man2/poll.2

.\" Copyright (C) 2006, 2019 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.\" manual page may be incorrect or out-of-date.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
.\"
.\" Additions from Richard Gooch <rgooch@atnf.CSIRO.AU> and aeb, 971207
.\" 2006-03-13, mtk, Added ppoll() + various other rewordings
.\" 2006-07-01, mtk, Added POLLRDHUP + various other wording and
.\"	formatting changes.
.\"
.TH POLL 2 2020-04-11 "Linux" "Linux Programmer's Manual"
.SH NAME
poll, ppoll \- wait for some event on a file descriptor
.SH SYNOPSIS
.nf
.B #include <poll.h>
.PP
.BI "int poll(struct pollfd *" fds ", nfds_t " nfds ", int " timeout );

.BR "#define _GNU_SOURCE" "         /* See feature_test_macros(7) */"
.B #include <signal.h>
.B #include <poll.h>
.PP
.BI "int ppoll(struct pollfd *" fds ", nfds_t " nfds ,
.BI "        const struct timespec *" tmo_p ", const sigset_t *" sigmask );
.fi
.SH DESCRIPTION
.BR poll ()
performs a similar task to
.BR select (2):
it waits for one of a set of file descriptors to become ready
to perform I/O.
The Linux-specific
.BR epoll (7)
API performs a similar task, but offers features beyond those found in
.BR poll ().
.PP
The set of file descriptors to be monitored is specified in the
.I fds
argument, which is an array of structures of the following form:
.PP
.in +4n
.EX
struct pollfd {
    int   fd;         /* file descriptor */
    short events;     /* requested events */
    short revents;    /* returned events */
};
.EE
.in
.PP
The caller should specify the number of items in the
.I fds
array in
.IR nfds .
.PP
The field
.I fd
contains a file descriptor for an open file.
If this field is negative, then the corresponding
.I events
field is ignored and the
.I revents
field returns zero.
(This provides an easy way of ignoring a
file descriptor for a single
.BR poll ()
call: simply negate the
.I fd
field.
Note, however, that this technique can't be used to ignore file descriptor 0.)
.PP
The field
.I events
is an input parameter, a bit mask specifying the events the application
is interested in for the file descriptor
.IR fd .
This field may be specified as zero,
in which case the only events that can be returned in
.I revents
are
.BR POLLHUP ,
.BR POLLERR ,
and
.B POLLNVAL
(see below).
.PP
The field
.I revents
is an output parameter, filled by the kernel with the events that
actually occurred.
The bits returned in
.I revents
can include any of those specified in
.IR events ,
or one of the values
.BR POLLERR ,
.BR POLLHUP ,
or
.BR POLLNVAL .
(These three bits are meaningless in the
.I events
field, and will be set in the
.I revents
field whenever the corresponding condition is true.)
.PP
If none of the events requested (and no error) has occurred for any
of the file descriptors, then
.BR poll ()
blocks until one of the events occurs.
.PP
The
.I timeout
argument specifies the number of milliseconds that
.BR poll ()
should block waiting for a file descriptor to become ready.
The call will block until either:
.IP \(bu 2
a file descriptor becomes ready;
.IP \(bu
the call is interrupted by a signal handler; or
.IP \(bu
the timeout expires.
.PP
Note that the
.I timeout
interval will be rounded up to the system clock granularity,
and kernel scheduling delays mean that the blocking interval
may overrun by a small amount.
Specifying a negative value in
.I timeout
means an infinite timeout.
Specifying a
.I timeout
of zero causes
.BR poll ()
to return immediately, even if no file descriptors are ready.
.PP
The bits that may be set/returned in
.I events
and
.I revents
are defined in \fI<poll.h>\fP:
.TP
.B POLLIN
There is data to read.
.TP
.B POLLPRI
There is some exceptional condition on the file descriptor.
Possibilities include:
.RS
.IP \(bu 2
There is out-of-band data on a TCP socket (see
.BR tcp (7)).
.IP \(bu
A pseudoterminal device in packet mode has seen a state change on
the corresponding terminal device
(see
.BR ioctl_tty (2)).
.IP \(bu
A
.I cgroup.events
file has been modified (see
.BR cgroups (7)).
.RE
.TP
.B POLLOUT
Writing is now possible, though a write larger than the available space
in a socket or pipe will still block (unless
.B O_NONBLOCK
is set).
.TP
.BR POLLRDHUP " (since Linux 2.6.17)"
Stream socket peer closed connection,
or shut down writing half of connection.
The
.B _GNU_SOURCE
feature test macro must be defined
(before including
.I any
header files)
in order to obtain this definition.
.TP
.B POLLERR
Error condition (only returned in
.IR revents ;
ignored in
.IR events ).
This bit is also set for a file descriptor referring
to the write end of a pipe when the read end has been closed.
.TP
.B POLLHUP
Hang up (only returned in
.IR revents ;
ignored in
.IR events ).
Note that when reading from a channel such as a pipe or a stream socket,
this event merely indicates that the peer closed its end of the channel.
Subsequent reads from the channel will return 0 (end of file)
only after all outstanding data in the channel has been consumed.
.TP
.B POLLNVAL
Invalid request:
.I fd
not open (only returned in
.IR revents ;
ignored in
.IR events ).
.PP
When compiling with
.B _XOPEN_SOURCE
defined, one also has the following,
which convey no further information beyond the bits listed above:
.TP
.B POLLRDNORM
Equivalent to
.BR POLLIN .
.TP
.B POLLRDBAND
Priority band data can be read (generally unused on Linux).
.\" POLLRDBAND is used in the DECnet protocol.
.TP
.B POLLWRNORM
Equivalent to
.BR POLLOUT .
.TP
.B POLLWRBAND
Priority data may be written.
.PP
Linux also knows about, but does not use
.BR POLLMSG .
.SS ppoll()
The relationship between
.BR poll ()
and
.BR ppoll ()
is analogous to the relationship between
.BR select (2)
and
.BR pselect (2):
like
.BR pselect (2),
.BR ppoll ()
allows an application to safely wait until either a file descriptor
becomes ready or until a signal is caught.
.PP
Other than the difference in the precision of the
.I timeout
argument, the following
.BR ppoll ()
call:
.PP
.in +4n
.EX
ready = ppoll(&fds, nfds, tmo_p, &sigmask);
.EE
.in
.PP
is nearly equivalent to
.I atomically
executing the following calls:
.PP
.in +4n
.EX
sigset_t origmask;
int timeout;

timeout = (tmo_p == NULL) ? \-1 :
          (tmo_p\->tv_sec * 1000 + tmo_p\->tv_nsec / 1000000);
pthread_sigmask(SIG_SETMASK, &sigmask, &origmask);
ready = poll(&fds, nfds, timeout);
pthread_sigmask(SIG_SETMASK, &origmask, NULL);
.EE
.in
.PP
The above code segment is described as
.I nearly
equivalent because whereas a negative
.I timeout
value for
.BR poll ()
is interpreted as an infinite timeout, a negative value expressed in
.IR *tmo_p
results in an error from
.BR ppoll ().
.PP
See the description of
.BR pselect (2)
for an explanation of why
.BR ppoll ()
is necessary.
.PP
If the
.I sigmask
argument is specified as NULL, then
no signal mask manipulation is performed
(and thus
.BR ppoll ()
differs from
.BR poll ()
only in the precision of the
.I timeout
argument).
.PP
The
.I tmo_p
argument specifies an upper limit on the amount of time that
.BR ppoll ()
will block.
This argument is a pointer to a structure of the following form:
.PP
.in +4n
.EX
struct timespec {
    long    tv_sec;         /* seconds */
    long    tv_nsec;        /* nanoseconds */
};
.EE
.in
.PP
If
.I tmo_p
is specified as NULL, then
.BR ppoll ()
can block indefinitely.
.SH RETURN VALUE
On success,
.BR poll ()
returns a nonnegative value which is the number of elements in the
.I pollfds
whose
.I revents
fields have been set to a nonzero value (indicating an event or an error).
A return value of zero indicates that the system call timed out
before any file descriptors became read.
.PP
On error, \-1 is returned, and
.I errno
is set to indicate the cause of the error.
.SH ERRORS
.TP
.B EFAULT
.I fds
points outside the process's accessible address space.
The array given as argument was not contained in the calling program's
address space.
.TP
.B EINTR
A signal occurred before any requested event; see
.BR signal (7).
.TP
.B EINVAL
The
.I nfds
value exceeds the
.B RLIMIT_NOFILE
value.
.TP
.B EINVAL
.RB ( ppoll ())
The timeout value expressed in
.IR *ip
is invalid (negative).
.TP
.B ENOMEM
Unable to allocate memory for kernel data structures.
.SH VERSIONS
The
.BR poll ()
system call was introduced in Linux 2.1.23.
On older kernels that lack this system call,
the glibc
.BR poll ()
wrapper function provides emulation using
.BR select (2).
.PP
The
.BR ppoll ()
system call was added to Linux in kernel 2.6.16.
The
.BR ppoll ()
library call was added in glibc 2.4.
.SH CONFORMING TO
.BR poll ()
conforms to POSIX.1-2001 and POSIX.1-2008.
.BR ppoll ()
is Linux-specific.
.\" FIXME .
.\" ppoll() is proposed for inclusion in POSIX:
.\" https://www.austingroupbugs.net/view.php?id=1263
.\" NetBSD 3.0 has a pollts() which is like Linux ppoll().
.SH NOTES
The operation of
.BR poll ()
and
.BR ppoll ()
is not affected by the
.BR O_NONBLOCK
flag.
.PP
On some other UNIX systems,
.\" Darwin, according to a report by Jeremy Sequoia, relayed by Josh Triplett
.BR poll ()
can fail with the error
.B EAGAIN
if the system fails to allocate kernel-internal resources, rather than
.B ENOMEM
as Linux does.
POSIX permits this behavior.
Portable programs may wish to check for
.B EAGAIN
and loop, just as with
.BR EINTR .
.PP
Some implementations define the nonstandard constant
.B INFTIM
with the value \-1 for use as a
.IR timeout
for
.BR poll ().
This constant is not provided in glibc.
.PP
For a discussion of what may happen if a file descriptor being monitored by
.BR poll ()
is closed in another thread, see
.BR select (2).
.SS C library/kernel differences
The Linux
.BR ppoll ()
system call modifies its
.I tmo_p
argument.
However, the glibc wrapper function hides this behavior
by using a local variable for the timeout argument that
is passed to the system call.
Thus, the glibc
.BR ppoll ()
function does not modify its
.I tmo_p
argument.
.PP
The raw
.BR ppoll ()
system call has a fifth argument,
.IR "size_t sigsetsize" ,
which specifies the size in bytes of the
.IR sigmask
argument.
The glibc
.BR ppoll ()
wrapper function specifies this argument as a fixed value
(equal to
.IR sizeof(kernel_sigset_t) ).
See
.BR sigprocmask (2)
for a discussion on the differences between the kernel and the libc
notion of the sigset.
.SH BUGS
See the discussion of spurious readiness notifications under the
BUGS section of
.BR select (2).
.SH EXAMPLES
The program below opens each of the files named in its command-line
arguments and monitors the resulting file descriptors for readiness to read
.RB ( POLLIN ).
The program loops, repeatedly using
.BR poll ()
to monitor the file descriptors,
printing the number of ready file descriptors on return.
For each ready file descriptor, the program:
.IP \(bu 2
displays the returned
.I revents
field in a human-readable form;
.IP \(bu
if the file descriptor is readable, reads some data from it,
and displays that data on standard output; and
.IP \(bu
if the file descriptors was not readable,
but some other event occurred (presumably
.BR POLLHUP ),
closes the file descriptor.
.PP
Suppose we run the program in one terminal, asking it to open a FIFO:
.PP
.in +4n
.EX
$ \fBmkfifo myfifo\fP
$ \fB./poll_input myfifo\fP
.EE
.in
.PP
In a second terminal window, we then open the FIFO for writing,
write some data to it, and close the FIFO:
.PP
.in +4n
.EX
$ \fBecho aaaaabbbbbccccc > myfifo\fP
.EE
.in
.PP
In the terminal where we are running the program, we would then see:
.PP
.in +4n
.EX
Opened "myfifo" on fd 3
About to poll()
Ready: 1
  fd=3; events: POLLIN POLLHUP
    read 10 bytes: aaaaabbbbb
About to poll()
Ready: 1
  fd=3; events: POLLIN POLLHUP
    read 6 bytes: ccccc

About to poll()
Ready: 1
  fd=3; events: POLLHUP
    closing fd 3
All file descriptors closed; bye
.EE
.in
.PP
In the above output, we see that
.BR poll ()
returned three times:
.IP \(bu 2
On the first return, the bits returned in the
.I revents
field were
.BR POLLIN ,
indicating that the file descriptor is readable, and
.BR POLLHUP ,
indicating that the other end of the FIFO has been closed.
The program then consumed some of the available input.
.IP \(bu
The second return from
.BR poll ()
also indicated
.BR POLLIN
and
.BR POLLHUP ;
the program then consumed the last of the available input.
.IP \(bu
On the final return,
.BR poll ()
indicated only
.BR POLLHUP
on the FIFO,
at which point the file descriptor was closed and the program terminated.
.\"
.SS Program source
\&
.nf
/* poll_input.c

   Licensed under GNU General Public License v2 or later.
*/
#include <poll.h>
#include <fcntl.h>
#include <sys/types.h>
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>

#define errExit(msg)    do { perror(msg); exit(EXIT_FAILURE); \e
                        } while (0)

int
main(int argc, char *argv[])
{
    int nfds, num_open_fds;
    struct pollfd *pfds;

    if (argc < 2) {
       fprintf(stderr, "Usage: %s file...\en", argv[0]);
       exit(EXIT_FAILURE);
    }

    num_open_fds = nfds = argc \- 1;
    pfds = calloc(nfds, sizeof(struct pollfd));
    if (pfds == NULL)
        errExit("malloc");

    /* Open each file on command line, and add it \(aqpfds\(aq array */

    for (int j = 0; j < nfds; j++) {
        pfds[j].fd = open(argv[j + 1], O_RDONLY);
        if (pfds[j].fd == \-1)
            errExit("open");

        printf("Opened \e"%s\e" on fd %d\en", argv[j + 1], pfds[j].fd);

        pfds[j].events = POLLIN;
    }

    /* Keep calling poll() as long as at least one file descriptor is
       open */

    while (num_open_fds > 0) {
        int ready;

        printf("About to poll()\en");
        ready = poll(pfds, nfds, \-1);
        if (ready == \-1)
            errExit("poll");

        printf("Ready: %d\en", ready);

        /* Deal with array returned by poll() */

        for (int j = 0; j < nfds; j++) {
            char buf[10];

            if (pfds[j].revents != 0) {
                printf("  fd=%d; events: %s%s%s\en", pfds[j].fd,
                        (pfds[j].revents & POLLIN)  ? "POLLIN "  : "",
                        (pfds[j].revents & POLLHUP) ? "POLLHUP " : "",
                        (pfds[j].revents & POLLERR) ? "POLLERR " : "");

                if (pfds[j].revents & POLLIN) {
                    ssize_t s = read(pfds[j].fd, buf, sizeof(buf));
                    if (s == \-1)
                        errExit("read");
                    printf("    read %zd bytes: %.*s\en",
                            s, (int) s, buf);
                } else {                /* POLLERR | POLLHUP */
                    printf("    closing fd %d\en", pfds[j].fd);
                    if (close(pfds[j].fd) == \-1)
                        errExit("close");
                    num_open_fds\-\-;
                }
            }
        }
    }

    printf("All file descriptors closed; bye\en");
    exit(EXIT_SUCCESS);
}
.fi
.SH SEE ALSO
.BR restart_syscall (2),
.BR select (2),
.BR select_tut (2),
.BR epoll (7),
.BR time (7)