man7/inotify.7

   1 '\" t
   2 .\" Copyright (C) 2006 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  15 .\" manual page may be incorrect or out-of-date.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 2014-03-28 "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 .BR inotify_init (2)
  40 (or
  41 .BR inotify_init1 (2)),
  42 .BR inotify_add_watch (2),
  43 .BR inotify_rm_watch (2),
  44 .BR read (2),
  45 and
  46 .BR close (2).
  47
  48 .BR inotify_init (2)
  49 creates an inotify instance and returns a file descriptor
  50 referring to the inotify instance.
  51 The more recent
  52 .BR inotify_init1 (2)
  53 is like
  54 .BR inotify_init (2),
  55 but provides some extra functionality.
  56
  57 .BR inotify_add_watch (2)
  58 manipulates the "watch list" associated with an inotify instance.
  59 Each item ("watch") in the watch list specifies the pathname of
  60 a file or directory,
  61 along with some set of events that the kernel should monitor for the
  62 file referred to by that pathname.
  63 .BR inotify_add_watch (2)
  64 either creates a new watch item, or modifies an existing watch.
  65 Each watch has a unique "watch descriptor", an integer
  66 returned by
  67 .BR inotify_add_watch (2)
  68 when the watch is created.
  69
  70 .BR inotify_rm_watch (2)
  71 removes an item from an inotify watch list.
  72
  73 When all file descriptors referring to an inotify
  74 instance have been closed,
  75 the underlying object and its resources are
  76 freed for reuse by the kernel;
  77 all associated watches are automatically freed.
  78
  79 To determine what events have occurred, an application
  80 .BR read (2)s
  81 from the inotify file descriptor.
  82 If no events have so far occurred, then,
  83 assuming a blocking file descriptor,
  84 .BR read (2)
  85 will block until at least one event occurs
  86 (unless interrupted by a signal,
  87 in which case the call fails with the error
  88 .BR EINTR ;
  89 see
  90 .BR signal (7)).
  91
  92 Each successful
  93 .BR read (2)
  94 returns a buffer containing one or more of the following structures:
  95 .in +4n
  96 .nf
  97
  98 struct inotify_event {
  99     int      wd;       /* Watch descriptor */
 100 .\" FIXME . The type of the 'wd' field should probably be "int32_t".
 101 .\" I submitted a patch to fix this.  See the LKML thread
 102 .\" "[patch] Fix type errors in inotify interfaces", 18 Nov 2008
 103 .\" Glibc bug filed: http://sources.redhat.com/bugzilla/show_bug.cgi?id=7040
 104     uint32_t mask;     /* Mask of events */
 105     uint32_t cookie;   /* Unique cookie associating related
 106                           events (for rename(2)) */
 107     uint32_t len;      /* Size of \fIname\fP field */
 108     char     name[];   /* Optional null-terminated name */
 109 };
 110 .fi
 111 .in
 112
 113 .I wd
 114 identifies the watch for which this event occurs.
 115 It is one of the watch descriptors returned by a previous call to
 116 .BR inotify_add_watch (2).
 117
 118 .I mask
 119 contains bits that describe the event that occurred (see below).
 120
 121 .I cookie
 122 is a unique integer that connects related events.
 123 Currently this is used only for rename events, and
 124 allows the resulting pair of
 125 .B IN_MOVED_FROM
 126 and
 127 .B IN_MOVED_TO
 128 events to be connected by the application.
 129 For all other event types,
 130 .I cookie
 131 is set to 0.
 132
 133 The
 134 .I name
 135 field is present only when an event is returned
 136 for a file inside a watched directory;
 137 it identifies the file pathname relative to the watched directory.
 138 This pathname is null-terminated,
 139 and may include further null bytes (\(aq\\0\(aq) to align subsequent reads to a
 140 suitable address boundary.
 141
 142 The
 143 .I len
 144 field counts all of the bytes in
 145 .IR name ,
 146 including the null bytes;
 147 the length of each
 148 .I inotify_event
 149 structure is thus
 150 .IR "sizeof(struct inotify_event)+len" .
 151
 152 The behavior when the buffer given to
 153 .BR read (2)
 154 is too small to return information about the next event depends
 155 on the kernel version: in kernels before 2.6.21,
 156 .BR read (2)
 157 returns 0; since kernel 2.6.21,
 158 .BR read (2)
 159 fails with the error
 160 .BR EINVAL .
 161 Specifying a buffer of size
 162
 163     sizeof(struct inotify_event) + NAME_MAX + 1
 164
 165 will be sufficient to read at least one event.
 166 .SS inotify events
 167 The
 168 .BR inotify_add_watch (2)
 169 .I mask
 170 argument and the
 171 .I mask
 172 field of the
 173 .I inotify_event
 174 structure returned when
 175 .BR read (2)ing
 176 an inotify file descriptor are both bit masks identifying
 177 inotify events.
 178 The following bits can be specified in
 179 .I mask
 180 when calling
 181 .BR inotify_add_watch (2)
 182 and may be returned in the
 183 .I mask
 184 field returned by
 185 .BR read (2):
 186 .RS 4
 187 .TP
 188 .BR IN_ACCESS " (*)"
 189 File was accessed (e.g.,
 190 .BR read (2),
 191 .BR execve (2)).
 192 .TP
 193 .BR IN_ATTRIB " (*)"
 194 Metadata changed\(emfor example, permissions, timestamps, extended attributes,
 195 link count (since Linux 2.6.25), UID, or GID.
 196 .TP
 197 .BR IN_CLOSE_WRITE " (*)"
 198 File opened for writing was closed.
 199 .TP
 200 .BR IN_CLOSE_NOWRITE " (*)"
 201 File not opened for writing was closed.
 202 .TP
 203 .BR IN_CREATE " (*)"
 204 File/directory created in watched directory.
 205 .TP
 206 .BR IN_DELETE " (*)"
 207 File/directory deleted from watched directory.
 208 .TP
 209 .B IN_DELETE_SELF
 210 Watched file/directory was itself deleted.
 211 (This event also occurs if an object is moved to another filesystem,
 212 since
 213 .BR mv (1)
 214 in effect copies the file to the other filesystem and
 215 then deletes it from the original filesystem.)
 216 .TP
 217 .BR IN_MODIFY " (*)"
 218 File was modified.
 219 .TP
 220 .B IN_MOVE_SELF
 221 Watched file/directory was itself moved.
 222 .TP
 223 .BR IN_MOVED_FROM " (*)"
 224 Generated for the directory containing the old filename
 225 when a file is renamed.
 226 .TP
 227 .BR IN_MOVED_TO " (*)"
 228 Generated for the directory containing the new filename
 229 when a file is renamed.
 230 .TP
 231 .BR IN_OPEN " (*)"
 232 File was opened.
 233 .RE
 234 .PP
 235 When monitoring a directory,
 236 the events marked with an asterisk (*) above can occur for
 237 files in the directory, in which case the
 238 .I name
 239 field in the returned
 240 .I inotify_event
 241 structure identifies the name of the file within the directory.
 242 .PP
 243 The
 244 .B IN_ALL_EVENTS
 245 macro is defined as a bit mask of all of the above events.
 246 This macro can be used as the
 247 .I mask
 248 argument when calling
 249 .BR inotify_add_watch (2).
 250
 251 Two additional convenience macros are defined:
 252 .RS 4
 253 .TP
 254 .BR IN_MOVE
 255 Equates to
 256 .BR "IN_MOVED_FROM | IN_MOVED_TO" .
 257 .TP
 258 .BR IN_CLOSE
 259 Equates to
 260 .BR "IN_CLOSE_WRITE | IN_CLOSE_NOWRITE" .
 261 .RE
 262 .PP
 263 The following further bits can be specified in
 264 .I mask
 265 when calling
 266 .BR inotify_add_watch (2):
 267 .RS 4
 268 .TP
 269 .BR IN_DONT_FOLLOW " (since Linux 2.6.15)"
 270 Don't dereference
 271 .I pathname
 272 if it is a symbolic link.
 273 .TP
 274 .BR IN_EXCL_UNLINK " (since Linux 2.6.36)"
 275 .\" commit 8c1934c8d70b22ca8333b216aec6c7d09fdbd6a6
 276 By default, when watching events on the children of a directory,
 277 events are generated for children even after they have been unlinked
 278 from the directory.
 279 This can result in large numbers of uninteresting events for
 280 some applications (e.g., if watching
 281 .IR /tmp ,
 282 in which many applications create temporary files whose
 283 names are immediately unlinked).
 284 Specifying
 285 .B IN_EXCL_UNLINK
 286 changes the default behavior,
 287 so that events are not generated for children after
 288 they have been unlinked from the watched directory.
 289 .TP
 290 .B IN_MASK_ADD
 291 Add (OR) events to watch mask for this pathname if
 292 it already exists (instead of replacing mask).
 293 .TP
 294 .B IN_ONESHOT
 295 Monitor
 296 .I pathname
 297 for one event, then remove from
 298 watch list.
 299 .TP
 300 .BR IN_ONLYDIR " (since Linux 2.6.15)"
 301 Only watch
 302 .I pathname
 303 if it is a directory.
 304 .RE
 305 .PP
 306 The following bits may be set in the
 307 .I mask
 308 field returned by
 309 .BR read (2):
 310 .RS 4
 311 .TP
 312 .B IN_IGNORED
 313 Watch was removed explicitly
 314 .RB ( inotify_rm_watch (2))
 315 or automatically (file was deleted, or filesystem was unmounted).
 316 See also BUGS.
 317 .TP
 318 .B IN_ISDIR
 319 Subject of this event is a directory.
 320 .TP
 321 .B IN_Q_OVERFLOW
 322 Event queue overflowed
 323 .RI ( wd
 324 is \-1 for this event).
 325 .TP
 326 .B IN_UNMOUNT
 327 Filesystem containing watched object was unmounted.
 328 .RE
 329 .SS /proc interfaces
 330 The following interfaces can be used to limit the amount of
 331 kernel memory consumed by inotify:
 332 .TP
 333 .I /proc/sys/fs/inotify/max_queued_events
 334 The value in this file is used when an application calls
 335 .BR inotify_init (2)
 336 to set an upper limit on the number of events that can be
 337 queued to the corresponding inotify instance.
 338 Events in excess of this limit are dropped, but an
 339 .B IN_Q_OVERFLOW
 340 event is always generated.
 341 .TP
 342 .I /proc/sys/fs/inotify/max_user_instances
 343 This specifies an upper limit on the number of inotify instances
 344 that can be created per real user ID.
 345 .TP
 346 .I /proc/sys/fs/inotify/max_user_watches
 347 This specifies an upper limit on the number of watches
 348 that can be created per real user ID.
 349 .SH VERSIONS
 350 Inotify was merged into the 2.6.13 Linux kernel.
 351 The required library interfaces were added to glibc in version 2.4.
 352 .RB ( IN_DONT_FOLLOW ,
 353 .BR IN_MASK_ADD ,
 354 and
 355 .B IN_ONLYDIR
 356 were added in glibc version 2.5.)
 357 .SH CONFORMING TO
 358 The inotify API is Linux-specific.
 359 .SH NOTES
 360 Inotify file descriptors can be monitored using
 361 .BR select (2),
 362 .BR poll (2),
 363 and
 364 .BR epoll (7).
 365 When an event is available, the file descriptor indicates as readable.
 366
 367 Since Linux 2.6.25,
 368 signal-driven I/O notification is available for inotify file descriptors;
 369 see the discussion of
 370 .B F_SETFL
 371 (for setting the
 372 .B O_ASYNC
 373 flag),
 374 .BR F_SETOWN ,
 375 and
 376 .B F_SETSIG
 377 in
 378 .BR fcntl (2).
 379 The
 380 .I siginfo_t
 381 structure (described in
 382 .BR sigaction (2))
 383 that is passed to the signal handler has the following fields set:
 384 .IR si_fd
 385 is set to the inotify file descriptor number;
 386 .IR si_signo
 387 is set to the signal number;
 388 .IR si_code
 389 is set to
 390 .BR POLL_IN ;
 391 and
 392 .B POLLIN
 393 is set in
 394 .IR si_band .
 395
 396 If successive output inotify events produced on the
 397 inotify file descriptor are identical (same
 398 .IR wd ,
 399 .IR mask ,
 400 .IR cookie ,
 401 and
 402 .IR name ),
 403 then they are coalesced into a single event if the
 404 older event has not yet been read (but see BUGS).
 405 This reduces the amount of kernel memory required for the event queue,
 406 but also means that an application can't use inotify to reliably count
 407 file events.
 408
 409 The events returned by reading from an inotify file descriptor
 410 form an ordered queue.
 411 Thus, for example, it is guaranteed that when renaming from
 412 one directory to another, events will be produced in the
 413 correct order on the inotify file descriptor.
 414
 415 The
 416 .B FIONREAD
 417 .BR ioctl (2)
 418 returns the number of bytes available to read from an
 419 inotify file descriptor.
 420 .SS Limitations and caveats
 421 Inotify monitoring of directories is not recursive:
 422 to monitor subdirectories under a directory,
 423 additional watches must be created.
 424 This can take a significant amount time for large directory trees.
 425
 426 The inotify API provides no information about the user or process that
 427 triggered the inotify event.
 428 In particular, there is no easy
 429 way for a process that is monitoring events via inotify
 430 to distinguish events that it triggers
 431 itself from those that are triggered by other processes.
 432
 433 Note that the event queue can overflow.
 434 In this case, events are lost.
 435 Robust applications should handle the possibility of
 436 lost events gracefully.
 437
 438 The inotify API identifies affected files by filename.
 439 However, by the time an application processes an inotify event,
 440 the filename may already have been deleted or renamed.
 441
 442 If monitoring an entire directory subtree,
 443 and a new subdirectory is created in that tree or an existing directory
 444 is renamed into that tree,
 445 be aware that by the time you create a watch for the new subdirectory,
 446 new files (and subdirectories) may already exist inside the subdirectory.
 447 Therefore, you may want to scan the contents of the subdirectory
 448 immediately after adding the watch (and, if desired,
 449 recursively add watches for any subdirectories that it contains).
 450
 451 The inotify applications identifies events via watch descriptors.
 452 It is the application's responsibility to cache a mapping
 453 (if one is needed) between watch descriptors and pathnames.
 454 Be aware that directory renamings may affect multiple cached pathnames.
 455 .SH BUGS
 456 .\" FIXME kernel commit 611da04f7a31b2208e838be55a42c7a1310ae321
 457 .\" implies that unmount events were buggy 2.6.11 to 2.6.36
 458 .\"
 459 In kernels before 2.6.16, the
 460 .B IN_ONESHOT
 461 .I mask
 462 flag does not work.
 463
 464 As originally designed and implemented, the
 465 .B IN_ONESHOT
 466 flag did not cause an
 467 .B IN_IGNORED
 468 event to be generated when the watch was dropped after one event.
 469 However, as an unintended effect of other changes,
 470 since Linux 2.6.36, an
 471 .B IN_IGNORED
 472 event is generated in this case.
 473
 474 Before kernel 2.6.25,
 475 .\" commit 1c17d18e3775485bf1e0ce79575eb637a94494a2
 476 the kernel code that was intended to coalesce successive identical events
 477 (i.e., the two most recent events could potentially be coalesced
 478 if the older had not yet been read)
 479 instead checked if the most recent event could be coalesced with the
 480 .I oldest
 481 unread event.
 482 .SH SEE ALSO
 483 .BR inotifywait (1),
 484 .BR inotifywatch (1),
 485 .BR inotify_add_watch (2),
 486 .BR inotify_init (2),
 487 .BR inotify_init1 (2),
 488 .BR inotify_rm_watch (2),
 489 .BR read (2),
 490 .BR stat (2)
 491
 492 .IR Documentation/filesystems/inotify.txt
 493 in the Linux kernel source tree