.\" Copyright (C) 2008 Michael Kerrisk <mtk.manpages@gmail.com>
.\" starting from a version by Davide Libenzi <davidel@xmailserver.org>
.\"
.\" This program is free software; you can redistribute it and/or modify
.\" it under the terms of the GNU General Public License as published by
.\" the Free Software Foundation; either version 2 of the License, or
.\" (at your option) any later version.
.\"
.\" This program is distributed in the hope that it will be useful,
.\" but WITHOUT ANY WARRANTY; without even the implied warranty of
.\" MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
.\" GNU General Public License for more details.
.\"
.\" You should have received a copy of the GNU General Public License
.\" along with this program; if not, write to the Free Software
.\" Foundation, Inc., 59 Temple Place, Suite 330, Boston,
.\" MA  02111-1307  USA
.\"
.\" 2008-10-10, mtk: describe eventfd2(), and EFD_NONBLOCK and EFD_CLOEXEC
.\"
.TH EVENTFD 2 2008-10-10 Linux "Linux Programmer's Manual"
.SH NAME
eventfd \- create a file descriptor for event notification
.SH SYNOPSIS
.B #include <sys/eventfd.h>
.sp
.BI "int eventfd(unsigned int " initval ", int " flags );
.SH DESCRIPTION
.BR eventfd ()
creates an "eventfd object" that can be used as
an event wait/notify mechanism by userspace applications,
and by the kernel to notify userspace applications of events.
The object contains an unsigned 64-bit integer
.RI ( uint64_t )
counter that is maintained by the kernel.
This counter is initialized with the value specified in the argument
.IR initval .

Starting with Linux 2.6.27, the following values may be bitwise ORed in
.IR flags
to change the behaviour of
.BR eventfd ():
.TP 14
.B EFD_NONBLOCK
Set the
.BR O_NONBLOCK
file status flag on the new open file description.
Using this flag saves extra calls to
.BR fcntl (2)
to achieve the same result.
.TP
.B EFD_CLOEXEC
Set the close-on-exec
.RB ( FD_CLOEXEC )
flag on the new file descriptor.
See the description of the 
.B O_CLOEXEC
flag in
.BR open (2)
for reasons why this may be useful.
.PP
In Linux up to version 2.6.26, the
.I flags
argument is unused, and must be specified as zero.

As its return value,
.BR eventfd ()
returns a new file descriptor that can be used to refer to the
eventfd object.
The following operations can be performed on the file descriptor:
.TP
.BR read (2)
If the eventfd counter has a non-zero value, then a
.BR read (2)
returns 8 bytes containing that value,
and the counter's value is reset to zero.
(The returned value is in host byte order,
i.e., the native byte order for integers on the host machine.)
.IP
If the counter is zero at the time of the
.BR read (2),
then the call either blocks until the counter becomes non-zero,
or fails with the error
.B EAGAIN
if the file descriptor has been made non-blocking.
.IP
A
.BR read (2)
will fail with the error
.B EINVAL
if the size of the supplied buffer is less than 8 bytes.
.TP
.BR write (2)
A
.BR write (2)
call adds the 8-byte integer value supplied in its
buffer to the counter.
The maximum value that may be stored in the counter is the largest
unsigned 64-bit value minus 1 (i.e., 0xfffffffffffffffe).
If the addition would cause the counter's value to exceed
the maximum, then the
.BR write (2)
either blocks until a
.BR read (2)
is performed on the file descriptor,
or fails with the error
.B EAGAIN
if the file descriptor has been made non-blocking.
.IP
A
.BR write (2)
will fail with the error
.B EINVAL
if the size of the supplied buffer is less than 8 bytes,
or if an attempt is made to write the value 0xffffffffffffffff.
.TP
.BR poll "(2), " select "(2) (and similar)"
The returned file descriptor supports
.BR poll (2)
(and analogously
.BR epoll (7))
and
.BR select (2),
as follows:
.RS
.IP * 3
The file descriptor is readable
(the
.BR select (2)
.I readfds
argument; the
.BR poll (2)
.B POLLIN
flag)
if the counter has a value greater than 0.
.IP *
The file descriptor is writable
(the
.BR select (2)
.I writefds
argument; the
.BR poll (2)
.B POLLOUT
flag)
if it is possible to write a value of at least "1" without blocking.
.IP *
If an overflow of the counter value was detected,
then
.BR select (2)
indicates the file descriptor as being both readable and writable, and
.BR poll (2)
returns a
.B POLLERR
event.
As noted above,
.BR write (2)
can never overflow the counter.
However an overflow can occur if 2^64
eventfd "signal posts" were performed by the KAIO
subsystem (theoretically possible, but practically unlikely).
If an overflow has occurred, then
.BR read (2)
will return that maximum
.I uint64_t
value (i.e., 0xffffffffffffffff).
.RE
.IP
The eventfd file descriptor also supports the other file-descriptor
multiplexing APIs:
.BR pselect (2),
.BR ppoll (2),
and
.BR epoll (7).
.TP
.BR close (2)
When the file descriptor is no longer required it should be closed.
When all file descriptors associated with the same eventfd object
have been closed, the resources for object are freed by the kernel.
.PP
A copy of the file descriptor created by
.BR eventfd ()
is inherited by the child produced by
.BR fork (2).
The duplicate file descriptor is associated with the same
eventfd object.
File descriptors created by
.BR eventfd ()
are preserved across
.BR execve (2).
.SH "RETURN VALUE"
On success,
.BR eventfd ()
returns a new eventfd file descriptor.
On error, \-1 is returned and
.I errno
is set to indicate the error.
.SH ERRORS
.TP
.B EINVAL
.I flags
is invalid;
or, in Linux 2.6.26 or earlier,
.I flags
is non-zero.
.TP
.B EMFILE
The per-process limit on open file descriptors has been reached.
.TP
.B ENFILE
The system-wide limit on the total number of open files has been
reached.
.TP
.B ENODEV
.\" Note from Davide:
.\" The ENODEV error is basically never going to happen if
.\" the kernel boots correctly. That error happen only if during
.\" the kernel initialization, some error occur in the anonymous
.\" inode source initialization.
Could not mount (internal) anonymous inode device.
.TP
.B ENOMEM
There was insufficient memory to create a new
eventfd file descriptor.
.SH VERSIONS
.BR eventfd ()
is available on Linux since kernel 2.6.22.
Working support is provided in glibc since version 2.8.
.\" eventfd() is in glibc 2.7, but reportedly does not build
The
.BR eventfd2 ()
system call (see NOTES) is available on Linux since kernel 2.6.27.
.SH CONFORMING TO
.BR eventfd ()
and
.BR eventfd2 ()
are Linux-specific.
.SH NOTES
Applications can use an eventfd file descriptor instead of a pipe (see
.BR pipe (2))
in all cases where a pipe is used simply to signal events.
The kernel overhead of an eventfd file descriptor
is much lower than that of a pipe,
and only one file descriptor is
required (versus the two required for a pipe).

When used in the kernel, an eventfd
file descriptor can provide a kernel-userspace bridge allowing,
for example, functionalities like KAIO (kernel AIO)
.\" or eventually syslets/threadlets
to signal to a file descriptor that some operation is complete.

A key point about an eventfd file descriptor is that it can be
monitored just like any other file descriptor using
.BR select (2),
.BR poll (2),
or
.BR epoll (7).
This means that an application can simultaneously monitor the
readiness of "traditional" files and the readiness of other
kernel mechanisms that support the eventfd interface.
(Without the
.BR eventfd ()
interface, these mechanisms could not be multiplexed via
.BR select (2),
.BR poll (2),
or
.BR epoll (7).)
.SS Underlying Linux system calls
There are two underlying Linux system calls:
.BR eventfd ()
and the more recent
.BR eventfd2 ().
The former system call does not implement a
.I flags
argument.
The latter system call implements the
.I flags
values described above.
.\" FIXME as at 2008-10-10, the glibc snapshot doesn't seem
.\" to make use of eventfd2() yet.  Check later to see what
.\" glibc version adds support for this syscall.
The glibc wrapper function will use
.BR eventfd2 ()
where it is available.
.SS Additional glibc features
The GNU C library defines an additional type,
and two functions that attempt to abstract some of the details of
reading and writing on an eventfd file descriptor:
.in +4n
.nf

typedef uint64_t eventfd_t;

int eventfd_read(int fd, eventfd_t *value);
int eventfd_write(int fd, eventfd_t value);
.fi
.in

The functions perform the read and write operations on an
eventfd file descriptor,
returning 0 if the correct number of bytes was transferred,
or \-1 otherwise.
.SH EXAMPLE
.PP
The following program creates an eventfd file descriptor
and then forks to create a child process.
While the parent briefly sleeps,
the child writes each of the integers supplied in the program's
command-line arguments to the eventfd file descriptor.
When the parent has finished sleeping,
it reads from the eventfd file descriptor.

The following shell session shows a sample run of the program:
.in +4n
.nf

.RB "$" " ./a.out 1 2 4 7 14"
Child writing 1 to efd
Child writing 2 to efd
Child writing 4 to efd
Child writing 7 to efd
Child writing 14 to efd
Child completed write loop
Parent about to read
Parent read 28 (0x1c) from efd
.fi
.in
.nf

#include <sys/eventfd.h>
#include <unistd.h>
#include <stdlib.h>
#include <stdio.h>
#include <stdint.h>             /* Definition of uint64_t */

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

int
main(int argc, char *argv[])
{
    int efd, j;
    uint64_t u;
    ssize_t s;

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

    efd = eventfd(0, 0);
    if (efd == \-1)
        handle_error("eventfd");

    switch (fork()) {
    case 0:
        for (j = 1; j < argc; j++) {
            printf("Child writing %s to efd\\n", argv[j]);
            u = strtoull(argv[j], NULL, 0);
                    /* strtoull() allows various bases */
            s = write(efd, &u, sizeof(uint64_t));
            if (s != sizeof(uint64_t))
                handle_error("write");
        }
        printf("Child completed write loop\\n");

        exit(EXIT_SUCCESS);

    default:
        sleep(2);

        printf("Parent about to read\\n");
        s = read(efd, &u, sizeof(uint64_t));
        if (s != sizeof(uint64_t))
            handle_error("read");
        printf("Parent read %llu (0x%llx) from efd\\n",
                (unsigned long long) u, (unsigned long long) u);
        exit(EXIT_SUCCESS);

    case \-1:
        handle_error("fork");
    }
}
.fi
.SH "SEE ALSO"
.BR futex (2),
.BR pipe (2),
.BR poll (2),
.BR read (2),
.BR select (2),
.BR signalfd (2),
.BR timerfd_create (2),
.BR write (2),
.BR epoll (7),
.BR sem_overview (7)