1 #include <common.h> 2 3 #if 0 /* Moved to malloc.h */ 4 /* ---------- To make a malloc.h, start cutting here ------------ */ 5 6 /* 7 A version of malloc/free/realloc written by Doug Lea and released to the 8 public domain. Send questions/comments/complaints/performance data 9 to dl@cs.oswego.edu 10 11 * VERSION 2.6.6 Sun Mar 5 19:10:03 2000 Doug Lea (dl at gee) 12 13 Note: There may be an updated version of this malloc obtainable at 14 ftp://g.oswego.edu/pub/misc/malloc.c 15 Check before installing! 16 17 * Why use this malloc? 18 19 This is not the fastest, most space-conserving, most portable, or 20 most tunable malloc ever written. However it is among the fastest 21 while also being among the most space-conserving, portable and tunable. 22 Consistent balance across these factors results in a good general-purpose 23 allocator. For a high-level description, see 24 http://g.oswego.edu/dl/html/malloc.html 25 26 * Synopsis of public routines 27 28 (Much fuller descriptions are contained in the program documentation below.) 29 30 malloc(size_t n); 31 Return a pointer to a newly allocated chunk of at least n bytes, or null 32 if no space is available. 33 free(Void_t* p); 34 Release the chunk of memory pointed to by p, or no effect if p is null. 35 realloc(Void_t* p, size_t n); 36 Return a pointer to a chunk of size n that contains the same data 37 as does chunk p up to the minimum of (n, p's size) bytes, or null 38 if no space is available. The returned pointer may or may not be 39 the same as p. If p is null, equivalent to malloc. Unless the 40 #define REALLOC_ZERO_BYTES_FREES below is set, realloc with a 41 size argument of zero (re)allocates a minimum-sized chunk. 42 memalign(size_t alignment, size_t n); 43 Return a pointer to a newly allocated chunk of n bytes, aligned 44 in accord with the alignment argument, which must be a power of 45 two. 46 valloc(size_t n); 47 Equivalent to memalign(pagesize, n), where pagesize is the page 48 size of the system (or as near to this as can be figured out from 49 all the includes/defines below.) 50 pvalloc(size_t n); 51 Equivalent to valloc(minimum-page-that-holds(n)), that is, 52 round up n to nearest pagesize. 53 calloc(size_t unit, size_t quantity); 54 Returns a pointer to quantity * unit bytes, with all locations 55 set to zero. 56 cfree(Void_t* p); 57 Equivalent to free(p). 58 malloc_trim(size_t pad); 59 Release all but pad bytes of freed top-most memory back 60 to the system. Return 1 if successful, else 0. 61 malloc_usable_size(Void_t* p); 62 Report the number usable allocated bytes associated with allocated 63 chunk p. This may or may not report more bytes than were requested, 64 due to alignment and minimum size constraints. 65 malloc_stats(); 66 Prints brief summary statistics. 67 mallinfo() 68 Returns (by copy) a struct containing various summary statistics. 69 mallopt(int parameter_number, int parameter_value) 70 Changes one of the tunable parameters described below. Returns 71 1 if successful in changing the parameter, else 0. 72 73 * Vital statistics: 74 75 Alignment: 8-byte 76 8 byte alignment is currently hardwired into the design. This 77 seems to suffice for all current machines and C compilers. 78 79 Assumed pointer representation: 4 or 8 bytes 80 Code for 8-byte pointers is untested by me but has worked 81 reliably by Wolfram Gloger, who contributed most of the 82 changes supporting this. 83 84 Assumed size_t representation: 4 or 8 bytes 85 Note that size_t is allowed to be 4 bytes even if pointers are 8. 86 87 Minimum overhead per allocated chunk: 4 or 8 bytes 88 Each malloced chunk has a hidden overhead of 4 bytes holding size 89 and status information. 90 91 Minimum allocated size: 4-byte ptrs: 16 bytes (including 4 overhead) 92 8-byte ptrs: 24/32 bytes (including, 4/8 overhead) 93 94 When a chunk is freed, 12 (for 4byte ptrs) or 20 (for 8 byte 95 ptrs but 4 byte size) or 24 (for 8/8) additional bytes are 96 needed; 4 (8) for a trailing size field 97 and 8 (16) bytes for free list pointers. Thus, the minimum 98 allocatable size is 16/24/32 bytes. 99 100 Even a request for zero bytes (i.e., malloc(0)) returns a 101 pointer to something of the minimum allocatable size. 102 103 Maximum allocated size: 4-byte size_t: 2^31 - 8 bytes 104 8-byte size_t: 2^63 - 16 bytes 105 106 It is assumed that (possibly signed) size_t bit values suffice to 107 represent chunk sizes. `Possibly signed' is due to the fact 108 that `size_t' may be defined on a system as either a signed or 109 an unsigned type. To be conservative, values that would appear 110 as negative numbers are avoided. 111 Requests for sizes with a negative sign bit when the request 112 size is treaded as a long will return null. 113 114 Maximum overhead wastage per allocated chunk: normally 15 bytes 115 116 Alignnment demands, plus the minimum allocatable size restriction 117 make the normal worst-case wastage 15 bytes (i.e., up to 15 118 more bytes will be allocated than were requested in malloc), with 119 two exceptions: 120 1. Because requests for zero bytes allocate non-zero space, 121 the worst case wastage for a request of zero bytes is 24 bytes. 122 2. For requests >= mmap_threshold that are serviced via 123 mmap(), the worst case wastage is 8 bytes plus the remainder 124 from a system page (the minimal mmap unit); typically 4096 bytes. 125 126 * Limitations 127 128 Here are some features that are NOT currently supported 129 130 * No user-definable hooks for callbacks and the like. 131 * No automated mechanism for fully checking that all accesses 132 to malloced memory stay within their bounds. 133 * No support for compaction. 134 135 * Synopsis of compile-time options: 136 137 People have reported using previous versions of this malloc on all 138 versions of Unix, sometimes by tweaking some of the defines 139 below. It has been tested most extensively on Solaris and 140 Linux. It is also reported to work on WIN32 platforms. 141 People have also reported adapting this malloc for use in 142 stand-alone embedded systems. 143 144 The implementation is in straight, hand-tuned ANSI C. Among other 145 consequences, it uses a lot of macros. Because of this, to be at 146 all usable, this code should be compiled using an optimizing compiler 147 (for example gcc -O2) that can simplify expressions and control 148 paths. 149 150 __STD_C (default: derived from C compiler defines) 151 Nonzero if using ANSI-standard C compiler, a C++ compiler, or 152 a C compiler sufficiently close to ANSI to get away with it. 153 DEBUG (default: NOT defined) 154 Define to enable debugging. Adds fairly extensive assertion-based 155 checking to help track down memory errors, but noticeably slows down 156 execution. 157 REALLOC_ZERO_BYTES_FREES (default: NOT defined) 158 Define this if you think that realloc(p, 0) should be equivalent 159 to free(p). Otherwise, since malloc returns a unique pointer for 160 malloc(0), so does realloc(p, 0). 161 HAVE_MEMCPY (default: defined) 162 Define if you are not otherwise using ANSI STD C, but still 163 have memcpy and memset in your C library and want to use them. 164 Otherwise, simple internal versions are supplied. 165 USE_MEMCPY (default: 1 if HAVE_MEMCPY is defined, 0 otherwise) 166 Define as 1 if you want the C library versions of memset and 167 memcpy called in realloc and calloc (otherwise macro versions are used). 168 At least on some platforms, the simple macro versions usually 169 outperform libc versions. 170 HAVE_MMAP (default: defined as 1) 171 Define to non-zero to optionally make malloc() use mmap() to 172 allocate very large blocks. 173 HAVE_MREMAP (default: defined as 0 unless Linux libc set) 174 Define to non-zero to optionally make realloc() use mremap() to 175 reallocate very large blocks. 176 malloc_getpagesize (default: derived from system #includes) 177 Either a constant or routine call returning the system page size. 178 HAVE_USR_INCLUDE_MALLOC_H (default: NOT defined) 179 Optionally define if you are on a system with a /usr/include/malloc.h 180 that declares struct mallinfo. It is not at all necessary to 181 define this even if you do, but will ensure consistency. 182 INTERNAL_SIZE_T (default: size_t) 183 Define to a 32-bit type (probably `unsigned int') if you are on a 184 64-bit machine, yet do not want or need to allow malloc requests of 185 greater than 2^31 to be handled. This saves space, especially for 186 very small chunks. 187 INTERNAL_LINUX_C_LIB (default: NOT defined) 188 Defined only when compiled as part of Linux libc. 189 Also note that there is some odd internal name-mangling via defines 190 (for example, internally, `malloc' is named `mALLOc') needed 191 when compiling in this case. These look funny but don't otherwise 192 affect anything. 193 WIN32 (default: undefined) 194 Define this on MS win (95, nt) platforms to compile in sbrk emulation. 195 LACKS_UNISTD_H (default: undefined if not WIN32) 196 Define this if your system does not have a <unistd.h>. 197 LACKS_SYS_PARAM_H (default: undefined if not WIN32) 198 Define this if your system does not have a <sys/param.h>. 199 MORECORE (default: sbrk) 200 The name of the routine to call to obtain more memory from the system. 201 MORECORE_FAILURE (default: -1) 202 The value returned upon failure of MORECORE. 203 MORECORE_CLEARS (default 1) 204 True (1) if the routine mapped to MORECORE zeroes out memory (which 205 holds for sbrk). 206 DEFAULT_TRIM_THRESHOLD 207 DEFAULT_TOP_PAD 208 DEFAULT_MMAP_THRESHOLD 209 DEFAULT_MMAP_MAX 210 Default values of tunable parameters (described in detail below) 211 controlling interaction with host system routines (sbrk, mmap, etc). 212 These values may also be changed dynamically via mallopt(). The 213 preset defaults are those that give best performance for typical 214 programs/systems. 215 USE_DL_PREFIX (default: undefined) 216 Prefix all public routines with the string 'dl'. Useful to 217 quickly avoid procedure declaration conflicts and linker symbol 218 conflicts with existing memory allocation routines. 219 220 221 */ 222 223 224 225 /* Preliminaries */ 226 227 #ifndef __STD_C 228 #ifdef __STDC__ 229 #define __STD_C 1 230 #else 231 #if __cplusplus 232 #define __STD_C 1 233 #else 234 #define __STD_C 0 235 #endif /*__cplusplus*/ 236 #endif /*__STDC__*/ 237 #endif /*__STD_C*/ 238 239 #ifndef Void_t 240 #if (__STD_C || defined(WIN32)) 241 #define Void_t void 242 #else 243 #define Void_t char 244 #endif 245 #endif /*Void_t*/ 246 247 #if __STD_C 248 #include <stddef.h> /* for size_t */ 249 #else 250 #include <sys/types.h> 251 #endif 252 253 #ifdef __cplusplus 254 extern "C" { 255 #endif 256 257 #include <stdio.h> /* needed for malloc_stats */ 258 259 260 /* 261 Compile-time options 262 */ 263 264 265 /* 266 Debugging: 267 268 Because freed chunks may be overwritten with link fields, this 269 malloc will often die when freed memory is overwritten by user 270 programs. This can be very effective (albeit in an annoying way) 271 in helping track down dangling pointers. 272 273 If you compile with -DDEBUG, a number of assertion checks are 274 enabled that will catch more memory errors. You probably won't be 275 able to make much sense of the actual assertion errors, but they 276 should help you locate incorrectly overwritten memory. The 277 checking is fairly extensive, and will slow down execution 278 noticeably. Calling malloc_stats or mallinfo with DEBUG set will 279 attempt to check every non-mmapped allocated and free chunk in the 280 course of computing the summmaries. (By nature, mmapped regions 281 cannot be checked very much automatically.) 282 283 Setting DEBUG may also be helpful if you are trying to modify 284 this code. The assertions in the check routines spell out in more 285 detail the assumptions and invariants underlying the algorithms. 286 287 */ 288 289 /* 290 INTERNAL_SIZE_T is the word-size used for internal bookkeeping 291 of chunk sizes. On a 64-bit machine, you can reduce malloc 292 overhead by defining INTERNAL_SIZE_T to be a 32 bit `unsigned int' 293 at the expense of not being able to handle requests greater than 294 2^31. This limitation is hardly ever a concern; you are encouraged 295 to set this. However, the default version is the same as size_t. 296 */ 297 298 #ifndef INTERNAL_SIZE_T 299 #define INTERNAL_SIZE_T size_t 300 #endif 301 302 /* 303 REALLOC_ZERO_BYTES_FREES should be set if a call to 304 realloc with zero bytes should be the same as a call to free. 305 Some people think it should. Otherwise, since this malloc 306 returns a unique pointer for malloc(0), so does realloc(p, 0). 307 */ 308 309 310 /* #define REALLOC_ZERO_BYTES_FREES */ 311 312 313 /* 314 WIN32 causes an emulation of sbrk to be compiled in 315 mmap-based options are not currently supported in WIN32. 316 */ 317 318 /* #define WIN32 */ 319 #ifdef WIN32 320 #define MORECORE wsbrk 321 #define HAVE_MMAP 0 322 323 #define LACKS_UNISTD_H 324 #define LACKS_SYS_PARAM_H 325 326 /* 327 Include 'windows.h' to get the necessary declarations for the 328 Microsoft Visual C++ data structures and routines used in the 'sbrk' 329 emulation. 330 331 Define WIN32_LEAN_AND_MEAN so that only the essential Microsoft 332 Visual C++ header files are included. 333 */ 334 #define WIN32_LEAN_AND_MEAN 335 #include <windows.h> 336 #endif 337 338 339 /* 340 HAVE_MEMCPY should be defined if you are not otherwise using 341 ANSI STD C, but still have memcpy and memset in your C library 342 and want to use them in calloc and realloc. Otherwise simple 343 macro versions are defined here. 344 345 USE_MEMCPY should be defined as 1 if you actually want to 346 have memset and memcpy called. People report that the macro 347 versions are often enough faster than libc versions on many 348 systems that it is better to use them. 349 350 */ 351 352 #define HAVE_MEMCPY 353 354 #ifndef USE_MEMCPY 355 #ifdef HAVE_MEMCPY 356 #define USE_MEMCPY 1 357 #else 358 #define USE_MEMCPY 0 359 #endif 360 #endif 361 362 #if (__STD_C || defined(HAVE_MEMCPY)) 363 364 #if __STD_C 365 void* memset(void*, int, size_t); 366 void* memcpy(void*, const void*, size_t); 367 #else 368 #ifdef WIN32 369 /* On Win32 platforms, 'memset()' and 'memcpy()' are already declared in */ 370 /* 'windows.h' */ 371 #else 372 Void_t* memset(); 373 Void_t* memcpy(); 374 #endif 375 #endif 376 #endif 377 378 #if USE_MEMCPY 379 380 /* The following macros are only invoked with (2n+1)-multiples of 381 INTERNAL_SIZE_T units, with a positive integer n. This is exploited 382 for fast inline execution when n is small. */ 383 384 #define MALLOC_ZERO(charp, nbytes) \ 385 do { \ 386 INTERNAL_SIZE_T mzsz = (nbytes); \ 387 if(mzsz <= 9*sizeof(mzsz)) { \ 388 INTERNAL_SIZE_T* mz = (INTERNAL_SIZE_T*) (charp); \ 389 if(mzsz >= 5*sizeof(mzsz)) { *mz++ = 0; \ 390 *mz++ = 0; \ 391 if(mzsz >= 7*sizeof(mzsz)) { *mz++ = 0; \ 392 *mz++ = 0; \ 393 if(mzsz >= 9*sizeof(mzsz)) { *mz++ = 0; \ 394 *mz++ = 0; }}} \ 395 *mz++ = 0; \ 396 *mz++ = 0; \ 397 *mz = 0; \ 398 } else memset((charp), 0, mzsz); \ 399 } while(0) 400 401 #define MALLOC_COPY(dest,src,nbytes) \ 402 do { \ 403 INTERNAL_SIZE_T mcsz = (nbytes); \ 404 if(mcsz <= 9*sizeof(mcsz)) { \ 405 INTERNAL_SIZE_T* mcsrc = (INTERNAL_SIZE_T*) (src); \ 406 INTERNAL_SIZE_T* mcdst = (INTERNAL_SIZE_T*) (dest); \ 407 if(mcsz >= 5*sizeof(mcsz)) { *mcdst++ = *mcsrc++; \ 408 *mcdst++ = *mcsrc++; \ 409 if(mcsz >= 7*sizeof(mcsz)) { *mcdst++ = *mcsrc++; \ 410 *mcdst++ = *mcsrc++; \ 411 if(mcsz >= 9*sizeof(mcsz)) { *mcdst++ = *mcsrc++; \ 412 *mcdst++ = *mcsrc++; }}} \ 413 *mcdst++ = *mcsrc++; \ 414 *mcdst++ = *mcsrc++; \ 415 *mcdst = *mcsrc ; \ 416 } else memcpy(dest, src, mcsz); \ 417 } while(0) 418 419 #else /* !USE_MEMCPY */ 420 421 /* Use Duff's device for good zeroing/copying performance. */ 422 423 #define MALLOC_ZERO(charp, nbytes) \ 424 do { \ 425 INTERNAL_SIZE_T* mzp = (INTERNAL_SIZE_T*)(charp); \ 426 long mctmp = (nbytes)/sizeof(INTERNAL_SIZE_T), mcn; \ 427 if (mctmp < 8) mcn = 0; else { mcn = (mctmp-1)/8; mctmp %= 8; } \ 428 switch (mctmp) { \ 429 case 0: for(;;) { *mzp++ = 0; \ 430 case 7: *mzp++ = 0; \ 431 case 6: *mzp++ = 0; \ 432 case 5: *mzp++ = 0; \ 433 case 4: *mzp++ = 0; \ 434 case 3: *mzp++ = 0; \ 435 case 2: *mzp++ = 0; \ 436 case 1: *mzp++ = 0; if(mcn <= 0) break; mcn--; } \ 437 } \ 438 } while(0) 439 440 #define MALLOC_COPY(dest,src,nbytes) \ 441 do { \ 442 INTERNAL_SIZE_T* mcsrc = (INTERNAL_SIZE_T*) src; \ 443 INTERNAL_SIZE_T* mcdst = (INTERNAL_SIZE_T*) dest; \ 444 long mctmp = (nbytes)/sizeof(INTERNAL_SIZE_T), mcn; \ 445 if (mctmp < 8) mcn = 0; else { mcn = (mctmp-1)/8; mctmp %= 8; } \ 446 switch (mctmp) { \ 447 case 0: for(;;) { *mcdst++ = *mcsrc++; \ 448 case 7: *mcdst++ = *mcsrc++; \ 449 case 6: *mcdst++ = *mcsrc++; \ 450 case 5: *mcdst++ = *mcsrc++; \ 451 case 4: *mcdst++ = *mcsrc++; \ 452 case 3: *mcdst++ = *mcsrc++; \ 453 case 2: *mcdst++ = *mcsrc++; \ 454 case 1: *mcdst++ = *mcsrc++; if(mcn <= 0) break; mcn--; } \ 455 } \ 456 } while(0) 457 458 #endif 459 460 461 /* 462 Define HAVE_MMAP to optionally make malloc() use mmap() to 463 allocate very large blocks. These will be returned to the 464 operating system immediately after a free(). 465 */ 466 467 #ifndef HAVE_MMAP 468 #define HAVE_MMAP 1 469 #endif 470 471 /* 472 Define HAVE_MREMAP to make realloc() use mremap() to re-allocate 473 large blocks. This is currently only possible on Linux with 474 kernel versions newer than 1.3.77. 475 */ 476 477 #ifndef HAVE_MREMAP 478 #ifdef INTERNAL_LINUX_C_LIB 479 #define HAVE_MREMAP 1 480 #else 481 #define HAVE_MREMAP 0 482 #endif 483 #endif 484 485 #if HAVE_MMAP 486 487 #include <unistd.h> 488 #include <fcntl.h> 489 #include <sys/mman.h> 490 491 #if !defined(MAP_ANONYMOUS) && defined(MAP_ANON) 492 #define MAP_ANONYMOUS MAP_ANON 493 #endif 494 495 #endif /* HAVE_MMAP */ 496 497 /* 498 Access to system page size. To the extent possible, this malloc 499 manages memory from the system in page-size units. 500 501 The following mechanics for getpagesize were adapted from 502 bsd/gnu getpagesize.h 503 */ 504 505 #ifndef LACKS_UNISTD_H 506 # include <unistd.h> 507 #endif 508 509 #ifndef malloc_getpagesize 510 # ifdef _SC_PAGESIZE /* some SVR4 systems omit an underscore */ 511 # ifndef _SC_PAGE_SIZE 512 # define _SC_PAGE_SIZE _SC_PAGESIZE 513 # endif 514 # endif 515 # ifdef _SC_PAGE_SIZE 516 # define malloc_getpagesize sysconf(_SC_PAGE_SIZE) 517 # else 518 # if defined(BSD) || defined(DGUX) || defined(HAVE_GETPAGESIZE) 519 extern size_t getpagesize(); 520 # define malloc_getpagesize getpagesize() 521 # else 522 # ifdef WIN32 523 # define malloc_getpagesize (4096) /* TBD: Use 'GetSystemInfo' instead */ 524 # else 525 # ifndef LACKS_SYS_PARAM_H 526 # include <sys/param.h> 527 # endif 528 # ifdef EXEC_PAGESIZE 529 # define malloc_getpagesize EXEC_PAGESIZE 530 # else 531 # ifdef NBPG 532 # ifndef CLSIZE 533 # define malloc_getpagesize NBPG 534 # else 535 # define malloc_getpagesize (NBPG * CLSIZE) 536 # endif 537 # else 538 # ifdef NBPC 539 # define malloc_getpagesize NBPC 540 # else 541 # ifdef PAGESIZE 542 # define malloc_getpagesize PAGESIZE 543 # else 544 # define malloc_getpagesize (4096) /* just guess */ 545 # endif 546 # endif 547 # endif 548 # endif 549 # endif 550 # endif 551 # endif 552 #endif 553 554 555 /* 556 557 This version of malloc supports the standard SVID/XPG mallinfo 558 routine that returns a struct containing the same kind of 559 information you can get from malloc_stats. It should work on 560 any SVID/XPG compliant system that has a /usr/include/malloc.h 561 defining struct mallinfo. (If you'd like to install such a thing 562 yourself, cut out the preliminary declarations as described above 563 and below and save them in a malloc.h file. But there's no 564 compelling reason to bother to do this.) 565 566 The main declaration needed is the mallinfo struct that is returned 567 (by-copy) by mallinfo(). The SVID/XPG malloinfo struct contains a 568 bunch of fields, most of which are not even meaningful in this 569 version of malloc. Some of these fields are are instead filled by 570 mallinfo() with other numbers that might possibly be of interest. 571 572 HAVE_USR_INCLUDE_MALLOC_H should be set if you have a 573 /usr/include/malloc.h file that includes a declaration of struct 574 mallinfo. If so, it is included; else an SVID2/XPG2 compliant 575 version is declared below. These must be precisely the same for 576 mallinfo() to work. 577 578 */ 579 580 /* #define HAVE_USR_INCLUDE_MALLOC_H */ 581 582 #if HAVE_USR_INCLUDE_MALLOC_H 583 #include "/usr/include/malloc.h" 584 #else 585 586 /* SVID2/XPG mallinfo structure */ 587 588 struct mallinfo { 589 int arena; /* total space allocated from system */ 590 int ordblks; /* number of non-inuse chunks */ 591 int smblks; /* unused -- always zero */ 592 int hblks; /* number of mmapped regions */ 593 int hblkhd; /* total space in mmapped regions */ 594 int usmblks; /* unused -- always zero */ 595 int fsmblks; /* unused -- always zero */ 596 int uordblks; /* total allocated space */ 597 int fordblks; /* total non-inuse space */ 598 int keepcost; /* top-most, releasable (via malloc_trim) space */ 599 }; 600 601 /* SVID2/XPG mallopt options */ 602 603 #define M_MXFAST 1 /* UNUSED in this malloc */ 604 #define M_NLBLKS 2 /* UNUSED in this malloc */ 605 #define M_GRAIN 3 /* UNUSED in this malloc */ 606 #define M_KEEP 4 /* UNUSED in this malloc */ 607 608 #endif 609 610 /* mallopt options that actually do something */ 611 612 #define M_TRIM_THRESHOLD -1 613 #define M_TOP_PAD -2 614 #define M_MMAP_THRESHOLD -3 615 #define M_MMAP_MAX -4 616 617 618 #ifndef DEFAULT_TRIM_THRESHOLD 619 #define DEFAULT_TRIM_THRESHOLD (128 * 1024) 620 #endif 621 622 /* 623 M_TRIM_THRESHOLD is the maximum amount of unused top-most memory 624 to keep before releasing via malloc_trim in free(). 625 626 Automatic trimming is mainly useful in long-lived programs. 627 Because trimming via sbrk can be slow on some systems, and can 628 sometimes be wasteful (in cases where programs immediately 629 afterward allocate more large chunks) the value should be high 630 enough so that your overall system performance would improve by 631 releasing. 632 633 The trim threshold and the mmap control parameters (see below) 634 can be traded off with one another. Trimming and mmapping are 635 two different ways of releasing unused memory back to the 636 system. Between these two, it is often possible to keep 637 system-level demands of a long-lived program down to a bare 638 minimum. For example, in one test suite of sessions measuring 639 the XF86 X server on Linux, using a trim threshold of 128K and a 640 mmap threshold of 192K led to near-minimal long term resource 641 consumption. 642 643 If you are using this malloc in a long-lived program, it should 644 pay to experiment with these values. As a rough guide, you 645 might set to a value close to the average size of a process 646 (program) running on your system. Releasing this much memory 647 would allow such a process to run in memory. Generally, it's 648 worth it to tune for trimming rather tham memory mapping when a 649 program undergoes phases where several large chunks are 650 allocated and released in ways that can reuse each other's 651 storage, perhaps mixed with phases where there are no such 652 chunks at all. And in well-behaved long-lived programs, 653 controlling release of large blocks via trimming versus mapping 654 is usually faster. 655 656 However, in most programs, these parameters serve mainly as 657 protection against the system-level effects of carrying around 658 massive amounts of unneeded memory. Since frequent calls to 659 sbrk, mmap, and munmap otherwise degrade performance, the default 660 parameters are set to relatively high values that serve only as 661 safeguards. 662 663 The default trim value is high enough to cause trimming only in 664 fairly extreme (by current memory consumption standards) cases. 665 It must be greater than page size to have any useful effect. To 666 disable trimming completely, you can set to (unsigned long)(-1); 667 668 669 */ 670 671 672 #ifndef DEFAULT_TOP_PAD 673 #define DEFAULT_TOP_PAD (0) 674 #endif 675 676 /* 677 M_TOP_PAD is the amount of extra `padding' space to allocate or 678 retain whenever sbrk is called. It is used in two ways internally: 679 680 * When sbrk is called to extend the top of the arena to satisfy 681 a new malloc request, this much padding is added to the sbrk 682 request. 683 684 * When malloc_trim is called automatically from free(), 685 it is used as the `pad' argument. 686 687 In both cases, the actual amount of padding is rounded 688 so that the end of the arena is always a system page boundary. 689 690 The main reason for using padding is to avoid calling sbrk so 691 often. Having even a small pad greatly reduces the likelihood 692 that nearly every malloc request during program start-up (or 693 after trimming) will invoke sbrk, which needlessly wastes 694 time. 695 696 Automatic rounding-up to page-size units is normally sufficient 697 to avoid measurable overhead, so the default is 0. However, in 698 systems where sbrk is relatively slow, it can pay to increase 699 this value, at the expense of carrying around more memory than 700 the program needs. 701 702 */ 703 704 705 #ifndef DEFAULT_MMAP_THRESHOLD 706 #define DEFAULT_MMAP_THRESHOLD (128 * 1024) 707 #endif 708 709 /* 710 711 M_MMAP_THRESHOLD is the request size threshold for using mmap() 712 to service a request. Requests of at least this size that cannot 713 be allocated using already-existing space will be serviced via mmap. 714 (If enough normal freed space already exists it is used instead.) 715 716 Using mmap segregates relatively large chunks of memory so that 717 they can be individually obtained and released from the host 718 system. A request serviced through mmap is never reused by any 719 other request (at least not directly; the system may just so 720 happen to remap successive requests to the same locations). 721 722 Segregating space in this way has the benefit that mmapped space 723 can ALWAYS be individually released back to the system, which 724 helps keep the system level memory demands of a long-lived 725 program low. Mapped memory can never become `locked' between 726 other chunks, as can happen with normally allocated chunks, which 727 menas that even trimming via malloc_trim would not release them. 728 729 However, it has the disadvantages that: 730 731 1. The space cannot be reclaimed, consolidated, and then 732 used to service later requests, as happens with normal chunks. 733 2. It can lead to more wastage because of mmap page alignment 734 requirements 735 3. It causes malloc performance to be more dependent on host 736 system memory management support routines which may vary in 737 implementation quality and may impose arbitrary 738 limitations. Generally, servicing a request via normal 739 malloc steps is faster than going through a system's mmap. 740 741 All together, these considerations should lead you to use mmap 742 only for relatively large requests. 743 744 745 */ 746 747 748 #ifndef DEFAULT_MMAP_MAX 749 #if HAVE_MMAP 750 #define DEFAULT_MMAP_MAX (64) 751 #else 752 #define DEFAULT_MMAP_MAX (0) 753 #endif 754 #endif 755 756 /* 757 M_MMAP_MAX is the maximum number of requests to simultaneously 758 service using mmap. This parameter exists because: 759 760 1. Some systems have a limited number of internal tables for 761 use by mmap. 762 2. In most systems, overreliance on mmap can degrade overall 763 performance. 764 3. If a program allocates many large regions, it is probably 765 better off using normal sbrk-based allocation routines that 766 can reclaim and reallocate normal heap memory. Using a 767 small value allows transition into this mode after the 768 first few allocations. 769 770 Setting to 0 disables all use of mmap. If HAVE_MMAP is not set, 771 the default value is 0, and attempts to set it to non-zero values 772 in mallopt will fail. 773 */ 774 775 776 /* 777 USE_DL_PREFIX will prefix all public routines with the string 'dl'. 778 Useful to quickly avoid procedure declaration conflicts and linker 779 symbol conflicts with existing memory allocation routines. 780 781 */ 782 783 /* #define USE_DL_PREFIX */ 784 785 786 /* 787 788 Special defines for linux libc 789 790 Except when compiled using these special defines for Linux libc 791 using weak aliases, this malloc is NOT designed to work in 792 multithreaded applications. No semaphores or other concurrency 793 control are provided to ensure that multiple malloc or free calls 794 don't run at the same time, which could be disasterous. A single 795 semaphore could be used across malloc, realloc, and free (which is 796 essentially the effect of the linux weak alias approach). It would 797 be hard to obtain finer granularity. 798 799 */ 800 801 802 #ifdef INTERNAL_LINUX_C_LIB 803 804 #if __STD_C 805 806 Void_t * __default_morecore_init (ptrdiff_t); 807 Void_t *(*__morecore)(ptrdiff_t) = __default_morecore_init; 808 809 #else 810 811 Void_t * __default_morecore_init (); 812 Void_t *(*__morecore)() = __default_morecore_init; 813 814 #endif 815 816 #define MORECORE (*__morecore) 817 #define MORECORE_FAILURE 0 818 #define MORECORE_CLEARS 1 819 820 #else /* INTERNAL_LINUX_C_LIB */ 821 822 #if __STD_C 823 extern Void_t* sbrk(ptrdiff_t); 824 #else 825 extern Void_t* sbrk(); 826 #endif 827 828 #ifndef MORECORE 829 #define MORECORE sbrk 830 #endif 831 832 #ifndef MORECORE_FAILURE 833 #define MORECORE_FAILURE -1 834 #endif 835 836 #ifndef MORECORE_CLEARS 837 #define MORECORE_CLEARS 1 838 #endif 839 840 #endif /* INTERNAL_LINUX_C_LIB */ 841 842 #if defined(INTERNAL_LINUX_C_LIB) && defined(__ELF__) 843 844 #define cALLOc __libc_calloc 845 #define fREe __libc_free 846 #define mALLOc __libc_malloc 847 #define mEMALIGn __libc_memalign 848 #define rEALLOc __libc_realloc 849 #define vALLOc __libc_valloc 850 #define pvALLOc __libc_pvalloc 851 #define mALLINFo __libc_mallinfo 852 #define mALLOPt __libc_mallopt 853 854 #pragma weak calloc = __libc_calloc 855 #pragma weak free = __libc_free 856 #pragma weak cfree = __libc_free 857 #pragma weak malloc = __libc_malloc 858 #pragma weak memalign = __libc_memalign 859 #pragma weak realloc = __libc_realloc 860 #pragma weak valloc = __libc_valloc 861 #pragma weak pvalloc = __libc_pvalloc 862 #pragma weak mallinfo = __libc_mallinfo 863 #pragma weak mallopt = __libc_mallopt 864 865 #else 866 867 #ifdef USE_DL_PREFIX 868 #define cALLOc dlcalloc 869 #define fREe dlfree 870 #define mALLOc dlmalloc 871 #define mEMALIGn dlmemalign 872 #define rEALLOc dlrealloc 873 #define vALLOc dlvalloc 874 #define pvALLOc dlpvalloc 875 #define mALLINFo dlmallinfo 876 #define mALLOPt dlmallopt 877 #else /* USE_DL_PREFIX */ 878 #define cALLOc calloc 879 #define fREe free 880 #define mALLOc malloc 881 #define mEMALIGn memalign 882 #define rEALLOc realloc 883 #define vALLOc valloc 884 #define pvALLOc pvalloc 885 #define mALLINFo mallinfo 886 #define mALLOPt mallopt 887 #endif /* USE_DL_PREFIX */ 888 889 #endif 890 891 /* Public routines */ 892 893 #if __STD_C 894 895 Void_t* mALLOc(size_t); 896 void fREe(Void_t*); 897 Void_t* rEALLOc(Void_t*, size_t); 898 Void_t* mEMALIGn(size_t, size_t); 899 Void_t* vALLOc(size_t); 900 Void_t* pvALLOc(size_t); 901 Void_t* cALLOc(size_t, size_t); 902 void cfree(Void_t*); 903 int malloc_trim(size_t); 904 size_t malloc_usable_size(Void_t*); 905 void malloc_stats(); 906 int mALLOPt(int, int); 907 struct mallinfo mALLINFo(void); 908 #else 909 Void_t* mALLOc(); 910 void fREe(); 911 Void_t* rEALLOc(); 912 Void_t* mEMALIGn(); 913 Void_t* vALLOc(); 914 Void_t* pvALLOc(); 915 Void_t* cALLOc(); 916 void cfree(); 917 int malloc_trim(); 918 size_t malloc_usable_size(); 919 void malloc_stats(); 920 int mALLOPt(); 921 struct mallinfo mALLINFo(); 922 #endif 923 924 925 #ifdef __cplusplus 926 }; /* end of extern "C" */ 927 #endif 928 929 /* ---------- To make a malloc.h, end cutting here ------------ */ 930 #endif /* 0 */ /* Moved to malloc.h */ 931 932 #include <malloc.h> 933 #ifdef DEBUG 934 #if __STD_C 935 static void malloc_update_mallinfo (void); 936 void malloc_stats (void); 937 #else 938 static void malloc_update_mallinfo (); 939 void malloc_stats(); 940 #endif 941 #endif /* DEBUG */ 942 943 DECLARE_GLOBAL_DATA_PTR; 944 945 /* 946 Emulation of sbrk for WIN32 947 All code within the ifdef WIN32 is untested by me. 948 949 Thanks to Martin Fong and others for supplying this. 950 */ 951 952 953 #ifdef WIN32 954 955 #define AlignPage(add) (((add) + (malloc_getpagesize-1)) & \ 956 ~(malloc_getpagesize-1)) 957 #define AlignPage64K(add) (((add) + (0x10000 - 1)) & ~(0x10000 - 1)) 958 959 /* resrve 64MB to insure large contiguous space */ 960 #define RESERVED_SIZE (1024*1024*64) 961 #define NEXT_SIZE (2048*1024) 962 #define TOP_MEMORY ((unsigned long)2*1024*1024*1024) 963 964 struct GmListElement; 965 typedef struct GmListElement GmListElement; 966 967 struct GmListElement 968 { 969 GmListElement* next; 970 void* base; 971 }; 972 973 static GmListElement* head = 0; 974 static unsigned int gNextAddress = 0; 975 static unsigned int gAddressBase = 0; 976 static unsigned int gAllocatedSize = 0; 977 978 static 979 GmListElement* makeGmListElement (void* bas) 980 { 981 GmListElement* this; 982 this = (GmListElement*)(void*)LocalAlloc (0, sizeof (GmListElement)); 983 assert (this); 984 if (this) 985 { 986 this->base = bas; 987 this->next = head; 988 head = this; 989 } 990 return this; 991 } 992 993 void gcleanup () 994 { 995 BOOL rval; 996 assert ( (head == NULL) || (head->base == (void*)gAddressBase)); 997 if (gAddressBase && (gNextAddress - gAddressBase)) 998 { 999 rval = VirtualFree ((void*)gAddressBase, 1000 gNextAddress - gAddressBase, 1001 MEM_DECOMMIT); 1002 assert (rval); 1003 } 1004 while (head) 1005 { 1006 GmListElement* next = head->next; 1007 rval = VirtualFree (head->base, 0, MEM_RELEASE); 1008 assert (rval); 1009 LocalFree (head); 1010 head = next; 1011 } 1012 } 1013 1014 static 1015 void* findRegion (void* start_address, unsigned long size) 1016 { 1017 MEMORY_BASIC_INFORMATION info; 1018 if (size >= TOP_MEMORY) return NULL; 1019 1020 while ((unsigned long)start_address + size < TOP_MEMORY) 1021 { 1022 VirtualQuery (start_address, &info, sizeof (info)); 1023 if ((info.State == MEM_FREE) && (info.RegionSize >= size)) 1024 return start_address; 1025 else 1026 { 1027 /* Requested region is not available so see if the */ 1028 /* next region is available. Set 'start_address' */ 1029 /* to the next region and call 'VirtualQuery()' */ 1030 /* again. */ 1031 1032 start_address = (char*)info.BaseAddress + info.RegionSize; 1033 1034 /* Make sure we start looking for the next region */ 1035 /* on the *next* 64K boundary. Otherwise, even if */ 1036 /* the new region is free according to */ 1037 /* 'VirtualQuery()', the subsequent call to */ 1038 /* 'VirtualAlloc()' (which follows the call to */ 1039 /* this routine in 'wsbrk()') will round *down* */ 1040 /* the requested address to a 64K boundary which */ 1041 /* we already know is an address in the */ 1042 /* unavailable region. Thus, the subsequent call */ 1043 /* to 'VirtualAlloc()' will fail and bring us back */ 1044 /* here, causing us to go into an infinite loop. */ 1045 1046 start_address = 1047 (void *) AlignPage64K((unsigned long) start_address); 1048 } 1049 } 1050 return NULL; 1051 1052 } 1053 1054 1055 void* wsbrk (long size) 1056 { 1057 void* tmp; 1058 if (size > 0) 1059 { 1060 if (gAddressBase == 0) 1061 { 1062 gAllocatedSize = max (RESERVED_SIZE, AlignPage (size)); 1063 gNextAddress = gAddressBase = 1064 (unsigned int)VirtualAlloc (NULL, gAllocatedSize, 1065 MEM_RESERVE, PAGE_NOACCESS); 1066 } else if (AlignPage (gNextAddress + size) > (gAddressBase + 1067 gAllocatedSize)) 1068 { 1069 long new_size = max (NEXT_SIZE, AlignPage (size)); 1070 void* new_address = (void*)(gAddressBase+gAllocatedSize); 1071 do 1072 { 1073 new_address = findRegion (new_address, new_size); 1074 1075 if (new_address == 0) 1076 return (void*)-1; 1077 1078 gAddressBase = gNextAddress = 1079 (unsigned int)VirtualAlloc (new_address, new_size, 1080 MEM_RESERVE, PAGE_NOACCESS); 1081 /* repeat in case of race condition */ 1082 /* The region that we found has been snagged */ 1083 /* by another thread */ 1084 } 1085 while (gAddressBase == 0); 1086 1087 assert (new_address == (void*)gAddressBase); 1088 1089 gAllocatedSize = new_size; 1090 1091 if (!makeGmListElement ((void*)gAddressBase)) 1092 return (void*)-1; 1093 } 1094 if ((size + gNextAddress) > AlignPage (gNextAddress)) 1095 { 1096 void* res; 1097 res = VirtualAlloc ((void*)AlignPage (gNextAddress), 1098 (size + gNextAddress - 1099 AlignPage (gNextAddress)), 1100 MEM_COMMIT, PAGE_READWRITE); 1101 if (res == 0) 1102 return (void*)-1; 1103 } 1104 tmp = (void*)gNextAddress; 1105 gNextAddress = (unsigned int)tmp + size; 1106 return tmp; 1107 } 1108 else if (size < 0) 1109 { 1110 unsigned int alignedGoal = AlignPage (gNextAddress + size); 1111 /* Trim by releasing the virtual memory */ 1112 if (alignedGoal >= gAddressBase) 1113 { 1114 VirtualFree ((void*)alignedGoal, gNextAddress - alignedGoal, 1115 MEM_DECOMMIT); 1116 gNextAddress = gNextAddress + size; 1117 return (void*)gNextAddress; 1118 } 1119 else 1120 { 1121 VirtualFree ((void*)gAddressBase, gNextAddress - gAddressBase, 1122 MEM_DECOMMIT); 1123 gNextAddress = gAddressBase; 1124 return (void*)-1; 1125 } 1126 } 1127 else 1128 { 1129 return (void*)gNextAddress; 1130 } 1131 } 1132 1133 #endif 1134 1135 1136 1137 /* 1138 Type declarations 1139 */ 1140 1141 1142 struct malloc_chunk 1143 { 1144 INTERNAL_SIZE_T prev_size; /* Size of previous chunk (if free). */ 1145 INTERNAL_SIZE_T size; /* Size in bytes, including overhead. */ 1146 struct malloc_chunk* fd; /* double links -- used only if free. */ 1147 struct malloc_chunk* bk; 1148 } __attribute__((__may_alias__)) ; 1149 1150 typedef struct malloc_chunk* mchunkptr; 1151 1152 /* 1153 1154 malloc_chunk details: 1155 1156 (The following includes lightly edited explanations by Colin Plumb.) 1157 1158 Chunks of memory are maintained using a `boundary tag' method as 1159 described in e.g., Knuth or Standish. (See the paper by Paul 1160 Wilson ftp://ftp.cs.utexas.edu/pub/garbage/allocsrv.ps for a 1161 survey of such techniques.) Sizes of free chunks are stored both 1162 in the front of each chunk and at the end. This makes 1163 consolidating fragmented chunks into bigger chunks very fast. The 1164 size fields also hold bits representing whether chunks are free or 1165 in use. 1166 1167 An allocated chunk looks like this: 1168 1169 1170 chunk-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1171 | Size of previous chunk, if allocated | | 1172 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1173 | Size of chunk, in bytes |P| 1174 mem-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1175 | User data starts here... . 1176 . . 1177 . (malloc_usable_space() bytes) . 1178 . | 1179 nextchunk-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1180 | Size of chunk | 1181 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1182 1183 1184 Where "chunk" is the front of the chunk for the purpose of most of 1185 the malloc code, but "mem" is the pointer that is returned to the 1186 user. "Nextchunk" is the beginning of the next contiguous chunk. 1187 1188 Chunks always begin on even word boundries, so the mem portion 1189 (which is returned to the user) is also on an even word boundary, and 1190 thus double-word aligned. 1191 1192 Free chunks are stored in circular doubly-linked lists, and look like this: 1193 1194 chunk-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1195 | Size of previous chunk | 1196 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1197 `head:' | Size of chunk, in bytes |P| 1198 mem-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1199 | Forward pointer to next chunk in list | 1200 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1201 | Back pointer to previous chunk in list | 1202 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1203 | Unused space (may be 0 bytes long) . 1204 . . 1205 . | 1206 nextchunk-> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1207 `foot:' | Size of chunk, in bytes | 1208 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1209 1210 The P (PREV_INUSE) bit, stored in the unused low-order bit of the 1211 chunk size (which is always a multiple of two words), is an in-use 1212 bit for the *previous* chunk. If that bit is *clear*, then the 1213 word before the current chunk size contains the previous chunk 1214 size, and can be used to find the front of the previous chunk. 1215 (The very first chunk allocated always has this bit set, 1216 preventing access to non-existent (or non-owned) memory.) 1217 1218 Note that the `foot' of the current chunk is actually represented 1219 as the prev_size of the NEXT chunk. (This makes it easier to 1220 deal with alignments etc). 1221 1222 The two exceptions to all this are 1223 1224 1. The special chunk `top', which doesn't bother using the 1225 trailing size field since there is no 1226 next contiguous chunk that would have to index off it. (After 1227 initialization, `top' is forced to always exist. If it would 1228 become less than MINSIZE bytes long, it is replenished via 1229 malloc_extend_top.) 1230 1231 2. Chunks allocated via mmap, which have the second-lowest-order 1232 bit (IS_MMAPPED) set in their size fields. Because they are 1233 never merged or traversed from any other chunk, they have no 1234 foot size or inuse information. 1235 1236 Available chunks are kept in any of several places (all declared below): 1237 1238 * `av': An array of chunks serving as bin headers for consolidated 1239 chunks. Each bin is doubly linked. The bins are approximately 1240 proportionally (log) spaced. There are a lot of these bins 1241 (128). This may look excessive, but works very well in 1242 practice. All procedures maintain the invariant that no 1243 consolidated chunk physically borders another one. Chunks in 1244 bins are kept in size order, with ties going to the 1245 approximately least recently used chunk. 1246 1247 The chunks in each bin are maintained in decreasing sorted order by 1248 size. This is irrelevant for the small bins, which all contain 1249 the same-sized chunks, but facilitates best-fit allocation for 1250 larger chunks. (These lists are just sequential. Keeping them in 1251 order almost never requires enough traversal to warrant using 1252 fancier ordered data structures.) Chunks of the same size are 1253 linked with the most recently freed at the front, and allocations 1254 are taken from the back. This results in LRU or FIFO allocation 1255 order, which tends to give each chunk an equal opportunity to be 1256 consolidated with adjacent freed chunks, resulting in larger free 1257 chunks and less fragmentation. 1258 1259 * `top': The top-most available chunk (i.e., the one bordering the 1260 end of available memory) is treated specially. It is never 1261 included in any bin, is used only if no other chunk is 1262 available, and is released back to the system if it is very 1263 large (see M_TRIM_THRESHOLD). 1264 1265 * `last_remainder': A bin holding only the remainder of the 1266 most recently split (non-top) chunk. This bin is checked 1267 before other non-fitting chunks, so as to provide better 1268 locality for runs of sequentially allocated chunks. 1269 1270 * Implicitly, through the host system's memory mapping tables. 1271 If supported, requests greater than a threshold are usually 1272 serviced via calls to mmap, and then later released via munmap. 1273 1274 */ 1275 1276 /* sizes, alignments */ 1277 1278 #define SIZE_SZ (sizeof(INTERNAL_SIZE_T)) 1279 #define MALLOC_ALIGNMENT (SIZE_SZ + SIZE_SZ) 1280 #define MALLOC_ALIGN_MASK (MALLOC_ALIGNMENT - 1) 1281 #define MINSIZE (sizeof(struct malloc_chunk)) 1282 1283 /* conversion from malloc headers to user pointers, and back */ 1284 1285 #define chunk2mem(p) ((Void_t*)((char*)(p) + 2*SIZE_SZ)) 1286 #define mem2chunk(mem) ((mchunkptr)((char*)(mem) - 2*SIZE_SZ)) 1287 1288 /* pad request bytes into a usable size */ 1289 1290 #define request2size(req) \ 1291 (((long)((req) + (SIZE_SZ + MALLOC_ALIGN_MASK)) < \ 1292 (long)(MINSIZE + MALLOC_ALIGN_MASK)) ? MINSIZE : \ 1293 (((req) + (SIZE_SZ + MALLOC_ALIGN_MASK)) & ~(MALLOC_ALIGN_MASK))) 1294 1295 /* Check if m has acceptable alignment */ 1296 1297 #define aligned_OK(m) (((unsigned long)((m)) & (MALLOC_ALIGN_MASK)) == 0) 1298 1299 1300 1301 1302 /* 1303 Physical chunk operations 1304 */ 1305 1306 1307 /* size field is or'ed with PREV_INUSE when previous adjacent chunk in use */ 1308 1309 #define PREV_INUSE 0x1 1310 1311 /* size field is or'ed with IS_MMAPPED if the chunk was obtained with mmap() */ 1312 1313 #define IS_MMAPPED 0x2 1314 1315 /* Bits to mask off when extracting size */ 1316 1317 #define SIZE_BITS (PREV_INUSE|IS_MMAPPED) 1318 1319 1320 /* Ptr to next physical malloc_chunk. */ 1321 1322 #define next_chunk(p) ((mchunkptr)( ((char*)(p)) + ((p)->size & ~PREV_INUSE) )) 1323 1324 /* Ptr to previous physical malloc_chunk */ 1325 1326 #define prev_chunk(p)\ 1327 ((mchunkptr)( ((char*)(p)) - ((p)->prev_size) )) 1328 1329 1330 /* Treat space at ptr + offset as a chunk */ 1331 1332 #define chunk_at_offset(p, s) ((mchunkptr)(((char*)(p)) + (s))) 1333 1334 1335 1336 1337 /* 1338 Dealing with use bits 1339 */ 1340 1341 /* extract p's inuse bit */ 1342 1343 #define inuse(p)\ 1344 ((((mchunkptr)(((char*)(p))+((p)->size & ~PREV_INUSE)))->size) & PREV_INUSE) 1345 1346 /* extract inuse bit of previous chunk */ 1347 1348 #define prev_inuse(p) ((p)->size & PREV_INUSE) 1349 1350 /* check for mmap()'ed chunk */ 1351 1352 #define chunk_is_mmapped(p) ((p)->size & IS_MMAPPED) 1353 1354 /* set/clear chunk as in use without otherwise disturbing */ 1355 1356 #define set_inuse(p)\ 1357 ((mchunkptr)(((char*)(p)) + ((p)->size & ~PREV_INUSE)))->size |= PREV_INUSE 1358 1359 #define clear_inuse(p)\ 1360 ((mchunkptr)(((char*)(p)) + ((p)->size & ~PREV_INUSE)))->size &= ~(PREV_INUSE) 1361 1362 /* check/set/clear inuse bits in known places */ 1363 1364 #define inuse_bit_at_offset(p, s)\ 1365 (((mchunkptr)(((char*)(p)) + (s)))->size & PREV_INUSE) 1366 1367 #define set_inuse_bit_at_offset(p, s)\ 1368 (((mchunkptr)(((char*)(p)) + (s)))->size |= PREV_INUSE) 1369 1370 #define clear_inuse_bit_at_offset(p, s)\ 1371 (((mchunkptr)(((char*)(p)) + (s)))->size &= ~(PREV_INUSE)) 1372 1373 1374 1375 1376 /* 1377 Dealing with size fields 1378 */ 1379 1380 /* Get size, ignoring use bits */ 1381 1382 #define chunksize(p) ((p)->size & ~(SIZE_BITS)) 1383 1384 /* Set size at head, without disturbing its use bit */ 1385 1386 #define set_head_size(p, s) ((p)->size = (((p)->size & PREV_INUSE) | (s))) 1387 1388 /* Set size/use ignoring previous bits in header */ 1389 1390 #define set_head(p, s) ((p)->size = (s)) 1391 1392 /* Set size at footer (only when chunk is not in use) */ 1393 1394 #define set_foot(p, s) (((mchunkptr)((char*)(p) + (s)))->prev_size = (s)) 1395 1396 1397 1398 1399 1400 /* 1401 Bins 1402 1403 The bins, `av_' are an array of pairs of pointers serving as the 1404 heads of (initially empty) doubly-linked lists of chunks, laid out 1405 in a way so that each pair can be treated as if it were in a 1406 malloc_chunk. (This way, the fd/bk offsets for linking bin heads 1407 and chunks are the same). 1408 1409 Bins for sizes < 512 bytes contain chunks of all the same size, spaced 1410 8 bytes apart. Larger bins are approximately logarithmically 1411 spaced. (See the table below.) The `av_' array is never mentioned 1412 directly in the code, but instead via bin access macros. 1413 1414 Bin layout: 1415 1416 64 bins of size 8 1417 32 bins of size 64 1418 16 bins of size 512 1419 8 bins of size 4096 1420 4 bins of size 32768 1421 2 bins of size 262144 1422 1 bin of size what's left 1423 1424 There is actually a little bit of slop in the numbers in bin_index 1425 for the sake of speed. This makes no difference elsewhere. 1426 1427 The special chunks `top' and `last_remainder' get their own bins, 1428 (this is implemented via yet more trickery with the av_ array), 1429 although `top' is never properly linked to its bin since it is 1430 always handled specially. 1431 1432 */ 1433 1434 #define NAV 128 /* number of bins */ 1435 1436 typedef struct malloc_chunk* mbinptr; 1437 1438 /* access macros */ 1439 1440 #define bin_at(i) ((mbinptr)((char*)&(av_[2*(i) + 2]) - 2*SIZE_SZ)) 1441 #define next_bin(b) ((mbinptr)((char*)(b) + 2 * sizeof(mbinptr))) 1442 #define prev_bin(b) ((mbinptr)((char*)(b) - 2 * sizeof(mbinptr))) 1443 1444 /* 1445 The first 2 bins are never indexed. The corresponding av_ cells are instead 1446 used for bookkeeping. This is not to save space, but to simplify 1447 indexing, maintain locality, and avoid some initialization tests. 1448 */ 1449 1450 #define top (av_[2]) /* The topmost chunk */ 1451 #define last_remainder (bin_at(1)) /* remainder from last split */ 1452 1453 1454 /* 1455 Because top initially points to its own bin with initial 1456 zero size, thus forcing extension on the first malloc request, 1457 we avoid having any special code in malloc to check whether 1458 it even exists yet. But we still need to in malloc_extend_top. 1459 */ 1460 1461 #define initial_top ((mchunkptr)(bin_at(0))) 1462 1463 /* Helper macro to initialize bins */ 1464 1465 #define IAV(i) bin_at(i), bin_at(i) 1466 1467 static mbinptr av_[NAV * 2 + 2] = { 1468 0, 0, 1469 IAV(0), IAV(1), IAV(2), IAV(3), IAV(4), IAV(5), IAV(6), IAV(7), 1470 IAV(8), IAV(9), IAV(10), IAV(11), IAV(12), IAV(13), IAV(14), IAV(15), 1471 IAV(16), IAV(17), IAV(18), IAV(19), IAV(20), IAV(21), IAV(22), IAV(23), 1472 IAV(24), IAV(25), IAV(26), IAV(27), IAV(28), IAV(29), IAV(30), IAV(31), 1473 IAV(32), IAV(33), IAV(34), IAV(35), IAV(36), IAV(37), IAV(38), IAV(39), 1474 IAV(40), IAV(41), IAV(42), IAV(43), IAV(44), IAV(45), IAV(46), IAV(47), 1475 IAV(48), IAV(49), IAV(50), IAV(51), IAV(52), IAV(53), IAV(54), IAV(55), 1476 IAV(56), IAV(57), IAV(58), IAV(59), IAV(60), IAV(61), IAV(62), IAV(63), 1477 IAV(64), IAV(65), IAV(66), IAV(67), IAV(68), IAV(69), IAV(70), IAV(71), 1478 IAV(72), IAV(73), IAV(74), IAV(75), IAV(76), IAV(77), IAV(78), IAV(79), 1479 IAV(80), IAV(81), IAV(82), IAV(83), IAV(84), IAV(85), IAV(86), IAV(87), 1480 IAV(88), IAV(89), IAV(90), IAV(91), IAV(92), IAV(93), IAV(94), IAV(95), 1481 IAV(96), IAV(97), IAV(98), IAV(99), IAV(100), IAV(101), IAV(102), IAV(103), 1482 IAV(104), IAV(105), IAV(106), IAV(107), IAV(108), IAV(109), IAV(110), IAV(111), 1483 IAV(112), IAV(113), IAV(114), IAV(115), IAV(116), IAV(117), IAV(118), IAV(119), 1484 IAV(120), IAV(121), IAV(122), IAV(123), IAV(124), IAV(125), IAV(126), IAV(127) 1485 }; 1486 1487 #ifdef CONFIG_NEEDS_MANUAL_RELOC 1488 void malloc_bin_reloc (void) 1489 { 1490 unsigned long *p = (unsigned long *)(&av_[2]); 1491 int i; 1492 for (i=2; i<(sizeof(av_)/sizeof(mbinptr)); ++i) { 1493 *p++ += gd->reloc_off; 1494 } 1495 } 1496 #endif 1497 1498 ulong mem_malloc_start = 0; 1499 ulong mem_malloc_end = 0; 1500 ulong mem_malloc_brk = 0; 1501 1502 void *sbrk(ptrdiff_t increment) 1503 { 1504 ulong old = mem_malloc_brk; 1505 ulong new = old + increment; 1506 1507 /* 1508 * if we are giving memory back make sure we clear it out since 1509 * we set MORECORE_CLEARS to 1 1510 */ 1511 if (increment < 0) 1512 memset((void *)new, 0, -increment); 1513 1514 if ((new < mem_malloc_start) || (new > mem_malloc_end)) 1515 return (void *)MORECORE_FAILURE; 1516 1517 mem_malloc_brk = new; 1518 1519 return (void *)old; 1520 } 1521 1522 void mem_malloc_init(ulong start, ulong size) 1523 { 1524 mem_malloc_start = start; 1525 mem_malloc_end = start + size; 1526 mem_malloc_brk = start; 1527 1528 memset((void *)mem_malloc_start, 0, size); 1529 } 1530 1531 /* field-extraction macros */ 1532 1533 #define first(b) ((b)->fd) 1534 #define last(b) ((b)->bk) 1535 1536 /* 1537 Indexing into bins 1538 */ 1539 1540 #define bin_index(sz) \ 1541 (((((unsigned long)(sz)) >> 9) == 0) ? (((unsigned long)(sz)) >> 3): \ 1542 ((((unsigned long)(sz)) >> 9) <= 4) ? 56 + (((unsigned long)(sz)) >> 6): \ 1543 ((((unsigned long)(sz)) >> 9) <= 20) ? 91 + (((unsigned long)(sz)) >> 9): \ 1544 ((((unsigned long)(sz)) >> 9) <= 84) ? 110 + (((unsigned long)(sz)) >> 12): \ 1545 ((((unsigned long)(sz)) >> 9) <= 340) ? 119 + (((unsigned long)(sz)) >> 15): \ 1546 ((((unsigned long)(sz)) >> 9) <= 1364) ? 124 + (((unsigned long)(sz)) >> 18): \ 1547 126) 1548 /* 1549 bins for chunks < 512 are all spaced 8 bytes apart, and hold 1550 identically sized chunks. This is exploited in malloc. 1551 */ 1552 1553 #define MAX_SMALLBIN 63 1554 #define MAX_SMALLBIN_SIZE 512 1555 #define SMALLBIN_WIDTH 8 1556 1557 #define smallbin_index(sz) (((unsigned long)(sz)) >> 3) 1558 1559 /* 1560 Requests are `small' if both the corresponding and the next bin are small 1561 */ 1562 1563 #define is_small_request(nb) (nb < MAX_SMALLBIN_SIZE - SMALLBIN_WIDTH) 1564 1565 1566 1567 /* 1568 To help compensate for the large number of bins, a one-level index 1569 structure is used for bin-by-bin searching. `binblocks' is a 1570 one-word bitvector recording whether groups of BINBLOCKWIDTH bins 1571 have any (possibly) non-empty bins, so they can be skipped over 1572 all at once during during traversals. The bits are NOT always 1573 cleared as soon as all bins in a block are empty, but instead only 1574 when all are noticed to be empty during traversal in malloc. 1575 */ 1576 1577 #define BINBLOCKWIDTH 4 /* bins per block */ 1578 1579 #define binblocks_r ((INTERNAL_SIZE_T)av_[1]) /* bitvector of nonempty blocks */ 1580 #define binblocks_w (av_[1]) 1581 1582 /* bin<->block macros */ 1583 1584 #define idx2binblock(ix) ((unsigned)1 << (ix / BINBLOCKWIDTH)) 1585 #define mark_binblock(ii) (binblocks_w = (mbinptr)(binblocks_r | idx2binblock(ii))) 1586 #define clear_binblock(ii) (binblocks_w = (mbinptr)(binblocks_r & ~(idx2binblock(ii)))) 1587 1588 1589 1590 1591 1592 /* Other static bookkeeping data */ 1593 1594 /* variables holding tunable values */ 1595 1596 static unsigned long trim_threshold = DEFAULT_TRIM_THRESHOLD; 1597 static unsigned long top_pad = DEFAULT_TOP_PAD; 1598 static unsigned int n_mmaps_max = DEFAULT_MMAP_MAX; 1599 static unsigned long mmap_threshold = DEFAULT_MMAP_THRESHOLD; 1600 1601 /* The first value returned from sbrk */ 1602 static char* sbrk_base = (char*)(-1); 1603 1604 /* The maximum memory obtained from system via sbrk */ 1605 static unsigned long max_sbrked_mem = 0; 1606 1607 /* The maximum via either sbrk or mmap */ 1608 static unsigned long max_total_mem = 0; 1609 1610 /* internal working copy of mallinfo */ 1611 static struct mallinfo current_mallinfo = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }; 1612 1613 /* The total memory obtained from system via sbrk */ 1614 #define sbrked_mem (current_mallinfo.arena) 1615 1616 /* Tracking mmaps */ 1617 1618 #ifdef DEBUG 1619 static unsigned int n_mmaps = 0; 1620 #endif /* DEBUG */ 1621 static unsigned long mmapped_mem = 0; 1622 #if HAVE_MMAP 1623 static unsigned int max_n_mmaps = 0; 1624 static unsigned long max_mmapped_mem = 0; 1625 #endif 1626 1627 1628 1629 /* 1630 Debugging support 1631 */ 1632 1633 #ifdef DEBUG 1634 1635 1636 /* 1637 These routines make a number of assertions about the states 1638 of data structures that should be true at all times. If any 1639 are not true, it's very likely that a user program has somehow 1640 trashed memory. (It's also possible that there is a coding error 1641 in malloc. In which case, please report it!) 1642 */ 1643 1644 #if __STD_C 1645 static void do_check_chunk(mchunkptr p) 1646 #else 1647 static void do_check_chunk(p) mchunkptr p; 1648 #endif 1649 { 1650 INTERNAL_SIZE_T sz = p->size & ~PREV_INUSE; 1651 1652 /* No checkable chunk is mmapped */ 1653 assert(!chunk_is_mmapped(p)); 1654 1655 /* Check for legal address ... */ 1656 assert((char*)p >= sbrk_base); 1657 if (p != top) 1658 assert((char*)p + sz <= (char*)top); 1659 else 1660 assert((char*)p + sz <= sbrk_base + sbrked_mem); 1661 1662 } 1663 1664 1665 #if __STD_C 1666 static void do_check_free_chunk(mchunkptr p) 1667 #else 1668 static void do_check_free_chunk(p) mchunkptr p; 1669 #endif 1670 { 1671 INTERNAL_SIZE_T sz = p->size & ~PREV_INUSE; 1672 mchunkptr next = chunk_at_offset(p, sz); 1673 1674 do_check_chunk(p); 1675 1676 /* Check whether it claims to be free ... */ 1677 assert(!inuse(p)); 1678 1679 /* Unless a special marker, must have OK fields */ 1680 if ((long)sz >= (long)MINSIZE) 1681 { 1682 assert((sz & MALLOC_ALIGN_MASK) == 0); 1683 assert(aligned_OK(chunk2mem(p))); 1684 /* ... matching footer field */ 1685 assert(next->prev_size == sz); 1686 /* ... and is fully consolidated */ 1687 assert(prev_inuse(p)); 1688 assert (next == top || inuse(next)); 1689 1690 /* ... and has minimally sane links */ 1691 assert(p->fd->bk == p); 1692 assert(p->bk->fd == p); 1693 } 1694 else /* markers are always of size SIZE_SZ */ 1695 assert(sz == SIZE_SZ); 1696 } 1697 1698 #if __STD_C 1699 static void do_check_inuse_chunk(mchunkptr p) 1700 #else 1701 static void do_check_inuse_chunk(p) mchunkptr p; 1702 #endif 1703 { 1704 mchunkptr next = next_chunk(p); 1705 do_check_chunk(p); 1706 1707 /* Check whether it claims to be in use ... */ 1708 assert(inuse(p)); 1709 1710 /* ... and is surrounded by OK chunks. 1711 Since more things can be checked with free chunks than inuse ones, 1712 if an inuse chunk borders them and debug is on, it's worth doing them. 1713 */ 1714 if (!prev_inuse(p)) 1715 { 1716 mchunkptr prv = prev_chunk(p); 1717 assert(next_chunk(prv) == p); 1718 do_check_free_chunk(prv); 1719 } 1720 if (next == top) 1721 { 1722 assert(prev_inuse(next)); 1723 assert(chunksize(next) >= MINSIZE); 1724 } 1725 else if (!inuse(next)) 1726 do_check_free_chunk(next); 1727 1728 } 1729 1730 #if __STD_C 1731 static void do_check_malloced_chunk(mchunkptr p, INTERNAL_SIZE_T s) 1732 #else 1733 static void do_check_malloced_chunk(p, s) mchunkptr p; INTERNAL_SIZE_T s; 1734 #endif 1735 { 1736 INTERNAL_SIZE_T sz = p->size & ~PREV_INUSE; 1737 long room = sz - s; 1738 1739 do_check_inuse_chunk(p); 1740 1741 /* Legal size ... */ 1742 assert((long)sz >= (long)MINSIZE); 1743 assert((sz & MALLOC_ALIGN_MASK) == 0); 1744 assert(room >= 0); 1745 assert(room < (long)MINSIZE); 1746 1747 /* ... and alignment */ 1748 assert(aligned_OK(chunk2mem(p))); 1749 1750 1751 /* ... and was allocated at front of an available chunk */ 1752 assert(prev_inuse(p)); 1753 1754 } 1755 1756 1757 #define check_free_chunk(P) do_check_free_chunk(P) 1758 #define check_inuse_chunk(P) do_check_inuse_chunk(P) 1759 #define check_chunk(P) do_check_chunk(P) 1760 #define check_malloced_chunk(P,N) do_check_malloced_chunk(P,N) 1761 #else 1762 #define check_free_chunk(P) 1763 #define check_inuse_chunk(P) 1764 #define check_chunk(P) 1765 #define check_malloced_chunk(P,N) 1766 #endif 1767 1768 1769 1770 /* 1771 Macro-based internal utilities 1772 */ 1773 1774 1775 /* 1776 Linking chunks in bin lists. 1777 Call these only with variables, not arbitrary expressions, as arguments. 1778 */ 1779 1780 /* 1781 Place chunk p of size s in its bin, in size order, 1782 putting it ahead of others of same size. 1783 */ 1784 1785 1786 #define frontlink(P, S, IDX, BK, FD) \ 1787 { \ 1788 if (S < MAX_SMALLBIN_SIZE) \ 1789 { \ 1790 IDX = smallbin_index(S); \ 1791 mark_binblock(IDX); \ 1792 BK = bin_at(IDX); \ 1793 FD = BK->fd; \ 1794 P->bk = BK; \ 1795 P->fd = FD; \ 1796 FD->bk = BK->fd = P; \ 1797 } \ 1798 else \ 1799 { \ 1800 IDX = bin_index(S); \ 1801 BK = bin_at(IDX); \ 1802 FD = BK->fd; \ 1803 if (FD == BK) mark_binblock(IDX); \ 1804 else \ 1805 { \ 1806 while (FD != BK && S < chunksize(FD)) FD = FD->fd; \ 1807 BK = FD->bk; \ 1808 } \ 1809 P->bk = BK; \ 1810 P->fd = FD; \ 1811 FD->bk = BK->fd = P; \ 1812 } \ 1813 } 1814 1815 1816 /* take a chunk off a list */ 1817 1818 #define unlink(P, BK, FD) \ 1819 { \ 1820 BK = P->bk; \ 1821 FD = P->fd; \ 1822 FD->bk = BK; \ 1823 BK->fd = FD; \ 1824 } \ 1825 1826 /* Place p as the last remainder */ 1827 1828 #define link_last_remainder(P) \ 1829 { \ 1830 last_remainder->fd = last_remainder->bk = P; \ 1831 P->fd = P->bk = last_remainder; \ 1832 } 1833 1834 /* Clear the last_remainder bin */ 1835 1836 #define clear_last_remainder \ 1837 (last_remainder->fd = last_remainder->bk = last_remainder) 1838 1839 1840 1841 1842 1843 /* Routines dealing with mmap(). */ 1844 1845 #if HAVE_MMAP 1846 1847 #if __STD_C 1848 static mchunkptr mmap_chunk(size_t size) 1849 #else 1850 static mchunkptr mmap_chunk(size) size_t size; 1851 #endif 1852 { 1853 size_t page_mask = malloc_getpagesize - 1; 1854 mchunkptr p; 1855 1856 #ifndef MAP_ANONYMOUS 1857 static int fd = -1; 1858 #endif 1859 1860 if(n_mmaps >= n_mmaps_max) return 0; /* too many regions */ 1861 1862 /* For mmapped chunks, the overhead is one SIZE_SZ unit larger, because 1863 * there is no following chunk whose prev_size field could be used. 1864 */ 1865 size = (size + SIZE_SZ + page_mask) & ~page_mask; 1866 1867 #ifdef MAP_ANONYMOUS 1868 p = (mchunkptr)mmap(0, size, PROT_READ|PROT_WRITE, 1869 MAP_PRIVATE|MAP_ANONYMOUS, -1, 0); 1870 #else /* !MAP_ANONYMOUS */ 1871 if (fd < 0) 1872 { 1873 fd = open("/dev/zero", O_RDWR); 1874 if(fd < 0) return 0; 1875 } 1876 p = (mchunkptr)mmap(0, size, PROT_READ|PROT_WRITE, MAP_PRIVATE, fd, 0); 1877 #endif 1878 1879 if(p == (mchunkptr)-1) return 0; 1880 1881 n_mmaps++; 1882 if (n_mmaps > max_n_mmaps) max_n_mmaps = n_mmaps; 1883 1884 /* We demand that eight bytes into a page must be 8-byte aligned. */ 1885 assert(aligned_OK(chunk2mem(p))); 1886 1887 /* The offset to the start of the mmapped region is stored 1888 * in the prev_size field of the chunk; normally it is zero, 1889 * but that can be changed in memalign(). 1890 */ 1891 p->prev_size = 0; 1892 set_head(p, size|IS_MMAPPED); 1893 1894 mmapped_mem += size; 1895 if ((unsigned long)mmapped_mem > (unsigned long)max_mmapped_mem) 1896 max_mmapped_mem = mmapped_mem; 1897 if ((unsigned long)(mmapped_mem + sbrked_mem) > (unsigned long)max_total_mem) 1898 max_total_mem = mmapped_mem + sbrked_mem; 1899 return p; 1900 } 1901 1902 #if __STD_C 1903 static void munmap_chunk(mchunkptr p) 1904 #else 1905 static void munmap_chunk(p) mchunkptr p; 1906 #endif 1907 { 1908 INTERNAL_SIZE_T size = chunksize(p); 1909 int ret; 1910 1911 assert (chunk_is_mmapped(p)); 1912 assert(! ((char*)p >= sbrk_base && (char*)p < sbrk_base + sbrked_mem)); 1913 assert((n_mmaps > 0)); 1914 assert(((p->prev_size + size) & (malloc_getpagesize-1)) == 0); 1915 1916 n_mmaps--; 1917 mmapped_mem -= (size + p->prev_size); 1918 1919 ret = munmap((char *)p - p->prev_size, size + p->prev_size); 1920 1921 /* munmap returns non-zero on failure */ 1922 assert(ret == 0); 1923 } 1924 1925 #if HAVE_MREMAP 1926 1927 #if __STD_C 1928 static mchunkptr mremap_chunk(mchunkptr p, size_t new_size) 1929 #else 1930 static mchunkptr mremap_chunk(p, new_size) mchunkptr p; size_t new_size; 1931 #endif 1932 { 1933 size_t page_mask = malloc_getpagesize - 1; 1934 INTERNAL_SIZE_T offset = p->prev_size; 1935 INTERNAL_SIZE_T size = chunksize(p); 1936 char *cp; 1937 1938 assert (chunk_is_mmapped(p)); 1939 assert(! ((char*)p >= sbrk_base && (char*)p < sbrk_base + sbrked_mem)); 1940 assert((n_mmaps > 0)); 1941 assert(((size + offset) & (malloc_getpagesize-1)) == 0); 1942 1943 /* Note the extra SIZE_SZ overhead as in mmap_chunk(). */ 1944 new_size = (new_size + offset + SIZE_SZ + page_mask) & ~page_mask; 1945 1946 cp = (char *)mremap((char *)p - offset, size + offset, new_size, 1); 1947 1948 if (cp == (char *)-1) return 0; 1949 1950 p = (mchunkptr)(cp + offset); 1951 1952 assert(aligned_OK(chunk2mem(p))); 1953 1954 assert((p->prev_size == offset)); 1955 set_head(p, (new_size - offset)|IS_MMAPPED); 1956 1957 mmapped_mem -= size + offset; 1958 mmapped_mem += new_size; 1959 if ((unsigned long)mmapped_mem > (unsigned long)max_mmapped_mem) 1960 max_mmapped_mem = mmapped_mem; 1961 if ((unsigned long)(mmapped_mem + sbrked_mem) > (unsigned long)max_total_mem) 1962 max_total_mem = mmapped_mem + sbrked_mem; 1963 return p; 1964 } 1965 1966 #endif /* HAVE_MREMAP */ 1967 1968 #endif /* HAVE_MMAP */ 1969 1970 1971 1972 1973 /* 1974 Extend the top-most chunk by obtaining memory from system. 1975 Main interface to sbrk (but see also malloc_trim). 1976 */ 1977 1978 #if __STD_C 1979 static void malloc_extend_top(INTERNAL_SIZE_T nb) 1980 #else 1981 static void malloc_extend_top(nb) INTERNAL_SIZE_T nb; 1982 #endif 1983 { 1984 char* brk; /* return value from sbrk */ 1985 INTERNAL_SIZE_T front_misalign; /* unusable bytes at front of sbrked space */ 1986 INTERNAL_SIZE_T correction; /* bytes for 2nd sbrk call */ 1987 char* new_brk; /* return of 2nd sbrk call */ 1988 INTERNAL_SIZE_T top_size; /* new size of top chunk */ 1989 1990 mchunkptr old_top = top; /* Record state of old top */ 1991 INTERNAL_SIZE_T old_top_size = chunksize(old_top); 1992 char* old_end = (char*)(chunk_at_offset(old_top, old_top_size)); 1993 1994 /* Pad request with top_pad plus minimal overhead */ 1995 1996 INTERNAL_SIZE_T sbrk_size = nb + top_pad + MINSIZE; 1997 unsigned long pagesz = malloc_getpagesize; 1998 1999 /* If not the first time through, round to preserve page boundary */ 2000 /* Otherwise, we need to correct to a page size below anyway. */ 2001 /* (We also correct below if an intervening foreign sbrk call.) */ 2002 2003 if (sbrk_base != (char*)(-1)) 2004 sbrk_size = (sbrk_size + (pagesz - 1)) & ~(pagesz - 1); 2005 2006 brk = (char*)(MORECORE (sbrk_size)); 2007 2008 /* Fail if sbrk failed or if a foreign sbrk call killed our space */ 2009 if (brk == (char*)(MORECORE_FAILURE) || 2010 (brk < old_end && old_top != initial_top)) 2011 return; 2012 2013 sbrked_mem += sbrk_size; 2014 2015 if (brk == old_end) /* can just add bytes to current top */ 2016 { 2017 top_size = sbrk_size + old_top_size; 2018 set_head(top, top_size | PREV_INUSE); 2019 } 2020 else 2021 { 2022 if (sbrk_base == (char*)(-1)) /* First time through. Record base */ 2023 sbrk_base = brk; 2024 else /* Someone else called sbrk(). Count those bytes as sbrked_mem. */ 2025 sbrked_mem += brk - (char*)old_end; 2026 2027 /* Guarantee alignment of first new chunk made from this space */ 2028 front_misalign = (unsigned long)chunk2mem(brk) & MALLOC_ALIGN_MASK; 2029 if (front_misalign > 0) 2030 { 2031 correction = (MALLOC_ALIGNMENT) - front_misalign; 2032 brk += correction; 2033 } 2034 else 2035 correction = 0; 2036 2037 /* Guarantee the next brk will be at a page boundary */ 2038 2039 correction += ((((unsigned long)(brk + sbrk_size))+(pagesz-1)) & 2040 ~(pagesz - 1)) - ((unsigned long)(brk + sbrk_size)); 2041 2042 /* Allocate correction */ 2043 new_brk = (char*)(MORECORE (correction)); 2044 if (new_brk == (char*)(MORECORE_FAILURE)) return; 2045 2046 sbrked_mem += correction; 2047 2048 top = (mchunkptr)brk; 2049 top_size = new_brk - brk + correction; 2050 set_head(top, top_size | PREV_INUSE); 2051 2052 if (old_top != initial_top) 2053 { 2054 2055 /* There must have been an intervening foreign sbrk call. */ 2056 /* A double fencepost is necessary to prevent consolidation */ 2057 2058 /* If not enough space to do this, then user did something very wrong */ 2059 if (old_top_size < MINSIZE) 2060 { 2061 set_head(top, PREV_INUSE); /* will force null return from malloc */ 2062 return; 2063 } 2064 2065 /* Also keep size a multiple of MALLOC_ALIGNMENT */ 2066 old_top_size = (old_top_size - 3*SIZE_SZ) & ~MALLOC_ALIGN_MASK; 2067 set_head_size(old_top, old_top_size); 2068 chunk_at_offset(old_top, old_top_size )->size = 2069 SIZE_SZ|PREV_INUSE; 2070 chunk_at_offset(old_top, old_top_size + SIZE_SZ)->size = 2071 SIZE_SZ|PREV_INUSE; 2072 /* If possible, release the rest. */ 2073 if (old_top_size >= MINSIZE) 2074 fREe(chunk2mem(old_top)); 2075 } 2076 } 2077 2078 if ((unsigned long)sbrked_mem > (unsigned long)max_sbrked_mem) 2079 max_sbrked_mem = sbrked_mem; 2080 if ((unsigned long)(mmapped_mem + sbrked_mem) > (unsigned long)max_total_mem) 2081 max_total_mem = mmapped_mem + sbrked_mem; 2082 2083 /* We always land on a page boundary */ 2084 assert(((unsigned long)((char*)top + top_size) & (pagesz - 1)) == 0); 2085 } 2086 2087 2088 2089 2090 /* Main public routines */ 2091 2092 2093 /* 2094 Malloc Algorthim: 2095 2096 The requested size is first converted into a usable form, `nb'. 2097 This currently means to add 4 bytes overhead plus possibly more to 2098 obtain 8-byte alignment and/or to obtain a size of at least 2099 MINSIZE (currently 16 bytes), the smallest allocatable size. 2100 (All fits are considered `exact' if they are within MINSIZE bytes.) 2101 2102 From there, the first successful of the following steps is taken: 2103 2104 1. The bin corresponding to the request size is scanned, and if 2105 a chunk of exactly the right size is found, it is taken. 2106 2107 2. The most recently remaindered chunk is used if it is big 2108 enough. This is a form of (roving) first fit, used only in 2109 the absence of exact fits. Runs of consecutive requests use 2110 the remainder of the chunk used for the previous such request 2111 whenever possible. This limited use of a first-fit style 2112 allocation strategy tends to give contiguous chunks 2113 coextensive lifetimes, which improves locality and can reduce 2114 fragmentation in the long run. 2115 2116 3. Other bins are scanned in increasing size order, using a 2117 chunk big enough to fulfill the request, and splitting off 2118 any remainder. This search is strictly by best-fit; i.e., 2119 the smallest (with ties going to approximately the least 2120 recently used) chunk that fits is selected. 2121 2122 4. If large enough, the chunk bordering the end of memory 2123 (`top') is split off. (This use of `top' is in accord with 2124 the best-fit search rule. In effect, `top' is treated as 2125 larger (and thus less well fitting) than any other available 2126 chunk since it can be extended to be as large as necessary 2127 (up to system limitations). 2128 2129 5. If the request size meets the mmap threshold and the 2130 system supports mmap, and there are few enough currently 2131 allocated mmapped regions, and a call to mmap succeeds, 2132 the request is allocated via direct memory mapping. 2133 2134 6. Otherwise, the top of memory is extended by 2135 obtaining more space from the system (normally using sbrk, 2136 but definable to anything else via the MORECORE macro). 2137 Memory is gathered from the system (in system page-sized 2138 units) in a way that allows chunks obtained across different 2139 sbrk calls to be consolidated, but does not require 2140 contiguous memory. Thus, it should be safe to intersperse 2141 mallocs with other sbrk calls. 2142 2143 2144 All allocations are made from the the `lowest' part of any found 2145 chunk. (The implementation invariant is that prev_inuse is 2146 always true of any allocated chunk; i.e., that each allocated 2147 chunk borders either a previously allocated and still in-use chunk, 2148 or the base of its memory arena.) 2149 2150 */ 2151 2152 #if __STD_C 2153 Void_t* mALLOc(size_t bytes) 2154 #else 2155 Void_t* mALLOc(bytes) size_t bytes; 2156 #endif 2157 { 2158 mchunkptr victim; /* inspected/selected chunk */ 2159 INTERNAL_SIZE_T victim_size; /* its size */ 2160 int idx; /* index for bin traversal */ 2161 mbinptr bin; /* associated bin */ 2162 mchunkptr remainder; /* remainder from a split */ 2163 long remainder_size; /* its size */ 2164 int remainder_index; /* its bin index */ 2165 unsigned long block; /* block traverser bit */ 2166 int startidx; /* first bin of a traversed block */ 2167 mchunkptr fwd; /* misc temp for linking */ 2168 mchunkptr bck; /* misc temp for linking */ 2169 mbinptr q; /* misc temp */ 2170 2171 INTERNAL_SIZE_T nb; 2172 2173 /* check if mem_malloc_init() was run */ 2174 if ((mem_malloc_start == 0) && (mem_malloc_end == 0)) { 2175 /* not initialized yet */ 2176 return 0; 2177 } 2178 2179 if ((long)bytes < 0) return 0; 2180 2181 nb = request2size(bytes); /* padded request size; */ 2182 2183 /* Check for exact match in a bin */ 2184 2185 if (is_small_request(nb)) /* Faster version for small requests */ 2186 { 2187 idx = smallbin_index(nb); 2188 2189 /* No traversal or size check necessary for small bins. */ 2190 2191 q = bin_at(idx); 2192 victim = last(q); 2193 2194 /* Also scan the next one, since it would have a remainder < MINSIZE */ 2195 if (victim == q) 2196 { 2197 q = next_bin(q); 2198 victim = last(q); 2199 } 2200 if (victim != q) 2201 { 2202 victim_size = chunksize(victim); 2203 unlink(victim, bck, fwd); 2204 set_inuse_bit_at_offset(victim, victim_size); 2205 check_malloced_chunk(victim, nb); 2206 return chunk2mem(victim); 2207 } 2208 2209 idx += 2; /* Set for bin scan below. We've already scanned 2 bins. */ 2210 2211 } 2212 else 2213 { 2214 idx = bin_index(nb); 2215 bin = bin_at(idx); 2216 2217 for (victim = last(bin); victim != bin; victim = victim->bk) 2218 { 2219 victim_size = chunksize(victim); 2220 remainder_size = victim_size - nb; 2221 2222 if (remainder_size >= (long)MINSIZE) /* too big */ 2223 { 2224 --idx; /* adjust to rescan below after checking last remainder */ 2225 break; 2226 } 2227 2228 else if (remainder_size >= 0) /* exact fit */ 2229 { 2230 unlink(victim, bck, fwd); 2231 set_inuse_bit_at_offset(victim, victim_size); 2232 check_malloced_chunk(victim, nb); 2233 return chunk2mem(victim); 2234 } 2235 } 2236 2237 ++idx; 2238 2239 } 2240 2241 /* Try to use the last split-off remainder */ 2242 2243 if ( (victim = last_remainder->fd) != last_remainder) 2244 { 2245 victim_size = chunksize(victim); 2246 remainder_size = victim_size - nb; 2247 2248 if (remainder_size >= (long)MINSIZE) /* re-split */ 2249 { 2250 remainder = chunk_at_offset(victim, nb); 2251 set_head(victim, nb | PREV_INUSE); 2252 link_last_remainder(remainder); 2253 set_head(remainder, remainder_size | PREV_INUSE); 2254 set_foot(remainder, remainder_size); 2255 check_malloced_chunk(victim, nb); 2256 return chunk2mem(victim); 2257 } 2258 2259 clear_last_remainder; 2260 2261 if (remainder_size >= 0) /* exhaust */ 2262 { 2263 set_inuse_bit_at_offset(victim, victim_size); 2264 check_malloced_chunk(victim, nb); 2265 return chunk2mem(victim); 2266 } 2267 2268 /* Else place in bin */ 2269 2270 frontlink(victim, victim_size, remainder_index, bck, fwd); 2271 } 2272 2273 /* 2274 If there are any possibly nonempty big-enough blocks, 2275 search for best fitting chunk by scanning bins in blockwidth units. 2276 */ 2277 2278 if ( (block = idx2binblock(idx)) <= binblocks_r) 2279 { 2280 2281 /* Get to the first marked block */ 2282 2283 if ( (block & binblocks_r) == 0) 2284 { 2285 /* force to an even block boundary */ 2286 idx = (idx & ~(BINBLOCKWIDTH - 1)) + BINBLOCKWIDTH; 2287 block <<= 1; 2288 while ((block & binblocks_r) == 0) 2289 { 2290 idx += BINBLOCKWIDTH; 2291 block <<= 1; 2292 } 2293 } 2294 2295 /* For each possibly nonempty block ... */ 2296 for (;;) 2297 { 2298 startidx = idx; /* (track incomplete blocks) */ 2299 q = bin = bin_at(idx); 2300 2301 /* For each bin in this block ... */ 2302 do 2303 { 2304 /* Find and use first big enough chunk ... */ 2305 2306 for (victim = last(bin); victim != bin; victim = victim->bk) 2307 { 2308 victim_size = chunksize(victim); 2309 remainder_size = victim_size - nb; 2310 2311 if (remainder_size >= (long)MINSIZE) /* split */ 2312 { 2313 remainder = chunk_at_offset(victim, nb); 2314 set_head(victim, nb | PREV_INUSE); 2315 unlink(victim, bck, fwd); 2316 link_last_remainder(remainder); 2317 set_head(remainder, remainder_size | PREV_INUSE); 2318 set_foot(remainder, remainder_size); 2319 check_malloced_chunk(victim, nb); 2320 return chunk2mem(victim); 2321 } 2322 2323 else if (remainder_size >= 0) /* take */ 2324 { 2325 set_inuse_bit_at_offset(victim, victim_size); 2326 unlink(victim, bck, fwd); 2327 check_malloced_chunk(victim, nb); 2328 return chunk2mem(victim); 2329 } 2330 2331 } 2332 2333 bin = next_bin(bin); 2334 2335 } while ((++idx & (BINBLOCKWIDTH - 1)) != 0); 2336 2337 /* Clear out the block bit. */ 2338 2339 do /* Possibly backtrack to try to clear a partial block */ 2340 { 2341 if ((startidx & (BINBLOCKWIDTH - 1)) == 0) 2342 { 2343 av_[1] = (mbinptr)(binblocks_r & ~block); 2344 break; 2345 } 2346 --startidx; 2347 q = prev_bin(q); 2348 } while (first(q) == q); 2349 2350 /* Get to the next possibly nonempty block */ 2351 2352 if ( (block <<= 1) <= binblocks_r && (block != 0) ) 2353 { 2354 while ((block & binblocks_r) == 0) 2355 { 2356 idx += BINBLOCKWIDTH; 2357 block <<= 1; 2358 } 2359 } 2360 else 2361 break; 2362 } 2363 } 2364 2365 2366 /* Try to use top chunk */ 2367 2368 /* Require that there be a remainder, ensuring top always exists */ 2369 if ( (remainder_size = chunksize(top) - nb) < (long)MINSIZE) 2370 { 2371 2372 #if HAVE_MMAP 2373 /* If big and would otherwise need to extend, try to use mmap instead */ 2374 if ((unsigned long)nb >= (unsigned long)mmap_threshold && 2375 (victim = mmap_chunk(nb)) != 0) 2376 return chunk2mem(victim); 2377 #endif 2378 2379 /* Try to extend */ 2380 malloc_extend_top(nb); 2381 if ( (remainder_size = chunksize(top) - nb) < (long)MINSIZE) 2382 return 0; /* propagate failure */ 2383 } 2384 2385 victim = top; 2386 set_head(victim, nb | PREV_INUSE); 2387 top = chunk_at_offset(victim, nb); 2388 set_head(top, remainder_size | PREV_INUSE); 2389 check_malloced_chunk(victim, nb); 2390 return chunk2mem(victim); 2391 2392 } 2393 2394 2395 2396 2397 /* 2398 2399 free() algorithm : 2400 2401 cases: 2402 2403 1. free(0) has no effect. 2404 2405 2. If the chunk was allocated via mmap, it is release via munmap(). 2406 2407 3. If a returned chunk borders the current high end of memory, 2408 it is consolidated into the top, and if the total unused 2409 topmost memory exceeds the trim threshold, malloc_trim is 2410 called. 2411 2412 4. Other chunks are consolidated as they arrive, and 2413 placed in corresponding bins. (This includes the case of 2414 consolidating with the current `last_remainder'). 2415 2416 */ 2417 2418 2419 #if __STD_C 2420 void fREe(Void_t* mem) 2421 #else 2422 void fREe(mem) Void_t* mem; 2423 #endif 2424 { 2425 mchunkptr p; /* chunk corresponding to mem */ 2426 INTERNAL_SIZE_T hd; /* its head field */ 2427 INTERNAL_SIZE_T sz; /* its size */ 2428 int idx; /* its bin index */ 2429 mchunkptr next; /* next contiguous chunk */ 2430 INTERNAL_SIZE_T nextsz; /* its size */ 2431 INTERNAL_SIZE_T prevsz; /* size of previous contiguous chunk */ 2432 mchunkptr bck; /* misc temp for linking */ 2433 mchunkptr fwd; /* misc temp for linking */ 2434 int islr; /* track whether merging with last_remainder */ 2435 2436 if (mem == 0) /* free(0) has no effect */ 2437 return; 2438 2439 p = mem2chunk(mem); 2440 hd = p->size; 2441 2442 #if HAVE_MMAP 2443 if (hd & IS_MMAPPED) /* release mmapped memory. */ 2444 { 2445 munmap_chunk(p); 2446 return; 2447 } 2448 #endif 2449 2450 check_inuse_chunk(p); 2451 2452 sz = hd & ~PREV_INUSE; 2453 next = chunk_at_offset(p, sz); 2454 nextsz = chunksize(next); 2455 2456 if (next == top) /* merge with top */ 2457 { 2458 sz += nextsz; 2459 2460 if (!(hd & PREV_INUSE)) /* consolidate backward */ 2461 { 2462 prevsz = p->prev_size; 2463 p = chunk_at_offset(p, -((long) prevsz)); 2464 sz += prevsz; 2465 unlink(p, bck, fwd); 2466 } 2467 2468 set_head(p, sz | PREV_INUSE); 2469 top = p; 2470 if ((unsigned long)(sz) >= (unsigned long)trim_threshold) 2471 malloc_trim(top_pad); 2472 return; 2473 } 2474 2475 set_head(next, nextsz); /* clear inuse bit */ 2476 2477 islr = 0; 2478 2479 if (!(hd & PREV_INUSE)) /* consolidate backward */ 2480 { 2481 prevsz = p->prev_size; 2482 p = chunk_at_offset(p, -((long) prevsz)); 2483 sz += prevsz; 2484 2485 if (p->fd == last_remainder) /* keep as last_remainder */ 2486 islr = 1; 2487 else 2488 unlink(p, bck, fwd); 2489 } 2490 2491 if (!(inuse_bit_at_offset(next, nextsz))) /* consolidate forward */ 2492 { 2493 sz += nextsz; 2494 2495 if (!islr && next->fd == last_remainder) /* re-insert last_remainder */ 2496 { 2497 islr = 1; 2498 link_last_remainder(p); 2499 } 2500 else 2501 unlink(next, bck, fwd); 2502 } 2503 2504 2505 set_head(p, sz | PREV_INUSE); 2506 set_foot(p, sz); 2507 if (!islr) 2508 frontlink(p, sz, idx, bck, fwd); 2509 } 2510 2511 2512 2513 2514 2515 /* 2516 2517 Realloc algorithm: 2518 2519 Chunks that were obtained via mmap cannot be extended or shrunk 2520 unless HAVE_MREMAP is defined, in which case mremap is used. 2521 Otherwise, if their reallocation is for additional space, they are 2522 copied. If for less, they are just left alone. 2523 2524 Otherwise, if the reallocation is for additional space, and the 2525 chunk can be extended, it is, else a malloc-copy-free sequence is 2526 taken. There are several different ways that a chunk could be 2527 extended. All are tried: 2528 2529 * Extending forward into following adjacent free chunk. 2530 * Shifting backwards, joining preceding adjacent space 2531 * Both shifting backwards and extending forward. 2532 * Extending into newly sbrked space 2533 2534 Unless the #define REALLOC_ZERO_BYTES_FREES is set, realloc with a 2535 size argument of zero (re)allocates a minimum-sized chunk. 2536 2537 If the reallocation is for less space, and the new request is for 2538 a `small' (<512 bytes) size, then the newly unused space is lopped 2539 off and freed. 2540 2541 The old unix realloc convention of allowing the last-free'd chunk 2542 to be used as an argument to realloc is no longer supported. 2543 I don't know of any programs still relying on this feature, 2544 and allowing it would also allow too many other incorrect 2545 usages of realloc to be sensible. 2546 2547 2548 */ 2549 2550 2551 #if __STD_C 2552 Void_t* rEALLOc(Void_t* oldmem, size_t bytes) 2553 #else 2554 Void_t* rEALLOc(oldmem, bytes) Void_t* oldmem; size_t bytes; 2555 #endif 2556 { 2557 INTERNAL_SIZE_T nb; /* padded request size */ 2558 2559 mchunkptr oldp; /* chunk corresponding to oldmem */ 2560 INTERNAL_SIZE_T oldsize; /* its size */ 2561 2562 mchunkptr newp; /* chunk to return */ 2563 INTERNAL_SIZE_T newsize; /* its size */ 2564 Void_t* newmem; /* corresponding user mem */ 2565 2566 mchunkptr next; /* next contiguous chunk after oldp */ 2567 INTERNAL_SIZE_T nextsize; /* its size */ 2568 2569 mchunkptr prev; /* previous contiguous chunk before oldp */ 2570 INTERNAL_SIZE_T prevsize; /* its size */ 2571 2572 mchunkptr remainder; /* holds split off extra space from newp */ 2573 INTERNAL_SIZE_T remainder_size; /* its size */ 2574 2575 mchunkptr bck; /* misc temp for linking */ 2576 mchunkptr fwd; /* misc temp for linking */ 2577 2578 #ifdef REALLOC_ZERO_BYTES_FREES 2579 if (bytes == 0) { fREe(oldmem); return 0; } 2580 #endif 2581 2582 if ((long)bytes < 0) return 0; 2583 2584 /* realloc of null is supposed to be same as malloc */ 2585 if (oldmem == 0) return mALLOc(bytes); 2586 2587 newp = oldp = mem2chunk(oldmem); 2588 newsize = oldsize = chunksize(oldp); 2589 2590 2591 nb = request2size(bytes); 2592 2593 #if HAVE_MMAP 2594 if (chunk_is_mmapped(oldp)) 2595 { 2596 #if HAVE_MREMAP 2597 newp = mremap_chunk(oldp, nb); 2598 if(newp) return chunk2mem(newp); 2599 #endif 2600 /* Note the extra SIZE_SZ overhead. */ 2601 if(oldsize - SIZE_SZ >= nb) return oldmem; /* do nothing */ 2602 /* Must alloc, copy, free. */ 2603 newmem = mALLOc(bytes); 2604 if (newmem == 0) return 0; /* propagate failure */ 2605 MALLOC_COPY(newmem, oldmem, oldsize - 2*SIZE_SZ); 2606 munmap_chunk(oldp); 2607 return newmem; 2608 } 2609 #endif 2610 2611 check_inuse_chunk(oldp); 2612 2613 if ((long)(oldsize) < (long)(nb)) 2614 { 2615 2616 /* Try expanding forward */ 2617 2618 next = chunk_at_offset(oldp, oldsize); 2619 if (next == top || !inuse(next)) 2620 { 2621 nextsize = chunksize(next); 2622 2623 /* Forward into top only if a remainder */ 2624 if (next == top) 2625 { 2626 if ((long)(nextsize + newsize) >= (long)(nb + MINSIZE)) 2627 { 2628 newsize += nextsize; 2629 top = chunk_at_offset(oldp, nb); 2630 set_head(top, (newsize - nb) | PREV_INUSE); 2631 set_head_size(oldp, nb); 2632 return chunk2mem(oldp); 2633 } 2634 } 2635 2636 /* Forward into next chunk */ 2637 else if (((long)(nextsize + newsize) >= (long)(nb))) 2638 { 2639 unlink(next, bck, fwd); 2640 newsize += nextsize; 2641 goto split; 2642 } 2643 } 2644 else 2645 { 2646 next = 0; 2647 nextsize = 0; 2648 } 2649 2650 /* Try shifting backwards. */ 2651 2652 if (!prev_inuse(oldp)) 2653 { 2654 prev = prev_chunk(oldp); 2655 prevsize = chunksize(prev); 2656 2657 /* try forward + backward first to save a later consolidation */ 2658 2659 if (next != 0) 2660 { 2661 /* into top */ 2662 if (next == top) 2663 { 2664 if ((long)(nextsize + prevsize + newsize) >= (long)(nb + MINSIZE)) 2665 { 2666 unlink(prev, bck, fwd); 2667 newp = prev; 2668 newsize += prevsize + nextsize; 2669 newmem = chunk2mem(newp); 2670 MALLOC_COPY(newmem, oldmem, oldsize - SIZE_SZ); 2671 top = chunk_at_offset(newp, nb); 2672 set_head(top, (newsize - nb) | PREV_INUSE); 2673 set_head_size(newp, nb); 2674 return newmem; 2675 } 2676 } 2677 2678 /* into next chunk */ 2679 else if (((long)(nextsize + prevsize + newsize) >= (long)(nb))) 2680 { 2681 unlink(next, bck, fwd); 2682 unlink(prev, bck, fwd); 2683 newp = prev; 2684 newsize += nextsize + prevsize; 2685 newmem = chunk2mem(newp); 2686 MALLOC_COPY(newmem, oldmem, oldsize - SIZE_SZ); 2687 goto split; 2688 } 2689 } 2690 2691 /* backward only */ 2692 if (prev != 0 && (long)(prevsize + newsize) >= (long)nb) 2693 { 2694 unlink(prev, bck, fwd); 2695 newp = prev; 2696 newsize += prevsize; 2697 newmem = chunk2mem(newp); 2698 MALLOC_COPY(newmem, oldmem, oldsize - SIZE_SZ); 2699 goto split; 2700 } 2701 } 2702 2703 /* Must allocate */ 2704 2705 newmem = mALLOc (bytes); 2706 2707 if (newmem == 0) /* propagate failure */ 2708 return 0; 2709 2710 /* Avoid copy if newp is next chunk after oldp. */ 2711 /* (This can only happen when new chunk is sbrk'ed.) */ 2712 2713 if ( (newp = mem2chunk(newmem)) == next_chunk(oldp)) 2714 { 2715 newsize += chunksize(newp); 2716 newp = oldp; 2717 goto split; 2718 } 2719 2720 /* Otherwise copy, free, and exit */ 2721 MALLOC_COPY(newmem, oldmem, oldsize - SIZE_SZ); 2722 fREe(oldmem); 2723 return newmem; 2724 } 2725 2726 2727 split: /* split off extra room in old or expanded chunk */ 2728 2729 if (newsize - nb >= MINSIZE) /* split off remainder */ 2730 { 2731 remainder = chunk_at_offset(newp, nb); 2732 remainder_size = newsize - nb; 2733 set_head_size(newp, nb); 2734 set_head(remainder, remainder_size | PREV_INUSE); 2735 set_inuse_bit_at_offset(remainder, remainder_size); 2736 fREe(chunk2mem(remainder)); /* let free() deal with it */ 2737 } 2738 else 2739 { 2740 set_head_size(newp, newsize); 2741 set_inuse_bit_at_offset(newp, newsize); 2742 } 2743 2744 check_inuse_chunk(newp); 2745 return chunk2mem(newp); 2746 } 2747 2748 2749 2750 2751 /* 2752 2753 memalign algorithm: 2754 2755 memalign requests more than enough space from malloc, finds a spot 2756 within that chunk that meets the alignment request, and then 2757 possibly frees the leading and trailing space. 2758 2759 The alignment argument must be a power of two. This property is not 2760 checked by memalign, so misuse may result in random runtime errors. 2761 2762 8-byte alignment is guaranteed by normal malloc calls, so don't 2763 bother calling memalign with an argument of 8 or less. 2764 2765 Overreliance on memalign is a sure way to fragment space. 2766 2767 */ 2768 2769 2770 #if __STD_C 2771 Void_t* mEMALIGn(size_t alignment, size_t bytes) 2772 #else 2773 Void_t* mEMALIGn(alignment, bytes) size_t alignment; size_t bytes; 2774 #endif 2775 { 2776 INTERNAL_SIZE_T nb; /* padded request size */ 2777 char* m; /* memory returned by malloc call */ 2778 mchunkptr p; /* corresponding chunk */ 2779 char* brk; /* alignment point within p */ 2780 mchunkptr newp; /* chunk to return */ 2781 INTERNAL_SIZE_T newsize; /* its size */ 2782 INTERNAL_SIZE_T leadsize; /* leading space befor alignment point */ 2783 mchunkptr remainder; /* spare room at end to split off */ 2784 long remainder_size; /* its size */ 2785 2786 if ((long)bytes < 0) return 0; 2787 2788 /* If need less alignment than we give anyway, just relay to malloc */ 2789 2790 if (alignment <= MALLOC_ALIGNMENT) return mALLOc(bytes); 2791 2792 /* Otherwise, ensure that it is at least a minimum chunk size */ 2793 2794 if (alignment < MINSIZE) alignment = MINSIZE; 2795 2796 /* Call malloc with worst case padding to hit alignment. */ 2797 2798 nb = request2size(bytes); 2799 m = (char*)(mALLOc(nb + alignment + MINSIZE)); 2800 2801 if (m == 0) return 0; /* propagate failure */ 2802 2803 p = mem2chunk(m); 2804 2805 if ((((unsigned long)(m)) % alignment) == 0) /* aligned */ 2806 { 2807 #if HAVE_MMAP 2808 if(chunk_is_mmapped(p)) 2809 return chunk2mem(p); /* nothing more to do */ 2810 #endif 2811 } 2812 else /* misaligned */ 2813 { 2814 /* 2815 Find an aligned spot inside chunk. 2816 Since we need to give back leading space in a chunk of at 2817 least MINSIZE, if the first calculation places us at 2818 a spot with less than MINSIZE leader, we can move to the 2819 next aligned spot -- we've allocated enough total room so that 2820 this is always possible. 2821 */ 2822 2823 brk = (char*)mem2chunk(((unsigned long)(m + alignment - 1)) & -((signed) alignment)); 2824 if ((long)(brk - (char*)(p)) < MINSIZE) brk = brk + alignment; 2825 2826 newp = (mchunkptr)brk; 2827 leadsize = brk - (char*)(p); 2828 newsize = chunksize(p) - leadsize; 2829 2830 #if HAVE_MMAP 2831 if(chunk_is_mmapped(p)) 2832 { 2833 newp->prev_size = p->prev_size + leadsize; 2834 set_head(newp, newsize|IS_MMAPPED); 2835 return chunk2mem(newp); 2836 } 2837 #endif 2838 2839 /* give back leader, use the rest */ 2840 2841 set_head(newp, newsize | PREV_INUSE); 2842 set_inuse_bit_at_offset(newp, newsize); 2843 set_head_size(p, leadsize); 2844 fREe(chunk2mem(p)); 2845 p = newp; 2846 2847 assert (newsize >= nb && (((unsigned long)(chunk2mem(p))) % alignment) == 0); 2848 } 2849 2850 /* Also give back spare room at the end */ 2851 2852 remainder_size = chunksize(p) - nb; 2853 2854 if (remainder_size >= (long)MINSIZE) 2855 { 2856 remainder = chunk_at_offset(p, nb); 2857 set_head(remainder, remainder_size | PREV_INUSE); 2858 set_head_size(p, nb); 2859 fREe(chunk2mem(remainder)); 2860 } 2861 2862 check_inuse_chunk(p); 2863 return chunk2mem(p); 2864 2865 } 2866 2867 2868 2869 2870 /* 2871 valloc just invokes memalign with alignment argument equal 2872 to the page size of the system (or as near to this as can 2873 be figured out from all the includes/defines above.) 2874 */ 2875 2876 #if __STD_C 2877 Void_t* vALLOc(size_t bytes) 2878 #else 2879 Void_t* vALLOc(bytes) size_t bytes; 2880 #endif 2881 { 2882 return mEMALIGn (malloc_getpagesize, bytes); 2883 } 2884 2885 /* 2886 pvalloc just invokes valloc for the nearest pagesize 2887 that will accommodate request 2888 */ 2889 2890 2891 #if __STD_C 2892 Void_t* pvALLOc(size_t bytes) 2893 #else 2894 Void_t* pvALLOc(bytes) size_t bytes; 2895 #endif 2896 { 2897 size_t pagesize = malloc_getpagesize; 2898 return mEMALIGn (pagesize, (bytes + pagesize - 1) & ~(pagesize - 1)); 2899 } 2900 2901 /* 2902 2903 calloc calls malloc, then zeroes out the allocated chunk. 2904 2905 */ 2906 2907 #if __STD_C 2908 Void_t* cALLOc(size_t n, size_t elem_size) 2909 #else 2910 Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size; 2911 #endif 2912 { 2913 mchunkptr p; 2914 INTERNAL_SIZE_T csz; 2915 2916 INTERNAL_SIZE_T sz = n * elem_size; 2917 2918 2919 /* check if expand_top called, in which case don't need to clear */ 2920 #if MORECORE_CLEARS 2921 mchunkptr oldtop = top; 2922 INTERNAL_SIZE_T oldtopsize = chunksize(top); 2923 #endif 2924 Void_t* mem = mALLOc (sz); 2925 2926 if ((long)n < 0) return 0; 2927 2928 if (mem == 0) 2929 return 0; 2930 else 2931 { 2932 p = mem2chunk(mem); 2933 2934 /* Two optional cases in which clearing not necessary */ 2935 2936 2937 #if HAVE_MMAP 2938 if (chunk_is_mmapped(p)) return mem; 2939 #endif 2940 2941 csz = chunksize(p); 2942 2943 #if MORECORE_CLEARS 2944 if (p == oldtop && csz > oldtopsize) 2945 { 2946 /* clear only the bytes from non-freshly-sbrked memory */ 2947 csz = oldtopsize; 2948 } 2949 #endif 2950 2951 MALLOC_ZERO(mem, csz - SIZE_SZ); 2952 return mem; 2953 } 2954 } 2955 2956 /* 2957 2958 cfree just calls free. It is needed/defined on some systems 2959 that pair it with calloc, presumably for odd historical reasons. 2960 2961 */ 2962 2963 #if !defined(INTERNAL_LINUX_C_LIB) || !defined(__ELF__) 2964 #if __STD_C 2965 void cfree(Void_t *mem) 2966 #else 2967 void cfree(mem) Void_t *mem; 2968 #endif 2969 { 2970 fREe(mem); 2971 } 2972 #endif 2973 2974 2975 2976 /* 2977 2978 Malloc_trim gives memory back to the system (via negative 2979 arguments to sbrk) if there is unused memory at the `high' end of 2980 the malloc pool. You can call this after freeing large blocks of 2981 memory to potentially reduce the system-level memory requirements 2982 of a program. However, it cannot guarantee to reduce memory. Under 2983 some allocation patterns, some large free blocks of memory will be 2984 locked between two used chunks, so they cannot be given back to 2985 the system. 2986 2987 The `pad' argument to malloc_trim represents the amount of free 2988 trailing space to leave untrimmed. If this argument is zero, 2989 only the minimum amount of memory to maintain internal data 2990 structures will be left (one page or less). Non-zero arguments 2991 can be supplied to maintain enough trailing space to service 2992 future expected allocations without having to re-obtain memory 2993 from the system. 2994 2995 Malloc_trim returns 1 if it actually released any memory, else 0. 2996 2997 */ 2998 2999 #if __STD_C 3000 int malloc_trim(size_t pad) 3001 #else 3002 int malloc_trim(pad) size_t pad; 3003 #endif 3004 { 3005 long top_size; /* Amount of top-most memory */ 3006 long extra; /* Amount to release */ 3007 char* current_brk; /* address returned by pre-check sbrk call */ 3008 char* new_brk; /* address returned by negative sbrk call */ 3009 3010 unsigned long pagesz = malloc_getpagesize; 3011 3012 top_size = chunksize(top); 3013 extra = ((top_size - pad - MINSIZE + (pagesz-1)) / pagesz - 1) * pagesz; 3014 3015 if (extra < (long)pagesz) /* Not enough memory to release */ 3016 return 0; 3017 3018 else 3019 { 3020 /* Test to make sure no one else called sbrk */ 3021 current_brk = (char*)(MORECORE (0)); 3022 if (current_brk != (char*)(top) + top_size) 3023 return 0; /* Apparently we don't own memory; must fail */ 3024 3025 else 3026 { 3027 new_brk = (char*)(MORECORE (-extra)); 3028 3029 if (new_brk == (char*)(MORECORE_FAILURE)) /* sbrk failed? */ 3030 { 3031 /* Try to figure out what we have */ 3032 current_brk = (char*)(MORECORE (0)); 3033 top_size = current_brk - (char*)top; 3034 if (top_size >= (long)MINSIZE) /* if not, we are very very dead! */ 3035 { 3036 sbrked_mem = current_brk - sbrk_base; 3037 set_head(top, top_size | PREV_INUSE); 3038 } 3039 check_chunk(top); 3040 return 0; 3041 } 3042 3043 else 3044 { 3045 /* Success. Adjust top accordingly. */ 3046 set_head(top, (top_size - extra) | PREV_INUSE); 3047 sbrked_mem -= extra; 3048 check_chunk(top); 3049 return 1; 3050 } 3051 } 3052 } 3053 } 3054 3055 3056 3057 /* 3058 malloc_usable_size: 3059 3060 This routine tells you how many bytes you can actually use in an 3061 allocated chunk, which may be more than you requested (although 3062 often not). You can use this many bytes without worrying about 3063 overwriting other allocated objects. Not a particularly great 3064 programming practice, but still sometimes useful. 3065 3066 */ 3067 3068 #if __STD_C 3069 size_t malloc_usable_size(Void_t* mem) 3070 #else 3071 size_t malloc_usable_size(mem) Void_t* mem; 3072 #endif 3073 { 3074 mchunkptr p; 3075 if (mem == 0) 3076 return 0; 3077 else 3078 { 3079 p = mem2chunk(mem); 3080 if(!chunk_is_mmapped(p)) 3081 { 3082 if (!inuse(p)) return 0; 3083 check_inuse_chunk(p); 3084 return chunksize(p) - SIZE_SZ; 3085 } 3086 return chunksize(p) - 2*SIZE_SZ; 3087 } 3088 } 3089 3090 3091 3092 3093 /* Utility to update current_mallinfo for malloc_stats and mallinfo() */ 3094 3095 #ifdef DEBUG 3096 static void malloc_update_mallinfo() 3097 { 3098 int i; 3099 mbinptr b; 3100 mchunkptr p; 3101 #ifdef DEBUG 3102 mchunkptr q; 3103 #endif 3104 3105 INTERNAL_SIZE_T avail = chunksize(top); 3106 int navail = ((long)(avail) >= (long)MINSIZE)? 1 : 0; 3107 3108 for (i = 1; i < NAV; ++i) 3109 { 3110 b = bin_at(i); 3111 for (p = last(b); p != b; p = p->bk) 3112 { 3113 #ifdef DEBUG 3114 check_free_chunk(p); 3115 for (q = next_chunk(p); 3116 q < top && inuse(q) && (long)(chunksize(q)) >= (long)MINSIZE; 3117 q = next_chunk(q)) 3118 check_inuse_chunk(q); 3119 #endif 3120 avail += chunksize(p); 3121 navail++; 3122 } 3123 } 3124 3125 current_mallinfo.ordblks = navail; 3126 current_mallinfo.uordblks = sbrked_mem - avail; 3127 current_mallinfo.fordblks = avail; 3128 current_mallinfo.hblks = n_mmaps; 3129 current_mallinfo.hblkhd = mmapped_mem; 3130 current_mallinfo.keepcost = chunksize(top); 3131 3132 } 3133 #endif /* DEBUG */ 3134 3135 3136 3137 /* 3138 3139 malloc_stats: 3140 3141 Prints on the amount of space obtain from the system (both 3142 via sbrk and mmap), the maximum amount (which may be more than 3143 current if malloc_trim and/or munmap got called), the maximum 3144 number of simultaneous mmap regions used, and the current number 3145 of bytes allocated via malloc (or realloc, etc) but not yet 3146 freed. (Note that this is the number of bytes allocated, not the 3147 number requested. It will be larger than the number requested 3148 because of alignment and bookkeeping overhead.) 3149 3150 */ 3151 3152 #ifdef DEBUG 3153 void malloc_stats() 3154 { 3155 malloc_update_mallinfo(); 3156 printf("max system bytes = %10u\n", 3157 (unsigned int)(max_total_mem)); 3158 printf("system bytes = %10u\n", 3159 (unsigned int)(sbrked_mem + mmapped_mem)); 3160 printf("in use bytes = %10u\n", 3161 (unsigned int)(current_mallinfo.uordblks + mmapped_mem)); 3162 #if HAVE_MMAP 3163 printf("max mmap regions = %10u\n", 3164 (unsigned int)max_n_mmaps); 3165 #endif 3166 } 3167 #endif /* DEBUG */ 3168 3169 /* 3170 mallinfo returns a copy of updated current mallinfo. 3171 */ 3172 3173 #ifdef DEBUG 3174 struct mallinfo mALLINFo() 3175 { 3176 malloc_update_mallinfo(); 3177 return current_mallinfo; 3178 } 3179 #endif /* DEBUG */ 3180 3181 3182 3183 3184 /* 3185 mallopt: 3186 3187 mallopt is the general SVID/XPG interface to tunable parameters. 3188 The format is to provide a (parameter-number, parameter-value) pair. 3189 mallopt then sets the corresponding parameter to the argument 3190 value if it can (i.e., so long as the value is meaningful), 3191 and returns 1 if successful else 0. 3192 3193 See descriptions of tunable parameters above. 3194 3195 */ 3196 3197 #if __STD_C 3198 int mALLOPt(int param_number, int value) 3199 #else 3200 int mALLOPt(param_number, value) int param_number; int value; 3201 #endif 3202 { 3203 switch(param_number) 3204 { 3205 case M_TRIM_THRESHOLD: 3206 trim_threshold = value; return 1; 3207 case M_TOP_PAD: 3208 top_pad = value; return 1; 3209 case M_MMAP_THRESHOLD: 3210 mmap_threshold = value; return 1; 3211 case M_MMAP_MAX: 3212 #if HAVE_MMAP 3213 n_mmaps_max = value; return 1; 3214 #else 3215 if (value != 0) return 0; else n_mmaps_max = value; return 1; 3216 #endif 3217 3218 default: 3219 return 0; 3220 } 3221 } 3222 3223 /* 3224 3225 History: 3226 3227 V2.6.6 Sun Dec 5 07:42:19 1999 Doug Lea (dl at gee) 3228 * return null for negative arguments 3229 * Added Several WIN32 cleanups from Martin C. Fong <mcfong@yahoo.com> 3230 * Add 'LACKS_SYS_PARAM_H' for those systems without 'sys/param.h' 3231 (e.g. WIN32 platforms) 3232 * Cleanup up header file inclusion for WIN32 platforms 3233 * Cleanup code to avoid Microsoft Visual C++ compiler complaints 3234 * Add 'USE_DL_PREFIX' to quickly allow co-existence with existing 3235 memory allocation routines 3236 * Set 'malloc_getpagesize' for WIN32 platforms (needs more work) 3237 * Use 'assert' rather than 'ASSERT' in WIN32 code to conform to 3238 usage of 'assert' in non-WIN32 code 3239 * Improve WIN32 'sbrk()' emulation's 'findRegion()' routine to 3240 avoid infinite loop 3241 * Always call 'fREe()' rather than 'free()' 3242 3243 V2.6.5 Wed Jun 17 15:57:31 1998 Doug Lea (dl at gee) 3244 * Fixed ordering problem with boundary-stamping 3245 3246 V2.6.3 Sun May 19 08:17:58 1996 Doug Lea (dl at gee) 3247 * Added pvalloc, as recommended by H.J. Liu 3248 * Added 64bit pointer support mainly from Wolfram Gloger 3249 * Added anonymously donated WIN32 sbrk emulation 3250 * Malloc, calloc, getpagesize: add optimizations from Raymond Nijssen 3251 * malloc_extend_top: fix mask error that caused wastage after 3252 foreign sbrks 3253 * Add linux mremap support code from HJ Liu 3254 3255 V2.6.2 Tue Dec 5 06:52:55 1995 Doug Lea (dl at gee) 3256 * Integrated most documentation with the code. 3257 * Add support for mmap, with help from 3258 Wolfram Gloger (Gloger@lrz.uni-muenchen.de). 3259 * Use last_remainder in more cases. 3260 * Pack bins using idea from colin@nyx10.cs.du.edu 3261 * Use ordered bins instead of best-fit threshhold 3262 * Eliminate block-local decls to simplify tracing and debugging. 3263 * Support another case of realloc via move into top 3264 * Fix error occuring when initial sbrk_base not word-aligned. 3265 * Rely on page size for units instead of SBRK_UNIT to 3266 avoid surprises about sbrk alignment conventions. 3267 * Add mallinfo, mallopt. Thanks to Raymond Nijssen 3268 (raymond@es.ele.tue.nl) for the suggestion. 3269 * Add `pad' argument to malloc_trim and top_pad mallopt parameter. 3270 * More precautions for cases where other routines call sbrk, 3271 courtesy of Wolfram Gloger (Gloger@lrz.uni-muenchen.de). 3272 * Added macros etc., allowing use in linux libc from 3273 H.J. Lu (hjl@gnu.ai.mit.edu) 3274 * Inverted this history list 3275 3276 V2.6.1 Sat Dec 2 14:10:57 1995 Doug Lea (dl at gee) 3277 * Re-tuned and fixed to behave more nicely with V2.6.0 changes. 3278 * Removed all preallocation code since under current scheme 3279 the work required to undo bad preallocations exceeds 3280 the work saved in good cases for most test programs. 3281 * No longer use return list or unconsolidated bins since 3282 no scheme using them consistently outperforms those that don't 3283 given above changes. 3284 * Use best fit for very large chunks to prevent some worst-cases. 3285 * Added some support for debugging 3286 3287 V2.6.0 Sat Nov 4 07:05:23 1995 Doug Lea (dl at gee) 3288 * Removed footers when chunks are in use. Thanks to 3289 Paul Wilson (wilson@cs.texas.edu) for the suggestion. 3290 3291 V2.5.4 Wed Nov 1 07:54:51 1995 Doug Lea (dl at gee) 3292 * Added malloc_trim, with help from Wolfram Gloger 3293 (wmglo@Dent.MED.Uni-Muenchen.DE). 3294 3295 V2.5.3 Tue Apr 26 10:16:01 1994 Doug Lea (dl at g) 3296 3297 V2.5.2 Tue Apr 5 16:20:40 1994 Doug Lea (dl at g) 3298 * realloc: try to expand in both directions 3299 * malloc: swap order of clean-bin strategy; 3300 * realloc: only conditionally expand backwards 3301 * Try not to scavenge used bins 3302 * Use bin counts as a guide to preallocation 3303 * Occasionally bin return list chunks in first scan 3304 * Add a few optimizations from colin@nyx10.cs.du.edu 3305 3306 V2.5.1 Sat Aug 14 15:40:43 1993 Doug Lea (dl at g) 3307 * faster bin computation & slightly different binning 3308 * merged all consolidations to one part of malloc proper 3309 (eliminating old malloc_find_space & malloc_clean_bin) 3310 * Scan 2 returns chunks (not just 1) 3311 * Propagate failure in realloc if malloc returns 0 3312 * Add stuff to allow compilation on non-ANSI compilers 3313 from kpv@research.att.com 3314 3315 V2.5 Sat Aug 7 07:41:59 1993 Doug Lea (dl at g.oswego.edu) 3316 * removed potential for odd address access in prev_chunk 3317 * removed dependency on getpagesize.h 3318 * misc cosmetics and a bit more internal documentation 3319 * anticosmetics: mangled names in macros to evade debugger strangeness 3320 * tested on sparc, hp-700, dec-mips, rs6000 3321 with gcc & native cc (hp, dec only) allowing 3322 Detlefs & Zorn comparison study (in SIGPLAN Notices.) 3323 3324 Trial version Fri Aug 28 13:14:29 1992 Doug Lea (dl at g.oswego.edu) 3325 * Based loosely on libg++-1.2X malloc. (It retains some of the overall 3326 structure of old version, but most details differ.) 3327 3328 */ 3329