1 .\" Copyright (C) 1996 Andries Brouwer <aeb@cwi.nl>
2 .\" and Copyright (C) 2006, 2007 Michael Kerrisk <mtk.manpages@gmail.com>
4 .\" %%%LICENSE_START(VERBATIM)
5 .\" Permission is granted to make and distribute verbatim copies of this
6 .\" manual provided the copyright notice and this permission notice are
7 .\" preserved on all copies.
9 .\" Permission is granted to copy and distribute modified versions of this
10 .\" manual under the conditions for verbatim copying, provided that the
11 .\" entire resulting derived work is distributed under the terms of a
12 .\" permission notice identical to this one.
14 .\" Since the Linux kernel and libraries are constantly changing, this
15 .\" manual page may be incorrect or out-of-date. The author(s) assume no
16 .\" responsibility for errors or omissions, or for damages resulting from
17 .\" the use of the information contained herein. The author(s) may not
18 .\" have taken the same level of care in the production of this manual,
19 .\" which is licensed free of charge, as they might when working
22 .\" Formatted or processed versions of this manual, if unaccompanied by
23 .\" 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 2015-01-22 "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 );
53 See NOTES for information on feature test macro requirements.
56 creates a new mapping in the virtual address space of
58 The starting address for the new mapping is specified in
62 argument specifies the length of the mapping.
67 then the kernel chooses the address at which to create the mapping;
68 this is the most portable method of creating a new mapping.
72 then the kernel takes it as a hint about where to place the mapping;
73 on Linux, the mapping will be created at a nearby page boundary.
74 .\" Before Linux 2.6.24, the address was rounded up to the next page
75 .\" boundary; since 2.6.24, it is rounded down!
76 The address of the new mapping is returned as the result of the call.
78 The contents of a file mapping (as opposed to an anonymous mapping; see
80 below), are initialized using
82 bytes starting at offset
84 in the file (or other object) referred to by the file descriptor
87 must be a multiple of the page size as returned by
88 .IR sysconf(_SC_PAGE_SIZE) .
92 argument describes the desired memory protection of the mapping
93 (and must not conflict with the open mode of the file).
96 or the bitwise OR of one or more of the following flags:
99 Pages may be executed.
105 Pages may be written.
108 Pages may not be accessed.
112 argument determines whether updates to the mapping
113 are visible to other processes mapping the same region,
114 and whether updates are carried through to the underlying file.
115 This behavior is determined by including exactly one
116 of the following values in
121 Updates to the mapping are visible to other processes that map this file,
122 and are carried through to the underlying file.
123 The file may not actually be updated until
130 Create a private copy-on-write mapping.
131 Updates to the mapping are not visible to other processes
132 mapping the same file, and are not carried through to
134 It is unspecified whether changes made to the file after the
136 call are visible in the mapped region.
138 Both of these flags are described in POSIX.1-2001.
140 In addition, zero or more of the following values can be ORed in
143 .BR MAP_32BIT " (since Linux 2.4.20, 2.6)"
144 Put the mapping into the first 2 Gigabytes of the process address space.
145 This flag is supported only on x86-64, for 64-bit programs.
146 It was added to allow thread stacks to be allocated somewhere
147 in the first 2GB of memory,
148 so as to improve context-switch performance on some early
150 .\" See http://lwn.net/Articles/294642 "Tangled up in threads", 19 Aug 08
151 Modern x86-64 processors no longer have this performance problem,
152 so use of this flag is not required on those systems.
165 The mapping is not backed by any file;
166 its contents are initialized to zero.
171 arguments are ignored;
172 however, some implementations require
179 and portable applications should ensure this.
184 is supported on Linux only since kernel 2.4.
187 This flag is ignored.
188 .\" Introduced in 1.1.36, removed in 1.3.24.
189 (Long ago, it signaled that attempts to write to the underlying file
192 But this was a source of denial-of-service attacks.)
195 This flag is ignored.
196 .\" Introduced in 1.1.38, removed in 1.3.24. Flag tested in proc_follow_link.
197 .\" (Long ago, it signaled that the underlying file is an executable.
198 .\" However, that information was not really used anywhere.)
199 .\" Linus talked about DOS related to MAP_EXECUTABLE, but he was thinking of
205 .\" On some systems, this was required as the opposite of
206 .\" MAP_ANONYMOUS -- mtk, 1 May 2007
211 as a hint: place the mapping at exactly that address.
213 must be a multiple of the page size.
214 If the memory region specified by
218 overlaps pages of any existing mapping(s), then the overlapped
219 part of the existing mapping(s) will be discarded.
220 If the specified address cannot be used,
223 Because requiring a fixed address for a mapping is less portable,
224 the use of this option is discouraged.
228 Indicates to the kernel virtual memory system that the mapping
229 should extend downward in memory.
231 .BR MAP_HUGETLB " (since Linux 2.6.32)"
232 Allocate the mapping using "huge pages."
233 See the Linux kernel source file
234 .I Documentation/vm/hugetlbpage.txt
235 for further information.
237 .BR MAP_LOCKED " (since Linux 2.5.37)"
238 Lock the pages of the mapped region into memory in the manner of
240 This flag is ignored in older kernels.
241 .\" If set, the mapped pages will not be swapped out.
243 .BR MAP_NONBLOCK " (since Linux 2.5.46)"
244 Only meaningful in conjunction with
246 Don't perform read-ahead:
247 create page tables entries only for pages
248 that are already present in RAM.
249 Since Linux 2.6.23, this flag causes
252 One day the combination of
256 may be reimplemented.
259 Do not reserve swap space for this mapping.
260 When swap space is reserved, one has the guarantee
261 that it is possible to modify the mapping.
262 When swap space is not reserved one might get
265 if no physical memory is available.
266 See also the discussion of the file
267 .I /proc/sys/vm/overcommit_memory
270 In kernels before 2.6, this flag had effect only for
271 private writable mappings.
273 .BR MAP_POPULATE " (since Linux 2.5.46)"
274 Populate (prefault) page tables for a mapping.
275 For a file mapping, this causes read-ahead on the file.
276 Later accesses to the mapping will not be blocked by page faults.
278 is supported for private mappings only since Linux 2.6.23.
280 .BR MAP_STACK " (since Linux 2.6.27)"
281 Allocate the mapping at an address suitable for a process
283 This flag is currently a no-op,
284 but is used in the glibc threading implementation so that
285 if some architectures require special treatment for stack allocations,
286 support can later be transparently implemented for glibc.
287 .\" See http://lwn.net/Articles/294642 "Tangled up in threads", 19 Aug 08
288 .\" commit cd98a04a59e2f94fa64d5bf1e26498d27427d5e7
289 .\" http://thread.gmane.org/gmane.linux.kernel/720412
290 .\" "pthread_create() slow for many threads; also time to revisit 64b
291 .\" context switch optimization?"
293 .BR MAP_UNINITIALIZED " (since Linux 2.6.33)"
294 Don't clear anonymous pages.
295 This flag is intended to improve performance on embedded devices.
296 This flag is honored only if the kernel was configured with the
297 .B CONFIG_MMAP_ALLOW_UNINITIALIZED
299 Because of the security implications,
300 that option is normally enabled only on embedded devices
301 (i.e., devices where one has complete control of the contents of user memory).
303 Of the above flags, only
305 is specified in POSIX.1-2001.
306 However, most systems also support
311 Some systems document the additional flags
322 with the same attributes.
324 A file is mapped in multiples of the page size.
325 For a file that is not
326 a multiple of the page size, the remaining memory is zeroed when mapped,
327 and writes to that region are not written out to the file.
329 changing the size of the underlying file of a mapping on the pages that
330 correspond to added or removed regions of the file is unspecified.
334 system call deletes the mappings for the specified address range, and
335 causes further references to addresses within the range to generate
336 invalid memory references.
337 The region is also automatically unmapped
338 when the process is terminated.
339 On the other hand, closing the file
340 descriptor does not unmap the region.
344 must be a multiple of the page size.
345 All pages containing a part
346 of the indicated range are unmapped, and subsequent references
347 to these pages will generate
349 It is not an error if the
350 indicated range does not contain any mapped pages.
351 .SS Timestamps changes for file-backed mappings
352 For file-backed mappings, the
354 field for the mapped file may be updated at any time between the
356 and the corresponding unmapping; the first reference to a mapped
357 page will update the field if it has not been already.
363 field for a file mapped with
367 will be updated after
368 a write to the mapped region, and before a subsequent
378 returns a pointer to the mapped area.
382 .IR "(void\ *)\ \-1" )
385 is set appropriately.
388 returns 0, on failure \-1, and
395 A file descriptor refers to a non-regular file.
396 Or a file mapping was requested, but
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 filesystem 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 filesystem that
466 .\" (Since 2.4.25 / 2.6.0.)
469 The operation was prevented by a file seal; see
474 was set but the object specified by
479 On 32-bit architecture together with the large file extension
482 the number of pages used for
484 plus number of pages used for
490 Use of a mapped region can result in these signals:
493 Attempted write into a region mapped as read-only.
496 Attempted access to a portion of the buffer that does not correspond
497 to the file (for example, beyond the end of the file, including the
498 case where another process has truncated the file).
500 SVr4, 4.4BSD, POSIX.1-2001.
501 .\" SVr4 documents additional error codes ENXIO and ENODEV.
502 .\" SUSv2 documents additional error codes EMFILE and EOVERFLOW.
504 On POSIX systems on which
510 .B _POSIX_MAPPED_FILES
511 is defined in \fI<unistd.h>\fP to a value greater than 0.
514 .\" POSIX.1-2001: It shall be defined to -1 or 0 or 200112L.
515 .\" -1: unavailable, 0: ask using sysconf().
516 .\" glibc defines it to 1.
518 On some hardware architectures (e.g., i386),
522 It is architecture dependent whether
527 Portable programs should always set
529 if they intend to execute code in the new mapping.
531 The portable way to create a mapping is to specify
533 as 0 (NULL), and omit
537 In this case, the system chooses the address for the mapping;
538 the address is chosen so as not to conflict with any existing mapping,
542 flag is specified, and
544 is 0 (NULL), then the mapped address will be 0 (NULL).
548 constants are defined only if either
556 and requiring that macro specifically would have been more logical,
557 since these flags are all Linux-specific.)
558 The relevant flags are:
575 .SS C library/kernel ABI differences
576 This page describes the interface provided by the glibc
579 Originally, this function invoked a system call of the same name.
580 Since kernel 2.4, that system call has been superseded by
583 .\" Since around glibc 2.1/2.2, depending on the platform.
586 wrapper function invokes
588 with a suitably adjusted value for
591 On Linux there are no guarantees like those suggested above under
593 By default, any process can be killed
594 at any moment when the system runs out of memory.
596 In kernels before 2.6.7, the
598 flag has effect only if
608 However, in kernels before 2.6.12,
610 succeeded in this case: no mapping was created and the call returned
618 POSIX specifies that the system shall always
619 zero fill any partial page at the end
620 of the object and that system will never write any modification of the
621 object beyond its end.
622 On Linux, when you write data to such partial page after the end
623 of the object, the data stays in the page cache even after the file
624 is closed and unmapped
625 and even though the data is never written to the file itself,
626 subsequent mappings may see the modified content.
627 In some cases, this could be fixed by calling
629 before the unmap takes place;
630 however, this doesn't work on tmpfs
631 (for example, when using POSIX shared memory interface documented in
632 .BR shm_overview (7)).
634 .\" FIXME . Add an example here that uses an anonymous shared region for
635 .\" IPC between parent and child.
637 The following program prints part of the file specified in
638 its first command-line argument to standard output.
639 The range of bytes to be printed is specified via offset and length
640 values in the second and third command-line arguments.
641 The program creates a memory mapping of the required
642 pages of the file and then uses
644 to output the desired bytes.
647 #include <sys/mman.h>
648 #include <sys/stat.h>
654 #define handle_error(msg) \\
655 do { perror(msg); exit(EXIT_FAILURE); } while (0)
658 main(int argc, char *argv[])
663 off_t offset, pa_offset;
667 if (argc < 3 || argc > 4) {
668 fprintf(stderr, "%s file offset [length]\\n", argv[0]);
672 fd = open(argv[1], O_RDONLY);
674 handle_error("open");
676 if (fstat(fd, &sb) == \-1) /* To obtain file size */
677 handle_error("fstat");
679 offset = atoi(argv[2]);
680 pa_offset = offset & ~(sysconf(_SC_PAGE_SIZE) \- 1);
681 /* offset for mmap() must be page aligned */
683 if (offset >= sb.st_size) {
684 fprintf(stderr, "offset is past end of file\\n");
689 length = atoi(argv[3]);
690 if (offset + length > sb.st_size)
691 length = sb.st_size \- offset;
692 /* Can\(aqt display bytes past end of file */
694 } else { /* No length arg ==> display to end of file */
695 length = sb.st_size \- offset;
698 addr = mmap(NULL, length + offset \- pa_offset, PROT_READ,
699 MAP_PRIVATE, fd, pa_offset);
700 if (addr == MAP_FAILED)
701 handle_error("mmap");
703 s = write(STDOUT_FILENO, addr + offset \- pa_offset, length);
706 handle_error("write");
708 fprintf(stderr, "partial write");
717 .BR memfd_create (2),
724 .BR remap_file_pages (2),
730 The descriptions of the following files in
732 .IR /proc/[pid]/maps ,
733 .IR /proc/[pid]/map_files ,
735 .IR /proc/[pid]/smaps .
737 B.O. Gallmeister, POSIX.4, O'Reilly, pp. 128-129 and 389-391.
739 .\" Repeat after me: private read-only mappings are 100% equivalent to
740 .\" shared read-only mappings. No ifs, buts, or maybes. -- Linus
742 This page is part of release 3.79 of the Linux
745 A description of the project,
746 information about reporting bugs,
747 and the latest version of this page,
749 \%http://www.kernel.org/doc/man\-pages/.