fork(2)fork(2)NAME
fork, vfork - Create a new process
SYNOPSIS
#include <unistd.h>
pid_t fork(
void ); pid_t vfork(
void );
Application developers may want to specify an #include statement for
<sys/types.h> before the one for <unistd.h> if programs are being
developed for multiple platforms. The additional #include statement is
not required on Tru64 UNIX systems or by ISO or XSH standards, but may
be on other vendors' systems that conform to these standards.
STANDARDS
Interfaces documented on this reference page conform to industry stan‐
dards as follows:
fork(): XSH4.0, XSH4.2, XSH5.0
vfork(): XSH4.2, XSH5.0
Refer to the standards(5) reference page for more information about
industry standards and associated tags.
DESCRIPTION
The fork() and vfork() functions create a new process (child process)
that is identical to the calling process (parent process).
The child process inherits the following from the parent process: Envi‐
ronment Close-on-exec flags Signal-handling settings Set user ID mode
bit Set group ID mode bit Trusted state Profiling on/off status Nice
value All attached shared libraries Process group ID tty group ID Cur‐
rent directory Root directory File mode creation mask File size limit
Attached shared memory segments Attached mapped file segments All
mapped regions, with the same protection and sharing mode as in the
parent process [Tru64 UNIX] Message catalog descriptors. These are
shared by parent and child processes until a modification is made. The
parent's policy and priority settings for the SCHED_FIFO and SCHED_RR
scheduling policies (fork() call) Open semaphores. Any semaphores open
in the parent process are also open in the child process.
The child process differs from the parent process in the following
ways: The child process has a unique process ID that does not match any
active process group ID. The parent process ID of the child process
matches the process ID of the parent. The child process has its own
copy of the parent process's file descriptors. Each of the child's file
descriptors refers to the same open file description with the corre‐
sponding file descriptor of the parent process. The child process has
its own copy of the parent's open directory streams. Each open direc‐
tory stream in the child process may share directory stream positioning
with the corresponding directory stream of the parent. The child
process has its own copy of the parent's message queue descriptors,
each of which refers to the same open message description as referred
to by the corresponding message queue descriptor of the parent. All
semadj values are cleared. Process locks, text locks, and data locks
are not inherited by the child process. The child process' values of
tms_utime, tms_stime, tms_cutime, and tms_cstime are set to 0 (zero).
Any pending alarms are cleared in the child process. [Tru64 UNIX] Any
interval timers enabled by the parent process are reset in the child
process. Any signals pending for the parent process are cleared for
the child process. Address space memory locks established by the par‐
ent process through calls to mlockall() or mlock() are not inherited by
the child process. Per-process timers created by the parent process
are not inherited by the child process. Asynchronous input or asyn‐
chronous output operations started by the parent process are not inher‐
ited by the child process.
NOTES
If a multithreaded process forks a child process, the new process con‐
tains a replica of the calling thread and its entire address space,
possibly including the states of mutexes and other resources. Conse‐
quently, to avoid errors, the child process should only execute opera‐
tions it knows will not cause deadlock.
The fork() and vfork() functions have essentially the same implementa‐
tion at the level of the operating system kernel but may differ in how
they are supported through different libraries. Some libraries, such
as libpthread and libc, support fork handler routines that can acquire
and release resources that are critical to the child process. Fork han‐
dlers therefore allow an application to manage potential deadlock situ‐
ations that might occur between the parent and child processes. Fork
handlers do not work correctly if the application calls vfork() to cre‐
ate the child process. Therefore, applications using libpthread and
libc should call fork() to create a child process.
For more information about fork handler routines, see
pthread_atfork(3). For a complete list of system calls that are reen‐
trant with respect to signals, see signal(4).
RETURN VALUES
Upon successful completion, the fork() and vfork() functions return a
value of 0 (zero) 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, no child process is created, and errno is set
to indicate the error.
ERRORS
The fork() and vfork() functions set errno to the specified values for
the following conditions: The limit on the total number of processes
executing for a single user would be exceeded. This limit can be
exceeded by a process with superuser privilege. There is not enough
space left for this process.
SEE ALSO
Functions: exec(2), exit(2), getpriority(2), getrusage(2), plock(2),
ptrace(2), semop(2), shmat(2), sigaction(2), sigvec(2), umask(2),
wait(2), nice(3), pthread_atfork(3), raise(3), times(3), ulimit(3)
Files: signal(4)
Others: standards(5)fork(2)
