1WAIT(2) Linux Programmer's Manual WAIT(2)
2
6 wait, waitpid, waitid - 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);
15 int waitid(idtype_t idtype, id_t id, siginfo_t *infop, int options);
17 Feature Test Macro Requirements for glibc (see feature_test_macros(7)):
19 waitid():
21 _SVID_SOURCE || _XOPEN_SOURCE >= 500 ||
22 _XOPEN_SOURCE && _XOPEN_SOURCE_EXTENDED
23 || /* Since glibc 2.12: */ _POSIX_C_SOURCE >= 200809L
24
26 All of these system calls are used to wait for state changes in a child
27 of the calling process, and obtain information about the child whose
28 state has changed. A state change is considered to be: the child ter‐
29 minated; the child was stopped by a signal; or the child was resumed by
30 a signal. In the case of a terminated child, performing a wait allows
31 the system to release the resources associated with the child; if a
32 wait is not performed, then the terminated child remains in a "zombie"
33 state (see NOTES below).
34 If a child has already changed state, then these calls return immedi‐
36 ately. Otherwise they block until either a child changes state or a
37 signal handler interrupts the call (assuming that system calls are not
38 automatically restarted using the SA_RESTART flag of sigaction(2)). In
39 the remainder of this page, a child whose state has changed and which
40 has not yet been waited upon by one of these system calls is termed
41 waitable.
42 wait() and waitpid()
44 The wait() system call suspends execution of the calling process until
45 one of its children terminates. The call wait(&status) is equivalent
46 to:
47 waitpid(-1, &status, 0);
49 The waitpid() system call suspends execution of the calling process
51 until a child specified by pid argument has changed state. By default,
52 waitpid() waits only for terminated children, but this behavior is mod‐
53 ifiable via the options argument, as described below.
54 The value of pid can be:
56 < -1 meaning wait for any child process whose process group ID is
58 equal to the absolute value of pid.
59 -1 meaning wait for any child process.
61 0 meaning wait for any child process whose process group ID is
63 equal to that of the calling process.
64 > 0 meaning wait for the child whose process ID is equal to the
66 value of pid.
67 The value of options is an OR of zero or more of the following con‐
69 stants:
70 WNOHANG return immediately if no child has exited.
72 WUNTRACED also return if a child has stopped (but not traced via
74 ptrace(2)). Status for traced children which have stopped
75 is provided even if this option is not specified.
76 WCONTINUED (since Linux 2.6.10)
78 also return if a stopped child has been resumed by delivery
79 of SIGCONT.
80 (For Linux-only options, see below.)
82 If status is not NULL, wait() and waitpid() store status information in
84 the int to which it points. This integer can be inspected with the
85 following macros (which take the integer itself as an argument, not a
86 pointer to it, as is done in wait() and waitpid()!):
87 WIFEXITED(status)
89 returns true if the child terminated normally, that is, by call‐
90 ing exit(3) or _exit(2), or by returning from main().
91 WEXITSTATUS(status)
93 returns the exit status of the child. This consists of the
94 least significant 8 bits of the status argument that the child
95 specified in a call to exit(3) or _exit(2) or as the argument
96 for a return statement in main(). This macro should be employed
97 only if WIFEXITED returned true.
98 WIFSIGNALED(status)
100 returns true if the child process was terminated by a signal.
101 WTERMSIG(status)
103 returns the number of the signal that caused the child process
104 to terminate. This macro should be employed only if WIFSIGNALED
105 returned true.
106 WCOREDUMP(status)
108 returns true if the child produced a core dump. This macro
109 should be employed only if WIFSIGNALED returned true. This
110 macro is not specified in POSIX.1-2001 and is not available on
111 some UNIX implementations (e.g., AIX, SunOS). Only use this
112 enclosed in #ifdef WCOREDUMP ... #endif.
113 WIFSTOPPED(status)
115 returns true if the child process was stopped by delivery of a
116 signal; this is possible only if the call was done using WUN‐
117 TRACED or when the child is being traced (see ptrace(2)).
118 WSTOPSIG(status)
120 returns the number of the signal which caused the child to stop.
121 This macro should be employed only if WIFSTOPPED returned true.
122 WIFCONTINUED(status)
124 (since Linux 2.6.10) returns true if the child process was
125 resumed by delivery of SIGCONT.
126 waitid()
128 The waitid() system call (available since Linux 2.6.9) provides more
129 precise control over which child state changes to wait for.
130 The idtype and id arguments select the child(ren) to wait for, as fol‐
132 lows:
133 idtype == P_PID
135 Wait for the child whose process ID matches id.
136 idtype == P_PGID
138 Wait for any child whose process group ID matches id.
139 idtype == P_ALL
141 Wait for any child; id is ignored.
142 The child state changes to wait for are specified by ORing one or more
144 of the following flags in options:
145 WEXITED Wait for children that have terminated.
147 WSTOPPED Wait for children that have been stopped by delivery of a
149 signal.
150 WCONTINUED Wait for (previously stopped) children that have been
152 resumed by delivery of SIGCONT.
153 The following flags may additionally be ORed in options:
155 WNOHANG As for waitpid().
157 WNOWAIT Leave the child in a waitable state; a later wait call can
159 be used to again retrieve the child status information.
160 Upon successful return, waitid() fills in the following fields of the
162 siginfo_t structure pointed to by infop:
163 si_pid The process ID of the child.
165 si_uid The real user ID of the child. (This field is not set on
167 most other implementations.)
168 si_signo Always set to SIGCHLD.
170 si_status Either the exit status of the child, as given to _exit(2)
172 (or exit(3)), or the signal that caused the child to termi‐
173 nate, stop, or continue. The si_code field can be used to
174 determine how to interpret this field.
175 si_code Set to one of: CLD_EXITED (child called _exit(2));
177 CLD_KILLED (child killed by signal); CLD_DUMPED (child
178 killed by signal, and dumped core); CLD_STOPPED (child
179 stopped by signal); CLD_TRAPPED (traced child has trapped);
180 or CLD_CONTINUED (child continued by SIGCONT).
181 If WNOHANG was specified in options and there were no children in a
183 waitable state, then waitid() returns 0 immediately and the state of
184 the siginfo_t structure pointed to by infop is unspecified. To distin‐
185 guish this case from that where a child was in a waitable state, zero
186 out the si_pid field before the call and check for a nonzero value in
187 this field after the call returns.
188
190 wait(): on success, returns the process ID of the terminated child; on
191 error, -1 is returned.
192 waitpid(): on success, returns the process ID of the child whose state
194 has changed; if WNOHANG was specified and one or more child(ren) speci‐
195 fied by pid exist, but have not yet changed state, then 0 is returned.
196 On error, -1 is returned.
197 waitid(): returns 0 on success or if WNOHANG was specified and no
199 child(ren) specified by id has yet changed state; on error, -1 is
200 returned. Each of these calls sets errno to an appropriate value in
201 the case of an error.
202
204 ECHILD (for wait()) The calling process does not have any unwaited-for
205 children.
206 ECHILD (for waitpid() or waitid()) The process specified by pid (wait‐
208 pid()) or idtype and id (waitid()) does not exist or is not a
209 child of the calling process. (This can happen for one's own
210 child if the action for SIGCHLD is set to SIG_IGN. See also the
211 Linux Notes section about threads.)
212 EINTR WNOHANG was not set and an unblocked signal or a SIGCHLD was
214 caught; see signal(7).
215 EINVAL The options argument was invalid.
217
219 SVr4, 4.3BSD, POSIX.1-2001.
220
222 A child that terminates, but has not been waited for becomes a "zom‐
223 bie". The kernel maintains a minimal set of information about the zom‐
224 bie process (PID, termination status, resource usage information) in
225 order to allow the parent to later perform a wait to obtain information
226 about the child. As long as a zombie is not removed from the system
227 via a wait, it will consume a slot in the kernel process table, and if
228 this table fills, it will not be possible to create further processes.
229 If a parent process terminates, then its "zombie" children (if any) are
230 adopted by init(8), which automatically performs a wait to remove the
231 zombies.
232 POSIX.1-2001 specifies that if the disposition of SIGCHLD is set to
234 SIG_IGN or the SA_NOCLDWAIT flag is set for SIGCHLD (see sigaction(2)),
235 then children that terminate do not become zombies and a call to wait()
236 or waitpid() will block until all children have terminated, and then
237 fail with errno set to ECHILD. (The original POSIX standard left the
238 behavior of setting SIGCHLD to SIG_IGN unspecified. Note that even
239 though the default disposition of SIGCHLD is "ignore", explicitly set‐
240 ting the disposition to SIG_IGN results in different treatment of zom‐
241 bie process children.) Linux 2.6 conforms to this specification. How‐
242 ever, Linux 2.4 (and earlier) does not: if a wait() or waitpid() call
243 is made while SIGCHLD is being ignored, the call behaves just as though
244 SIGCHLD were not being ignored, that is, the call blocks until the next
245 child terminates and then returns the process ID and status of that
246 child.
247 Linux notes
249 In the Linux kernel, a kernel-scheduled thread is not a distinct con‐
250 struct from a process. Instead, a thread is simply a process that is
251 created using the Linux-unique clone(2) system call; other routines
252 such as the portable pthread_create(3) call are implemented using
253 clone(2). Before Linux 2.4, a thread was just a special case of a
254 process, and as a consequence one thread could not wait on the children
255 of another thread, even when the latter belongs to the same thread
256 group. However, POSIX prescribes such functionality, and since Linux
257 2.4 a thread can, and by default will, wait on children of other
258 threads in the same thread group.
259 The following Linux-specific options are for use with children created
261 using clone(2); they cannot be used with waitid():
262 __WCLONE
264 Wait for "clone" children only. If omitted then wait for "non-
265 clone" children only. (A "clone" child is one which delivers no
266 signal, or a signal other than SIGCHLD to its parent upon termi‐
267 nation.) This option is ignored if __WALL is also specified.
268 __WALL (since Linux 2.4)
270 Wait for all children, regardless of type ("clone" or "non-
271 clone").
272 __WNOTHREAD (since Linux 2.4)
274 Do not wait for children of other threads in the same thread
275 group. This was the default before Linux 2.4.
276
278 According to POSIX.1-2008, an application calling waitid() must ensure
279 that infop points to a siginfo_t structure (i.e., that it is a non-NULL
280 pointer). On Linux, if infop is NULL, waitid() succeeds, and returns
281 the process ID of the waited-for child. Applications should avoid
282 relying on this inconsistent, nonstandard, and unnecessary feature.
283
285 The following program demonstrates the use of fork(2) and waitpid().
286 The program creates a child process. If no command-line argument is
287 supplied to the program, then the child suspends its execution using
288 pause(2), to allow the user to send signals to the child. Otherwise,
289 if a command-line argument is supplied, then the child exits immedi‐
290 ately, using the integer supplied on the command line as the exit sta‐
291 tus. The parent process executes a loop that monitors the child using
292 waitpid(), and uses the W*() macros described above to analyze the wait
293 status value.
294 The following shell session demonstrates the use of the program:
296 $ ./a.out &
298 Child PID is 32360
299 [1] 32359
300 $ kill -STOP 32360
301 stopped by signal 19
302 $ kill -CONT 32360
303 continued
304 $ kill -TERM 32360
305 killed by signal 15
306 [1]+ Done ./a.out
307 $
308 Program source
310 #include <sys/wait.h>
312 #include <stdlib.h>
313 #include <unistd.h>
314 #include <stdio.h>
315 int
317 main(int argc, char *argv[])
318 {
319 pid_t cpid, w;
320 int status;
321 cpid = fork();
323 if (cpid == -1) {
324 perror("fork");
325 exit(EXIT_FAILURE);
326 }
327 if (cpid == 0) { /* Code executed by child */
329 printf("Child PID is %ld\n", (long) getpid());
330 if (argc == 1)
331 pause(); /* Wait for signals */
332 _exit(atoi(argv[1]));
333 } else { /* Code executed by parent */
335 do {
336 w = waitpid(cpid, &status, WUNTRACED | WCONTINUED);
337 if (w == -1) {
338 perror("waitpid");
339 exit(EXIT_FAILURE);
340 }
341 if (WIFEXITED(status)) {
343 printf("exited, status=%d\n", WEXITSTATUS(status));
344 } else if (WIFSIGNALED(status)) {
345 printf("killed by signal %d\n", WTERMSIG(status));
346 } else if (WIFSTOPPED(status)) {
347 printf("stopped by signal %d\n", WSTOPSIG(status));
348 } else if (WIFCONTINUED(status)) {
349 printf("continued\n");
350 }
351 } while (!WIFEXITED(status) && !WIFSIGNALED(status));
352 exit(EXIT_SUCCESS);
353 }
354 }
355
357 _exit(2), clone(2), fork(2), kill(2), ptrace(2), sigaction(2), sig‐
358 nal(2), wait4(2), pthread_create(3), credentials(7), signal(7)
359
361 This page is part of release 3.53 of the Linux man-pages project. A
362 description of the project, and information about reporting bugs, can
363 be found at http://www.kernel.org/doc/man-pages/.
364Linux 2012-12-21 WAIT(2)