man7/inotify.7

   1 '\" t
   2 .\" Copyright (C) 2006, 2014 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  15 .\" manual page may be incorrect or out-of-date.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 .TH INOTIFY 7 2014-04-01 "Linux" "Linux Programmer's Manual"
  27 .SH NAME
  28 inotify \- monitoring filesystem events
  29 .SH DESCRIPTION
  30 The
  31 .I inotify
  32 API provides a mechanism for monitoring filesystem events.
  33 Inotify can be used to monitor individual files,
  34 or to monitor directories.
  35 When a directory is monitored, inotify will return events
  36 for the directory itself, and for files inside the directory.
  37
  38 The following system calls are used with this API:
  39 .IP * 3
  40 .BR inotify_init (2)
  41 creates an inotify instance and returns a file descriptor
  42 referring to the inotify instance.
  43 The more recent
  44 .BR inotify_init1 (2)
  45 is like
  46 .BR inotify_init (2),
  47 but has a
  48 .IR flags
  49 argument that provides access to some extra functionality.
  50 .IP *
  51 .BR inotify_add_watch (2)
  52 manipulates the "watch list" associated with an inotify instance.
  53 Each item ("watch") in the watch list specifies the pathname of
  54 a file or directory,
  55 along with some set of events that the kernel should monitor for the
  56 file referred to by that pathname.
  57 .BR inotify_add_watch (2)
  58 either creates a new watch item, or modifies an existing watch.
  59 Each watch has a unique "watch descriptor", an integer
  60 returned by
  61 .BR inotify_add_watch (2)
  62 when the watch is created.
  63 .IP *
  64 When events occur for monitored files and directories,
  65 those events are made available to the application as structured data that
  66 can be read from the inotify file descriptor using
  67 .BR read (2)
  68 (see below).
  69 .IP *
  70 .BR inotify_rm_watch (2)
  71 removes an item from an inotify watch list.
  72 .IP *
  73 When all file descriptors referring to an inotify
  74 instance have been closed (using
  75 .BR close (2)),
  76 the underlying object and its resources are
  77 freed for reuse by the kernel;
  78 all associated watches are automatically freed.
  79 .SS Reading events from an inotify file descriptor
  80 To determine what events have occurred, an application
  81 .BR read (2)s
  82 from the inotify file descriptor.
  83 If no events have so far occurred, then,
  84 assuming a blocking file descriptor,
  85 .BR read (2)
  86 will block until at least one event occurs
  87 (unless interrupted by a signal,
  88 in which case the call fails with the error
  89 .BR EINTR ;
  90 see
  91 .BR signal (7)).
  92
  93 Each successful
  94 .BR read (2)
  95 returns a buffer containing one or more of the following structures:
  96 .in +4n
  97 .nf
  98
  99 struct inotify_event {
 100     int      wd;       /* Watch descriptor */
 101 .\" FIXME . The type of the 'wd' field should probably be "int32_t".
 102 .\" I submitted a patch to fix this.  See the LKML thread
 103 .\" "[patch] Fix type errors in inotify interfaces", 18 Nov 2008
 104 .\" Glibc bug filed: http://sources.redhat.com/bugzilla/show_bug.cgi?id=7040
 105     uint32_t mask;     /* Mask of events */
 106     uint32_t cookie;   /* Unique cookie associating related
 107                           events (for rename(2)) */
 108     uint32_t len;      /* Size of \fIname\fP field */
 109     char     name[];   /* Optional null-terminated name */
 110 };
 111 .fi
 112 .in
 113
 114 .I wd
 115 identifies the watch for which this event occurs.
 116 It is one of the watch descriptors returned by a previous call to
 117 .BR inotify_add_watch (2).
 118
 119 .I mask
 120 contains bits that describe the event that occurred (see below).
 121
 122 .I cookie
 123 is a unique integer that connects related events.
 124 Currently this is used only for rename events, and
 125 allows the resulting pair of
 126 .B IN_MOVED_FROM
 127 and
 128 .B IN_MOVED_TO
 129 events to be connected by the application.
 130 For all other event types,
 131 .I cookie
 132 is set to 0.
 133
 134 The
 135 .I name
 136 field is present only when an event is returned
 137 for a file inside a watched directory;
 138 it identifies the file pathname relative to the watched directory.
 139 This pathname is null-terminated,
 140 and may include further null bytes (\(aq\\0\(aq) to align subsequent reads to a
 141 suitable address boundary.
 142
 143 The
 144 .I len
 145 field counts all of the bytes in
 146 .IR name ,
 147 including the null bytes;
 148 the length of each
 149 .I inotify_event
 150 structure is thus
 151 .IR "sizeof(struct inotify_event)+len" .
 152
 153 The behavior when the buffer given to
 154 .BR read (2)
 155 is too small to return information about the next event depends
 156 on the kernel version: in kernels before 2.6.21,
 157 .BR read (2)
 158 returns 0; since kernel 2.6.21,
 159 .BR read (2)
 160 fails with the error
 161 .BR EINVAL .
 162 Specifying a buffer of size
 163
 164     sizeof(struct inotify_event) + NAME_MAX + 1
 165
 166 will be sufficient to read at least one event.
 167 .SS inotify events
 168 The
 169 .BR inotify_add_watch (2)
 170 .I mask
 171 argument and the
 172 .I mask
 173 field of the
 174 .I inotify_event
 175 structure returned when
 176 .BR read (2)ing
 177 an inotify file descriptor are both bit masks identifying
 178 inotify events.
 179 The following bits can be specified in
 180 .I mask
 181 when calling
 182 .BR inotify_add_watch (2)
 183 and may be returned in the
 184 .I mask
 185 field returned by
 186 .BR read (2):
 187 .RS 4
 188 .TP
 189 .BR IN_ACCESS " (*)"
 190 File was accessed (e.g.,
 191 .BR read (2),
 192 .BR execve (2)).
 193 .TP
 194 .BR IN_ATTRIB " (*)"
 195 Metadata changed\(emfor example, permissions (e.g.,
 196 .BR chmod (2)),
 197 timestamps (e.g.,
 198 .BR utimensat (2)),
 199 extended attributes
 200 .RB ( setxattr (2)),
 201 link count (since Linux 2.6.25; e.g.,
 202 for the target of
 203 .BR link (2)
 204 and for
 205 .BR unlink (2)),
 206 and user/group ID (e.g.,
 207 .BR chown (2)).
 208 .TP
 209 .BR IN_CLOSE_WRITE " (*)"
 210 File opened for writing was closed.
 211 .TP
 212 .BR IN_CLOSE_NOWRITE " (*)"
 213 File not opened for writing was closed.
 214 .TP
 215 .BR IN_CREATE " (*)"
 216 File/directory created in watched directory (e.g.,
 217 .BR open (2)
 218 .BR O_CREAT ,
 219 .BR mkdir (2),
 220 .BR link (2),
 221 .BR symlink (2),
 222 .BR bind (2)
 223 on a UNIX domain socket).
 224 .TP
 225 .BR IN_DELETE " (*)"
 226 File/directory deleted from watched directory.
 227 .TP
 228 .B IN_DELETE_SELF
 229 Watched file/directory was itself deleted.
 230 (This event also occurs if an object is moved to another filesystem,
 231 since
 232 .BR mv (1)
 233 in effect copies the file to the other filesystem and
 234 then deletes it from the original filesystem.)
 235 In addition, an
 236 .B IN_IGNORED
 237 event will subsequently be generated for the watch descriptor.
 238 .TP
 239 .BR IN_MODIFY " (*)"
 240 File was modified (e.g.,
 241 .BR write (2),
 242 .BR truncate (2)).
 243 .TP
 244 .B IN_MOVE_SELF
 245 Watched file/directory was itself moved.
 246 .TP
 247 .BR IN_MOVED_FROM " (*)"
 248 Generated for the directory containing the old filename
 249 when a file is renamed.
 250 .TP
 251 .BR IN_MOVED_TO " (*)"
 252 Generated for the directory containing the new filename
 253 when a file is renamed.
 254 .TP
 255 .BR IN_OPEN " (*)"
 256 File was opened.
 257 .RE
 258 .PP
 259 When monitoring a directory,
 260 the events marked with an asterisk (*) above can occur for
 261 files in the directory, in which case the
 262 .I name
 263 field in the returned
 264 .I inotify_event
 265 structure identifies the name of the file within the directory.
 266 .PP
 267 The
 268 .B IN_ALL_EVENTS
 269 macro is defined as a bit mask of all of the above events.
 270 This macro can be used as the
 271 .I mask
 272 argument when calling
 273 .BR inotify_add_watch (2).
 274
 275 Two additional convenience macros are defined:
 276 .RS 4
 277 .TP
 278 .BR IN_MOVE
 279 Equates to
 280 .BR "IN_MOVED_FROM | IN_MOVED_TO" .
 281 .TP
 282 .BR IN_CLOSE
 283 Equates to
 284 .BR "IN_CLOSE_WRITE | IN_CLOSE_NOWRITE" .
 285 .RE
 286 .PP
 287 The following further bits can be specified in
 288 .I mask
 289 when calling
 290 .BR inotify_add_watch (2):
 291 .RS 4
 292 .TP
 293 .BR IN_DONT_FOLLOW " (since Linux 2.6.15)"
 294 Don't dereference
 295 .I pathname
 296 if it is a symbolic link.
 297 .TP
 298 .BR IN_EXCL_UNLINK " (since Linux 2.6.36)"
 299 .\" commit 8c1934c8d70b22ca8333b216aec6c7d09fdbd6a6
 300 By default, when watching events on the children of a directory,
 301 events are generated for children even after they have been unlinked
 302 from the directory.
 303 This can result in large numbers of uninteresting events for
 304 some applications (e.g., if watching
 305 .IR /tmp ,
 306 in which many applications create temporary files whose
 307 names are immediately unlinked).
 308 Specifying
 309 .B IN_EXCL_UNLINK
 310 changes the default behavior,
 311 so that events are not generated for children after
 312 they have been unlinked from the watched directory.
 313 .TP
 314 .B IN_MASK_ADD
 315 Add (OR) events to watch mask for this pathname if
 316 it already exists (instead of replacing mask).
 317 .TP
 318 .B IN_ONESHOT
 319 Monitor
 320 .I pathname
 321 for one event, then remove from
 322 watch list.
 323 .TP
 324 .BR IN_ONLYDIR " (since Linux 2.6.15)"
 325 Only watch
 326 .I pathname
 327 if it is a directory.
 328 .RE
 329 .PP
 330 The following bits may be set in the
 331 .I mask
 332 field returned by
 333 .BR read (2):
 334 .RS 4
 335 .TP
 336 .B IN_IGNORED
 337 Watch was removed explicitly
 338 .RB ( inotify_rm_watch (2))
 339 or automatically (file was deleted, or filesystem was unmounted).
 340 See also BUGS.
 341 .TP
 342 .B IN_ISDIR
 343 Subject of this event is a directory.
 344 .TP
 345 .B IN_Q_OVERFLOW
 346 Event queue overflowed
 347 .RI ( wd
 348 is \-1 for this event).
 349 .TP
 350 .B IN_UNMOUNT
 351 Filesystem containing watched object was unmounted.
 352 In addition, an
 353 .B IN_IGNORED
 354 event will subsequently be generated for the watch descriptor.
 355 .RE
 356 .SS Examples
 357 Suppose an application is watching the directory
 358 .I dir
 359 and the file
 360 .IR dir/myfile
 361 for all events.
 362 The examples below show some events that will be generated
 363 for these two objects.
 364 .RS 4
 365 .TP
 366 fd = open("dir/myfile", O_RDWR);
 367 Generates
 368 .B IN_OPEN
 369 events for both
 370 .I dir
 371 and
 372 .IR dir/myfile .
 373 .TP
 374 read(fd, buf, count);
 375 Generates
 376 .B IN_ACCESS
 377 events for both
 378 .I dir
 379 and
 380 .IR dir/myfile .
 381 .TP
 382 write(fd, buf, count);
 383 Generates
 384 .B IN_MODIFY
 385 events for both
 386 .I dir
 387 and
 388 .IR dir/myfile .
 389 .TP
 390 fchmod(fd, mode);
 391 Generates
 392 .B IN_ATTRIB
 393 events for both
 394 .I dir
 395 and
 396 .IR dir/myfile .
 397 .TP
 398 close(fd);
 399 Generates
 400 .B IN_CLOSE_WRITE
 401 events for both
 402 .I dir
 403 and
 404 .IR dir/myfile .
 405 .RE
 406 .PP
 407 Suppose an application is watching the directories
 408 .I dir1
 409 and
 410 .IR dir ,
 411 and the file
 412 .IR dir1/myfile .
 413 The following examples show some events that may be generated.
 414 .RS 4
 415 .TP
 416 link("dir/myfile", "dir2/new");
 417 Generates an
 418 .B IN_ATTRIB
 419 event for
 420 .IR myfile
 421 and an
 422 .B IN_CREATE
 423 event for
 424 .IR dir2 .
 425 .TP
 426 rename("dir1/myfile", "dir2/myfile");
 427 Generates an
 428 .B IN_MOVED_FROM
 429 event for
 430 .IR dir1 ,
 431 an
 432 .B IN_MOVED_TO
 433 event for
 434 .IR dir2 ,
 435 and an
 436 .B IN_MOVE_SELF
 437 event for
 438 .IR myfile .
 439 The
 440 .B IN_MOVED_FROM
 441 and
 442 .B IN_MOVED_TO
 443 events will have the same
 444 .I cookie
 445 value.
 446 .RE
 447 .PP
 448 Suppose that
 449 .IR dir1/xx
 450 and
 451 .IR dir2/yy
 452 are (the only) links to the same file, and an application is watching
 453 .IR dir1 ,
 454 .IR dir2 ,
 455 .IR dir1/xx ,
 456 and
 457 .IR dir2/yy .
 458 Executing the following calls in the order given below will generate
 459 the following events:
 460 .RS 4
 461 .TP
 462 unlink("dir2/yy");
 463 Generates
 464 .BR IN_ATTRIB
 465 event for
 466 .IR xx
 467 (because its link count changes)
 468 and an
 469 .B IN_DELETE
 470 event for
 471 .IR dir2 .
 472 .TP
 473 unlink("dir1/xx");
 474 Generates
 475 .BR IN_ATTRIB ,
 476 .BR IN_DELETE_SELF ,
 477 and
 478 .BR IN_IGNORED
 479 events for
 480 .IR xx ,
 481 and an
 482 .BR IN_DELETE
 483 for
 484 .IR dir1 .
 485 .RE
 486 .PP
 487 Suppose an application is watching the directory
 488 .IR dir
 489 and (the empty) directory
 490 .IR dir/subdir .
 491 The following examples show some events that may be generated.
 492 .RS 4
 493 .TP
 494 mkdir("dir/new", mode);
 495 Generates an
 496 .B "IN_CREATE | IN_ISDIR"
 497 event for
 498 .IR dir .
 499 .TP
 500 rmdir("dir/sub");
 501 Generates
 502 .B IN_DELETE_SELF
 503 and
 504 .B IN_IGNORED
 505 events for
 506 .IR subdir ,
 507 and an
 508 .B "IN_DELETE | IN_ISDIR"
 509 event for
 510 .IR dir .
 511 .RE
 512 .SS /proc interfaces
 513 The following interfaces can be used to limit the amount of
 514 kernel memory consumed by inotify:
 515 .TP
 516 .I /proc/sys/fs/inotify/max_queued_events
 517 The value in this file is used when an application calls
 518 .BR inotify_init (2)
 519 to set an upper limit on the number of events that can be
 520 queued to the corresponding inotify instance.
 521 Events in excess of this limit are dropped, but an
 522 .B IN_Q_OVERFLOW
 523 event is always generated.
 524 .TP
 525 .I /proc/sys/fs/inotify/max_user_instances
 526 This specifies an upper limit on the number of inotify instances
 527 that can be created per real user ID.
 528 .TP
 529 .I /proc/sys/fs/inotify/max_user_watches
 530 This specifies an upper limit on the number of watches
 531 that can be created per real user ID.
 532 .SH VERSIONS
 533 Inotify was merged into the 2.6.13 Linux kernel.
 534 The required library interfaces were added to glibc in version 2.4.
 535 .RB ( IN_DONT_FOLLOW ,
 536 .BR IN_MASK_ADD ,
 537 and
 538 .B IN_ONLYDIR
 539 were added in glibc version 2.5.)
 540 .SH CONFORMING TO
 541 The inotify API is Linux-specific.
 542 .SH NOTES
 543 Inotify file descriptors can be monitored using
 544 .BR select (2),
 545 .BR poll (2),
 546 and
 547 .BR epoll (7).
 548 When an event is available, the file descriptor indicates as readable.
 549
 550 Since Linux 2.6.25,
 551 signal-driven I/O notification is available for inotify file descriptors;
 552 see the discussion of
 553 .B F_SETFL
 554 (for setting the
 555 .B O_ASYNC
 556 flag),
 557 .BR F_SETOWN ,
 558 and
 559 .B F_SETSIG
 560 in
 561 .BR fcntl (2).
 562 The
 563 .I siginfo_t
 564 structure (described in
 565 .BR sigaction (2))
 566 that is passed to the signal handler has the following fields set:
 567 .IR si_fd
 568 is set to the inotify file descriptor number;
 569 .IR si_signo
 570 is set to the signal number;
 571 .IR si_code
 572 is set to
 573 .BR POLL_IN ;
 574 and
 575 .B POLLIN
 576 is set in
 577 .IR si_band .
 578
 579 If successive output inotify events produced on the
 580 inotify file descriptor are identical (same
 581 .IR wd ,
 582 .IR mask ,
 583 .IR cookie ,
 584 and
 585 .IR name ),
 586 then they are coalesced into a single event if the
 587 older event has not yet been read (but see BUGS).
 588 This reduces the amount of kernel memory required for the event queue,
 589 but also means that an application can't use inotify to reliably count
 590 file events.
 591
 592 The events returned by reading from an inotify file descriptor
 593 form an ordered queue.
 594 Thus, for example, it is guaranteed that when renaming from
 595 one directory to another, events will be produced in the
 596 correct order on the inotify file descriptor.
 597
 598 The
 599 .B FIONREAD
 600 .BR ioctl (2)
 601 returns the number of bytes available to read from an
 602 inotify file descriptor.
 603 .SS Limitations and caveats
 604 The inotify API provides no information about the user or process that
 605 triggered the inotify event.
 606 In particular, there is no easy
 607 way for a process that is monitoring events via inotify
 608 to distinguish events that it triggers
 609 itself from those that are triggered by other processes.
 610
 611 The inotify API identifies affected files by filename.
 612 However, by the time an application processes an inotify event,
 613 the filename may already have been deleted or renamed.
 614
 615 The inotify API identifies events via watch descriptors.
 616 It is the application's responsibility to cache a mapping
 617 (if one is needed) between watch descriptors and pathnames.
 618 Be aware that directory renamings may affect multiple cached pathnames.
 619
 620 Inotify monitoring of directories is not recursive:
 621 to monitor subdirectories under a directory,
 622 additional watches must be created.
 623 This can take a significant amount time for large directory trees.
 624
 625 If monitoring an entire directory subtree,
 626 and a new subdirectory is created in that tree or an existing directory
 627 is renamed into that tree,
 628 be aware that by the time you create a watch for the new subdirectory,
 629 new files (and subdirectories) may already exist inside the subdirectory.
 630 Therefore, you may want to scan the contents of the subdirectory
 631 immediately after adding the watch (and, if desired,
 632 recursively add watches for any subdirectories that it contains).
 633
 634 Note that the event queue can overflow.
 635 In this case, events are lost.
 636 Robust applications should handle the possibility of
 637 lost events gracefully.
 638 For example, it may be necessary to rebuild part or all of
 639 the application cache.
 640 (One simple, but possibly expensive,
 641 approach is to close the inotify file descriptor, empty the cache,
 642 create a new inotify file descriptor,
 643 and then re-create watches and cache entries
 644 for the objects to be monitored.)
 645 .SS Dealing with rename() events
 646 The
 647 .B IN_MOVED_FROM
 648 and
 649 .B IN_MOVED_TO
 650 events that are generated by
 651 .BR rename (2)
 652 are usually available as consecutive events when reading
 653 from the inotify file descriptor.
 654 However, this is not guaranteed.
 655 If multiple processes are triggering events for monitored objects,
 656 then (on rare occasions) an arbitrary number of
 657 other events may appear between the
 658 .B IN_MOVED_FROM
 659 and
 660 .B IN_MOVED_TO
 661 events.
 662
 663 Matching up the
 664 .B IN_MOVED_FROM
 665 and
 666 .B IN_MOVED_TO
 667 event pair generated by
 668 .BR rename (2)
 669 is thus inherently racy.
 670 (Don't forget that if an object is renamed outside of a monitored directory,
 671 there may not even be an
 672 .BR IN_MOVED_TO
 673 event.)
 674 Heuristic approaches (e.g., assume the events are always consecutive)
 675 can be used to ensure a match in most cases,
 676 but will inevitably miss some cases,
 677 causing the application to perceive the
 678 .B IN_MOVED_FROM
 679 and
 680 .B IN_MOVED_TO
 681 events as being unrelated.
 682 If watch descriptors are destroyed and re-created as a result,
 683 then those watch descriptors will be inconsistent with
 684 the watch descriptors in any pending events.
 685 (Rebuilding the cache and re-creating the inotify file descriptor may
 686 be useful to deal with this scenario.)
 687
 688 Applications should also allow for the possibility that the
 689 .B IN_MOVED_FROM
 690 event was the last event that could fit in the buffer
 691 returned by the current call to
 692 .BR read (2),
 693 and the accompanying
 694 .B IN_MOVED_TO
 695 event might be fetched only on the next
 696 .BR read (2).
 697 .SH BUGS
 698 .\" FIXME kernel commit 611da04f7a31b2208e838be55a42c7a1310ae321
 699 .\" implies that unmount events were buggy 2.6.11 to 2.6.36
 700 .\"
 701 In kernels before 2.6.16, the
 702 .B IN_ONESHOT
 703 .I mask
 704 flag does not work.
 705
 706 As originally designed and implemented, the
 707 .B IN_ONESHOT
 708 flag did not cause an
 709 .B IN_IGNORED
 710 event to be generated when the watch was dropped after one event.
 711 However, as an unintended effect of other changes,
 712 since Linux 2.6.36, an
 713 .B IN_IGNORED
 714 event is generated in this case.
 715
 716 Before kernel 2.6.25,
 717 .\" commit 1c17d18e3775485bf1e0ce79575eb637a94494a2
 718 the kernel code that was intended to coalesce successive identical events
 719 (i.e., the two most recent events could potentially be coalesced
 720 if the older had not yet been read)
 721 instead checked if the most recent event could be coalesced with the
 722 .I oldest
 723 unread event.
 724 .SH SEE ALSO
 725 .BR inotifywait (1),
 726 .BR inotifywatch (1),
 727 .BR inotify_add_watch (2),
 728 .BR inotify_init (2),
 729 .BR inotify_init1 (2),
 730 .BR inotify_rm_watch (2),
 731 .BR read (2),
 732 .BR stat (2)
 733
 734 .IR Documentation/filesystems/inotify.txt
 735 in the Linux kernel source tree