    ar_ptr = ar_ptr->next;
    if(ar_ptr == &main_arena) break;
  }
  save_malloc_hook = __malloc_hook;
  save_free_hook = __free_hook;
  __malloc_hook = malloc_atfork;
  __free_hook = free_atfork;
  /* Only the current thread may perform malloc/free calls now. */
  tsd_getspecific(arena_key, save_arena);
  tsd_setspecific(arena_key, ATFORK_ARENA_PTR);
}

static void
ptmalloc_unlock_all (void)
{
  mstate ar_ptr;

  if(__malloc_initialized < 1)
    return;
  tsd_setspecific(arena_key, save_arena);
  __malloc_hook = save_malloc_hook;
  __free_hook = save_free_hook;
  for(ar_ptr = &main_arena;;) {
    (void)mutex_unlock(&ar_ptr->mutex);
    ar_ptr = ar_ptr->next;
    if(ar_ptr == &main_arena) break;
  }
  (void)mutex_unlock(&list_lock);
}

#ifdef __linux__

/* In LinuxThreads, unlocking a mutex in the child process after a
   fork() is currently unsafe, whereas re-initializing it is safe and
   does not leak resources.  Therefore, a special atfork handler is
   installed for the child. */

static void
ptmalloc_unlock_all2 (void)
{
  mstate ar_ptr;

  if(__malloc_initialized < 1)
    return;
#if defined _LIBC || defined MALLOC_HOOKS
  tsd_setspecific(arena_key, save_arena);
  __malloc_hook = save_malloc_hook;
  __free_hook = save_free_hook;
#endif
  for(ar_ptr = &main_arena;;) {
    mutex_init(&ar_ptr->mutex);
    ar_ptr = ar_ptr->next;
    if(ar_ptr == &main_arena) break;
  }
  mutex_init(&list_lock);
}

#else

#define ptmalloc_unlock_all2 ptmalloc_unlock_all

#endif

#endif /* !defined NO_THREADS */

/* Initialization routine. */
#ifdef _LIBC
#include <string.h>
extern char **_environ;

static char *
internal_function
next_env_entry (char ***position)
{
  char **current = *position;
  char *result = NULL;

  while (*current != NULL)
    {
      if (__builtin_expect ((*current)[0] == 'M', 0)
	  && (*current)[1] == 'A'
	  && (*current)[2] == 'L'
	  && (*current)[3] == 'L'
	  && (*current)[4] == 'O'
	  && (*current)[5] == 'C'
	  && (*current)[6] == '_')
	{
	  result = &(*current)[7];

	  /* Save current position for next visit.  */
	  *position = ++current;

	  break;
	}

      ++current;
    }

  return result;
}
#endif /* _LIBC */

/* Set up basic state so that _int_malloc et al can work.  */
static void
ptmalloc_init_minimal (void)
{
#if DEFAULT_TOP_PAD != 0
  mp_.top_pad        = DEFAULT_TOP_PAD;
#endif
  mp_.n_mmaps_max    = DEFAULT_MMAP_MAX;
  mp_.mmap_threshold = DEFAULT_MMAP_THRESHOLD;
  mp_.trim_threshold = DEFAULT_TRIM_THRESHOLD;
  mp_.pagesize       = malloc_getpagesize;
}


#ifdef _LIBC
# ifdef SHARED
static void *
__failing_morecore (ptrdiff_t d)
{
  return (void *) MORECORE_FAILURE;
}
# endif

# if defined SHARED && defined USE_TLS && !USE___THREAD
# include <stdbool.h>

/* This is called by __pthread_initialize_minimal when it needs to use
   malloc to set up the TLS state.  We cannot do the full work of
   ptmalloc_init (below) until __pthread_initialize_minimal has finished,
   so it has to switch to using the special startup-time hooks while doing
   those allocations.  */
void
__libc_malloc_pthread_startup (bool first_time)
{
  if (first_time)
    {
      ptmalloc_init_minimal ();
      save_malloc_hook = __malloc_hook;
      save_memalign_hook = __memalign_hook;
      save_free_hook = __free_hook;
      __malloc_hook = malloc_starter;
      __memalign_hook = memalign_starter;
      __free_hook = free_starter;
    }
  else
    {
      __malloc_hook = save_malloc_hook;
      __memalign_hook = save_memalign_hook;
      __free_hook = save_free_hook;
    }
}
# endif
#endif

static void
ptmalloc_init (void)
{
#if __STD_C
  const char* s;
#else
  char* s;
#endif
  int secure = 0;

  if(__malloc_initialized >= 0) return;
  __malloc_initialized = 0;

#ifdef _LIBC
# if defined SHARED && defined USE_TLS && !USE___THREAD
  /* ptmalloc_init_minimal may already have been called via
     __libc_malloc_pthread_startup, above.  */
  if (mp_.pagesize == 0)
# endif
#endif
    ptmalloc_init_minimal();

#ifndef NO_THREADS
# if defined _LIBC && defined USE_TLS
  /* We know __pthread_initialize_minimal has already been called,
     and that is enough.  */
#   define NO_STARTER
# endif
# ifndef NO_STARTER
  /* With some threads implementations, creating thread-specific data
     or initializing a mutex may call malloc() itself.  Provide a
     simple starter version (realloc() won't work). */
  save_malloc_hook = __malloc_hook;
  save_memalign_hook = __memalign_hook;
  save_free_hook = __free_hook;
  __malloc_hook = malloc_starter;
  __memalign_hook = memalign_starter;
  __free_hook = free_starter;
#  ifdef _LIBC
  /* Initialize the pthreads interface. */
  if (__pthread_initialize != NULL)
    __pthread_initialize();
#  endif /* !defined _LIBC */
# endif	/* !defined NO_STARTER */
#endif /* !defined NO_THREADS */
  mutex_init(&main_arena.mutex);
  main_arena.next = &main_arena;

#if defined _LIBC && defined SHARED
  /* In case this libc copy is in a non-default namespace, never use brk.  */
  Dl_info di;
  struct link_map *l;
  if (_dl_addr (ptmalloc_init, &di, &l, NULL) != 0 && l->l_ns != LM_ID_BASE)
    __morecore = __failing_morecore;
#endif

  mutex_init(&list_lock);
  tsd_key_create(&arena_key, NULL);
  tsd_setspecific(arena_key, (Void_t *)&main_arena);
  thread_atfork(ptmalloc_lock_all, ptmalloc_unlock_all, ptmalloc_unlock_all2);
#ifndef NO_THREADS
# ifndef NO_STARTER
  __malloc_hook = save_malloc_hook;
  __memalign_hook = save_memalign_hook;
  __free_hook = save_free_hook;
# else
#  undef NO_STARTER
# endif
#endif
#ifdef _LIBC
  secure = __libc_enable_secure;
  s = NULL;
  if (__builtin_expect (_environ != NULL, 1))
    {
      char **runp = _environ;
      char *envline;

      while (__builtin_expect ((envline = next_env_entry (&runp)) != NULL,
			       0))
	{
	  size_t len = strcspn (envline, "=");

	  if (envline[len] != '=')
	    /* This is a "MALLOC_" variable at the end of the string
	       without a '=' character.  Ignore it since otherwise we
	       will access invalid memory below.  */
	    continue;

	  switch (len)
	    {
	    case 6:
	      if (memcmp (envline, "CHECK_", 6) == 0)
		s = &envline[7];
	      break;
	    case 8:
	      if (! secure && memcmp (envline, "TOP_PAD_", 8) == 0)
		mALLOPt(M_TOP_PAD, atoi(&envline[9]));
	      break;
	    case 9:
	      if (! secure && memcmp (envline, "MMAP_MAX_", 9) == 0)
		mALLOPt(M_MMAP_MAX, atoi(&envline[10]));
	      break;
	    case 15:
	      if (! secure)
		{
		  if (memcmp (envline, "TRIM_THRESHOLD_", 15) == 0)
		    mALLOPt(M_TRIM_THRESHOLD, atoi(&envline[16]));
		  else if (memcmp (envline, "MMAP_THRESHOLD_", 15) == 0)
		    mALLOPt(M_MMAP_THRESHOLD, atoi(&envline[16]));
		}
	      break;
	    default:
	      break;
	    }
	}
    }
#else
  if (! secure)
    {
      if((s = getenv("MALLOC_TRIM_THRESHOLD_")))
	mALLOPt(M_TRIM_THRESHOLD, atoi(s));
      if((s = getenv("MALLOC_TOP_PAD_")))
	mALLOPt(M_TOP_PAD, atoi(s));
      if((s = getenv("MALLOC_MMAP_THRESHOLD_")))
	mALLOPt(M_MMAP_THRESHOLD, atoi(s));
      if((s = getenv("MALLOC_MMAP_MAX_")))
	mALLOPt(M_MMAP_MAX, atoi(s));
    }
  s = getenv("MALLOC_CHECK_");
#endif
  if(s) {
    if(s[0]) mALLOPt(M_CHECK_ACTION, (int)(s[0] - '0'));
    if (check_action != 0)
      __malloc_check_init();
  }
  if(__malloc_initialize_hook != NULL)
    (*__malloc_initialize_hook)();
  __malloc_initialized = 1;
}

/* There are platforms (e.g. Hurd) with a link-time hook mechanism. */
#ifdef thread_atfork_static
thread_atfork_static(ptmalloc_lock_all, ptmalloc_unlock_all, \
                     ptmalloc_unlock_all2)
#endif



/* Managing heaps and arenas (for concurrent threads) */

#if USE_ARENAS

#if MALLOC_DEBUG > 1

/* Print the complete contents of a single heap to stderr. */

static void
#if __STD_C
dump_heap(heap_info *heap)
#else
dump_heap(heap) heap_info *heap;
#endif
{
  char *ptr;
  mchunkptr p;

  fprintf(stderr, "Heap %p, size %10lx:\n", heap, (long)heap->size);
  ptr = (heap->ar_ptr != (mstate)(heap+1)) ?
    (char*)(heap + 1) : (char*)(heap + 1) + sizeof(struct malloc_state);
  p = (mchunkptr)(((unsigned long)ptr + MALLOC_ALIGN_MASK) &
                  ~MALLOC_ALIGN_MASK);
  for(;;) {
    fprintf(stderr, "chunk %p size %10lx", p, (long)p->size);
    if(p == top(heap->ar_ptr)) {
      fprintf(stderr, " (top)\n");
      break;
    } else if(p->size == (0|PREV_INUSE)) {
      fprintf(stderr, " (fence)\n");
      break;
    }
    fprintf(stderr, "\n");
    p = next_chunk(p);
  }
}

#endif /* MALLOC_DEBUG > 1 */

/* If consecutive mmap (0, HEAP_MAX_SIZE << 1, ...) calls return decreasing
   addresses as opposed to increasing, new_heap would badly fragment the
   address space.  In that case remember the second HEAP_MAX_SIZE part
   aligned to HEAP_MAX_SIZE from last mmap (0, HEAP_MAX_SIZE << 1, ...)
   call (if it is already aligned) and try to reuse it next time.  We need
   no locking for it, as kernel ensures the atomicity for us - worst case
   we'll call mmap (addr, HEAP_MAX_SIZE, ...) for some value of addr in
   multiple threads, but only one will succeed.  */
static char *aligned_heap_area;

/* Create a new heap.  size is automatically rounded up to a multiple
   of the page size. */

static heap_info *
internal_function
#if __STD_C
new_heap(size_t size, size_t top_pad)
#else
new_heap(size, top_pad) size_t size, top_pad;
#endif
{
  size_t page_mask = malloc_getpagesize - 1;
  char *p1, *p2;
  unsigned long ul;
  heap_info *h;

  if(size+top_pad < HEAP_MIN_SIZE)
    size = HEAP_MIN_SIZE;
  else if(size+top_pad <= HEAP_MAX_SIZE)
    size += top_pad;
  else if(size > HEAP_MAX_SIZE)
    return 0;
  else
    size = HEAP_MAX_SIZE;
  size = (size + page_mask) & ~page_mask;

  /* A memory region aligned to a multiple of HEAP_MAX_SIZE is needed.
     No swap space needs to be reserved for the following large
     mapping (on Linux, this is the case for all non-writable mappings
     anyway). */
  p2 = MAP_FAILED;
  if(aligned_heap_area) {
    p2 = (char *)MMAP(aligned_heap_area, HEAP_MAX_SIZE, PROT_NONE,
		      MAP_PRIVATE|MAP_NORESERVE);
    aligned_heap_area = NULL;
    if (p2 != MAP_FAILED && ((unsigned long)p2 & (HEAP_MAX_SIZE-1))) {
      munmap(p2, HEAP_MAX_SIZE);
      p2 = MAP_FAILED;
    }
  }
  if(p2 == MAP_FAILED) {
    p1 = (char *)MMAP(0, HEAP_MAX_SIZE<<1, PROT_NONE,
		      MAP_PRIVATE|MAP_NORESERVE);
    if(p1 != MAP_FAILED) {
      p2 = (char *)(((unsigned long)p1 + (HEAP_MAX_SIZE-1))
		    & ~(HEAP_MAX_SIZE-1));
      ul = p2 - p1;
      if (ul)
	munmap(p1, ul);
      else
	aligned_heap_area = p2 + HEAP_MAX_SIZE;
      munmap(p2 + HEAP_MAX_SIZE, HEAP_MAX_SIZE - ul);
    } else {
      /* Try to take the chance that an allocation of only HEAP_MAX_SIZE
	 is already aligned. */
      p2 = (char *)MMAP(0, HEAP_MAX_SIZE, PROT_NONE, MAP_PRIVATE|MAP_NORESERVE);
      if(p2 == MAP_FAILED)
	return 0;
      if((unsigned long)p2 & (HEAP_MAX_SIZE-1)) {
	munmap(p2, HEAP_MAX_SIZE);
	return 0;
      }
    }
  }
  if(mprotect(p2, size, PROT_READ|PROT_WRITE) != 0) {
    munmap(p2, HEAP_MAX_SIZE);
    return 0;
  }
  h = (heap_info *)p2;
  h->size = size;
  THREAD_STAT(stat_n_heaps++);
  return h;
}

/* Grow or shrink a heap.  size is automatically rounded up to a
   multiple of the page size if it is positive. */

static int
#if __STD_C
grow_heap(heap_info *h, long diff)
#else
grow_heap(h, diff) heap_info *h; long diff;
#endif
{
  size_t page_mask = malloc_getpagesize - 1;
  long new_size;

  if(diff >= 0) {
    diff = (diff + page_mask) & ~page_mask;
    new_size = (long)h->size + diff;
    if(new_size > HEAP_MAX_SIZE)
      return -1;
    if(mprotect((char *)h + h->size, diff, PROT_READ|PROT_WRITE) != 0)
      return -2;
  } else {
    new_size = (long)h->size + diff;
    if(new_size < (long)sizeof(*h))
      return -1;
    /* Try to re-map the extra heap space freshly to save memory, and
       make it inaccessible. */
    if((char *)MMAP((char *)h + new_size, -diff, PROT_NONE,
                    MAP_PRIVATE|MAP_FIXED) == (char *) MAP_FAILED)
      return -2;
    /*fprintf(stderr, "shrink %p %08lx\n", h, new_size);*/
  }
  h->size = new_size;
  return 0;
}

/* Delete a heap. */

#define delete_heap(heap) \
  do {								\
    if ((char *)(heap) + HEAP_MAX_SIZE == aligned_heap_area)	\
      aligned_heap_area = NULL;					\
    munmap((char*)(heap), HEAP_MAX_SIZE);			\
  } while (0)

static int
internal_function
#if __STD_C
heap_trim(heap_info *heap, size_t pad)
#else
heap_trim(heap, pad) heap_info *heap; size_t pad;
#endif
{
  mstate ar_ptr = heap->ar_ptr;
  unsigned long pagesz = mp_.pagesize;
  mchunkptr top_chunk = top(ar_ptr), p, bck, fwd;
  heap_info *prev_heap;
  long new_size, top_size, extra;

  /* Can this heap go away completely? */
  while(top_chunk == chunk_at_offset(heap, sizeof(*heap))) {
    prev_heap = heap->prev;
    p = chunk_at_offset(prev_heap, prev_heap->size - (MINSIZE-2*SIZE_SZ));
    assert(p->size == (0|PREV_INUSE)); /* must be fencepost */
    p = prev_chunk(p);
    new_size = chunksize(p) + (MINSIZE-2*SIZE_SZ);
    assert(new_size>0 && new_size<(long)(2*MINSIZE));
    if(!prev_inuse(p))
      new_size += p->prev_size;
    assert(new_size>0 && new_size<HEAP_MAX_SIZE);
    if(new_size + (HEAP_MAX_SIZE - prev_heap->size) < pad + MINSIZE + pagesz)
      break;
    ar_ptr->system_mem -= heap->size;
    arena_mem -= heap->size;
    delete_heap(heap);
    heap = prev_heap;
    if(!prev_inuse(p)) { /* consolidate backward */
      p = prev_chunk(p);
      unlink(p, bck, fwd);
    }
    assert(((unsigned long)((char*)p + new_size) & (pagesz-1)) == 0);
    assert( ((char*)p + new_size) == ((char*)heap + heap->size) );
    top(ar_ptr) = top_chunk = p;
    set_head(top_chunk, new_size | PREV_INUSE);
    /*check_chunk(ar_ptr, top_chunk);*/
  }
  top_size = chunksize(top_chunk);
  extra = ((top_size - pad - MINSIZE + (pagesz-1))/pagesz - 1) * pagesz;
  if(extra < (long)pagesz)
    return 0;
  /* Try to shrink. */
  if(grow_heap(heap, -extra) != 0)
    return 0;
  ar_ptr->system_mem -= extra;
  arena_mem -= extra;

  /* Success. Adjust top accordingly. */
  set_head(top_chunk, (top_size - extra) | PREV_INUSE);
  /*check_chunk(ar_ptr, top_chunk);*/
  return 1;
}

static mstate
internal_function
#if __STD_C
arena_get2(mstate a_tsd, size_t size)
#else
arena_get2(a_tsd, size) mstate a_tsd; size_t size;
#endif
{
  mstate a;

  if(!a_tsd)
    a = a_tsd = &main_arena;
  else {
    a = a_tsd->next;
    if(!a) {
      /* This can only happen while initializing the new arena. */
      (void)mutex_lock(&main_arena.mutex);
      THREAD_STAT(++(main_arena.stat_lock_wait));
      return &main_arena;
    }
  }

  /* Check the global, circularly linked list for available arenas. */
 repeat:
  do {
    if(!mutex_trylock(&a->mutex)) {
      THREAD_STAT(++(a->stat_lock_loop));
      tsd_setspecific(arena_key, (Void_t *)a);
      return a;
    }
    a = a->next;
  } while(a != a_tsd);

  /* If not even the list_lock can be obtained, try again.  This can
     happen during `atfork', or for example on systems where thread
     creation makes it temporarily impossible to obtain _any_
     locks. */
  if(mutex_trylock(&list_lock)) {
    a = a_tsd;
    goto repeat;
  }
  (void)mutex_unlock(&list_lock);

  /* Nothing immediately available, so generate a new arena.  */
  a = _int_new_arena(size);
  if(!a)
    return 0;

  tsd_setspecific(arena_key, (Void_t *)a);
  mutex_init(&a->mutex);
  mutex_lock(&a->mutex); /* remember result */

  /* Add the new arena to the global list.  */
  (void)mutex_lock(&list_lock);
  a->next = main_arena.next;
  atomic_write_barrier ();
  main_arena.next = a;
  (void)mutex_unlock(&list_lock);

  THREAD_STAT(++(a->stat_lock_loop));
  return a;
}

/* Create a new arena with initial size "size".  */

mstate
_int_new_arena(size_t size)
{
  mstate a;
  heap_info *h;
  char *ptr;
  unsigned long misalign;

  h = new_heap(size + (sizeof(*h) + sizeof(*a) + MALLOC_ALIGNMENT),
	       mp_.top_pad);
  if(!h) {
    /* Maybe size is too large to fit in a single heap.  So, just try
       to create a minimally-sized arena and let _int_malloc() attempt
       to deal with the large request via mmap_chunk().  */
    h = new_heap(sizeof(*h) + sizeof(*a) + MALLOC_ALIGNMENT, mp_.top_pad);
    if(!h)
      return 0;
  }
  a = h->ar_ptr = (mstate)(h+1);
  malloc_init_state(a);
  /*a->next = NULL;*/
  a->system_mem = a->max_system_mem = h->size;
  arena_mem += h->size;
#ifdef NO_THREADS
  if((unsigned long)(mp_.mmapped_mem + arena_mem + main_arena.system_mem) >
     mp_.max_total_mem)
    mp_.max_total_mem = mp_.mmapped_mem + arena_mem + main_arena.system_mem;
#endif

  /* Set up the top chunk, with proper alignment. */
  ptr = (char *)(a + 1);
  misalign = (unsigned long)chunk2mem(ptr) & MALLOC_ALIGN_MASK;
  if (misalign > 0)
    ptr += MALLOC_ALIGNMENT - misalign;
  top(a) = (mchunkptr)ptr;
  set_head(top(a), (((char*)h + h->size) - ptr) | PREV_INUSE);

  return a;
}

#endif /* USE_ARENAS */

/*
 * Local variables:
 * c-basic-offset: 2
 * End:
 */