1 .\" Hey Emacs! This file is -*- nroff -*- source.
3 .\" Copyright (C) 1996 Andries Brouwer <aeb@cwi.nl>
4 .\" and Copyright (C) 2006, 2007 Michael Kerrisk <mtk.manpages@gmail.com>
6 .\" Permission is granted to make and distribute verbatim copies of this
7 .\" manual provided the copyright notice and this permission notice are
8 .\" preserved on all copies.
10 .\" Permission is granted to copy and distribute modified versions of this
11 .\" manual under the conditions for verbatim copying, provided that the
12 .\" entire resulting derived work is distributed under the terms of a
13 .\" permission notice identical to this one.
15 .\" Since the Linux kernel and libraries are constantly changing, this
16 .\" manual page may be incorrect or out-of-date. The author(s) assume no
17 .\" responsibility for errors or omissions, or for damages resulting from
18 .\" the use of the information contained herein. The author(s) may not
19 .\" have taken the same level of care in the production of this manual,
20 .\" which is licensed free of charge, as they might when working
23 .\" Formatted or processed versions of this manual, if unaccompanied by
24 .\" the source, must acknowledge the copyright and authors of this work.
26 .\" Modified 1997-01-31 by Eric S. Raymond <esr@thyrsus.com>
27 .\" Modified 2000-03-25 by Jim Van Zandt <jrv@vanzandt.mv.com>
28 .\" Modified 2001-10-04 by John Levon <moz@compsoc.man.ac.uk>
29 .\" Modified 2003-02-02 by Andi Kleen <ak@muc.de>
30 .\" Modified 2003-05-21 by Michael Kerrisk <mtk.manpages@gmail.com>
31 .\" MAP_LOCKED works from 2.5.37
32 .\" Modified 2004-06-17 by Michael Kerrisk <mtk.manpages@gmail.com>
33 .\" Modified 2004-09-11 by aeb
34 .\" Modified 2004-12-08, from Eric Estievenart <eric.estievenart@free.fr>
35 .\" Modified 2004-12-08, mtk, formatting tidy-ups
36 .\" Modified 2006-12-04, mtk, various parts rewritten
37 .\" 2007-07-10, mtk, Added an example program.
38 .\" 2008-11-18, mtk, document MAP_STACK
40 .TH MMAP 2 2012-04-16 "Linux" "Linux Programmer's Manual"
42 mmap, munmap \- map or unmap files or devices into memory
45 .B #include <sys/mman.h>
47 .BI "void *mmap(void *" addr ", size_t " length \
48 ", int " prot ", int " flags ,
49 .BI " int " fd ", off_t " offset );
50 .BI "int munmap(void *" addr ", size_t " length );
54 creates a new mapping in the virtual address space of
56 The starting address for the new mapping is specified in
60 argument specifies the length of the mapping.
65 then the kernel chooses the address at which to create the mapping;
66 this is the most portable method of creating a new mapping.
70 then the kernel takes it as a hint about where to place the mapping;
71 on Linux, the mapping will be created at a nearby page boundary.
72 .\" Before Linux 2.6.24, the address was rounded up to the next page
73 .\" boundary; since 2.6.24, it is rounded down!
74 The address of the new mapping is returned as the result of the call.
76 The contents of a file mapping (as opposed to an anonymous mapping; see
78 below), are initialized using
80 bytes starting at offset
82 in the file (or other object) referred to by the file descriptor
85 must be a multiple of the page size as returned by
86 .IR sysconf(_SC_PAGE_SIZE) .
90 argument describes the desired memory protection of the mapping
91 (and must not conflict with the open mode of the file).
94 or the bitwise OR of one or more of the following flags:
97 Pages may be executed.
103 Pages may be written.
106 Pages may not be accessed.
110 argument determines whether updates to the mapping
111 are visible to other processes mapping the same region,
112 and whether updates are carried through to the underlying file.
113 This behavior is determined by including exactly one
114 of the following values in
119 Updates to the mapping are visible to other processes that map this file,
120 and are carried through to the underlying file.
121 The file may not actually be updated until
128 Create a private copy-on-write mapping.
129 Updates to the mapping are not visible to other processes
130 mapping the same file, and are not carried through to
132 It is unspecified whether changes made to the file after the
134 call are visible in the mapped region.
136 Both of these flags are described in POSIX.1-2001.
138 In addition, zero or more of the following values can be ORed in
141 .BR MAP_32BIT " (since Linux 2.4.20, 2.6)"
142 Put the mapping into the first 2 Gigabytes of the process address space.
143 This flag is only supported on x86-64, for 64-bit programs.
144 It was added to allow thread stacks to be allocated somewhere
145 in the first 2GB of memory,
146 so as to improve context-switch performance on some early
148 .\" See http://lwn.net/Articles/294642 "Tangled up in threads", 19 Aug 08
149 Modern x86-64 processors no longer have this performance problem,
150 so use of this flag is not required on those systems.
163 The mapping is not backed by any file;
164 its contents are initialized to zero.
169 arguments are ignored;
170 however, some implementations require
177 and portable applications should ensure this.
182 is only supported on Linux since kernel 2.4.
185 This flag is ignored.
186 .\" Introduced in 1.1.36, removed in 1.3.24.
187 (Long ago, it signaled that attempts to write to the underlying file
190 But this was a source of denial-of-service attacks.)
193 This flag is ignored.
194 .\" Introduced in 1.1.38, removed in 1.3.24. Flag tested in proc_follow_link.
195 .\" (Long ago, it signaled that the underlying file is an executable.
196 .\" However, that information was not really used anywhere.)
197 .\" Linus talked about DOS related to MAP_EXECUTABLE, but he was thinking of
203 .\" On some systems, this was required as the opposite of
204 .\" MAP_ANONYMOUS -- mtk, 1 May 2007
209 as a hint: place the mapping at exactly that address.
211 must be a multiple of the page size.
212 If the memory region specified by
216 overlaps pages of any existing mapping(s), then the overlapped
217 part of the existing mapping(s) will be discarded.
218 If the specified address cannot be used,
221 Because requiring a fixed address for a mapping is less portable,
222 the use of this option is discouraged.
226 Indicates to the kernel virtual memory system that the mapping
227 should extend downward in memory.
229 .BR MAP_HUGETLB " (since Linux 2.6.32)"
230 Allocate the mapping using "huge pages."
231 See the kernel source file
232 .I Documentation/vm/hugetlbpage.txt
233 for further information.
235 .BR MAP_LOCKED " (since Linux 2.5.37)"
236 Lock the pages of the mapped region into memory in the manner of
238 This flag is ignored in older kernels.
239 .\" If set, the mapped pages will not be swapped out.
241 .BR MAP_NONBLOCK " (since Linux 2.5.46)"
242 Only meaningful in conjunction with
244 Don't perform read-ahead:
245 only create page tables entries for pages
246 that are already present in RAM.
247 Since Linux 2.6.23, this flag causes
250 One day the combination of
254 may be reimplemented.
257 Do not reserve swap space for this mapping.
258 When swap space is reserved, one has the guarantee
259 that it is possible to modify the mapping.
260 When swap space is not reserved one might get
263 if no physical memory is available.
264 See also the discussion of the file
265 .I /proc/sys/vm/overcommit_memory
268 In kernels before 2.6, this flag only had effect for
269 private writable mappings.
271 .BR MAP_POPULATE " (since Linux 2.5.46)"
272 Populate (prefault) page tables for a mapping.
273 For a file mapping, this causes read-ahead on the file.
274 Later accesses to the mapping will not be blocked by page faults.
276 is only supported for private mappings since Linux 2.6.23.
278 .BR MAP_STACK " (since Linux 2.6.27)"
279 Allocate the mapping at an address suitable for a process
281 This flag is currently a no-op,
282 but is used in the glibc threading implementation so that
283 if some architectures require special treatment for stack allocations,
284 support can later be transparently implemented for glibc.
285 .\" See http://lwn.net/Articles/294642 "Tangled up in threads", 19 Aug 08
286 .\" commit cd98a04a59e2f94fa64d5bf1e26498d27427d5e7
287 .\" http://thread.gmane.org/gmane.linux.kernel/720412
288 .\" "pthread_create() slow for many threads; also time to revisit 64b
289 .\" context switch optimization?"
291 .BR MAP_UNINITIALIZED " (since Linux 2.6.33)"
292 Don't clear anonymous pages.
293 This flag is intended to improve performance on embedded devices.
294 This flag is only honored if the kernel was configured with the
295 .B CONFIG_MMAP_ALLOW_UNINITIALIZED
297 Because of the security implications,
298 that option is normally enabled only on embedded devices
299 (i.e., devices where one has complete control of the contents of user memory).
301 Of the above flags, only
303 is specified in POSIX.1-2001.
304 However, most systems also support
309 Some systems document the additional flags
320 with the same attributes.
322 A file is mapped in multiples of the page size.
323 For a file that is not
324 a multiple of the page size, the remaining memory is zeroed when mapped,
325 and writes to that region are not written out to the file.
327 changing the size of the underlying file of a mapping on the pages that
328 correspond to added or removed regions of the file is unspecified.
332 system call deletes the mappings for the specified address range, and
333 causes further references to addresses within the range to generate
334 invalid memory references.
335 The region is also automatically unmapped
336 when the process is terminated.
337 On the other hand, closing the file
338 descriptor does not unmap the region.
342 must be a multiple of the page size.
343 All pages containing a part
344 of the indicated range are unmapped, and subsequent references
345 to these pages will generate
347 It is not an error if the
348 indicated range does not contain any mapped pages.
349 .SS Timestamps changes for file-backed mappings
350 For file-backed mappings, the
352 field for the mapped file may be updated at any time between the
354 and the corresponding unmapping; the first reference to a mapped
355 page will update the field if it has not been already.
361 field for a file mapped with
365 will be updated after
366 a write to the mapped region, and before a subsequent
376 returns a pointer to the mapped area.
380 .IR "(void\ *)\ \-1" )
383 is set appropriately.
386 returns 0, on failure \-1, and
393 A file descriptor refers to a non-regular file.
398 is not open for reading.
405 is not open in read/write
410 is set, but the file is append-only.
413 The file has been locked, or too much memory has been locked (see
418 is not a valid file descriptor (and
428 (e.g., they are too large, or not aligned on a page boundary).
441 or contained both of these values.
444 .\" This is for shared anonymous segments
445 .\" [2.6.7] shmem_zero_setup()-->shmem_file_setup()-->get_empty_filp()
446 The system limit on the total number of open files has been reached.
449 .\" A file could not be mapped for reading.
452 The underlying file system of the specified file does not support
456 No memory is available, or the process's maximum number of mappings would
464 but the mapped area belongs to a file on a file system that
466 .\" (Since 2.4.25 / 2.6.0.)
470 was set but the object specified by
474 Use of a mapped region can result in these signals:
477 Attempted write into a region mapped as read-only.
480 Attempted access to a portion of the buffer that does not correspond
481 to the file (for example, beyond the end of the file, including the
482 case where another process has truncated the file).
484 SVr4, 4.4BSD, POSIX.1-2001.
485 .\" SVr4 documents additional error codes ENXIO and ENODEV.
486 .\" SUSv2 documents additional error codes EMFILE and EOVERFLOW.
488 On POSIX systems on which
494 .B _POSIX_MAPPED_FILES
495 is defined in \fI<unistd.h>\fP to a value greater than 0.
498 .\" POSIX.1-2001: It shall be defined to -1 or 0 or 200112L.
499 .\" -1: unavailable, 0: ask using sysconf().
500 .\" glibc defines it to 1.
502 This page describes the interface provided by the glibc
505 Originally, this function invoked a system call of the same name.
506 Since kernel 2.4, that system call has been superseded by
509 .\" Since around glibc 2.1/2.2, depending on the platform.
512 wrapper function invokes
514 with a suitably adjusted value for
517 On some hardware architectures (e.g., i386),
521 It is architecture dependent whether
526 Portable programs should always set
528 if they intend to execute code in the new mapping.
530 The portable way to create a mapping is to specify
532 as 0 (NULL), and omit
536 In this case, the system chooses the address for the mapping;
537 the address is chosen so as not to conflict with any existing mapping,
541 flag is specified, and
543 is 0 (NULL), then the mapped address will be 0 (NULL).
545 On Linux there are no guarantees like those suggested above under
547 By default, any process can be killed
548 at any moment when the system runs out of memory.
550 In kernels before 2.6.7, the
552 flag only has effect if
562 However, in kernels before 2.6.12,
564 succeeded in this case: no mapping was created and the call returned
572 .\" FIXME . Add an example here that uses an anonymous shared region for
573 .\" IPC between parent and child.
575 The following program prints part of the file specified in
576 its first command-line argument to standard output.
577 The range of bytes to be printed is specified via offset and length
578 values in the second and third command-line arguments.
579 The program creates a memory mapping of the required
580 pages of the file and then uses
582 to output the desired bytes.
585 #include <sys/mman.h>
586 #include <sys/stat.h>
592 #define handle_error(msg) \\
593 do { perror(msg); exit(EXIT_FAILURE); } while (0)
596 main(int argc, char *argv[])
601 off_t offset, pa_offset;
605 if (argc < 3 || argc > 4) {
606 fprintf(stderr, "%s file offset [length]\\n", argv[0]);
610 fd = open(argv[1], O_RDONLY);
612 handle_error("open");
614 if (fstat(fd, &sb) == \-1) /* To obtain file size */
615 handle_error("fstat");
617 offset = atoi(argv[2]);
618 pa_offset = offset & ~(sysconf(_SC_PAGE_SIZE) \- 1);
619 /* offset for mmap() must be page aligned */
621 if (offset >= sb.st_size) {
622 fprintf(stderr, "offset is past end of file\\n");
627 length = atoi(argv[3]);
628 if (offset + length > sb.st_size)
629 length = sb.st_size \- offset;
630 /* Can\(aqt display bytes past end of file */
632 } else { /* No length arg ==> display to end of file */
633 length = sb.st_size \- offset;
636 addr = mmap(NULL, length + offset \- pa_offset, PROT_READ,
637 MAP_PRIVATE, fd, pa_offset);
638 if (addr == MAP_FAILED)
639 handle_error("mmap");
641 s = write(STDOUT_FILENO, addr + offset \- pa_offset, length);
644 handle_error("write");
646 fprintf(stderr, "partial write");
661 .BR remap_file_pages (2),
667 B.O. Gallmeister, POSIX.4, O'Reilly, pp. 128-129 and 389-391.
669 .\" Repeat after me: private read-only mappings are 100% equivalent to
670 .\" shared read-only mappings. No ifs, buts, or maybes. -- Linus