1WAIT(2) Linux Programmer's Manual WAIT(2)
2
3
4
6 wait, waitpid, waitid - wait for process to change state
7
9 #include <sys/types.h>
10 #include <sys/wait.h>
11
12 pid_t wait(int *wstatus);
13
14 pid_t waitpid(pid_t pid, int *wstatus, int options);
15
16 int waitid(idtype_t idtype, id_t id, siginfo_t *infop, int options);
17 /* This is the glibc and POSIX interface; see
18 NOTES for information on the raw system call. */
19
20 Feature Test Macro Requirements for glibc (see feature_test_macros(7)):
21
22 waitid():
23 Since glibc 2.26: _XOPEN_SOURCE >= 500 ||
24 _POSIX_C_SOURCE >= 200809L
25 Glibc 2.25 and earlier:
26 _XOPEN_SOURCE
27 || /* Since glibc 2.12: */ _POSIX_C_SOURCE >= 200809L
28 || /* Glibc versions <= 2.19: */ _BSD_SOURCE
29
31 All of these system calls are used to wait for state changes in a child
32 of the calling process, and obtain information about the child whose
33 state has changed. A state change is considered to be: the child ter‐
34 minated; the child was stopped by a signal; or the child was resumed by
35 a signal. In the case of a terminated child, performing a wait allows
36 the system to release the resources associated with the child; if a
37 wait is not performed, then the terminated child remains in a "zombie"
38 state (see NOTES below).
39
40 If a child has already changed state, then these calls return immedi‐
41 ately. Otherwise, they block until either a child changes state or a
42 signal handler interrupts the call (assuming that system calls are not
43 automatically restarted using the SA_RESTART flag of sigaction(2)). In
44 the remainder of this page, a child whose state has changed and which
45 has not yet been waited upon by one of these system calls is termed
46 waitable.
47
48 wait() and waitpid()
49 The wait() system call suspends execution of the calling thread until
50 one of its children terminates. The call wait(&wstatus) is equivalent
51 to:
52
53 waitpid(-1, &wstatus, 0);
54
55 The waitpid() system call suspends execution of the calling thread
56 until a child specified by pid argument has changed state. By default,
57 waitpid() waits only for terminated children, but this behavior is mod‐
58 ifiable via the options argument, as described below.
59
60 The value of pid can be:
61
62 < -1 meaning wait for any child process whose process group ID is
63 equal to the absolute value of pid.
64
65 -1 meaning wait for any child process.
66
67 0 meaning wait for any child process whose process group ID is
68 equal to that of the calling process at the time of the call to
69 waitpid().
70
71 > 0 meaning wait for the child whose process ID is equal to the
72 value of pid.
73
74 The value of options is an OR of zero or more of the following con‐
75 stants:
76
77 WNOHANG return immediately if no child has exited.
78
79 WUNTRACED also return if a child has stopped (but not traced via
80 ptrace(2)). Status for traced children which have stopped
81 is provided even if this option is not specified.
82
83 WCONTINUED (since Linux 2.6.10)
84 also return if a stopped child has been resumed by delivery
85 of SIGCONT.
86
87 (For Linux-only options, see below.)
88
89 If wstatus is not NULL, wait() and waitpid() store status information
90 in the int to which it points. This integer can be inspected with the
91 following macros (which take the integer itself as an argument, not a
92 pointer to it, as is done in wait() and waitpid()!):
93
94 WIFEXITED(wstatus)
95 returns true if the child terminated normally, that is, by call‐
96 ing exit(3) or _exit(2), or by returning from main().
97
98 WEXITSTATUS(wstatus)
99 returns the exit status of the child. This consists of the
100 least significant 8 bits of the status argument that the child
101 specified in a call to exit(3) or _exit(2) or as the argument
102 for a return statement in main(). This macro should be employed
103 only if WIFEXITED returned true.
104
105 WIFSIGNALED(wstatus)
106 returns true if the child process was terminated by a signal.
107
108 WTERMSIG(wstatus)
109 returns the number of the signal that caused the child process
110 to terminate. This macro should be employed only if WIFSIGNALED
111 returned true.
112
113 WCOREDUMP(wstatus)
114 returns true if the child produced a core dump (see core(5)).
115 This macro should be employed only if WIFSIGNALED returned true.
116
117 This macro is not specified in POSIX.1-2001 and is not available
118 on some UNIX implementations (e.g., AIX, SunOS). Therefore,
119 enclose its use inside #ifdef WCOREDUMP ... #endif.
120
121 WIFSTOPPED(wstatus)
122 returns true if the child process was stopped by delivery of a
123 signal; this is possible only if the call was done using WUN‐
124 TRACED or when the child is being traced (see ptrace(2)).
125
126 WSTOPSIG(wstatus)
127 returns the number of the signal which caused the child to stop.
128 This macro should be employed only if WIFSTOPPED returned true.
129
130 WIFCONTINUED(wstatus)
131 (since Linux 2.6.10) returns true if the child process was
132 resumed by delivery of SIGCONT.
133
134 waitid()
135 The waitid() system call (available since Linux 2.6.9) provides more
136 precise control over which child state changes to wait for.
137
138 The idtype and id arguments select the child(ren) to wait for, as fol‐
139 lows:
140
141 idtype == P_PID
142 Wait for the child whose process ID matches id.
143
144 idtype == P_PIDFD (since Linux 5.4)
145 Wait for the child referred to by the PID file descriptor speci‐
146 fied in id. (See pidfd_open(2) for further information on PID
147 file descriptors.)
148
149 idtype == P_PGID
150 Wait for any child whose process group ID matches id. Since
151 Linux 5.4, if id is zero, then wait for any child that is in the
152 same process group as the caller's process group at the time of
153 the call.
154
155 idtype == P_ALL
156 Wait for any child; id is ignored.
157
158 The child state changes to wait for are specified by ORing one or more
159 of the following flags in options:
160
161 WEXITED Wait for children that have terminated.
162
163 WSTOPPED Wait for children that have been stopped by delivery of a
164 signal.
165
166 WCONTINUED Wait for (previously stopped) children that have been
167 resumed by delivery of SIGCONT.
168
169 The following flags may additionally be ORed in options:
170
171 WNOHANG As for waitpid().
172
173 WNOWAIT Leave the child in a waitable state; a later wait call can
174 be used to again retrieve the child status information.
175
176 Upon successful return, waitid() fills in the following fields of the
177 siginfo_t structure pointed to by infop:
178
179 si_pid The process ID of the child.
180
181 si_uid The real user ID of the child. (This field is not set on
182 most other implementations.)
183
184 si_signo Always set to SIGCHLD.
185
186 si_status Either the exit status of the child, as given to _exit(2)
187 (or exit(3)), or the signal that caused the child to termi‐
188 nate, stop, or continue. The si_code field can be used to
189 determine how to interpret this field.
190
191 si_code Set to one of: CLD_EXITED (child called _exit(2));
192 CLD_KILLED (child killed by signal); CLD_DUMPED (child
193 killed by signal, and dumped core); CLD_STOPPED (child
194 stopped by signal); CLD_TRAPPED (traced child has trapped);
195 or CLD_CONTINUED (child continued by SIGCONT).
196
197 If WNOHANG was specified in options and there were no children in a
198 waitable state, then waitid() returns 0 immediately and the state of
199 the siginfo_t structure pointed to by infop depends on the implementa‐
200 tion. To (portably) distinguish this case from that where a child was
201 in a waitable state, zero out the si_pid field before the call and
202 check for a nonzero value in this field after the call returns.
203
204 POSIX.1-2008 Technical Corrigendum 1 (2013) adds the requirement that
205 when WNOHANG is specified in options and there were no children in a
206 waitable state, then waitid() should zero out the si_pid and si_signo
207 fields of the structure. On Linux and other implementations that
208 adhere to this requirement, it is not necessary to zero out the si_pid
209 field before calling waitid(). However, not all implementations follow
210 the POSIX.1 specification on this point.
211
213 wait(): on success, returns the process ID of the terminated child; on
214 error, -1 is returned.
215
216 waitpid(): on success, returns the process ID of the child whose state
217 has changed; if WNOHANG was specified and one or more child(ren) speci‐
218 fied by pid exist, but have not yet changed state, then 0 is returned.
219 On error, -1 is returned.
220
221 waitid(): returns 0 on success or if WNOHANG was specified and no
222 child(ren) specified by id has yet changed state; on error, -1 is
223 returned.
224
225 Each of these calls sets errno to an appropriate value in the case of
226 an error.
227
229 ECHILD (for wait()) The calling process does not have any unwaited-for
230 children.
231
232 ECHILD (for waitpid() or waitid()) The process specified by pid (wait‐
233 pid()) or idtype and id (waitid()) does not exist or is not a
234 child of the calling process. (This can happen for one's own
235 child if the action for SIGCHLD is set to SIG_IGN. See also the
236 Linux Notes section about threads.)
237
238 EINTR WNOHANG was not set and an unblocked signal or a SIGCHLD was
239 caught; see signal(7).
240
241 EINVAL The options argument was invalid.
242
244 SVr4, 4.3BSD, POSIX.1-2001.
245
247 A child that terminates, but has not been waited for becomes a "zom‐
248 bie". The kernel maintains a minimal set of information about the zom‐
249 bie process (PID, termination status, resource usage information) in
250 order to allow the parent to later perform a wait to obtain information
251 about the child. As long as a zombie is not removed from the system
252 via a wait, it will consume a slot in the kernel process table, and if
253 this table fills, it will not be possible to create further processes.
254 If a parent process terminates, then its "zombie" children (if any) are
255 adopted by init(1), (or by the nearest "subreaper" process as defined
256 through the use of the prctl(2) PR_SET_CHILD_SUBREAPER operation);
257 init(1) automatically performs a wait to remove the zombies.
258
259 POSIX.1-2001 specifies that if the disposition of SIGCHLD is set to
260 SIG_IGN or the SA_NOCLDWAIT flag is set for SIGCHLD (see sigaction(2)),
261 then children that terminate do not become zombies and a call to wait()
262 or waitpid() will block until all children have terminated, and then
263 fail with errno set to ECHILD. (The original POSIX standard left the
264 behavior of setting SIGCHLD to SIG_IGN unspecified. Note that even
265 though the default disposition of SIGCHLD is "ignore", explicitly set‐
266 ting the disposition to SIG_IGN results in different treatment of zom‐
267 bie process children.)
268
269 Linux 2.6 conforms to the POSIX requirements. However, Linux 2.4 (and
270 earlier) does not: if a wait() or waitpid() call is made while SIGCHLD
271 is being ignored, the call behaves just as though SIGCHLD were not
272 being ignored, that is, the call blocks until the next child terminates
273 and then returns the process ID and status of that child.
274
275 Linux notes
276 In the Linux kernel, a kernel-scheduled thread is not a distinct con‐
277 struct from a process. Instead, a thread is simply a process that is
278 created using the Linux-unique clone(2) system call; other routines
279 such as the portable pthread_create(3) call are implemented using
280 clone(2). Before Linux 2.4, a thread was just a special case of a
281 process, and as a consequence one thread could not wait on the children
282 of another thread, even when the latter belongs to the same thread
283 group. However, POSIX prescribes such functionality, and since Linux
284 2.4 a thread can, and by default will, wait on children of other
285 threads in the same thread group.
286
287 The following Linux-specific options are for use with children created
288 using clone(2); they can also, since Linux 4.7, be used with waitid():
289
290 __WCLONE
291 Wait for "clone" children only. If omitted, then wait for "non-
292 clone" children only. (A "clone" child is one which delivers no
293 signal, or a signal other than SIGCHLD to its parent upon termi‐
294 nation.) This option is ignored if __WALL is also specified.
295
296 __WALL (since Linux 2.4)
297 Wait for all children, regardless of type ("clone" or "non-
298 clone").
299
300 __WNOTHREAD (since Linux 2.4)
301 Do not wait for children of other threads in the same thread
302 group. This was the default before Linux 2.4.
303
304 Since Linux 4.7, the __WALL flag is automatically implied if the child
305 is being ptraced.
306
307 C library/kernel differences
308 wait() is actually a library function that (in glibc) is implemented as
309 a call to wait4(2).
310
311 On some architectures, there is no waitpid() system call; instead, this
312 interface is implemented via a C library wrapper function that calls
313 wait4(2).
314
315 The raw waitid() system call takes a fifth argument, of type struct
316 rusage *. If this argument is non-NULL, then it is used to return
317 resource usage information about the child, in the same manner as
318 wait4(2). See getrusage(2) for details.
319
321 According to POSIX.1-2008, an application calling waitid() must ensure
322 that infop points to a siginfo_t structure (i.e., that it is a non-null
323 pointer). On Linux, if infop is NULL, waitid() succeeds, and returns
324 the process ID of the waited-for child. Applications should avoid
325 relying on this inconsistent, nonstandard, and unnecessary feature.
326
328 The following program demonstrates the use of fork(2) and waitpid().
329 The program creates a child process. If no command-line argument is
330 supplied to the program, then the child suspends its execution using
331 pause(2), to allow the user to send signals to the child. Otherwise,
332 if a command-line argument is supplied, then the child exits immedi‐
333 ately, using the integer supplied on the command line as the exit sta‐
334 tus. The parent process executes a loop that monitors the child using
335 waitpid(), and uses the W*() macros described above to analyze the wait
336 status value.
337
338 The following shell session demonstrates the use of the program:
339
340 $ ./a.out &
341 Child PID is 32360
342 [1] 32359
343 $ kill -STOP 32360
344 stopped by signal 19
345 $ kill -CONT 32360
346 continued
347 $ kill -TERM 32360
348 killed by signal 15
349 [1]+ Done ./a.out
350 $
351
352 Program source
353
354 #include <sys/wait.h>
355 #include <stdlib.h>
356 #include <unistd.h>
357 #include <stdio.h>
358
359 int
360 main(int argc, char *argv[])
361 {
362 pid_t cpid, w;
363 int wstatus;
364
365 cpid = fork();
366 if (cpid == -1) {
367 perror("fork");
368 exit(EXIT_FAILURE);
369 }
370
371 if (cpid == 0) { /* Code executed by child */
372 printf("Child PID is %ld\n", (long) getpid());
373 if (argc == 1)
374 pause(); /* Wait for signals */
375 _exit(atoi(argv[1]));
376
377 } else { /* Code executed by parent */
378 do {
379 w = waitpid(cpid, &wstatus, WUNTRACED | WCONTINUED);
380 if (w == -1) {
381 perror("waitpid");
382 exit(EXIT_FAILURE);
383 }
384
385 if (WIFEXITED(wstatus)) {
386 printf("exited, status=%d\n", WEXITSTATUS(wstatus));
387 } else if (WIFSIGNALED(wstatus)) {
388 printf("killed by signal %d\n", WTERMSIG(wstatus));
389 } else if (WIFSTOPPED(wstatus)) {
390 printf("stopped by signal %d\n", WSTOPSIG(wstatus));
391 } else if (WIFCONTINUED(wstatus)) {
392 printf("continued\n");
393 }
394 } while (!WIFEXITED(wstatus) && !WIFSIGNALED(wstatus));
395 exit(EXIT_SUCCESS);
396 }
397 }
398
400 _exit(2), clone(2), fork(2), kill(2), ptrace(2), sigaction(2), sig‐
401 nal(2), wait4(2), pthread_create(3), core(5), credentials(7), signal(7)
402
404 This page is part of release 5.04 of the Linux man-pages project. A
405 description of the project, information about reporting bugs, and the
406 latest version of this page, can be found at
407 https://www.kernel.org/doc/man-pages/.
408
409
410
411Linux 2019-11-19 WAIT(2)