1WAIT(2) Linux Programmer's Manual WAIT(2)
2
6 wait, waitpid - wait for process to change state
7
9 #include <sys/types.h>
10 #include <sys/wait.h>
11 pid_t wait(int *status);
13 pid_t waitpid(pid_t pid, int *status, int options);
14 int waitid(idtype_t idtype, id_t id, siginfo_t *infop, int options);
15
17 All of these system calls are used to wait for state changes in a child
18 of the calling process, and obtain information about the child whose
19 state has changed. A state change is considered to be: the child ter‐
20 minated; the child was stopped by a signal; or the child was resumed by
21 a signal. In the case of a terminated child, performing a wait allows
22 the system to release the resources associated with the child; if a
23 wait is not performed, then terminated the child remains in a "zombie"
24 state (see NOTES below).
25 If a child has already changed state, then these calls return immedi‐
27 ately. Otherwise they block until either a child changes state or a
28 signal handler interrupts the call (assuming that system calls are not
29 automatically restarted using the SA_RESTART flag of sigaction(2)). In
30 the remainder of this page, a child whose state has changed and which
31 has not yet been waited upon by one of these system calls is termed
32 waitable.
33 wait() and waitpid()
35 The wait() system call suspends execution of the current process until
36 one of its children terminates. The call wait(&status) is equivalent
37 to:
38 waitpid(-1, &status, 0);
40 The waitpid() system call suspends execution of the current process
42 until a child specified by pid argument has changed state. By default,
43 waitpid() waits only for terminated children, but this behaviour is
44 modifiable via the options argument, as described below.
45 The value of pid can be:
47 < -1 meaning wait for any child process whose process group ID is
49 equal to the absolute value of pid.
50 -1 meaning wait for any child process.
52 0 meaning wait for any child process whose process group ID is
54 equal to that of the calling process.
55 > 0 meaning wait for the child whose process ID is equal to the
57 value of pid.
58 The value of options is an OR of zero or more of the following con‐
60 stants:
61 WNOHANG
63 return immediately if no child has exited.
64 WUNTRACED
66 also return if a child has stopped (but not traced via
67 ptrace(2)). Status for traced children which have stopped is
68 provided even if this option is not specified.
69 WCONTINUED
71 (Since Linux 2.6.10) also return if a stopped child has been
72 resumed by delivery of SIGCONT.
73 (For Linux-only options, see below.)
75 The WUNTRACED and WCONTINUED options are only effective if the
77 SA_NOCLDSTOP flag has not been set for the SIGCHLD signal (see sigac‐
78 tion(2)).
79 If status is not NULL, wait() and waitpid() store status information in
81 the int to which it points. This integer can be inspected with the
82 following macros (which take the integer itself as an argument, not a
83 pointer to it, as is done in wait() and waitpid()!):
84 WIFEXITED(status)
86 returns true if the child terminated normally, that is, by call‐
87 ing exit(3) or _exit(2), or by returning from main().
88 WEXITSTATUS(status)
90 returns the exit status of the child. This consists of the
91 least significant 8 bits of the status argument that the child
92 specified in a call to exit() or _exit() or as the argument for
93 a return statement in main(). This macro should only be
94 employed if WIFEXITED returned true.
95 WIFSIGNALED(status)
97 returns true if the child process was terminated by a signal.
98 WTERMSIG(status)
100 returns the number of the signal that caused the child process
101 to terminate. This macro should only be employed if WIFSIGNALED
102 returned true.
103 WCOREDUMP(status)
105 returns true if the child produced a core dump. This macro
106 should only be employed if WIFSIGNALED returned true. This
107 macro is not specified in POSIX.1-2001 and is not available on
108 some Unix implementations (e.g., AIX, SunOS). Only use this
109 enclosed in #ifdef WCOREDUMP ... #endif.
110 WIFSTOPPED(status)
112 returns true if the child process was stopped by delivery of a
113 signal; this is only possible if the call was done using WUN‐
114 TRACED or when the child is being traced (see ptrace(2)).
115 WSTOPSIG(status)
117 returns the number of the signal which caused the child to stop.
118 This macro should only be employed if WIFSTOPPED returned true.
119 WIFCONTINUED(status)
121 (Since Linux 2.6.10) returns true if the child process was
122 resumed by delivery of SIGCONT.
123 waitid()
125 The waitid() system call (available since Linux 2.6.9) provides more
126 precise control over which child state changes to wait for.
127 The idtype and id arguments select the child(ren) to wait for, as fol‐
129 lows:
130 idtype == P_PID
132 Wait for the child whose process ID matches id.
133 idtype == P_PGID
135 Wait for any child whose process group ID matches id.
136 idtype == P_ALL
138 Wait for any child; id is ignored.
139 The child state changes to wait for are specified by ORing one or more
141 of the following flags in options:
142 WEXITED
144 Wait for children that have terminated.
145 WSTOPPED
147 Wait for children that have been stopped by delivery of a sig‐
148 nal.
149 WCONTINUED
151 Wait for (previously stopped) children that have been resumed by
152 delivery of SIGCONT.
153 The following flags may additionally be ORed in options:
155 WNOHANG
157 As for waitpid().
158 WNOWAIT
160 Leave the child in a waitable state; a later wait call can be
161 used to again retrieve the child status information.
162 Upon successful return, waitid() fills in the following fields of the
164 siginfo_t structure pointed to by infop:
165 si_pid The process ID of the child.
167 si_uid The real user ID of the child. (This field is not set on most
169 other implementations.)
170 si_signo
172 Always set to SIGCHLD.
173 si_status
175 Either the exit status of the child, as given to _exit(2) (or
176 exit(3)), or the signal that caused the child to terminate,
177 stop, or continue. The si_code field can be used to determine
178 how to interpret this field.
179 si_code
181 Set to one of: CLD_EXITED (child called _exit(2)); CLD_KILLED
182 (child killed by signal); CLD_STOPPED (child stopped by signal);
183 or CLD_CONTINUED (child continued by SIGCONT).
184 If WNOHANG was specified in options and there were no children in a
186 waitable state, then waitid() returns 0 immediately and the state of
187 the siginfo_t structure pointed to by infop is unspecified. To distin‐
188 guish this case from that where a child was in a waitable state, zero
189 out the si_pid field before the call and check for a non-zero value in
190 this field after the call returns.
191
193 wait(): on success, returns the process ID of the terminated child; on
194 error, -1 is returned.
195 waitpid(): on success, returns the process ID of the child whose state
197 has changed; on error, -1 is returned; if WNOHANG was specified and no
198 child(ren) specified by pid has yet changed state, then 0 is returned.
199 waitid(): returns 0 on success or if WNOHANG was specified and no
201 child(ren) specified by id has yet changed state; on error, -1 is
202 returned.
203 Each of these calls sets errno to an appropriate value in the case of
205 an error.
206
208 ECHILD (for wait()) The calling process does not have any unwaited-for
209 children.
210 ECHILD (for waitpid() or waitid()) The process specified by pid (wait‐
212 pid()) or idtype and id (waitid()) does not exist or is not a
213 child of the calling process. (This can happen for one's own
214 child if the action for SIGCHLD is set to SIG_IGN. See also the
215 LINUX NOTES section about threads.)
216 EINTR WNOHANG was not set and an unblocked signal or a SIGCHLD was
218 caught.
219 EINVAL The options argument was invalid.
221
223 A child that terminates, but has not been waited for becomes a "zom‐
224 bie". The kernel maintains a minimal set of information about the zom‐
225 bie process (PID, termination status, resource usage information) in
226 order to allow the parent to later perform a wait to obtain information
227 about the child. As long as a zombie is not removed from the system
228 via a wait, it will consume a slot in the kernel process table, and if
229 this table fills, it will not be possible to create further processes.
230 If a parent process terminates, then its "zombie" children (if any) are
231 adopted by init(8), which automatically performs a wait to remove the
232 zombies.
233 POSIX.1-2001 specifies that if the disposition of SIGCHLD is set to
235 SIG_IGN or the SA_NOCLDWAIT flag is set for SIGCHLD (see sigaction(2)),
236 then children that terminate do not become zombies and a call to wait()
237 or waitpid() will block until all children have terminated, and then
238 fail with errno set to ECHILD. (The original POSIX standard left the
239 behaviour of setting SIGCHLD to SIG_IGN unspecified.) Linux 2.6 con‐
240 forms to this specification. However, Linux 2.4 (and earlier) does
241 not: if a wait() or waitpid() call is made while SIGCHLD is being
242 ignored, the call behaves just as though SIGCHLD were not being
243 ignored, that is, the call blocks until the next child terminates and
244 then returns the process ID and status of that child.
245
247 In the Linux kernel, a kernel-scheduled thread is not a distinct con‐
248 struct from a process. Instead, a thread is simply a process that is
249 created using the Linux-unique clone(2) system call; other routines
250 such as the portable pthread_create(3) call are implemented using
251 clone(2). Before Linux 2.4, a thread was just a special case of a
252 process, and as a consequence one thread could not wait on the children
253 of another thread, even when the latter belongs to the same thread
254 group. However, POSIX prescribes such functionality, and since Linux
255 2.4 a thread can, and by default will, wait on children of other
256 threads in the same thread group.
257 The following Linux specific options are for use with children created
259 using clone(2); they cannot be used with waitid():
260 __WCLONE
262 Wait for "clone" children only. If omitted then wait for "non-
263 clone" children only. (A "clone" child is one which delivers no
264 signal, or a signal other than SIGCHLD to its parent upon termi‐
265 nation.) This option is ignored if __WALL is also specified.
266 __WALL (Since Linux 2.4) Wait for all children, regardless of type
268 ("clone" or "non-clone").
269 __WNOTHREAD
271 (Since Linux 2.4) Do not wait for children of other threads in
272 the same thread group. This was the default before Linux 2.4.
273
275 The following program demonstrates the use of fork(2) and waitpid(2).
276 The program creates a child process. If no command-line argument is
277 supplied to the program, then the child suspends its execution using
278 pause(2), to allow the user to send signals to the child. Otherwise,
279 if a command-line argument is supplied, then the child exits immedi‐
280 ately, using the integer supplied on the command line as the exit sta‐
281 tus. The parent process executes a loop that monitors the child using
282 waitpid(2), and uses the W*() macros described above to analyse the
283 wait status value.
284 The following shell session demonstrates the use of the program:
286 $ ./a.out &
288 Child PID is 32360
289 [1] 32359
290 $ kill -STOP 32360
291 stopped by signal 19
292 $ kill -CONT 32360
293 continued
294 $ kill -TERM 32360
295 killed by signal 15
296 [1]+ Done ./a.out
297 $
298 #include <sys/wait.h>
301 #include <stdlib.h>
302 #include <unistd.h>
303 #include <stdio.h>
304 int
306 main(int argc, char *argv[])
307 {
308 pid_t cpid, w;
309 int status;
310 cpid = fork();
312 if (cpid == -1) { perror("fork"); exit(EXIT_FAILURE); }
313 if (cpid == 0) { /* Code executed by child */
315 printf("Child PID is %ld\n", (long) getpid());
316 if (argc == 1)
317 pause(); /* Wait for signals */
318 _exit(atoi(argv[1]));
319 } else { /* Code executed by parent */
321 do {
322 w = waitpid(cpid, &status, WUNTRACED | WCONTINUED);
323 if (w == -1) { perror("waitpid"); exit(EXIT_FAILURE); }
324 if (WIFEXITED(status)) {
326 printf("exited, status=%d\n", WEXITSTATUS(status));
327 } else if (WIFSIGNALED(status)) {
328 printf("killed by signal %d\n", WTERMSIG(status));
329 } else if (WIFSTOPPED(status)) {
330 printf("stopped by signal %d\n", WSTOPSIG(status));
331 } else if (WIFCONTINUED(status)) {
332 printf("continued\n");
333 }
334 } while (!WIFEXITED(status) && !WIFSIGNALED(status));
335 exit(EXIT_SUCCESS);
336 }
337 }
338
340 SVr4, 4.3BSD, POSIX.1-2001.
341
343 _exit(2), clone(2), fork(2), kill(2), ptrace(2), sigaction(2), sig‐
344 nal(2), wait4(2), pthread_create(3), signal(7)
345Linux 2004-11-11 WAIT(2)