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  15 .\" manual page may be incorrect or out-of-date.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 2014-03-25 "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 .sp
 188 .PD 0
 189 .TP 18
 190 .B IN_ACCESS
 191 File was accessed (read,
 192 .BR execve (2)) (*).
 193 .TP
 194 .B IN_ATTRIB
 195 Metadata changed\(emfor example, permissions, timestamps, extended attributes,
 196 link count (since Linux 2.6.25), UID, or GID. (*).
 197 .TP
 198 .B IN_CLOSE_WRITE
 199 File opened for writing was closed (*).
 200 .TP
 201 .B IN_CLOSE_NOWRITE
 202 File not opened for writing was closed (*).
 203 .TP
 204 .B IN_CREATE
 205 File/directory created in watched directory (*).
 206 .TP
 207 .B IN_DELETE
 208 File/directory deleted from watched directory (*).
 209 .TP
 210 .B IN_DELETE_SELF
 211 Watched file/directory was itself deleted.
 212 .TP
 213 .B IN_MODIFY
 214 File was modified (*).
 215 .TP
 216 .B IN_MOVE_SELF
 217 Watched file/directory was itself moved.
 218 .TP
 219 .B IN_MOVED_FROM
 220 Generated for the directory containing the old filename
 221 when a file is renamed (*).
 222 .TP
 223 .B IN_MOVED_TO
 224 Generated for the directory containing the new filename
 225 when a file is renamed (*).
 226 .TP
 227 .B IN_OPEN
 228 File was opened (*).
 229 .PD
 230 .RE
 231 .PP
 232 When monitoring a directory,
 233 the events marked with an asterisk (*) above can occur for
 234 files in the directory, in which case the
 235 .I name
 236 field in the returned
 237 .I inotify_event
 238 structure identifies the name of the file within the directory.
 239 .PP
 240 The
 241 .B IN_ALL_EVENTS
 242 macro is defined as a bit mask of all of the above events.
 243 This macro can be used as the
 244 .I mask
 245 argument when calling
 246 .BR inotify_add_watch (2).
 247
 248 Two additional convenience macros are
 249 .BR IN_MOVE ,
 250 which equates to
 251 IN_MOVED_FROM|IN_MOVED_TO,
 252 and
 253 .BR IN_CLOSE ,
 254 which equates to
 255 IN_CLOSE_WRITE|IN_CLOSE_NOWRITE.
 256 .PP
 257 The following further bits can be specified in
 258 .I mask
 259 when calling
 260 .BR inotify_add_watch (2):
 261 .RS 4
 262 .sp
 263 .PD 0
 264 .TP 18
 265 .BR IN_DONT_FOLLOW " (since Linux 2.6.15)"
 266 Don't dereference
 267 .I pathname
 268 if it is a symbolic link.
 269 .TP
 270 .BR IN_EXCL_UNLINK " (since Linux 2.6.36)"
 271 .\" commit 8c1934c8d70b22ca8333b216aec6c7d09fdbd6a6
 272 By default, when watching events on the children of a directory,
 273 events are generated for children even after they have been unlinked
 274 from the directory.
 275 This can result in large numbers of uninteresting events for
 276 some applications (e.g., if watching
 277 .IR /tmp ,
 278 in which many applications create temporary files whose
 279 names are immediately unlinked).
 280 Specifying
 281 .B IN_EXCL_UNLINK
 282 changes the default behavior,
 283 so that events are not generated for children after
 284 they have been unlinked from the watched directory.
 285 .TP
 286 .B IN_MASK_ADD
 287 Add (OR) events to watch mask for this pathname if
 288 it already exists (instead of replacing mask).
 289 .TP
 290 .B IN_ONESHOT
 291 Monitor
 292 .I pathname
 293 for one event, then remove from
 294 watch list.
 295 .TP
 296 .BR IN_ONLYDIR " (since Linux 2.6.15)"
 297 Only watch
 298 .I pathname
 299 if it is a directory.
 300 .PD
 301 .RE
 302 .PP
 303 The following bits may be set in the
 304 .I mask
 305 field returned by
 306 .BR read (2):
 307 .RS 4
 308 .sp
 309 .PD 0
 310 .TP 18
 311 .B IN_IGNORED
 312 Watch was removed explicitly
 313 .RB ( inotify_rm_watch (2))
 314 or automatically (file was deleted, or filesystem was unmounted).
 315 .TP
 316 .B IN_ISDIR
 317 Subject of this event is a directory.
 318 .TP
 319 .B IN_Q_OVERFLOW
 320 Event queue overflowed
 321 .RI ( wd
 322 is \-1 for this event).
 323 .TP
 324 .B IN_UNMOUNT
 325 Filesystem containing watched object was unmounted.
 326 .PD
 327 .RE
 328 .SS /proc interfaces
 329 The following interfaces can be used to limit the amount of
 330 kernel memory consumed by inotify:
 331 .TP
 332 .I /proc/sys/fs/inotify/max_queued_events
 333 The value in this file is used when an application calls
 334 .BR inotify_init (2)
 335 to set an upper limit on the number of events that can be
 336 queued to the corresponding inotify instance.
 337 Events in excess of this limit are dropped, but an
 338 .B IN_Q_OVERFLOW
 339 event is always generated.
 340 .TP
 341 .I /proc/sys/fs/inotify/max_user_instances
 342 This specifies an upper limit on the number of inotify instances
 343 that can be created per real user ID.
 344 .TP
 345 .I /proc/sys/fs/inotify/max_user_watches
 346 This specifies an upper limit on the number of watches
 347 that can be created per real user ID.
 348 .SH VERSIONS
 349 Inotify was merged into the 2.6.13 Linux kernel.
 350 The required library interfaces were added to glibc in version 2.4.
 351 .RB ( IN_DONT_FOLLOW ,
 352 .BR IN_MASK_ADD ,
 353 and
 354 .B IN_ONLYDIR
 355 were added in version 2.5.)
 356 .SH CONFORMING TO
 357 The inotify API is Linux-specific.
 358 .SH NOTES
 359 Inotify file descriptors can be monitored using
 360 .BR select (2),
 361 .BR poll (2),
 362 and
 363 .BR epoll (7).
 364 When an event is available, the file descriptor indicates as readable.
 365
 366 Since Linux 2.6.25,
 367 signal-driven I/O notification is available for inotify file descriptors;
 368 see the discussion of
 369 .B F_SETFL
 370 (for setting the
 371 .B O_ASYNC
 372 flag),
 373 .BR F_SETOWN ,
 374 and
 375 .B F_SETSIG
 376 in
 377 .BR fcntl (2).
 378 The
 379 .I siginfo_t
 380 structure (described in
 381 .BR sigaction (2))
 382 that is passed to the signal handler has the following fields set:
 383 .IR si_fd
 384 is set to the inotify file descriptor number;
 385 .IR si_signo
 386 is set to the signal number;
 387 .IR si_code
 388 is set to
 389 .BR POLL_IN ;
 390 and
 391 .B POLLIN
 392 is set in
 393 .IR si_band .
 394
 395 If successive output inotify events produced on the
 396 inotify file descriptor are identical (same
 397 .IR wd ,
 398 .IR mask ,
 399 .IR cookie ,
 400 and
 401 .IR name ),
 402 then they are coalesced into a single event if the
 403 older event has not yet been read (but see BUGS).
 404 This reduces the amount of kernel memory required for the event queue,
 405 but also means that an application can't use inotify to reliably count
 406 file events.
 407
 408 The events returned by reading from an inotify file descriptor
 409 form an ordered queue.
 410 Thus, for example, it is guaranteed that when renaming from
 411 one directory to another, events will be produced in the
 412 correct order on the inotify file descriptor.
 413
 414 The
 415 .B FIONREAD
 416 .BR ioctl (2)
 417 returns the number of bytes available to read from an
 418 inotify file descriptor.
 419 .SS Limitations and caveats
 420 Inotify monitoring of directories is not recursive:
 421 to monitor subdirectories under a directory,
 422 additional watches must be created.
 423 This can take a significant amount time for large directory trees.
 424
 425 The inotify API provides no information about the user or process that
 426 triggered the inotify event.
 427 In particular, there is no easy
 428 way for a process that is monitoring events via inotify
 429 to distinguish events that it triggers
 430 itself from those that are triggered by other processes.
 431
 432 Note that the event queue can overflow.
 433 In this case, events are lost.
 434 Robust applications should handle the possibility of
 435 lost events gracefully.
 436
 437 The inotify API identifies affected files by filename.
 438 However, by the time an application processes an inotify event,
 439 the filename may already have been deleted or renamed.
 440
 441 If monitoring an entire directory subtree,
 442 and a new subdirectory is created in that tree or an existing directory
 443 is renamed into that tree,
 444 be aware that by the time you create a watch for the new subdirectory,
 445 new files (and subdirectories) may already exist inside the subdirectory.
 446 Therefore, you may want to scan the contents of the subdirectory
 447 immediately after adding the watch (and, if desired,
 448 recursively add watches for any subdirectories that it contains).
 449
 450 The inotify applications identifies events via watch descriptors.
 451 It is the application's responsibility to cache a mapping
 452 (if one is needed) between watch descriptors and pathnames.
 453 Be aware that directory renamings may affect multiple cached pathnames.
 454 .SH BUGS
 455 In kernels before 2.6.16, the
 456 .B IN_ONESHOT
 457 .I mask
 458 flag does not work.
 459
 460 Before kernel 2.6.25,
 461 .\" commit 1c17d18e3775485bf1e0ce79575eb637a94494a2
 462 the kernel code that was intended to coalesce successive identical events
 463 (i.e., the two most recent events could potentially be coalesced
 464 if the older had not yet been read)
 465 instead checked if the most recent event could be coalesced with the
 466 .I oldest
 467 unread event.
 468 .SH SEE ALSO
 469 .BR inotifywait (1),
 470 .BR inotifywatch (1),
 471 .BR inotify_add_watch (2),
 472 .BR inotify_init (2),
 473 .BR inotify_init1 (2),
 474 .BR inotify_rm_watch (2),
 475 .BR read (2),
 476 .BR stat (2)
 477
 478 .IR Documentation/filesystems/inotify.txt
 479 in the Linux kernel source tree