man-pages/man2/fork.2

328 lines
9.2 KiB
Groff

.\" Copyright (C) 2006 Michael Kerrisk <mtk.manpages@gmail.com>
.\" A few fragments remain from an earlier (1992) page by
.\" Drew Eckhardt (drew@cs.colorado.edu),
.\"
.\" %%%LICENSE_START(VERBATIM)
.\" Permission is granted to make and distribute verbatim copies of this
.\" manual provided the copyright notice and this permission notice are
.\" preserved on all copies.
.\"
.\" Permission is granted to copy and distribute modified versions of this
.\" manual under the conditions for verbatim copying, provided that the
.\" entire resulting derived work is distributed under the terms of a
.\" permission notice identical to this one.
.\"
.\" Since the Linux kernel and libraries are constantly changing, this
.\" manual page may be incorrect or out-of-date. The author(s) assume no
.\" responsibility for errors or omissions, or for damages resulting from
.\" the use of the information contained herein. The author(s) may not
.\" have taken the same level of care in the production of this manual,
.\" which is licensed free of charge, as they might when working
.\" professionally.
.\"
.\" Formatted or processed versions of this manual, if unaccompanied by
.\" the source, must acknowledge the copyright and authors of this work.
.\" %%%LICENSE_END
.\"
.\" Modified by Michael Haardt (michael@moria.de)
.\" Modified Sat Jul 24 13:22:07 1993 by Rik Faith (faith@cs.unc.edu)
.\" Modified 21 Aug 1994 by Michael Chastain (mec@shell.portal.com):
.\" Referenced 'clone(2)'.
.\" Modified 1995-06-10, 1996-04-18, 1999-11-01, 2000-12-24
.\" by Andries Brouwer (aeb@cwi.nl)
.\" Modified, 27 May 2004, Michael Kerrisk <mtk.manpages@gmail.com>
.\" Added notes on capability requirements
.\" 2006-09-04, Michael Kerrisk
.\" Greatly expanded, to describe all attributes that differ
.\" parent and child.
.\"
.TH FORK 2 2021-03-22 "Linux" "Linux Programmer's Manual"
.SH NAME
fork \- create a child process
.SH SYNOPSIS
.nf
.B #include <unistd.h>
.PP
.B pid_t fork(void);
.fi
.SH DESCRIPTION
.BR fork ()
creates a new process by duplicating the calling process.
The new process is referred to as the
.I child
process.
The calling process is referred to as the
.I parent
process.
.PP
The child process and the parent process run in separate memory spaces.
At the time of
.BR fork ()
both memory spaces have the same content.
Memory writes, file mappings
.RB ( mmap (2)),
and unmappings
.RB ( munmap (2))
performed by one of the processes do not affect the other.
.PP
The child process is an exact duplicate of the parent
process except for the following points:
.IP * 3
The child has its own unique process ID,
and this PID does not match the ID of any existing process group
.RB ( setpgid (2))
or session.
.IP *
The child's parent process ID is the same as the parent's process ID.
.IP *
The child does not inherit its parent's memory locks
.RB ( mlock (2),
.BR mlockall (2)).
.IP *
Process resource utilizations
.RB ( getrusage (2))
and CPU time counters
.RB ( times (2))
are reset to zero in the child.
.IP *
The child's set of pending signals is initially empty
.RB ( sigpending (2)).
.IP *
The child does not inherit semaphore adjustments from its parent
.RB ( semop (2)).
.IP *
The child does not inherit process-associated record locks from its parent
.RB ( fcntl (2)).
(On the other hand, it does inherit
.BR fcntl (2)
open file description locks and
.BR flock (2)
locks from its parent.)
.IP *
The child does not inherit timers from its parent
.RB ( setitimer (2),
.BR alarm (2),
.BR timer_create (2)).
.IP *
The child does not inherit outstanding asynchronous I/O operations
from its parent
.RB ( aio_read (3),
.BR aio_write (3)),
nor does it inherit any asynchronous I/O contexts from its parent (see
.BR io_setup (2)).
.PP
The process attributes in the preceding list are all specified
in POSIX.1.
The parent and child also differ with respect to the following
Linux-specific process attributes:
.IP * 3
The child does not inherit directory change notifications (dnotify)
from its parent
(see the description of
.B F_NOTIFY
in
.BR fcntl (2)).
.IP *
The
.BR prctl (2)
.B PR_SET_PDEATHSIG
setting is reset so that the child does not receive a signal
when its parent terminates.
.IP *
The default timer slack value is set to the parent's
current timer slack value.
See the description of
.BR PR_SET_TIMERSLACK
in
.BR prctl (2).
.IP *
Memory mappings that have been marked with the
.BR madvise (2)
.B MADV_DONTFORK
flag are not inherited across a
.BR fork ().
.IP *
Memory in address ranges that have been marked with the
.BR madvise (2)
.B MADV_WIPEONFORK
flag is zeroed in the child after a
.BR fork ().
(The
.B MADV_WIPEONFORK
setting remains in place for those address ranges in the child.)
.IP *
The termination signal of the child is always
.B SIGCHLD
(see
.BR clone (2)).
.IP *
The port access permission bits set by
.BR ioperm (2)
are not inherited by the child;
the child must turn on any bits that it requires using
.BR ioperm (2).
.PP
Note the following further points:
.IP * 3
The child process is created with a single thread\(emthe
one that called
.BR fork ().
The entire virtual address space of the parent is replicated in the child,
including the states of mutexes, condition variables,
and other pthreads objects; the use of
.BR pthread_atfork (3)
may be helpful for dealing with problems that this can cause.
.IP *
After a
.BR fork ()
in a multithreaded program,
the child can safely call only async-signal-safe functions (see
.BR signal\-safety (7))
until such time as it calls
.BR execve (2).
.IP *
The child inherits copies of the parent's set of open file descriptors.
Each file descriptor in the child refers to the same
open file description (see
.BR open (2))
as the corresponding file descriptor in the parent.
This means that the two file descriptors share open file status flags,
file offset,
and signal-driven I/O attributes (see the description of
.B F_SETOWN
and
.B F_SETSIG
in
.BR fcntl (2)).
.IP *
The child inherits copies of the parent's set of open message
queue descriptors (see
.BR mq_overview (7)).
Each file descriptor in the child refers to the same
open message queue description
as the corresponding file descriptor in the parent.
This means that the two file descriptors share the same flags
.RI ( mq_flags ).
.IP *
The child inherits copies of the parent's set of open directory streams (see
.BR opendir (3)).
POSIX.1 says that the corresponding directory streams
in the parent and child
.I may
share the directory stream positioning;
on Linux/glibc they do not.
.SH RETURN VALUE
On success, the PID of the child process is returned in the parent,
and 0 is returned in the child.
On failure, \-1 is returned in the parent,
no child process is created, and
.I errno
is set to indicate the error.
.SH ERRORS
.TP
.B EAGAIN
.\" NOTE! The following should match the description in pthread_create(3)
A system-imposed limit on the number of threads was encountered.
There are a number of limits that may trigger this error:
.RS
.IP * 3
the
.BR RLIMIT_NPROC
soft resource limit (set via
.BR setrlimit (2)),
which limits the number of processes and threads for a real user ID,
was reached;
.IP *
the kernel's system-wide limit on the number of processes and threads,
.IR /proc/sys/kernel/threads\-max ,
was reached (see
.BR proc (5));
.IP *
the maximum number of PIDs,
.IR /proc/sys/kernel/pid_max ,
was reached (see
.BR proc (5));
or
.IP *
the PID limit
.RI ( pids.max )
imposed by the cgroup "process number" (PIDs) controller was reached.
.RE
.TP
.B EAGAIN
The caller is operating under the
.BR SCHED_DEADLINE
scheduling policy and does not have the reset-on-fork flag set.
See
.BR sched (7).
.TP
.B ENOMEM
.BR fork ()
failed to allocate the necessary kernel structures because memory is tight.
.TP
.B ENOMEM
An attempt was made to create a child process in a PID namespace
whose "init" process has terminated.
See
.BR pid_namespaces (7).
.TP
.B ENOSYS
.BR fork ()
is not supported on this platform (for example,
.\" e.g., arm (optionally), blackfin, c6x, frv, h8300, microblaze, xtensa
hardware without a Memory-Management Unit).
.TP
.BR ERESTARTNOINTR " (since Linux 2.6.17)"
.\" commit 4a2c7a7837da1b91468e50426066d988050e4d56
System call was interrupted by a signal and will be restarted.
(This can be seen only during a trace.)
.SH CONFORMING TO
POSIX.1-2001, POSIX.1-2008, SVr4, 4.3BSD.
.SH NOTES
Under Linux,
.BR fork ()
is implemented using copy-on-write pages, so the only penalty that it incurs
is the time and memory required to duplicate the parent's page tables,
and to create a unique task structure for the child.
.SS C library/kernel differences
Since version 2.3.3,
.\" nptl/sysdeps/unix/sysv/linux/fork.c
rather than invoking the kernel's
.BR fork ()
system call,
the glibc
.BR fork ()
wrapper that is provided as part of the
NPTL threading implementation invokes
.BR clone (2)
with flags that provide the same effect as the traditional system call.
(A call to
.BR fork ()
is equivalent to a call to
.BR clone (2)
specifying
.I flags
as just
.BR SIGCHLD .)
The glibc wrapper invokes any fork handlers that have been
established using
.BR pthread_atfork (3).
.\" and does some magic to ensure that getpid(2) returns the right value.
.SH EXAMPLES
See
.BR pipe (2)
and
.BR wait (2).
.SH SEE ALSO
.BR clone (2),
.BR execve (2),
.BR exit (2),
.BR setrlimit (2),
.BR unshare (2),
.BR vfork (2),
.BR wait (2),
.BR daemon (3),
.BR pthread_atfork (3),
.BR capabilities (7),
.BR credentials (7)