1 .\" Copyright (c) 1999 Andries Brouwer (aeb@cwi.nl), 1 Nov 1999
3 .\" Permission is granted to make and distribute verbatim copies of this
4 .\" manual provided the copyright notice and this permission notice are
5 .\" preserved on all copies.
7 .\" Permission is granted to copy and distribute modified versions of this
8 .\" manual under the conditions for verbatim copying, provided that the
9 .\" entire resulting derived work is distributed under the terms of a
10 .\" permission notice identical to this one.
12 .\" Since the Linux kernel and libraries are constantly changing, this
13 .\" manual page may be incorrect or out-of-date. The author(s) assume no
14 .\" responsibility for errors or omissions, or for damages resulting from
15 .\" the use of the information contained herein. The author(s) may not
16 .\" have taken the same level of care in the production of this manual,
17 .\" which is licensed free of charge, as they might when working
20 .\" Formatted or processed versions of this manual, if unaccompanied by
21 .\" the source, must acknowledge the copyright and authors of this work.
23 .\" 1999-11-10: Merged text taken from the page contributed by
24 .\" Reed H. Petty (rhp@draper.net)
26 .TH VFORK 2 1999-11-01 "Linux" "Linux Programmer's Manual"
28 vfork \- create a child process and block parent
30 .B #include <sys/types.h>
32 .B #include <unistd.h>
36 .SS "Standard Description"
37 (From SUSv2 / POSIX draft.)
40 function has the same effect as
42 except that the behavior is undefined if the process created by
44 either modifies any data other than a variable of type
46 used to store the return value from
48 or returns from the function in which
50 was called, or calls any other function before successfully calling
55 .SS "Linux Description"
59 creates a child process of the calling process.
60 For details and return value and errors, see
66 It is used to create new processes without copying the page tables of
68 It may be useful in performance sensitive applications
69 where a child will be created which then immediately issues an
75 in that the parent is suspended until the child makes a call to
79 The child shares all memory with its parent, including the stack, until
81 is issued by the child.
82 The child must not return from the current function or call
87 Signal handlers are inherited, but not shared.
89 arrive after the child releases the parent's memory.
90 .SS "Historic Description"
93 is implemented using copy-on-write pages, so the only penalty incurred by
95 is the time and memory required to duplicate the parent's page tables,
96 and to create a unique task structure for the child.
97 However, in the bad old days a
99 would require making a complete copy of the caller's data space,
100 often needlessly, since usually immediately afterwards an
103 Thus, for greater efficiency, BSD introduced the
105 system call, that did not fully copy the address space of
106 the parent process, but borrowed the parent's memory and thread
107 of control until a call to
110 The parent process was suspended while the
111 child was using its resources.
114 was tricky: for example, not modifying data
115 in the parent process depended on knowing which variables are
118 4.3BSD, POSIX.1-2001.
119 .\" FIXME Mar 07: in the draft of the next POSIX revision, the spec for
120 .\" vfork() has been removed. See if this remains so in final version.
121 The requirements put on
123 by the standards are weaker than those put on
125 so an implementation where the two are synonymous is compliant.
126 In particular, the programmer cannot
127 rely on the parent remaining blocked until a call of
131 and cannot rely on any specific behavior w.r.t. shared memory.
132 .\" In AIXv3.1 vfork is equivalent to fork.
135 Fork handlers established using
136 .BR pthread_atfork (3)
137 are not called when a multithreaded program employing
138 the NPTL threading library calls
140 Fork handlers are called in this case in a program using the
141 LinuxThreads threading library.
144 for a description of Linux threading libraries.)
148 system call appeared in 3.0BSD.
149 .\" In the release notes for 4.2BSD Sam Leffler wrote: `vfork: Is still
150 .\" present, but definitely on its way out'.
151 In 4.4BSD it was made synonymous to
153 but NetBSD introduced it again,
154 cf. http://www.netbsd.org/Documentation/kernel/vfork.html .
155 In Linux, it has been equivalent to
157 until 2.2.0-pre6 or so.
158 Since 2.2.0-pre9 (on i386, somewhat later on
159 other architectures) it is an independent system call.
160 Support was added in glibc 2.0.112.
162 It is rather unfortunate that Linux revived this specter from the past.
163 The BSD man page states:
164 "This system call will be eliminated when proper system sharing mechanisms
166 Users should not depend on the memory sharing semantics of
168 as it will, in that case, be made synonymous to
172 Details of the signal handling are obscure and differ between systems.
173 The BSD man page states:
174 "To avoid a possible deadlock situation, processes that are children
177 are never sent SIGTTOU or SIGTTIN signals; rather, output or
179 are allowed and input attempts result in an end-of-file indication."
181 .\" As far as I can tell, the following is not true in 2.6.19:
182 .\" Currently (Linux 2.3.25),
186 .\" and requires a kernel patch.