freebsd-dev/lib/libc/sys/fork.2
Konstantin Belousov 2a51e8823a fork(2): comment about doubtful use of stdio and exit(3) in example
Add fflush(stdout) as the common idiom.  Explain the need to use exit()
but advise against it.

Reviewed by:	emaste, markj
Sponsored by:	The FreeBSD Foundation
MFC after:	3 days
Differential revision:	https://reviews.freebsd.org/D31425
2021-08-08 22:38:59 +03:00

272 lines
7.3 KiB
Groff

.\" Copyright (c) 1980, 1991, 1993
.\" The Regents of the University of California. All rights reserved.
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions
.\" are met:
.\" 1. Redistributions of source code must retain the above copyright
.\" notice, this list of conditions and the following disclaimer.
.\" 2. Redistributions in binary form must reproduce the above copyright
.\" notice, this list of conditions and the following disclaimer in the
.\" documentation and/or other materials provided with the distribution.
.\" 3. Neither the name of the University nor the names of its contributors
.\" may be used to endorse or promote products derived from this software
.\" without specific prior written permission.
.\"
.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
.\" ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
.\" SUCH DAMAGE.
.\"
.\" @(#)fork.2 8.1 (Berkeley) 6/4/93
.\" $FreeBSD$
.\"
.Dd August 5, 2021
.Dt FORK 2
.Os
.Sh NAME
.Nm fork
.Nd create a new process
.Sh LIBRARY
.Lb libc
.Sh SYNOPSIS
.In unistd.h
.Ft pid_t
.Fn fork void
.Ft pid_t
.Fn _Fork void
.Sh DESCRIPTION
The
.Fn fork
function causes creation of a new process.
The new process (child process) is an exact copy of the
calling process (parent process) except for the following:
.Bl -bullet -offset indent
.It
The child process has a unique process ID.
.It
The child process has a different parent
process ID (i.e., the process ID of the parent process).
.It
The child process has its own copy of the parent's descriptors,
except for descriptors returned by
.Xr kqueue 2 ,
which are not inherited from the parent process.
These descriptors reference the same underlying objects, so that,
for instance, file pointers in file objects are shared between
the child and the parent, so that an
.Xr lseek 2
on a descriptor in the child process can affect a subsequent
.Xr read 2
or
.Xr write 2
by the parent.
This descriptor copying is also used by the shell to
establish standard input and output for newly created processes
as well as to set up pipes.
.It
The child process' resource utilizations
are set to 0; see
.Xr setrlimit 2 .
.It
All interval timers are cleared; see
.Xr setitimer 2 .
.It
The robust mutexes list (see
.Xr pthread_mutexattr_setrobust 3 )
is cleared for the child.
.It
The atfork handlers established with the
.Xr pthread_atfork 3
function are called as appropriate before fork in the parent process,
and after the child is created, in parent and child.
.It
The child process has only one thread,
corresponding to the calling thread in the parent process.
If the process has more than one thread,
locks and other resources held by the other threads are not released
and therefore only async-signal-safe functions
(see
.Xr sigaction 2 )
are guaranteed to work in the child process until a call to
.Xr execve 2
or a similar function.
The
.Fx
implementation of
.Fn fork
provides a usable
.Xr malloc 3 ,
and
.Xr rtld 1
services in the child process.
.El
.Pp
The
.Fn fork
function is not async-signal safe and creates a cancellation point
in the parent process.
It cannot be safely used from signal handlers, and the atfork handlers
established by
.Xr pthread_atfork 3
do not need to be async-signal safe either.
.Pp
The
.Fn _Fork
function creates a new process, similarly to
.Fn fork ,
but it is async-signal safe.
.Fn _Fork
does not call atfork handlers, and does not create a cancellation point.
It can be used safely from signal handlers, but then no userspace
services (
.Xr malloc 3
or
.Xr rtld 1 )
are available in the child if forked from multi-threaded parent.
In particular, if using dynamic linking, all dynamic symbols used by the
child after
.Fn _Fork
must be pre-resolved.
Note: resolving can be done globally by specifying the
.Ev LD_BIND_NOW
environment variable to the dynamic linker, or per-binary by passing the
.Fl z Ar now
option to the static linker
.Xr ld 1 ,
or by using each symbol before the
.Fn _Fork
call to force the binding.
.Sh RETURN VALUES
Upon successful completion,
.Fn fork
and
.Fn _Fork
return a value
of 0 to the child process and return the process ID of the child
process to the parent process.
Otherwise, a value of -1 is returned
to the parent process, no child process is created, and the global
variable
.Va errno
is set to indicate the error.
.Sh EXAMPLES
The following example shows a common pattern of how
.Fn fork
is used in practice.
.Bd -literal -offset indent
#include <err.h>
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
int
main(void)
{
pid_t pid;
/*
* If child is expected to use stdio(3), state of
* the reused io streams must be synchronized between
* parent and child, to avoid double output and other
* possible issues.
*/
fflush(stdout);
switch (pid = fork()) {
case -1:
err(1, "Failed to fork");
case 0:
printf("Hello from child process!\en");
/*
* Since we wrote into stdout, child needs to use
* exit(3) and not _exit(2). This causes handlers
* registered with atexit(3) to be called twice,
* once in parent, and once in the child. If such
* behavior is undesirable, consider
* terminating child with _exit(2) or _Exit(3).
*/
exit(0);
default:
break;
}
printf("Hello from parent process (child's PID: %d)!\en", pid);
return (0);
}
.Ed
.Pp
The output of such a program is along the lines of:
.Bd -literal -offset indent
Hello from parent process (child's PID: 27804)!
Hello from child process!
.Ed
.Sh ERRORS
The
.Fn fork
system call will fail and no child process will be created if:
.Bl -tag -width Er
.It Bq Er EAGAIN
The system-imposed limit on the total
number of processes under execution would be exceeded.
The limit is given by the
.Xr sysctl 3
MIB variable
.Dv KERN_MAXPROC .
(The limit is actually ten less than this
except for the super user).
.It Bq Er EAGAIN
The user is not the super user, and
the system-imposed limit
on the total number of
processes under execution by a single user would be exceeded.
The limit is given by the
.Xr sysctl 3
MIB variable
.Dv KERN_MAXPROCPERUID .
.It Bq Er EAGAIN
The user is not the super user, and
the soft resource limit corresponding to the
.Fa resource
argument
.Dv RLIMIT_NPROC
would be exceeded (see
.Xr getrlimit 2 ) .
.It Bq Er ENOMEM
There is insufficient swap space for the new process.
.El
.Sh SEE ALSO
.Xr execve 2 ,
.Xr rfork 2 ,
.Xr setitimer 2 ,
.Xr setrlimit 2 ,
.Xr sigaction 2 ,
.Xr vfork 2 ,
.Xr wait 2 ,
.Xr pthread_atfork 3
.Sh HISTORY
The
.Fn fork
function appeared in
.At v1 .
.Pp
The
.Fn _Fork
function was defined by Austin Group together with the removal
of a requirement that the
.Fn fork
implementation must be async-signal safe.
The
.Fn _Fork
function appeared in
.Fx 14.0 .