1.\" Copyright (c) 1980, 1991, 1993 2.\" The Regents of the University of California. All rights reserved. 3.\" 4.\" Redistribution and use in source and binary forms, with or without 5.\" modification, are permitted provided that the following conditions 6.\" are met: 7.\" 1. Redistributions of source code must retain the above copyright 8.\" notice, this list of conditions and the following disclaimer. 9.\" 2. Redistributions in binary form must reproduce the above copyright 10.\" notice, this list of conditions and the following disclaimer in the 11.\" documentation and/or other materials provided with the distribution. 12.\" 3. Neither the name of the University nor the names of its contributors 13.\" may be used to endorse or promote products derived from this software 14.\" without specific prior written permission. 15.\" 16.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 17.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 18.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 19.\" ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 20.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 21.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 22.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 23.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 24.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 25.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 26.\" SUCH DAMAGE. 27.\" 28.\" @(#)fork.2 8.1 (Berkeley) 6/4/93 29.\" 30.Dd August 5, 2021 31.Dt FORK 2 32.Os 33.Sh NAME 34.Nm fork 35.Nd create a new process 36.Sh LIBRARY 37.Lb libc 38.Sh SYNOPSIS 39.In unistd.h 40.Ft pid_t 41.Fn fork void 42.Ft pid_t 43.Fn _Fork void 44.Sh DESCRIPTION 45The 46.Fn fork 47function causes creation of a new process. 48The new process (child process) is an exact copy of the 49calling process (parent process) except for the following: 50.Bl -bullet -offset indent 51.It 52The child process has a unique process ID. 53.It 54The child process has a different parent 55process ID (i.e., the process ID of the parent process). 56.It 57The child process has its own copy of the parent's descriptors, 58except for descriptors returned by 59.Xr kqueue 2 , 60which are not inherited from the parent process. 61These descriptors reference the same underlying objects, so that, 62for instance, file pointers in file objects are shared between 63the child and the parent, so that an 64.Xr lseek 2 65on a descriptor in the child process can affect a subsequent 66.Xr read 2 67or 68.Xr write 2 69by the parent. 70This descriptor copying is also used by the shell to 71establish standard input and output for newly created processes 72as well as to set up pipes. 73.It 74The child process' resource utilizations 75are set to 0; see 76.Xr setrlimit 2 . 77.It 78All interval timers are cleared; see 79.Xr setitimer 2 . 80.It 81The robust mutexes list (see 82.Xr pthread_mutexattr_setrobust 3 ) 83is cleared for the child. 84.It 85The atfork handlers established with the 86.Xr pthread_atfork 3 87function are called as appropriate before fork in the parent process, 88and after the child is created, in parent and child. 89.It 90The child process has only one thread, 91corresponding to the calling thread in the parent process. 92If the process has more than one thread, 93locks and other resources held by the other threads are not released 94and therefore only async-signal-safe functions 95(see 96.Xr sigaction 2 ) 97are guaranteed to work in the child process until a call to 98.Xr execve 2 99or a similar function. 100The 101.Fx 102implementation of 103.Fn fork 104provides a usable 105.Xr malloc 3 , 106and 107.Xr rtld 1 108services in the child process. 109.El 110.Pp 111The 112.Fn fork 113function is not async-signal safe and creates a cancellation point 114in the parent process. 115It cannot be safely used from signal handlers, and the atfork handlers 116established by 117.Xr pthread_atfork 3 118do not need to be async-signal safe either. 119.Pp 120The 121.Fn _Fork 122function creates a new process, similarly to 123.Fn fork , 124but it is async-signal safe. 125.Fn _Fork 126does not call atfork handlers, and does not create a cancellation point. 127It can be used safely from signal handlers, but then no userspace 128services ( 129.Xr malloc 3 130or 131.Xr rtld 1 ) 132are available in the child if forked from multi-threaded parent. 133In particular, if using dynamic linking, all dynamic symbols used by the 134child after 135.Fn _Fork 136must be pre-resolved. 137Note: resolving can be done globally by specifying the 138.Ev LD_BIND_NOW 139environment variable to the dynamic linker, or per-binary by passing the 140.Fl z Ar now 141option to the static linker 142.Xr ld 1 , 143or by using each symbol before the 144.Fn _Fork 145call to force the binding. 146.Sh RETURN VALUES 147Upon successful completion, 148.Fn fork 149and 150.Fn _Fork 151return a value 152of 0 to the child process and return the process ID of the child 153process to the parent process. 154Otherwise, a value of -1 is returned 155to the parent process, no child process is created, and the global 156variable 157.Va errno 158is set to indicate the error. 159.Sh EXAMPLES 160The following example shows a common pattern of how 161.Fn fork 162is used in practice. 163.Bd -literal -offset indent 164#include <err.h> 165#include <stdio.h> 166#include <stdlib.h> 167#include <unistd.h> 168 169int 170main(void) 171{ 172 pid_t pid; 173 174 /* 175 * If child is expected to use stdio(3), state of 176 * the reused io streams must be synchronized between 177 * parent and child, to avoid double output and other 178 * possible issues. 179 */ 180 fflush(stdout); 181 182 switch (pid = fork()) { 183 case -1: 184 err(1, "Failed to fork"); 185 case 0: 186 printf("Hello from child process!\en"); 187 188 /* 189 * Since we wrote into stdout, child needs to use 190 * exit(3) and not _exit(2). This causes handlers 191 * registered with atexit(3) to be called twice, 192 * once in parent, and once in the child. If such 193 * behavior is undesirable, consider 194 * terminating child with _exit(2) or _Exit(3). 195 */ 196 exit(0); 197 default: 198 break; 199 } 200 201 printf("Hello from parent process (child's PID: %d)!\en", pid); 202 203 return (0); 204} 205.Ed 206.Pp 207The output of such a program is along the lines of: 208.Bd -literal -offset indent 209Hello from parent process (child's PID: 27804)! 210Hello from child process! 211.Ed 212.Sh ERRORS 213The 214.Fn fork 215system call will fail and no child process will be created if: 216.Bl -tag -width Er 217.It Bq Er EAGAIN 218The system-imposed limit on the total 219number of processes under execution would be exceeded. 220The limit is given by the 221.Xr sysctl 3 222MIB variable 223.Dv KERN_MAXPROC . 224(The limit is actually ten less than this 225except for the super user). 226.It Bq Er EAGAIN 227The user is not the super user, and 228the system-imposed limit 229on the total number of 230processes under execution by a single user would be exceeded. 231The limit is given by the 232.Xr sysctl 3 233MIB variable 234.Dv KERN_MAXPROCPERUID . 235.It Bq Er EAGAIN 236The user is not the super user, and 237the soft resource limit corresponding to the 238.Fa resource 239argument 240.Dv RLIMIT_NPROC 241would be exceeded (see 242.Xr getrlimit 2 ) . 243.It Bq Er ENOMEM 244There is insufficient swap space for the new process. 245.El 246.Sh SEE ALSO 247.Xr execve 2 , 248.Xr rfork 2 , 249.Xr setitimer 2 , 250.Xr setrlimit 2 , 251.Xr sigaction 2 , 252.Xr vfork 2 , 253.Xr wait 2 , 254.Xr pthread_atfork 3 255.Sh HISTORY 256The 257.Fn fork 258function appeared in 259.At v1 . 260.Pp 261The 262.Fn _Fork 263function was defined by Austin Group together with the removal 264of a requirement that the 265.Fn fork 266implementation must be async-signal safe. 267The 268.Fn _Fork 269function appeared in 270.Fx 13.1 . 271