1FORK(2)                    Linux Programmer's Manual                   FORK(2)
2
3
4

NAME

6       fork - create a child process
7

SYNOPSIS

9       #include <unistd.h>
10
11       pid_t fork(void);
12

DESCRIPTION

14       fork()  creates  a new process by duplicating the calling process.  The
15       new process, referred to as the child, is an  exact  duplicate  of  the
16       calling  process,  referred  to as the parent, except for the following
17       points:
18
19       *  The child has its own unique process ID, and this PID does not match
20          the ID of any existing process group (setpgid(2)).
21
22       *  The  child's  parent  process ID is the same as the parent's process
23          ID.
24
25       *  The child does not inherit  its  parent's  memory  locks  (mlock(2),
26          mlockall(2)).
27
28       *  Process  resource  utilizations (getrusage(2)) and CPU time counters
29          (times(2)) are reset to zero in the child.
30
31       *  The child's set of pending  signals  is  initially  empty  (sigpend‐
32          ing(2)).
33
34       *  The  child  does  not  inherit semaphore adjustments from its parent
35          (semop(2)).
36
37       *  The child does not inherit record locks from its parent (fcntl(2)).
38
39       *  The child does not inherit timers  from  its  parent  (setitimer(2),
40          alarm(2), timer_create(2)).
41
42       *  The  child  does not inherit outstanding asynchronous I/O operations
43          from its parent (aio_read(3), aio_write(3)), nor does it inherit any
44          asynchronous I/O contexts from its parent (seeio_setup(2)).
45
46       The  process  attributes  in  the  preceding  list are all specified in
47       POSIX.1-2001.  The parent and child also differ  with  respect  to  the
48       following Linux-specific process attributes:
49
50       *  The  child does not inherit directory change notifications (dnotify)
51          from its parent (see the description of F_NOTIFY in fcntl(2)).
52
53       *  The prctl(2) PR_SET_PDEATHSIG setting is reset  so  that  the  child
54          does not receive a signal when its parent terminates.
55
56       *  Memory mappings that have been marked with the madvise(2) MADV_DONT‐
57          FORK flag are not inherited across a fork().
58
59       *  The  termination  signal  of  the  child  is  always  SIGCHLD   (see
60          clone(2)).
61
62       Note the following further points:
63
64       *  The  child  process  is  created with a single thread — the one that
65          called fork().  The entire virtual address space of  the  parent  is
66          replicated  in the child, including the states of mutexes, condition
67          variables, and other pthreads objects; the use of  pthread_atfork(3)
68          may be helpful for dealing with problems that this can cause.
69
70       *  The  child inherits copies of the parent's set of open file descrip‐
71          tors.  Each file descriptor in the child refers  to  the  same  open
72          file  description (see open(2)) as the corresponding file descriptor
73          in the parent.  This means that the two descriptors share open  file
74          status  flags, current file offset, and signal-driven I/O attributes
75          (see the description of F_SETOWN and F_SETSIG in fcntl(2)).
76
77       *  The child inherits copies of the parent's set of open message  queue
78          descriptors  (see  mq_overview(7)).   Each  descriptor  in the child
79          refers to the same open message queue description as the correspond‐
80          ing  descriptor  in the parent.  This means that the two descriptors
81          share the same flags (mq_flags).
82
83       *  The child inherits copies of the  parent's  set  of  open  directory
84          streams  (see opendir(3)).  POSIX.1-2001 says that the corresponding
85          directory streams in the parent and child may  share  the  directory
86          stream positioning; on Linux/glibc they do not.
87

RETURN VALUE

89       On success, the PID of the child process is returned in the parent, and
90       0 is returned in the child.  On failure, -1 is returned in the  parent,
91       no child process is created, and errno is set appropriately.
92

ERRORS

94       EAGAIN fork()  cannot  allocate  sufficient memory to copy the parent's
95              page tables and allocate a task structure for the child.
96
97       EAGAIN It was not possible to create a new process because the caller's
98              RLIMIT_NPROC  resource  limit  was  encountered.  To exceed this
99              limit, the process must have either  the  CAP_SYS_ADMIN  or  the
100              CAP_SYS_RESOURCE capability.
101
102       ENOMEM fork()  failed  to  allocate  the  necessary  kernel  structures
103              because memory is tight.
104

CONFORMING TO

106       SVr4, 4.3BSD, POSIX.1-2001.
107

NOTES

109       Under Linux, fork() is implemented using copy-on-write  pages,  so  the
110       only  penalty  that it incurs is the time and memory required to dupli‐
111       cate the parent's page tables, and to create a  unique  task  structure
112       for the child.
113
114       Since  version  2.3.3,  rather than invoking the kernel's fork() system
115       call, the glibc fork() wrapper that is provided as  part  of  the  NPTL
116       threading  implementation  invokes clone(2) with flags that provide the
117       same effect as the traditional system call.  The glibc wrapper  invokes
118       any fork handlers that have been established using pthread_atfork(3).
119

EXAMPLE

121       See pipe(2) and wait(2).
122

SEE ALSO

124       clone(2),  execve(2), setrlimit(2), unshare(2), vfork(2), wait(2), dae‐
125       mon(3), capabilities(7), credentials(7)
126

COLOPHON

128       This page is part of release 3.25 of the Linux  man-pages  project.   A
129       description  of  the project, and information about reporting bugs, can
130       be found at http://www.kernel.org/doc/man-pages/.
131
132
133
134Linux                             2009-04-27                           FORK(2)
Impressum