Debian Codegraph @

upstream/debian-trixie/man2/fcntl.2

raw
1 .\" This manpage is Copyright (C) 1992 Drew Eckhardt;
2 .\" and Copyright (C) 1993 Michael Haardt, Ian Jackson;
3 .\" and Copyright (C) 1998 Jamie Lokier;
4 .\" and Copyright (C) 2002-2010, 2014 Michael Kerrisk;
5 .\" and Copyright (C) 2014 Jeff Layton
6 .\" and Copyright (C) 2014 David Herrmann
7 .\" and Copyright (C) 2017 Jens Axboe
8 .\"
9 .\" SPDX-License-Identifier: Linux-man-pages-copyleft
10 .\"
11 .\" Modified 1993-07-24 by Rik Faith <[email protected]>
12 .\" Modified 1995-09-26 by Andries Brouwer <[email protected]>
13 .\" and again on 960413 and 980804 and 981223.
14 .\" Modified 1998-12-11 by Jamie Lokier <[email protected]>
15 .\" Applied correction by Christian Ehrhardt - aeb, 990712
16 .\" Modified 2002-04-23 by Michael Kerrisk <[email protected]>
17 .\" Added note on F_SETFL and O_DIRECT
18 .\" Complete rewrite + expansion of material on file locking
19 .\" Incorporated description of F_NOTIFY, drawing on
20 .\" Stephen Rothwell's notes in Documentation/dnotify.txt.
21 .\" Added description of F_SETLEASE and F_GETLEASE
22 .\" Corrected and polished, aeb, 020527.
23 .\" Modified 2004-03-03 by Michael Kerrisk <[email protected]>
24 .\" Modified description of file leases: fixed some errors of detail
25 .\" Replaced the term "lease contestant" by "lease breaker"
26 .\" Modified, 27 May 2004, Michael Kerrisk <[email protected]>
27 .\" Added notes on capability requirements
28 .\" Modified 2004-12-08, added O_NOATIME after note from Martin Pool
29 .\" 2004-12-10, mtk, noted F_GETOWN bug after suggestion from aeb.
30 .\" 2005-04-08 Jamie Lokier <[email protected]>, mtk
31 .\" Described behavior of F_SETOWN/F_SETSIG in
32 .\" multithreaded processes, and generally cleaned
33 .\" up the discussion of F_SETOWN.
34 .\" 2005-05-20, Johannes Nicolai <[email protected]>,
35 .\" mtk: Noted F_SETOWN bug for socket file descriptor in Linux 2.4
36 .\" and earlier. Added text on permissions required to send signal.
37 .\" 2009-09-30, Michael Kerrisk
38 .\" Note obsolete F_SETOWN behavior with threads.
39 .\" Document F_SETOWN_EX and F_GETOWN_EX
40 .\" 2010-06-17, Michael Kerrisk
41 .\" Document F_SETPIPE_SZ and F_GETPIPE_SZ.
42 .\" 2014-07-08, David Herrmann <[email protected]>
43 .\" Document F_ADD_SEALS and F_GET_SEALS
44 .\" 2017-06-26, Jens Axboe <[email protected]>
45 .\" Document F_{GET,SET}_RW_HINT and F_{GET,SET}_FILE_RW_HINT
46 .\"
47 .TH fcntl 2 2024-05-02 "Linux man-pages 6.9.1"
48 .SH NAME
49 fcntl \- manipulate file descriptor
50 .SH LIBRARY
51 Standard C library
52 .RI ( libc ", " \-lc )
53 .SH SYNOPSIS
54 .nf
55 .B #include <fcntl.h>
56 .P
57 .BI "int fcntl(int " fd ", int " op ", ... /* " arg " */ );"
58 .fi
59 .SH DESCRIPTION
60 .BR fcntl ()
61 performs one of the operations described below on the open file descriptor
62 .IR fd .
63 The operation is determined by
64 .IR op .
65 .P
66 .BR fcntl ()
67 can take an optional third argument.
68 Whether or not this argument is required is determined by
69 .IR op .
70 The required argument type is indicated in parentheses after each
71 .I op
72 name (in most cases, the required type is
73 .IR int ,
74 and we identify the argument using the name
75 .IR arg ),
76 or
77 .I void
78 is specified if the argument is not required.
79 .P
80 Certain of the operations below are supported only since a particular
81 Linux kernel version.
82 The preferred method of checking whether the host kernel supports
83 a particular operation is to invoke
84 .BR fcntl ()
85 with the desired
86 .I op
87 value and then test whether the call failed with
88 .BR EINVAL ,
89 indicating that the kernel does not recognize this value.
90 .SS Duplicating a file descriptor
91 .TP
92 .BR F_DUPFD " (\fIint\fP)"
93 Duplicate the file descriptor
94 .I fd
95 using the lowest-numbered available file descriptor greater than or equal to
96 .IR arg .
97 This is different from
98 .BR dup2 (2),
99 which uses exactly the file descriptor specified.
100 .IP
101 On success, the new file descriptor is returned.
102 .IP
103 See
104 .BR dup (2)
105 for further details.
106 .TP
107 .BR F_DUPFD_CLOEXEC " (\fIint\fP; since Linux 2.6.24)"
108 As for
109 .BR F_DUPFD ,
110 but additionally set the
111 close-on-exec flag for the duplicate file descriptor.
112 Specifying this flag permits a program to avoid an additional
113 .BR fcntl ()
114 .B F_SETFD
115 operation to set the
116 .B FD_CLOEXEC
117 flag.
118 For an explanation of why this flag is useful,
119 see the description of
120 .B O_CLOEXEC
121 in
122 .BR open (2).
123 .SS File descriptor flags
124 The following operations manipulate the flags associated with
125 a file descriptor.
126 Currently, only one such flag is defined:
127 .BR FD_CLOEXEC ,
128 the close-on-exec flag.
129 If the
130 .B FD_CLOEXEC
131 bit is set,
132 the file descriptor will automatically be closed during a successful
133 .BR execve (2).
134 (If the
135 .BR execve (2)
136 fails, the file descriptor is left open.)
137 If the
138 .B FD_CLOEXEC
139 bit is not set, the file descriptor will remain open across an
140 .BR execve (2).
141 .TP
142 .BR F_GETFD " (\fIvoid\fP)"
143 Return (as the function result) the file descriptor flags;
144 .I arg
145 is ignored.
146 .TP
147 .BR F_SETFD " (\fIint\fP)"
148 Set the file descriptor flags to the value specified by
149 .IR arg .
150 .P
151 In multithreaded programs, using
152 .BR fcntl ()
153 .B F_SETFD
154 to set the close-on-exec flag at the same time as another thread performs a
155 .BR fork (2)
156 plus
157 .BR execve (2)
158 is vulnerable to a race condition that may unintentionally leak
159 the file descriptor to the program executed in the child process.
160 See the discussion of the
161 .B O_CLOEXEC
162 flag in
163 .BR open (2)
164 for details and a remedy to the problem.
165 .SS File status flags
166 Each open file description has certain associated status flags,
167 initialized by
168 .BR open (2)
169 .\" or
170 .\" .BR creat (2),
171 and possibly modified by
172 .BR fcntl ().
173 Duplicated file descriptors
174 (made with
175 .BR dup (2),
176 .BR fcntl (F_DUPFD),
177 .BR fork (2),
178 etc.) refer to the same open file description, and thus
179 share the same file status flags.
180 .P
181 The file status flags and their semantics are described in
182 .BR open (2).
183 .TP
184 .BR F_GETFL " (\fIvoid\fP)"
185 Return (as the function result)
186 the file access mode and the file status flags;
187 .I arg
188 is ignored.
189 .TP
190 .BR F_SETFL " (\fIint\fP)"
191 Set the file status flags to the value specified by
192 .IR arg .
193 File access mode
194 .RB ( O_RDONLY ", " O_WRONLY ", " O_RDWR )
195 and file creation flags
196 (i.e.,
197 .BR O_CREAT ", " O_EXCL ", " O_NOCTTY ", " O_TRUNC )
198 in
199 .I arg
200 are ignored.
201 On Linux, this operation can change only the
202 .BR O_APPEND ,
203 .BR O_ASYNC ,
204 .BR O_DIRECT ,
205 .BR O_NOATIME ,
206 and
207 .B O_NONBLOCK
208 flags.
209 It is not possible to change the
210 .B O_DSYNC
211 and
212 .B O_SYNC
213 flags; see BUGS, below.
214 .SS Advisory record locking
215 Linux implements traditional ("process-associated") UNIX record locks,
216 as standardized by POSIX.
217 For a Linux-specific alternative with better semantics,
218 see the discussion of open file description locks below.
219 .P
220 .BR F_SETLK ,
221 .BR F_SETLKW ,
222 and
223 .B F_GETLK
224 are used to acquire, release, and test for the existence of record
225 locks (also known as byte-range, file-segment, or file-region locks).
226 The third argument,
227 .IR lock ,
228 is a pointer to a structure that has at least the following fields
229 (in unspecified order).
230 .P
231 .in +4n
232 .EX
233 struct flock {
234 ...
235 short l_type; /* Type of lock: F_RDLCK,
236 F_WRLCK, F_UNLCK */
237 short l_whence; /* How to interpret l_start:
238 SEEK_SET, SEEK_CUR, SEEK_END */
239 off_t l_start; /* Starting offset for lock */
240 off_t l_len; /* Number of bytes to lock */
241 pid_t l_pid; /* PID of process blocking our lock
242 (set by F_GETLK and F_OFD_GETLK) */
243 ...
244 };
245 .EE
246 .in
247 .P
248 The
249 .IR l_whence ", " l_start ", and " l_len
250 fields of this structure specify the range of bytes we wish to lock.
251 Bytes past the end of the file may be locked,
252 but not bytes before the start of the file.
253 .P
254 .I l_start
255 is the starting offset for the lock, and is interpreted
256 relative to either:
257 the start of the file (if
258 .I l_whence
259 is
260 .BR SEEK_SET );
261 the current file offset (if
262 .I l_whence
263 is
264 .BR SEEK_CUR );
265 or the end of the file (if
266 .I l_whence
267 is
268 .BR SEEK_END ).
269 In the final two cases,
270 .I l_start
271 can be a negative number provided the
272 offset does not lie before the start of the file.
273 .P
274 .I l_len
275 specifies the number of bytes to be locked.
276 If
277 .I l_len
278 is positive, then the range to be locked covers bytes
279 .I l_start
280 up to and including
281 .IR l_start + l_len \-1.
282 Specifying 0 for
283 .I l_len
284 has the special meaning: lock all bytes starting at the
285 location specified by
286 .IR l_whence " and " l_start
287 through to the end of file, no matter how large the file grows.
288 .P
289 POSIX.1-2001 allows (but does not require)
290 an implementation to support a negative
291 .I l_len
292 value; if
293 .I l_len
294 is negative, the interval described by
295 .I lock
296 covers bytes
297 .IR l_start + l_len
298 up to and including
299 .IR l_start \-1.
300 This is supported since Linux 2.4.21 and Linux 2.5.49.
301 .P
302 The
303 .I l_type
304 field can be used to place a read
305 .RB ( F_RDLCK )
306 or a write
307 .RB ( F_WRLCK )
308 lock on a file.
309 Any number of processes may hold a read lock (shared lock)
310 on a file region, but only one process may hold a write lock
311 (exclusive lock).
312 An exclusive lock excludes all other locks,
313 both shared and exclusive.
314 A single process can hold only one type of lock on a file region;
315 if a new lock is applied to an already-locked region,
316 then the existing lock is converted to the new lock type.
317 (Such conversions may involve splitting, shrinking, or coalescing with
318 an existing lock if the byte range specified by the new lock does not
319 precisely coincide with the range of the existing lock.)
320 .TP
321 .BR F_SETLK " (\fIstruct flock *\fP)"
322 Acquire a lock (when
323 .I l_type
324 is
325 .B F_RDLCK
326 or
327 .BR F_WRLCK )
328 or release a lock (when
329 .I l_type
330 is
331 .BR F_UNLCK )
332 on the bytes specified by the
333 .IR l_whence ", " l_start ", and " l_len
334 fields of
335 .IR lock .
336 If a conflicting lock is held by another process,
337 this call returns \-1 and sets
338 .I errno
339 to
340 .B EACCES
341 or
342 .BR EAGAIN .
343 (The error returned in this case differs across implementations,
344 so POSIX requires a portable application to check for both errors.)
345 .TP
346 .BR F_SETLKW " (\fIstruct flock *\fP)"
347 As for
348 .BR F_SETLK ,
349 but if a conflicting lock is held on the file, then wait for that
350 lock to be released.
351 If a signal is caught while waiting, then the call is interrupted
352 and (after the signal handler has returned)
353 returns immediately (with return value \-1 and
354 .I errno
355 set to
356 .BR EINTR ;
357 see
358 .BR signal (7)).
359 .TP
360 .BR F_GETLK " (\fIstruct flock *\fP)"
361 On input to this call,
362 .I lock
363 describes a lock we would like to place on the file.
364 If the lock could be placed,
365 .BR fcntl ()
366 does not actually place it, but returns
367 .B F_UNLCK
368 in the
369 .I l_type
370 field of
371 .I lock
372 and leaves the other fields of the structure unchanged.
373 .IP
374 If one or more incompatible locks would prevent
375 this lock being placed, then
376 .BR fcntl ()
377 returns details about one of those locks in the
378 .IR l_type ", " l_whence ", " l_start ", and " l_len
379 fields of
380 .IR lock .
381 If the conflicting lock is a traditional (process-associated) record lock,
382 then the
383 .I l_pid
384 field is set to the PID of the process holding that lock.
385 If the conflicting lock is an open file description lock, then
386 .I l_pid
387 is set to \-1.
388 Note that the returned information
389 may already be out of date by the time the caller inspects it.
390 .P
391 In order to place a read lock,
392 .I fd
393 must be open for reading.
394 In order to place a write lock,
395 .I fd
396 must be open for writing.
397 To place both types of lock, open a file read-write.
398 .P
399 When placing locks with
400 .BR F_SETLKW ,
401 the kernel detects
402 .IR deadlocks ,
403 whereby two or more processes have their
404 lock requests mutually blocked by locks held by the other processes.
405 For example, suppose process A holds a write lock on byte 100 of a file,
406 and process B holds a write lock on byte 200.
407 If each process then attempts to lock the byte already
408 locked by the other process using
409 .BR F_SETLKW ,
410 then, without deadlock detection,
411 both processes would remain blocked indefinitely.
412 When the kernel detects such deadlocks,
413 it causes one of the blocking lock requests to immediately fail with the error
414 .BR EDEADLK ;
415 an application that encounters such an error should release
416 some of its locks to allow other applications to proceed before
417 attempting regain the locks that it requires.
418 Circular deadlocks involving more than two processes are also detected.
419 Note, however, that there are limitations to the kernel's
420 deadlock-detection algorithm; see BUGS.
421 .P
422 As well as being removed by an explicit
423 .BR F_UNLCK ,
424 record locks are automatically released when the process terminates.
425 .P
426 Record locks are not inherited by a child created via
427 .BR fork (2),
428 but are preserved across an
429 .BR execve (2).
430 .P
431 Because of the buffering performed by the
432 .BR stdio (3)
433 library, the use of record locking with routines in that package
434 should be avoided; use
435 .BR read (2)
436 and
437 .BR write (2)
438 instead.
439 .P
440 The record locks described above are associated with the process
441 (unlike the open file description locks described below).
442 This has some unfortunate consequences:
443 .IP \[bu] 3
444 If a process closes
445 .I any
446 file descriptor referring to a file,
447 then all of the process's locks on that file are released,
448 regardless of the file descriptor(s) on which the locks were obtained.
449 .\" (Additional file descriptors referring to the same file
450 .\" may have been obtained by calls to
451 .\" .BR open "(2), " dup "(2), " dup2 "(2), or " fcntl ().)
452 This is bad: it means that a process can lose its locks on
453 a file such as
454 .I /etc/passwd
455 or
456 .I /etc/mtab
457 when for some reason a library function decides to open, read,
458 and close the same file.
459 .IP \[bu]
460 The threads in a process share locks.
461 In other words,
462 a multithreaded program can't use record locking to ensure
463 that threads don't simultaneously access the same region of a file.
464 .P
465 Open file description locks solve both of these problems.
466 .SS Open file description locks (non-POSIX)
467 Open file description locks are advisory byte-range locks whose operation is
468 in most respects identical to the traditional record locks described above.
469 This lock type is Linux-specific,
470 and available since Linux 3.15.
471 (There is a proposal with the Austin Group
472 .\" FIXME . Review progress into POSIX
473 .\" http://austingroupbugs.net/view.php?id=768
474 to include this lock type in the next revision of POSIX.1.)
475 For an explanation of open file descriptions, see
476 .BR open (2).
477 .P
478 The principal difference between the two lock types
479 is that whereas traditional record locks
480 are associated with a process,
481 open file description locks are associated with the
482 open file description on which they are acquired,
483 much like locks acquired with
484 .BR flock (2).
485 Consequently (and unlike traditional advisory record locks),
486 open file description locks are inherited across
487 .BR fork (2)
488 (and
489 .BR clone (2)
490 with
491 .BR CLONE_FILES ),
492 and are only automatically released on the last close
493 of the open file description,
494 instead of being released on any close of the file.
495 .P
496 Conflicting lock combinations
497 (i.e., a read lock and a write lock or two write locks)
498 where one lock is an open file description lock and the other
499 is a traditional record lock conflict
500 even when they are acquired by the same process on the same file descriptor.
501 .P
502 Open file description locks placed via the same open file description
503 (i.e., via the same file descriptor,
504 or via a duplicate of the file descriptor created by
505 .BR fork (2),
506 .BR dup (2),
507 .BR fcntl ()
508 .BR F_DUPFD ,
509 and so on) are always compatible:
510 if a new lock is placed on an already locked region,
511 then the existing lock is converted to the new lock type.
512 (Such conversions may result in splitting, shrinking, or coalescing with
513 an existing lock as discussed above.)
514 .P
515 On the other hand, open file description locks may conflict with
516 each other when they are acquired via different open file descriptions.
517 Thus, the threads in a multithreaded program can use
518 open file description locks to synchronize access to a file region
519 by having each thread perform its own
520 .BR open (2)
521 on the file and applying locks via the resulting file descriptor.
522 .P
523 As with traditional advisory locks, the third argument to
524 .BR fcntl (),
525 .IR lock ,
526 is a pointer to an
527 .I flock
528 structure.
529 By contrast with traditional record locks, the
530 .I l_pid
531 field of that structure must be set to zero
532 when using the operations described below.
533 .P
534 The operations for working with open file description locks are analogous
535 to those used with traditional locks:
536 .TP
537 .BR F_OFD_SETLK " (\fIstruct flock *\fP)"
538 Acquire an open file description lock (when
539 .I l_type
540 is
541 .B F_RDLCK
542 or
543 .BR F_WRLCK )
544 or release an open file description lock (when
545 .I l_type
546 is
547 .BR F_UNLCK )
548 on the bytes specified by the
549 .IR l_whence ", " l_start ", and " l_len
550 fields of
551 .IR lock .
552 If a conflicting lock is held by another process,
553 this call returns \-1 and sets
554 .I errno
555 to
556 .BR EAGAIN .
557 .TP
558 .BR F_OFD_SETLKW " (\fIstruct flock *\fP)"
559 As for
560 .BR F_OFD_SETLK ,
561 but if a conflicting lock is held on the file, then wait for that lock to be
562 released.
563 If a signal is caught while waiting, then the call is interrupted
564 and (after the signal handler has returned) returns immediately
565 (with return value \-1 and
566 .I errno
567 set to
568 .BR EINTR ;
569 see
570 .BR signal (7)).
571 .TP
572 .BR F_OFD_GETLK " (\fIstruct flock *\fP)"
573 On input to this call,
574 .I lock
575 describes an open file description lock we would like to place on the file.
576 If the lock could be placed,
577 .BR fcntl ()
578 does not actually place it, but returns
579 .B F_UNLCK
580 in the
581 .I l_type
582 field of
583 .I lock
584 and leaves the other fields of the structure unchanged.
585 If one or more incompatible locks would prevent this lock being placed,
586 then details about one of these locks are returned via
587 .IR lock ,
588 as described above for
589 .BR F_GETLK .
590 .P
591 In the current implementation,
592 .\" commit 57b65325fe34ec4c917bc4e555144b4a94d9e1f7
593 no deadlock detection is performed for open file description locks.
594 (This contrasts with process-associated record locks,
595 for which the kernel does perform deadlock detection.)
596 .\"
597 .SS Mandatory locking
598 .IR Warning :
599 the Linux implementation of mandatory locking is unreliable.
600 See BUGS below.
601 Because of these bugs,
602 and the fact that the feature is believed to be little used,
603 since Linux 4.5, mandatory locking has been made an optional feature,
604 governed by a configuration option
605 .RB ( CONFIG_MANDATORY_FILE_LOCKING ).
606 This feature is no longer supported at all in Linux 5.15 and above.
607 .P
608 By default, both traditional (process-associated) and open file description
609 record locks are advisory.
610 Advisory locks are not enforced and are useful only between
611 cooperating processes.
612 .P
613 Both lock types can also be mandatory.
614 Mandatory locks are enforced for all processes.
615 If a process tries to perform an incompatible access (e.g.,
616 .BR read (2)
617 or
618 .BR write (2))
619 on a file region that has an incompatible mandatory lock,
620 then the result depends upon whether the
621 .B O_NONBLOCK
622 flag is enabled for its open file description.
623 If the
624 .B O_NONBLOCK
625 flag is not enabled, then
626 the system call is blocked until the lock is removed
627 or converted to a mode that is compatible with the access.
628 If the
629 .B O_NONBLOCK
630 flag is enabled, then the system call fails with the error
631 .BR EAGAIN .
632 .P
633 To make use of mandatory locks, mandatory locking must be enabled
634 both on the filesystem that contains the file to be locked,
635 and on the file itself.
636 Mandatory locking is enabled on a filesystem
637 using the "\-o mand" option to
638 .BR mount (8),
639 or the
640 .B MS_MANDLOCK
641 flag for
642 .BR mount (2).
643 Mandatory locking is enabled on a file by disabling
644 group execute permission on the file and enabling the set-group-ID
645 permission bit (see
646 .BR chmod (1)
647 and
648 .BR chmod (2)).
649 .P
650 Mandatory locking is not specified by POSIX.
651 Some other systems also support mandatory locking,
652 although the details of how to enable it vary across systems.
653 .\"
654 .SS Lost locks
655 When an advisory lock is obtained on a networked filesystem such as
656 NFS it is possible that the lock might get lost.
657 This may happen due to administrative action on the server, or due to a
658 network partition (i.e., loss of network connectivity with the server)
659 which lasts long enough for the server to assume
660 that the client is no longer functioning.
661 .P
662 When the filesystem determines that a lock has been lost, future
663 .BR read (2)
664 or
665 .BR write (2)
666 requests may fail with the error
667 .BR EIO .
668 This error will persist until the lock is removed or the file
669 descriptor is closed.
670 Since Linux 3.12,
671 .\" commit ef1820f9be27b6ad158f433ab38002ab8131db4d
672 this happens at least for NFSv4 (including all minor versions).
673 .P
674 Some versions of UNIX send a signal
675 .RB ( SIGLOST )
676 in this circumstance.
677 Linux does not define this signal, and does not provide any
678 asynchronous notification of lost locks.
679 .\"
680 .SS Managing signals
681 .BR F_GETOWN ,
682 .BR F_SETOWN ,
683 .BR F_GETOWN_EX ,
684 .BR F_SETOWN_EX ,
685 .BR F_GETSIG ,
686 and
687 .B F_SETSIG
688 are used to manage I/O availability signals:
689 .TP
690 .BR F_GETOWN " (\fIvoid\fP)"
691 Return (as the function result)
692 the process ID or process group ID currently receiving
693 .B SIGIO
694 and
695 .B SIGURG
696 signals for events on file descriptor
697 .IR fd .
698 Process IDs are returned as positive values;
699 process group IDs are returned as negative values (but see BUGS below).
700 .I arg
701 is ignored.
702 .TP
703 .BR F_SETOWN " (\fIint\fP)"
704 Set the process ID or process group ID that will receive
705 .B SIGIO
706 and
707 .B SIGURG
708 signals for events on the file descriptor
709 .IR fd .
710 The target process or process group ID is specified in
711 .IR arg .
712 A process ID is specified as a positive value;
713 a process group ID is specified as a negative value.
714 Most commonly, the calling process specifies itself as the owner
715 (that is,
716 .I arg
717 is specified as
718 .BR getpid (2)).
719 .IP
720 As well as setting the file descriptor owner,
721 one must also enable generation of signals on the file descriptor.
722 This is done by using the
723 .BR fcntl ()
724 .B F_SETFL
725 operation to set the
726 .B O_ASYNC
727 file status flag on the file descriptor.
728 Subsequently, a
729 .B SIGIO
730 signal is sent whenever input or output becomes possible
731 on the file descriptor.
732 The
733 .BR fcntl ()
734 .B F_SETSIG
735 operation can be used to obtain delivery of a signal other than
736 .BR SIGIO .
737 .IP
738 Sending a signal to the owner process (group) specified by
739 .B F_SETOWN
740 is subject to the same permissions checks as are described for
741 .BR kill (2),
742 where the sending process is the one that employs
743 .B F_SETOWN
744 (but see BUGS below).
745 If this permission check fails, then the signal is
746 silently discarded.
747 .IR Note :
748 The
749 .B F_SETOWN
750 operation records the caller's credentials at the time of the
751 .BR fcntl ()
752 call,
753 and it is these saved credentials that are used for the permission checks.
754 .IP
755 If the file descriptor
756 .I fd
757 refers to a socket,
758 .B F_SETOWN
759 also selects
760 the recipient of
761 .B SIGURG
762 signals that are delivered when out-of-band
763 data arrives on that socket.
764 .RB ( SIGURG
765 is sent in any situation where
766 .BR select (2)
767 would report the socket as having an "exceptional condition".)
768 .\" The following appears to be rubbish. It doesn't seem to
769 .\" be true according to the kernel source, and I can write
770 .\" a program that gets a terminal-generated SIGIO even though
771 .\" it is not the foreground process group of the terminal.
772 .\" -- MTK, 8 Apr 05
773 .\"
774 .\" If the file descriptor
775 .\" .I fd
776 .\" refers to a terminal device, then SIGIO
777 .\" signals are sent to the foreground process group of the terminal.
778 .IP
779 The following was true in Linux 2.6.x up to and including Linux 2.6.11:
780 .RS
781 .IP
782 If a nonzero value is given to
783 .B F_SETSIG
784 in a multithreaded process running with a threading library
785 that supports thread groups (e.g., NPTL),
786 then a positive value given to
787 .B F_SETOWN
788 has a different meaning:
789 .\" The relevant place in the (2.6) kernel source is the
790 .\" 'switch' in fs/fcntl.c::send_sigio_to_task() -- MTK, Apr 2005
791 instead of being a process ID identifying a whole process,
792 it is a thread ID identifying a specific thread within a process.
793 Consequently, it may be necessary to pass
794 .B F_SETOWN
795 the result of
796 .BR gettid (2)
797 instead of
798 .BR getpid (2)
799 to get sensible results when
800 .B F_SETSIG
801 is used.
802 (In current Linux threading implementations,
803 a main thread's thread ID is the same as its process ID.
804 This means that a single-threaded program can equally use
805 .BR gettid (2)
806 or
807 .BR getpid (2)
808 in this scenario.)
809 Note, however, that the statements in this paragraph do not apply
810 to the
811 .B SIGURG
812 signal generated for out-of-band data on a socket:
813 this signal is always sent to either a process or a process group,
814 depending on the value given to
815 .BR F_SETOWN .
816 .\" send_sigurg()/send_sigurg_to_task() bypasses
817 .\" kill_fasync()/send_sigio()/send_sigio_to_task()
818 .\" to directly call send_group_sig_info()
819 .\" -- MTK, Apr 2005 (kernel 2.6.11)
820 .RE
821 .IP
822 The above behavior was accidentally dropped in Linux 2.6.12,
823 and won't be restored.
824 From Linux 2.6.32 onward, use
825 .B F_SETOWN_EX
826 to target
827 .B SIGIO
828 and
829 .B SIGURG
830 signals at a particular thread.
831 .TP
832 .BR F_GETOWN_EX " (\fIstruct f_owner_ex *\fP) (since Linux 2.6.32)"
833 Return the current file descriptor owner settings
834 as defined by a previous
835 .B F_SETOWN_EX
836 operation.
837 The information is returned in the structure pointed to by
838 .IR arg ,
839 which has the following form:
840 .IP
841 .in +4n
842 .EX
843 struct f_owner_ex {
844 int type;
845 pid_t pid;
846 };
847 .EE
848 .in
849 .IP
850 The
851 .I type
852 field will have one of the values
853 .BR F_OWNER_TID ,
854 .BR F_OWNER_PID ,
855 or
856 .BR F_OWNER_PGRP .
857 The
858 .I pid
859 field is a positive integer representing a thread ID, process ID,
860 or process group ID.
861 See
862 .B F_SETOWN_EX
863 for more details.
864 .TP
865 .BR F_SETOWN_EX " (\fIstruct f_owner_ex *\fP) (since Linux 2.6.32)"
866 This operation performs a similar task to
867 .BR F_SETOWN .
868 It allows the caller to direct I/O availability signals
869 to a specific thread, process, or process group.
870 The caller specifies the target of signals via
871 .IR arg ,
872 which is a pointer to a
873 .I f_owner_ex
874 structure.
875 The
876 .I type
877 field has one of the following values, which define how
878 .I pid
879 is interpreted:
880 .RS
881 .TP
882 .B F_OWNER_TID
883 Send the signal to the thread whose thread ID
884 (the value returned by a call to
885 .BR clone (2)
886 or
887 .BR gettid (2))
888 is specified in
889 .IR pid .
890 .TP
891 .B F_OWNER_PID
892 Send the signal to the process whose ID
893 is specified in
894 .IR pid .
895 .TP
896 .B F_OWNER_PGRP
897 Send the signal to the process group whose ID
898 is specified in
899 .IR pid .
900 (Note that, unlike with
901 .BR F_SETOWN ,
902 a process group ID is specified as a positive value here.)
903 .RE
904 .TP
905 .BR F_GETSIG " (\fIvoid\fP)"
906 Return (as the function result)
907 the signal sent when input or output becomes possible.
908 A value of zero means
909 .B SIGIO
910 is sent.
911 Any other value (including
912 .BR SIGIO )
913 is the
914 signal sent instead, and in this case additional info is available to
915 the signal handler if installed with
916 .BR SA_SIGINFO .
917 .I arg
918 is ignored.
919 .TP
920 .BR F_SETSIG " (\fIint\fP)"
921 Set the signal sent when input or output becomes possible
922 to the value given in
923 .IR arg .
924 A value of zero means to send the default
925 .B SIGIO
926 signal.
927 Any other value (including
928 .BR SIGIO )
929 is the signal to send instead, and in this case additional info
930 is available to the signal handler if installed with
931 .BR SA_SIGINFO .
932 .\"
933 .\" The following was true only up until Linux 2.6.11:
934 .\"
935 .\" Additionally, passing a nonzero value to
936 .\" .B F_SETSIG
937 .\" changes the signal recipient from a whole process to a specific thread
938 .\" within a process.
939 .\" See the description of
940 .\" .B F_SETOWN
941 .\" for more details.
942 .IP
943 By using
944 .B F_SETSIG
945 with a nonzero value, and setting
946 .B SA_SIGINFO
947 for the
948 signal handler (see
949 .BR sigaction (2)),
950 extra information about I/O events is passed to
951 the handler in a
952 .I siginfo_t
953 structure.
954 If the
955 .I si_code
956 field indicates the source is
957 .BR SI_SIGIO ,
958 the
959 .I si_fd
960 field gives the file descriptor associated with the event.
961 Otherwise,
962 there is no indication which file descriptors are pending, and you
963 should use the usual mechanisms
964 .RB ( select (2),
965 .BR poll (2),
966 .BR read (2)
967 with
968 .B O_NONBLOCK
969 set etc.) to determine which file descriptors are available for I/O.
970 .IP
971 Note that the file descriptor provided in
972 .I si_fd
973 is the one that was specified during the
974 .B F_SETSIG
975 operation.
976 This can lead to an unusual corner case.
977 If the file descriptor is duplicated
978 .RB ( dup (2)
979 or similar), and the original file descriptor is closed,
980 then I/O events will continue to be generated, but the
981 .I si_fd
982 field will contain the number of the now closed file descriptor.
983 .IP
984 By selecting a real time signal (value >=
985 .BR SIGRTMIN ),
986 multiple I/O events may be queued using the same signal numbers.
987 (Queuing is dependent on available memory.)
988 Extra information is available
989 if
990 .B SA_SIGINFO
991 is set for the signal handler, as above.
992 .IP
993 Note that Linux imposes a limit on the
994 number of real-time signals that may be queued to a
995 process (see
996 .BR getrlimit (2)
997 and
998 .BR signal (7))
999 and if this limit is reached, then the kernel reverts to
1000 delivering
1001 .BR SIGIO ,
1002 and this signal is delivered to the entire
1003 process rather than to a specific thread.
1004 .\" See fs/fcntl.c::send_sigio_to_task() (2.4/2.6) sources -- MTK, Apr 05
1005 .P
1006 Using these mechanisms, a program can implement fully asynchronous I/O
1007 without using
1008 .BR select (2)
1009 or
1010 .BR poll (2)
1011 most of the time.
1012 .P
1013 The use of
1014 .B O_ASYNC
1015 is specific to BSD and Linux.
1016 The only use of
1017 .B F_GETOWN
1018 and
1019 .B F_SETOWN
1020 specified in POSIX.1 is in conjunction with the use of the
1021 .B SIGURG
1022 signal on sockets.
1023 (POSIX does not specify the
1024 .B SIGIO
1025 signal.)
1026 .BR F_GETOWN_EX ,
1027 .BR F_SETOWN_EX ,
1028 .BR F_GETSIG ,
1029 and
1030 .B F_SETSIG
1031 are Linux-specific.
1032 POSIX has asynchronous I/O and the
1033 .I aio_sigevent
1034 structure to achieve similar things; these are also available
1035 in Linux as part of the GNU C Library (glibc).
1036 .SS Leases
1037 .B F_SETLEASE
1038 and
1039 .B F_GETLEASE
1040 (Linux 2.4 onward) are used to establish a new lease,
1041 and retrieve the current lease, on the open file description
1042 referred to by the file descriptor
1043 .IR fd .
1044 A file lease provides a mechanism whereby the process holding
1045 the lease (the "lease holder") is notified (via delivery of a signal)
1046 when a process (the "lease breaker") tries to
1047 .BR open (2)
1048 or
1049 .BR truncate (2)
1050 the file referred to by that file descriptor.
1051 .TP
1052 .BR F_SETLEASE " (\fIint\fP)"
1053 Set or remove a file lease according to which of the following
1054 values is specified in the integer
1055 .IR arg :
1056 .RS
1057 .TP
1058 .B F_RDLCK
1059 Take out a read lease.
1060 This will cause the calling process to be notified when
1061 the file is opened for writing or is truncated.
1062 .\" The following became true in Linux 2.6.10:
1063 .\" See the man-pages-2.09 Changelog for further info.
1064 A read lease can be placed only on a file descriptor that
1065 is opened read-only.
1066 .TP
1067 .B F_WRLCK
1068 Take out a write lease.
1069 This will cause the caller to be notified when
1070 the file is opened for reading or writing or is truncated.
1071 A write lease may be placed on a file only if there are no
1072 other open file descriptors for the file.
1073 .TP
1074 .B F_UNLCK
1075 Remove our lease from the file.
1076 .RE
1077 .P
1078 Leases are associated with an open file description (see
1079 .BR open (2)).
1080 This means that duplicate file descriptors (created by, for example,
1081 .BR fork (2)
1082 or
1083 .BR dup (2))
1084 refer to the same lease, and this lease may be modified
1085 or released using any of these descriptors.
1086 Furthermore, the lease is released by either an explicit
1087 .B F_UNLCK
1088 operation on any of these duplicate file descriptors, or when all
1089 such file descriptors have been closed.
1090 .P
1091 Leases may be taken out only on regular files.
1092 An unprivileged process may take out a lease only on a file whose
1093 UID (owner) matches the filesystem UID of the process.
1094 A process with the
1095 .B CAP_LEASE
1096 capability may take out leases on arbitrary files.
1097 .TP
1098 .BR F_GETLEASE " (\fIvoid\fP)"
1099 Indicates what type of lease is associated with the file descriptor
1100 .I fd
1101 by returning either
1102 .BR F_RDLCK ", " F_WRLCK ", or " F_UNLCK ,
1103 indicating, respectively, a read lease , a write lease, or no lease.
1104 .I arg
1105 is ignored.
1106 .P
1107 When a process (the "lease breaker") performs an
1108 .BR open (2)
1109 or
1110 .BR truncate (2)
1111 that conflicts with a lease established via
1112 .BR F_SETLEASE ,
1113 the system call is blocked by the kernel and
1114 the kernel notifies the lease holder by sending it a signal
1115 .RB ( SIGIO
1116 by default).
1117 The lease holder should respond to receipt of this signal by doing
1118 whatever cleanup is required in preparation for the file to be
1119 accessed by another process (e.g., flushing cached buffers) and
1120 then either remove or downgrade its lease.
1121 A lease is removed by performing an
1122 .B F_SETLEASE
1123 operation specifying
1124 .I arg
1125 as
1126 .BR F_UNLCK .
1127 If the lease holder currently holds a write lease on the file,
1128 and the lease breaker is opening the file for reading,
1129 then it is sufficient for the lease holder to downgrade
1130 the lease to a read lease.
1131 This is done by performing an
1132 .B F_SETLEASE
1133 operation specifying
1134 .I arg
1135 as
1136 .BR F_RDLCK .
1137 .P
1138 If the lease holder fails to downgrade or remove the lease within
1139 the number of seconds specified in
1140 .IR /proc/sys/fs/lease\-break\-time ,
1141 then the kernel forcibly removes or downgrades the lease holder's lease.
1142 .P
1143 Once a lease break has been initiated,
1144 .B F_GETLEASE
1145 returns the target lease type (either
1146 .B F_RDLCK
1147 or
1148 .BR F_UNLCK ,
1149 depending on what would be compatible with the lease breaker)
1150 until the lease holder voluntarily downgrades or removes the lease or
1151 the kernel forcibly does so after the lease break timer expires.
1152 .P
1153 Once the lease has been voluntarily or forcibly removed or downgraded,
1154 and assuming the lease breaker has not unblocked its system call,
1155 the kernel permits the lease breaker's system call to proceed.
1156 .P
1157 If the lease breaker's blocked
1158 .BR open (2)
1159 or
1160 .BR truncate (2)
1161 is interrupted by a signal handler,
1162 then the system call fails with the error
1163 .BR EINTR ,
1164 but the other steps still occur as described above.
1165 If the lease breaker is killed by a signal while blocked in
1166 .BR open (2)
1167 or
1168 .BR truncate (2),
1169 then the other steps still occur as described above.
1170 If the lease breaker specifies the
1171 .B O_NONBLOCK
1172 flag when calling
1173 .BR open (2),
1174 then the call immediately fails with the error
1175 .BR EWOULDBLOCK ,
1176 but the other steps still occur as described above.
1177 .P
1178 The default signal used to notify the lease holder is
1179 .BR SIGIO ,
1180 but this can be changed using the
1181 .B F_SETSIG
1182 operation to
1183 .BR fcntl ().
1184 If a
1185 .B F_SETSIG
1186 operation is performed (even one specifying
1187 .BR SIGIO ),
1188 and the signal
1189 handler is established using
1190 .BR SA_SIGINFO ,
1191 then the handler will receive a
1192 .I siginfo_t
1193 structure as its second argument, and the
1194 .I si_fd
1195 field of this argument will hold the file descriptor of the leased file
1196 that has been accessed by another process.
1197 (This is useful if the caller holds leases against multiple files.)
1198 .SS File and directory change notification (dnotify)
1199 .TP
1200 .BR F_NOTIFY " (\fIint\fP)"
1201 (Linux 2.4 onward)
1202 Provide notification when the directory referred to by
1203 .I fd
1204 or any of the files that it contains is changed.
1205 The events to be notified are specified in
1206 .IR arg ,
1207 which is a bit mask specified by ORing together zero or more of
1208 the following bits:
1209 .P
1210 .RS
1211 .PD 0
1212 .TP
1213 .B DN_ACCESS
1214 A file was accessed
1215 .RB ( read (2),
1216 .BR pread (2),
1217 .BR readv (2),
1218 and similar)
1219 .TP
1220 .B DN_MODIFY
1221 A file was modified
1222 .RB ( write (2),
1223 .BR pwrite (2),
1224 .BR writev (2),
1225 .BR truncate (2),
1226 .BR ftruncate (2),
1227 and similar).
1228 .TP
1229 .B DN_CREATE
1230 A file was created
1231 .RB ( open (2),
1232 .BR creat (2),
1233 .BR mknod (2),
1234 .BR mkdir (2),
1235 .BR link (2),
1236 .BR symlink (2),
1237 .BR rename (2)
1238 into this directory).
1239 .TP
1240 .B DN_DELETE
1241 A file was unlinked
1242 .RB ( unlink (2),
1243 .BR rename (2)
1244 to another directory,
1245 .BR rmdir (2)).
1246 .TP
1247 .B DN_RENAME
1248 A file was renamed within this directory
1249 .RB ( rename (2)).
1250 .TP
1251 .B DN_ATTRIB
1252 The attributes of a file were changed
1253 .RB ( chown (2),
1254 .BR chmod (2),
1255 .BR utime (2),
1256 .BR utimensat (2),
1257 and similar).
1258 .PD
1259 .RE
1260 .IP
1261 (In order to obtain these definitions, the
1262 .B _GNU_SOURCE
1263 feature test macro must be defined before including
1264 .I any
1265 header files.)
1266 .IP
1267 Directory notifications are normally "one-shot", and the application
1268 must reregister to receive further notifications.
1269 Alternatively, if
1270 .B DN_MULTISHOT
1271 is included in
1272 .IR arg ,
1273 then notification will remain in effect until explicitly removed.
1274 .IP
1275 .\" The following does seem a poor API-design choice...
1276 A series of
1277 .B F_NOTIFY
1278 requests is cumulative, with the events in
1279 .I arg
1280 being added to the set already monitored.
1281 To disable notification of all events, make an
1282 .B F_NOTIFY
1283 call specifying
1284 .I arg
1285 as 0.
1286 .IP
1287 Notification occurs via delivery of a signal.
1288 The default signal is
1289 .BR SIGIO ,
1290 but this can be changed using the
1291 .B F_SETSIG
1292 operation to
1293 .BR fcntl ().
1294 (Note that
1295 .B SIGIO
1296 is one of the nonqueuing standard signals;
1297 switching to the use of a real-time signal means that
1298 multiple notifications can be queued to the process.)
1299 In the latter case, the signal handler receives a
1300 .I siginfo_t
1301 structure as its second argument (if the handler was
1302 established using
1303 .BR SA_SIGINFO )
1304 and the
1305 .I si_fd
1306 field of this structure contains the file descriptor which
1307 generated the notification (useful when establishing notification
1308 on multiple directories).
1309 .IP
1310 Especially when using
1311 .BR DN_MULTISHOT ,
1312 a real time signal should be used for notification,
1313 so that multiple notifications can be queued.
1314 .IP
1315 .B NOTE:
1316 New applications should use the
1317 .I inotify
1318 interface (available since Linux 2.6.13),
1319 which provides a much superior interface for obtaining notifications of
1320 filesystem events.
1321 See
1322 .BR inotify (7).
1323 .SS Changing the capacity of a pipe
1324 .TP
1325 .BR F_SETPIPE_SZ " (\fIint\fP; since Linux 2.6.35)"
1326 Change the capacity of the pipe referred to by
1327 .I fd
1328 to be at least
1329 .I arg
1330 bytes.
1331 An unprivileged process can adjust the pipe capacity to any value
1332 between the system page size and the limit defined in
1333 .I /proc/sys/fs/pipe\-max\-size
1334 (see
1335 .BR proc (5)).
1336 Attempts to set the pipe capacity below the page size are silently
1337 rounded up to the page size.
1338 Attempts by an unprivileged process to set the pipe capacity above the limit in
1339 .I /proc/sys/fs/pipe\-max\-size
1340 yield the error
1341 .BR EPERM ;
1342 a privileged process
1343 .RB ( CAP_SYS_RESOURCE )
1344 can override the limit.
1345 .IP
1346 When allocating the buffer for the pipe,
1347 the kernel may use a capacity larger than
1348 .IR arg ,
1349 if that is convenient for the implementation.
1350 (In the current implementation,
1351 the allocation is the next higher power-of-two page-size multiple
1352 of the requested size.)
1353 The actual capacity (in bytes) that is set is returned as the function result.
1354 .IP
1355 Attempting to set the pipe capacity smaller than the amount
1356 of buffer space currently used to store data produces the error
1357 .BR EBUSY .
1358 .IP
1359 Note that because of the way the pages of the pipe buffer
1360 are employed when data is written to the pipe,
1361 the number of bytes that can be written may be less than the nominal size,
1362 depending on the size of the writes.
1363 .TP
1364 .BR F_GETPIPE_SZ " (\fIvoid\fP; since Linux 2.6.35)"
1365 Return (as the function result) the capacity of the pipe referred to by
1366 .IR fd .
1367 .\"
1368 .SS File Sealing
1369 File seals limit the set of allowed operations on a given file.
1370 For each seal that is set on a file,
1371 a specific set of operations will fail with
1372 .B EPERM
1373 on this file from now on.
1374 The file is said to be sealed.
1375 The default set of seals depends on the type of the underlying
1376 file and filesystem.
1377 For an overview of file sealing, a discussion of its purpose,
1378 and some code examples, see
1379 .BR memfd_create (2).
1380 .P
1381 Currently,
1382 file seals can be applied only to a file descriptor returned by
1383 .BR memfd_create (2)
1384 (if the
1385 .B MFD_ALLOW_SEALING
1386 was employed).
1387 On other filesystems, all
1388 .BR fcntl ()
1389 operations that operate on seals will return
1390 .BR EINVAL .
1391 .P
1392 Seals are a property of an inode.
1393 Thus, all open file descriptors referring to the same inode share
1394 the same set of seals.
1395 Furthermore, seals can never be removed, only added.
1396 .TP
1397 .BR F_ADD_SEALS " (\fIint\fP; since Linux 3.17)"
1398 Add the seals given in the bit-mask argument
1399 .I arg
1400 to the set of seals of the inode referred to by the file descriptor
1401 .IR fd .
1402 Seals cannot be removed again.
1403 Once this call succeeds, the seals are enforced by the kernel immediately.
1404 If the current set of seals includes
1405 .B F_SEAL_SEAL
1406 (see below), then this call will be rejected with
1407 .BR EPERM .
1408 Adding a seal that is already set is a no-op, in case
1409 .B F_SEAL_SEAL
1410 is not set already.
1411 In order to place a seal, the file descriptor
1412 .I fd
1413 must be writable.
1414 .TP
1415 .BR F_GET_SEALS " (\fIvoid\fP; since Linux 3.17)"
1416 Return (as the function result) the current set of seals
1417 of the inode referred to by
1418 .IR fd .
1419 If no seals are set, 0 is returned.
1420 If the file does not support sealing, \-1 is returned and
1421 .I errno
1422 is set to
1423 .BR EINVAL .
1424 .P
1425 The following seals are available:
1426 .TP
1427 .B F_SEAL_SEAL
1428 If this seal is set, any further call to
1429 .BR fcntl ()
1430 with
1431 .B F_ADD_SEALS
1432 fails with the error
1433 .BR EPERM .
1434 Therefore, this seal prevents any modifications to the set of seals itself.
1435 If the initial set of seals of a file includes
1436 .BR F_SEAL_SEAL ,
1437 then this effectively causes the set of seals to be constant and locked.
1438 .TP
1439 .B F_SEAL_SHRINK
1440 If this seal is set, the file in question cannot be reduced in size.
1441 This affects
1442 .BR open (2)
1443 with the
1444 .B O_TRUNC
1445 flag as well as
1446 .BR truncate (2)
1447 and
1448 .BR ftruncate (2).
1449 Those calls fail with
1450 .B EPERM
1451 if you try to shrink the file in question.
1452 Increasing the file size is still possible.
1453 .TP
1454 .B F_SEAL_GROW
1455 If this seal is set, the size of the file in question cannot be increased.
1456 This affects
1457 .BR write (2)
1458 beyond the end of the file,
1459 .BR truncate (2),
1460 .BR ftruncate (2),
1461 and
1462 .BR fallocate (2).
1463 These calls fail with
1464 .B EPERM
1465 if you use them to increase the file size.
1466 If you keep the size or shrink it, those calls still work as expected.
1467 .TP
1468 .B F_SEAL_WRITE
1469 If this seal is set, you cannot modify the contents of the file.
1470 Note that shrinking or growing the size of the file is
1471 still possible and allowed.
1472 .\" One or more other seals are typically used with F_SEAL_WRITE
1473 .\" because, given a file with the F_SEAL_WRITE seal set, then,
1474 .\" while it would no longer be possible to (say) write zeros into
1475 .\" the last 100 bytes of a file, it would still be possible
1476 .\" to (say) shrink the file by 100 bytes using ftruncate(), and
1477 .\" then increase the file size by 100 bytes, which would have
1478 .\" the effect of replacing the last hundred bytes by zeros.
1479 .\"
1480 Thus, this seal is normally used in combination with one of the other seals.
1481 This seal affects
1482 .BR write (2)
1483 and
1484 .BR fallocate (2)
1485 (only in combination with the
1486 .B FALLOC_FL_PUNCH_HOLE
1487 flag).
1488 Those calls fail with
1489 .B EPERM
1490 if this seal is set.
1491 Furthermore, trying to create new shared, writable memory-mappings via
1492 .BR mmap (2)
1493 will also fail with
1494 .BR EPERM .
1495 .IP
1496 Using the
1497 .B F_ADD_SEALS
1498 operation to set the
1499 .B F_SEAL_WRITE
1500 seal fails with
1501 .B EBUSY
1502 if any writable, shared mapping exists.
1503 Such mappings must be unmapped before you can add this seal.
1504 Furthermore, if there are any asynchronous I/O operations
1505 .RB ( io_submit (2))
1506 pending on the file,
1507 all outstanding writes will be discarded.
1508 .TP
1509 .BR F_SEAL_FUTURE_WRITE " (since Linux 5.1)"
1510 The effect of this seal is similar to
1511 .BR F_SEAL_WRITE ,
1512 but the contents of the file can still be modified via
1513 shared writable mappings that were created prior to the seal being set.
1514 Any attempt to create a new writable mapping on the file via
1515 .BR mmap (2)
1516 will fail with
1517 .BR EPERM .
1518 Likewise, an attempt to write to the file via
1519 .BR write (2)
1520 will fail with
1521 .BR EPERM .
1522 .IP
1523 Using this seal,
1524 one process can create a memory buffer that it can continue to modify
1525 while sharing that buffer on a "read-only" basis with other processes.
1526 .\"
1527 .SS File read/write hints
1528 Write lifetime hints can be used to inform the kernel about the relative
1529 expected lifetime of writes on a given inode or
1530 via a particular open file description.
1531 (See
1532 .BR open (2)
1533 for an explanation of open file descriptions.)
1534 In this context, the term "write lifetime" means
1535 the expected time the data will live on media, before
1536 being overwritten or erased.
1537 .P
1538 An application may use the different hint values specified below to
1539 separate writes into different write classes,
1540 so that multiple users or applications running on a single storage back-end
1541 can aggregate their I/O patterns in a consistent manner.
1542 However, there are no functional semantics implied by these flags,
1543 and different I/O classes can use the write lifetime hints
1544 in arbitrary ways, so long as the hints are used consistently.
1545 .P
1546 The following operations can be applied to the file descriptor,
1547 .IR fd :
1548 .TP
1549 .BR F_GET_RW_HINT " (\fIuint64_t *\fP; since Linux 4.13)"
1550 Returns the value of the read/write hint associated with the underlying inode
1551 referred to by
1552 .IR fd .
1553 .TP
1554 .BR F_SET_RW_HINT " (\fIuint64_t *\fP; since Linux 4.13)"
1555 Sets the read/write hint value associated with the
1556 underlying inode referred to by
1557 .IR fd .
1558 This hint persists until either it is explicitly modified or
1559 the underlying filesystem is unmounted.
1560 .TP
1561 .BR F_GET_FILE_RW_HINT " (\fIuint64_t *\fP; since Linux 4.13)"
1562 Returns the value of the read/write hint associated with
1563 the open file description referred to by
1564 .IR fd .
1565 .TP
1566 .BR F_SET_FILE_RW_HINT " (\fIuint64_t *\fP; since Linux 4.13)"
1567 Sets the read/write hint value associated with the open file description
1568 referred to by
1569 .IR fd .
1570 .P
1571 If an open file description has not been assigned a read/write hint,
1572 then it shall use the value assigned to the inode, if any.
1573 .P
1574 The following read/write
1575 hints are valid since Linux 4.13:
1576 .TP
1577 .B RWH_WRITE_LIFE_NOT_SET
1578 No specific hint has been set.
1579 This is the default value.
1580 .TP
1581 .B RWH_WRITE_LIFE_NONE
1582 No specific write lifetime is associated with this file or inode.
1583 .TP
1584 .B RWH_WRITE_LIFE_SHORT
1585 Data written to this inode or via this open file description
1586 is expected to have a short lifetime.
1587 .TP
1588 .B RWH_WRITE_LIFE_MEDIUM
1589 Data written to this inode or via this open file description
1590 is expected to have a lifetime longer than
1591 data written with
1592 .BR RWH_WRITE_LIFE_SHORT .
1593 .TP
1594 .B RWH_WRITE_LIFE_LONG
1595 Data written to this inode or via this open file description
1596 is expected to have a lifetime longer than
1597 data written with
1598 .BR RWH_WRITE_LIFE_MEDIUM .
1599 .TP
1600 .B RWH_WRITE_LIFE_EXTREME
1601 Data written to this inode or via this open file description
1602 is expected to have a lifetime longer than
1603 data written with
1604 .BR RWH_WRITE_LIFE_LONG .
1605 .P
1606 All the write-specific hints are relative to each other,
1607 and no individual absolute meaning should be attributed to them.
1608 .SH RETURN VALUE
1609 For a successful call, the return value depends on the operation:
1610 .TP
1611 .B F_DUPFD
1612 The new file descriptor.
1613 .TP
1614 .B F_GETFD
1615 Value of file descriptor flags.
1616 .TP
1617 .B F_GETFL
1618 Value of file status flags.
1619 .TP
1620 .B F_GETLEASE
1621 Type of lease held on file descriptor.
1622 .TP
1623 .B F_GETOWN
1624 Value of file descriptor owner.
1625 .TP
1626 .B F_GETSIG
1627 Value of signal sent when read or write becomes possible, or zero
1628 for traditional
1629 .B SIGIO
1630 behavior.
1631 .TP
1632 .B F_GETPIPE_SZ
1633 .TQ
1634 .B F_SETPIPE_SZ
1635 The pipe capacity.
1636 .TP
1637 .B F_GET_SEALS
1638 A bit mask identifying the seals that have been set
1639 for the inode referred to by
1640 .IR fd .
1641 .TP
1642 All other operations
1643 Zero.
1644 .P
1645 On error, \-1 is returned, and
1646 .I errno
1647 is set to indicate the error.
1648 .SH ERRORS
1649 .TP
1650 .BR EACCES " or " EAGAIN
1651 Operation is prohibited by locks held by other processes.
1652 .TP
1653 .B EAGAIN
1654 The operation is prohibited because the file has been memory-mapped by
1655 another process.
1656 .TP
1657 .B EBADF
1658 .I fd
1659 is not an open file descriptor
1660 .TP
1661 .B EBADF
1662 .I op
1663 is
1664 .B F_SETLK
1665 or
1666 .B F_SETLKW
1667 and the file descriptor open mode doesn't match with the
1668 type of lock requested.
1669 .TP
1670 .B EBUSY
1671 .I op
1672 is
1673 .B F_SETPIPE_SZ
1674 and the new pipe capacity specified in
1675 .I arg
1676 is smaller than the amount of buffer space currently
1677 used to store data in the pipe.
1678 .TP
1679 .B EBUSY
1680 .I op
1681 is
1682 .BR F_ADD_SEALS ,
1683 .I arg
1684 includes
1685 .BR F_SEAL_WRITE ,
1686 and there exists a writable, shared mapping on the file referred to by
1687 .IR fd .
1688 .TP
1689 .B EDEADLK
1690 It was detected that the specified
1691 .B F_SETLKW
1692 operation would cause a deadlock.
1693 .TP
1694 .B EFAULT
1695 .I lock
1696 is outside your accessible address space.
1697 .TP
1698 .B EINTR
1699 .I op
1700 is
1701 .B F_SETLKW
1702 or
1703 .B F_OFD_SETLKW
1704 and the operation was interrupted by a signal; see
1705 .BR signal (7).
1706 .TP
1707 .B EINTR
1708 .I op
1709 is
1710 .BR F_GETLK ,
1711 .BR F_SETLK ,
1712 .BR F_OFD_GETLK ,
1713 or
1714 .BR F_OFD_SETLK ,
1715 and the operation was interrupted by a signal before the lock was checked or
1716 acquired.
1717 Most likely when locking a remote file (e.g., locking over
1718 NFS), but can sometimes happen locally.
1719 .TP
1720 .B EINVAL
1721 The value specified in
1722 .I op
1723 is not recognized by this kernel.
1724 .TP
1725 .B EINVAL
1726 .I op
1727 is
1728 .B F_ADD_SEALS
1729 and
1730 .I arg
1731 includes an unrecognized sealing bit.
1732 .TP
1733 .B EINVAL
1734 .I op
1735 is
1736 .B F_ADD_SEALS
1737 or
1738 .B F_GET_SEALS
1739 and the filesystem containing the inode referred to by
1740 .I fd
1741 does not support sealing.
1742 .TP
1743 .B EINVAL
1744 .I op
1745 is
1746 .B F_DUPFD
1747 and
1748 .I arg
1749 is negative or is greater than the maximum allowable value
1750 (see the discussion of
1751 .B RLIMIT_NOFILE
1752 in
1753 .BR getrlimit (2)).
1754 .TP
1755 .B EINVAL
1756 .I op
1757 is
1758 .B F_SETSIG
1759 and
1760 .I arg
1761 is not an allowable signal number.
1762 .TP
1763 .B EINVAL
1764 .I op
1765 is
1766 .BR F_OFD_SETLK ,
1767 .BR F_OFD_SETLKW ,
1768 or
1769 .BR F_OFD_GETLK ,
1770 and
1771 .I l_pid
1772 was not specified as zero.
1773 .TP
1774 .B EMFILE
1775 .I op
1776 is
1777 .B F_DUPFD
1778 and the per-process limit on the number of open file descriptors
1779 has been reached.
1780 .TP
1781 .B ENOLCK
1782 Too many segment locks open, lock table is full, or a remote locking
1783 protocol failed (e.g., locking over NFS).
1784 .TP
1785 .B ENOTDIR
1786 .B F_NOTIFY
1787 was specified in
1788 .IR op ,
1789 but
1790 .I fd
1791 does not refer to a directory.
1792 .TP
1793 .B EPERM
1794 .I op
1795 is
1796 .B F_SETPIPE_SZ
1797 and the soft or hard user pipe limit has been reached; see
1798 .BR pipe (7).
1799 .TP
1800 .B EPERM
1801 Attempted to clear the
1802 .B O_APPEND
1803 flag on a file that has the append-only attribute set.
1804 .TP
1805 .B EPERM
1806 .I op
1807 was
1808 .BR F_ADD_SEALS ,
1809 but
1810 .I fd
1811 was not open for writing
1812 or the current set of seals on the file already includes
1813 .BR F_SEAL_SEAL .
1814 .SH STANDARDS
1815 POSIX.1-2008.
1816 .P
1817 .BR F_GETOWN_EX ,
1818 .BR F_SETOWN_EX ,
1819 .BR F_SETPIPE_SZ ,
1820 .BR F_GETPIPE_SZ ,
1821 .BR F_GETSIG ,
1822 .BR F_SETSIG ,
1823 .BR F_NOTIFY ,
1824 .BR F_GETLEASE ,
1825 and
1826 .B F_SETLEASE
1827 are Linux-specific.
1828 (Define the
1829 .B _GNU_SOURCE
1830 macro to obtain these definitions.)
1831 .\" .P
1832 .\" SVr4 documents additional EIO, ENOLINK and EOVERFLOW error conditions.
1833 .P
1834 .BR F_OFD_SETLK ,
1835 .BR F_OFD_SETLKW ,
1836 and
1837 .B F_OFD_GETLK
1838 are Linux-specific (and one must define
1839 .B _GNU_SOURCE
1840 to obtain their definitions),
1841 but work is being done to have them included in the next version of POSIX.1.
1842 .P
1843 .B F_ADD_SEALS
1844 and
1845 .B F_GET_SEALS
1846 are Linux-specific.
1847 .\" FIXME . Once glibc adds support, add a note about FTM requirements
1848 .SH HISTORY
1849 SVr4, 4.3BSD, POSIX.1-2001.
1850 .P
1851 Only the operations
1852 .BR F_DUPFD ,
1853 .BR F_GETFD ,
1854 .BR F_SETFD ,
1855 .BR F_GETFL ,
1856 .BR F_SETFL ,
1857 .BR F_GETLK ,
1858 .BR F_SETLK ,
1859 and
1860 .B F_SETLKW
1861 are specified in POSIX.1-2001.
1862 .P
1863 .B F_GETOWN
1864 and
1865 .B F_SETOWN
1866 are specified in POSIX.1-2001.
1867 (To get their definitions, define either
1868 .\" .BR _BSD_SOURCE ,
1869 .\" or
1870 .B _XOPEN_SOURCE
1871 with the value 500 or greater, or
1872 .B _POSIX_C_SOURCE
1873 with the value 200809L or greater.)
1874 .P
1875 .B F_DUPFD_CLOEXEC
1876 is specified in POSIX.1-2008.
1877 (To get this definition, define
1878 .B _POSIX_C_SOURCE
1879 with the value 200809L or greater, or
1880 .B _XOPEN_SOURCE
1881 with the value 700 or greater.)
1882 .SH NOTES
1883 The errors returned by
1884 .BR dup2 (2)
1885 are different from those returned by
1886 .BR F_DUPFD .
1887 .\"
1888 .SS File locking
1889 The original Linux
1890 .BR fcntl ()
1891 system call was not designed to handle large file offsets
1892 (in the
1893 .I flock
1894 structure).
1895 Consequently, an
1896 .BR fcntl64 ()
1897 system call was added in Linux 2.4.
1898 The newer system call employs a different structure for file locking,
1899 .IR flock64 ,
1900 and corresponding operations,
1901 .BR F_GETLK64 ,
1902 .BR F_SETLK64 ,
1903 and
1904 .BR F_SETLKW64 .
1905 However, these details can be ignored by applications using glibc, whose
1906 .BR fcntl ()
1907 wrapper function transparently employs the more recent system call
1908 where it is available.
1909 .\"
1910 .SS Record locks
1911 Since Linux 2.0, there is no interaction between the types of lock
1912 placed by
1913 .BR flock (2)
1914 and
1915 .BR fcntl ().
1916 .P
1917 Several systems have more fields in
1918 .I "struct flock"
1919 such as, for example,
1920 .I l_sysid
1921 (to identify the machine where the lock is held).
1922 .\" e.g., Solaris 8 documents this field in fcntl(2), and Irix 6.5
1923 .\" documents it in fcntl(5). mtk, May 2007
1924 .\" Also, FreeBSD documents it (Apr 2014).
1925 Clearly,
1926 .I l_pid
1927 alone is not going to be very useful if the process holding the lock
1928 may live on a different machine;
1929 on Linux, while present on some architectures (such as MIPS32),
1930 this field is not used.
1931 .P
1932 The original Linux
1933 .BR fcntl ()
1934 system call was not designed to handle large file offsets
1935 (in the
1936 .I flock
1937 structure).
1938 Consequently, an
1939 .BR fcntl64 ()
1940 system call was added in Linux 2.4.
1941 The newer system call employs a different structure for file locking,
1942 .IR flock64 ,
1943 and corresponding operations,
1944 .BR F_GETLK64 ,
1945 .BR F_SETLK64 ,
1946 and
1947 .BR F_SETLKW64 .
1948 However, these details can be ignored by applications using glibc, whose
1949 .BR fcntl ()
1950 wrapper function transparently employs the more recent system call
1951 where it is available.
1952 .SS Record locking and NFS
1953 Before Linux 3.12, if an NFSv4 client
1954 loses contact with the server for a period of time
1955 (defined as more than 90 seconds with no communication),
1956 .\"
1957 .\" Neil Brown: With NFSv3 the failure mode is the reverse. If
1958 .\" the server loses contact with a client then any lock stays in place
1959 .\" indefinitely ("why can't I read my mail"... I remember it well).
1960 .\"
1961 it might lose and regain a lock without ever being aware of the fact.
1962 (The period of time after which contact is assumed lost is known as
1963 the NFSv4 leasetime.
1964 On a Linux NFS server, this can be determined by looking at
1965 .IR /proc/fs/nfsd/nfsv4leasetime ,
1966 which expresses the period in seconds.
1967 The default value for this file is 90.)
1968 .\"
1969 .\" Jeff Layton:
1970 .\" Note that this is not a firm timeout. The server runs a job
1971 .\" periodically to clean out expired stateful objects, and it's likely
1972 .\" that there is some time (maybe even up to another whole lease period)
1973 .\" between when the timeout expires and the job actually runs. If the
1974 .\" client gets a RENEW in there within that window, its lease will be
1975 .\" renewed and its state preserved.
1976 .\"
1977 This scenario potentially risks data corruption,
1978 since another process might acquire a lock in the intervening period
1979 and perform file I/O.
1980 .P
1981 Since Linux 3.12,
1982 .\" commit ef1820f9be27b6ad158f433ab38002ab8131db4d
1983 if an NFSv4 client loses contact with the server,
1984 any I/O to the file by a process which "thinks" it holds
1985 a lock will fail until that process closes and reopens the file.
1986 A kernel parameter,
1987 .IR nfs.recover_lost_locks ,
1988 can be set to 1 to obtain the pre-3.12 behavior,
1989 whereby the client will attempt to recover lost locks
1990 when contact is reestablished with the server.
1991 Because of the attendant risk of data corruption,
1992 .\" commit f6de7a39c181dfb8a2c534661a53c73afb3081cd
1993 this parameter defaults to 0 (disabled).
1994 .SH BUGS
1995 .SS F_SETFL
1996 It is not possible to use
1997 .B F_SETFL
1998 to change the state of the
1999 .B O_DSYNC
2000 and
2001 .B O_SYNC
2002 flags.
2003 .\" FIXME . According to POSIX.1-2001, O_SYNC should also be modifiable
2004 .\" via fcntl(2), but currently Linux does not permit this
2005 .\" See http://bugzilla.kernel.org/show_bug.cgi?id=5994
2006 Attempts to change the state of these flags are silently ignored.
2007 .SS F_GETOWN
2008 A limitation of the Linux system call conventions on some
2009 architectures (notably i386) means that if a (negative)
2010 process group ID to be returned by
2011 .B F_GETOWN
2012 falls in the range \-1 to \-4095, then the return value is wrongly
2013 interpreted by glibc as an error in the system call;
2014 .\" glibc source: sysdeps/unix/sysv/linux/i386/sysdep.h
2015 that is, the return value of
2016 .BR fcntl ()
2017 will be \-1, and
2018 .I errno
2019 will contain the (positive) process group ID.
2020 The Linux-specific
2021 .B F_GETOWN_EX
2022 operation avoids this problem.
2023 .\" mtk, Dec 04: some limited testing on alpha and ia64 seems to
2024 .\" indicate that ANY negative PGID value will cause F_GETOWN
2025 .\" to misinterpret the return as an error. Some other architectures
2026 .\" seem to have the same range check as i386.
2027 Since glibc 2.11, glibc makes the kernel
2028 .B F_GETOWN
2029 problem invisible by implementing
2030 .B F_GETOWN
2031 using
2032 .BR F_GETOWN_EX .
2033 .SS F_SETOWN
2034 In Linux 2.4 and earlier, there is bug that can occur
2035 when an unprivileged process uses
2036 .B F_SETOWN
2037 to specify the owner
2038 of a socket file descriptor
2039 as a process (group) other than the caller.
2040 In this case,
2041 .BR fcntl ()
2042 can return \-1 with
2043 .I errno
2044 set to
2045 .BR EPERM ,
2046 even when the owner process (group) is one that the caller
2047 has permission to send signals to.
2048 Despite this error return, the file descriptor owner is set,
2049 and signals will be sent to the owner.
2050 .\"
2051 .SS Deadlock detection
2052 The deadlock-detection algorithm employed by the kernel when dealing with
2053 .B F_SETLKW
2054 requests can yield both
2055 false negatives (failures to detect deadlocks,
2056 leaving a set of deadlocked processes blocked indefinitely)
2057 and false positives
2058 .RB ( EDEADLK
2059 errors when there is no deadlock).
2060 For example,
2061 the kernel limits the lock depth of its dependency search to 10 steps,
2062 meaning that circular deadlock chains that exceed
2063 that size will not be detected.
2064 In addition, the kernel may falsely indicate a deadlock
2065 when two or more processes created using the
2066 .BR clone (2)
2067 .B CLONE_FILES
2068 flag place locks that appear (to the kernel) to conflict.
2069 .\"
2070 .SS Mandatory locking
2071 The Linux implementation of mandatory locking
2072 is subject to race conditions which render it unreliable:
2073 .\" http://marc.info/?l=linux-kernel&m=119013491707153&w=2
2074 .\"
2075 .\" Reconfirmed by Jeff Layton
2076 .\" From: Jeff Layton <jlayton <at> redhat.com>
2077 .\" Subject: Re: Status of fcntl() mandatory locking
2078 .\" Newsgroups: gmane.linux.file-systems
2079 .\" Date: 2014-04-28 10:07:57 GMT
2080 .\" http://thread.gmane.org/gmane.linux.file-systems/84481/focus=84518
2081 a
2082 .BR write (2)
2083 call that overlaps with a lock may modify data after the mandatory lock is
2084 acquired;
2085 a
2086 .BR read (2)
2087 call that overlaps with a lock may detect changes to data that were made
2088 only after a write lock was acquired.
2089 Similar races exist between mandatory locks and
2090 .BR mmap (2).
2091 It is therefore inadvisable to rely on mandatory locking.
2092 .SH SEE ALSO
2093 .BR dup2 (2),
2094 .BR flock (2),
2095 .BR open (2),
2096 .BR socket (2),
2097 .BR lockf (3),
2098 .BR capabilities (7),
2099 .BR feature_test_macros (7),
2100 .BR lslocks (8)
2101 .P
2102 .IR locks.txt ,
2103 .IR mandatory\-locking.txt ,
2104 and
2105 .I dnotify.txt
2106 in the Linux kernel source directory
2107 .I Documentation/filesystems/
2108 (on older kernels, these files are directly under the
2109 .I Documentation/
2110 directory, and
2111 .I mandatory\-locking.txt
2112 is called
2113 .IR mandatory.txt )