1 .\" Copyright (c) 1993 by Thomas Koenig <ig25@rz.uni-karlsruhe.de>
2 .\" and Copyright (c) 2004 by Michael Kerrisk <mtk.manpages@gmail.com>
4 .\" %%%LICENSE_START(VERBATIM)
5 .\" Permission is granted to make and distribute verbatim copies of this
6 .\" manual provided the copyright notice and this permission notice are
7 .\" preserved on all copies.
9 .\" Permission is granted to copy and distribute modified versions of this
10 .\" manual under the conditions for verbatim copying, provided that the
11 .\" entire resulting derived work is distributed under the terms of a
12 .\" permission notice identical to this one.
14 .\" Since the Linux kernel and libraries are constantly changing, this
15 .\" manual page may be incorrect or out-of-date. The author(s) assume no
16 .\" responsibility for errors or omissions, or for damages resulting from
17 .\" the use of the information contained herein. The author(s) may not
18 .\" have taken the same level of care in the production of this manual,
19 .\" which is licensed free of charge, as they might when working
22 .\" Formatted or processed versions of this manual, if unaccompanied by
23 .\" the source, must acknowledge the copyright and authors of this work.
26 .\" Modified Sat Jul 24 13:30:06 1993 by Rik Faith <faith@cs.unc.edu>
27 .\" Modified Sun Aug 21 17:42:42 1994 by Rik Faith <faith@cs.unc.edu>
28 .\" (Thanks to Koen Holtman <koen@win.tue.nl>)
29 .\" Modified Wed May 17 15:54:12 1995 by Rik Faith <faith@cs.unc.edu>
30 .\" To remove *'s from status in macros (Thanks to Michael Shields).
31 .\" Modified as suggested by Nick Duffek <nsd@bbc.com>, aeb, 960426
32 .\" Modified Mon Jun 23 14:09:52 1997 by aeb - add EINTR.
33 .\" Modified Thu Nov 26 02:12:45 1998 by aeb - add SIGCHLD stuff.
34 .\" Modified Mon Jul 24 21:37:38 2000 by David A. Wheeler
35 .\" <dwheeler@dwheeler.com> - noted thread issues.
36 .\" Modified 26 Jun 01 by Michael Kerrisk
37 .\" Added __WCLONE, __WALL, and __WNOTHREAD descriptions
38 .\" Modified 2001-09-25, aeb
39 .\" Modified 26 Jun 01 by Michael Kerrisk, <mtk.manpages@gmail.com>
40 .\" Updated notes on setting disposition of SIGCHLD to SIG_IGN
42 .\" Added waitid(2); added WCONTINUED and WIFCONTINUED()
43 .\" Added text on SA_NOCLDSTOP
44 .\" Updated discussion of SA_NOCLDWAIT to reflect 2.6 behavior
45 .\" Much other text rewritten
46 .\" 2005-05-10, mtk, __W* flags can't be used with waitid()
47 .\" 2008-07-04, mtk, removed erroneous text about SA_NOCLDSTOP
49 .TH WAIT 2 2021-03-22 "Linux" "Linux Programmer's Manual"
51 wait, waitpid, waitid \- wait for process to change state
54 .B #include <sys/wait.h>
56 .BI "pid_t wait(int *" "wstatus" );
57 .BI "pid_t waitpid(pid_t " pid ", int *" wstatus ", int " options );
59 .BI "int waitid(idtype_t " idtype ", id_t " id \
60 ", siginfo_t *" infop ", int " options );
61 /* This is the glibc and POSIX interface; see
62 NOTES for information on the raw system call. */
66 Feature Test Macro Requirements for glibc (see
67 .BR feature_test_macros (7)):
73 _XOPEN_SOURCE >= 500 || _POSIX_C_SOURCE >= 200809L
74 .\" (_XOPEN_SOURCE && _XOPEN_SOURCE_EXTENDED)
75 Glibc 2.25 and earlier:
77 || /* Since glibc 2.12: */ _POSIX_C_SOURCE >= 200809L
78 || /* Glibc <= 2.19: */ _BSD_SOURCE
81 All of these system calls are used to wait for state changes
82 in a child of the calling process, and obtain information
83 about the child whose state has changed.
84 A state change is considered to be: the child terminated;
85 the child was stopped by a signal; or the child was resumed by a signal.
86 In the case of a terminated child, performing a wait allows
87 the system to release the resources associated with the child;
88 if a wait is not performed, then the terminated child remains in
89 a "zombie" state (see NOTES below).
91 If a child has already changed state, then these calls return immediately.
92 Otherwise, they block until either a child changes state or
93 a signal handler interrupts the call (assuming that system calls
94 are not automatically restarted using the
98 In the remainder of this page, a child whose state has changed
99 and which has not yet been waited upon by one of these system
102 .SS wait() and waitpid()
105 system call suspends execution of the calling thread until one of its
113 waitpid(\-1, &wstatus, 0);
119 system call suspends execution of the calling thread 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 at the time of the call to
144 meaning wait for the child whose process ID is equal to the
150 is an OR of zero or more of the following constants:
153 return immediately if no child has exited.
156 also return if a child has stopped
161 children which have stopped is provided
162 even if this option is not specified.
164 .BR WCONTINUED " (since Linux 2.6.10)"
165 also return if a stopped child has been resumed by delivery of
168 (For Linux-only options, see below.)
176 store status information in the \fIint\fP to which it points.
177 This integer can be inspected with the following macros (which
178 take the integer itself as an argument, not a pointer to it,
184 .BI WIFEXITED( wstatus )
185 returns true if the child terminated normally, that is,
190 or by returning from main().
192 .BI WEXITSTATUS( wstatus )
193 returns the exit status of the child.
194 This consists of the least significant 8 bits of the
196 argument that the child specified in a call to
200 or as the argument for a return statement in main().
201 This macro should be employed only if
205 .BI WIFSIGNALED( wstatus )
206 returns true if the child process was terminated by a signal.
208 .BI WTERMSIG( wstatus )
209 returns the number of the signal that caused the child process to
211 This macro should be employed only if
215 .BI WCOREDUMP( wstatus )
216 returns true if the child produced a core dump (see
218 This macro should be employed only if
222 This macro is not specified in POSIX.1-2001 and is not available on
223 some UNIX implementations (e.g., AIX, SunOS).
224 Therefore, enclose its use inside
225 .IR "#ifdef WCOREDUMP ... #endif" .
227 .BI WIFSTOPPED( wstatus )
228 returns true if the child process was stopped by delivery of a signal;
229 this is possible only if the call was done using
231 or when the child is being traced (see
234 .BI WSTOPSIG( wstatus )
235 returns the number of the signal which caused the child to stop.
236 This macro should be employed only if
240 .BI WIFCONTINUED( wstatus )
242 returns true if the child process was resumed by delivery of
247 system call (available since Linux 2.6.9) provides more precise
248 control over which child state changes to wait for.
254 arguments select the child(ren) to wait for, as follows:
255 .IP "\fIidtype\fP == \fBP_PID\fP"
256 Wait for the child whose process ID matches
258 .IP "\fIidtype\fP == \fBP_PIDFD\fP (since Linux 5.4)"
259 .\" commit 3695eae5fee0605f316fbaad0b9e3de791d7dfaf
260 Wait for the child referred to by the PID file descriptor specified in
264 for further information on PID file descriptors.)
265 .IP "\fIidtype\fP == \fBP_PGID\fP"
266 Wait for any child whose process group ID matches
269 .\" commit 821cc7b0b205c0df64cce59aacc330af251fa8f7
272 is zero, then wait for any child that is in the same process group
273 as the caller's process group at the time of the call.
274 .IP "\fIidtype\fP == \fBP_ALL\fP"
279 The child state changes to wait for are specified by ORing
280 one or more of the following flags in
284 Wait for children that have terminated.
287 Wait for children that have been stopped by delivery of a signal.
290 Wait for (previously stopped) children that have been
291 resumed by delivery of
294 The following flags may additionally be ORed in
302 Leave the child in a waitable state; a later wait call
303 can be used to again retrieve the child status information.
305 Upon successful return,
307 fills in the following fields of the
309 structure pointed to by
313 The process ID of the child.
316 The real user ID of the child.
317 (This field is not set on most other implementations.)
324 Either the exit status of the child, as given to
328 or the signal that caused the child to terminate, stop, or continue.
331 field can be used to determine how to interpret this field.
339 (child killed by signal);
341 (child killed by signal, and dumped core);
343 (child stopped by signal);
345 (traced child has trapped); or
354 and there were no children in a waitable state, then
356 returns 0 immediately and
359 structure pointed to by
361 depends on the implementation.
362 To (portably) distinguish this case from that where a child was in a
363 waitable state, zero out the
365 field before the call and check for a nonzero value in this field
366 after the call returns.
368 POSIX.1-2008 Technical Corrigendum 1 (2013) adds the requirement that when
372 and there were no children in a waitable state, then
378 fields of the structure.
379 On Linux and other implementations that adhere to this requirement,
380 it is not necessary to zero out the
385 not all implementations follow the POSIX.1 specification on this point.
386 .\" POSIX.1-2001 leaves this possibility unspecified; most
387 .\" implementations (including Linux) zero out the structure
388 .\" in this case, but at least one implementation (AIX 5.1)
389 .\" does not -- MTK Nov 04
392 on success, returns the process ID of the terminated child;
393 on failure, \-1 is returned.
396 on success, returns the process ID of the child whose state has changed;
399 was specified and one or more child(ren) specified by
401 exist, but have not yet changed state, then 0 is returned.
402 On failure, \-1 is returned.
405 returns 0 on success or
408 was specified and no child(ren) specified by
410 has yet changed state;
411 on failure, \-1 is returned.
412 .\" FIXME As reported by Vegard Nossum, if infop is NULL, then waitid()
413 .\" returns the PID of the child. Either this is a bug, or it is intended
414 .\" behavior that needs to be documented. See my Jan 2009 LKML mail
415 .\" "waitid() return value strangeness when infop is NULL".
417 On failure, each of these calls sets
419 to indicate the error.
425 The calling process does not have any unwaited-for children.
432 The process specified by
440 does not exist or is not a child of the calling process.
441 (This can happen for one's own child if the action for
445 See also the \fILinux Notes\fP section about threads.)
449 was not set and an unblocked signal or a
457 argument was invalid.
459 SVr4, 4.3BSD, POSIX.1-2001.
461 A child that terminates, but has not been waited for becomes a "zombie".
462 The kernel maintains a minimal set of information about the zombie
463 process (PID, termination status, resource usage information)
464 in order to allow the parent to later perform a wait to obtain
465 information about the child.
466 As long as a zombie is not removed from the system via a wait,
467 it will consume a slot in the kernel process table, and if
468 this table fills, it will not be possible to create further processes.
469 If a parent process terminates, then its "zombie" children (if any)
472 (or by the nearest "subreaper" process as defined through the use of the
474 .B PR_SET_CHILD_SUBREAPER
477 automatically performs a wait to remove the zombies.
479 POSIX.1-2001 specifies that if the disposition of
489 then children that terminate do not become zombies and a call to
493 will block until all children have terminated, and then fail with
497 (The original POSIX standard left the behavior of setting
502 Note that even though the default disposition of
504 is "ignore", explicitly setting the disposition to
506 results in different treatment of zombie process children.)
508 Linux 2.6 conforms to the POSIX requirements.
509 However, Linux 2.4 (and earlier) does not:
516 is being ignored, the call behaves just as though
518 were not being ignored, that is, the call blocks until the next child
519 terminates and then returns the process ID and status of that child.
521 In the Linux kernel, a kernel-scheduled thread is not a distinct
522 construct from a process.
523 Instead, a thread is simply a process
524 that is created using the Linux-unique
526 system call; other routines such as the portable
527 .BR pthread_create (3)
528 call are implemented using
530 Before Linux 2.4, a thread was just a special case of a process,
531 and as a consequence one thread could not wait on the children
532 of another thread, even when the latter belongs to the same thread group.
533 However, POSIX prescribes such functionality, and since Linux 2.4
534 a thread can, and by default will, wait on children of other threads
535 in the same thread group.
537 The following Linux-specific
539 are for use with children created using
541 they can also, since Linux 4.7,
542 .\" commit 91c4e8ea8f05916df0c8a6f383508ac7c9e10dba
548 Wait for "clone" children only.
549 If omitted, then wait for "non-clone" children only.
550 (A "clone" child is one which delivers no signal, or a signal other than
552 to its parent upon termination.)
553 This option is ignored if
557 .BR __WALL " (since Linux 2.4)"
558 .\" since patch-2.3.48
559 Wait for all children, regardless of
560 type ("clone" or "non-clone").
562 .BR __WNOTHREAD " (since Linux 2.4)"
563 .\" since patch-2.4.0-test8
564 Do not wait for children of other threads in
565 the same thread group.
566 This was the default before Linux 2.4.
569 .\" commit bf959931ddb88c4e4366e96dd22e68fa0db9527c
570 .\" prevents cases where an unreapable zombie is created if
571 .\" /sbin/init doesn't use __WALL.
574 flag is automatically implied if the child is being ptraced.
575 .SS C library/kernel differences
577 is actually a library function that (in glibc) is implemented as a call to
580 On some architectures, there is no
583 .\" e.g., i386 has the system call, but not x86-64
584 instead, this interface is implemented via a C library
585 wrapper function that calls
590 system call takes a fifth argument, of type
591 .IR "struct rusage\ *" .
592 If this argument is non-NULL,
593 then it is used to return resource usage information about the child,
594 in the same manner as
600 According to POSIX.1-2008, an application calling
606 structure (i.e., that it is a non-null pointer).
611 succeeds, and returns the process ID of the waited-for child.
612 Applications should avoid relying on this inconsistent,
613 nonstandard, and unnecessary feature.
615 .\" fork.2 refers to this example program.
616 The following program demonstrates the use of
620 The program creates a child process.
621 If no command-line argument is supplied to the program,
622 then the child suspends its execution using
624 to allow the user to send signals to the child.
625 Otherwise, if a command-line argument is supplied,
626 then the child exits immediately,
627 using the integer supplied on the command line as the exit status.
628 The parent process executes a loop that monitors the child using
630 and uses the W*() macros described above to analyze the wait status value.
632 The following shell session demonstrates the use of the program:
639 .RB "$" " kill \-STOP 32360"
641 .RB "$" " kill \-CONT 32360"
643 .RB "$" " kill \-TERM 32360"
652 #include <sys/wait.h>
659 main(int argc, char *argv[])
670 if (cpid == 0) { /* Code executed by child */
671 printf("Child PID is %jd\en", (intmax_t) getpid());
673 pause(); /* Wait for signals */
674 _exit(atoi(argv[1]));
676 } else { /* Code executed by parent */
678 w = waitpid(cpid, &wstatus, WUNTRACED | WCONTINUED);
684 if (WIFEXITED(wstatus)) {
685 printf("exited, status=%d\en", WEXITSTATUS(wstatus));
686 } else if (WIFSIGNALED(wstatus)) {
687 printf("killed by signal %d\en", WTERMSIG(wstatus));
688 } else if (WIFSTOPPED(wstatus)) {
689 printf("stopped by signal %d\en", WSTOPSIG(wstatus));
690 } else if (WIFCONTINUED(wstatus)) {
691 printf("continued\en");
693 } while (!WIFEXITED(wstatus) && !WIFSIGNALED(wstatus));
707 .BR pthread_create (3),