man2/mmap.2

   1 .\" Copyright (C) 1996 Andries Brouwer <aeb@cwi.nl>
   2 .\" and Copyright (C) 2006, 2007 Michael Kerrisk <mtk.manpages@gmail.com>
   3 .\"
   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.
   8 .\"
   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.
  13 .\"
  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  15 .\" manual page may be incorrect or out-of-date.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
  20 .\" professionally.
  21 .\"
  22 .\" Formatted or processed versions of this manual, if unaccompanied by
  23 .\" the source, must acknowledge the copyright and authors of this work.
  24 .\" %%%LICENSE_END
  25 .\"
  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
  39 .\"
  40 .TH MMAP 2 2016-07-17 "Linux" "Linux Programmer's Manual"
  41 .SH NAME
  42 mmap, munmap \- map or unmap files or devices into memory
  43 .SH SYNOPSIS
  44 .nf
  45 .B #include <sys/mman.h>
  46 .sp
  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 );
  51 .fi
  52
  53 See NOTES for information on feature test macro requirements.
  54 .SH DESCRIPTION
  55 .BR mmap ()
  56 creates a new mapping in the virtual address space of
  57 the calling process.
  58 The starting address for the new mapping is specified in
  59 .IR addr .
  60 The
  61 .I length
  62 argument specifies the length of the mapping.
  63
  64 If
  65 .I addr
  66 is NULL,
  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.
  69 If
  70 .I addr
  71 is not NULL,
  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.
  77
  78 The contents of a file mapping (as opposed to an anonymous mapping; see
  79 .B MAP_ANONYMOUS
  80 below), are initialized using
  81 .I length
  82 bytes starting at offset
  83 .I offset
  84 in the file (or other object) referred to by the file descriptor
  85 .IR fd .
  86 .I offset
  87 must be a multiple of the page size as returned by
  88 .IR sysconf(_SC_PAGE_SIZE) .
  89 .LP
  90 The
  91 .I prot
  92 argument describes the desired memory protection of the mapping
  93 (and must not conflict with the open mode of the file).
  94 It is either
  95 .B PROT_NONE
  96 or the bitwise OR of one or more of the following flags:
  97 .TP 1.1i
  98 .B PROT_EXEC
  99 Pages may be executed.
 100 .TP
 101 .B PROT_READ
 102 Pages may be read.
 103 .TP
 104 .B PROT_WRITE
 105 Pages may be written.
 106 .TP
 107 .B PROT_NONE
 108 Pages may not be accessed.
 109 .LP
 110 The
 111 .I flags
 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
 117 .IR flags :
 118 .TP
 119 .B MAP_SHARED
 120 Share this mapping.
 121 Updates to the mapping are visible to other processes that map this file,
 122 and are carried through to the underlying file.
 123 (To precisely control when updates are carried through
 124 to the underlying file requires the use of
 125 .BR msync (2).)
 126 .TP
 127 .B MAP_PRIVATE
 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
 131 the underlying file.
 132 It is unspecified whether changes made to the file after the
 133 .BR mmap ()
 134 call are visible in the mapped region.
 135 .LP
 136 Both of these flags are described in POSIX.1-2001 and POSIX.1-2008.
 137
 138 In addition, zero or more of the following values can be ORed in
 139 .IR flags :
 140 .TP
 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 supported only 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
 147 64-bit processors.
 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.
 151 The
 152 .B MAP_32BIT
 153 flag is ignored when
 154 .B MAP_FIXED
 155 is set.
 156 .TP
 157 .B MAP_ANON
 158 Synonym for
 159 .BR MAP_ANONYMOUS .
 160 Deprecated.
 161 .TP
 162 .B MAP_ANONYMOUS
 163 The mapping is not backed by any file;
 164 its contents are initialized to zero.
 165 The
 166 .I fd
 167 and
 168 .I offset
 169 arguments are ignored;
 170 however, some implementations require
 171 .I fd
 172 to be \-1 if
 173 .B MAP_ANONYMOUS
 174 (or
 175 .BR MAP_ANON )
 176 is specified,
 177 and portable applications should ensure this.
 178 The use of
 179 .B MAP_ANONYMOUS
 180 in conjunction with
 181 .B MAP_SHARED
 182 is supported on Linux only since kernel 2.4.
 183 .TP
 184 .B MAP_DENYWRITE
 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
 188 should fail with
 189 .BR ETXTBUSY .
 190 But this was a source of denial-of-service attacks.)
 191 .TP
 192 .B MAP_EXECUTABLE
 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
 198 .\" MAP_DENYWRITE?
 199 .TP
 200 .B MAP_FILE
 201 Compatibility flag.
 202 Ignored.
 203 .\" On some systems, this was required as the opposite of
 204 .\" MAP_ANONYMOUS -- mtk, 1 May 2007
 205 .TP
 206 .B MAP_FIXED
 207 Don't interpret
 208 .I addr
 209 as a hint: place the mapping at exactly that address.
 210 .I addr
 211 must be a multiple of the page size.
 212 If the memory region specified by
 213 .I addr
 214 and
 215 .I len
 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,
 219 .BR mmap ()
 220 will fail.
 221 Because requiring a fixed address for a mapping is less portable,
 222 the use of this option is discouraged.
 223 .TP
 224 .B MAP_GROWSDOWN
 225 Used for stacks.
 226 Indicates to the kernel virtual memory system that the mapping
 227 should extend downward in memory.
 228 .TP
 229 .BR MAP_HUGETLB " (since Linux 2.6.32)"
 230 Allocate the mapping using "huge pages."
 231 See the Linux kernel source file
 232 .I Documentation/vm/hugetlbpage.txt
 233 for further information, as well as NOTES, below.
 234 .TP
 235 .BR MAP_HUGE_2MB ", " MAP_HUGE_1GB " (since Linux 3.8)"
 236 .\" See https://lwn.net/Articles/533499/
 237 Used in conjunction with
 238 .B MAP_HUGETLB
 239 to select alternative hugetlb page sizes (respectively, 2 MB and 1 GB)
 240 on systems that support multiple hugetlb page sizes.
 241
 242 More generally, the desired huge page size can be configured by encoding
 243 the base-2 logarithm of the desired page size in the six bits at the offset
 244 .BR MAP_HUGE_SHIFT .
 245 (A value of zero in this bit field provides the default huge page size;
 246 the default huge page size can be discovered vie the
 247 .I Hugepagesize
 248 field exposed by
 249 .IR /proc/meminfo .)
 250 Thus, the above two constants are defined as:
 251
 252 .nf
 253 .in +4n
 254 #define MAP_HUGE_2MB    (21 << MAP_HUGE_SHIFT)
 255 #define MAP_HUGE_1GB    (30 << MAP_HUGE_SHIFT)
 256 .in
 257 .fi
 258
 259 The range of huge page sizes that are supported by the system
 260 can be discovered by listing the subdirectories in
 261 .IR /sys/kernel/mm/hugepages .
 262 .TP
 263 .BR MAP_LOCKED " (since Linux 2.5.37)"
 264 Mark the mmaped region to be locked in the same way as
 265 .BR mlock (2).
 266 This implementation will try to populate (prefault) the whole range but
 267 the mmap call doesn't fail with
 268 .B ENOMEM
 269 if this fails.
 270 Therefore major faults might happen later on.
 271 So the semantic is not as strong as
 272 .BR mlock (2).
 273 One should use
 274 .BR mmap (2)
 275 plus
 276 .BR mlock (2)
 277 when major faults are not acceptable after the initialization of the mapping.
 278 The
 279 .BR MAP_LOCKED
 280 flag is ignored in older kernels.
 281 .\" If set, the mapped pages will not be swapped out.
 282 .TP
 283 .BR MAP_NONBLOCK " (since Linux 2.5.46)"
 284 Only meaningful in conjunction with
 285 .BR MAP_POPULATE .
 286 Don't perform read-ahead:
 287 create page tables entries only for pages
 288 that are already present in RAM.
 289 Since Linux 2.6.23, this flag causes
 290 .BR MAP_POPULATE
 291 to do nothing.
 292 One day, the combination of
 293 .BR MAP_POPULATE
 294 and
 295 .BR MAP_NONBLOCK
 296 may be reimplemented.
 297 .TP
 298 .B MAP_NORESERVE
 299 Do not reserve swap space for this mapping.
 300 When swap space is reserved, one has the guarantee
 301 that it is possible to modify the mapping.
 302 When swap space is not reserved one might get
 303 .B SIGSEGV
 304 upon a write
 305 if no physical memory is available.
 306 See also the discussion of the file
 307 .I /proc/sys/vm/overcommit_memory
 308 in
 309 .BR proc (5).
 310 In kernels before 2.6, this flag had effect only for
 311 private writable mappings.
 312 .TP
 313 .BR MAP_POPULATE " (since Linux 2.5.46)"
 314 Populate (prefault) page tables for a mapping.
 315 For a file mapping, this causes read-ahead on the file.
 316 This will help to reduce blocking on page faults later.
 317 .BR MAP_POPULATE
 318 is supported for private mappings only since Linux 2.6.23.
 319 .TP
 320 .BR MAP_STACK " (since Linux 2.6.27)"
 321 Allocate the mapping at an address suitable for a process
 322 or thread stack.
 323 This flag is currently a no-op,
 324 but is used in the glibc threading implementation so that
 325 if some architectures require special treatment for stack allocations,
 326 support can later be transparently implemented for glibc.
 327 .\" See http://lwn.net/Articles/294642 "Tangled up in threads", 19 Aug 08
 328 .\" commit cd98a04a59e2f94fa64d5bf1e26498d27427d5e7
 329 .\" http://thread.gmane.org/gmane.linux.kernel/720412
 330 .\" "pthread_create() slow for many threads; also time to revisit 64b
 331 .\"  context switch optimization?"
 332 .TP
 333 .BR MAP_UNINITIALIZED " (since Linux 2.6.33)"
 334 Don't clear anonymous pages.
 335 This flag is intended to improve performance on embedded devices.
 336 This flag is honored only if the kernel was configured with the
 337 .B CONFIG_MMAP_ALLOW_UNINITIALIZED
 338 option.
 339 Because of the security implications,
 340 that option is normally enabled only on embedded devices
 341 (i.e., devices where one has complete control of the contents of user memory).
 342 .LP
 343 Of the above flags, only
 344 .B MAP_FIXED
 345 is specified in POSIX.1-2001 and POSIX.1-2008.
 346 However, most systems also support
 347 .B MAP_ANONYMOUS
 348 (or its synonym
 349 .BR MAP_ANON ).
 350 .LP
 351 Some systems document the additional flags
 352 .BR MAP_AUTOGROW ,
 353 .BR MAP_AUTORESRV ,
 354 .BR MAP_COPY ,
 355 and
 356 .BR MAP_LOCAL .
 357 .LP
 358 Memory mapped by
 359 .BR mmap ()
 360 is preserved across
 361 .BR fork (2),
 362 with the same attributes.
 363 .LP
 364 A file is mapped in multiples of the page size.
 365 For a file that is not
 366 a multiple of the page size, the remaining memory is zeroed when mapped,
 367 and writes to that region are not written out to the file.
 368 The effect of
 369 changing the size of the underlying file of a mapping on the pages that
 370 correspond to added or removed regions of the file is unspecified.
 371 .SS munmap()
 372 The
 373 .BR munmap ()
 374 system call deletes the mappings for the specified address range, and
 375 causes further references to addresses within the range to generate
 376 invalid memory references.
 377 The region is also automatically unmapped
 378 when the process is terminated.
 379 On the other hand, closing the file
 380 descriptor does not unmap the region.
 381 .LP
 382 The address
 383 .I addr
 384 must be a multiple of the page size (but
 385 .I length
 386 need not be).
 387 All pages containing a part
 388 of the indicated range are unmapped, and subsequent references
 389 to these pages will generate
 390 .BR SIGSEGV .
 391 It is not an error if the
 392 indicated range does not contain any mapped pages.
 393 .SH RETURN VALUE
 394 On success,
 395 .BR mmap ()
 396 returns a pointer to the mapped area.
 397 On error, the value
 398 .B MAP_FAILED
 399 (that is,
 400 .IR "(void\ *)\ \-1" )
 401 is returned, and
 402 .I errno
 403 is set to indicate the cause of the error.
 404
 405 On success,
 406 .BR munmap ()
 407 returns 0.
 408 On failure, it returns \-1, and
 409 .I errno
 410 is set to indicate the cause of the error (probably to
 411 .BR EINVAL ).
 412 .SH ERRORS
 413 .TP
 414 .B EACCES
 415 A file descriptor refers to a non-regular file.
 416 Or a file mapping was requested, but
 417 .I fd
 418 is not open for reading.
 419 Or
 420 .B MAP_SHARED
 421 was requested and
 422 .B PROT_WRITE
 423 is set, but
 424 .I fd
 425 is not open in read/write
 426 .RB ( O_RDWR )
 427 mode.
 428 Or
 429 .B PROT_WRITE
 430 is set, but the file is append-only.
 431 .TP
 432 .B EAGAIN
 433 The file has been locked, or too much memory has been locked (see
 434 .BR setrlimit (2)).
 435 .TP
 436 .B EBADF
 437 .I fd
 438 is not a valid file descriptor (and
 439 .B MAP_ANONYMOUS
 440 was not set).
 441 .TP
 442 .B EINVAL
 443 We don't like
 444 .IR addr ,
 445 .IR length ,
 446 or
 447 .I offset
 448 (e.g., they are too large, or not aligned on a page boundary).
 449 .TP
 450 .B EINVAL
 451 (since Linux 2.6.12)
 452 .I length
 453 was 0.
 454 .TP
 455 .B EINVAL
 456 .I flags
 457 contained neither
 458 .B MAP_PRIVATE
 459 or
 460 .BR MAP_SHARED ,
 461 or contained both of these values.
 462 .TP
 463 .B ENFILE
 464 .\" This is for shared anonymous segments
 465 .\" [2.6.7] shmem_zero_setup()-->shmem_file_setup()-->get_empty_filp()
 466 The system-wide limit on the total number of open files has been reached.
 467 .\" .TP
 468 .\" .B ENOEXEC
 469 .\" A file could not be mapped for reading.
 470 .TP
 471 .B ENODEV
 472 The underlying filesystem of the specified file does not support
 473 memory mapping.
 474 .TP
 475 .B ENOMEM
 476 No memory is available.
 477 .TP
 478 .B ENOMEM
 479 The process's maximum number of mappings would have been exceeded.
 480 This error can also occur for
 481 .BR munmap (2),
 482 when unmapping a region in the middle of an existing mapping,
 483 since this results in two smaller mappings on either side of
 484 the region being unmapped.
 485 .TP
 486 .B EPERM
 487 The
 488 .I prot
 489 argument asks for
 490 .B PROT_EXEC
 491 but the mapped area belongs to a file on a filesystem that
 492 was mounted no-exec.
 493 .\" (Since 2.4.25 / 2.6.0.)
 494 .TP
 495 .B EPERM
 496 The operation was prevented by a file seal; see
 497 .BR fcntl (2).
 498 .TP
 499 .B ETXTBSY
 500 .B MAP_DENYWRITE
 501 was set but the object specified by
 502 .I fd
 503 is open for writing.
 504 .TP
 505 .B EOVERFLOW
 506 On 32-bit architecture together with the large file extension
 507 (i.e., using 64-bit
 508 .IR off_t ):
 509 the number of pages used for
 510 .I length
 511 plus number of pages used for
 512 .I offset
 513 would overflow
 514 .I "unsigned long"
 515 (32 bits).
 516 .LP
 517 Use of a mapped region can result in these signals:
 518 .TP
 519 .B SIGSEGV
 520 Attempted write into a region mapped as read-only.
 521 .TP
 522 .B SIGBUS
 523 Attempted access to a portion of the buffer that does not correspond
 524 to the file (for example, beyond the end of the file, including the
 525 case where another process has truncated the file).
 526 .SH ATTRIBUTES
 527 For an explanation of the terms used in this section, see
 528 .BR attributes (7).
 529 .TS
 530 allbox;
 531 lbw18 lb lb
 532 l l l.
 533 Interface       Attribute       Value
 534 T{
 535 .BR mmap (),
 536 .BR munmap ()
 537 T}      Thread safety   MT-Safe
 538 .TE
 539 .SH CONFORMING TO
 540 POSIX.1-2001, POSIX.1-2008, SVr4, 4.4BSD.
 541 .\" SVr4 documents additional error codes ENXIO and ENODEV.
 542 .\" SUSv2 documents additional error codes EMFILE and EOVERFLOW.
 543 .SH AVAILABILITY
 544 On POSIX systems on which
 545 .BR mmap (),
 546 .BR msync (2),
 547 and
 548 .BR munmap ()
 549 are available,
 550 .B _POSIX_MAPPED_FILES
 551 is defined in \fI<unistd.h>\fP to a value greater than 0.
 552 (See also
 553 .BR sysconf (3).)
 554 .\" POSIX.1-2001: It shall be defined to -1 or 0 or 200112L.
 555 .\" -1: unavailable, 0: ask using sysconf().
 556 .\" glibc defines it to 1.
 557 .SH NOTES
 558 On some hardware architectures (e.g., i386),
 559 .B PROT_WRITE
 560 implies
 561 .BR PROT_READ .
 562 It is architecture dependent whether
 563 .B PROT_READ
 564 implies
 565 .B PROT_EXEC
 566 or not.
 567 Portable programs should always set
 568 .B PROT_EXEC
 569 if they intend to execute code in the new mapping.
 570
 571 The portable way to create a mapping is to specify
 572 .I addr
 573 as 0 (NULL), and omit
 574 .B MAP_FIXED
 575 from
 576 .IR flags .
 577 In this case, the system chooses the address for the mapping;
 578 the address is chosen so as not to conflict with any existing mapping,
 579 and will not be 0.
 580 If the
 581 .B MAP_FIXED
 582 flag is specified, and
 583 .I addr
 584 is 0 (NULL), then the mapped address will be 0 (NULL).
 585
 586 Certain
 587 .I flags
 588 constants are defined only if suitable feature test macros are defined
 589 (possibly by default):
 590 .BR _DEFAULT_SOURCE
 591 with glibc 2.19 or later;
 592 or
 593 .BR _BSD_SOURCE
 594 or
 595 .BR _SVID_SOURCE
 596 in glibc 2.19 and earlier.
 597 (Requiring
 598 .BR _GNU_SOURCE
 599 also suffices,
 600 and requiring that macro specifically would have been more logical,
 601 since these flags are all Linux-specific.)
 602 The relevant flags are:
 603 .BR MAP_32BIT ,
 604 .BR MAP_ANONYMOUS
 605 (and the synonym
 606 .BR MAP_ANON ),
 607 .BR MAP_DENYWRITE ,
 608 .BR MAP_EXECUTABLE ,
 609 .BR MAP_FILE ,
 610 .BR MAP_GROWSDOWN ,
 611 .BR MAP_HUGETLB ,
 612 .BR MAP_LOCKED ,
 613 .BR MAP_NONBLOCK ,
 614 .BR MAP_NORESERVE ,
 615 .BR MAP_POPULATE ,
 616 and
 617 .BR MAP_STACK .
 618 .\"
 619 .SS Timestamps changes for file-backed mappings
 620 For file-backed mappings, the
 621 .I st_atime
 622 field for the mapped file may be updated at any time between the
 623 .BR mmap ()
 624 and the corresponding unmapping; the first reference to a mapped
 625 page will update the field if it has not been already.
 626 .LP
 627 The
 628 .I st_ctime
 629 and
 630 .I st_mtime
 631 field for a file mapped with
 632 .B PROT_WRITE
 633 and
 634 .B MAP_SHARED
 635 will be updated after
 636 a write to the mapped region, and before a subsequent
 637 .BR msync (2)
 638 with the
 639 .B MS_SYNC
 640 or
 641 .B MS_ASYNC
 642 flag, if one occurs.
 643 .\"
 644 .SS Huge page (Huge TLB) mappings
 645 For mappings that employ huge pages, the requirements for the arguments of
 646 .BR mmap ()
 647 and
 648 .BR munmap ()
 649 differ somewhat from the requirements for mappings
 650 that use the native system page size.
 651
 652 For
 653 .BR mmap (),
 654 .I offset
 655 must be a multiple of the underlying huge page size.
 656 The system automatically aligns
 657 .I length
 658 to be a multiple of the underlying huge page size.
 659
 660 For
 661 .BR munmap (),
 662 .I addr
 663 and
 664 .I length
 665 must both be a multiple of the underlying huge page size.
 666 .\"
 667 .SS C library/kernel differences
 668 This page describes the interface provided by the glibc
 669 .BR mmap ()
 670 wrapper function.
 671 Originally, this function invoked a system call of the same name.
 672 Since kernel 2.4, that system call has been superseded by
 673 .BR mmap2 (2),
 674 and nowadays
 675 .\" Since around glibc 2.1/2.2, depending on the platform.
 676 the glibc
 677 .BR mmap ()
 678 wrapper function invokes
 679 .BR mmap2 (2)
 680 with a suitably adjusted value for
 681 .IR offset .
 682 .SH BUGS
 683 On Linux, there are no guarantees like those suggested above under
 684 .BR MAP_NORESERVE .
 685 By default, any process can be killed
 686 at any moment when the system runs out of memory.
 687
 688 In kernels before 2.6.7, the
 689 .B MAP_POPULATE
 690 flag has effect only if
 691 .I prot
 692 is specified as
 693 .BR PROT_NONE .
 694
 695 SUSv3 specifies that
 696 .BR mmap ()
 697 should fail if
 698 .I length
 699 is 0.
 700 However, in kernels before 2.6.12,
 701 .BR mmap ()
 702 succeeded in this case: no mapping was created and the call returned
 703 .IR addr .
 704 Since kernel 2.6.12,
 705 .BR mmap ()
 706 fails with the error
 707 .B EINVAL
 708 for this case.
 709
 710 POSIX specifies that the system shall always
 711 zero fill any partial page at the end
 712 of the object and that system will never write any modification of the
 713 object beyond its end.
 714 On Linux, when you write data to such partial page after the end
 715 of the object, the data stays in the page cache even after the file
 716 is closed and unmapped
 717 and even though the data is never written to the file itself,
 718 subsequent mappings may see the modified content.
 719 In some cases, this could be fixed by calling
 720 .BR msync (2)
 721 before the unmap takes place;
 722 however, this doesn't work on tmpfs
 723 (for example, when using POSIX shared memory interface documented in
 724 .BR shm_overview (7)).
 725 .SH EXAMPLE
 726 .\" FIXME . Add an example here that uses an anonymous shared region for
 727 .\" IPC between parent and child.
 728 .PP
 729 The following program prints part of the file specified in
 730 its first command-line argument to standard output.
 731 The range of bytes to be printed is specified via offset and length
 732 values in the second and third command-line arguments.
 733 The program creates a memory mapping of the required
 734 pages of the file and then uses
 735 .BR write (2)
 736 to output the desired bytes.
 737 .SS Program source
 738 .nf
 739 #include <sys/mman.h>
 740 #include <sys/stat.h>
 741 #include <fcntl.h>
 742 #include <stdio.h>
 743 #include <stdlib.h>
 744 #include <unistd.h>
 745
 746 #define handle_error(msg) \\
 747     do { perror(msg); exit(EXIT_FAILURE); } while (0)
 748
 749 int
 750 main(int argc, char *argv[])
 751 {
 752     char *addr;
 753     int fd;
 754     struct stat sb;
 755     off_t offset, pa_offset;
 756     size_t length;
 757     ssize_t s;
 758
 759     if (argc < 3 || argc > 4) {
 760         fprintf(stderr, "%s file offset [length]\\n", argv[0]);
 761         exit(EXIT_FAILURE);
 762     }
 763
 764     fd = open(argv[1], O_RDONLY);
 765     if (fd == \-1)
 766         handle_error("open");
 767
 768     if (fstat(fd, &sb) == \-1)           /* To obtain file size */
 769         handle_error("fstat");
 770
 771     offset = atoi(argv[2]);
 772     pa_offset = offset & ~(sysconf(_SC_PAGE_SIZE) \- 1);
 773         /* offset for mmap() must be page aligned */
 774
 775     if (offset >= sb.st_size) {
 776         fprintf(stderr, "offset is past end of file\\n");
 777         exit(EXIT_FAILURE);
 778     }
 779
 780     if (argc == 4) {
 781         length = atoi(argv[3]);
 782         if (offset + length > sb.st_size)
 783             length = sb.st_size \- offset;
 784                 /* Can\(aqt display bytes past end of file */
 785
 786     } else {    /* No length arg ==> display to end of file */
 787         length = sb.st_size \- offset;
 788     }
 789
 790     addr = mmap(NULL, length + offset \- pa_offset, PROT_READ,
 791                 MAP_PRIVATE, fd, pa_offset);
 792     if (addr == MAP_FAILED)
 793         handle_error("mmap");
 794
 795     s = write(STDOUT_FILENO, addr + offset \- pa_offset, length);
 796     if (s != length) {
 797         if (s == \-1)
 798             handle_error("write");
 799
 800         fprintf(stderr, "partial write");
 801         exit(EXIT_FAILURE);
 802     }
 803
 804     munmap(addr, length + offset \- pa_offset);
 805     close(fd);
 806
 807     exit(EXIT_SUCCESS);
 808 }
 809 .fi
 810 .SH SEE ALSO
 811 .BR getpagesize (2),
 812 .BR memfd_create (2),
 813 .BR mincore (2),
 814 .BR mlock (2),
 815 .BR mmap2 (2),
 816 .BR mprotect (2),
 817 .BR mremap (2),
 818 .BR msync (2),
 819 .BR remap_file_pages (2),
 820 .BR setrlimit (2),
 821 .BR shmat (2),
 822 .BR shm_open (3),
 823 .BR shm_overview (7)
 824
 825 The descriptions of the following files in
 826 .BR proc (5):
 827 .IR /proc/[pid]/maps ,
 828 .IR /proc/[pid]/map_files ,
 829 and
 830 .IR /proc/[pid]/smaps .
 831
 832 B.O. Gallmeister, POSIX.4, O'Reilly, pp. 128-129 and 389-391.
 833 .\"
 834 .\" Repeat after me: private read-only mappings are 100% equivalent to
 835 .\" shared read-only mappings. No ifs, buts, or maybes. -- Linus