]>
Commit | Line | Data |
---|---|---|
99a20616 UD |
1 | @node Memory, Character Handling, Error Reporting, Top |
2 | @chapter Virtual Memory Allocation And Paging | |
3 | @c %MENU% Allocating virtual memory and controlling paging | |
28f540f4 RM |
4 | @cindex memory allocation |
5 | @cindex storage allocation | |
6 | ||
99a20616 | 7 | This chapter describes how processes manage and use memory in a system |
1f77f049 | 8 | that uses @theglibc{}. |
99a20616 | 9 | |
1f77f049 | 10 | @Theglibc{} has several functions for dynamically allocating |
99a20616 UD |
11 | virtual memory in various ways. They vary in generality and in |
12 | efficiency. The library also provides functions for controlling paging | |
13 | and allocation of real memory. | |
14 | ||
15 | ||
16 | @menu | |
17 | * Memory Concepts:: An introduction to concepts and terminology. | |
18 | * Memory Allocation:: Allocating storage for your program data | |
99a20616 | 19 | * Resizing the Data Segment:: @code{brk}, @code{sbrk} |
4c23fed5 | 20 | * Locking Pages:: Preventing page faults |
99a20616 UD |
21 | @end menu |
22 | ||
23 | Memory mapped I/O is not discussed in this chapter. @xref{Memory-mapped I/O}. | |
24 | ||
25 | ||
26 | ||
27 | @node Memory Concepts | |
28 | @section Process Memory Concepts | |
29 | ||
30 | One of the most basic resources a process has available to it is memory. | |
31 | There are a lot of different ways systems organize memory, but in a | |
32 | typical one, each process has one linear virtual address space, with | |
33 | addresses running from zero to some huge maximum. It need not be | |
11bf311e | 34 | contiguous; i.e., not all of these addresses actually can be used to |
99a20616 UD |
35 | store data. |
36 | ||
37 | The virtual memory is divided into pages (4 kilobytes is typical). | |
38 | Backing each page of virtual memory is a page of real memory (called a | |
39 | @dfn{frame}) or some secondary storage, usually disk space. The disk | |
40 | space might be swap space or just some ordinary disk file. Actually, a | |
41 | page of all zeroes sometimes has nothing at all backing it -- there's | |
42 | just a flag saying it is all zeroes. | |
43 | @cindex page frame | |
44 | @cindex frame, real memory | |
45 | @cindex swap space | |
46 | @cindex page, virtual memory | |
47 | ||
48 | The same frame of real memory or backing store can back multiple virtual | |
49 | pages belonging to multiple processes. This is normally the case, for | |
1f77f049 | 50 | example, with virtual memory occupied by @glibcadj{} code. The same |
99a20616 UD |
51 | real memory frame containing the @code{printf} function backs a virtual |
52 | memory page in each of the existing processes that has a @code{printf} | |
53 | call in its program. | |
54 | ||
55 | In order for a program to access any part of a virtual page, the page | |
56 | must at that moment be backed by (``connected to'') a real frame. But | |
57 | because there is usually a lot more virtual memory than real memory, the | |
58 | pages must move back and forth between real memory and backing store | |
59 | regularly, coming into real memory when a process needs to access them | |
60 | and then retreating to backing store when not needed anymore. This | |
61 | movement is called @dfn{paging}. | |
62 | ||
63 | When a program attempts to access a page which is not at that moment | |
64 | backed by real memory, this is known as a @dfn{page fault}. When a page | |
65 | fault occurs, the kernel suspends the process, places the page into a | |
66 | real page frame (this is called ``paging in'' or ``faulting in''), then | |
67 | resumes the process so that from the process' point of view, the page | |
68 | was in real memory all along. In fact, to the process, all pages always | |
69 | seem to be in real memory. Except for one thing: the elapsed execution | |
70 | time of an instruction that would normally be a few nanoseconds is | |
71 | suddenly much, much, longer (because the kernel normally has to do I/O | |
72 | to complete the page-in). For programs sensitive to that, the functions | |
73 | described in @ref{Locking Pages} can control it. | |
74 | @cindex page fault | |
75 | @cindex paging | |
76 | ||
77 | Within each virtual address space, a process has to keep track of what | |
78 | is at which addresses, and that process is called memory allocation. | |
79 | Allocation usually brings to mind meting out scarce resources, but in | |
80 | the case of virtual memory, that's not a major goal, because there is | |
81 | generally much more of it than anyone needs. Memory allocation within a | |
68979757 | 82 | process is mainly just a matter of making sure that the same byte of |
99a20616 UD |
83 | memory isn't used to store two different things. |
84 | ||
85 | Processes allocate memory in two major ways: by exec and | |
86 | programmatically. Actually, forking is a third way, but it's not very | |
87 | interesting. @xref{Creating a Process}. | |
88 | ||
89 | Exec is the operation of creating a virtual address space for a process, | |
90 | loading its basic program into it, and executing the program. It is | |
91 | done by the ``exec'' family of functions (e.g. @code{execl}). The | |
92 | operation takes a program file (an executable), it allocates space to | |
93 | load all the data in the executable, loads it, and transfers control to | |
94 | it. That data is most notably the instructions of the program (the | |
95 | @dfn{text}), but also literals and constants in the program and even | |
96 | some variables: C variables with the static storage class (@pxref{Memory | |
97 | Allocation and C}). | |
98 | @cindex executable | |
99 | @cindex literals | |
100 | @cindex constants | |
101 | ||
102 | Once that program begins to execute, it uses programmatic allocation to | |
1f77f049 | 103 | gain additional memory. In a C program with @theglibc{}, there |
99a20616 UD |
104 | are two kinds of programmatic allocation: automatic and dynamic. |
105 | @xref{Memory Allocation and C}. | |
106 | ||
107 | Memory-mapped I/O is another form of dynamic virtual memory allocation. | |
108 | Mapping memory to a file means declaring that the contents of certain | |
109 | range of a process' addresses shall be identical to the contents of a | |
110 | specified regular file. The system makes the virtual memory initially | |
111 | contain the contents of the file, and if you modify the memory, the | |
112 | system writes the same modification to the file. Note that due to the | |
113 | magic of virtual memory and page faults, there is no reason for the | |
114 | system to do I/O to read the file, or allocate real memory for its | |
115 | contents, until the program accesses the virtual memory. | |
116 | @xref{Memory-mapped I/O}. | |
117 | @cindex memory mapped I/O | |
118 | @cindex memory mapped file | |
119 | @cindex files, accessing | |
120 | ||
121 | Just as it programmatically allocates memory, the program can | |
122 | programmatically deallocate (@dfn{free}) it. You can't free the memory | |
123 | that was allocated by exec. When the program exits or execs, you might | |
124 | say that all its memory gets freed, but since in both cases the address | |
125 | space ceases to exist, the point is really moot. @xref{Program | |
126 | Termination}. | |
127 | @cindex execing a program | |
128 | @cindex freeing memory | |
129 | @cindex exiting a program | |
130 | ||
131 | A process' virtual address space is divided into segments. A segment is | |
132 | a contiguous range of virtual addresses. Three important segments are: | |
28f540f4 | 133 | |
28f540f4 | 134 | @itemize @bullet |
28f540f4 | 135 | |
68979757 | 136 | @item |
99a20616 UD |
137 | |
138 | The @dfn{text segment} contains a program's instructions and literals and | |
139 | static constants. It is allocated by exec and stays the same size for | |
68979757 | 140 | the life of the virtual address space. |
28f540f4 RM |
141 | |
142 | @item | |
99a20616 UD |
143 | The @dfn{data segment} is working storage for the program. It can be |
144 | preallocated and preloaded by exec and the process can extend or shrink | |
145 | it by calling functions as described in @xref{Resizing the Data | |
146 | Segment}. Its lower end is fixed. | |
147 | ||
68979757 | 148 | @item |
99a20616 UD |
149 | The @dfn{stack segment} contains a program stack. It grows as the stack |
150 | grows, but doesn't shrink when the stack shrinks. | |
151 | ||
28f540f4 | 152 | @end itemize |
99a20616 UD |
153 | |
154 | ||
155 | ||
156 | @node Memory Allocation | |
68979757 | 157 | @section Allocating Storage For Program Data |
99a20616 UD |
158 | |
159 | This section covers how ordinary programs manage storage for their data, | |
160 | including the famous @code{malloc} function and some fancier facilities | |
3ef569c7 | 161 | special to @theglibc{} and GNU Compiler. |
28f540f4 RM |
162 | |
163 | @menu | |
99a20616 | 164 | * Memory Allocation and C:: How to get different kinds of allocation in C. |
28f540f4 RM |
165 | * Unconstrained Allocation:: The @code{malloc} facility allows fully general |
166 | dynamic allocation. | |
bd355af0 | 167 | * Allocation Debugging:: Finding memory leaks and not freed memory. |
28f540f4 RM |
168 | * Obstacks:: Obstacks are less general than malloc |
169 | but more efficient and convenient. | |
170 | * Variable Size Automatic:: Allocation of variable-sized blocks | |
171 | of automatic storage that are freed when the | |
172 | calling function returns. | |
28f540f4 RM |
173 | @end menu |
174 | ||
28f540f4 | 175 | |
99a20616 UD |
176 | @node Memory Allocation and C |
177 | @subsection Memory Allocation in C Programs | |
28f540f4 | 178 | |
99a20616 UD |
179 | The C language supports two kinds of memory allocation through the |
180 | variables in C programs: | |
28f540f4 RM |
181 | |
182 | @itemize @bullet | |
183 | @item | |
184 | @dfn{Static allocation} is what happens when you declare a static or | |
185 | global variable. Each static or global variable defines one block of | |
186 | space, of a fixed size. The space is allocated once, when your program | |
99a20616 UD |
187 | is started (part of the exec operation), and is never freed. |
188 | @cindex static memory allocation | |
189 | @cindex static storage class | |
28f540f4 RM |
190 | |
191 | @item | |
192 | @dfn{Automatic allocation} happens when you declare an automatic | |
193 | variable, such as a function argument or a local variable. The space | |
194 | for an automatic variable is allocated when the compound statement | |
195 | containing the declaration is entered, and is freed when that | |
196 | compound statement is exited. | |
99a20616 UD |
197 | @cindex automatic memory allocation |
198 | @cindex automatic storage class | |
28f540f4 | 199 | |
99a20616 | 200 | In GNU C, the size of the automatic storage can be an expression |
28f540f4 RM |
201 | that varies. In other C implementations, it must be a constant. |
202 | @end itemize | |
203 | ||
99a20616 | 204 | A third important kind of memory allocation, @dfn{dynamic allocation}, |
1f77f049 | 205 | is not supported by C variables but is available via @glibcadj{} |
99a20616 UD |
206 | functions. |
207 | @cindex dynamic memory allocation | |
208 | ||
209 | @subsubsection Dynamic Memory Allocation | |
210 | @cindex dynamic memory allocation | |
211 | ||
212 | @dfn{Dynamic memory allocation} is a technique in which programs | |
213 | determine as they are running where to store some information. You need | |
214 | dynamic allocation when the amount of memory you need, or how long you | |
215 | continue to need it, depends on factors that are not known before the | |
216 | program runs. | |
217 | ||
218 | For example, you may need a block to store a line read from an input | |
219 | file; since there is no limit to how long a line can be, you must | |
220 | allocate the memory dynamically and make it dynamically larger as you | |
221 | read more of the line. | |
222 | ||
223 | Or, you may need a block for each record or each definition in the input | |
224 | data; since you can't know in advance how many there will be, you must | |
225 | allocate a new block for each record or definition as you read it. | |
226 | ||
227 | When you use dynamic allocation, the allocation of a block of memory is | |
228 | an action that the program requests explicitly. You call a function or | |
229 | macro when you want to allocate space, and specify the size with an | |
230 | argument. If you want to free the space, you do so by calling another | |
231 | function or macro. You can do these things whenever you want, as often | |
232 | as you want. | |
233 | ||
28f540f4 RM |
234 | Dynamic allocation is not supported by C variables; there is no storage |
235 | class ``dynamic'', and there can never be a C variable whose value is | |
99a20616 | 236 | stored in dynamically allocated space. The only way to get dynamically |
1f77f049 JM |
237 | allocated memory is via a system call (which is generally via a @glibcadj{} |
238 | function call), and the only way to refer to dynamically | |
99a20616 UD |
239 | allocated space is through a pointer. Because it is less convenient, |
240 | and because the actual process of dynamic allocation requires more | |
241 | computation time, programmers generally use dynamic allocation only when | |
242 | neither static nor automatic allocation will serve. | |
28f540f4 RM |
243 | |
244 | For example, if you want to allocate dynamically some space to hold a | |
245 | @code{struct foobar}, you cannot declare a variable of type @code{struct | |
246 | foobar} whose contents are the dynamically allocated space. But you can | |
247 | declare a variable of pointer type @code{struct foobar *} and assign it the | |
248 | address of the space. Then you can use the operators @samp{*} and | |
249 | @samp{->} on this pointer variable to refer to the contents of the space: | |
250 | ||
251 | @smallexample | |
252 | @{ | |
253 | struct foobar *ptr | |
254 | = (struct foobar *) malloc (sizeof (struct foobar)); | |
255 | ptr->name = x; | |
256 | ptr->next = current_foobar; | |
257 | current_foobar = ptr; | |
258 | @} | |
259 | @end smallexample | |
260 | ||
261 | @node Unconstrained Allocation | |
99a20616 UD |
262 | @subsection Unconstrained Allocation |
263 | @cindex unconstrained memory allocation | |
28f540f4 RM |
264 | @cindex @code{malloc} function |
265 | @cindex heap, dynamic allocation from | |
266 | ||
267 | The most general dynamic allocation facility is @code{malloc}. It | |
268 | allows you to allocate blocks of memory of any size at any time, make | |
269 | them bigger or smaller at any time, and free the blocks individually at | |
270 | any time (or never). | |
271 | ||
272 | @menu | |
273 | * Basic Allocation:: Simple use of @code{malloc}. | |
274 | * Malloc Examples:: Examples of @code{malloc}. @code{xmalloc}. | |
275 | * Freeing after Malloc:: Use @code{free} to free a block you | |
276 | got with @code{malloc}. | |
277 | * Changing Block Size:: Use @code{realloc} to make a block | |
278 | bigger or smaller. | |
279 | * Allocating Cleared Space:: Use @code{calloc} to allocate a | |
280 | block and clear it. | |
281 | * Efficiency and Malloc:: Efficiency considerations in use of | |
282 | these functions. | |
68979757 | 283 | * Aligned Memory Blocks:: Allocating specially aligned memory. |
c131718c UD |
284 | * Malloc Tunable Parameters:: Use @code{mallopt} to adjust allocation |
285 | parameters. | |
28f540f4 RM |
286 | * Heap Consistency Checking:: Automatic checking for errors. |
287 | * Hooks for Malloc:: You can use these hooks for debugging | |
288 | programs that use @code{malloc}. | |
289 | * Statistics of Malloc:: Getting information about how much | |
290 | memory your program is using. | |
291 | * Summary of Malloc:: Summary of @code{malloc} and related functions. | |
292 | @end menu | |
293 | ||
294 | @node Basic Allocation | |
99a20616 | 295 | @subsubsection Basic Memory Allocation |
28f540f4 RM |
296 | @cindex allocation of memory with @code{malloc} |
297 | ||
298 | To allocate a block of memory, call @code{malloc}. The prototype for | |
299 | this function is in @file{stdlib.h}. | |
300 | @pindex stdlib.h | |
301 | ||
302 | @comment malloc.h stdlib.h | |
f65fd747 | 303 | @comment ISO |
28f540f4 | 304 | @deftypefun {void *} malloc (size_t @var{size}) |
9f529d7c AO |
305 | @safety{@prelim{}@mtsafe{}@asunsafe{@asulock{}}@acunsafe{@aculock{} @acsfd{} @acsmem{}}} |
306 | @c Malloc hooks and __morecore pointers, as well as such parameters as | |
307 | @c max_n_mmaps and max_mmapped_mem, are accessed without guards, so they | |
308 | @c could pose a thread safety issue; in order to not declare malloc | |
309 | @c MT-unsafe, it's modifying the hooks and parameters while multiple | |
310 | @c threads are active that is regarded as unsafe. An arena's next field | |
311 | @c is initialized and never changed again, except for main_arena's, | |
312 | @c that's protected by list_lock; next_free is only modified while | |
313 | @c list_lock is held too. All other data members of an arena, as well | |
314 | @c as the metadata of the memory areas assigned to it, are only modified | |
315 | @c while holding the arena's mutex (fastbin pointers use catomic ops | |
316 | @c because they may be modified by free without taking the arena's | |
317 | @c lock). Some reassurance was needed for fastbins, for it wasn't clear | |
318 | @c how they were initialized. It turns out they are always | |
319 | @c zero-initialized: main_arena's, for being static data, and other | |
320 | @c arena's, for being just-mmapped memory. | |
321 | ||
322 | @c Leaking file descriptors and memory in case of cancellation is | |
323 | @c unavoidable without disabling cancellation, but the lock situation is | |
324 | @c a bit more complicated: we don't have fallback arenas for malloc to | |
325 | @c be safe to call from within signal handlers. Error-checking mutexes | |
326 | @c or trylock could enable us to try and use alternate arenas, even with | |
327 | @c -DPER_THREAD (enabled by default), but supporting interruption | |
328 | @c (cancellation or signal handling) while holding the arena list mutex | |
329 | @c would require more work; maybe blocking signals and disabling async | |
330 | @c cancellation while manipulating the arena lists? | |
331 | ||
332 | @c __libc_malloc @asulock @aculock @acsfd @acsmem | |
333 | @c force_reg ok | |
334 | @c *malloc_hook unguarded | |
9f529d7c AO |
335 | @c arena_lock @asulock @aculock @acsfd @acsmem |
336 | @c mutex_lock @asulock @aculock | |
337 | @c arena_get2 @asulock @aculock @acsfd @acsmem | |
338 | @c get_free_list @asulock @aculock | |
339 | @c mutex_lock (list_lock) dup @asulock @aculock | |
340 | @c mutex_unlock (list_lock) dup @aculock | |
341 | @c mutex_lock (arena lock) dup @asulock @aculock [returns locked] | |
9f529d7c AO |
342 | @c __get_nprocs ext ok @acsfd |
343 | @c NARENAS_FROM_NCORES ok | |
344 | @c catomic_compare_and_exchange_bool_acq ok | |
345 | @c _int_new_arena ok @asulock @aculock @acsmem | |
346 | @c new_heap ok @acsmem | |
347 | @c mmap ok @acsmem | |
348 | @c munmap ok @acsmem | |
349 | @c mprotect ok | |
350 | @c chunk2mem ok | |
351 | @c set_head ok | |
352 | @c tsd_setspecific dup ok | |
353 | @c mutex_init ok | |
354 | @c mutex_lock (just-created mutex) ok, returns locked | |
355 | @c mutex_lock (list_lock) dup @asulock @aculock | |
356 | @c atomic_write_barrier ok | |
357 | @c mutex_unlock (list_lock) @aculock | |
358 | @c catomic_decrement ok | |
359 | @c reused_arena @asulock @aculock | |
360 | @c reads&writes next_to_use and iterates over arena next without guards | |
361 | @c those are harmless as long as we don't drop arenas from the | |
362 | @c NEXT list, and we never do; when a thread terminates, | |
363 | @c arena_thread_freeres prepends the arena to the free_list | |
364 | @c NEXT_FREE list, but NEXT is never modified, so it's safe! | |
365 | @c mutex_trylock (arena lock) @asulock @aculock | |
366 | @c mutex_lock (arena lock) dup @asulock @aculock | |
367 | @c tsd_setspecific dup ok | |
368 | @c _int_malloc @acsfd @acsmem | |
369 | @c checked_request2size ok | |
370 | @c REQUEST_OUT_OF_RANGE ok | |
371 | @c request2size ok | |
372 | @c get_max_fast ok | |
373 | @c fastbin_index ok | |
374 | @c fastbin ok | |
375 | @c catomic_compare_and_exhange_val_acq ok | |
376 | @c malloc_printerr dup @mtsenv | |
377 | @c if we get to it, we're toast already, undefined behavior must have | |
378 | @c been invoked before | |
379 | @c libc_message @mtsenv [no leaks with cancellation disabled] | |
380 | @c FATAL_PREPARE ok | |
381 | @c pthread_setcancelstate disable ok | |
382 | @c libc_secure_getenv @mtsenv | |
383 | @c getenv @mtsenv | |
384 | @c open_not_cancel_2 dup @acsfd | |
385 | @c strchrnul ok | |
386 | @c WRITEV_FOR_FATAL ok | |
387 | @c writev ok | |
388 | @c mmap ok @acsmem | |
389 | @c munmap ok @acsmem | |
390 | @c BEFORE_ABORT @acsfd | |
391 | @c backtrace ok | |
392 | @c write_not_cancel dup ok | |
393 | @c backtrace_symbols_fd @aculock | |
394 | @c open_not_cancel_2 dup @acsfd | |
395 | @c read_not_cancel dup ok | |
396 | @c close_not_cancel_no_status dup @acsfd | |
397 | @c abort ok | |
398 | @c itoa_word ok | |
399 | @c abort ok | |
400 | @c check_remalloced_chunk ok/disabled | |
401 | @c chunk2mem dup ok | |
402 | @c alloc_perturb ok | |
403 | @c in_smallbin_range ok | |
404 | @c smallbin_index ok | |
405 | @c bin_at ok | |
406 | @c last ok | |
407 | @c malloc_consolidate ok | |
408 | @c get_max_fast dup ok | |
409 | @c clear_fastchunks ok | |
410 | @c unsorted_chunks dup ok | |
411 | @c fastbin dup ok | |
412 | @c atomic_exchange_acq ok | |
413 | @c check_inuse_chunk dup ok/disabled | |
414 | @c chunk_at_offset dup ok | |
415 | @c chunksize dup ok | |
416 | @c inuse_bit_at_offset dup ok | |
417 | @c unlink dup ok | |
418 | @c clear_inuse_bit_at_offset dup ok | |
419 | @c in_smallbin_range dup ok | |
420 | @c set_head dup ok | |
421 | @c malloc_init_state ok | |
422 | @c bin_at dup ok | |
423 | @c set_noncontiguous dup ok | |
424 | @c set_max_fast dup ok | |
425 | @c initial_top ok | |
426 | @c unsorted_chunks dup ok | |
427 | @c check_malloc_state ok/disabled | |
428 | @c set_inuse_bit_at_offset ok | |
429 | @c check_malloced_chunk ok/disabled | |
430 | @c largebin_index ok | |
431 | @c have_fastchunks ok | |
432 | @c unsorted_chunks ok | |
433 | @c bin_at ok | |
434 | @c chunksize ok | |
435 | @c chunk_at_offset ok | |
436 | @c set_head ok | |
437 | @c set_foot ok | |
438 | @c mark_bin ok | |
439 | @c idx2bit ok | |
440 | @c first ok | |
441 | @c unlink ok | |
442 | @c malloc_printerr dup ok | |
443 | @c in_smallbin_range dup ok | |
444 | @c idx2block ok | |
445 | @c idx2bit dup ok | |
446 | @c next_bin ok | |
447 | @c sysmalloc @acsfd @acsmem | |
448 | @c MMAP @acsmem | |
449 | @c set_head dup ok | |
450 | @c check_chunk ok/disabled | |
451 | @c chunk2mem dup ok | |
452 | @c chunksize dup ok | |
453 | @c chunk_at_offset dup ok | |
454 | @c heap_for_ptr ok | |
455 | @c grow_heap ok | |
456 | @c mprotect ok | |
457 | @c set_head dup ok | |
458 | @c new_heap @acsmem | |
459 | @c MMAP dup @acsmem | |
460 | @c munmap @acsmem | |
461 | @c top ok | |
462 | @c set_foot dup ok | |
463 | @c contiguous ok | |
464 | @c MORECORE ok | |
465 | @c *__morecore ok unguarded | |
466 | @c __default_morecore | |
467 | @c sbrk ok | |
468 | @c force_reg dup ok | |
469 | @c *__after_morecore_hook unguarded | |
470 | @c set_noncontiguous ok | |
471 | @c malloc_printerr dup ok | |
472 | @c _int_free (have_lock) @acsfd @acsmem [@asulock @aculock] | |
473 | @c chunksize dup ok | |
474 | @c mutex_unlock dup @aculock/!have_lock | |
475 | @c malloc_printerr dup ok | |
476 | @c check_inuse_chunk ok/disabled | |
477 | @c chunk_at_offset dup ok | |
478 | @c mutex_lock dup @asulock @aculock/@have_lock | |
479 | @c chunk2mem dup ok | |
480 | @c free_perturb ok | |
481 | @c set_fastchunks ok | |
482 | @c catomic_and ok | |
483 | @c fastbin_index dup ok | |
484 | @c fastbin dup ok | |
485 | @c catomic_compare_and_exchange_val_rel ok | |
486 | @c chunk_is_mmapped ok | |
487 | @c contiguous dup ok | |
488 | @c prev_inuse ok | |
489 | @c unlink dup ok | |
490 | @c inuse_bit_at_offset dup ok | |
491 | @c clear_inuse_bit_at_offset ok | |
492 | @c unsorted_chunks dup ok | |
493 | @c in_smallbin_range dup ok | |
494 | @c set_head dup ok | |
495 | @c set_foot dup ok | |
496 | @c check_free_chunk ok/disabled | |
497 | @c check_chunk dup ok/disabled | |
498 | @c have_fastchunks dup ok | |
499 | @c malloc_consolidate dup ok | |
500 | @c systrim ok | |
501 | @c MORECORE dup ok | |
502 | @c *__after_morecore_hook dup unguarded | |
503 | @c set_head dup ok | |
504 | @c check_malloc_state ok/disabled | |
505 | @c top dup ok | |
506 | @c heap_for_ptr dup ok | |
507 | @c heap_trim @acsfd @acsmem | |
508 | @c top dup ok | |
509 | @c chunk_at_offset dup ok | |
510 | @c prev_chunk ok | |
511 | @c chunksize dup ok | |
512 | @c prev_inuse dup ok | |
513 | @c delete_heap @acsmem | |
514 | @c munmap dup @acsmem | |
515 | @c unlink dup ok | |
516 | @c set_head dup ok | |
517 | @c shrink_heap @acsfd | |
518 | @c check_may_shrink_heap @acsfd | |
519 | @c open_not_cancel_2 @acsfd | |
520 | @c read_not_cancel ok | |
521 | @c close_not_cancel_no_status @acsfd | |
522 | @c MMAP dup ok | |
523 | @c madvise ok | |
524 | @c munmap_chunk @acsmem | |
525 | @c chunksize dup ok | |
526 | @c chunk_is_mmapped dup ok | |
527 | @c chunk2mem dup ok | |
528 | @c malloc_printerr dup ok | |
529 | @c munmap dup @acsmem | |
530 | @c check_malloc_state ok/disabled | |
531 | @c arena_get_retry @asulock @aculock @acsfd @acsmem | |
532 | @c mutex_unlock dup @aculock | |
533 | @c mutex_lock dup @asulock @aculock | |
534 | @c arena_get2 dup @asulock @aculock @acsfd @acsmem | |
535 | @c mutex_unlock @aculock | |
536 | @c mem2chunk ok | |
537 | @c chunk_is_mmapped ok | |
538 | @c arena_for_chunk ok | |
539 | @c chunk_non_main_arena ok | |
540 | @c heap_for_ptr ok | |
28f540f4 RM |
541 | This function returns a pointer to a newly allocated block @var{size} |
542 | bytes long, or a null pointer if the block could not be allocated. | |
543 | @end deftypefun | |
544 | ||
545 | The contents of the block are undefined; you must initialize it yourself | |
546 | (or use @code{calloc} instead; @pxref{Allocating Cleared Space}). | |
547 | Normally you would cast the value as a pointer to the kind of object | |
548 | that you want to store in the block. Here we show an example of doing | |
549 | so, and of initializing the space with zeros using the library function | |
0a13c9e9 | 550 | @code{memset} (@pxref{Copying Strings and Arrays}): |
28f540f4 RM |
551 | |
552 | @smallexample | |
553 | struct foo *ptr; | |
554 | @dots{} | |
555 | ptr = (struct foo *) malloc (sizeof (struct foo)); | |
556 | if (ptr == 0) abort (); | |
557 | memset (ptr, 0, sizeof (struct foo)); | |
558 | @end smallexample | |
559 | ||
560 | You can store the result of @code{malloc} into any pointer variable | |
f65fd747 | 561 | without a cast, because @w{ISO C} automatically converts the type |
28f540f4 RM |
562 | @code{void *} to another type of pointer when necessary. But the cast |
563 | is necessary in contexts other than assignment operators or if you might | |
564 | want your code to run in traditional C. | |
565 | ||
566 | Remember that when allocating space for a string, the argument to | |
567 | @code{malloc} must be one plus the length of the string. This is | |
568 | because a string is terminated with a null character that doesn't count | |
569 | in the ``length'' of the string but does need space. For example: | |
570 | ||
571 | @smallexample | |
572 | char *ptr; | |
573 | @dots{} | |
574 | ptr = (char *) malloc (length + 1); | |
575 | @end smallexample | |
576 | ||
577 | @noindent | |
578 | @xref{Representation of Strings}, for more information about this. | |
579 | ||
580 | @node Malloc Examples | |
99a20616 | 581 | @subsubsection Examples of @code{malloc} |
28f540f4 RM |
582 | |
583 | If no more space is available, @code{malloc} returns a null pointer. | |
584 | You should check the value of @emph{every} call to @code{malloc}. It is | |
585 | useful to write a subroutine that calls @code{malloc} and reports an | |
586 | error if the value is a null pointer, returning only if the value is | |
587 | nonzero. This function is conventionally called @code{xmalloc}. Here | |
588 | it is: | |
589 | ||
590 | @smallexample | |
591 | void * | |
592 | xmalloc (size_t size) | |
593 | @{ | |
e256c421 | 594 | void *value = malloc (size); |
28f540f4 RM |
595 | if (value == 0) |
596 | fatal ("virtual memory exhausted"); | |
597 | return value; | |
598 | @} | |
599 | @end smallexample | |
600 | ||
601 | Here is a real example of using @code{malloc} (by way of @code{xmalloc}). | |
602 | The function @code{savestring} will copy a sequence of characters into | |
603 | a newly allocated null-terminated string: | |
604 | ||
605 | @smallexample | |
606 | @group | |
607 | char * | |
608 | savestring (const char *ptr, size_t len) | |
609 | @{ | |
e256c421 | 610 | char *value = (char *) xmalloc (len + 1); |
28f540f4 | 611 | value[len] = '\0'; |
390955cb | 612 | return (char *) memcpy (value, ptr, len); |
28f540f4 RM |
613 | @} |
614 | @end group | |
615 | @end smallexample | |
616 | ||
617 | The block that @code{malloc} gives you is guaranteed to be aligned so | |
a7a93d50 | 618 | that it can hold any type of data. On @gnusystems{}, the address is |
0a096e44 | 619 | always a multiple of eight on 32-bit systems, and a multiple of 16 on |
c131718c | 620 | 64-bit systems. Only rarely is any higher boundary (such as a page |
5764c27f WN |
621 | boundary) necessary; for those cases, use @code{aligned_alloc} or |
622 | @code{posix_memalign} (@pxref{Aligned Memory Blocks}). | |
28f540f4 RM |
623 | |
624 | Note that the memory located after the end of the block is likely to be | |
625 | in use for something else; perhaps a block already allocated by another | |
626 | call to @code{malloc}. If you attempt to treat the block as longer than | |
627 | you asked for it to be, you are liable to destroy the data that | |
628 | @code{malloc} uses to keep track of its blocks, or you may destroy the | |
629 | contents of another block. If you have already allocated a block and | |
630 | discover you want it to be bigger, use @code{realloc} (@pxref{Changing | |
631 | Block Size}). | |
632 | ||
633 | @node Freeing after Malloc | |
99a20616 | 634 | @subsubsection Freeing Memory Allocated with @code{malloc} |
28f540f4 RM |
635 | @cindex freeing memory allocated with @code{malloc} |
636 | @cindex heap, freeing memory from | |
637 | ||
638 | When you no longer need a block that you got with @code{malloc}, use the | |
639 | function @code{free} to make the block available to be allocated again. | |
640 | The prototype for this function is in @file{stdlib.h}. | |
641 | @pindex stdlib.h | |
642 | ||
643 | @comment malloc.h stdlib.h | |
f65fd747 | 644 | @comment ISO |
28f540f4 | 645 | @deftypefun void free (void *@var{ptr}) |
9f529d7c AO |
646 | @safety{@prelim{}@mtsafe{}@asunsafe{@asulock{}}@acunsafe{@aculock{} @acsfd{} @acsmem{}}} |
647 | @c __libc_free @asulock @aculock @acsfd @acsmem | |
648 | @c releasing memory into fastbins modifies the arena without taking | |
649 | @c its mutex, but catomic operations ensure safety. If two (or more) | |
650 | @c threads are running malloc and have their own arenas locked when | |
651 | @c each gets a signal whose handler free()s large (non-fastbin-able) | |
652 | @c blocks from each other's arena, we deadlock; this is a more general | |
653 | @c case of @asulock. | |
654 | @c *__free_hook unguarded | |
655 | @c mem2chunk ok | |
656 | @c chunk_is_mmapped ok, chunk bits not modified after allocation | |
657 | @c chunksize ok | |
658 | @c munmap_chunk dup @acsmem | |
659 | @c arena_for_chunk dup ok | |
660 | @c _int_free (!have_lock) dup @asulock @aculock @acsfd @acsmem | |
99a20616 | 661 | The @code{free} function deallocates the block of memory pointed at |
28f540f4 RM |
662 | by @var{ptr}. |
663 | @end deftypefun | |
664 | ||
665 | @comment stdlib.h | |
666 | @comment Sun | |
667 | @deftypefun void cfree (void *@var{ptr}) | |
9f529d7c AO |
668 | @safety{@prelim{}@mtsafe{}@asunsafe{@asulock{}}@acunsafe{@aculock{} @acsfd{} @acsmem{}}} |
669 | @c alias to free | |
28f540f4 RM |
670 | This function does the same thing as @code{free}. It's provided for |
671 | backward compatibility with SunOS; you should use @code{free} instead. | |
672 | @end deftypefun | |
673 | ||
674 | Freeing a block alters the contents of the block. @strong{Do not expect to | |
675 | find any data (such as a pointer to the next block in a chain of blocks) in | |
676 | the block after freeing it.} Copy whatever you need out of the block before | |
677 | freeing it! Here is an example of the proper way to free all the blocks in | |
678 | a chain, and the strings that they point to: | |
679 | ||
680 | @smallexample | |
681 | struct chain | |
682 | @{ | |
683 | struct chain *next; | |
684 | char *name; | |
685 | @} | |
686 | ||
687 | void | |
688 | free_chain (struct chain *chain) | |
689 | @{ | |
690 | while (chain != 0) | |
691 | @{ | |
692 | struct chain *next = chain->next; | |
693 | free (chain->name); | |
694 | free (chain); | |
695 | chain = next; | |
696 | @} | |
697 | @} | |
698 | @end smallexample | |
699 | ||
700 | Occasionally, @code{free} can actually return memory to the operating | |
701 | system and make the process smaller. Usually, all it can do is allow a | |
702 | later call to @code{malloc} to reuse the space. In the meantime, the | |
703 | space remains in your program as part of a free-list used internally by | |
704 | @code{malloc}. | |
705 | ||
706 | There is no point in freeing blocks at the end of a program, because all | |
707 | of the program's space is given back to the system when the process | |
708 | terminates. | |
709 | ||
710 | @node Changing Block Size | |
99a20616 | 711 | @subsubsection Changing the Size of a Block |
28f540f4 RM |
712 | @cindex changing the size of a block (@code{malloc}) |
713 | ||
714 | Often you do not know for certain how big a block you will ultimately need | |
715 | at the time you must begin to use the block. For example, the block might | |
716 | be a buffer that you use to hold a line being read from a file; no matter | |
717 | how long you make the buffer initially, you may encounter a line that is | |
718 | longer. | |
719 | ||
720 | You can make the block longer by calling @code{realloc}. This function | |
721 | is declared in @file{stdlib.h}. | |
722 | @pindex stdlib.h | |
723 | ||
724 | @comment malloc.h stdlib.h | |
f65fd747 | 725 | @comment ISO |
28f540f4 | 726 | @deftypefun {void *} realloc (void *@var{ptr}, size_t @var{newsize}) |
9f529d7c AO |
727 | @safety{@prelim{}@mtsafe{}@asunsafe{@asulock{}}@acunsafe{@aculock{} @acsfd{} @acsmem{}}} |
728 | @c It may call the implementations of malloc and free, so all of their | |
729 | @c issues arise, plus the realloc hook, also accessed without guards. | |
730 | ||
731 | @c __libc_realloc @asulock @aculock @acsfd @acsmem | |
732 | @c *__realloc_hook unguarded | |
733 | @c __libc_free dup @asulock @aculock @acsfd @acsmem | |
734 | @c __libc_malloc dup @asulock @aculock @acsfd @acsmem | |
735 | @c mem2chunk dup ok | |
736 | @c chunksize dup ok | |
737 | @c malloc_printerr dup ok | |
738 | @c checked_request2size dup ok | |
739 | @c chunk_is_mmapped dup ok | |
740 | @c mremap_chunk | |
741 | @c chunksize dup ok | |
742 | @c __mremap ok | |
743 | @c set_head dup ok | |
744 | @c MALLOC_COPY ok | |
745 | @c memcpy ok | |
746 | @c munmap_chunk dup @acsmem | |
747 | @c arena_for_chunk dup ok | |
748 | @c mutex_lock (arena mutex) dup @asulock @aculock | |
749 | @c _int_realloc @acsfd @acsmem | |
750 | @c malloc_printerr dup ok | |
751 | @c check_inuse_chunk dup ok/disabled | |
752 | @c chunk_at_offset dup ok | |
753 | @c chunksize dup ok | |
754 | @c set_head_size dup ok | |
755 | @c chunk_at_offset dup ok | |
756 | @c set_head dup ok | |
757 | @c chunk2mem dup ok | |
758 | @c inuse dup ok | |
759 | @c unlink dup ok | |
760 | @c _int_malloc dup @acsfd @acsmem | |
761 | @c mem2chunk dup ok | |
762 | @c MALLOC_COPY dup ok | |
763 | @c _int_free (have_lock) dup @acsfd @acsmem | |
764 | @c set_inuse_bit_at_offset dup ok | |
765 | @c set_head dup ok | |
766 | @c mutex_unlock (arena mutex) dup @aculock | |
767 | @c _int_free (!have_lock) dup @asulock @aculock @acsfd @acsmem | |
768 | ||
28f540f4 RM |
769 | The @code{realloc} function changes the size of the block whose address is |
770 | @var{ptr} to be @var{newsize}. | |
771 | ||
772 | Since the space after the end of the block may be in use, @code{realloc} | |
773 | may find it necessary to copy the block to a new address where more free | |
774 | space is available. The value of @code{realloc} is the new address of the | |
775 | block. If the block needs to be moved, @code{realloc} copies the old | |
776 | contents. | |
777 | ||
778 | If you pass a null pointer for @var{ptr}, @code{realloc} behaves just | |
779 | like @samp{malloc (@var{newsize})}. This can be convenient, but beware | |
f65fd747 | 780 | that older implementations (before @w{ISO C}) may not support this |
28f540f4 RM |
781 | behavior, and will probably crash when @code{realloc} is passed a null |
782 | pointer. | |
783 | @end deftypefun | |
784 | ||
785 | Like @code{malloc}, @code{realloc} may return a null pointer if no | |
786 | memory space is available to make the block bigger. When this happens, | |
787 | the original block is untouched; it has not been modified or relocated. | |
788 | ||
789 | In most cases it makes no difference what happens to the original block | |
790 | when @code{realloc} fails, because the application program cannot continue | |
791 | when it is out of memory, and the only thing to do is to give a fatal error | |
792 | message. Often it is convenient to write and use a subroutine, | |
793 | conventionally called @code{xrealloc}, that takes care of the error message | |
794 | as @code{xmalloc} does for @code{malloc}: | |
795 | ||
796 | @smallexample | |
797 | void * | |
798 | xrealloc (void *ptr, size_t size) | |
799 | @{ | |
e256c421 | 800 | void *value = realloc (ptr, size); |
28f540f4 RM |
801 | if (value == 0) |
802 | fatal ("Virtual memory exhausted"); | |
803 | return value; | |
804 | @} | |
805 | @end smallexample | |
806 | ||
807 | You can also use @code{realloc} to make a block smaller. The reason you | |
ed277b4e | 808 | would do this is to avoid tying up a lot of memory space when only a little |
c131718c UD |
809 | is needed. |
810 | @comment The following is no longer true with the new malloc. | |
811 | @comment But it seems wise to keep the warning for other implementations. | |
812 | In several allocation implementations, making a block smaller sometimes | |
813 | necessitates copying it, so it can fail if no other space is available. | |
28f540f4 RM |
814 | |
815 | If the new size you specify is the same as the old size, @code{realloc} | |
816 | is guaranteed to change nothing and return the same address that you gave. | |
817 | ||
818 | @node Allocating Cleared Space | |
99a20616 | 819 | @subsubsection Allocating Cleared Space |
28f540f4 RM |
820 | |
821 | The function @code{calloc} allocates memory and clears it to zero. It | |
822 | is declared in @file{stdlib.h}. | |
823 | @pindex stdlib.h | |
824 | ||
825 | @comment malloc.h stdlib.h | |
f65fd747 | 826 | @comment ISO |
28f540f4 | 827 | @deftypefun {void *} calloc (size_t @var{count}, size_t @var{eltsize}) |
9f529d7c AO |
828 | @safety{@prelim{}@mtsafe{}@asunsafe{@asulock{}}@acunsafe{@aculock{} @acsfd{} @acsmem{}}} |
829 | @c Same caveats as malloc. | |
830 | ||
831 | @c __libc_calloc @asulock @aculock @acsfd @acsmem | |
832 | @c *__malloc_hook dup unguarded | |
833 | @c memset dup ok | |
834 | @c arena_get @asulock @aculock @acsfd @acsmem | |
9f529d7c AO |
835 | @c arena_lock dup @asulock @aculock @acsfd @acsmem |
836 | @c top dup ok | |
837 | @c chunksize dup ok | |
838 | @c heap_for_ptr dup ok | |
839 | @c _int_malloc dup @acsfd @acsmem | |
840 | @c arena_get_retry dup @asulock @aculock @acsfd @acsmem | |
841 | @c mutex_unlock dup @aculock | |
842 | @c mem2chunk dup ok | |
843 | @c chunk_is_mmapped dup ok | |
844 | @c MALLOC_ZERO ok | |
845 | @c memset dup ok | |
28f540f4 RM |
846 | This function allocates a block long enough to contain a vector of |
847 | @var{count} elements, each of size @var{eltsize}. Its contents are | |
848 | cleared to zero before @code{calloc} returns. | |
849 | @end deftypefun | |
850 | ||
851 | You could define @code{calloc} as follows: | |
852 | ||
853 | @smallexample | |
854 | void * | |
855 | calloc (size_t count, size_t eltsize) | |
856 | @{ | |
857 | size_t size = count * eltsize; | |
858 | void *value = malloc (size); | |
859 | if (value != 0) | |
860 | memset (value, 0, size); | |
861 | return value; | |
862 | @} | |
863 | @end smallexample | |
864 | ||
c131718c UD |
865 | But in general, it is not guaranteed that @code{calloc} calls |
866 | @code{malloc} internally. Therefore, if an application provides its own | |
867 | @code{malloc}/@code{realloc}/@code{free} outside the C library, it | |
868 | should always define @code{calloc}, too. | |
869 | ||
28f540f4 | 870 | @node Efficiency and Malloc |
99a20616 | 871 | @subsubsection Efficiency Considerations for @code{malloc} |
28f540f4 RM |
872 | @cindex efficiency and @code{malloc} |
873 | ||
99a20616 UD |
874 | |
875 | ||
876 | ||
c131718c UD |
877 | @ignore |
878 | ||
879 | @c No longer true, see below instead. | |
28f540f4 RM |
880 | To make the best use of @code{malloc}, it helps to know that the GNU |
881 | version of @code{malloc} always dispenses small amounts of memory in | |
882 | blocks whose sizes are powers of two. It keeps separate pools for each | |
883 | power of two. This holds for sizes up to a page size. Therefore, if | |
884 | you are free to choose the size of a small block in order to make | |
885 | @code{malloc} more efficient, make it a power of two. | |
886 | @c !!! xref getpagesize | |
887 | ||
888 | Once a page is split up for a particular block size, it can't be reused | |
889 | for another size unless all the blocks in it are freed. In many | |
890 | programs, this is unlikely to happen. Thus, you can sometimes make a | |
891 | program use memory more efficiently by using blocks of the same size for | |
892 | many different purposes. | |
893 | ||
894 | When you ask for memory blocks of a page or larger, @code{malloc} uses a | |
895 | different strategy; it rounds the size up to a multiple of a page, and | |
896 | it can coalesce and split blocks as needed. | |
897 | ||
898 | The reason for the two strategies is that it is important to allocate | |
899 | and free small blocks as fast as possible, but speed is less important | |
900 | for a large block since the program normally spends a fair amount of | |
901 | time using it. Also, large blocks are normally fewer in number. | |
902 | Therefore, for large blocks, it makes sense to use a method which takes | |
903 | more time to minimize the wasted space. | |
904 | ||
c131718c UD |
905 | @end ignore |
906 | ||
1f77f049 | 907 | As opposed to other versions, the @code{malloc} in @theglibc{} |
99a20616 UD |
908 | does not round up block sizes to powers of two, neither for large nor |
909 | for small sizes. Neighboring chunks can be coalesced on a @code{free} | |
910 | no matter what their size is. This makes the implementation suitable | |
911 | for all kinds of allocation patterns without generally incurring high | |
912 | memory waste through fragmentation. | |
c131718c UD |
913 | |
914 | Very large blocks (much larger than a page) are allocated with | |
915 | @code{mmap} (anonymous or via @code{/dev/zero}) by this implementation. | |
916 | This has the great advantage that these chunks are returned to the | |
917 | system immediately when they are freed. Therefore, it cannot happen | |
918 | that a large chunk becomes ``locked'' in between smaller ones and even | |
919 | after calling @code{free} wastes memory. The size threshold for | |
920 | @code{mmap} to be used can be adjusted with @code{mallopt}. The use of | |
921 | @code{mmap} can also be disabled completely. | |
922 | ||
28f540f4 | 923 | @node Aligned Memory Blocks |
99a20616 | 924 | @subsubsection Allocating Aligned Memory Blocks |
28f540f4 RM |
925 | |
926 | @cindex page boundary | |
927 | @cindex alignment (with @code{malloc}) | |
928 | @pindex stdlib.h | |
929 | The address of a block returned by @code{malloc} or @code{realloc} in | |
a7a93d50 | 930 | @gnusystems{} is always a multiple of eight (or sixteen on 64-bit |
c131718c | 931 | systems). If you need a block whose address is a multiple of a higher |
5764c27f WN |
932 | power of two than that, use @code{aligned_alloc} or @code{posix_memalign}. |
933 | @code{aligned_alloc} and @code{posix_memalign} are declared in | |
934 | @file{stdlib.h}. | |
935 | ||
936 | @comment stdlib.h | |
937 | @deftypefun {void *} aligned_alloc (size_t @var{alignment}, size_t @var{size}) | |
9f529d7c AO |
938 | @safety{@prelim{}@mtsafe{}@asunsafe{@asulock{}}@acunsafe{@aculock{} @acsfd{} @acsmem{}}} |
939 | @c Alias to memalign. | |
5764c27f WN |
940 | The @code{aligned_alloc} function allocates a block of @var{size} bytes whose |
941 | address is a multiple of @var{alignment}. The @var{alignment} must be a | |
942 | power of two and @var{size} must be a multiple of @var{alignment}. | |
943 | ||
944 | The @code{aligned_alloc} function returns a null pointer on error and sets | |
945 | @code{errno} to one of the following values: | |
946 | ||
947 | @table @code | |
948 | @item ENOMEM | |
949 | There was insufficient memory available to satisfy the request. | |
950 | ||
951 | @item EINVAL | |
952 | @var{alignment} is not a power of two. | |
953 | ||
954 | This function was introduced in @w{ISO C11} and hence may have better | |
955 | portability to modern non-POSIX systems than @code{posix_memalign}. | |
956 | @end table | |
957 | ||
958 | @end deftypefun | |
28f540f4 | 959 | |
eab0f04c | 960 | @comment malloc.h |
28f540f4 | 961 | @comment BSD |
22a1292a | 962 | @deftypefun {void *} memalign (size_t @var{boundary}, size_t @var{size}) |
9f529d7c AO |
963 | @safety{@prelim{}@mtsafe{}@asunsafe{@asulock{}}@acunsafe{@aculock{} @acsfd{} @acsmem{}}} |
964 | @c Same issues as malloc. The padding bytes are safely freed in | |
965 | @c _int_memalign, with the arena still locked. | |
966 | ||
967 | @c __libc_memalign @asulock @aculock @acsfd @acsmem | |
968 | @c *__memalign_hook dup unguarded | |
969 | @c __libc_malloc dup @asulock @aculock @acsfd @acsmem | |
970 | @c arena_get dup @asulock @aculock @acsfd @acsmem | |
971 | @c _int_memalign @acsfd @acsmem | |
972 | @c _int_malloc dup @acsfd @acsmem | |
973 | @c checked_request2size dup ok | |
974 | @c mem2chunk dup ok | |
975 | @c chunksize dup ok | |
976 | @c chunk_is_mmapped dup ok | |
977 | @c set_head dup ok | |
978 | @c chunk2mem dup ok | |
979 | @c set_inuse_bit_at_offset dup ok | |
980 | @c set_head_size dup ok | |
981 | @c _int_free (have_lock) dup @acsfd @acsmem | |
982 | @c chunk_at_offset dup ok | |
983 | @c check_inuse_chunk dup ok | |
984 | @c arena_get_retry dup @asulock @aculock @acsfd @acsmem | |
985 | @c mutex_unlock dup @aculock | |
28f540f4 RM |
986 | The @code{memalign} function allocates a block of @var{size} bytes whose |
987 | address is a multiple of @var{boundary}. The @var{boundary} must be a | |
c131718c UD |
988 | power of two! The function @code{memalign} works by allocating a |
989 | somewhat larger block, and then returning an address within the block | |
990 | that is on the specified boundary. | |
0a096e44 WN |
991 | |
992 | The @code{memalign} function returns a null pointer on error and sets | |
993 | @code{errno} to one of the following values: | |
994 | ||
995 | @table @code | |
996 | @item ENOMEM | |
997 | There was insufficient memory available to satisfy the request. | |
998 | ||
999 | @item EINVAL | |
3ef569c7 | 1000 | @var{boundary} is not a power of two. |
0a096e44 WN |
1001 | |
1002 | @end table | |
1003 | ||
5764c27f WN |
1004 | The @code{memalign} function is obsolete and @code{aligned_alloc} or |
1005 | @code{posix_memalign} should be used instead. | |
28f540f4 RM |
1006 | @end deftypefun |
1007 | ||
68979757 UD |
1008 | @comment stdlib.h |
1009 | @comment POSIX | |
1010 | @deftypefun int posix_memalign (void **@var{memptr}, size_t @var{alignment}, size_t @var{size}) | |
9f529d7c AO |
1011 | @safety{@prelim{}@mtsafe{}@asunsafe{@asulock{}}@acunsafe{@aculock{} @acsfd{} @acsmem{}}} |
1012 | @c Calls memalign unless the requirements are not met (powerof2 macro is | |
1013 | @c safe given an automatic variable as an argument) or there's a | |
1014 | @c memalign hook (accessed unguarded, but safely). | |
68979757 UD |
1015 | The @code{posix_memalign} function is similar to the @code{memalign} |
1016 | function in that it returns a buffer of @var{size} bytes aligned to a | |
1017 | multiple of @var{alignment}. But it adds one requirement to the | |
1018 | parameter @var{alignment}: the value must be a power of two multiple of | |
1019 | @code{sizeof (void *)}. | |
1020 | ||
1021 | If the function succeeds in allocation memory a pointer to the allocated | |
1022 | memory is returned in @code{*@var{memptr}} and the return value is zero. | |
1023 | Otherwise the function returns an error value indicating the problem. | |
0a096e44 WN |
1024 | The possible error values returned are: |
1025 | ||
1026 | @table @code | |
1027 | @item ENOMEM | |
1028 | There was insufficient memory available to satisfy the request. | |
1029 | ||
1030 | @item EINVAL | |
1031 | @var{alignment} is not a power of two multiple of @code{sizeof (void *)}. | |
1032 | ||
1033 | @end table | |
68979757 | 1034 | |
cf822e3c | 1035 | This function was introduced in POSIX 1003.1d. Although this function is |
5764c27f WN |
1036 | superseded by @code{aligned_alloc}, it is more portable to older POSIX |
1037 | systems that do not support @w{ISO C11}. | |
68979757 UD |
1038 | @end deftypefun |
1039 | ||
28f540f4 RM |
1040 | @comment malloc.h stdlib.h |
1041 | @comment BSD | |
1042 | @deftypefun {void *} valloc (size_t @var{size}) | |
9f529d7c AO |
1043 | @safety{@prelim{}@mtunsafe{@mtuinit{}}@asunsafe{@asuinit{} @asulock{}}@acunsafe{@acuinit{} @aculock{} @acsfd{} @acsmem{}}} |
1044 | @c __libc_valloc @mtuinit @asuinit @asulock @aculock @acsfd @acsmem | |
1045 | @c ptmalloc_init (once) @mtsenv @asulock @aculock @acsfd @acsmem | |
1046 | @c _dl_addr @asucorrupt? @aculock | |
1047 | @c __rtld_lock_lock_recursive (dl_load_lock) @asucorrupt? @aculock | |
1048 | @c _dl_find_dso_for_object ok, iterates over dl_ns and its _ns_loaded objs | |
1049 | @c the ok above assumes no partial updates on dl_ns and _ns_loaded | |
1050 | @c that could confuse a _dl_addr call in a signal handler | |
1051 | @c _dl_addr_inside_object ok | |
1052 | @c determine_info ok | |
1053 | @c __rtld_lock_unlock_recursive (dl_load_lock) @aculock | |
9f529d7c AO |
1054 | @c *_environ @mtsenv |
1055 | @c next_env_entry ok | |
1056 | @c strcspn dup ok | |
1057 | @c __libc_mallopt dup @mtasuconst:mallopt [setting mp_] | |
1058 | @c __malloc_check_init @mtasuconst:malloc_hooks [setting hooks] | |
1059 | @c *__malloc_initialize_hook unguarded, ok | |
1060 | @c *__memalign_hook dup ok, unguarded | |
1061 | @c arena_get dup @asulock @aculock @acsfd @acsmem | |
1062 | @c _int_valloc @acsfd @acsmem | |
1063 | @c malloc_consolidate dup ok | |
1064 | @c _int_memalign dup @acsfd @acsmem | |
1065 | @c arena_get_retry dup @asulock @aculock @acsfd @acsmem | |
1066 | @c _int_memalign dup @acsfd @acsmem | |
1067 | @c mutex_unlock dup @aculock | |
28f540f4 | 1068 | Using @code{valloc} is like using @code{memalign} and passing the page size |
3ef569c7 | 1069 | as the value of the first argument. It is implemented like this: |
28f540f4 RM |
1070 | |
1071 | @smallexample | |
1072 | void * | |
1073 | valloc (size_t size) | |
1074 | @{ | |
22a1292a | 1075 | return memalign (getpagesize (), size); |
28f540f4 RM |
1076 | @} |
1077 | @end smallexample | |
b642f101 UD |
1078 | |
1079 | @ref{Query Memory Parameters} for more information about the memory | |
1080 | subsystem. | |
0a096e44 | 1081 | |
5764c27f WN |
1082 | The @code{valloc} function is obsolete and @code{aligned_alloc} or |
1083 | @code{posix_memalign} should be used instead. | |
28f540f4 RM |
1084 | @end deftypefun |
1085 | ||
c131718c | 1086 | @node Malloc Tunable Parameters |
99a20616 | 1087 | @subsubsection Malloc Tunable Parameters |
c131718c UD |
1088 | |
1089 | You can adjust some parameters for dynamic memory allocation with the | |
1090 | @code{mallopt} function. This function is the general SVID/XPG | |
1091 | interface, defined in @file{malloc.h}. | |
1092 | @pindex malloc.h | |
1093 | ||
1094 | @deftypefun int mallopt (int @var{param}, int @var{value}) | |
9f529d7c AO |
1095 | @safety{@prelim{}@mtunsafe{@mtuinit{} @mtasuconst{:mallopt}}@asunsafe{@asuinit{} @asulock{}}@acunsafe{@acuinit{} @aculock{}}} |
1096 | @c __libc_mallopt @mtuinit @mtasuconst:mallopt @asuinit @asulock @aculock | |
1097 | @c ptmalloc_init (once) dup @mtsenv @asulock @aculock @acsfd @acsmem | |
1098 | @c mutex_lock (main_arena->mutex) @asulock @aculock | |
1099 | @c malloc_consolidate dup ok | |
1100 | @c set_max_fast ok | |
1101 | @c mutex_unlock dup @aculock | |
1102 | ||
c131718c UD |
1103 | When calling @code{mallopt}, the @var{param} argument specifies the |
1104 | parameter to be set, and @var{value} the new value to be set. Possible | |
1105 | choices for @var{param}, as defined in @file{malloc.h}, are: | |
1106 | ||
1107 | @table @code | |
ec4ff04d CD |
1108 | @comment TODO: @item M_ARENA_MAX |
1109 | @comment - Document ARENA_MAX env var. | |
1110 | @comment TODO: @item M_ARENA_TEST | |
1111 | @comment - Document ARENA_TEST env var. | |
1112 | @comment TODO: @item M_CHECK_ACTION | |
1113 | @item M_MMAP_MAX | |
1114 | The maximum number of chunks to allocate with @code{mmap}. Setting this | |
1115 | to zero disables all use of @code{mmap}. | |
2bce3035 SP |
1116 | |
1117 | The default value of this parameter is @code{65536}. | |
1118 | ||
1119 | This parameter can also be set for the process at startup by setting the | |
1120 | environment variable @env{MALLOC_MMAP_MAX_} to the desired value. | |
1121 | ||
c131718c UD |
1122 | @item M_MMAP_THRESHOLD |
1123 | All chunks larger than this value are allocated outside the normal | |
1124 | heap, using the @code{mmap} system call. This way it is guaranteed | |
1125 | that the memory for these chunks can be returned to the system on | |
13c0f771 AJ |
1126 | @code{free}. Note that requests smaller than this threshold might still |
1127 | be allocated via @code{mmap}. | |
2bce3035 SP |
1128 | |
1129 | If this parameter is not set, the default value is set as 128 KiB and the | |
1130 | threshold is adjusted dynamically to suit the allocation patterns of the | |
1131 | program. If the parameter is set, the dynamic adjustment is disabled and the | |
1132 | value is set statically to the input value. | |
1133 | ||
1134 | This parameter can also be set for the process at startup by setting the | |
1135 | environment variable @env{MALLOC_MMAP_THRESHOLD_} to the desired value. | |
ec4ff04d | 1136 | @comment TODO: @item M_MXFAST |
2bce3035 | 1137 | |
deb9cabb AS |
1138 | @item M_PERTURB |
1139 | If non-zero, memory blocks are filled with values depending on some | |
1140 | low order bits of this parameter when they are allocated (except when | |
1141 | allocated by @code{calloc}) and freed. This can be used to debug the | |
b741de23 SP |
1142 | use of uninitialized or freed heap memory. Note that this option does not |
1143 | guarantee that the freed block will have any specific values. It only | |
1144 | guarantees that the content the block had before it was freed will be | |
1145 | overwritten. | |
2bce3035 SP |
1146 | |
1147 | The default value of this parameter is @code{0}. | |
1148 | ||
1149 | This parameter can also be set for the process at startup by setting the | |
1150 | environment variable @env{MALLOC_MMAP_PERTURB_} to the desired value. | |
1151 | ||
ec4ff04d CD |
1152 | @item M_TOP_PAD |
1153 | This parameter determines the amount of extra memory to obtain from the | |
1154 | system when a call to @code{sbrk} is required. It also specifies the | |
1155 | number of bytes to retain when shrinking the heap by calling @code{sbrk} | |
1156 | with a negative argument. This provides the necessary hysteresis in | |
1157 | heap size such that excessive amounts of system calls can be avoided. | |
2bce3035 SP |
1158 | |
1159 | The default value of this parameter is @code{0}. | |
1160 | ||
1161 | This parameter can also be set for the process at startup by setting the | |
1162 | environment variable @env{MALLOC_TOP_PAD_} to the desired value. | |
1163 | ||
ec4ff04d CD |
1164 | @item M_TRIM_THRESHOLD |
1165 | This is the minimum size (in bytes) of the top-most, releasable chunk | |
1166 | that will cause @code{sbrk} to be called with a negative argument in | |
1167 | order to return memory to the system. | |
2bce3035 SP |
1168 | |
1169 | If this parameter is not set, the default value is set as 128 KiB and the | |
1170 | threshold is adjusted dynamically to suit the allocation patterns of the | |
1171 | program. If the parameter is set, the dynamic adjustment is disabled and the | |
1172 | value is set statically to the provided input. | |
1173 | ||
1174 | This parameter can also be set for the process at startup by setting the | |
1175 | environment variable @env{MALLOC_TRIM_THRESHOLD_} to the desired value. | |
1176 | ||
c131718c UD |
1177 | @end table |
1178 | ||
1179 | @end deftypefun | |
1180 | ||
28f540f4 | 1181 | @node Heap Consistency Checking |
99a20616 | 1182 | @subsubsection Heap Consistency Checking |
28f540f4 RM |
1183 | |
1184 | @cindex heap consistency checking | |
1185 | @cindex consistency checking, of heap | |
1186 | ||
99a20616 | 1187 | You can ask @code{malloc} to check the consistency of dynamic memory by |
28f540f4 | 1188 | using the @code{mcheck} function. This function is a GNU extension, |
4775243a UD |
1189 | declared in @file{mcheck.h}. |
1190 | @pindex mcheck.h | |
28f540f4 | 1191 | |
4775243a | 1192 | @comment mcheck.h |
28f540f4 RM |
1193 | @comment GNU |
1194 | @deftypefun int mcheck (void (*@var{abortfn}) (enum mcheck_status @var{status})) | |
9f529d7c AO |
1195 | @safety{@prelim{}@mtunsafe{@mtasurace{:mcheck} @mtasuconst{:malloc_hooks}}@asunsafe{@asucorrupt{}}@acunsafe{@acucorrupt{}}} |
1196 | @c The hooks must be set up before malloc is first used, which sort of | |
1197 | @c implies @mtuinit/@asuinit but since the function is a no-op if malloc | |
1198 | @c was already used, that doesn't pose any safety issues. The actual | |
1199 | @c problem is with the hooks, designed for single-threaded | |
1200 | @c fully-synchronous operation: they manage an unguarded linked list of | |
1201 | @c allocated blocks, and get temporarily overwritten before calling the | |
1202 | @c allocation functions recursively while holding the old hooks. There | |
1203 | @c are no guards for thread safety, and inconsistent hooks may be found | |
1204 | @c within signal handlers or left behind in case of cancellation. | |
1205 | ||
28f540f4 RM |
1206 | Calling @code{mcheck} tells @code{malloc} to perform occasional |
1207 | consistency checks. These will catch things such as writing | |
1208 | past the end of a block that was allocated with @code{malloc}. | |
1209 | ||
1210 | The @var{abortfn} argument is the function to call when an inconsistency | |
1211 | is found. If you supply a null pointer, then @code{mcheck} uses a | |
1212 | default function which prints a message and calls @code{abort} | |
1213 | (@pxref{Aborting a Program}). The function you supply is called with | |
1214 | one argument, which says what sort of inconsistency was detected; its | |
1215 | type is described below. | |
1216 | ||
1217 | It is too late to begin allocation checking once you have allocated | |
1218 | anything with @code{malloc}. So @code{mcheck} does nothing in that | |
1219 | case. The function returns @code{-1} if you call it too late, and | |
1220 | @code{0} otherwise (when it is successful). | |
1221 | ||
1222 | The easiest way to arrange to call @code{mcheck} early enough is to use | |
1223 | the option @samp{-lmcheck} when you link your program; then you don't | |
bc938d3d | 1224 | need to modify your program source at all. Alternatively you might use |
3cb07217 UD |
1225 | a debugger to insert a call to @code{mcheck} whenever the program is |
1226 | started, for example these gdb commands will automatically call @code{mcheck} | |
1227 | whenever the program starts: | |
1228 | ||
1229 | @smallexample | |
1230 | (gdb) break main | |
1231 | Breakpoint 1, main (argc=2, argv=0xbffff964) at whatever.c:10 | |
1232 | (gdb) command 1 | |
1233 | Type commands for when breakpoint 1 is hit, one per line. | |
1234 | End with a line saying just "end". | |
1235 | >call mcheck(0) | |
1236 | >continue | |
1237 | >end | |
95fdc6a0 | 1238 | (gdb) @dots{} |
3cb07217 UD |
1239 | @end smallexample |
1240 | ||
1241 | This will however only work if no initialization function of any object | |
1242 | involved calls any of the @code{malloc} functions since @code{mcheck} | |
1243 | must be called before the first such function. | |
1244 | ||
28f540f4 RM |
1245 | @end deftypefun |
1246 | ||
1247 | @deftypefun {enum mcheck_status} mprobe (void *@var{pointer}) | |
9f529d7c AO |
1248 | @safety{@prelim{}@mtunsafe{@mtasurace{:mcheck} @mtasuconst{:malloc_hooks}}@asunsafe{@asucorrupt{}}@acunsafe{@acucorrupt{}}} |
1249 | @c The linked list of headers may be modified concurrently by other | |
1250 | @c threads, and it may find a partial update if called from a signal | |
1251 | @c handler. It's mostly read only, so cancelling it might be safe, but | |
1252 | @c it will modify global state that, if cancellation hits at just the | |
1253 | @c right spot, may be left behind inconsistent. This path is only taken | |
1254 | @c if checkhdr finds an inconsistency. If the inconsistency could only | |
1255 | @c occur because of earlier undefined behavior, that wouldn't be an | |
1256 | @c additional safety issue problem, but because of the other concurrency | |
1257 | @c issues in the mcheck hooks, the apparent inconsistency could be the | |
1258 | @c result of mcheck's own internal data race. So, AC-Unsafe it is. | |
1259 | ||
28f540f4 RM |
1260 | The @code{mprobe} function lets you explicitly check for inconsistencies |
1261 | in a particular allocated block. You must have already called | |
1262 | @code{mcheck} at the beginning of the program, to do its occasional | |
1263 | checks; calling @code{mprobe} requests an additional consistency check | |
1264 | to be done at the time of the call. | |
1265 | ||
1266 | The argument @var{pointer} must be a pointer returned by @code{malloc} | |
1267 | or @code{realloc}. @code{mprobe} returns a value that says what | |
1268 | inconsistency, if any, was found. The values are described below. | |
1269 | @end deftypefun | |
1270 | ||
1271 | @deftp {Data Type} {enum mcheck_status} | |
1272 | This enumerated type describes what kind of inconsistency was detected | |
1273 | in an allocated block, if any. Here are the possible values: | |
1274 | ||
1275 | @table @code | |
1276 | @item MCHECK_DISABLED | |
1277 | @code{mcheck} was not called before the first allocation. | |
1278 | No consistency checking can be done. | |
1279 | @item MCHECK_OK | |
1280 | No inconsistency detected. | |
1281 | @item MCHECK_HEAD | |
1282 | The data immediately before the block was modified. | |
1283 | This commonly happens when an array index or pointer | |
1284 | is decremented too far. | |
1285 | @item MCHECK_TAIL | |
1286 | The data immediately after the block was modified. | |
1287 | This commonly happens when an array index or pointer | |
1288 | is incremented too far. | |
1289 | @item MCHECK_FREE | |
1290 | The block was already freed. | |
1291 | @end table | |
1292 | @end deftp | |
1293 | ||
7551a1e5 UD |
1294 | Another possibility to check for and guard against bugs in the use of |
1295 | @code{malloc}, @code{realloc} and @code{free} is to set the environment | |
1296 | variable @code{MALLOC_CHECK_}. When @code{MALLOC_CHECK_} is set, a | |
1297 | special (less efficient) implementation is used which is designed to be | |
1298 | tolerant against simple errors, such as double calls of @code{free} with | |
1299 | the same argument, or overruns of a single byte (off-by-one bugs). Not | |
bc938d3d | 1300 | all such errors can be protected against, however, and memory leaks can |
7551a1e5 UD |
1301 | result. If @code{MALLOC_CHECK_} is set to @code{0}, any detected heap |
1302 | corruption is silently ignored; if set to @code{1}, a diagnostic is | |
1303 | printed on @code{stderr}; if set to @code{2}, @code{abort} is called | |
1304 | immediately. This can be useful because otherwise a crash may happen | |
1305 | much later, and the true cause for the problem is then very hard to | |
1306 | track down. | |
1307 | ||
68979757 UD |
1308 | There is one problem with @code{MALLOC_CHECK_}: in SUID or SGID binaries |
1309 | it could possibly be exploited since diverging from the normal programs | |
0bc93a2f | 1310 | behavior it now writes something to the standard error descriptor. |
68979757 UD |
1311 | Therefore the use of @code{MALLOC_CHECK_} is disabled by default for |
1312 | SUID and SGID binaries. It can be enabled again by the system | |
1313 | administrator by adding a file @file{/etc/suid-debug} (the content is | |
1314 | not important it could be empty). | |
1315 | ||
789b13c4 | 1316 | So, what's the difference between using @code{MALLOC_CHECK_} and linking |
bc938d3d | 1317 | with @samp{-lmcheck}? @code{MALLOC_CHECK_} is orthogonal with respect to |
789b13c4 UD |
1318 | @samp{-lmcheck}. @samp{-lmcheck} has been added for backward |
1319 | compatibility. Both @code{MALLOC_CHECK_} and @samp{-lmcheck} should | |
1320 | uncover the same bugs - but using @code{MALLOC_CHECK_} you don't need to | |
1321 | recompile your application. | |
1322 | ||
28f540f4 | 1323 | @node Hooks for Malloc |
99a20616 | 1324 | @subsubsection Memory Allocation Hooks |
28f540f4 RM |
1325 | @cindex allocation hooks, for @code{malloc} |
1326 | ||
1f77f049 | 1327 | @Theglibc{} lets you modify the behavior of @code{malloc}, |
28f540f4 RM |
1328 | @code{realloc}, and @code{free} by specifying appropriate hook |
1329 | functions. You can use these hooks to help you debug programs that use | |
99a20616 | 1330 | dynamic memory allocation, for example. |
28f540f4 RM |
1331 | |
1332 | The hook variables are declared in @file{malloc.h}. | |
1333 | @pindex malloc.h | |
1334 | ||
1335 | @comment malloc.h | |
1336 | @comment GNU | |
1337 | @defvar __malloc_hook | |
bc938d3d UD |
1338 | The value of this variable is a pointer to the function that |
1339 | @code{malloc} uses whenever it is called. You should define this | |
1340 | function to look like @code{malloc}; that is, like: | |
28f540f4 RM |
1341 | |
1342 | @smallexample | |
18a3a9a3 | 1343 | void *@var{function} (size_t @var{size}, const void *@var{caller}) |
28f540f4 | 1344 | @end smallexample |
bd355af0 UD |
1345 | |
1346 | The value of @var{caller} is the return address found on the stack when | |
bc938d3d UD |
1347 | the @code{malloc} function was called. This value allows you to trace |
1348 | the memory consumption of the program. | |
28f540f4 RM |
1349 | @end defvar |
1350 | ||
1351 | @comment malloc.h | |
1352 | @comment GNU | |
1353 | @defvar __realloc_hook | |
1354 | The value of this variable is a pointer to function that @code{realloc} | |
1355 | uses whenever it is called. You should define this function to look | |
1356 | like @code{realloc}; that is, like: | |
1357 | ||
1358 | @smallexample | |
18a3a9a3 | 1359 | void *@var{function} (void *@var{ptr}, size_t @var{size}, const void *@var{caller}) |
28f540f4 | 1360 | @end smallexample |
bd355af0 UD |
1361 | |
1362 | The value of @var{caller} is the return address found on the stack when | |
e8b1163e | 1363 | the @code{realloc} function was called. This value allows you to trace the |
bd355af0 | 1364 | memory consumption of the program. |
28f540f4 RM |
1365 | @end defvar |
1366 | ||
1367 | @comment malloc.h | |
1368 | @comment GNU | |
1369 | @defvar __free_hook | |
1370 | The value of this variable is a pointer to function that @code{free} | |
1371 | uses whenever it is called. You should define this function to look | |
1372 | like @code{free}; that is, like: | |
1373 | ||
1374 | @smallexample | |
18a3a9a3 | 1375 | void @var{function} (void *@var{ptr}, const void *@var{caller}) |
28f540f4 | 1376 | @end smallexample |
bd355af0 UD |
1377 | |
1378 | The value of @var{caller} is the return address found on the stack when | |
e8b1163e | 1379 | the @code{free} function was called. This value allows you to trace the |
bd355af0 | 1380 | memory consumption of the program. |
28f540f4 RM |
1381 | @end defvar |
1382 | ||
3cb07217 UD |
1383 | @comment malloc.h |
1384 | @comment GNU | |
1385 | @defvar __memalign_hook | |
5764c27f WN |
1386 | The value of this variable is a pointer to function that @code{aligned_alloc}, |
1387 | @code{memalign}, @code{posix_memalign} and @code{valloc} use whenever they | |
cf822e3c | 1388 | are called. You should define this function to look like @code{aligned_alloc}; |
5764c27f | 1389 | that is, like: |
3cb07217 UD |
1390 | |
1391 | @smallexample | |
46ca7a1c | 1392 | void *@var{function} (size_t @var{alignment}, size_t @var{size}, const void *@var{caller}) |
3cb07217 | 1393 | @end smallexample |
18a3a9a3 UD |
1394 | |
1395 | The value of @var{caller} is the return address found on the stack when | |
5764c27f WN |
1396 | the @code{aligned_alloc}, @code{memalign}, @code{posix_memalign} or |
1397 | @code{valloc} functions are called. This value allows you to trace the | |
1398 | memory consumption of the program. | |
3cb07217 UD |
1399 | @end defvar |
1400 | ||
28f540f4 RM |
1401 | You must make sure that the function you install as a hook for one of |
1402 | these functions does not call that function recursively without restoring | |
1403 | the old value of the hook first! Otherwise, your program will get stuck | |
3cb07217 UD |
1404 | in an infinite recursion. Before calling the function recursively, one |
1405 | should make sure to restore all the hooks to their previous value. When | |
1406 | coming back from the recursive call, all the hooks should be resaved | |
1407 | since a hook might modify itself. | |
28f540f4 | 1408 | |
b2f46c3c UD |
1409 | An issue to look out for is the time at which the malloc hook functions |
1410 | can be safely installed. If the hook functions call the malloc-related | |
1411 | functions recursively, it is necessary that malloc has already properly | |
1412 | initialized itself at the time when @code{__malloc_hook} etc. is | |
1413 | assigned to. On the other hand, if the hook functions provide a | |
1414 | complete malloc implementation of their own, it is vital that the hooks | |
1415 | are assigned to @emph{before} the very first @code{malloc} call has | |
1416 | completed, because otherwise a chunk obtained from the ordinary, | |
1417 | un-hooked malloc may later be handed to @code{__free_hook}, for example. | |
1418 | ||
3cb07217 UD |
1419 | Here is an example showing how to use @code{__malloc_hook} and |
1420 | @code{__free_hook} properly. It installs a function that prints out | |
1421 | information every time @code{malloc} or @code{free} is called. We just | |
1422 | assume here that @code{realloc} and @code{memalign} are not used in our | |
1423 | program. | |
28f540f4 RM |
1424 | |
1425 | @smallexample | |
18a3a9a3 UD |
1426 | /* Prototypes for __malloc_hook, __free_hook */ |
1427 | #include <malloc.h> | |
3cb07217 UD |
1428 | |
1429 | /* Prototypes for our hooks. */ | |
2ac057a0 | 1430 | static void my_init_hook (void); |
18a3a9a3 UD |
1431 | static void *my_malloc_hook (size_t, const void *); |
1432 | static void my_free_hook (void*, const void *); | |
b2f46c3c | 1433 | |
b2f46c3c | 1434 | static void |
2ba3cfa1 | 1435 | my_init (void) |
b2f46c3c UD |
1436 | @{ |
1437 | old_malloc_hook = __malloc_hook; | |
1438 | old_free_hook = __free_hook; | |
1439 | __malloc_hook = my_malloc_hook; | |
1440 | __free_hook = my_free_hook; | |
1441 | @} | |
3cb07217 | 1442 | |
28f540f4 | 1443 | static void * |
18a3a9a3 | 1444 | my_malloc_hook (size_t size, const void *caller) |
28f540f4 RM |
1445 | @{ |
1446 | void *result; | |
3cb07217 | 1447 | /* Restore all old hooks */ |
28f540f4 | 1448 | __malloc_hook = old_malloc_hook; |
3cb07217 UD |
1449 | __free_hook = old_free_hook; |
1450 | /* Call recursively */ | |
28f540f4 | 1451 | result = malloc (size); |
0bc93a2f | 1452 | /* Save underlying hooks */ |
3cb07217 UD |
1453 | old_malloc_hook = __malloc_hook; |
1454 | old_free_hook = __free_hook; | |
28f540f4 RM |
1455 | /* @r{@code{printf} might call @code{malloc}, so protect it too.} */ |
1456 | printf ("malloc (%u) returns %p\n", (unsigned int) size, result); | |
3cb07217 | 1457 | /* Restore our own hooks */ |
28f540f4 | 1458 | __malloc_hook = my_malloc_hook; |
3cb07217 | 1459 | __free_hook = my_free_hook; |
28f540f4 RM |
1460 | return result; |
1461 | @} | |
1462 | ||
2ac057a0 | 1463 | static void |
18a3a9a3 | 1464 | my_free_hook (void *ptr, const void *caller) |
3cb07217 UD |
1465 | @{ |
1466 | /* Restore all old hooks */ | |
1467 | __malloc_hook = old_malloc_hook; | |
1468 | __free_hook = old_free_hook; | |
1469 | /* Call recursively */ | |
1470 | free (ptr); | |
0bc93a2f | 1471 | /* Save underlying hooks */ |
3cb07217 UD |
1472 | old_malloc_hook = __malloc_hook; |
1473 | old_free_hook = __free_hook; | |
1474 | /* @r{@code{printf} might call @code{free}, so protect it too.} */ | |
1475 | printf ("freed pointer %p\n", ptr); | |
1476 | /* Restore our own hooks */ | |
1477 | __malloc_hook = my_malloc_hook; | |
1478 | __free_hook = my_free_hook; | |
1479 | @} | |
1480 | ||
28f540f4 RM |
1481 | main () |
1482 | @{ | |
2ba3cfa1 | 1483 | my_init (); |
95fdc6a0 | 1484 | @dots{} |
28f540f4 RM |
1485 | @} |
1486 | @end smallexample | |
1487 | ||
1488 | The @code{mcheck} function (@pxref{Heap Consistency Checking}) works by | |
1489 | installing such hooks. | |
1490 | ||
1491 | @c __morecore, __after_morecore_hook are undocumented | |
1492 | @c It's not clear whether to document them. | |
1493 | ||
1494 | @node Statistics of Malloc | |
99a20616 | 1495 | @subsubsection Statistics for Memory Allocation with @code{malloc} |
28f540f4 RM |
1496 | |
1497 | @cindex allocation statistics | |
99a20616 | 1498 | You can get information about dynamic memory allocation by calling the |
c131718c UD |
1499 | @code{mallinfo} function. This function and its associated data type |
1500 | are declared in @file{malloc.h}; they are an extension of the standard | |
1501 | SVID/XPG version. | |
28f540f4 RM |
1502 | @pindex malloc.h |
1503 | ||
1504 | @comment malloc.h | |
1505 | @comment GNU | |
c131718c | 1506 | @deftp {Data Type} {struct mallinfo} |
28f540f4 | 1507 | This structure type is used to return information about the dynamic |
99a20616 | 1508 | memory allocator. It contains the following members: |
28f540f4 RM |
1509 | |
1510 | @table @code | |
c131718c UD |
1511 | @item int arena |
1512 | This is the total size of memory allocated with @code{sbrk} by | |
1513 | @code{malloc}, in bytes. | |
1514 | ||
1515 | @item int ordblks | |
99a20616 | 1516 | This is the number of chunks not in use. (The memory allocator |
c131718c UD |
1517 | internally gets chunks of memory from the operating system, and then |
1518 | carves them up to satisfy individual @code{malloc} requests; see | |
1519 | @ref{Efficiency and Malloc}.) | |
1520 | ||
1521 | @item int smblks | |
1522 | This field is unused. | |
1523 | ||
1524 | @item int hblks | |
1525 | This is the total number of chunks allocated with @code{mmap}. | |
1526 | ||
1527 | @item int hblkhd | |
1528 | This is the total size of memory allocated with @code{mmap}, in bytes. | |
1529 | ||
1530 | @item int usmblks | |
ca135f82 | 1531 | This field is unused and always 0. |
28f540f4 | 1532 | |
c131718c UD |
1533 | @item int fsmblks |
1534 | This field is unused. | |
28f540f4 | 1535 | |
c131718c UD |
1536 | @item int uordblks |
1537 | This is the total size of memory occupied by chunks handed out by | |
1538 | @code{malloc}. | |
1539 | ||
1540 | @item int fordblks | |
1541 | This is the total size of memory occupied by free (not in use) chunks. | |
28f540f4 | 1542 | |
c131718c | 1543 | @item int keepcost |
e8b1163e | 1544 | This is the size of the top-most releasable chunk that normally |
11bf311e | 1545 | borders the end of the heap (i.e., the high end of the virtual address |
99a20616 | 1546 | space's data segment). |
28f540f4 | 1547 | |
28f540f4 RM |
1548 | @end table |
1549 | @end deftp | |
1550 | ||
1551 | @comment malloc.h | |
c131718c UD |
1552 | @comment SVID |
1553 | @deftypefun {struct mallinfo} mallinfo (void) | |
9f529d7c AO |
1554 | @safety{@prelim{}@mtunsafe{@mtuinit{} @mtasuconst{:mallopt}}@asunsafe{@asuinit{} @asulock{}}@acunsafe{@acuinit{} @aculock{}}} |
1555 | @c Accessing mp_.n_mmaps and mp_.max_mmapped_mem, modified with atomics | |
1556 | @c but non-atomically elsewhere, may get us inconsistent results. We | |
1557 | @c mark the statistics as unsafe, rather than the fast-path functions | |
1558 | @c that collect the possibly inconsistent data. | |
1559 | ||
1560 | @c __libc_mallinfo @mtuinit @mtasuconst:mallopt @asuinit @asulock @aculock | |
1561 | @c ptmalloc_init (once) dup @mtsenv @asulock @aculock @acsfd @acsmem | |
1562 | @c mutex_lock dup @asulock @aculock | |
1563 | @c int_mallinfo @mtasuconst:mallopt [mp_ access on main_arena] | |
1564 | @c malloc_consolidate dup ok | |
1565 | @c check_malloc_state dup ok/disabled | |
1566 | @c chunksize dup ok | |
1567 | @c fastbin dupo ok | |
1568 | @c bin_at dup ok | |
1569 | @c last dup ok | |
1570 | @c mutex_unlock @aculock | |
1571 | ||
28f540f4 | 1572 | This function returns information about the current dynamic memory usage |
c131718c | 1573 | in a structure of type @code{struct mallinfo}. |
28f540f4 RM |
1574 | @end deftypefun |
1575 | ||
1576 | @node Summary of Malloc | |
99a20616 | 1577 | @subsubsection Summary of @code{malloc}-Related Functions |
28f540f4 RM |
1578 | |
1579 | Here is a summary of the functions that work with @code{malloc}: | |
1580 | ||
1581 | @table @code | |
1582 | @item void *malloc (size_t @var{size}) | |
1583 | Allocate a block of @var{size} bytes. @xref{Basic Allocation}. | |
1584 | ||
1585 | @item void free (void *@var{addr}) | |
1586 | Free a block previously allocated by @code{malloc}. @xref{Freeing after | |
1587 | Malloc}. | |
1588 | ||
1589 | @item void *realloc (void *@var{addr}, size_t @var{size}) | |
1590 | Make a block previously allocated by @code{malloc} larger or smaller, | |
1591 | possibly by copying it to a new location. @xref{Changing Block Size}. | |
1592 | ||
1593 | @item void *calloc (size_t @var{count}, size_t @var{eltsize}) | |
1594 | Allocate a block of @var{count} * @var{eltsize} bytes using | |
1595 | @code{malloc}, and set its contents to zero. @xref{Allocating Cleared | |
1596 | Space}. | |
1597 | ||
1598 | @item void *valloc (size_t @var{size}) | |
1599 | Allocate a block of @var{size} bytes, starting on a page boundary. | |
1600 | @xref{Aligned Memory Blocks}. | |
1601 | ||
5764c27f WN |
1602 | @item void *aligned_alloc (size_t @var{size}, size_t @var{alignment}) |
1603 | Allocate a block of @var{size} bytes, starting on an address that is a | |
1604 | multiple of @var{alignment}. @xref{Aligned Memory Blocks}. | |
1605 | ||
0a096e44 WN |
1606 | @item int posix_memalign (void **@var{memptr}, size_t @var{alignment}, size_t @var{size}) |
1607 | Allocate a block of @var{size} bytes, starting on an address that is a | |
1608 | multiple of @var{alignment}. @xref{Aligned Memory Blocks}. | |
1609 | ||
28f540f4 RM |
1610 | @item void *memalign (size_t @var{size}, size_t @var{boundary}) |
1611 | Allocate a block of @var{size} bytes, starting on an address that is a | |
1612 | multiple of @var{boundary}. @xref{Aligned Memory Blocks}. | |
1613 | ||
c131718c | 1614 | @item int mallopt (int @var{param}, int @var{value}) |
8b7fb588 | 1615 | Adjust a tunable parameter. @xref{Malloc Tunable Parameters}. |
c131718c | 1616 | |
28f540f4 RM |
1617 | @item int mcheck (void (*@var{abortfn}) (void)) |
1618 | Tell @code{malloc} to perform occasional consistency checks on | |
1619 | dynamically allocated memory, and to call @var{abortfn} when an | |
1620 | inconsistency is found. @xref{Heap Consistency Checking}. | |
1621 | ||
18a3a9a3 | 1622 | @item void *(*__malloc_hook) (size_t @var{size}, const void *@var{caller}) |
28f540f4 RM |
1623 | A pointer to a function that @code{malloc} uses whenever it is called. |
1624 | ||
18a3a9a3 | 1625 | @item void *(*__realloc_hook) (void *@var{ptr}, size_t @var{size}, const void *@var{caller}) |
28f540f4 RM |
1626 | A pointer to a function that @code{realloc} uses whenever it is called. |
1627 | ||
18a3a9a3 | 1628 | @item void (*__free_hook) (void *@var{ptr}, const void *@var{caller}) |
28f540f4 RM |
1629 | A pointer to a function that @code{free} uses whenever it is called. |
1630 | ||
18a3a9a3 | 1631 | @item void (*__memalign_hook) (size_t @var{size}, size_t @var{alignment}, const void *@var{caller}) |
5764c27f WN |
1632 | A pointer to a function that @code{aligned_alloc}, @code{memalign}, |
1633 | @code{posix_memalign} and @code{valloc} use whenever they are called. | |
3cb07217 | 1634 | |
c131718c | 1635 | @item struct mallinfo mallinfo (void) |
28f540f4 RM |
1636 | Return information about the current dynamic memory usage. |
1637 | @xref{Statistics of Malloc}. | |
1638 | @end table | |
1639 | ||
bd355af0 | 1640 | @node Allocation Debugging |
99a20616 | 1641 | @subsection Allocation Debugging |
bd355af0 UD |
1642 | @cindex allocation debugging |
1643 | @cindex malloc debugger | |
1644 | ||
bc938d3d | 1645 | A complicated task when programming with languages which do not use |
bd355af0 | 1646 | garbage collected dynamic memory allocation is to find memory leaks. |
3ef569c7 | 1647 | Long running programs must ensure that dynamically allocated objects are |
bd355af0 UD |
1648 | freed at the end of their lifetime. If this does not happen the system |
1649 | runs out of memory, sooner or later. | |
1650 | ||
1f77f049 | 1651 | The @code{malloc} implementation in @theglibc{} provides some |
bc938d3d | 1652 | simple means to detect such leaks and obtain some information to find |
bd355af0 UD |
1653 | the location. To do this the application must be started in a special |
1654 | mode which is enabled by an environment variable. There are no speed | |
bc938d3d | 1655 | penalties for the program if the debugging mode is not enabled. |
bd355af0 UD |
1656 | |
1657 | @menu | |
1658 | * Tracing malloc:: How to install the tracing functionality. | |
1659 | * Using the Memory Debugger:: Example programs excerpts. | |
1660 | * Tips for the Memory Debugger:: Some more or less clever ideas. | |
1661 | * Interpreting the traces:: What do all these lines mean? | |
1662 | @end menu | |
1663 | ||
1664 | @node Tracing malloc | |
99a20616 | 1665 | @subsubsection How to install the tracing functionality |
bd355af0 UD |
1666 | |
1667 | @comment mcheck.h | |
1668 | @comment GNU | |
1669 | @deftypefun void mtrace (void) | |
9f529d7c AO |
1670 | @safety{@prelim{}@mtunsafe{@mtsenv{} @mtasurace{:mtrace} @mtasuconst{:malloc_hooks} @mtuinit{}}@asunsafe{@asuinit{} @ascuheap{} @asucorrupt{} @asulock{}}@acunsafe{@acuinit{} @acucorrupt{} @aculock{} @acsfd{} @acsmem{}}} |
1671 | @c Like the mcheck hooks, these are not designed with thread safety in | |
1672 | @c mind, because the hook pointers are temporarily modified without | |
1673 | @c regard to other threads, signals or cancellation. | |
1674 | ||
1675 | @c mtrace @mtuinit @mtasurace:mtrace @mtsenv @asuinit @ascuheap @asucorrupt @acuinit @acucorrupt @aculock @acsfd @acsmem | |
1676 | @c __libc_secure_getenv dup @mtsenv | |
1677 | @c malloc dup @ascuheap @acsmem | |
1678 | @c fopen dup @ascuheap @asulock @aculock @acsmem @acsfd | |
1679 | @c fcntl dup ok | |
1680 | @c setvbuf dup @aculock | |
1681 | @c fprintf dup (on newly-created stream) @aculock | |
1682 | @c __cxa_atexit (once) dup @asulock @aculock @acsmem | |
1683 | @c free dup @ascuheap @acsmem | |
bd355af0 UD |
1684 | When the @code{mtrace} function is called it looks for an environment |
1685 | variable named @code{MALLOC_TRACE}. This variable is supposed to | |
1686 | contain a valid file name. The user must have write access. If the | |
1687 | file already exists it is truncated. If the environment variable is not | |
1688 | set or it does not name a valid file which can be opened for writing | |
0bc93a2f | 1689 | nothing is done. The behavior of @code{malloc} etc. is not changed. |
bc938d3d UD |
1690 | For obvious reasons this also happens if the application is installed |
1691 | with the SUID or SGID bit set. | |
bd355af0 | 1692 | |
e8b1163e | 1693 | If the named file is successfully opened, @code{mtrace} installs special |
bd355af0 | 1694 | handlers for the functions @code{malloc}, @code{realloc}, and |
e8b1163e | 1695 | @code{free} (@pxref{Hooks for Malloc}). From then on, all uses of these |
bd355af0 | 1696 | functions are traced and protocolled into the file. There is now of |
bc938d3d | 1697 | course a speed penalty for all calls to the traced functions so tracing |
e8b1163e | 1698 | should not be enabled during normal use. |
bd355af0 UD |
1699 | |
1700 | This function is a GNU extension and generally not available on other | |
1701 | systems. The prototype can be found in @file{mcheck.h}. | |
1702 | @end deftypefun | |
1703 | ||
1704 | @comment mcheck.h | |
1705 | @comment GNU | |
1706 | @deftypefun void muntrace (void) | |
9f529d7c AO |
1707 | @safety{@prelim{}@mtunsafe{@mtasurace{:mtrace} @mtasuconst{:malloc_hooks} @mtslocale{}}@asunsafe{@asucorrupt{} @ascuheap{}}@acunsafe{@acucorrupt{} @acsmem{} @aculock{} @acsfd{}}} |
1708 | ||
1709 | @c muntrace @mtasurace:mtrace @mtslocale @asucorrupt @ascuheap @acucorrupt @acsmem @aculock @acsfd | |
1710 | @c fprintf (fputs) dup @mtslocale @asucorrupt @ascuheap @acsmem @aculock @acucorrupt | |
1711 | @c fclose dup @ascuheap @asulock @aculock @acsmem @acsfd | |
bd355af0 | 1712 | The @code{muntrace} function can be called after @code{mtrace} was used |
0bc93a2f | 1713 | to enable tracing the @code{malloc} calls. If no (successful) call of |
bd355af0 UD |
1714 | @code{mtrace} was made @code{muntrace} does nothing. |
1715 | ||
1716 | Otherwise it deinstalls the handlers for @code{malloc}, @code{realloc}, | |
1717 | and @code{free} and then closes the protocol file. No calls are | |
bc938d3d | 1718 | protocolled anymore and the program runs again at full speed. |
bd355af0 UD |
1719 | |
1720 | This function is a GNU extension and generally not available on other | |
1721 | systems. The prototype can be found in @file{mcheck.h}. | |
1722 | @end deftypefun | |
1723 | ||
1724 | @node Using the Memory Debugger | |
99a20616 | 1725 | @subsubsection Example program excerpts |
bd355af0 UD |
1726 | |
1727 | Even though the tracing functionality does not influence the runtime | |
0bc93a2f | 1728 | behavior of the program it is not a good idea to call @code{mtrace} in |
bc938d3d UD |
1729 | all programs. Just imagine that you debug a program using @code{mtrace} |
1730 | and all other programs used in the debugging session also trace their | |
1731 | @code{malloc} calls. The output file would be the same for all programs | |
1732 | and thus is unusable. Therefore one should call @code{mtrace} only if | |
1733 | compiled for debugging. A program could therefore start like this: | |
bd355af0 UD |
1734 | |
1735 | @example | |
1736 | #include <mcheck.h> | |
1737 | ||
1738 | int | |
1739 | main (int argc, char *argv[]) | |
1740 | @{ | |
1741 | #ifdef DEBUGGING | |
1742 | mtrace (); | |
1743 | #endif | |
1744 | @dots{} | |
1745 | @} | |
1746 | @end example | |
1747 | ||
3ef569c7 | 1748 | This is all that is needed if you want to trace the calls during the |
bd355af0 UD |
1749 | whole runtime of the program. Alternatively you can stop the tracing at |
1750 | any time with a call to @code{muntrace}. It is even possible to restart | |
bc938d3d UD |
1751 | the tracing again with a new call to @code{mtrace}. But this can cause |
1752 | unreliable results since there may be calls of the functions which are | |
1753 | not called. Please note that not only the application uses the traced | |
1754 | functions, also libraries (including the C library itself) use these | |
1755 | functions. | |
bd355af0 | 1756 | |
3ef569c7 RJ |
1757 | This last point is also why it is not a good idea to call @code{muntrace} |
1758 | before the program terminates. The libraries are informed about the | |
bd355af0 UD |
1759 | termination of the program only after the program returns from |
1760 | @code{main} or calls @code{exit} and so cannot free the memory they use | |
1761 | before this time. | |
1762 | ||
1763 | So the best thing one can do is to call @code{mtrace} as the very first | |
1764 | function in the program and never call @code{muntrace}. So the program | |
1765 | traces almost all uses of the @code{malloc} functions (except those | |
1766 | calls which are executed by constructors of the program or used | |
1767 | libraries). | |
1768 | ||
1769 | @node Tips for the Memory Debugger | |
99a20616 | 1770 | @subsubsection Some more or less clever ideas |
bd355af0 UD |
1771 | |
1772 | You know the situation. The program is prepared for debugging and in | |
1773 | all debugging sessions it runs well. But once it is started without | |
bc938d3d UD |
1774 | debugging the error shows up. A typical example is a memory leak that |
1775 | becomes visible only when we turn off the debugging. If you foresee | |
1776 | such situations you can still win. Simply use something equivalent to | |
1777 | the following little program: | |
bd355af0 UD |
1778 | |
1779 | @example | |
1780 | #include <mcheck.h> | |
1781 | #include <signal.h> | |
1782 | ||
1783 | static void | |
1784 | enable (int sig) | |
1785 | @{ | |
1786 | mtrace (); | |
1787 | signal (SIGUSR1, enable); | |
1788 | @} | |
1789 | ||
1790 | static void | |
1791 | disable (int sig) | |
1792 | @{ | |
1793 | muntrace (); | |
1794 | signal (SIGUSR2, disable); | |
1795 | @} | |
1796 | ||
1797 | int | |
1798 | main (int argc, char *argv[]) | |
1799 | @{ | |
1800 | @dots{} | |
1801 | ||
1802 | signal (SIGUSR1, enable); | |
1803 | signal (SIGUSR2, disable); | |
1804 | ||
1805 | @dots{} | |
1806 | @} | |
1807 | @end example | |
1808 | ||
9756dfe1 | 1809 | I.e., the user can start the memory debugger any time s/he wants if the |
bd355af0 UD |
1810 | program was started with @code{MALLOC_TRACE} set in the environment. |
1811 | The output will of course not show the allocations which happened before | |
1812 | the first signal but if there is a memory leak this will show up | |
1813 | nevertheless. | |
1814 | ||
1815 | @node Interpreting the traces | |
99a20616 | 1816 | @subsubsection Interpreting the traces |
bd355af0 UD |
1817 | |
1818 | If you take a look at the output it will look similar to this: | |
1819 | ||
1820 | @example | |
1821 | = Start | |
1822 | @ [0x8048209] - 0x8064cc8 | |
1823 | @ [0x8048209] - 0x8064ce0 | |
1824 | @ [0x8048209] - 0x8064cf8 | |
1825 | @ [0x80481eb] + 0x8064c48 0x14 | |
1826 | @ [0x80481eb] + 0x8064c60 0x14 | |
1827 | @ [0x80481eb] + 0x8064c78 0x14 | |
1828 | @ [0x80481eb] + 0x8064c90 0x14 | |
1829 | = End | |
1830 | @end example | |
1831 | ||
1832 | What this all means is not really important since the trace file is not | |
bc938d3d | 1833 | meant to be read by a human. Therefore no attention is given to |
1f77f049 JM |
1834 | readability. Instead there is a program which comes with @theglibc{} |
1835 | which interprets the traces and outputs a summary in an | |
bd355af0 UD |
1836 | user-friendly way. The program is called @code{mtrace} (it is in fact a |
1837 | Perl script) and it takes one or two arguments. In any case the name of | |
bc938d3d UD |
1838 | the file with the trace output must be specified. If an optional |
1839 | argument precedes the name of the trace file this must be the name of | |
1840 | the program which generated the trace. | |
bd355af0 UD |
1841 | |
1842 | @example | |
1843 | drepper$ mtrace tst-mtrace log | |
1844 | No memory leaks. | |
1845 | @end example | |
1846 | ||
1847 | In this case the program @code{tst-mtrace} was run and it produced a | |
1848 | trace file @file{log}. The message printed by @code{mtrace} shows there | |
1849 | are no problems with the code, all allocated memory was freed | |
1850 | afterwards. | |
1851 | ||
1852 | If we call @code{mtrace} on the example trace given above we would get a | |
1853 | different outout: | |
1854 | ||
1855 | @example | |
1856 | drepper$ mtrace errlog | |
1857 | - 0x08064cc8 Free 2 was never alloc'd 0x8048209 | |
1858 | - 0x08064ce0 Free 3 was never alloc'd 0x8048209 | |
1859 | - 0x08064cf8 Free 4 was never alloc'd 0x8048209 | |
1860 | ||
1861 | Memory not freed: | |
1862 | ----------------- | |
1863 | Address Size Caller | |
1864 | 0x08064c48 0x14 at 0x80481eb | |
1865 | 0x08064c60 0x14 at 0x80481eb | |
1866 | 0x08064c78 0x14 at 0x80481eb | |
1867 | 0x08064c90 0x14 at 0x80481eb | |
1868 | @end example | |
1869 | ||
1870 | We have called @code{mtrace} with only one argument and so the script | |
1871 | has no chance to find out what is meant with the addresses given in the | |
1872 | trace. We can do better: | |
1873 | ||
1874 | @example | |
bc938d3d UD |
1875 | drepper$ mtrace tst errlog |
1876 | - 0x08064cc8 Free 2 was never alloc'd /home/drepper/tst.c:39 | |
1877 | - 0x08064ce0 Free 3 was never alloc'd /home/drepper/tst.c:39 | |
1878 | - 0x08064cf8 Free 4 was never alloc'd /home/drepper/tst.c:39 | |
bd355af0 UD |
1879 | |
1880 | Memory not freed: | |
1881 | ----------------- | |
1882 | Address Size Caller | |
bc938d3d UD |
1883 | 0x08064c48 0x14 at /home/drepper/tst.c:33 |
1884 | 0x08064c60 0x14 at /home/drepper/tst.c:33 | |
1885 | 0x08064c78 0x14 at /home/drepper/tst.c:33 | |
1886 | 0x08064c90 0x14 at /home/drepper/tst.c:33 | |
bd355af0 UD |
1887 | @end example |
1888 | ||
1889 | Suddenly the output makes much more sense and the user can see | |
1890 | immediately where the function calls causing the trouble can be found. | |
1891 | ||
9756dfe1 UD |
1892 | Interpreting this output is not complicated. There are at most two |
1893 | different situations being detected. First, @code{free} was called for | |
1894 | pointers which were never returned by one of the allocation functions. | |
bc938d3d | 1895 | This is usually a very bad problem and what this looks like is shown in |
9756dfe1 UD |
1896 | the first three lines of the output. Situations like this are quite |
1897 | rare and if they appear they show up very drastically: the program | |
1898 | normally crashes. | |
1899 | ||
1900 | The other situation which is much harder to detect are memory leaks. As | |
1901 | you can see in the output the @code{mtrace} function collects all this | |
1902 | information and so can say that the program calls an allocation function | |
1903 | from line 33 in the source file @file{/home/drepper/tst-mtrace.c} four | |
1904 | times without freeing this memory before the program terminates. | |
bc938d3d | 1905 | Whether this is a real problem remains to be investigated. |
9756dfe1 | 1906 | |
28f540f4 | 1907 | @node Obstacks |
99a20616 | 1908 | @subsection Obstacks |
28f540f4 RM |
1909 | @cindex obstacks |
1910 | ||
1911 | An @dfn{obstack} is a pool of memory containing a stack of objects. You | |
1912 | can create any number of separate obstacks, and then allocate objects in | |
1913 | specified obstacks. Within each obstack, the last object allocated must | |
1914 | always be the first one freed, but distinct obstacks are independent of | |
1915 | each other. | |
1916 | ||
1917 | Aside from this one constraint of order of freeing, obstacks are totally | |
1918 | general: an obstack can contain any number of objects of any size. They | |
1919 | are implemented with macros, so allocation is usually very fast as long as | |
1920 | the objects are usually small. And the only space overhead per object is | |
1921 | the padding needed to start each object on a suitable boundary. | |
1922 | ||
1923 | @menu | |
1924 | * Creating Obstacks:: How to declare an obstack in your program. | |
1925 | * Preparing for Obstacks:: Preparations needed before you can | |
1926 | use obstacks. | |
1927 | * Allocation in an Obstack:: Allocating objects in an obstack. | |
1928 | * Freeing Obstack Objects:: Freeing objects in an obstack. | |
1929 | * Obstack Functions:: The obstack functions are both | |
1930 | functions and macros. | |
1931 | * Growing Objects:: Making an object bigger by stages. | |
1932 | * Extra Fast Growing:: Extra-high-efficiency (though more | |
1933 | complicated) growing objects. | |
1934 | * Status of an Obstack:: Inquiries about the status of an obstack. | |
1935 | * Obstacks Data Alignment:: Controlling alignment of objects in obstacks. | |
1936 | * Obstack Chunks:: How obstacks obtain and release chunks; | |
1937 | efficiency considerations. | |
a5113b14 | 1938 | * Summary of Obstacks:: |
28f540f4 RM |
1939 | @end menu |
1940 | ||
1941 | @node Creating Obstacks | |
99a20616 | 1942 | @subsubsection Creating Obstacks |
28f540f4 RM |
1943 | |
1944 | The utilities for manipulating obstacks are declared in the header | |
1945 | file @file{obstack.h}. | |
1946 | @pindex obstack.h | |
1947 | ||
1948 | @comment obstack.h | |
1949 | @comment GNU | |
1950 | @deftp {Data Type} {struct obstack} | |
1951 | An obstack is represented by a data structure of type @code{struct | |
1952 | obstack}. This structure has a small fixed size; it records the status | |
1953 | of the obstack and how to find the space in which objects are allocated. | |
1954 | It does not contain any of the objects themselves. You should not try | |
1955 | to access the contents of the structure directly; use only the functions | |
1956 | described in this chapter. | |
1957 | @end deftp | |
1958 | ||
1959 | You can declare variables of type @code{struct obstack} and use them as | |
1960 | obstacks, or you can allocate obstacks dynamically like any other kind | |
1961 | of object. Dynamic allocation of obstacks allows your program to have a | |
1962 | variable number of different stacks. (You can even allocate an | |
1963 | obstack structure in another obstack, but this is rarely useful.) | |
1964 | ||
1965 | All the functions that work with obstacks require you to specify which | |
1966 | obstack to use. You do this with a pointer of type @code{struct obstack | |
1967 | *}. In the following, we often say ``an obstack'' when strictly | |
1968 | speaking the object at hand is such a pointer. | |
1969 | ||
1970 | The objects in the obstack are packed into large blocks called | |
1971 | @dfn{chunks}. The @code{struct obstack} structure points to a chain of | |
1972 | the chunks currently in use. | |
1973 | ||
1974 | The obstack library obtains a new chunk whenever you allocate an object | |
1975 | that won't fit in the previous chunk. Since the obstack library manages | |
1976 | chunks automatically, you don't need to pay much attention to them, but | |
1977 | you do need to supply a function which the obstack library should use to | |
1978 | get a chunk. Usually you supply a function which uses @code{malloc} | |
1979 | directly or indirectly. You must also supply a function to free a chunk. | |
1980 | These matters are described in the following section. | |
1981 | ||
1982 | @node Preparing for Obstacks | |
99a20616 | 1983 | @subsubsection Preparing for Using Obstacks |
28f540f4 RM |
1984 | |
1985 | Each source file in which you plan to use the obstack functions | |
1986 | must include the header file @file{obstack.h}, like this: | |
1987 | ||
1988 | @smallexample | |
1989 | #include <obstack.h> | |
1990 | @end smallexample | |
1991 | ||
1992 | @findex obstack_chunk_alloc | |
1993 | @findex obstack_chunk_free | |
1994 | Also, if the source file uses the macro @code{obstack_init}, it must | |
1995 | declare or define two functions or macros that will be called by the | |
1996 | obstack library. One, @code{obstack_chunk_alloc}, is used to allocate | |
1997 | the chunks of memory into which objects are packed. The other, | |
1998 | @code{obstack_chunk_free}, is used to return chunks when the objects in | |
1999 | them are freed. These macros should appear before any use of obstacks | |
2000 | in the source file. | |
2001 | ||
2002 | Usually these are defined to use @code{malloc} via the intermediary | |
2003 | @code{xmalloc} (@pxref{Unconstrained Allocation}). This is done with | |
2004 | the following pair of macro definitions: | |
2005 | ||
2006 | @smallexample | |
2007 | #define obstack_chunk_alloc xmalloc | |
2008 | #define obstack_chunk_free free | |
2009 | @end smallexample | |
2010 | ||
2011 | @noindent | |
99a20616 | 2012 | Though the memory you get using obstacks really comes from @code{malloc}, |
28f540f4 RM |
2013 | using obstacks is faster because @code{malloc} is called less often, for |
2014 | larger blocks of memory. @xref{Obstack Chunks}, for full details. | |
2015 | ||
2016 | At run time, before the program can use a @code{struct obstack} object | |
2017 | as an obstack, it must initialize the obstack by calling | |
2018 | @code{obstack_init}. | |
2019 | ||
2020 | @comment obstack.h | |
2021 | @comment GNU | |
2022 | @deftypefun int obstack_init (struct obstack *@var{obstack-ptr}) | |
9f529d7c AO |
2023 | @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acsafe{@acsmem{}}} |
2024 | @c obstack_init @mtsrace:obstack-ptr @acsmem | |
2025 | @c _obstack_begin @acsmem | |
2026 | @c chunkfun = obstack_chunk_alloc (suggested malloc) | |
2027 | @c freefun = obstack_chunk_free (suggested free) | |
2028 | @c *chunkfun @acsmem | |
2029 | @c obstack_chunk_alloc user-supplied | |
2030 | @c *obstack_alloc_failed_handler user-supplied | |
2031 | @c -> print_and_abort (default) | |
2032 | @c | |
2033 | @c print_and_abort | |
2034 | @c _ dup @ascuintl | |
2035 | @c fxprintf dup @asucorrupt @aculock @acucorrupt | |
2036 | @c exit @acucorrupt? | |
28f540f4 | 2037 | Initialize obstack @var{obstack-ptr} for allocation of objects. This |
3cb07217 UD |
2038 | function calls the obstack's @code{obstack_chunk_alloc} function. If |
2039 | allocation of memory fails, the function pointed to by | |
2040 | @code{obstack_alloc_failed_handler} is called. The @code{obstack_init} | |
2041 | function always returns 1 (Compatibility notice: Former versions of | |
2042 | obstack returned 0 if allocation failed). | |
28f540f4 RM |
2043 | @end deftypefun |
2044 | ||
2045 | Here are two examples of how to allocate the space for an obstack and | |
2046 | initialize it. First, an obstack that is a static variable: | |
2047 | ||
2048 | @smallexample | |
2049 | static struct obstack myobstack; | |
2050 | @dots{} | |
2051 | obstack_init (&myobstack); | |
2052 | @end smallexample | |
2053 | ||
2054 | @noindent | |
2055 | Second, an obstack that is itself dynamically allocated: | |
2056 | ||
2057 | @smallexample | |
2058 | struct obstack *myobstack_ptr | |
2059 | = (struct obstack *) xmalloc (sizeof (struct obstack)); | |
2060 | ||
2061 | obstack_init (myobstack_ptr); | |
2062 | @end smallexample | |
2063 | ||
3cb07217 UD |
2064 | @comment obstack.h |
2065 | @comment GNU | |
2066 | @defvar obstack_alloc_failed_handler | |
2067 | The value of this variable is a pointer to a function that | |
2068 | @code{obstack} uses when @code{obstack_chunk_alloc} fails to allocate | |
2069 | memory. The default action is to print a message and abort. | |
2070 | You should supply a function that either calls @code{exit} | |
2071 | (@pxref{Program Termination}) or @code{longjmp} (@pxref{Non-Local | |
2072 | Exits}) and doesn't return. | |
2073 | ||
2074 | @smallexample | |
2075 | void my_obstack_alloc_failed (void) | |
2076 | @dots{} | |
2077 | obstack_alloc_failed_handler = &my_obstack_alloc_failed; | |
2078 | @end smallexample | |
2079 | ||
2080 | @end defvar | |
2081 | ||
28f540f4 | 2082 | @node Allocation in an Obstack |
99a20616 | 2083 | @subsubsection Allocation in an Obstack |
28f540f4 RM |
2084 | @cindex allocation (obstacks) |
2085 | ||
2086 | The most direct way to allocate an object in an obstack is with | |
2087 | @code{obstack_alloc}, which is invoked almost like @code{malloc}. | |
2088 | ||
2089 | @comment obstack.h | |
2090 | @comment GNU | |
2091 | @deftypefun {void *} obstack_alloc (struct obstack *@var{obstack-ptr}, int @var{size}) | |
9f529d7c AO |
2092 | @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{} @acsmem{}}} |
2093 | @c obstack_alloc @mtsrace:obstack-ptr @acucorrupt @acsmem | |
2094 | @c obstack_blank dup @mtsrace:obstack-ptr @acucorrupt @acsmem | |
2095 | @c obstack_finish dup @mtsrace:obstack-ptr @acucorrupt | |
28f540f4 RM |
2096 | This allocates an uninitialized block of @var{size} bytes in an obstack |
2097 | and returns its address. Here @var{obstack-ptr} specifies which obstack | |
2098 | to allocate the block in; it is the address of the @code{struct obstack} | |
2099 | object which represents the obstack. Each obstack function or macro | |
2100 | requires you to specify an @var{obstack-ptr} as the first argument. | |
2101 | ||
2102 | This function calls the obstack's @code{obstack_chunk_alloc} function if | |
3cb07217 UD |
2103 | it needs to allocate a new chunk of memory; it calls |
2104 | @code{obstack_alloc_failed_handler} if allocation of memory by | |
2105 | @code{obstack_chunk_alloc} failed. | |
28f540f4 RM |
2106 | @end deftypefun |
2107 | ||
2108 | For example, here is a function that allocates a copy of a string @var{str} | |
2109 | in a specific obstack, which is in the variable @code{string_obstack}: | |
2110 | ||
2111 | @smallexample | |
2112 | struct obstack string_obstack; | |
2113 | ||
2114 | char * | |
2115 | copystring (char *string) | |
2116 | @{ | |
7cc27f44 UD |
2117 | size_t len = strlen (string) + 1; |
2118 | char *s = (char *) obstack_alloc (&string_obstack, len); | |
2119 | memcpy (s, string, len); | |
28f540f4 RM |
2120 | return s; |
2121 | @} | |
2122 | @end smallexample | |
2123 | ||
2124 | To allocate a block with specified contents, use the function | |
2125 | @code{obstack_copy}, declared like this: | |
2126 | ||
2127 | @comment obstack.h | |
2128 | @comment GNU | |
2129 | @deftypefun {void *} obstack_copy (struct obstack *@var{obstack-ptr}, void *@var{address}, int @var{size}) | |
9f529d7c AO |
2130 | @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{} @acsmem{}}} |
2131 | @c obstack_copy @mtsrace:obstack-ptr @acucorrupt @acsmem | |
2132 | @c obstack_grow dup @mtsrace:obstack-ptr @acucorrupt @acsmem | |
2133 | @c obstack_finish dup @mtsrace:obstack-ptr @acucorrupt | |
28f540f4 | 2134 | This allocates a block and initializes it by copying @var{size} |
3cb07217 UD |
2135 | bytes of data starting at @var{address}. It calls |
2136 | @code{obstack_alloc_failed_handler} if allocation of memory by | |
2137 | @code{obstack_chunk_alloc} failed. | |
28f540f4 RM |
2138 | @end deftypefun |
2139 | ||
2140 | @comment obstack.h | |
2141 | @comment GNU | |
2142 | @deftypefun {void *} obstack_copy0 (struct obstack *@var{obstack-ptr}, void *@var{address}, int @var{size}) | |
9f529d7c AO |
2143 | @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{} @acsmem{}}} |
2144 | @c obstack_copy0 @mtsrace:obstack-ptr @acucorrupt @acsmem | |
2145 | @c obstack_grow0 dup @mtsrace:obstack-ptr @acucorrupt @acsmem | |
2146 | @c obstack_finish dup @mtsrace:obstack-ptr @acucorrupt | |
28f540f4 RM |
2147 | Like @code{obstack_copy}, but appends an extra byte containing a null |
2148 | character. This extra byte is not counted in the argument @var{size}. | |
2149 | @end deftypefun | |
2150 | ||
2151 | The @code{obstack_copy0} function is convenient for copying a sequence | |
2152 | of characters into an obstack as a null-terminated string. Here is an | |
2153 | example of its use: | |
2154 | ||
2155 | @smallexample | |
2156 | char * | |
2157 | obstack_savestring (char *addr, int size) | |
2158 | @{ | |
2159 | return obstack_copy0 (&myobstack, addr, size); | |
2160 | @} | |
2161 | @end smallexample | |
2162 | ||
2163 | @noindent | |
2164 | Contrast this with the previous example of @code{savestring} using | |
2165 | @code{malloc} (@pxref{Basic Allocation}). | |
2166 | ||
2167 | @node Freeing Obstack Objects | |
99a20616 | 2168 | @subsubsection Freeing Objects in an Obstack |
28f540f4 RM |
2169 | @cindex freeing (obstacks) |
2170 | ||
2171 | To free an object allocated in an obstack, use the function | |
2172 | @code{obstack_free}. Since the obstack is a stack of objects, freeing | |
2173 | one object automatically frees all other objects allocated more recently | |
2174 | in the same obstack. | |
2175 | ||
2176 | @comment obstack.h | |
2177 | @comment GNU | |
2178 | @deftypefun void obstack_free (struct obstack *@var{obstack-ptr}, void *@var{object}) | |
9f529d7c AO |
2179 | @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{}}} |
2180 | @c obstack_free @mtsrace:obstack-ptr @acucorrupt | |
2181 | @c (obstack_free) @mtsrace:obstack-ptr @acucorrupt | |
2182 | @c *freefun dup user-supplied | |
28f540f4 RM |
2183 | If @var{object} is a null pointer, everything allocated in the obstack |
2184 | is freed. Otherwise, @var{object} must be the address of an object | |
2185 | allocated in the obstack. Then @var{object} is freed, along with | |
3ef569c7 | 2186 | everything allocated in @var{obstack-ptr} since @var{object}. |
28f540f4 RM |
2187 | @end deftypefun |
2188 | ||
2189 | Note that if @var{object} is a null pointer, the result is an | |
99a20616 | 2190 | uninitialized obstack. To free all memory in an obstack but leave it |
28f540f4 RM |
2191 | valid for further allocation, call @code{obstack_free} with the address |
2192 | of the first object allocated on the obstack: | |
2193 | ||
2194 | @smallexample | |
2195 | obstack_free (obstack_ptr, first_object_allocated_ptr); | |
2196 | @end smallexample | |
2197 | ||
2198 | Recall that the objects in an obstack are grouped into chunks. When all | |
2199 | the objects in a chunk become free, the obstack library automatically | |
2200 | frees the chunk (@pxref{Preparing for Obstacks}). Then other | |
2201 | obstacks, or non-obstack allocation, can reuse the space of the chunk. | |
2202 | ||
2203 | @node Obstack Functions | |
99a20616 | 2204 | @subsubsection Obstack Functions and Macros |
28f540f4 RM |
2205 | @cindex macros |
2206 | ||
2207 | The interfaces for using obstacks may be defined either as functions or | |
2208 | as macros, depending on the compiler. The obstack facility works with | |
f65fd747 | 2209 | all C compilers, including both @w{ISO C} and traditional C, but there are |
28f540f4 RM |
2210 | precautions you must take if you plan to use compilers other than GNU C. |
2211 | ||
f65fd747 | 2212 | If you are using an old-fashioned @w{non-ISO C} compiler, all the obstack |
28f540f4 RM |
2213 | ``functions'' are actually defined only as macros. You can call these |
2214 | macros like functions, but you cannot use them in any other way (for | |
2215 | example, you cannot take their address). | |
2216 | ||
2217 | Calling the macros requires a special precaution: namely, the first | |
2218 | operand (the obstack pointer) may not contain any side effects, because | |
2219 | it may be computed more than once. For example, if you write this: | |
2220 | ||
2221 | @smallexample | |
2222 | obstack_alloc (get_obstack (), 4); | |
2223 | @end smallexample | |
2224 | ||
2225 | @noindent | |
2226 | you will find that @code{get_obstack} may be called several times. | |
2227 | If you use @code{*obstack_list_ptr++} as the obstack pointer argument, | |
2228 | you will get very strange results since the incrementation may occur | |
2229 | several times. | |
2230 | ||
f65fd747 | 2231 | In @w{ISO C}, each function has both a macro definition and a function |
28f540f4 RM |
2232 | definition. The function definition is used if you take the address of the |
2233 | function without calling it. An ordinary call uses the macro definition by | |
2234 | default, but you can request the function definition instead by writing the | |
2235 | function name in parentheses, as shown here: | |
2236 | ||
2237 | @smallexample | |
2238 | char *x; | |
2239 | void *(*funcp) (); | |
2240 | /* @r{Use the macro}. */ | |
2241 | x = (char *) obstack_alloc (obptr, size); | |
2242 | /* @r{Call the function}. */ | |
2243 | x = (char *) (obstack_alloc) (obptr, size); | |
2244 | /* @r{Take the address of the function}. */ | |
2245 | funcp = obstack_alloc; | |
2246 | @end smallexample | |
2247 | ||
2248 | @noindent | |
f65fd747 | 2249 | This is the same situation that exists in @w{ISO C} for the standard library |
28f540f4 RM |
2250 | functions. @xref{Macro Definitions}. |
2251 | ||
2252 | @strong{Warning:} When you do use the macros, you must observe the | |
f65fd747 | 2253 | precaution of avoiding side effects in the first operand, even in @w{ISO C}. |
28f540f4 RM |
2254 | |
2255 | If you use the GNU C compiler, this precaution is not necessary, because | |
2256 | various language extensions in GNU C permit defining the macros so as to | |
2257 | compute each argument only once. | |
2258 | ||
2259 | @node Growing Objects | |
99a20616 | 2260 | @subsubsection Growing Objects |
28f540f4 RM |
2261 | @cindex growing objects (in obstacks) |
2262 | @cindex changing the size of a block (obstacks) | |
2263 | ||
99a20616 | 2264 | Because memory in obstack chunks is used sequentially, it is possible to |
28f540f4 RM |
2265 | build up an object step by step, adding one or more bytes at a time to the |
2266 | end of the object. With this technique, you do not need to know how much | |
2267 | data you will put in the object until you come to the end of it. We call | |
2268 | this the technique of @dfn{growing objects}. The special functions | |
2269 | for adding data to the growing object are described in this section. | |
2270 | ||
2271 | You don't need to do anything special when you start to grow an object. | |
2272 | Using one of the functions to add data to the object automatically | |
2273 | starts it. However, it is necessary to say explicitly when the object is | |
2274 | finished. This is done with the function @code{obstack_finish}. | |
2275 | ||
2276 | The actual address of the object thus built up is not known until the | |
2277 | object is finished. Until then, it always remains possible that you will | |
2278 | add so much data that the object must be copied into a new chunk. | |
2279 | ||
2280 | While the obstack is in use for a growing object, you cannot use it for | |
2281 | ordinary allocation of another object. If you try to do so, the space | |
2282 | already added to the growing object will become part of the other object. | |
2283 | ||
2284 | @comment obstack.h | |
2285 | @comment GNU | |
2286 | @deftypefun void obstack_blank (struct obstack *@var{obstack-ptr}, int @var{size}) | |
9f529d7c AO |
2287 | @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{} @acsmem{}}} |
2288 | @c obstack_blank @mtsrace:obstack-ptr @acucorrupt @acsmem | |
2289 | @c _obstack_newchunk @mtsrace:obstack-ptr @acucorrupt @acsmem | |
2290 | @c *chunkfun dup @acsmem | |
2291 | @c *obstack_alloc_failed_handler dup user-supplied | |
2292 | @c *freefun | |
2293 | @c obstack_blank_fast dup @mtsrace:obstack-ptr | |
28f540f4 RM |
2294 | The most basic function for adding to a growing object is |
2295 | @code{obstack_blank}, which adds space without initializing it. | |
2296 | @end deftypefun | |
2297 | ||
2298 | @comment obstack.h | |
2299 | @comment GNU | |
2300 | @deftypefun void obstack_grow (struct obstack *@var{obstack-ptr}, void *@var{data}, int @var{size}) | |
9f529d7c AO |
2301 | @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{} @acsmem{}}} |
2302 | @c obstack_grow @mtsrace:obstack-ptr @acucorrupt @acsmem | |
2303 | @c _obstack_newchunk dup @mtsrace:obstack-ptr @acucorrupt @acsmem | |
2304 | @c memcpy ok | |
28f540f4 RM |
2305 | To add a block of initialized space, use @code{obstack_grow}, which is |
2306 | the growing-object analogue of @code{obstack_copy}. It adds @var{size} | |
2307 | bytes of data to the growing object, copying the contents from | |
2308 | @var{data}. | |
2309 | @end deftypefun | |
2310 | ||
2311 | @comment obstack.h | |
2312 | @comment GNU | |
2313 | @deftypefun void obstack_grow0 (struct obstack *@var{obstack-ptr}, void *@var{data}, int @var{size}) | |
9f529d7c AO |
2314 | @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{} @acsmem{}}} |
2315 | @c obstack_grow0 @mtsrace:obstack-ptr @acucorrupt @acsmem | |
2316 | @c (no sequence point between storing NUL and incrementing next_free) | |
2317 | @c (multiple changes to next_free => @acucorrupt) | |
2318 | @c _obstack_newchunk dup @mtsrace:obstack-ptr @acucorrupt @acsmem | |
2319 | @c memcpy ok | |
28f540f4 RM |
2320 | This is the growing-object analogue of @code{obstack_copy0}. It adds |
2321 | @var{size} bytes copied from @var{data}, followed by an additional null | |
2322 | character. | |
2323 | @end deftypefun | |
2324 | ||
2325 | @comment obstack.h | |
2326 | @comment GNU | |
2327 | @deftypefun void obstack_1grow (struct obstack *@var{obstack-ptr}, char @var{c}) | |
9f529d7c AO |
2328 | @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{} @acsmem{}}} |
2329 | @c obstack_1grow @mtsrace:obstack-ptr @acucorrupt @acsmem | |
2330 | @c _obstack_newchunk dup @mtsrace:obstack-ptr @acucorrupt @acsmem | |
2331 | @c obstack_1grow_fast dup @mtsrace:obstack-ptr @acucorrupt @acsmem | |
28f540f4 RM |
2332 | To add one character at a time, use the function @code{obstack_1grow}. |
2333 | It adds a single byte containing @var{c} to the growing object. | |
2334 | @end deftypefun | |
2335 | ||
2c6fe0bd UD |
2336 | @comment obstack.h |
2337 | @comment GNU | |
2338 | @deftypefun void obstack_ptr_grow (struct obstack *@var{obstack-ptr}, void *@var{data}) | |
9f529d7c AO |
2339 | @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{} @acsmem{}}} |
2340 | @c obstack_ptr_grow @mtsrace:obstack-ptr @acucorrupt @acsmem | |
2341 | @c _obstack_newchunk dup @mtsrace:obstack-ptr @acucorrupt @acsmem | |
2342 | @c obstack_ptr_grow_fast dup @mtsrace:obstack-ptr | |
2c6fe0bd UD |
2343 | Adding the value of a pointer one can use the function |
2344 | @code{obstack_ptr_grow}. It adds @code{sizeof (void *)} bytes | |
2345 | containing the value of @var{data}. | |
2346 | @end deftypefun | |
2347 | ||
2348 | @comment obstack.h | |
2349 | @comment GNU | |
2350 | @deftypefun void obstack_int_grow (struct obstack *@var{obstack-ptr}, int @var{data}) | |
9f529d7c AO |
2351 | @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{} @acsmem{}}} |
2352 | @c obstack_int_grow @mtsrace:obstack-ptr @acucorrupt @acsmem | |
2353 | @c _obstack_newchunk dup @mtsrace:obstack-ptr @acucorrupt @acsmem | |
2354 | @c obstack_int_grow_fast dup @mtsrace:obstack-ptr | |
2c6fe0bd UD |
2355 | A single value of type @code{int} can be added by using the |
2356 | @code{obstack_int_grow} function. It adds @code{sizeof (int)} bytes to | |
2357 | the growing object and initializes them with the value of @var{data}. | |
2358 | @end deftypefun | |
2359 | ||
28f540f4 RM |
2360 | @comment obstack.h |
2361 | @comment GNU | |
2362 | @deftypefun {void *} obstack_finish (struct obstack *@var{obstack-ptr}) | |
9f529d7c AO |
2363 | @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{}}} |
2364 | @c obstack_finish @mtsrace:obstack-ptr @acucorrupt | |
28f540f4 RM |
2365 | When you are finished growing the object, use the function |
2366 | @code{obstack_finish} to close it off and return its final address. | |
2367 | ||
2368 | Once you have finished the object, the obstack is available for ordinary | |
2369 | allocation or for growing another object. | |
2370 | ||
2371 | This function can return a null pointer under the same conditions as | |
2372 | @code{obstack_alloc} (@pxref{Allocation in an Obstack}). | |
2373 | @end deftypefun | |
2374 | ||
2375 | When you build an object by growing it, you will probably need to know | |
2376 | afterward how long it became. You need not keep track of this as you grow | |
2377 | the object, because you can find out the length from the obstack just | |
2378 | before finishing the object with the function @code{obstack_object_size}, | |
2379 | declared as follows: | |
2380 | ||
2381 | @comment obstack.h | |
2382 | @comment GNU | |
2383 | @deftypefun int obstack_object_size (struct obstack *@var{obstack-ptr}) | |
9f529d7c | 2384 | @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acsafe{}} |
28f540f4 RM |
2385 | This function returns the current size of the growing object, in bytes. |
2386 | Remember to call this function @emph{before} finishing the object. | |
2387 | After it is finished, @code{obstack_object_size} will return zero. | |
2388 | @end deftypefun | |
2389 | ||
2390 | If you have started growing an object and wish to cancel it, you should | |
2391 | finish it and then free it, like this: | |
2392 | ||
2393 | @smallexample | |
2394 | obstack_free (obstack_ptr, obstack_finish (obstack_ptr)); | |
2395 | @end smallexample | |
2396 | ||
2397 | @noindent | |
2398 | This has no effect if no object was growing. | |
2399 | ||
2400 | @cindex shrinking objects | |
2401 | You can use @code{obstack_blank} with a negative size argument to make | |
2402 | the current object smaller. Just don't try to shrink it beyond zero | |
2403 | length---there's no telling what will happen if you do that. | |
2404 | ||
2405 | @node Extra Fast Growing | |
99a20616 | 2406 | @subsubsection Extra Fast Growing Objects |
28f540f4 RM |
2407 | @cindex efficiency and obstacks |
2408 | ||
2409 | The usual functions for growing objects incur overhead for checking | |
2410 | whether there is room for the new growth in the current chunk. If you | |
2411 | are frequently constructing objects in small steps of growth, this | |
2412 | overhead can be significant. | |
2413 | ||
2414 | You can reduce the overhead by using special ``fast growth'' | |
2415 | functions that grow the object without checking. In order to have a | |
2416 | robust program, you must do the checking yourself. If you do this checking | |
2417 | in the simplest way each time you are about to add data to the object, you | |
2418 | have not saved anything, because that is what the ordinary growth | |
2419 | functions do. But if you can arrange to check less often, or check | |
2420 | more efficiently, then you make the program faster. | |
2421 | ||
2422 | The function @code{obstack_room} returns the amount of room available | |
2423 | in the current chunk. It is declared as follows: | |
2424 | ||
2425 | @comment obstack.h | |
2426 | @comment GNU | |
2427 | @deftypefun int obstack_room (struct obstack *@var{obstack-ptr}) | |
9f529d7c | 2428 | @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acsafe{}} |
28f540f4 RM |
2429 | This returns the number of bytes that can be added safely to the current |
2430 | growing object (or to an object about to be started) in obstack | |
3ef569c7 | 2431 | @var{obstack-ptr} using the fast growth functions. |
28f540f4 RM |
2432 | @end deftypefun |
2433 | ||
2434 | While you know there is room, you can use these fast growth functions | |
2435 | for adding data to a growing object: | |
2436 | ||
2437 | @comment obstack.h | |
2438 | @comment GNU | |
2439 | @deftypefun void obstack_1grow_fast (struct obstack *@var{obstack-ptr}, char @var{c}) | |
9f529d7c AO |
2440 | @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acunsafe{@acucorrupt{} @acsmem{}}} |
2441 | @c obstack_1grow_fast @mtsrace:obstack-ptr @acucorrupt @acsmem | |
2442 | @c (no sequence point between copying c and incrementing next_free) | |
28f540f4 RM |
2443 | The function @code{obstack_1grow_fast} adds one byte containing the |
2444 | character @var{c} to the growing object in obstack @var{obstack-ptr}. | |
2445 | @end deftypefun | |
2446 | ||
2c6fe0bd UD |
2447 | @comment obstack.h |
2448 | @comment GNU | |
2449 | @deftypefun void obstack_ptr_grow_fast (struct obstack *@var{obstack-ptr}, void *@var{data}) | |
9f529d7c AO |
2450 | @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acsafe{}} |
2451 | @c obstack_ptr_grow_fast @mtsrace:obstack-ptr | |
2c6fe0bd UD |
2452 | The function @code{obstack_ptr_grow_fast} adds @code{sizeof (void *)} |
2453 | bytes containing the value of @var{data} to the growing object in | |
2454 | obstack @var{obstack-ptr}. | |
2455 | @end deftypefun | |
2456 | ||
2457 | @comment obstack.h | |
2458 | @comment GNU | |
2459 | @deftypefun void obstack_int_grow_fast (struct obstack *@var{obstack-ptr}, int @var{data}) | |
9f529d7c AO |
2460 | @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acsafe{}} |
2461 | @c obstack_int_grow_fast @mtsrace:obstack-ptr | |
2c6fe0bd UD |
2462 | The function @code{obstack_int_grow_fast} adds @code{sizeof (int)} bytes |
2463 | containing the value of @var{data} to the growing object in obstack | |
2464 | @var{obstack-ptr}. | |
2465 | @end deftypefun | |
2466 | ||
28f540f4 RM |
2467 | @comment obstack.h |
2468 | @comment GNU | |
2469 | @deftypefun void obstack_blank_fast (struct obstack *@var{obstack-ptr}, int @var{size}) | |
9f529d7c AO |
2470 | @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acsafe{}} |
2471 | @c obstack_blank_fast @mtsrace:obstack-ptr | |
28f540f4 RM |
2472 | The function @code{obstack_blank_fast} adds @var{size} bytes to the |
2473 | growing object in obstack @var{obstack-ptr} without initializing them. | |
2474 | @end deftypefun | |
2475 | ||
2476 | When you check for space using @code{obstack_room} and there is not | |
2477 | enough room for what you want to add, the fast growth functions | |
2478 | are not safe. In this case, simply use the corresponding ordinary | |
2479 | growth function instead. Very soon this will copy the object to a | |
a5113b14 | 2480 | new chunk; then there will be lots of room available again. |
28f540f4 RM |
2481 | |
2482 | So, each time you use an ordinary growth function, check afterward for | |
2483 | sufficient space using @code{obstack_room}. Once the object is copied | |
2484 | to a new chunk, there will be plenty of space again, so the program will | |
2485 | start using the fast growth functions again. | |
2486 | ||
2487 | Here is an example: | |
2488 | ||
2489 | @smallexample | |
2490 | @group | |
2491 | void | |
2492 | add_string (struct obstack *obstack, const char *ptr, int len) | |
2493 | @{ | |
2494 | while (len > 0) | |
2495 | @{ | |
2496 | int room = obstack_room (obstack); | |
2497 | if (room == 0) | |
2498 | @{ | |
cf822e3c | 2499 | /* @r{Not enough room. Add one character slowly,} |
28f540f4 RM |
2500 | @r{which may copy to a new chunk and make room.} */ |
2501 | obstack_1grow (obstack, *ptr++); | |
2502 | len--; | |
2503 | @} | |
a5113b14 | 2504 | else |
28f540f4 RM |
2505 | @{ |
2506 | if (room > len) | |
2507 | room = len; | |
2508 | /* @r{Add fast as much as we have room for.} */ | |
2509 | len -= room; | |
2510 | while (room-- > 0) | |
2511 | obstack_1grow_fast (obstack, *ptr++); | |
2512 | @} | |
2513 | @} | |
2514 | @} | |
2515 | @end group | |
2516 | @end smallexample | |
2517 | ||
2518 | @node Status of an Obstack | |
99a20616 | 2519 | @subsubsection Status of an Obstack |
28f540f4 RM |
2520 | @cindex obstack status |
2521 | @cindex status of obstack | |
2522 | ||
2523 | Here are functions that provide information on the current status of | |
2524 | allocation in an obstack. You can use them to learn about an object while | |
2525 | still growing it. | |
2526 | ||
2527 | @comment obstack.h | |
2528 | @comment GNU | |
2529 | @deftypefun {void *} obstack_base (struct obstack *@var{obstack-ptr}) | |
9f529d7c | 2530 | @safety{@prelim{}@mtsafe{}@asunsafe{@asucorrupt{}}@acsafe{}} |
28f540f4 RM |
2531 | This function returns the tentative address of the beginning of the |
2532 | currently growing object in @var{obstack-ptr}. If you finish the object | |
2533 | immediately, it will have that address. If you make it larger first, it | |
2534 | may outgrow the current chunk---then its address will change! | |
2535 | ||
2536 | If no object is growing, this value says where the next object you | |
2537 | allocate will start (once again assuming it fits in the current | |
2538 | chunk). | |
2539 | @end deftypefun | |
2540 | ||
2541 | @comment obstack.h | |
2542 | @comment GNU | |
2543 | @deftypefun {void *} obstack_next_free (struct obstack *@var{obstack-ptr}) | |
9f529d7c | 2544 | @safety{@prelim{}@mtsafe{}@asunsafe{@asucorrupt{}}@acsafe{}} |
28f540f4 RM |
2545 | This function returns the address of the first free byte in the current |
2546 | chunk of obstack @var{obstack-ptr}. This is the end of the currently | |
2547 | growing object. If no object is growing, @code{obstack_next_free} | |
2548 | returns the same value as @code{obstack_base}. | |
2549 | @end deftypefun | |
2550 | ||
2551 | @comment obstack.h | |
2552 | @comment GNU | |
2553 | @deftypefun int obstack_object_size (struct obstack *@var{obstack-ptr}) | |
9f529d7c AO |
2554 | @c dup |
2555 | @safety{@prelim{}@mtsafe{@mtsrace{:obstack-ptr}}@assafe{}@acsafe{}} | |
28f540f4 RM |
2556 | This function returns the size in bytes of the currently growing object. |
2557 | This is equivalent to | |
2558 | ||
2559 | @smallexample | |
2560 | obstack_next_free (@var{obstack-ptr}) - obstack_base (@var{obstack-ptr}) | |
2561 | @end smallexample | |
2562 | @end deftypefun | |
2563 | ||
2564 | @node Obstacks Data Alignment | |
99a20616 | 2565 | @subsubsection Alignment of Data in Obstacks |
28f540f4 RM |
2566 | @cindex alignment (in obstacks) |
2567 | ||
2568 | Each obstack has an @dfn{alignment boundary}; each object allocated in | |
2569 | the obstack automatically starts on an address that is a multiple of the | |
11883883 RM |
2570 | specified boundary. By default, this boundary is aligned so that |
2571 | the object can hold any type of data. | |
28f540f4 RM |
2572 | |
2573 | To access an obstack's alignment boundary, use the macro | |
2574 | @code{obstack_alignment_mask}, whose function prototype looks like | |
2575 | this: | |
2576 | ||
2577 | @comment obstack.h | |
2578 | @comment GNU | |
2579 | @deftypefn Macro int obstack_alignment_mask (struct obstack *@var{obstack-ptr}) | |
9f529d7c | 2580 | @safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}} |
28f540f4 RM |
2581 | The value is a bit mask; a bit that is 1 indicates that the corresponding |
2582 | bit in the address of an object should be 0. The mask value should be one | |
2583 | less than a power of 2; the effect is that all object addresses are | |
11883883 RM |
2584 | multiples of that power of 2. The default value of the mask is a value |
2585 | that allows aligned objects to hold any type of data: for example, if | |
2586 | its value is 3, any type of data can be stored at locations whose | |
28f540f4 RM |
2587 | addresses are multiples of 4. A mask value of 0 means an object can start |
2588 | on any multiple of 1 (that is, no alignment is required). | |
2589 | ||
2590 | The expansion of the macro @code{obstack_alignment_mask} is an lvalue, | |
2591 | so you can alter the mask by assignment. For example, this statement: | |
2592 | ||
2593 | @smallexample | |
2594 | obstack_alignment_mask (obstack_ptr) = 0; | |
2595 | @end smallexample | |
2596 | ||
2597 | @noindent | |
2598 | has the effect of turning off alignment processing in the specified obstack. | |
2599 | @end deftypefn | |
2600 | ||
2601 | Note that a change in alignment mask does not take effect until | |
2602 | @emph{after} the next time an object is allocated or finished in the | |
2603 | obstack. If you are not growing an object, you can make the new | |
2604 | alignment mask take effect immediately by calling @code{obstack_finish}. | |
2605 | This will finish a zero-length object and then do proper alignment for | |
2606 | the next object. | |
2607 | ||
2608 | @node Obstack Chunks | |
99a20616 | 2609 | @subsubsection Obstack Chunks |
28f540f4 RM |
2610 | @cindex efficiency of chunks |
2611 | @cindex chunks | |
2612 | ||
2613 | Obstacks work by allocating space for themselves in large chunks, and | |
2614 | then parceling out space in the chunks to satisfy your requests. Chunks | |
2615 | are normally 4096 bytes long unless you specify a different chunk size. | |
2616 | The chunk size includes 8 bytes of overhead that are not actually used | |
2617 | for storing objects. Regardless of the specified size, longer chunks | |
2618 | will be allocated when necessary for long objects. | |
2619 | ||
2620 | The obstack library allocates chunks by calling the function | |
2621 | @code{obstack_chunk_alloc}, which you must define. When a chunk is no | |
2622 | longer needed because you have freed all the objects in it, the obstack | |
2623 | library frees the chunk by calling @code{obstack_chunk_free}, which you | |
2624 | must also define. | |
2625 | ||
2626 | These two must be defined (as macros) or declared (as functions) in each | |
2627 | source file that uses @code{obstack_init} (@pxref{Creating Obstacks}). | |
2628 | Most often they are defined as macros like this: | |
2629 | ||
2630 | @smallexample | |
bd355af0 | 2631 | #define obstack_chunk_alloc malloc |
28f540f4 RM |
2632 | #define obstack_chunk_free free |
2633 | @end smallexample | |
2634 | ||
2635 | Note that these are simple macros (no arguments). Macro definitions with | |
2636 | arguments will not work! It is necessary that @code{obstack_chunk_alloc} | |
2637 | or @code{obstack_chunk_free}, alone, expand into a function name if it is | |
2638 | not itself a function name. | |
2639 | ||
2640 | If you allocate chunks with @code{malloc}, the chunk size should be a | |
2641 | power of 2. The default chunk size, 4096, was chosen because it is long | |
2642 | enough to satisfy many typical requests on the obstack yet short enough | |
2643 | not to waste too much memory in the portion of the last chunk not yet used. | |
2644 | ||
2645 | @comment obstack.h | |
2646 | @comment GNU | |
2647 | @deftypefn Macro int obstack_chunk_size (struct obstack *@var{obstack-ptr}) | |
9f529d7c | 2648 | @safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}} |
28f540f4 RM |
2649 | This returns the chunk size of the given obstack. |
2650 | @end deftypefn | |
2651 | ||
2652 | Since this macro expands to an lvalue, you can specify a new chunk size by | |
2653 | assigning it a new value. Doing so does not affect the chunks already | |
2654 | allocated, but will change the size of chunks allocated for that particular | |
2655 | obstack in the future. It is unlikely to be useful to make the chunk size | |
2656 | smaller, but making it larger might improve efficiency if you are | |
2657 | allocating many objects whose size is comparable to the chunk size. Here | |
2658 | is how to do so cleanly: | |
2659 | ||
2660 | @smallexample | |
2661 | if (obstack_chunk_size (obstack_ptr) < @var{new-chunk-size}) | |
2662 | obstack_chunk_size (obstack_ptr) = @var{new-chunk-size}; | |
2663 | @end smallexample | |
2664 | ||
2665 | @node Summary of Obstacks | |
99a20616 | 2666 | @subsubsection Summary of Obstack Functions |
28f540f4 RM |
2667 | |
2668 | Here is a summary of all the functions associated with obstacks. Each | |
2669 | takes the address of an obstack (@code{struct obstack *}) as its first | |
2670 | argument. | |
2671 | ||
2672 | @table @code | |
2673 | @item void obstack_init (struct obstack *@var{obstack-ptr}) | |
2674 | Initialize use of an obstack. @xref{Creating Obstacks}. | |
2675 | ||
2676 | @item void *obstack_alloc (struct obstack *@var{obstack-ptr}, int @var{size}) | |
2677 | Allocate an object of @var{size} uninitialized bytes. | |
2678 | @xref{Allocation in an Obstack}. | |
2679 | ||
2680 | @item void *obstack_copy (struct obstack *@var{obstack-ptr}, void *@var{address}, int @var{size}) | |
2681 | Allocate an object of @var{size} bytes, with contents copied from | |
2682 | @var{address}. @xref{Allocation in an Obstack}. | |
2683 | ||
2684 | @item void *obstack_copy0 (struct obstack *@var{obstack-ptr}, void *@var{address}, int @var{size}) | |
2685 | Allocate an object of @var{size}+1 bytes, with @var{size} of them copied | |
2686 | from @var{address}, followed by a null character at the end. | |
2687 | @xref{Allocation in an Obstack}. | |
2688 | ||
2689 | @item void obstack_free (struct obstack *@var{obstack-ptr}, void *@var{object}) | |
2690 | Free @var{object} (and everything allocated in the specified obstack | |
2691 | more recently than @var{object}). @xref{Freeing Obstack Objects}. | |
2692 | ||
2693 | @item void obstack_blank (struct obstack *@var{obstack-ptr}, int @var{size}) | |
2694 | Add @var{size} uninitialized bytes to a growing object. | |
2695 | @xref{Growing Objects}. | |
2696 | ||
2697 | @item void obstack_grow (struct obstack *@var{obstack-ptr}, void *@var{address}, int @var{size}) | |
2698 | Add @var{size} bytes, copied from @var{address}, to a growing object. | |
2699 | @xref{Growing Objects}. | |
2700 | ||
2701 | @item void obstack_grow0 (struct obstack *@var{obstack-ptr}, void *@var{address}, int @var{size}) | |
2702 | Add @var{size} bytes, copied from @var{address}, to a growing object, | |
2703 | and then add another byte containing a null character. @xref{Growing | |
2704 | Objects}. | |
2705 | ||
2706 | @item void obstack_1grow (struct obstack *@var{obstack-ptr}, char @var{data-char}) | |
2707 | Add one byte containing @var{data-char} to a growing object. | |
2708 | @xref{Growing Objects}. | |
2709 | ||
2710 | @item void *obstack_finish (struct obstack *@var{obstack-ptr}) | |
2711 | Finalize the object that is growing and return its permanent address. | |
2712 | @xref{Growing Objects}. | |
2713 | ||
2714 | @item int obstack_object_size (struct obstack *@var{obstack-ptr}) | |
2715 | Get the current size of the currently growing object. @xref{Growing | |
2716 | Objects}. | |
2717 | ||
2718 | @item void obstack_blank_fast (struct obstack *@var{obstack-ptr}, int @var{size}) | |
2719 | Add @var{size} uninitialized bytes to a growing object without checking | |
2720 | that there is enough room. @xref{Extra Fast Growing}. | |
2721 | ||
2722 | @item void obstack_1grow_fast (struct obstack *@var{obstack-ptr}, char @var{data-char}) | |
2723 | Add one byte containing @var{data-char} to a growing object without | |
2724 | checking that there is enough room. @xref{Extra Fast Growing}. | |
2725 | ||
2726 | @item int obstack_room (struct obstack *@var{obstack-ptr}) | |
2727 | Get the amount of room now available for growing the current object. | |
2728 | @xref{Extra Fast Growing}. | |
2729 | ||
2730 | @item int obstack_alignment_mask (struct obstack *@var{obstack-ptr}) | |
2731 | The mask used for aligning the beginning of an object. This is an | |
2732 | lvalue. @xref{Obstacks Data Alignment}. | |
2733 | ||
2734 | @item int obstack_chunk_size (struct obstack *@var{obstack-ptr}) | |
2735 | The size for allocating chunks. This is an lvalue. @xref{Obstack Chunks}. | |
2736 | ||
2737 | @item void *obstack_base (struct obstack *@var{obstack-ptr}) | |
2738 | Tentative starting address of the currently growing object. | |
2739 | @xref{Status of an Obstack}. | |
2740 | ||
2741 | @item void *obstack_next_free (struct obstack *@var{obstack-ptr}) | |
2742 | Address just after the end of the currently growing object. | |
2743 | @xref{Status of an Obstack}. | |
2744 | @end table | |
2745 | ||
2746 | @node Variable Size Automatic | |
99a20616 | 2747 | @subsection Automatic Storage with Variable Size |
28f540f4 RM |
2748 | @cindex automatic freeing |
2749 | @cindex @code{alloca} function | |
2750 | @cindex automatic storage with variable size | |
2751 | ||
2752 | The function @code{alloca} supports a kind of half-dynamic allocation in | |
2753 | which blocks are allocated dynamically but freed automatically. | |
2754 | ||
2755 | Allocating a block with @code{alloca} is an explicit action; you can | |
2756 | allocate as many blocks as you wish, and compute the size at run time. But | |
2757 | all the blocks are freed when you exit the function that @code{alloca} was | |
2758 | called from, just as if they were automatic variables declared in that | |
2759 | function. There is no way to free the space explicitly. | |
2760 | ||
2761 | The prototype for @code{alloca} is in @file{stdlib.h}. This function is | |
2762 | a BSD extension. | |
2763 | @pindex stdlib.h | |
2764 | ||
2765 | @comment stdlib.h | |
2766 | @comment GNU, BSD | |
cc6e48bc | 2767 | @deftypefun {void *} alloca (size_t @var{size}) |
9f529d7c | 2768 | @safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}} |
28f540f4 | 2769 | The return value of @code{alloca} is the address of a block of @var{size} |
99a20616 | 2770 | bytes of memory, allocated in the stack frame of the calling function. |
28f540f4 RM |
2771 | @end deftypefun |
2772 | ||
2773 | Do not use @code{alloca} inside the arguments of a function call---you | |
2774 | will get unpredictable results, because the stack space for the | |
2775 | @code{alloca} would appear on the stack in the middle of the space for | |
2776 | the function arguments. An example of what to avoid is @code{foo (x, | |
2777 | alloca (4), y)}. | |
2778 | @c This might get fixed in future versions of GCC, but that won't make | |
2779 | @c it safe with compilers generally. | |
2780 | ||
2781 | @menu | |
2782 | * Alloca Example:: Example of using @code{alloca}. | |
2783 | * Advantages of Alloca:: Reasons to use @code{alloca}. | |
2784 | * Disadvantages of Alloca:: Reasons to avoid @code{alloca}. | |
2785 | * GNU C Variable-Size Arrays:: Only in GNU C, here is an alternative | |
2786 | method of allocating dynamically and | |
2787 | freeing automatically. | |
2788 | @end menu | |
2789 | ||
2790 | @node Alloca Example | |
99a20616 | 2791 | @subsubsection @code{alloca} Example |
28f540f4 | 2792 | |
bc938d3d UD |
2793 | As an example of the use of @code{alloca}, here is a function that opens |
2794 | a file name made from concatenating two argument strings, and returns a | |
2795 | file descriptor or minus one signifying failure: | |
28f540f4 RM |
2796 | |
2797 | @smallexample | |
2798 | int | |
2799 | open2 (char *str1, char *str2, int flags, int mode) | |
2800 | @{ | |
2801 | char *name = (char *) alloca (strlen (str1) + strlen (str2) + 1); | |
a5113b14 | 2802 | stpcpy (stpcpy (name, str1), str2); |
28f540f4 RM |
2803 | return open (name, flags, mode); |
2804 | @} | |
2805 | @end smallexample | |
2806 | ||
2807 | @noindent | |
2808 | Here is how you would get the same results with @code{malloc} and | |
2809 | @code{free}: | |
2810 | ||
2811 | @smallexample | |
2812 | int | |
2813 | open2 (char *str1, char *str2, int flags, int mode) | |
2814 | @{ | |
2815 | char *name = (char *) malloc (strlen (str1) + strlen (str2) + 1); | |
2816 | int desc; | |
2817 | if (name == 0) | |
2818 | fatal ("virtual memory exceeded"); | |
a5113b14 | 2819 | stpcpy (stpcpy (name, str1), str2); |
28f540f4 RM |
2820 | desc = open (name, flags, mode); |
2821 | free (name); | |
2822 | return desc; | |
2823 | @} | |
2824 | @end smallexample | |
2825 | ||
2826 | As you can see, it is simpler with @code{alloca}. But @code{alloca} has | |
2827 | other, more important advantages, and some disadvantages. | |
2828 | ||
2829 | @node Advantages of Alloca | |
99a20616 | 2830 | @subsubsection Advantages of @code{alloca} |
28f540f4 RM |
2831 | |
2832 | Here are the reasons why @code{alloca} may be preferable to @code{malloc}: | |
2833 | ||
2834 | @itemize @bullet | |
2835 | @item | |
2836 | Using @code{alloca} wastes very little space and is very fast. (It is | |
2837 | open-coded by the GNU C compiler.) | |
2838 | ||
2839 | @item | |
2840 | Since @code{alloca} does not have separate pools for different sizes of | |
3ef569c7 | 2841 | blocks, space used for any size block can be reused for any other size. |
99a20616 | 2842 | @code{alloca} does not cause memory fragmentation. |
28f540f4 RM |
2843 | |
2844 | @item | |
2845 | @cindex longjmp | |
2846 | Nonlocal exits done with @code{longjmp} (@pxref{Non-Local Exits}) | |
2847 | automatically free the space allocated with @code{alloca} when they exit | |
2848 | through the function that called @code{alloca}. This is the most | |
2849 | important reason to use @code{alloca}. | |
2850 | ||
2851 | To illustrate this, suppose you have a function | |
2852 | @code{open_or_report_error} which returns a descriptor, like | |
2853 | @code{open}, if it succeeds, but does not return to its caller if it | |
2854 | fails. If the file cannot be opened, it prints an error message and | |
2855 | jumps out to the command level of your program using @code{longjmp}. | |
2856 | Let's change @code{open2} (@pxref{Alloca Example}) to use this | |
2857 | subroutine:@refill | |
2858 | ||
2859 | @smallexample | |
2860 | int | |
2861 | open2 (char *str1, char *str2, int flags, int mode) | |
2862 | @{ | |
2863 | char *name = (char *) alloca (strlen (str1) + strlen (str2) + 1); | |
a5113b14 | 2864 | stpcpy (stpcpy (name, str1), str2); |
28f540f4 RM |
2865 | return open_or_report_error (name, flags, mode); |
2866 | @} | |
2867 | @end smallexample | |
2868 | ||
2869 | @noindent | |
99a20616 | 2870 | Because of the way @code{alloca} works, the memory it allocates is |
28f540f4 RM |
2871 | freed even when an error occurs, with no special effort required. |
2872 | ||
2873 | By contrast, the previous definition of @code{open2} (which uses | |
99a20616 | 2874 | @code{malloc} and @code{free}) would develop a memory leak if it were |
28f540f4 RM |
2875 | changed in this way. Even if you are willing to make more changes to |
2876 | fix it, there is no easy way to do so. | |
2877 | @end itemize | |
2878 | ||
2879 | @node Disadvantages of Alloca | |
99a20616 | 2880 | @subsubsection Disadvantages of @code{alloca} |
28f540f4 RM |
2881 | |
2882 | @cindex @code{alloca} disadvantages | |
2883 | @cindex disadvantages of @code{alloca} | |
2884 | These are the disadvantages of @code{alloca} in comparison with | |
2885 | @code{malloc}: | |
2886 | ||
2887 | @itemize @bullet | |
2888 | @item | |
99a20616 | 2889 | If you try to allocate more memory than the machine can provide, you |
28f540f4 RM |
2890 | don't get a clean error message. Instead you get a fatal signal like |
2891 | the one you would get from an infinite recursion; probably a | |
2892 | segmentation violation (@pxref{Program Error Signals}). | |
2893 | ||
2894 | @item | |
a7a93d50 | 2895 | Some @nongnusystems{} fail to support @code{alloca}, so it is less |
28f540f4 RM |
2896 | portable. However, a slower emulation of @code{alloca} written in C |
2897 | is available for use on systems with this deficiency. | |
2898 | @end itemize | |
2899 | ||
2900 | @node GNU C Variable-Size Arrays | |
99a20616 | 2901 | @subsubsection GNU C Variable-Size Arrays |
28f540f4 RM |
2902 | @cindex variable-sized arrays |
2903 | ||
2904 | In GNU C, you can replace most uses of @code{alloca} with an array of | |
2905 | variable size. Here is how @code{open2} would look then: | |
2906 | ||
2907 | @smallexample | |
2908 | int open2 (char *str1, char *str2, int flags, int mode) | |
2909 | @{ | |
2910 | char name[strlen (str1) + strlen (str2) + 1]; | |
a5113b14 | 2911 | stpcpy (stpcpy (name, str1), str2); |
28f540f4 RM |
2912 | return open (name, flags, mode); |
2913 | @} | |
2914 | @end smallexample | |
2915 | ||
2916 | But @code{alloca} is not always equivalent to a variable-sized array, for | |
2917 | several reasons: | |
2918 | ||
2919 | @itemize @bullet | |
2920 | @item | |
2921 | A variable size array's space is freed at the end of the scope of the | |
2922 | name of the array. The space allocated with @code{alloca} | |
2923 | remains until the end of the function. | |
2924 | ||
2925 | @item | |
2926 | It is possible to use @code{alloca} within a loop, allocating an | |
2927 | additional block on each iteration. This is impossible with | |
2928 | variable-sized arrays. | |
2929 | @end itemize | |
2930 | ||
48b22986 | 2931 | @strong{NB:} If you mix use of @code{alloca} and variable-sized arrays |
28f540f4 RM |
2932 | within one function, exiting a scope in which a variable-sized array was |
2933 | declared frees all blocks allocated with @code{alloca} during the | |
2934 | execution of that scope. | |
2935 | ||
99a20616 UD |
2936 | |
2937 | @node Resizing the Data Segment | |
2938 | @section Resizing the Data Segment | |
2939 | ||
2940 | The symbols in this section are declared in @file{unistd.h}. | |
2941 | ||
2942 | You will not normally use the functions in this section, because the | |
2943 | functions described in @ref{Memory Allocation} are easier to use. Those | |
1f77f049 | 2944 | are interfaces to a @glibcadj{} memory allocator that uses the |
99a20616 UD |
2945 | functions below itself. The functions below are simple interfaces to |
2946 | system calls. | |
2947 | ||
2948 | @comment unistd.h | |
2949 | @comment BSD | |
2950 | @deftypefun int brk (void *@var{addr}) | |
9f529d7c | 2951 | @safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}} |
99a20616 UD |
2952 | |
2953 | @code{brk} sets the high end of the calling process' data segment to | |
2954 | @var{addr}. | |
2955 | ||
2956 | The address of the end of a segment is defined to be the address of the | |
2957 | last byte in the segment plus 1. | |
2958 | ||
2959 | The function has no effect if @var{addr} is lower than the low end of | |
3ef569c7 | 2960 | the data segment. (This is considered success, by the way.) |
99a20616 UD |
2961 | |
2962 | The function fails if it would cause the data segment to overlap another | |
68979757 | 2963 | segment or exceed the process' data storage limit (@pxref{Limits on |
99a20616 UD |
2964 | Resources}). |
2965 | ||
2966 | The function is named for a common historical case where data storage | |
2967 | and the stack are in the same segment. Data storage allocation grows | |
2968 | upward from the bottom of the segment while the stack grows downward | |
2969 | toward it from the top of the segment and the curtain between them is | |
2970 | called the @dfn{break}. | |
2971 | ||
2972 | The return value is zero on success. On failure, the return value is | |
68979757 | 2973 | @code{-1} and @code{errno} is set accordingly. The following @code{errno} |
99a20616 UD |
2974 | values are specific to this function: |
2975 | ||
2976 | @table @code | |
2977 | @item ENOMEM | |
2978 | The request would cause the data segment to overlap another segment or | |
2979 | exceed the process' data storage limit. | |
2980 | @end table | |
2981 | ||
2982 | @c The Brk system call in Linux (as opposed to the GNU C Library function) | |
2983 | @c is considerably different. It always returns the new end of the data | |
2984 | @c segment, whether it succeeds or fails. The GNU C library Brk determines | |
bbf70ae9 | 2985 | @c it's a failure if and only if the system call returns an address less |
99a20616 UD |
2986 | @c than the address requested. |
2987 | ||
2988 | @end deftypefun | |
2989 | ||
2990 | ||
2991 | @comment unistd.h | |
2992 | @comment BSD | |
d6868416 | 2993 | @deftypefun void *sbrk (ptrdiff_t @var{delta}) |
9f529d7c AO |
2994 | @safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}} |
2995 | ||
99a20616 UD |
2996 | This function is the same as @code{brk} except that you specify the new |
2997 | end of the data segment as an offset @var{delta} from the current end | |
2998 | and on success the return value is the address of the resulting end of | |
2999 | the data segment instead of zero. | |
3000 | ||
3001 | This means you can use @samp{sbrk(0)} to find out what the current end | |
3002 | of the data segment is. | |
3003 | ||
3004 | @end deftypefun | |
3005 | ||
3006 | ||
3007 | ||
3008 | @node Locking Pages | |
3009 | @section Locking Pages | |
3010 | @cindex locking pages | |
3011 | @cindex memory lock | |
3012 | @cindex paging | |
3013 | ||
3014 | You can tell the system to associate a particular virtual memory page | |
11bf311e | 3015 | with a real page frame and keep it that way --- i.e., cause the page to |
99a20616 UD |
3016 | be paged in if it isn't already and mark it so it will never be paged |
3017 | out and consequently will never cause a page fault. This is called | |
3018 | @dfn{locking} a page. | |
3019 | ||
3020 | The functions in this chapter lock and unlock the calling process' | |
3021 | pages. | |
3022 | ||
3023 | @menu | |
3024 | * Why Lock Pages:: Reasons to read this section. | |
3025 | * Locked Memory Details:: Everything you need to know locked | |
3026 | memory | |
3027 | * Page Lock Functions:: Here's how to do it. | |
3028 | @end menu | |
3029 | ||
3030 | @node Why Lock Pages | |
3031 | @subsection Why Lock Pages | |
3032 | ||
3033 | Because page faults cause paged out pages to be paged in transparently, | |
68979757 | 3034 | a process rarely needs to be concerned about locking pages. However, |
99a20616 UD |
3035 | there are two reasons people sometimes are: |
3036 | ||
3037 | @itemize @bullet | |
3038 | ||
3039 | @item | |
3040 | Speed. A page fault is transparent only insofar as the process is not | |
3041 | sensitive to how long it takes to do a simple memory access. Time-critical | |
3042 | processes, especially realtime processes, may not be able to wait or | |
3043 | may not be able to tolerate variance in execution speed. | |
3044 | @cindex realtime processing | |
3045 | @cindex speed of execution | |
3046 | ||
3047 | A process that needs to lock pages for this reason probably also needs | |
3048 | priority among other processes for use of the CPU. @xref{Priority}. | |
3049 | ||
3050 | In some cases, the programmer knows better than the system's demand | |
3051 | paging allocator which pages should remain in real memory to optimize | |
3052 | system performance. In this case, locking pages can help. | |
3053 | ||
3054 | @item | |
3055 | Privacy. If you keep secrets in virtual memory and that virtual memory | |
3056 | gets paged out, that increases the chance that the secrets will get out. | |
3057 | If a password gets written out to disk swap space, for example, it might | |
3058 | still be there long after virtual and real memory have been wiped clean. | |
3059 | ||
3060 | @end itemize | |
3061 | ||
3062 | Be aware that when you lock a page, that's one fewer page frame that can | |
3063 | be used to back other virtual memory (by the same or other processes), | |
3064 | which can mean more page faults, which means the system runs more | |
3065 | slowly. In fact, if you lock enough memory, some programs may not be | |
3066 | able to run at all for lack of real memory. | |
3067 | ||
3068 | @node Locked Memory Details | |
3069 | @subsection Locked Memory Details | |
3070 | ||
3071 | A memory lock is associated with a virtual page, not a real frame. The | |
3072 | paging rule is: If a frame backs at least one locked page, don't page it | |
3073 | out. | |
3074 | ||
11bf311e | 3075 | Memory locks do not stack. I.e., you can't lock a particular page twice |
99a20616 UD |
3076 | so that it has to be unlocked twice before it is truly unlocked. It is |
3077 | either locked or it isn't. | |
3078 | ||
3079 | A memory lock persists until the process that owns the memory explicitly | |
3080 | unlocks it. (But process termination and exec cause the virtual memory | |
3081 | to cease to exist, which you might say means it isn't locked any more). | |
3082 | ||
3083 | Memory locks are not inherited by child processes. (But note that on a | |
3084 | modern Unix system, immediately after a fork, the parent's and the | |
3085 | child's virtual address space are backed by the same real page frames, | |
3086 | so the child enjoys the parent's locks). @xref{Creating a Process}. | |
3087 | ||
3088 | Because of its ability to impact other processes, only the superuser can | |
3089 | lock a page. Any process can unlock its own page. | |
3090 | ||
3091 | The system sets limits on the amount of memory a process can have locked | |
3092 | and the amount of real memory it can have dedicated to it. @xref{Limits | |
3093 | on Resources}. | |
3094 | ||
3095 | In Linux, locked pages aren't as locked as you might think. | |
3096 | Two virtual pages that are not shared memory can nonetheless be backed | |
3097 | by the same real frame. The kernel does this in the name of efficiency | |
3098 | when it knows both virtual pages contain identical data, and does it | |
68979757 | 3099 | even if one or both of the virtual pages are locked. |
99a20616 UD |
3100 | |
3101 | But when a process modifies one of those pages, the kernel must get it a | |
3102 | separate frame and fill it with the page's data. This is known as a | |
3103 | @dfn{copy-on-write page fault}. It takes a small amount of time and in | |
3104 | a pathological case, getting that frame may require I/O. | |
3105 | @cindex copy-on-write page fault | |
3106 | @cindex page fault, copy-on-write | |
3107 | ||
3108 | To make sure this doesn't happen to your program, don't just lock the | |
3109 | pages. Write to them as well, unless you know you won't write to them | |
3110 | ever. And to make sure you have pre-allocated frames for your stack, | |
3111 | enter a scope that declares a C automatic variable larger than the | |
3112 | maximum stack size you will need, set it to something, then return from | |
3113 | its scope. | |
3114 | ||
3115 | @node Page Lock Functions | |
3116 | @subsection Functions To Lock And Unlock Pages | |
3117 | ||
3118 | The symbols in this section are declared in @file{sys/mman.h}. These | |
3119 | functions are defined by POSIX.1b, but their availability depends on | |
3120 | your kernel. If your kernel doesn't allow these functions, they exist | |
3121 | but always fail. They @emph{are} available with a Linux kernel. | |
3122 | ||
3123 | @strong{Portability Note:} POSIX.1b requires that when the @code{mlock} | |
3124 | and @code{munlock} functions are available, the file @file{unistd.h} | |
3125 | define the macro @code{_POSIX_MEMLOCK_RANGE} and the file | |
3126 | @code{limits.h} define the macro @code{PAGESIZE} to be the size of a | |
3127 | memory page in bytes. It requires that when the @code{mlockall} and | |
3128 | @code{munlockall} functions are available, the @file{unistd.h} file | |
1f77f049 | 3129 | define the macro @code{_POSIX_MEMLOCK}. @Theglibc{} conforms to |
99a20616 UD |
3130 | this requirement. |
3131 | ||
3132 | @comment sys/mman.h | |
3133 | @comment POSIX.1b | |
3134 | @deftypefun int mlock (const void *@var{addr}, size_t @var{len}) | |
9f529d7c | 3135 | @safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}} |
99a20616 UD |
3136 | |
3137 | @code{mlock} locks a range of the calling process' virtual pages. | |
3138 | ||
3139 | The range of memory starts at address @var{addr} and is @var{len} bytes | |
3140 | long. Actually, since you must lock whole pages, it is the range of | |
3141 | pages that include any part of the specified range. | |
3142 | ||
3143 | When the function returns successfully, each of those pages is backed by | |
3144 | (connected to) a real frame (is resident) and is marked to stay that | |
3145 | way. This means the function may cause page-ins and have to wait for | |
3146 | them. | |
3147 | ||
3148 | When the function fails, it does not affect the lock status of any | |
3149 | pages. | |
3150 | ||
3151 | The return value is zero if the function succeeds. Otherwise, it is | |
3152 | @code{-1} and @code{errno} is set accordingly. @code{errno} values | |
3153 | specific to this function are: | |
3154 | ||
3155 | @table @code | |
3156 | @item ENOMEM | |
3157 | @itemize @bullet | |
3158 | @item | |
3159 | At least some of the specified address range does not exist in the | |
3160 | calling process' virtual address space. | |
3161 | @item | |
3162 | The locking would cause the process to exceed its locked page limit. | |
3163 | @end itemize | |
3164 | ||
3165 | @item EPERM | |
3166 | The calling process is not superuser. | |
3167 | ||
3168 | @item EINVAL | |
3169 | @var{len} is not positive. | |
3170 | ||
3171 | @item ENOSYS | |
3172 | The kernel does not provide @code{mlock} capability. | |
3173 | ||
3174 | @end table | |
3175 | ||
3176 | You can lock @emph{all} a process' memory with @code{mlockall}. You | |
3177 | unlock memory with @code{munlock} or @code{munlockall}. | |
3178 | ||
3179 | To avoid all page faults in a C program, you have to use | |
3180 | @code{mlockall}, because some of the memory a program uses is hidden | |
3181 | from the C code, e.g. the stack and automatic variables, and you | |
3182 | wouldn't know what address to tell @code{mlock}. | |
3183 | ||
3184 | @end deftypefun | |
3185 | ||
3186 | @comment sys/mman.h | |
3187 | @comment POSIX.1b | |
3188 | @deftypefun int munlock (const void *@var{addr}, size_t @var{len}) | |
9f529d7c | 3189 | @safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}} |
99a20616 | 3190 | |
10e0498e | 3191 | @code{munlock} unlocks a range of the calling process' virtual pages. |
99a20616 UD |
3192 | |
3193 | @code{munlock} is the inverse of @code{mlock} and functions completely | |
3194 | analogously to @code{mlock}, except that there is no @code{EPERM} | |
3195 | failure. | |
3196 | ||
3197 | @end deftypefun | |
3198 | ||
3199 | @comment sys/mman.h | |
3200 | @comment POSIX.1b | |
3201 | @deftypefun int mlockall (int @var{flags}) | |
9f529d7c | 3202 | @safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}} |
99a20616 UD |
3203 | |
3204 | @code{mlockall} locks all the pages in a process' virtual memory address | |
3205 | space, and/or any that are added to it in the future. This includes the | |
3206 | pages of the code, data and stack segment, as well as shared libraries, | |
3207 | user space kernel data, shared memory, and memory mapped files. | |
3208 | ||
3209 | @var{flags} is a string of single bit flags represented by the following | |
3210 | macros. They tell @code{mlockall} which of its functions you want. All | |
3211 | other bits must be zero. | |
3212 | ||
3213 | @table @code | |
3214 | ||
3215 | @item MCL_CURRENT | |
3216 | Lock all pages which currently exist in the calling process' virtual | |
3217 | address space. | |
3218 | ||
3219 | @item MCL_FUTURE | |
3220 | Set a mode such that any pages added to the process' virtual address | |
3221 | space in the future will be locked from birth. This mode does not | |
3222 | affect future address spaces owned by the same process so exec, which | |
3223 | replaces a process' address space, wipes out @code{MCL_FUTURE}. | |
3224 | @xref{Executing a File}. | |
3225 | ||
3226 | @end table | |
3227 | ||
3228 | When the function returns successfully, and you specified | |
3229 | @code{MCL_CURRENT}, all of the process' pages are backed by (connected | |
3230 | to) real frames (they are resident) and are marked to stay that way. | |
3231 | This means the function may cause page-ins and have to wait for them. | |
3232 | ||
3233 | When the process is in @code{MCL_FUTURE} mode because it successfully | |
3234 | executed this function and specified @code{MCL_CURRENT}, any system call | |
3235 | by the process that requires space be added to its virtual address space | |
3236 | fails with @code{errno} = @code{ENOMEM} if locking the additional space | |
3237 | would cause the process to exceed its locked page limit. In the case | |
0bc93a2f | 3238 | that the address space addition that can't be accommodated is stack |
99a20616 UD |
3239 | expansion, the stack expansion fails and the kernel sends a |
3240 | @code{SIGSEGV} signal to the process. | |
3241 | ||
3242 | When the function fails, it does not affect the lock status of any pages | |
3243 | or the future locking mode. | |
3244 | ||
3245 | The return value is zero if the function succeeds. Otherwise, it is | |
3246 | @code{-1} and @code{errno} is set accordingly. @code{errno} values | |
3247 | specific to this function are: | |
3248 | ||
3249 | @table @code | |
3250 | @item ENOMEM | |
3251 | @itemize @bullet | |
3252 | @item | |
3253 | At least some of the specified address range does not exist in the | |
3254 | calling process' virtual address space. | |
3255 | @item | |
3256 | The locking would cause the process to exceed its locked page limit. | |
3257 | @end itemize | |
3258 | ||
3259 | @item EPERM | |
3260 | The calling process is not superuser. | |
3261 | ||
3262 | @item EINVAL | |
3263 | Undefined bits in @var{flags} are not zero. | |
3264 | ||
3265 | @item ENOSYS | |
3266 | The kernel does not provide @code{mlockall} capability. | |
3267 | ||
3268 | @end table | |
3269 | ||
3270 | You can lock just specific pages with @code{mlock}. You unlock pages | |
3271 | with @code{munlockall} and @code{munlock}. | |
3272 | ||
3273 | @end deftypefun | |
3274 | ||
3275 | ||
3276 | @comment sys/mman.h | |
3277 | @comment POSIX.1b | |
3278 | @deftypefun int munlockall (void) | |
9f529d7c | 3279 | @safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}} |
99a20616 UD |
3280 | |
3281 | @code{munlockall} unlocks every page in the calling process' virtual | |
3ef569c7 | 3282 | address space and turns off @code{MCL_FUTURE} future locking mode. |
99a20616 UD |
3283 | |
3284 | The return value is zero if the function succeeds. Otherwise, it is | |
68979757 | 3285 | @code{-1} and @code{errno} is set accordingly. The only way this |
99a20616 UD |
3286 | function can fail is for generic reasons that all functions and system |
3287 | calls can fail, so there are no specific @code{errno} values. | |
3288 | ||
3289 | @end deftypefun | |
3290 | ||
3291 | ||
3292 | ||
3293 | ||
a9ddb793 UD |
3294 | @ignore |
3295 | @c This was never actually implemented. -zw | |
28f540f4 RM |
3296 | @node Relocating Allocator |
3297 | @section Relocating Allocator | |
3298 | ||
3299 | @cindex relocating memory allocator | |
3300 | Any system of dynamic memory allocation has overhead: the amount of | |
3301 | space it uses is more than the amount the program asks for. The | |
3302 | @dfn{relocating memory allocator} achieves very low overhead by moving | |
3303 | blocks in memory as necessary, on its own initiative. | |
3304 | ||
a9ddb793 UD |
3305 | @c @menu |
3306 | @c * Relocator Concepts:: How to understand relocating allocation. | |
3307 | @c * Using Relocator:: Functions for relocating allocation. | |
3308 | @c @end menu | |
28f540f4 RM |
3309 | |
3310 | @node Relocator Concepts | |
3311 | @subsection Concepts of Relocating Allocation | |
3312 | ||
3313 | @ifinfo | |
3314 | The @dfn{relocating memory allocator} achieves very low overhead by | |
3315 | moving blocks in memory as necessary, on its own initiative. | |
3316 | @end ifinfo | |
3317 | ||
3318 | When you allocate a block with @code{malloc}, the address of the block | |
3319 | never changes unless you use @code{realloc} to change its size. Thus, | |
3320 | you can safely store the address in various places, temporarily or | |
3321 | permanently, as you like. This is not safe when you use the relocating | |
3322 | memory allocator, because any and all relocatable blocks can move | |
3323 | whenever you allocate memory in any fashion. Even calling @code{malloc} | |
3324 | or @code{realloc} can move the relocatable blocks. | |
3325 | ||
3326 | @cindex handle | |
3327 | For each relocatable block, you must make a @dfn{handle}---a pointer | |
3328 | object in memory, designated to store the address of that block. The | |
3329 | relocating allocator knows where each block's handle is, and updates the | |
3330 | address stored there whenever it moves the block, so that the handle | |
3331 | always points to the block. Each time you access the contents of the | |
3332 | block, you should fetch its address anew from the handle. | |
3333 | ||
3334 | To call any of the relocating allocator functions from a signal handler | |
3335 | is almost certainly incorrect, because the signal could happen at any | |
3336 | time and relocate all the blocks. The only way to make this safe is to | |
3337 | block the signal around any access to the contents of any relocatable | |
3338 | block---not a convenient mode of operation. @xref{Nonreentrancy}. | |
3339 | ||
3340 | @node Using Relocator | |
3341 | @subsection Allocating and Freeing Relocatable Blocks | |
3342 | ||
3343 | @pindex malloc.h | |
3344 | In the descriptions below, @var{handleptr} designates the address of the | |
3345 | handle. All the functions are declared in @file{malloc.h}; all are GNU | |
3346 | extensions. | |
3347 | ||
3348 | @comment malloc.h | |
3349 | @comment GNU | |
a9ddb793 | 3350 | @c @deftypefun {void *} r_alloc (void **@var{handleptr}, size_t @var{size}) |
28f540f4 RM |
3351 | This function allocates a relocatable block of size @var{size}. It |
3352 | stores the block's address in @code{*@var{handleptr}} and returns | |
3353 | a non-null pointer to indicate success. | |
3354 | ||
3355 | If @code{r_alloc} can't get the space needed, it stores a null pointer | |
3356 | in @code{*@var{handleptr}}, and returns a null pointer. | |
3357 | @end deftypefun | |
3358 | ||
3359 | @comment malloc.h | |
3360 | @comment GNU | |
a9ddb793 | 3361 | @c @deftypefun void r_alloc_free (void **@var{handleptr}) |
28f540f4 RM |
3362 | This function is the way to free a relocatable block. It frees the |
3363 | block that @code{*@var{handleptr}} points to, and stores a null pointer | |
3364 | in @code{*@var{handleptr}} to show it doesn't point to an allocated | |
3365 | block any more. | |
3366 | @end deftypefun | |
3367 | ||
3368 | @comment malloc.h | |
3369 | @comment GNU | |
a9ddb793 | 3370 | @c @deftypefun {void *} r_re_alloc (void **@var{handleptr}, size_t @var{size}) |
28f540f4 RM |
3371 | The function @code{r_re_alloc} adjusts the size of the block that |
3372 | @code{*@var{handleptr}} points to, making it @var{size} bytes long. It | |
3373 | stores the address of the resized block in @code{*@var{handleptr}} and | |
3374 | returns a non-null pointer to indicate success. | |
3375 | ||
3376 | If enough memory is not available, this function returns a null pointer | |
3377 | and does not modify @code{*@var{handleptr}}. | |
3378 | @end deftypefun | |
a9ddb793 | 3379 | @end ignore |
28f540f4 | 3380 | |
99a20616 UD |
3381 | |
3382 | ||
3383 | ||
c131718c UD |
3384 | @ignore |
3385 | @comment No longer available... | |
3386 | ||
3387 | @comment @node Memory Warnings | |
3388 | @comment @section Memory Usage Warnings | |
3389 | @comment @cindex memory usage warnings | |
3390 | @comment @cindex warnings of memory almost full | |
28f540f4 RM |
3391 | |
3392 | @pindex malloc.c | |
3393 | You can ask for warnings as the program approaches running out of memory | |
3394 | space, by calling @code{memory_warnings}. This tells @code{malloc} to | |
3395 | check memory usage every time it asks for more memory from the operating | |
3396 | system. This is a GNU extension declared in @file{malloc.h}. | |
3397 | ||
3398 | @comment malloc.h | |
3399 | @comment GNU | |
c131718c | 3400 | @comment @deftypefun void memory_warnings (void *@var{start}, void (*@var{warn-func}) (const char *)) |
28f540f4 RM |
3401 | Call this function to request warnings for nearing exhaustion of virtual |
3402 | memory. | |
3403 | ||
3404 | The argument @var{start} says where data space begins, in memory. The | |
3405 | allocator compares this against the last address used and against the | |
3406 | limit of data space, to determine the fraction of available memory in | |
3407 | use. If you supply zero for @var{start}, then a default value is used | |
3408 | which is right in most circumstances. | |
3409 | ||
3410 | For @var{warn-func}, supply a function that @code{malloc} can call to | |
3411 | warn you. It is called with a string (a warning message) as argument. | |
3412 | Normally it ought to display the string for the user to read. | |
3413 | @end deftypefun | |
3414 | ||
3415 | The warnings come when memory becomes 75% full, when it becomes 85% | |
3416 | full, and when it becomes 95% full. Above 95% you get another warning | |
3417 | each time memory usage increases. | |
c131718c UD |
3418 | |
3419 | @end ignore |