pow.3: Minor tweak to BUGS

[thirdparty/man-pages.git] / man2 / wait.2

.\" Copyright (c) 1993 by Thomas Koenig <ig25@rz.uni-karlsruhe.de>
.\" and Copyright (c) 2004 by Michael Kerrisk <mtk.manpages@gmail.com>
.\"
.\" %%%LICENSE_START(VERBATIM)
.\" Permission is granted to make and distribute verbatim copies of this
.\" manual provided the copyright notice and this permission notice are
.\" preserved on all copies.
.\"
.\" Permission is granted to copy and distribute modified versions of this
.\" manual under the conditions for verbatim copying, provided that the
.\" entire resulting derived work is distributed under the terms of a
.\" permission notice identical to this one.
.\"
.\" Since the Linux kernel and libraries are constantly changing, this
.\" manual page may be incorrect or out-of-date.  The author(s) assume no
.\" responsibility for errors or omissions, or for damages resulting from
.\" the use of the information contained herein.  The author(s) may not
.\" have taken the same level of care in the production of this manual,
.\" which is licensed free of charge, as they might when working
.\" professionally.
.\"
.\" Formatted or processed versions of this manual, if unaccompanied by
.\" the source, must acknowledge the copyright and authors of this work.
.\" %%%LICENSE_END
.\"
.\" Modified Sat Jul 24 13:30:06 1993 by Rik Faith <faith@cs.unc.edu>
.\" Modified Sun Aug 21 17:42:42 1994 by Rik Faith <faith@cs.unc.edu>
.\"          (Thanks to Koen Holtman <koen@win.tue.nl>)
.\" Modified Wed May 17 15:54:12 1995 by Rik Faith <faith@cs.unc.edu>
.\"           To remove *'s from status in macros (Thanks to Michael Shields).
.\" Modified as suggested by Nick Duffek <nsd@bbc.com>, aeb, 960426
.\" Modified Mon Jun 23 14:09:52 1997 by aeb - add EINTR.
.\" Modified Thu Nov 26 02:12:45 1998 by aeb - add SIGCHLD stuff.
.\" Modified Mon Jul 24 21:37:38 2000 by David A. Wheeler
.\"          <dwheeler@dwheeler.com> - noted thread issues.
.\" Modified 26 Jun 01 by Michael Kerrisk
.\"          Added __WCLONE, __WALL, and __WNOTHREAD descriptions
.\" Modified 2001-09-25, aeb
.\" Modified 26 Jun 01 by Michael Kerrisk, <mtk.manpages@gmail.com>
.\"	Updated notes on setting disposition of SIGCHLD to SIG_IGN
.\" 2004-11-11, mtk
.\"	Added waitid(2); added WCONTINUED and WIFCONTINUED()
.\"	Added text on SA_NOCLDSTOP
.\"	Updated discussion of SA_NOCLDWAIT to reflect 2.6 behavior
.\"	Much other text rewritten
.\" 2005-05-10, mtk, __W* flags can't be used with waitid()
.\" 2008-07-04, mtk, removed erroneous text about SA_NOCLDSTOP
.\"
.TH WAIT 2 2020-04-11 "Linux" "Linux Programmer's Manual"
.SH NAME
wait, waitpid, waitid \- wait for process to change state
.SH SYNOPSIS
.B #include <sys/types.h>
.br
.B #include <sys/wait.h>
.PP
.BI "pid_t wait(int *" "wstatus" );
.PP
.BI "pid_t waitpid(pid_t " pid ", int *" wstatus ", int " options );
.PP
.BI "int waitid(idtype_t " idtype ", id_t " id \
", siginfo_t *" infop ", int " options );
                /* This is the glibc and POSIX interface; see
                   NOTES for information on the raw system call. */
.PP
.in -4n
Feature Test Macro Requirements for glibc (see
.BR feature_test_macros (7)):
.in
.PP
.ad l
.PD 0
.BR waitid ():
.RS 4
Since glibc 2.26:
_XOPEN_SOURCE >= 500 ||
.\"    (_XOPEN_SOURCE && _XOPEN_SOURCE_EXTENDED)
    _POSIX_C_SOURCE\ >=\ 200809L
.br
Glibc 2.25 and earlier:
    _XOPEN_SOURCE
        || /* Since glibc 2.12: */ _POSIX_C_SOURCE\ >=\ 200809L
        || /* Glibc versions <= 2.19: */ _BSD_SOURCE
.RE
.PD
.ad
.SH DESCRIPTION
All of these system calls are used to wait for state changes
in a child of the calling process, and obtain information
about the child whose state has changed.
A state change is considered to be: the child terminated;
the child was stopped by a signal; or the child was resumed by a signal.
In the case of a terminated child, performing a wait allows
the system to release the resources associated with the child;
if a wait is not performed, then the terminated child remains in
a "zombie" state (see NOTES below).
.PP
If a child has already changed state, then these calls return immediately.
Otherwise, they block until either a child changes state or
a signal handler interrupts the call (assuming that system calls
are not automatically restarted using the
.B SA_RESTART
flag of
.BR sigaction (2)).
In the remainder of this page, a child whose state has changed
and which has not yet been waited upon by one of these system
calls is termed
.IR waitable .
.SS wait() and waitpid()
The
.BR wait ()
system call suspends execution of the calling thread until one of its
children terminates.
The call
.I wait(&wstatus)
is equivalent to:
.PP
.in +4n
.EX
waitpid(\-1, &wstatus, 0);
.EE
.in
.PP
The
.BR waitpid ()
system call suspends execution of the calling thread until a
child specified by
.I pid
argument has changed state.
By default,
.BR waitpid ()
waits only for terminated children, but this behavior is modifiable
via the
.I options
argument, as described below.
.PP
The value of
.I pid
can be:
.IP "< \-1"
meaning wait for any child process whose process group ID is
equal to the absolute value of
.IR pid .
.IP \-1
meaning wait for any child process.
.IP 0
meaning wait for any child process whose process group ID is
equal to that of the calling process at the time of the call to
.BR waitpid ().
.IP "> 0"
meaning wait for the child whose process ID is equal to the
value of
.IR pid .
.PP
The value of
.I options
is an OR of zero or more of the following constants:
.TP
.B WNOHANG
return immediately if no child has exited.
.TP
.B WUNTRACED
also return if a child has stopped
(but not traced via
.BR ptrace (2)).
Status for
.I traced
children which have stopped is provided
even if this option is not specified.
.TP
.BR WCONTINUED " (since Linux 2.6.10)"
also return if a stopped child has been resumed by delivery of
.BR SIGCONT .
.PP
(For Linux-only options, see below.)
.PP
If
.I wstatus
is not NULL,
.BR wait ()
and
.BR waitpid ()
store status information in the \fIint\fP to which it points.
This integer can be inspected with the following macros (which
take the integer itself as an argument, not a pointer to it,
as is done in
.BR wait ()
and
.BR waitpid ()!):
.TP
.BI WIFEXITED( wstatus )
returns true if the child terminated normally, that is,
by calling
.BR exit (3)
or
.BR _exit (2),
or by returning from main().
.TP
.BI WEXITSTATUS( wstatus )
returns the exit status of the child.
This consists of the least significant 8 bits of the
.I status
argument that the child specified in a call to
.BR exit (3)
or
.BR _exit (2)
or as the argument for a return statement in main().
This macro should be employed only if
.B WIFEXITED
returned true.
.TP
.BI WIFSIGNALED( wstatus )
returns true if the child process was terminated by a signal.
.TP
.BI WTERMSIG( wstatus )
returns the number of the signal that caused the child process to
terminate.
This macro should be employed only if
.B WIFSIGNALED
returned true.
.TP
.BI WCOREDUMP( wstatus )
returns true if the child produced a core dump (see
.BR core (5)).
This macro should be employed only if
.B WIFSIGNALED
returned true.
.IP
This macro is not specified in POSIX.1-2001 and is not available on
some UNIX implementations (e.g., AIX, SunOS).
Therefore, enclose its use inside
.IR "#ifdef WCOREDUMP ... #endif" .
.TP
.BI WIFSTOPPED( wstatus )
returns true if the child process was stopped by delivery of a signal;
this is possible only if the call was done using
.B WUNTRACED
or when the child is being traced (see
.BR ptrace (2)).
.TP
.BI WSTOPSIG( wstatus )
returns the number of the signal which caused the child to stop.
This macro should be employed only if
.B WIFSTOPPED
returned true.
.TP
.BI WIFCONTINUED( wstatus )
(since Linux 2.6.10)
returns true if the child process was resumed by delivery of
.BR SIGCONT .
.SS waitid()
The
.BR waitid ()
system call (available since Linux 2.6.9) provides more precise
control over which child state changes to wait for.
.PP
The
.I idtype
and
.I id
arguments select the child(ren) to wait for, as follows:
.IP "\fIidtype\fP == \fBP_PID\fP"
Wait for the child whose process ID matches
.IR id .
.IP "\fIidtype\fP == \fBP_PIDFD\fP (since Linux 5.4)"
.\" commit 3695eae5fee0605f316fbaad0b9e3de791d7dfaf
Wait for the child referred to by the PID file descriptor specified in
.IR id .
(See
.BR pidfd_open (2)
for further information on PID file descriptors.)
.IP "\fIidtype\fP == \fBP_PGID\fP"
Wait for any child whose process group ID matches
.IR id .
Since Linux 5.4,
.\" commit 821cc7b0b205c0df64cce59aacc330af251fa8f7
if
.I id
is zero, then wait for any child that is in the same process group
as the caller's process group at the time of the call.
.IP "\fIidtype\fP == \fBP_ALL\fP"
Wait for any child;
.I id
is ignored.
.PP
The child state changes to wait for are specified by ORing
one or more of the following flags in
.IR options :
.TP
.B WEXITED
Wait for children that have terminated.
.TP
.B WSTOPPED
Wait for children that have been stopped by delivery of a signal.
.TP
.B WCONTINUED
Wait for (previously stopped) children that have been
resumed by delivery of
.BR SIGCONT .
.PP
The following flags may additionally be ORed in
.IR options :
.TP
.B WNOHANG
As for
.BR waitpid ().
.TP
.B WNOWAIT
Leave the child in a waitable state; a later wait call
can be used to again retrieve the child status information.
.PP
Upon successful return,
.BR waitid ()
fills in the following fields of the
.I siginfo_t
structure pointed to by
.IR infop :
.TP
\fIsi_pid\fP
The process ID of the child.
.TP
\fIsi_uid\fP
The real user ID of the child.
(This field is not set on most other implementations.)
.TP
\fIsi_signo\fP
Always set to
.BR SIGCHLD .
.TP
\fIsi_status\fP
Either the exit status of the child, as given to
.BR _exit (2)
(or
.BR exit (3)),
or the signal that caused the child to terminate, stop, or continue.
The
.I si_code
field can be used to determine how to interpret this field.
.TP
\fIsi_code\fP
Set to one of:
.B CLD_EXITED
(child called
.BR _exit (2));
.B CLD_KILLED
(child killed by signal);
.B CLD_DUMPED
(child killed by signal, and dumped core);
.B CLD_STOPPED
(child stopped by signal);
.B CLD_TRAPPED
(traced child has trapped); or
.B CLD_CONTINUED
(child continued by
.BR SIGCONT ).
.PP
If
.B WNOHANG
was specified in
.I options
and there were no children in a waitable state, then
.BR waitid ()
returns 0 immediately and
the state of the
.I siginfo_t
structure pointed to by
.I infop
depends on the implementation.
To (portably) distinguish this case from that where a child was in a
waitable state, zero out the
.I si_pid
field before the call and check for a nonzero value in this field
after the call returns.
.PP
POSIX.1-2008 Technical Corrigendum 1 (2013) adds the requirement that when
.B WNOHANG
is specified in
.I options
and there were no children in a waitable state, then
.BR waitid ()
should zero out the
.I si_pid
and
.I si_signo
fields of the structure.
On Linux and other implementations that adhere to this requirement,
it is not necessary to zero out the
.I si_pid
field before calling
.BR waitid ().
However,
not all implementations follow the POSIX.1 specification on this point.
.\" POSIX.1-2001 leaves this possibility unspecified; most
.\" implementations (including Linux) zero out the structure
.\" in this case, but at least one implementation (AIX 5.1)
.\" does not -- MTK Nov 04
.SH RETURN VALUE
.BR wait ():
on success, returns the process ID of the terminated child;
on error, \-1 is returned.
.PP
.BR waitpid ():
on success, returns the process ID of the child whose state has changed;
if
.B WNOHANG
was specified and one or more child(ren) specified by
.I pid
exist, but have not yet changed state, then 0 is returned.
On error, \-1 is returned.
.PP
.BR waitid ():
returns 0 on success or
if
.B WNOHANG
was specified and no child(ren) specified by
.I id
has yet changed state;
on error, \-1 is returned.
.\" FIXME As reported by Vegard Nossum, if infop is NULL, then waitid()
.\" returns the PID of the child.  Either this is a bug, or it is intended
.\" behavior that needs to be documented.  See my Jan 2009 LKML mail
.\" "waitid() return value strangeness when infop is NULL".
.PP
Each of these calls sets
.I errno
to an appropriate value in the case of an error.
.SH ERRORS
.TP
.B ECHILD
(for
.BR wait ())
The calling process does not have any unwaited-for children.
.TP
.B ECHILD
(for
.BR waitpid ()
or
.BR waitid ())
The process specified by
.I pid
.RB ( waitpid ())
or
.I idtype
and
.I id
.RB ( waitid ())
does not exist or is not a child of the calling process.
(This can happen for one's own child if the action for
.B SIGCHLD
is set to
.BR SIG_IGN .
See also the \fILinux Notes\fP section about threads.)
.TP
.B EINTR
.B WNOHANG
was not set and an unblocked signal or a
.B SIGCHLD
was caught; see
.BR signal (7).
.TP
.B EINVAL
The
.I options
argument was invalid.
.SH CONFORMING TO
SVr4, 4.3BSD, POSIX.1-2001.
.SH NOTES
A child that terminates, but has not been waited for becomes a "zombie".
The kernel maintains a minimal set of information about the zombie
process (PID, termination status, resource usage information)
in order to allow the parent to later perform a wait to obtain
information about the child.
As long as a zombie is not removed from the system via a wait,
it will consume a slot in the kernel process table, and if
this table fills, it will not be possible to create further processes.
If a parent process terminates, then its "zombie" children (if any)
are adopted by
.BR init (1),
(or by the nearest "subreaper" process as defined through the use of the
.BR prctl (2)
.B PR_SET_CHILD_SUBREAPER
operation);
.BR init (1)
automatically performs a wait to remove the zombies.
.PP
POSIX.1-2001 specifies that if the disposition of
.B SIGCHLD
is set to
.B SIG_IGN
or the
.B SA_NOCLDWAIT
flag is set for
.B SIGCHLD
(see
.BR sigaction (2)),
then children that terminate do not become zombies and a call to
.BR wait ()
or
.BR waitpid ()
will block until all children have terminated, and then fail with
.I errno
set to
.BR ECHILD .
(The original POSIX standard left the behavior of setting
.B SIGCHLD
to
.B SIG_IGN
unspecified.
Note that even though the default disposition of
.B SIGCHLD
is "ignore", explicitly setting the disposition to
.B SIG_IGN
results in different treatment of zombie process children.)
.PP
Linux 2.6 conforms to the POSIX requirements.
However, Linux 2.4 (and earlier) does not:
if a
.BR wait ()
or
.BR waitpid ()
call is made while
.B SIGCHLD
is being ignored, the call behaves just as though
.B SIGCHLD
were not being ignored, that is, the call blocks until the next child
terminates and then returns the process ID and status of that child.
.SS Linux notes
In the Linux kernel, a kernel-scheduled thread is not a distinct
construct from a process.
Instead, a thread is simply a process
that is created using the Linux-unique
.BR clone (2)
system call; other routines such as the portable
.BR pthread_create (3)
call are implemented using
.BR clone (2).
Before Linux 2.4, a thread was just a special case of a process,
and as a consequence one thread could not wait on the children
of another thread, even when the latter belongs to the same thread group.
However, POSIX prescribes such functionality, and since Linux 2.4
a thread can, and by default will, wait on children of other threads
in the same thread group.
.PP
The following Linux-specific
.I options
are for use with children created using
.BR clone (2);
they can also, since Linux 4.7,
.\" commit 91c4e8ea8f05916df0c8a6f383508ac7c9e10dba
be used with
.BR waitid ():
.TP
.B __WCLONE
.\" since 0.99pl10
Wait for "clone" children only.
If omitted, then wait for "non-clone" children only.
(A "clone" child is one which delivers no signal, or a signal other than
.B SIGCHLD
to its parent upon termination.)
This option is ignored if
.B __WALL
is also specified.
.TP
.BR __WALL " (since Linux 2.4)"
.\" since patch-2.3.48
Wait for all children, regardless of
type ("clone" or "non-clone").
.TP
.BR __WNOTHREAD " (since Linux 2.4)"
.\" since patch-2.4.0-test8
Do not wait for children of other threads in
the same thread group.
This was the default before Linux 2.4.
.PP
Since Linux 4.7,
.\" commit bf959931ddb88c4e4366e96dd22e68fa0db9527c
.\" prevents cases where an unreapable zombie is created if
.\" /sbin/init doesn't use __WALL.
the
.B __WALL
flag is automatically implied if the child is being ptraced.
.SS C library/kernel differences
.BR wait ()
is actually a library function that (in glibc) is implemented as a call to
.BR wait4 (2).
.PP
On some architectures, there is no
.BR waitpid ()
system call;
.\" e.g., i386 has the system call, but not x86-64
instead, this interface is implemented via a C library
wrapper function that calls
.BR wait4 (2).
.PP
The raw
.BR waitid ()
system call takes a fifth argument, of type
.IR "struct rusage\ *" .
If this argument is non-NULL,
then it is used to return resource usage information about the child,
in the same manner as
.BR wait4 (2).
See
.BR getrusage (2)
for details.
.SH BUGS
According to POSIX.1-2008, an application calling
.BR waitid ()
must ensure that
.I infop
points to a
.I siginfo_t
structure (i.e., that it is a non-null pointer).
On Linux, if
.I infop
is NULL,
.BR waitid ()
succeeds, and returns the process ID of the waited-for child.
Applications should avoid relying on this inconsistent,
nonstandard, and unnecessary feature.
.SH EXAMPLE
.\" fork.2 refers to this example program.
The following program demonstrates the use of
.BR fork (2)
and
.BR waitpid ().
The program creates a child process.
If no command-line argument is supplied to the program,
then the child suspends its execution using
.BR pause (2),
to allow the user to send signals to the child.
Otherwise, if a command-line argument is supplied,
then the child exits immediately,
using the integer supplied on the command line as the exit status.
The parent process executes a loop that monitors the child using
.BR waitpid (),
and uses the W*() macros described above to analyze the wait status value.
.PP
The following shell session demonstrates the use of the program:
.PP
.in +4n
.EX
.RB "$" " ./a.out &"
Child PID is 32360
[1] 32359
.RB "$" " kill \-STOP 32360"
stopped by signal 19
.RB "$" " kill \-CONT 32360"
continued
.RB "$" " kill \-TERM 32360"
killed by signal 15
[1]+  Done                    ./a.out
$
.EE
.in
.SS Program source
\&
.EX
#include <sys/wait.h>
#include <stdlib.h>
#include <unistd.h>
#include <stdio.h>

int
main(int argc, char *argv[])
{
    pid_t cpid, w;
    int wstatus;

    cpid = fork();
    if (cpid == \-1) {
        perror("fork");
        exit(EXIT_FAILURE);
    }

    if (cpid == 0) {            /* Code executed by child */
        printf("Child PID is %ld\en", (long) getpid());
        if (argc == 1)
            pause();                    /* Wait for signals */
        _exit(atoi(argv[1]));

    } else {                    /* Code executed by parent */
        do {
            w = waitpid(cpid, &wstatus, WUNTRACED | WCONTINUED);
            if (w == \-1) {
                perror("waitpid");
                exit(EXIT_FAILURE);
            }

            if (WIFEXITED(wstatus)) {
                printf("exited, status=%d\en", WEXITSTATUS(wstatus));
            } else if (WIFSIGNALED(wstatus)) {
                printf("killed by signal %d\en", WTERMSIG(wstatus));
            } else if (WIFSTOPPED(wstatus)) {
                printf("stopped by signal %d\en", WSTOPSIG(wstatus));
            } else if (WIFCONTINUED(wstatus)) {
                printf("continued\en");
            }
        } while (!WIFEXITED(wstatus) && !WIFSIGNALED(wstatus));
        exit(EXIT_SUCCESS);
    }
}
.EE
.SH SEE ALSO
.BR _exit (2),
.BR clone (2),
.BR fork (2),
.BR kill (2),
.BR ptrace (2),
.BR sigaction (2),
.BR signal (2),
.BR wait4 (2),
.BR pthread_create (3),
.BR core (5),
.BR credentials (7),
.BR signal (7)