1 .\" Hey Emacs! This file is -*- nroff -*- source.
3 .\" Copyright (c) 1993 by Thomas Koenig <ig25@rz.uni-karlsruhe.de>
4 .\" and Copyright (c) 2004 by Michael Kerrisk <mtk.manpages@gmail.com>
6 .\" Permission is granted to make and distribute verbatim copies of this
7 .\" manual provided the copyright notice and this permission notice are
8 .\" preserved on all copies.
10 .\" Permission is granted to copy and distribute modified versions of this
11 .\" manual under the conditions for verbatim copying, provided that the
12 .\" entire resulting derived work is distributed under the terms of a
13 .\" permission notice identical to this one.
15 .\" Since the Linux kernel and libraries are constantly changing, this
16 .\" manual page may be incorrect or out-of-date. The author(s) assume no
17 .\" responsibility for errors or omissions, or for damages resulting from
18 .\" the use of the information contained herein. The author(s) may not
19 .\" have taken the same level of care in the production of this manual,
20 .\" which is licensed free of charge, as they might when working
23 .\" Formatted or processed versions of this manual, if unaccompanied by
24 .\" the source, must acknowledge the copyright and authors of this work.
27 .\" Modified Sat Jul 24 13:30:06 1993 by Rik Faith <faith@cs.unc.edu>
28 .\" Modified Sun Aug 21 17:42:42 1994 by Rik Faith <faith@cs.unc.edu>
29 .\" (Thanks to Koen Holtman <koen@win.tue.nl>)
30 .\" Modified Wed May 17 15:54:12 1995 by Rik Faith <faith@cs.unc.edu>
31 .\" To remove *'s from status in macros (Thanks to Michael Shields).
32 .\" Modified as suggested by Nick Duffek <nsd@bbc.com>, aeb, 960426
33 .\" Modified Mon Jun 23 14:09:52 1997 by aeb - add EINTR.
34 .\" Modified Thu Nov 26 02:12:45 1998 by aeb - add SIGCHLD stuff.
35 .\" Modified Mon Jul 24 21:37:38 2000 by David A. Wheeler
36 .\" <dwheeler@dwheeler.com> - noted thread issues.
37 .\" Modified 26 Jun 01 by Michael Kerrisk
38 .\" Added __WCLONE, __WALL, and __WNOTHREAD descriptions
39 .\" Modified 2001-09-25, aeb
40 .\" Modified 26 Jun 01 by Michael Kerrisk, <mtk.manpages@gmail.com>
41 .\" Updated notes on setting disposition of SIGCHLD to SIG_IGN
43 .\" Added waitid(2); added WCONTINUED and WIFCONTINUED()
44 .\" Added text on SA_NOCLDSTOP
45 .\" Updated discussion of SA_NOCLDWAIT to reflect 2.6 behavior
46 .\" Much other text rewritten
47 .\" 2005-05-10, mtk, __W* flags can't be used with waitid()
48 .\" 2008-07-04, mtk, removed erroneous text about SA_NOCLDSTOP
50 .TH WAIT 2 2012-12-21 "Linux" "Linux Programmer's Manual"
52 wait, waitpid, waitid \- wait for process to change state
54 .B #include <sys/types.h>
56 .B #include <sys/wait.h>
58 .BI "pid_t wait(int *" "status" );
60 .BI "pid_t waitpid(pid_t " pid ", int *" status ", int " options );
62 .BI "int waitid(idtype_t " idtype ", id_t " id \
63 ", siginfo_t *" infop ", int " options );
66 Feature Test Macro Requirements for glibc (see
67 .BR feature_test_macros (7)):
75 _XOPEN_SOURCE\ >=\ 500 ||
76 _XOPEN_SOURCE\ &&\ _XOPEN_SOURCE_EXTENDED
78 || /* Since glibc 2.12: */ _POSIX_C_SOURCE\ >=\ 200809L
83 All of these system calls are used to wait for state changes
84 in a child of the calling process, and obtain information
85 about the child whose state has changed.
86 A state change is considered to be: the child terminated;
87 the child was stopped by a signal; or the child was resumed by a signal.
88 In the case of a terminated child, performing a wait allows
89 the system to release the resources associated with the child;
90 if a wait is not performed, then the terminated child remains in
91 a "zombie" state (see NOTES below).
93 If a child has already changed state, then these calls return immediately.
94 Otherwise they block until either a child changes state or
95 a signal handler interrupts the call (assuming that system calls
96 are not automatically restarted using the
100 In the remainder of this page, a child whose state has changed
101 and which has not yet been waited upon by one of these system
104 .SS "wait() and waitpid()"
107 system call suspends execution of the calling process until one of its
114 waitpid(\-1, &status, 0);
119 system call suspends execution of the calling process until a
122 argument has changed state.
125 waits only for terminated children, but this behavior is modifiable
128 argument, as described below.
134 meaning wait for any child process whose process group ID is
135 equal to the absolute value of
138 meaning wait for any child process.
140 meaning wait for any child process whose process group ID is
141 equal to that of the calling process.
143 meaning wait for the child whose process ID is equal to the
149 is an OR of zero or more of the following constants:
152 return immediately if no child has exited.
155 also return if a child has stopped
160 children which have stopped is provided
161 even if this option is not specified.
163 .BR WCONTINUED " (since Linux 2.6.10)"
164 also return if a stopped child has been resumed by delivery of
167 (For Linux-only options, see below.)
175 store status information in the \fIint\fP to which it points.
176 This integer can be inspected with the following macros (which
177 take the integer itself as an argument, not a pointer to it,
183 .BI WIFEXITED( status )
184 returns true if the child terminated normally, that is,
189 or by returning from main().
191 .BI WEXITSTATUS( status )
192 returns the exit status of the child.
193 This consists of the least significant 8 bits of the
195 argument that the child specified in a call to
199 or as the argument for a return statement in main().
200 This macro should only be employed if
204 .BI WIFSIGNALED( status )
205 returns true if the child process was terminated by a signal.
207 .BI WTERMSIG( status )
208 returns the number of the signal that caused the child process to
210 This macro should only be employed if
214 .BI WCOREDUMP( status )
215 returns true if the child produced a core dump.
216 This macro should only be employed if
219 This macro is not specified in POSIX.1-2001 and is not available on
220 some UNIX implementations (e.g., AIX, SunOS).
221 Only use this enclosed in #ifdef WCOREDUMP ... #endif.
223 .BI WIFSTOPPED( status )
224 returns true if the child process was stopped by delivery of a signal;
225 this is only possible if the call was done using
227 or when the child is being traced (see
230 .BI WSTOPSIG( status )
231 returns the number of the signal which caused the child to stop.
232 This macro should only be employed if
236 .BI WIFCONTINUED( status )
238 returns true if the child process was resumed by delivery of
243 system call (available since Linux 2.6.9) provides more precise
244 control over which child state changes to wait for.
250 arguments select the child(ren) to wait for, as follows:
251 .IP "\fIidtype\fP == \fBP_PID\fP"
252 Wait for the child whose process ID matches
254 .IP "\fIidtype\fP == \fBP_PGID\fP"
255 Wait for any child whose process group ID matches
257 .IP "\fIidtype\fP == \fBP_ALL\fP"
262 The child state changes to wait for are specified by ORing
263 one or more of the following flags in
267 Wait for children that have terminated.
270 Wait for children that have been stopped by delivery of a signal.
273 Wait for (previously stopped) children that have been
274 resumed by delivery of
277 The following flags may additionally be ORed in
285 Leave the child in a waitable state; a later wait call
286 can be used to again retrieve the child status information.
288 Upon successful return,
290 fills in the following fields of the
292 structure pointed to by
296 The process ID of the child.
299 The real user ID of the child.
300 (This field is not set on most other implementations.)
307 Either the exit status of the child, as given to
311 or the signal that caused the child to terminate, stop, or continue.
314 field can be used to determine how to interpret this field.
322 (child killed by signal);
324 (child killed by signal, and dumped core);
326 (child stopped by signal);
328 (traced child has trapped); or
337 and there were no children in a waitable state, then
339 returns 0 immediately and
342 structure pointed to by
345 .\" POSIX.1-2001 leaves this possibility unspecified; most
346 .\" implementations (including Linux) zero out the structure
347 .\" in this case, but at least one implementation (AIX 5.1)
348 .\" does not -- MTK Nov 04
349 To distinguish this case from that where a child was in a
350 waitable state, zero out the
352 field before the call and check for a nonzero value in this field
353 after the call returns.
356 on success, returns the process ID of the terminated child;
357 on error, \-1 is returned.
360 on success, returns the process ID of the child whose state has changed;
363 was specified and one or more child(ren) specified by
365 exist, but have not yet changed state, then 0 is returned.
366 On error, \-1 is returned.
369 returns 0 on success or
372 was specified and no child(ren) specified by
374 has yet changed state;
375 on error, \-1 is returned.
376 .\" FIXME: As reported by Vegard Nossum, if infop is NULL, then waitid()
377 .\" returns the PID of the child. Either this is a bug, or it is intended
378 .\" behavior that needs to be documented. See my Jan 2009 LKML mail
379 .\" "waitid() return value strangeness when infop is NULL".
380 Each of these calls sets
382 to an appropriate value in the case of an error.
388 The calling process does not have any unwaited-for children.
395 The process specified by
403 does not exist or is not a child of the calling process.
404 (This can happen for one's own child if the action for
408 See also the \fILinux Notes\fP section about threads.)
412 was not set and an unblocked signal or a
420 argument was invalid.
422 SVr4, 4.3BSD, POSIX.1-2001.
424 A child that terminates, but has not been waited for becomes a "zombie".
425 The kernel maintains a minimal set of information about the zombie
426 process (PID, termination status, resource usage information)
427 in order to allow the parent to later perform a wait to obtain
428 information about the child.
429 As long as a zombie is not removed from the system via a wait,
430 it will consume a slot in the kernel process table, and if
431 this table fills, it will not be possible to create further processes.
432 If a parent process terminates, then its "zombie" children (if any)
435 which automatically performs a wait to remove the zombies.
437 POSIX.1-2001 specifies that if the disposition of
447 then children that terminate do not become zombies and a call to
451 will block until all children have terminated, and then fail with
455 (The original POSIX standard left the behavior of setting
460 Note that even though the default disposition of
462 is "ignore", explicitly setting the disposition to
464 results in different treatment of zombie process children.)
465 Linux 2.6 conforms to this specification.
466 However, Linux 2.4 (and earlier) does not:
473 is being ignored, the call behaves just as though
475 were not being ignored, that is, the call blocks until the next child
476 terminates and then returns the process ID and status of that child.
478 In the Linux kernel, a kernel-scheduled thread is not a distinct
479 construct from a process.
480 Instead, a thread is simply a process
481 that is created using the Linux-unique
483 system call; other routines such as the portable
484 .BR pthread_create (3)
485 call are implemented using
487 Before Linux 2.4, a thread was just a special case of a process,
488 and as a consequence one thread could not wait on the children
489 of another thread, even when the latter belongs to the same thread group.
490 However, POSIX prescribes such functionality, and since Linux 2.4
491 a thread can, and by default will, wait on children of other threads
492 in the same thread group.
494 The following Linux-specific
496 are for use with children created using
498 they cannot be used with
503 Wait for "clone" children only.
504 If omitted then wait for "non-clone" children only.
505 (A "clone" child is one which delivers no signal, or a signal other than
507 to its parent upon termination.)
508 This option is ignored if
512 .BR __WALL " (since Linux 2.4)"
513 .\" since patch-2.3.48
514 Wait for all children, regardless of
515 type ("clone" or "non-clone").
517 .BR __WNOTHREAD " (since Linux 2.4)"
518 .\" since patch-2.4.0-test8
519 Do not wait for children of other threads in
520 the same thread group.
521 This was the default before Linux 2.4.
523 According to POSIX.1-2008, an application calling
529 structure (i.e., that it is a non-NULL pointer).
534 succeeds, and returns the process ID of the waited-for child.
535 Applications should avoid relying on this inconsistent,
536 nonstandard, and unnecessary feature.
538 .\" fork.2 refers to this example program.
539 The following program demonstrates the use of
543 The program creates a child process.
544 If no command-line argument is supplied to the program,
545 then the child suspends its execution using
547 to allow the user to send signals to the child.
548 Otherwise, if a command-line argument is supplied,
549 then the child exits immediately,
550 using the integer supplied on the command line as the exit status.
551 The parent process executes a loop that monitors the child using
553 and uses the W*() macros described above to analyze the wait status value.
555 The following shell session demonstrates the use of the program:
562 .RB "$" " kill \-STOP 32360"
564 .RB "$" " kill \-CONT 32360"
566 .RB "$" " kill \-TERM 32360"
575 #include <sys/wait.h>
581 main(int argc, char *argv[])
592 if (cpid == 0) { /* Code executed by child */
593 printf("Child PID is %ld\\n", (long) getpid());
595 pause(); /* Wait for signals */
596 _exit(atoi(argv[1]));
598 } else { /* Code executed by parent */
600 w = waitpid(cpid, &status, WUNTRACED | WCONTINUED);
606 if (WIFEXITED(status)) {
607 printf("exited, status=%d\\n", WEXITSTATUS(status));
608 } else if (WIFSIGNALED(status)) {
609 printf("killed by signal %d\\n", WTERMSIG(status));
610 } else if (WIFSTOPPED(status)) {
611 printf("stopped by signal %d\\n", WSTOPSIG(status));
612 } else if (WIFCONTINUED(status)) {
613 printf("continued\\n");
615 } while (!WIFEXITED(status) && !WIFSIGNALED(status));
629 .BR pthread_create (3),