man2/vfork.2

   1 .\" Copyright (c) 1999 Andries Brouwer (aeb@cwi.nl), 1 Nov 1999
   2 .\"
   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.
   6 .\"
   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.
  11 .\"
  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  13 .\" manual page may be incorrect or out-of-date.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
  18 .\" professionally.
  19 .\"
  20 .\" Formatted or processed versions of this manual, if unaccompanied by
  21 .\" the source, must acknowledge the copyright and authors of this work.
  22 .\"
  23 .\" 1999-11-10: Merged text taken from the page contributed by
  24 .\" Reed H. Petty (rhp@draper.net)
  25 .\"
  26 .TH VFORK 2 1999-11-01 "Linux 2.2.0" "Linux Programmer's Manual"
  27 .SH NAME
  28 vfork \- create a child process and block parent
  29 .SH SYNOPSIS
  30 .B #include <sys/types.h>
  31 .br
  32 .B #include <unistd.h>
  33 .sp
  34 .B pid_t vfork(void);
  35 .SH "STANDARD DESCRIPTION"
  36 (From XPG4 / SUSv2 / POSIX draft.)
  37 The
  38 .BR vfork ()
  39 function has the same effect as
  40 .BR fork (),
  41 except that the behaviour is undefined if the process created by
  42 .BR vfork ()
  43 either modifies any data other than a variable of type pid_t used
  44 to store the return value from
  45 .BR vfork (),
  46 or returns from the function in which
  47 .BR vfork ()
  48 was called, or calls any other function before successfully calling
  49 .BR _exit ()
  50 or one of the
  51 .IR exec ()
  52 family of functions.
  53 .SH ERRORS
  54 .TP
  55 .B EAGAIN
  56 Too many processes; try again.
  57 .TP
  58 .B ENOMEM
  59 There is insufficient swap space for the new process.
  60 .SH "LINUX DESCRIPTION"
  61 .BR vfork (),
  62 just like
  63 .BR fork (2),
  64 creates a child process of the calling process.
  65 For details and return value and errors, see
  66 .BR fork (2).
  67 .PP
  68 .BR vfork ()
  69 is a special case of
  70 .BR clone (2).
  71 It is used to create new processes without copying the page tables of
  72 the parent process.  It may be useful in performance sensitive applications
  73 where a child will be created which then immediately issues an
  74 .BR execve () .
  75 .PP
  76 .BR vfork ()
  77 differs from
  78 .BR fork ()
  79 in that the parent is suspended until the child makes a call to
  80 .BR execve (2)
  81 or
  82 .BR _exit (2).
  83 The child shares all memory with its parent, including the stack, until
  84 .BR execve ()
  85 is issued by the child.  The child must not return from the
  86 current function or call
  87 .BR exit (),
  88 but may call
  89 .BR _exit ().
  90 .PP
  91 Signal handlers are inherited, but not shared.  Signals to the parent
  92 arrive after the child releases the parent.
  93 .SH "HISTORIC DESCRIPTION"
  94 Under Linux,
  95 .BR fork ()
  96 is implemented using copy-on-write pages, so the only penalty incurred by
  97 .BR fork ()
  98 is the time and memory required to duplicate the parent's page tables,
  99 and to create a unique task structure for the child.
 100 However, in the bad old days a
 101 .BR fork ()
 102 would require making a complete copy of the caller's data space,
 103 often needlessly, since usually immediately afterwards an
 104 .BR exec ()
 105 is done. Thus, for greater efficiency, BSD introduced the
 106 .BR vfork ()
 107 system call, that did not fully copy the address space of
 108 the parent process, but borrowed the parent's memory and thread
 109 of control until a call to
 110 .BR execve ()
 111 or an exit occurred. The parent process was suspended while the
 112 child was using its resources.
 113 The use of
 114 .BR vfork ()
 115 was tricky: for example, not modifying data
 116 in the parent process depended on knowing which variables are
 117 held in a register.
 118 .SH BUGS
 119 It is rather unfortunate that Linux revived this spectre from the past.
 120 The BSD manpage states:
 121 "This system call will be eliminated when proper system sharing mechanisms
 122 are implemented. Users should not depend on the memory sharing semantics of
 123 .BR vfork ()
 124 as it will, in that case, be made synonymous to
 125 .BR fork ().\c
 126 "
 127
 128 Formally speaking, the standard description given above does not allow
 129 one to use
 130 .BR vfork ()
 131 since a following
 132 .IR exec ()
 133 might fail, and then what happens is undefined.
 134
 135 Details of the signal handling are obscure and differ between systems.
 136 The BSD manpage states:
 137 "To avoid a possible deadlock situation, processes that are children
 138 in the middle of a
 139 .BR vfork ()
 140 are never sent SIGTTOU or SIGTTIN signals; rather, output or
 141 .IR ioctl s
 142 are allowed and input attempts result in an end-of-file indication."
 143
 144 Currently (Linux 2.3.25),
 145 .BR strace (1)
 146 cannot follow
 147 .BR vfork ()
 148 and requires a kernel patch.
 149 .SH HISTORY
 150 The
 151 .BR vfork ()
 152 system call appeared in 3.0BSD.
 153 .\" In the release notes for 4.2BSD Sam Leffler wrote: `vfork: Is still
 154 .\" present, but definitely on its way out'.
 155 In 4.4BSD it was made synonymous to
 156 .BR fork ()
 157 but NetBSD introduced it again,
 158 cf. http://www.netbsd.org/Documentation/kernel/vfork.html .
 159 In Linux, it has been equivalent to
 160 .BR fork ()
 161 until 2.2.0-pre6 or so. Since 2.2.0-pre9 (on i386, somewhat later on
 162 other architectures) it is an independent system call. Support was
 163 added in glibc 2.0.112.
 164 .SH "CONFORMING TO"
 165 The
 166 .BR vfork ()
 167 call may be a bit similar to calls with the same name in other
 168 operating systems. The requirements put on
 169 .BR vfork ()
 170 by the standards are weaker than those put on
 171 .BR fork (),
 172 so an implementation where the two are synonymous
 173 is compliant. In particular, the programmer cannot
 174 rely on the parent remaining blocked until a call of
 175 .BR execve ()
 176 or
 177 .BR _exit ()
 178 and cannot rely on any specific behaviour w.r.t. shared memory.
 179 .\" In AIXv3.1 vfork is equivalent to fork.
 180 .SH "SEE ALSO"
 181 .BR clone (2),
 182 .BR execve (2),
 183 .BR fork (2),
 184 .BR wait (2)