path: root/upstream/opensuse-leap-15-6/man2/fcntl.2

.\" This manpage is Copyright (C) 1992 Drew Eckhardt;
.\" and Copyright (C) 1993 Michael Haardt, Ian Jackson;
.\" and Copyright (C) 1998 Jamie Lokier;
.\" and Copyright (C) 2002-2010, 2014 Michael Kerrisk;
.\" and Copyright (C) 2014 Jeff Layton
.\" and Copyright (C) 2014 David Herrmann
.\" and Copyright (C) 2017 Jens Axboe
.\"
.\" SPDX-License-Identifier: Linux-man-pages-copyleft
.\"
.\" Modified 1993-07-24 by Rik Faith <faith@cs.unc.edu>
.\" Modified 1995-09-26 by Andries Brouwer <aeb@cwi.nl>
.\" and again on 960413 and 980804 and 981223.
.\" Modified 1998-12-11 by Jamie Lokier <jamie@imbolc.ucc.ie>
.\" Applied correction by Christian Ehrhardt - aeb, 990712
.\" Modified 2002-04-23 by Michael Kerrisk <mtk.manpages@gmail.com>
.\"	Added note on F_SETFL and O_DIRECT
.\"	Complete rewrite + expansion of material on file locking
.\"	Incorporated description of F_NOTIFY, drawing on
.\"		Stephen Rothwell's notes in Documentation/dnotify.txt.
.\"	Added description of F_SETLEASE and F_GETLEASE
.\" Corrected and polished, aeb, 020527.
.\" Modified 2004-03-03 by Michael Kerrisk <mtk.manpages@gmail.com>
.\"     Modified description of file leases: fixed some errors of detail
.\"     Replaced the term "lease contestant" by "lease breaker"
.\" Modified, 27 May 2004, Michael Kerrisk <mtk.manpages@gmail.com>
.\"     Added notes on capability requirements
.\" Modified 2004-12-08, added O_NOATIME after note from Martin Pool
.\" 2004-12-10, mtk, noted F_GETOWN bug after suggestion from aeb.
.\" 2005-04-08 Jamie Lokier <jamie@shareable.org>, mtk
.\"	Described behavior of F_SETOWN/F_SETSIG in
.\"	multithreaded processes, and generally cleaned
.\"	up the discussion of F_SETOWN.
.\" 2005-05-20, Johannes Nicolai <johannes.nicolai@hpi.uni-potsdam.de>,
.\"	mtk: Noted F_SETOWN bug for socket file descriptor in Linux 2.4
.\"	and earlier.  Added text on permissions required to send signal.
.\" 2009-09-30, Michael Kerrisk
.\"     Note obsolete F_SETOWN behavior with threads.
.\"     Document F_SETOWN_EX and F_GETOWN_EX
.\" 2010-06-17, Michael Kerrisk
.\"	Document F_SETPIPE_SZ and F_GETPIPE_SZ.
.\" 2014-07-08, David Herrmann <dh.herrmann@gmail.com>
.\"     Document F_ADD_SEALS and F_GET_SEALS
.\" 2017-06-26, Jens Axboe <axboe@kernel.dk>
.\"     Document F_{GET,SET}_RW_HINT and F_{GET,SET}_FILE_RW_HINT
.\"
.TH fcntl 2 2023-03-30 "Linux man-pages 6.04"
.SH NAME
fcntl \- manipulate file descriptor
.SH LIBRARY
Standard C library
.RI ( libc ", " \-lc )
.SH SYNOPSIS
.nf
.B #include <fcntl.h>
.PP
.BI "int fcntl(int " fd ", int " cmd ", ... /* " arg " */ );"
.fi
.SH DESCRIPTION
.BR fcntl ()
performs one of the operations described below on the open file descriptor
.IR fd .
The operation is determined by
.IR cmd .
.PP
.BR fcntl ()
can take an optional third argument.
Whether or not this argument is required is determined by
.IR cmd .
The required argument type is indicated in parentheses after each
.I cmd
name (in most cases, the required type is
.IR int ,
and we identify the argument using the name
.IR arg ),
or
.I void
is specified if the argument is not required.
.PP
Certain of the operations below are supported only since a particular
Linux kernel version.
The preferred method of checking whether the host kernel supports
a particular operation is to invoke
.BR fcntl ()
with the desired
.I cmd
value and then test whether the call failed with
.BR EINVAL ,
indicating that the kernel does not recognize this value.
.SS Duplicating a file descriptor
.TP
.BR F_DUPFD " (\fIint\fP)"
Duplicate the file descriptor
.I fd
using the lowest-numbered available file descriptor greater than or equal to
.IR arg .
This is different from
.BR dup2 (2),
which uses exactly the file descriptor specified.
.IP
On success, the new file descriptor is returned.
.IP
See
.BR dup (2)
for further details.
.TP
.BR F_DUPFD_CLOEXEC " (\fIint\fP; since Linux 2.6.24)"
As for
.BR F_DUPFD ,
but additionally set the
close-on-exec flag for the duplicate file descriptor.
Specifying this flag permits a program to avoid an additional
.BR fcntl ()
.B F_SETFD
operation to set the
.B FD_CLOEXEC
flag.
For an explanation of why this flag is useful,
see the description of
.B O_CLOEXEC
in
.BR open (2).
.SS File descriptor flags
The following commands manipulate the flags associated with
a file descriptor.
Currently, only one such flag is defined:
.BR FD_CLOEXEC ,
the close-on-exec flag.
If the
.B FD_CLOEXEC
bit is set,
the file descriptor will automatically be closed during a successful
.BR execve (2).
(If the
.BR execve (2)
fails, the file descriptor is left open.)
If the
.B FD_CLOEXEC
bit is not set, the file descriptor will remain open across an
.BR execve (2).
.TP
.BR F_GETFD " (\fIvoid\fP)"
Return (as the function result) the file descriptor flags;
.I arg
is ignored.
.TP
.BR F_SETFD " (\fIint\fP)"
Set the file descriptor flags to the value specified by
.IR arg .
.PP
In multithreaded programs, using
.BR fcntl ()
.B F_SETFD
to set the close-on-exec flag at the same time as another thread performs a
.BR fork (2)
plus
.BR execve (2)
is vulnerable to a race condition that may unintentionally leak
the file descriptor to the program executed in the child process.
See the discussion of the
.B O_CLOEXEC
flag in
.BR open (2)
for details and a remedy to the problem.
.SS File status flags
Each open file description has certain associated status flags,
initialized by
.BR open (2)
.\" or
.\" .BR creat (2),
and possibly modified by
.BR fcntl ().
Duplicated file descriptors
(made with
.BR dup (2),
.BR fcntl (F_DUPFD),
.BR fork (2),
etc.) refer to the same open file description, and thus
share the same file status flags.
.PP
The file status flags and their semantics are described in
.BR open (2).
.TP
.BR F_GETFL " (\fIvoid\fP)"
Return (as the function result)
the file access mode and the file status flags;
.I arg
is ignored.
.TP
.BR F_SETFL " (\fIint\fP)"
Set the file status flags to the value specified by
.IR arg .
File access mode
.RB ( O_RDONLY ", " O_WRONLY ", " O_RDWR )
and file creation flags
(i.e.,
.BR O_CREAT ", " O_EXCL ", " O_NOCTTY ", " O_TRUNC )
in
.I arg
are ignored.
On Linux, this command can change only the
.BR O_APPEND ,
.BR O_ASYNC ,
.BR O_DIRECT ,
.BR O_NOATIME ,
and
.B O_NONBLOCK
flags.
It is not possible to change the
.B O_DSYNC
and
.B O_SYNC
flags; see BUGS, below.
.SS Advisory record locking
Linux implements traditional ("process-associated") UNIX record locks,
as standardized by POSIX.
For a Linux-specific alternative with better semantics,
see the discussion of open file description locks below.
.PP
.BR F_SETLK ,
.BR F_SETLKW ,
and
.B F_GETLK
are used to acquire, release, and test for the existence of record
locks (also known as byte-range, file-segment, or file-region locks).
The third argument,
.IR lock ,
is a pointer to a structure that has at least the following fields
(in unspecified order).
.PP
.in +4n
.EX
struct flock {
    ...
    short l_type;    /* Type of lock: F_RDLCK,
                        F_WRLCK, F_UNLCK */
    short l_whence;  /* How to interpret l_start:
                        SEEK_SET, SEEK_CUR, SEEK_END */
    off_t l_start;   /* Starting offset for lock */
    off_t l_len;     /* Number of bytes to lock */
    pid_t l_pid;     /* PID of process blocking our lock
                        (set by F_GETLK and F_OFD_GETLK) */
    ...
};
.EE
.in
.PP
The
.IR l_whence ", " l_start ", and " l_len
fields of this structure specify the range of bytes we wish to lock.
Bytes past the end of the file may be locked,
but not bytes before the start of the file.
.PP
.I l_start
is the starting offset for the lock, and is interpreted
relative to either:
the start of the file (if
.I l_whence
is
.BR SEEK_SET );
the current file offset (if
.I l_whence
is
.BR SEEK_CUR );
or the end of the file (if
.I l_whence
is
.BR SEEK_END ).
In the final two cases,
.I l_start
can be a negative number provided the
offset does not lie before the start of the file.
.PP
.I l_len
specifies the number of bytes to be locked.
If
.I l_len
is positive, then the range to be locked covers bytes
.I l_start
up to and including
.IR l_start + l_len \-1.
Specifying 0 for
.I l_len
has the special meaning: lock all bytes starting at the
location specified by
.IR l_whence " and " l_start
through to the end of file, no matter how large the file grows.
.PP
POSIX.1-2001 allows (but does not require)
an implementation to support a negative
.I l_len
value; if
.I l_len
is negative, the interval described by
.I lock
covers bytes
.IR l_start + l_len
up to and including
.IR l_start \-1.
This is supported since Linux 2.4.21 and Linux 2.5.49.
.PP
The
.I l_type
field can be used to place a read
.RB ( F_RDLCK )
or a write
.RB ( F_WRLCK )
lock on a file.
Any number of processes may hold a read lock (shared lock)
on a file region, but only one process may hold a write lock
(exclusive lock).
An exclusive lock excludes all other locks,
both shared and exclusive.
A single process can hold only one type of lock on a file region;
if a new lock is applied to an already-locked region,
then the existing lock is converted to the new lock type.
(Such conversions may involve splitting, shrinking, or coalescing with
an existing lock if the byte range specified by the new lock does not
precisely coincide with the range of the existing lock.)
.TP
.BR F_SETLK " (\fIstruct flock *\fP)"
Acquire a lock (when
.I l_type
is
.B F_RDLCK
or
.BR F_WRLCK )
or release a lock (when
.I l_type
is
.BR F_UNLCK )
on the bytes specified by the
.IR l_whence ", " l_start ", and " l_len
fields of
.IR lock .
If a conflicting lock is held by another process,
this call returns \-1 and sets
.I errno
to
.B EACCES
or
.BR EAGAIN .
(The error returned in this case differs across implementations,
so POSIX requires a portable application to check for both errors.)
.TP
.BR F_SETLKW " (\fIstruct flock *\fP)"
As for
.BR F_SETLK ,
but if a conflicting lock is held on the file, then wait for that
lock to be released.
If a signal is caught while waiting, then the call is interrupted
and (after the signal handler has returned)
returns immediately (with return value \-1 and
.I errno
set to
.BR EINTR ;
see
.BR signal (7)).
.TP
.BR F_GETLK " (\fIstruct flock *\fP)"
On input to this call,
.I lock
describes a lock we would like to place on the file.
If the lock could be placed,
.BR fcntl ()
does not actually place it, but returns
.B F_UNLCK
in the
.I l_type
field of
.I lock
and leaves the other fields of the structure unchanged.
.IP
If one or more incompatible locks would prevent
this lock being placed, then
.BR fcntl ()
returns details about one of those locks in the
.IR l_type ", " l_whence ", " l_start ", and " l_len
fields of
.IR lock .
If the conflicting lock is a traditional (process-associated) record lock,
then the
.I l_pid
field is set to the PID of the process holding that lock.
If the conflicting lock is an open file description lock, then
.I l_pid
is set to \-1.
Note that the returned information
may already be out of date by the time the caller inspects it.
.PP
In order to place a read lock,
.I fd
must be open for reading.
In order to place a write lock,
.I fd
must be open for writing.
To place both types of lock, open a file read-write.
.PP
When placing locks with
.BR F_SETLKW ,
the kernel detects
.IR deadlocks ,
whereby two or more processes have their
lock requests mutually blocked by locks held by the other processes.
For example, suppose process A holds a write lock on byte 100 of a file,
and process B holds a write lock on byte 200.
If each process then attempts to lock the byte already
locked by the other process using
.BR F_SETLKW ,
then, without deadlock detection,
both processes would remain blocked indefinitely.
When the kernel detects such deadlocks,
it causes one of the blocking lock requests to immediately fail with the error
.BR EDEADLK ;
an application that encounters such an error should release
some of its locks to allow other applications to proceed before
attempting regain the locks that it requires.
Circular deadlocks involving more than two processes are also detected.
Note, however, that there are limitations to the kernel's
deadlock-detection algorithm; see BUGS.
.PP
As well as being removed by an explicit
.BR F_UNLCK ,
record locks are automatically released when the process terminates.
.PP
Record locks are not inherited by a child created via
.BR fork (2),
but are preserved across an
.BR execve (2).
.PP
Because of the buffering performed by the
.BR stdio (3)
library, the use of record locking with routines in that package
should be avoided; use
.BR read (2)
and
.BR write (2)
instead.
.PP
The record locks described above are associated with the process
(unlike the open file description locks described below).
This has some unfortunate consequences:
.IP \[bu] 3
If a process closes
.I any
file descriptor referring to a file,
then all of the process's locks on that file are released,
regardless of the file descriptor(s) on which the locks were obtained.
.\" (Additional file descriptors referring to the same file
.\" may have been obtained by calls to
.\" .BR open "(2), " dup "(2), " dup2 "(2), or " fcntl ().)
This is bad: it means that a process can lose its locks on
a file such as
.I /etc/passwd
or
.I /etc/mtab
when for some reason a library function decides to open, read,
and close the same file.
.IP \[bu]
The threads in a process share locks.
In other words,
a multithreaded program can't use record locking to ensure
that threads don't simultaneously access the same region of a file.
.PP
Open file description locks solve both of these problems.
.SS Open file description locks (non-POSIX)
Open file description locks are advisory byte-range locks whose operation is
in most respects identical to the traditional record locks described above.
This lock type is Linux-specific,
and available since Linux 3.15.
(There is a proposal with the Austin Group
.\" FIXME . Review progress into POSIX
.\" http://austingroupbugs.net/view.php?id=768
to include this lock type in the next revision of POSIX.1.)
For an explanation of open file descriptions, see
.BR open (2).
.PP
The principal difference between the two lock types
is that whereas traditional record locks
are associated with a process,
open file description locks are associated with the
open file description on which they are acquired,
much like locks acquired with
.BR flock (2).
Consequently (and unlike traditional advisory record locks),
open file description locks are inherited across
.BR fork (2)
(and
.BR clone (2)
with
.BR CLONE_FILES ),
and are only automatically released on the last close
of the open file description,
instead of being released on any close of the file.
.PP
Conflicting lock combinations
(i.e., a read lock and a write lock or two write locks)
where one lock is an open file description lock and the other
is a traditional record lock conflict
even when they are acquired by the same process on the same file descriptor.
.PP
Open file description locks placed via the same open file description
(i.e., via the same file descriptor,
or via a duplicate of the file descriptor created by
.BR fork (2),
.BR dup (2),
.BR fcntl ()
.BR F_DUPFD ,
and so on) are always compatible:
if a new lock is placed on an already locked region,
then the existing lock is converted to the new lock type.
(Such conversions may result in splitting, shrinking, or coalescing with
an existing lock as discussed above.)
.PP
On the other hand, open file description locks may conflict with
each other when they are acquired via different open file descriptions.
Thus, the threads in a multithreaded program can use
open file description locks to synchronize access to a file region
by having each thread perform its own
.BR open (2)
on the file and applying locks via the resulting file descriptor.
.PP
As with traditional advisory locks, the third argument to
.BR fcntl (),
.IR lock ,
is a pointer to an
.I flock
structure.
By contrast with traditional record locks, the
.I l_pid
field of that structure must be set to zero
when using the commands described below.
.PP
The commands for working with open file description locks are analogous
to those used with traditional locks:
.TP
.BR F_OFD_SETLK " (\fIstruct flock *\fP)"
Acquire an open file description lock (when
.I l_type
is
.B F_RDLCK
or
.BR F_WRLCK )
or release an open file description lock (when
.I l_type
is
.BR F_UNLCK )
on the bytes specified by the
.IR l_whence ", " l_start ", and " l_len
fields of
.IR lock .
If a conflicting lock is held by another process,
this call returns \-1 and sets
.I errno
to
.BR EAGAIN .
.TP
.BR F_OFD_SETLKW " (\fIstruct flock *\fP)"
As for
.BR F_OFD_SETLK ,
but if a conflicting lock is held on the file, then wait for that lock to be
released.
If a signal is caught while waiting, then the call is interrupted
and (after the signal handler has returned) returns immediately
(with return value \-1 and
.I errno
set to
.BR EINTR ;
see
.BR signal (7)).
.TP
.BR F_OFD_GETLK " (\fIstruct flock *\fP)"
On input to this call,
.I lock
describes an open file description lock we would like to place on the file.
If the lock could be placed,
.BR fcntl ()
does not actually place it, but returns
.B F_UNLCK
in the
.I l_type
field of
.I lock
and leaves the other fields of the structure unchanged.
If one or more incompatible locks would prevent this lock being placed,
then details about one of these locks are returned via
.IR lock ,
as described above for
.BR F_GETLK .
.PP
In the current implementation,
.\" commit 57b65325fe34ec4c917bc4e555144b4a94d9e1f7
no deadlock detection is performed for open file description locks.
(This contrasts with process-associated record locks,
for which the kernel does perform deadlock detection.)
.\"
.SS Mandatory locking
.IR Warning :
the Linux implementation of mandatory locking is unreliable.
See BUGS below.
Because of these bugs,
and the fact that the feature is believed to be little used,
since Linux 4.5, mandatory locking has been made an optional feature,
governed by a configuration option
.RB ( CONFIG_MANDATORY_FILE_LOCKING ).
This feature is no longer supported at all in Linux 5.15 and above.
.PP
By default, both traditional (process-associated) and open file description
record locks are advisory.
Advisory locks are not enforced and are useful only between
cooperating processes.
.PP
Both lock types can also be mandatory.
Mandatory locks are enforced for all processes.
If a process tries to perform an incompatible access (e.g.,
.BR read (2)
or
.BR write (2))
on a file region that has an incompatible mandatory lock,
then the result depends upon whether the
.B O_NONBLOCK
flag is enabled for its open file description.
If the
.B O_NONBLOCK
flag is not enabled, then
the system call is blocked until the lock is removed
or converted to a mode that is compatible with the access.
If the
.B O_NONBLOCK
flag is enabled, then the system call fails with the error
.BR EAGAIN .
.PP
To make use of mandatory locks, mandatory locking must be enabled
both on the filesystem that contains the file to be locked,
and on the file itself.
Mandatory locking is enabled on a filesystem
using the "\-o mand" option to
.BR mount (8),
or the
.B MS_MANDLOCK
flag for
.BR mount (2).
Mandatory locking is enabled on a file by disabling
group execute permission on the file and enabling the set-group-ID
permission bit (see
.BR chmod (1)
and
.BR chmod (2)).
.PP
Mandatory locking is not specified by POSIX.
Some other systems also support mandatory locking,
although the details of how to enable it vary across systems.
.\"
.SS Lost locks
When an advisory lock is obtained on a networked filesystem such as
NFS it is possible that the lock might get lost.
This may happen due to administrative action on the server, or due to a
network partition (i.e., loss of network connectivity with the server)
which lasts long enough for the server to assume
that the client is no longer functioning.
.PP
When the filesystem determines that a lock has been lost, future
.BR read (2)
or
.BR write (2)
requests may fail with the error
.BR EIO .
This error will persist until the lock is removed or the file
descriptor is closed.
Since Linux 3.12,
.\" commit ef1820f9be27b6ad158f433ab38002ab8131db4d
this happens at least for NFSv4 (including all minor versions).
.PP
Some versions of UNIX send a signal
.RB ( SIGLOST )
in this circumstance.
Linux does not define this signal, and does not provide any
asynchronous notification of lost locks.
.\"
.SS Managing signals
.BR F_GETOWN ,
.BR F_SETOWN ,
.BR F_GETOWN_EX ,
.BR F_SETOWN_EX ,
.BR F_GETSIG ,
and
.B F_SETSIG
are used to manage I/O availability signals:
.TP
.BR F_GETOWN " (\fIvoid\fP)"
Return (as the function result)
the process ID or process group ID currently receiving
.B SIGIO
and
.B SIGURG
signals for events on file descriptor
.IR fd .
Process IDs are returned as positive values;
process group IDs are returned as negative values (but see BUGS below).
.I arg
is ignored.
.TP
.BR F_SETOWN " (\fIint\fP)"
Set the process ID or process group ID that will receive
.B SIGIO
and
.B SIGURG
signals for events on the file descriptor
.IR fd .
The target process or process group ID is specified in
.IR arg .
A process ID is specified as a positive value;
a process group ID is specified as a negative value.
Most commonly, the calling process specifies itself as the owner
(that is,
.I arg
is specified as
.BR getpid (2)).
.IP
As well as setting the file descriptor owner,
one must also enable generation of signals on the file descriptor.
This is done by using the
.BR fcntl ()
.B F_SETFL
command to set the
.B O_ASYNC
file status flag on the file descriptor.
Subsequently, a
.B SIGIO
signal is sent whenever input or output becomes possible
on the file descriptor.
The
.BR fcntl ()
.B F_SETSIG
command can be used to obtain delivery of a signal other than
.BR SIGIO .
.IP
Sending a signal to the owner process (group) specified by
.B F_SETOWN
is subject to the same permissions checks as are described for
.BR kill (2),
where the sending process is the one that employs
.B F_SETOWN
(but see BUGS below).
If this permission check fails, then the signal is
silently discarded.
.IR Note :
The
.B F_SETOWN
operation records the caller's credentials at the time of the
.BR fcntl ()
call,
and it is these saved credentials that are used for the permission checks.
.IP
If the file descriptor
.I fd
refers to a socket,
.B F_SETOWN
also selects
the recipient of
.B SIGURG
signals that are delivered when out-of-band
data arrives on that socket.
.RB ( SIGURG
is sent in any situation where
.BR select (2)
would report the socket as having an "exceptional condition".)
.\" The following appears to be rubbish.  It doesn't seem to
.\" be true according to the kernel source, and I can write
.\" a program that gets a terminal-generated SIGIO even though
.\" it is not the foreground process group of the terminal.
.\" -- MTK, 8 Apr 05
.\"
.\" If the file descriptor
.\" .I fd
.\" refers to a terminal device, then SIGIO
.\" signals are sent to the foreground process group of the terminal.
.IP
The following was true in Linux 2.6.x up to and including Linux 2.6.11:
.RS
.IP
If a nonzero value is given to
.B F_SETSIG
in a multithreaded process running with a threading library
that supports thread groups (e.g., NPTL),
then a positive value given to
.B F_SETOWN
has a different meaning:
.\" The relevant place in the (2.6) kernel source is the
.\" 'switch' in fs/fcntl.c::send_sigio_to_task() -- MTK, Apr 2005
instead of being a process ID identifying a whole process,
it is a thread ID identifying a specific thread within a process.
Consequently, it may be necessary to pass
.B F_SETOWN
the result of
.BR gettid (2)
instead of
.BR getpid (2)
to get sensible results when
.B F_SETSIG
is used.
(In current Linux threading implementations,
a main thread's thread ID is the same as its process ID.
This means that a single-threaded program can equally use
.BR gettid (2)
or
.BR getpid (2)
in this scenario.)
Note, however, that the statements in this paragraph do not apply
to the
.B SIGURG
signal generated for out-of-band data on a socket:
this signal is always sent to either a process or a process group,
depending on the value given to
.BR F_SETOWN .
.\" send_sigurg()/send_sigurg_to_task() bypasses
.\" kill_fasync()/send_sigio()/send_sigio_to_task()
.\" to directly call send_group_sig_info()
.\"	-- MTK, Apr 2005 (kernel 2.6.11)
.RE
.IP
The above behavior was accidentally dropped in Linux 2.6.12,
and won't be restored.
From Linux 2.6.32 onward, use
.B F_SETOWN_EX
to target
.B SIGIO
and
.B SIGURG
signals at a particular thread.
.TP
.BR F_GETOWN_EX " (\fIstruct f_owner_ex *\fP) (since Linux 2.6.32)"
Return the current file descriptor owner settings
as defined by a previous
.B F_SETOWN_EX
operation.
The information is returned in the structure pointed to by
.IR arg ,
which has the following form:
.IP
.in +4n
.EX
struct f_owner_ex {
    int   type;
    pid_t pid;
};
.EE
.in
.IP
The
.I type
field will have one of the values
.BR F_OWNER_TID ,
.BR F_OWNER_PID ,
or
.BR F_OWNER_PGRP .
The
.I pid
field is a positive integer representing a thread ID, process ID,
or process group ID.
See
.B F_SETOWN_EX
for more details.
.TP
.BR F_SETOWN_EX " (\fIstruct f_owner_ex *\fP) (since Linux 2.6.32)"
This operation performs a similar task to
.BR F_SETOWN .
It allows the caller to direct I/O availability signals
to a specific thread, process, or process group.
The caller specifies the target of signals via
.IR arg ,
which is a pointer to a
.I f_owner_ex
structure.
The
.I type
field has one of the following values, which define how
.I pid
is interpreted:
.RS
.TP
.B F_OWNER_TID
Send the signal to the thread whose thread ID
(the value returned by a call to
.BR clone (2)
or
.BR gettid (2))
is specified in
.IR pid .
.TP
.B F_OWNER_PID
Send the signal to the process whose ID
is specified in
.IR pid .
.TP
.B F_OWNER_PGRP
Send the signal to the process group whose ID
is specified in
.IR pid .
(Note that, unlike with
.BR F_SETOWN ,
a process group ID is specified as a positive value here.)
.RE
.TP
.BR F_GETSIG " (\fIvoid\fP)"
Return (as the function result)
the signal sent when input or output becomes possible.
A value of zero means
.B SIGIO
is sent.
Any other value (including
.BR SIGIO )
is the
signal sent instead, and in this case additional info is available to
the signal handler if installed with
.BR SA_SIGINFO .
.I arg
is ignored.
.TP
.BR F_SETSIG " (\fIint\fP)"
Set the signal sent when input or output becomes possible
to the value given in
.IR arg .
A value of zero means to send the default
.B SIGIO
signal.
Any other value (including
.BR SIGIO )
is the signal to send instead, and in this case additional info
is available to the signal handler if installed with
.BR SA_SIGINFO .
.\"
.\" The following was true only up until Linux 2.6.11:
.\"
.\" Additionally, passing a nonzero value to
.\" .B F_SETSIG
.\" changes the signal recipient from a whole process to a specific thread
.\" within a process.
.\" See the description of
.\" .B F_SETOWN
.\" for more details.
.IP
By using
.B F_SETSIG
with a nonzero value, and setting
.B SA_SIGINFO
for the
signal handler (see
.BR sigaction (2)),
extra information about I/O events is passed to
the handler in a
.I siginfo_t
structure.
If the
.I si_code
field indicates the source is
.BR SI_SIGIO ,
the
.I si_fd
field gives the file descriptor associated with the event.
Otherwise,
there is no indication which file descriptors are pending, and you
should use the usual mechanisms
.RB ( select (2),
.BR poll (2),
.BR read (2)
with
.B O_NONBLOCK
set etc.) to determine which file descriptors are available for I/O.
.IP
Note that the file descriptor provided in
.I si_fd
is the one that was specified during the
.B F_SETSIG
operation.
This can lead to an unusual corner case.
If the file descriptor is duplicated
.RB ( dup (2)
or similar), and the original file descriptor is closed,
then I/O events will continue to be generated, but the
.I si_fd
field will contain the number of the now closed file descriptor.
.IP
By selecting a real time signal (value >=
.BR SIGRTMIN ),
multiple I/O events may be queued using the same signal numbers.
(Queuing is dependent on available memory.)
Extra information is available
if
.B SA_SIGINFO
is set for the signal handler, as above.
.IP
Note that Linux imposes a limit on the
number of real-time signals that may be queued to a
process (see
.BR getrlimit (2)
and
.BR signal (7))
and if this limit is reached, then the kernel reverts to
delivering
.BR SIGIO ,
and this signal is delivered to the entire
process rather than to a specific thread.
.\" See fs/fcntl.c::send_sigio_to_task() (2.4/2.6) sources -- MTK, Apr 05
.PP
Using these mechanisms, a program can implement fully asynchronous I/O
without using
.BR select (2)
or
.BR poll (2)
most of the time.
.PP
The use of
.B O_ASYNC
is specific to BSD and Linux.
The only use of
.B F_GETOWN
and
.B F_SETOWN
specified in POSIX.1 is in conjunction with the use of the
.B SIGURG
signal on sockets.
(POSIX does not specify the
.B SIGIO
signal.)
.BR F_GETOWN_EX ,
.BR F_SETOWN_EX ,
.BR F_GETSIG ,
and
.B F_SETSIG
are Linux-specific.
POSIX has asynchronous I/O and the
.I aio_sigevent
structure to achieve similar things; these are also available
in Linux as part of the GNU C Library (glibc).
.SS Leases
.B F_SETLEASE
and
.B F_GETLEASE
(Linux 2.4 onward) are used to establish a new lease,
and retrieve the current lease, on the open file description
referred to by the file descriptor
.IR fd .
A file lease provides a mechanism whereby the process holding
the lease (the "lease holder") is notified (via delivery of a signal)
when a process (the "lease breaker") tries to
.BR open (2)
or
.BR truncate (2)
the file referred to by that file descriptor.
.TP
.BR F_SETLEASE " (\fIint\fP)"
Set or remove a file lease according to which of the following
values is specified in the integer
.IR arg :
.RS
.TP
.B F_RDLCK
Take out a read lease.
This will cause the calling process to be notified when
the file is opened for writing or is truncated.
.\" The following became true in Linux 2.6.10:
.\" See the man-pages-2.09 Changelog for further info.
A read lease can be placed only on a file descriptor that
is opened read-only.
.TP
.B F_WRLCK
Take out a write lease.
This will cause the caller to be notified when
the file is opened for reading or writing or is truncated.
A write lease may be placed on a file only if there are no
other open file descriptors for the file.
.TP
.B F_UNLCK
Remove our lease from the file.
.RE
.PP
Leases are associated with an open file description (see
.BR open (2)).
This means that duplicate file descriptors (created by, for example,
.BR fork (2)
or
.BR dup (2))
refer to the same lease, and this lease may be modified
or released using any of these descriptors.
Furthermore, the lease is released by either an explicit
.B F_UNLCK
operation on any of these duplicate file descriptors, or when all
such file descriptors have been closed.
.PP
Leases may be taken out only on regular files.
An unprivileged process may take out a lease only on a file whose
UID (owner) matches the filesystem UID of the process.
A process with the
.B CAP_LEASE
capability may take out leases on arbitrary files.
.TP
.BR F_GETLEASE " (\fIvoid\fP)"
Indicates what type of lease is associated with the file descriptor
.I fd
by returning either
.BR F_RDLCK ", " F_WRLCK ", or " F_UNLCK ,
indicating, respectively, a read lease , a write lease, or no lease.
.I arg
is ignored.
.PP
When a process (the "lease breaker") performs an
.BR open (2)
or
.BR truncate (2)
that conflicts with a lease established via
.BR F_SETLEASE ,
the system call is blocked by the kernel and
the kernel notifies the lease holder by sending it a signal
.RB ( SIGIO
by default).
The lease holder should respond to receipt of this signal by doing
whatever cleanup is required in preparation for the file to be
accessed by another process (e.g., flushing cached buffers) and
then either remove or downgrade its lease.
A lease is removed by performing an
.B F_SETLEASE
command specifying
.I arg
as
.BR F_UNLCK .
If the lease holder currently holds a write lease on the file,
and the lease breaker is opening the file for reading,
then it is sufficient for the lease holder to downgrade
the lease to a read lease.
This is done by performing an
.B F_SETLEASE
command specifying
.I arg
as
.BR F_RDLCK .
.PP
If the lease holder fails to downgrade or remove the lease within
the number of seconds specified in
.IR /proc/sys/fs/lease\-break\-time ,
then the kernel forcibly removes or downgrades the lease holder's lease.
.PP
Once a lease break has been initiated,
.B F_GETLEASE
returns the target lease type (either
.B F_RDLCK
or
.BR F_UNLCK ,
depending on what would be compatible with the lease breaker)
until the lease holder voluntarily downgrades or removes the lease or
the kernel forcibly does so after the lease break timer expires.
.PP
Once the lease has been voluntarily or forcibly removed or downgraded,
and assuming the lease breaker has not unblocked its system call,
the kernel permits the lease breaker's system call to proceed.
.PP
If the lease breaker's blocked
.BR open (2)
or
.BR truncate (2)
is interrupted by a signal handler,
then the system call fails with the error
.BR EINTR ,
but the other steps still occur as described above.
If the lease breaker is killed by a signal while blocked in
.BR open (2)
or
.BR truncate (2),
then the other steps still occur as described above.
If the lease breaker specifies the
.B O_NONBLOCK
flag when calling
.BR open (2),
then the call immediately fails with the error
.BR EWOULDBLOCK ,
but the other steps still occur as described above.
.PP
The default signal used to notify the lease holder is
.BR SIGIO ,
but this can be changed using the
.B F_SETSIG
command to
.BR fcntl ().
If a
.B F_SETSIG
command is performed (even one specifying
.BR SIGIO ),
and the signal
handler is established using
.BR SA_SIGINFO ,
then the handler will receive a
.I siginfo_t
structure as its second argument, and the
.I si_fd
field of this argument will hold the file descriptor of the leased file
that has been accessed by another process.
(This is useful if the caller holds leases against multiple files.)
.SS File and directory change notification (dnotify)
.TP
.BR F_NOTIFY " (\fIint\fP)"
(Linux 2.4 onward)
Provide notification when the directory referred to by
.I fd
or any of the files that it contains is changed.
The events to be notified are specified in
.IR arg ,
which is a bit mask specified by ORing together zero or more of
the following bits:
.PP
.RS
.PD 0
.TP
.B DN_ACCESS
A file was accessed
.RB ( read (2),
.BR pread (2),
.BR readv (2),
and similar)
.TP
.B DN_MODIFY
A file was modified
.RB ( write (2),
.BR pwrite (2),
.BR writev (2),
.BR truncate (2),
.BR ftruncate (2),
and similar).
.TP
.B DN_CREATE
A file was created
.RB ( open (2),
.BR creat (2),
.BR mknod (2),
.BR mkdir (2),
.BR link (2),
.BR symlink (2),
.BR rename (2)
into this directory).
.TP
.B DN_DELETE
A file was unlinked
.RB ( unlink (2),
.BR rename (2)
to another directory,
.BR rmdir (2)).
.TP
.B DN_RENAME
A file was renamed within this directory
.RB ( rename (2)).
.TP
.B DN_ATTRIB
The attributes of a file were changed
.RB ( chown (2),
.BR chmod (2),
.BR utime (2),
.BR utimensat (2),
and similar).
.PD
.RE
.IP
(In order to obtain these definitions, the
.B _GNU_SOURCE
feature test macro must be defined before including
.I any
header files.)
.IP
Directory notifications are normally "one-shot", and the application
must reregister to receive further notifications.
Alternatively, if
.B DN_MULTISHOT
is included in
.IR arg ,
then notification will remain in effect until explicitly removed.
.IP
.\" The following does seem a poor API-design choice...
A series of
.B F_NOTIFY
requests is cumulative, with the events in
.I arg
being added to the set already monitored.
To disable notification of all events, make an
.B F_NOTIFY
call specifying
.I arg
as 0.
.IP
Notification occurs via delivery of a signal.
The default signal is
.BR SIGIO ,
but this can be changed using the
.B F_SETSIG
command to
.BR fcntl ().
(Note that
.B SIGIO
is one of the nonqueuing standard signals;
switching to the use of a real-time signal means that
multiple notifications can be queued to the process.)
In the latter case, the signal handler receives a
.I siginfo_t
structure as its second argument (if the handler was
established using
.BR SA_SIGINFO )
and the
.I si_fd
field of this structure contains the file descriptor which
generated the notification (useful when establishing notification
on multiple directories).
.IP
Especially when using
.BR DN_MULTISHOT ,
a real time signal should be used for notification,
so that multiple notifications can be queued.
.IP
.B NOTE:
New applications should use the
.I inotify
interface (available since Linux 2.6.13),
which provides a much superior interface for obtaining notifications of
filesystem events.
See
.BR inotify (7).
.SS Changing the capacity of a pipe
.TP
.BR F_SETPIPE_SZ " (\fIint\fP; since Linux 2.6.35)"
Change the capacity of the pipe referred to by
.I fd
to be at least
.I arg
bytes.
An unprivileged process can adjust the pipe capacity to any value
between the system page size and the limit defined in
.I /proc/sys/fs/pipe\-max\-size
(see
.BR proc (5)).
Attempts to set the pipe capacity below the page size are silently
rounded up to the page size.
Attempts by an unprivileged process to set the pipe capacity above the limit in
.I /proc/sys/fs/pipe\-max\-size
yield the error
.BR EPERM ;
a privileged process
.RB ( CAP_SYS_RESOURCE )
can override the limit.
.IP
When allocating the buffer for the pipe,
the kernel may use a capacity larger than
.IR arg ,
if that is convenient for the implementation.
(In the current implementation,
the allocation is the next higher power-of-two page-size multiple
of the requested size.)
The actual capacity (in bytes) that is set is returned as the function result.
.IP
Attempting to set the pipe capacity smaller than the amount
of buffer space currently used to store data produces the error
.BR EBUSY .
.IP
Note that because of the way the pages of the pipe buffer
are employed when data is written to the pipe,
the number of bytes that can be written may be less than the nominal size,
depending on the size of the writes.
.TP
.BR F_GETPIPE_SZ " (\fIvoid\fP; since Linux 2.6.35)"
Return (as the function result) the capacity of the pipe referred to by
.IR fd .
.\"
.SS File Sealing
File seals limit the set of allowed operations on a given file.
For each seal that is set on a file,
a specific set of operations will fail with
.B EPERM
on this file from now on.
The file is said to be sealed.
The default set of seals depends on the type of the underlying
file and filesystem.
For an overview of file sealing, a discussion of its purpose,
and some code examples, see
.BR memfd_create (2).
.PP
Currently,
file seals can be applied only to a file descriptor returned by
.BR memfd_create (2)
(if the
.B MFD_ALLOW_SEALING
was employed).
On other filesystems, all
.BR fcntl ()
operations that operate on seals will return
.BR EINVAL .
.PP
Seals are a property of an inode.
Thus, all open file descriptors referring to the same inode share
the same set of seals.
Furthermore, seals can never be removed, only added.
.TP
.BR F_ADD_SEALS " (\fIint\fP; since Linux 3.17)"
Add the seals given in the bit-mask argument
.I arg
to the set of seals of the inode referred to by the file descriptor
.IR fd .
Seals cannot be removed again.
Once this call succeeds, the seals are enforced by the kernel immediately.
If the current set of seals includes
.B F_SEAL_SEAL
(see below), then this call will be rejected with
.BR EPERM .
Adding a seal that is already set is a no-op, in case
.B F_SEAL_SEAL
is not set already.
In order to place a seal, the file descriptor
.I fd
must be writable.
.TP
.BR F_GET_SEALS " (\fIvoid\fP; since Linux 3.17)"
Return (as the function result) the current set of seals
of the inode referred to by
.IR fd .
If no seals are set, 0 is returned.
If the file does not support sealing, \-1 is returned and
.I errno
is set to
.BR EINVAL .
.PP
The following seals are available:
.TP
.B F_SEAL_SEAL
If this seal is set, any further call to
.BR fcntl ()
with
.B F_ADD_SEALS
fails with the error
.BR EPERM .
Therefore, this seal prevents any modifications to the set of seals itself.
If the initial set of seals of a file includes
.BR F_SEAL_SEAL ,
then this effectively causes the set of seals to be constant and locked.
.TP
.B F_SEAL_SHRINK
If this seal is set, the file in question cannot be reduced in size.
This affects
.BR open (2)
with the
.B O_TRUNC
flag as well as
.BR truncate (2)
and
.BR ftruncate (2).
Those calls fail with
.B EPERM
if you try to shrink the file in question.
Increasing the file size is still possible.
.TP
.B F_SEAL_GROW
If this seal is set, the size of the file in question cannot be increased.
This affects
.BR write (2)
beyond the end of the file,
.BR truncate (2),
.BR ftruncate (2),
and
.BR fallocate (2).
These calls fail with
.B EPERM
if you use them to increase the file size.
If you keep the size or shrink it, those calls still work as expected.
.TP
.B F_SEAL_WRITE
If this seal is set, you cannot modify the contents of the file.
Note that shrinking or growing the size of the file is
still possible and allowed.
.\" One or more other seals are typically used with F_SEAL_WRITE
.\" because, given a file with the F_SEAL_WRITE seal set, then,
.\" while it would no longer be possible to (say) write zeros into
.\" the last 100 bytes of a file, it would still be possible
.\" to (say) shrink the file by 100 bytes using ftruncate(), and
.\" then increase the file size by 100 bytes, which would have
.\" the effect of replacing the last hundred bytes by zeros.
.\"
Thus, this seal is normally used in combination with one of the other seals.
This seal affects
.BR write (2)
and
.BR fallocate (2)
(only in combination with the
.B FALLOC_FL_PUNCH_HOLE
flag).
Those calls fail with
.B EPERM
if this seal is set.
Furthermore, trying to create new shared, writable memory-mappings via
.BR mmap (2)
will also fail with
.BR EPERM .
.IP
Using the
.B F_ADD_SEALS
operation to set the
.B F_SEAL_WRITE
seal fails with
.B EBUSY
if any writable, shared mapping exists.
Such mappings must be unmapped before you can add this seal.
Furthermore, if there are any asynchronous I/O operations
.RB ( io_submit (2))
pending on the file,
all outstanding writes will be discarded.
.TP
.BR F_SEAL_FUTURE_WRITE " (since Linux 5.1)"
The effect of this seal is similar to
.BR F_SEAL_WRITE ,
but the contents of the file can still be modified via
shared writable mappings that were created prior to the seal being set.
Any attempt to create a new writable mapping on the file via
.BR mmap (2)
will fail with
.BR EPERM .
Likewise, an attempt to write to the file via
.BR write (2)
will fail with
.BR EPERM .
.IP
Using this seal,
one process can create a memory buffer that it can continue to modify
while sharing that buffer on a "read-only" basis with other processes.
.\"
.SS File read/write hints
Write lifetime hints can be used to inform the kernel about the relative
expected lifetime of writes on a given inode or
via a particular open file description.
(See
.BR open (2)
for an explanation of open file descriptions.)
In this context, the term "write lifetime" means
the expected time the data will live on media, before
being overwritten or erased.
.PP
An application may use the different hint values specified below to
separate writes into different write classes,
so that multiple users or applications running on a single storage back-end
can aggregate their I/O patterns in a consistent manner.
However, there are no functional semantics implied by these flags,
and different I/O classes can use the write lifetime hints
in arbitrary ways, so long as the hints are used consistently.
.PP
The following operations can be applied to the file descriptor,
.IR fd :
.TP
.BR F_GET_RW_HINT " (\fIuint64_t *\fP; since Linux 4.13)"
Returns the value of the read/write hint associated with the underlying inode
referred to by
.IR fd .
.TP
.BR F_SET_RW_HINT " (\fIuint64_t *\fP; since Linux 4.13)"
Sets the read/write hint value associated with the
underlying inode referred to by
.IR fd .
This hint persists until either it is explicitly modified or
the underlying filesystem is unmounted.
.TP
.BR F_GET_FILE_RW_HINT " (\fIuint64_t *\fP; since Linux 4.13)"
Returns the value of the read/write hint associated with
the open file description referred to by
.IR fd .
.TP
.BR F_SET_FILE_RW_HINT " (\fIuint64_t *\fP; since Linux 4.13)"
Sets the read/write hint value associated with the open file description
referred to by
.IR fd .
.PP
If an open file description has not been assigned a read/write hint,
then it shall use the value assigned to the inode, if any.
.PP
The following read/write
hints are valid since Linux 4.13:
.TP
.B RWH_WRITE_LIFE_NOT_SET
No specific hint has been set.
This is the default value.
.TP
.B RWH_WRITE_LIFE_NONE
No specific write lifetime is associated with this file or inode.
.TP
.B RWH_WRITE_LIFE_SHORT
Data written to this inode or via this open file description
is expected to have a short lifetime.
.TP
.B RWH_WRITE_LIFE_MEDIUM
Data written to this inode or via this open file description
is expected to have a lifetime longer than
data written with
.BR RWH_WRITE_LIFE_SHORT .
.TP
.B RWH_WRITE_LIFE_LONG
Data written to this inode or via this open file description
is expected to have a lifetime longer than
data written with
.BR RWH_WRITE_LIFE_MEDIUM .
.TP
.B RWH_WRITE_LIFE_EXTREME
Data written to this inode or via this open file description
is expected to have a lifetime longer than
data written with
.BR RWH_WRITE_LIFE_LONG .
.PP
All the write-specific hints are relative to each other,
and no individual absolute meaning should be attributed to them.
.SH RETURN VALUE
For a successful call, the return value depends on the operation:
.TP
.B F_DUPFD
The new file descriptor.
.TP
.B F_GETFD
Value of file descriptor flags.
.TP
.B F_GETFL
Value of file status flags.
.TP
.B F_GETLEASE
Type of lease held on file descriptor.
.TP
.B F_GETOWN
Value of file descriptor owner.
.TP
.B F_GETSIG
Value of signal sent when read or write becomes possible, or zero
for traditional
.B SIGIO
behavior.
.TP
.BR F_GETPIPE_SZ ", " F_SETPIPE_SZ
The pipe capacity.
.TP
.B F_GET_SEALS
A bit mask identifying the seals that have been set
for the inode referred to by
.IR fd .
.TP
All other commands
Zero.
.PP
On error, \-1 is returned, and
.I errno
is set to indicate the error.
.SH ERRORS
.TP
.BR EACCES " or " EAGAIN
Operation is prohibited by locks held by other processes.
.TP
.B EAGAIN
The operation is prohibited because the file has been memory-mapped by
another process.
.TP
.B EBADF
.I fd
is not an open file descriptor
.TP
.B EBADF
.I cmd
is
.B F_SETLK
or
.B F_SETLKW
and the file descriptor open mode doesn't match with the
type of lock requested.
.TP
.B EBUSY
.I cmd
is
.B F_SETPIPE_SZ
and the new pipe capacity specified in
.I arg
is smaller than the amount of buffer space currently
used to store data in the pipe.
.TP
.B EBUSY
.I cmd
is
.BR F_ADD_SEALS ,
.I arg
includes
.BR F_SEAL_WRITE ,
and there exists a writable, shared mapping on the file referred to by
.IR fd .
.TP
.B EDEADLK
It was detected that the specified
.B F_SETLKW
command would cause a deadlock.
.TP
.B EFAULT
.I lock
is outside your accessible address space.
.TP
.B EINTR
.I cmd
is
.B F_SETLKW
or
.B F_OFD_SETLKW
and the operation was interrupted by a signal; see
.BR signal (7).
.TP
.B EINTR
.I cmd
is
.BR F_GETLK ,
.BR F_SETLK ,
.BR F_OFD_GETLK ,
or
.BR F_OFD_SETLK ,
and the operation was interrupted by a signal before the lock was checked or
acquired.
Most likely when locking a remote file (e.g., locking over
NFS), but can sometimes happen locally.
.TP
.B EINVAL
The value specified in
.I cmd
is not recognized by this kernel.
.TP
.B EINVAL
.I cmd
is
.B F_ADD_SEALS
and
.I arg
includes an unrecognized sealing bit.
.TP
.B EINVAL
.I cmd
is
.B F_ADD_SEALS
or
.B F_GET_SEALS
and the filesystem containing the inode referred to by
.I fd
does not support sealing.
.TP
.B EINVAL
.I cmd
is
.B F_DUPFD
and
.I arg
is negative or is greater than the maximum allowable value
(see the discussion of
.B RLIMIT_NOFILE
in
.BR getrlimit (2)).
.TP
.B EINVAL
.I cmd
is
.B F_SETSIG
and
.I arg
is not an allowable signal number.
.TP
.B EINVAL
.I cmd
is
.BR F_OFD_SETLK ,
.BR F_OFD_SETLKW ,
or
.BR F_OFD_GETLK ,
and
.I l_pid
was not specified as zero.
.TP
.B EMFILE
.I cmd
is
.B F_DUPFD
and the per-process limit on the number of open file descriptors
has been reached.
.TP
.B ENOLCK
Too many segment locks open, lock table is full, or a remote locking
protocol failed (e.g., locking over NFS).
.TP
.B ENOTDIR
.B F_NOTIFY
was specified in
.IR cmd ,
but
.I fd
does not refer to a directory.
.TP
.B EPERM
.I cmd
is
.B F_SETPIPE_SZ
and the soft or hard user pipe limit has been reached; see
.BR pipe (7).
.TP
.B EPERM
Attempted to clear the
.B O_APPEND
flag on a file that has the append-only attribute set.
.TP
.B EPERM
.I cmd
was
.BR F_ADD_SEALS ,
but
.I fd
was not open for writing
or the current set of seals on the file already includes
.BR F_SEAL_SEAL .
.SH STANDARDS
POSIX.1-2008.
.PP
.BR F_GETOWN_EX ,
.BR F_SETOWN_EX ,
.BR F_SETPIPE_SZ ,
.BR F_GETPIPE_SZ ,
.BR F_GETSIG ,
.BR F_SETSIG ,
.BR F_NOTIFY ,
.BR F_GETLEASE ,
and
.B F_SETLEASE
are Linux-specific.
(Define the
.B _GNU_SOURCE
macro to obtain these definitions.)
.\" .PP
.\" SVr4 documents additional EIO, ENOLINK and EOVERFLOW error conditions.
.PP
.BR F_OFD_SETLK ,
.BR F_OFD_SETLKW ,
and
.B F_OFD_GETLK
are Linux-specific (and one must define
.B _GNU_SOURCE
to obtain their definitions),
but work is being done to have them included in the next version of POSIX.1.
.PP
.B F_ADD_SEALS
and
.B F_GET_SEALS
are Linux-specific.
.\" FIXME . Once glibc adds support, add a note about FTM requirements
.SH HISTORY
SVr4, 4.3BSD, POSIX.1-2001.
.PP
Only the operations
.BR F_DUPFD ,
.BR F_GETFD ,
.BR F_SETFD ,
.BR F_GETFL ,
.BR F_SETFL ,
.BR F_GETLK ,
.BR F_SETLK ,
and
.B F_SETLKW
are specified in POSIX.1-2001.
.PP
.B F_GETOWN
and
.B F_SETOWN
are specified in POSIX.1-2001.
(To get their definitions, define either
.\" .BR _BSD_SOURCE ,
.\" or
.B _XOPEN_SOURCE
with the value 500 or greater, or
.B _POSIX_C_SOURCE
with the value 200809L or greater.)
.PP
.B F_DUPFD_CLOEXEC
is specified in POSIX.1-2008.
(To get this definition, define
.B _POSIX_C_SOURCE
with the value 200809L or greater, or
.B _XOPEN_SOURCE
with the value 700 or greater.)
.SH NOTES
The errors returned by
.BR dup2 (2)
are different from those returned by
.BR F_DUPFD .
.\"
.SS File locking
The original Linux
.BR fcntl ()
system call was not designed to handle large file offsets
(in the
.I flock
structure).
Consequently, an
.BR fcntl64 ()
system call was added in Linux 2.4.
The newer system call employs a different structure for file locking,
.IR flock64 ,
and corresponding commands,
.BR F_GETLK64 ,
.BR F_SETLK64 ,
and
.BR F_SETLKW64 .
However, these details can be ignored by applications using glibc, whose
.BR fcntl ()
wrapper function transparently employs the more recent system call
where it is available.
.\"
.SS Record locks
Since Linux 2.0, there is no interaction between the types of lock
placed by
.BR flock (2)
and
.BR fcntl ().
.PP
Several systems have more fields in
.I "struct flock"
such as, for example,
.I l_sysid
(to identify the machine where the lock is held).
.\" e.g., Solaris 8 documents this field in fcntl(2), and Irix 6.5
.\" documents it in fcntl(5).  mtk, May 2007
.\" Also, FreeBSD documents it (Apr 2014).
Clearly,
.I l_pid
alone is not going to be very useful if the process holding the lock
may live on a different machine;
on Linux, while present on some architectures (such as MIPS32),
this field is not used.
.PP
The original Linux
.BR fcntl ()
system call was not designed to handle large file offsets
(in the
.I flock
structure).
Consequently, an
.BR fcntl64 ()
system call was added in Linux 2.4.
The newer system call employs a different structure for file locking,
.IR flock64 ,
and corresponding commands,
.BR F_GETLK64 ,
.BR F_SETLK64 ,
and
.BR F_SETLKW64 .
However, these details can be ignored by applications using glibc, whose
.BR fcntl ()
wrapper function transparently employs the more recent system call
where it is available.
.SS Record locking and NFS
Before Linux 3.12, if an NFSv4 client
loses contact with the server for a period of time
(defined as more than 90 seconds with no communication),
.\"
.\" Neil Brown: With NFSv3 the failure mode is the reverse.  If
.\"     the server loses contact with a client then any lock stays in place
.\"     indefinitely ("why can't I read my mail"... I remember it well).
.\"
it might lose and regain a lock without ever being aware of the fact.
(The period of time after which contact is assumed lost is known as
the NFSv4 leasetime.
On a Linux NFS server, this can be determined by looking at
.IR /proc/fs/nfsd/nfsv4leasetime ,
which expresses the period in seconds.
The default value for this file is 90.)
.\"
.\" Jeff Layton:
.\"     Note that this is not a firm timeout. The server runs a job
.\"     periodically to clean out expired stateful objects, and it's likely
.\"     that there is some time (maybe even up to another whole lease period)
.\"     between when the timeout expires and the job actually runs. If the
.\"     client gets a RENEW in there within that window, its lease will be
.\"     renewed and its state preserved.
.\"
This scenario potentially risks data corruption,
since another process might acquire a lock in the intervening period
and perform file I/O.
.PP
Since Linux 3.12,
.\" commit ef1820f9be27b6ad158f433ab38002ab8131db4d
if an NFSv4 client loses contact with the server,
any I/O to the file by a process which "thinks" it holds
a lock will fail until that process closes and reopens the file.
A kernel parameter,
.IR nfs.recover_lost_locks ,
can be set to 1 to obtain the pre-3.12 behavior,
whereby the client will attempt to recover lost locks
when contact is reestablished with the server.
Because of the attendant risk of data corruption,
.\" commit f6de7a39c181dfb8a2c534661a53c73afb3081cd
this parameter defaults to 0 (disabled).
.SH BUGS
.SS F_SETFL
It is not possible to use
.B F_SETFL
to change the state of the
.B O_DSYNC
and
.B O_SYNC
flags.
.\" FIXME . According to POSIX.1-2001, O_SYNC should also be modifiable
.\" via fcntl(2), but currently Linux does not permit this
.\" See http://bugzilla.kernel.org/show_bug.cgi?id=5994
Attempts to change the state of these flags are silently ignored.
.SS F_GETOWN
A limitation of the Linux system call conventions on some
architectures (notably i386) means that if a (negative)
process group ID to be returned by
.B F_GETOWN
falls in the range \-1 to \-4095, then the return value is wrongly
interpreted by glibc as an error in the system call;
.\" glibc source: sysdeps/unix/sysv/linux/i386/sysdep.h
that is, the return value of
.BR fcntl ()
will be \-1, and
.I errno
will contain the (positive) process group ID.
The Linux-specific
.B F_GETOWN_EX
operation avoids this problem.
.\" mtk, Dec 04: some limited testing on alpha and ia64 seems to
.\" indicate that ANY negative PGID value will cause F_GETOWN
.\" to misinterpret the return as an error. Some other architectures
.\" seem to have the same range check as i386.
Since glibc 2.11, glibc makes the kernel
.B F_GETOWN
problem invisible by implementing
.B F_GETOWN
using
.BR F_GETOWN_EX .
.SS F_SETOWN
In Linux 2.4 and earlier, there is bug that can occur
when an unprivileged process uses
.B F_SETOWN
to specify the owner
of a socket file descriptor
as a process (group) other than the caller.
In this case,
.BR fcntl ()
can return \-1 with
.I errno
set to
.BR EPERM ,
even when the owner process (group) is one that the caller
has permission to send signals to.
Despite this error return, the file descriptor owner is set,
and signals will be sent to the owner.
.\"
.SS Deadlock detection
The deadlock-detection algorithm employed by the kernel when dealing with
.B F_SETLKW
requests can yield both
false negatives (failures to detect deadlocks,
leaving a set of deadlocked processes blocked indefinitely)
and false positives
.RB ( EDEADLK
errors when there is no deadlock).
For example,
the kernel limits the lock depth of its dependency search to 10 steps,
meaning that circular deadlock chains that exceed
that size will not be detected.
In addition, the kernel may falsely indicate a deadlock
when two or more processes created using the
.BR clone (2)
.B CLONE_FILES
flag place locks that appear (to the kernel) to conflict.
.\"
.SS Mandatory locking
The Linux implementation of mandatory locking
is subject to race conditions which render it unreliable:
.\" http://marc.info/?l=linux-kernel&m=119013491707153&w=2
.\"
.\" Reconfirmed by Jeff Layton
.\"     From: Jeff Layton <jlayton <at> redhat.com>
.\"     Subject: Re: Status of fcntl() mandatory locking
.\"     Newsgroups: gmane.linux.file-systems
.\"     Date: 2014-04-28 10:07:57 GMT
.\"     http://thread.gmane.org/gmane.linux.file-systems/84481/focus=84518
a
.BR write (2)
call that overlaps with a lock may modify data after the mandatory lock is
acquired;
a
.BR read (2)
call that overlaps with a lock may detect changes to data that were made
only after a write lock was acquired.
Similar races exist between mandatory locks and
.BR mmap (2).
It is therefore inadvisable to rely on mandatory locking.
.SH SEE ALSO
.BR dup2 (2),
.BR flock (2),
.BR open (2),
.BR socket (2),
.BR lockf (3),
.BR capabilities (7),
.BR feature_test_macros (7),
.BR lslocks (8)
.PP
.IR locks.txt ,
.IR mandatory\-locking.txt ,
and
.I dnotify.txt
in the Linux kernel source directory
.I Documentation/filesystems/
(on older kernels, these files are directly under the
.I Documentation/
directory, and
.I mandatory\-locking.txt
is called
.IR mandatory.txt )