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