xref: /netbsd-src/lib/libc/sys/kqueue.2 (revision 627f7eb200a4419d89b531d55fccd2ee3ffdcde0)

.\"	$NetBSD: kqueue.2,v 1.53 2020/10/31 14:35:28 christos Exp $
.\"
.\" Copyright (c) 2000 Jonathan Lemon
.\" All rights reserved.
.\"
.\" Copyright (c) 2001, 2002, 2003 The NetBSD Foundation, Inc.
.\" All rights reserved.
.\"
.\" Portions of this documentation is derived from text contributed by
.\" Luke Mewburn.
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions
.\" are met:
.\" 1. Redistributions of source code must retain the above copyright
.\"    notice, this list of conditions and the following disclaimer.
.\" 2. Redistributions in binary form must reproduce the above copyright
.\"    notice, this list of conditions and the following disclaimer in the
.\"    documentation and/or other materials provided with the distribution.
.\"
.\" THIS SOFTWARE IS PROVIDED ``AS IS'' AND
.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
.\" SUCH DAMAGE.
.\"
.\" $FreeBSD: src/lib/libc/sys/kqueue.2,v 1.22 2001/06/27 19:55:57 dd Exp $
.\"
.Dd October 30, 2020
.Dt KQUEUE 2
.Os
.Sh NAME
.Nm kqueue ,
.Nm kqueue1 ,
.Nm kevent ,
.Nm EV_SET
.Nd kernel event notification mechanism
.Sh LIBRARY
.Lb libc
.Sh SYNOPSIS
.In sys/event.h
.In sys/time.h
.Ft int
.Fn kqueue "void"
.Ft int
.Fn kqueue1 "int flags"
.Ft int
.Fn kevent "int kq" "const struct kevent *changelist" "size_t nchanges" "struct kevent *eventlist" "size_t nevents" "const struct timespec *timeout"
.Fn EV_SET "&kev" ident filter flags fflags data udata
.Sh DESCRIPTION
The
.Fn kqueue
system call
provides a generic method of notifying the user when an event
happens or a condition holds, based on the results of small
pieces of kernel code termed filters.
A kevent is identified by the (ident, filter) pair; there may only
be one unique kevent per kqueue.
.Pp
The filter is executed upon the initial registration of a kevent
in order to detect whether a preexisting condition is present, and is also
executed whenever an event is passed to the filter for evaluation.
If the filter determines that the condition should be reported,
then the kevent is placed on the kqueue for the user to retrieve.
.Pp
The filter is also run when the user attempts to retrieve the kevent
from the kqueue.
If the filter indicates that the condition that triggered
the event no longer holds, the kevent is removed from the kqueue and
is not returned.
.Pp
Multiple events which trigger the filter do not result in multiple
kevents being placed on the kqueue; instead, the filter will aggregate
the events into a single struct kevent.
Calling
.Xr close 2
on a file descriptor will remove any kevents that reference the descriptor.
.Pp
The
.Fn kqueue
system call
creates a new kernel event queue and returns a descriptor.
.Pp
The
.Fn kqueue1
system call also allows to set the following
.Fa flags
on the returned file descriptor:
.Bl -column O_NONBLOCK -offset indent
.It Dv O_CLOEXEC
Set the close on exec property.
.It Dv O_NONBLOCK
Set non-blocking I/O.
.It Dv O_NOSIGPIPE
Return
.Er EPIPE
instead of raising
.Dv SIGPIPE .
.El
.Pp
The queue is not inherited by a child created with
.Xr fork 2 .
.\" However, if
.\" .Xr rfork 2
.\" is called without the
.\" .Dv RFFDG
.\" flag, then the descriptor table is shared,
.\" which will allow sharing of the kqueue between two processes.
.Pp
The
.Fn kevent
system call
is used to register events with the queue, and return any pending
events to the user.
The
.Fa changelist
argument
is a pointer to an array of
.Va kevent
structures, as defined in
.In sys/event.h .
All changes contained in the
.Fa changelist
are applied before any pending events are read from the queue.
The
.Fa nchanges
argument
gives the size of
.Fa changelist .
The
.Fa eventlist
argument
is a pointer to an array of kevent structures.
The
.Fa nevents
argument
determines the size of
.Fa eventlist .
When
.Fa nevents
is zero,
.Fn kevent
will return immediately even if there is a
.Fa timeout
specified unlike
.Xr select 2 .
If
.Fa timeout
is a
.No non- Ns Dv NULL
pointer, it specifies a maximum interval to wait
for an event, which will be interpreted as a
.Li struct timespec .
If
.Fa timeout
is a
.Dv NULL
pointer,
.Fn kevent
waits indefinitely.
To effect a poll, the
.Fa timeout
argument should be
.No non- Ns Dv NULL ,
pointing to a zero-valued
.Xr timespec 3
structure.
The same array may be used for the
.Fa changelist
and
.Fa eventlist .
.Pp
The
.Fn EV_SET
macro is provided for ease of initializing a kevent structure.
This macro does not evaluate its parameters multiple times.
.Pp
The
.Va kevent
structure is defined as:
.Bd -literal
struct kevent {
	uintptr_t ident;	/* identifier for this event */
	uint32_t  filter;	/* filter for event */
	uint32_t  flags;	/* action flags for kqueue */
	uint32_t  fflags;	/* filter flag value */
	int64_t   data;		/* filter data value */
	void     *udata;	/* opaque user data identifier */
};
.Ed
.Pp
The fields of
.Fa struct kevent
are:
.Bl -tag -width "Fa filter" -offset indent
.It ident
Value used to identify this event.
The exact interpretation is determined by the attached filter,
but often is a file descriptor.
.It Fa filter
Identifies the kernel filter used to process this event.
There are pre-defined system filters (which are described below), and
other filters may be added by kernel subsystems as necessary.
.It Fa flags
Actions to perform on the event.
.It Fa fflags
Filter-specific flags.
.It Fa data
Filter-specific data value.
.It Fa udata
Opaque user-defined value passed through the kernel unchanged.
.El
.Pp
The
.Va flags
field can contain the following values:
.Bl -tag -width XXXEV_ONESHOT -offset indent
.It Dv EV_ADD
Adds the event to the kqueue.
Re-adding an existing event will modify the parameters of the original
event, and not result in a duplicate entry.
Adding an event automatically enables it,
unless overridden by the EV_DISABLE flag.
.It Dv EV_ENABLE
Permit
.Fn kevent
to return the event if it is triggered.
.It Dv EV_DISABLE
Disable the event so
.Fn kevent
will not return it.
The filter itself is not disabled.
.It Dv EV_DISPATCH
Disable the event source immediately after delivery of an event.
See
.Dv EV_DISABLE
above.
.It Dv EV_DELETE
Removes the event from the kqueue.
Events which are attached to file descriptors are automatically deleted
on the last close of the descriptor.
.It Dv EV_RECEIPT
This flag is useful for making bulk changes to a kqueue without draining
any pending events.
When passed as input, it forces
.Dv EV_ERROR
to always be returned.
When a filter is successfully added the
.Va data
field will be zero.
Note that if this flag is encountered and there is no remaining space in
.Fa eventlist
to hold the
.Dv EV_ERROR
event, then subsequent changes will not get processed.
.It Dv EV_ONESHOT
Causes the event to return only the first occurrence of the filter
being triggered.
After the user retrieves the event from the kqueue, it is deleted.
.It Dv EV_CLEAR
After the event is retrieved by the user, its state is reset.
This is useful for filters which report state transitions
instead of the current state.
Note that some filters may automatically set this flag internally.
.It Dv EV_EOF
Filters may set this flag to indicate filter-specific EOF condition.
.It Dv EV_ERROR
See
.Sx RETURN VALUES
below.
.El
.Ss Filters
Filters are identified by a number.
There are two types of filters; pre-defined filters which
are described below, and third-party filters that may be added with
.Xr kfilter_register 9
by kernel sub-systems, third-party device drivers, or loadable
kernel modules.
.Pp
As a third-party filter is referenced by a well-known name instead
of a statically assigned number, two
.Xr ioctl 2 Ns s
are supported on the file descriptor returned by
.Fn kqueue
to map a filter name to a filter number, and vice-versa (passing
arguments in a structure described below):
.Bl -tag -width KFILTER_BYFILTER -offset indent
.It Dv KFILTER_BYFILTER
Map
.Va filter
to
.Va name ,
which is of size
.Va len .
.It Dv KFILTER_BYNAME
Map
.Va name
to
.Va filter .
.Va len
is ignored.
.El
.Pp
The following structure is used to pass arguments in and out of the
.Xr ioctl 2 :
.Bd -literal -offset indent
struct kfilter_mapping {
	char	 *name;		/* name to lookup or return */
	size_t	 len;		/* length of name */
	uint32_t filter;	/* filter to lookup or return */
};
.Ed
.Pp
The predefined system filters are listed below.
Arguments may be passed to and from the filter via the
.Va fflags
and
.Va data
fields in the kevent structure.
.Pp
The predefined system filters are:
.Bl -tag -width EVFILT_SIGNAL
.It Dv EVFILT_READ
Takes a descriptor as the identifier, and returns whenever
there is data available to read.
The behavior of the filter is slightly different depending
on the descriptor type.
.Bl -tag -width 2n
.It Sockets
Sockets which have previously been passed to
.Xr listen 2
return when there is an incoming connection pending.
.Va data
contains the size of the listen backlog (i.e., the number of
connections ready to be accepted with
.Xr accept 2 . )
.Pp
Other socket descriptors return when there is data to be read,
subject to the
.Dv SO_RCVLOWAT
value of the socket buffer.
This may be overridden with a per-filter low water mark at the
time the filter is added by setting the
NOTE_LOWAT
flag in
.Va fflags ,
and specifying the new low water mark in
.Va data .
On return,
.Va data
contains the number of bytes in the socket buffer.
.Pp
If the read direction of the socket has shutdown, then the filter
also sets EV_EOF in
.Va flags ,
and returns the socket error (if any) in
.Va fflags .
It is possible for EOF to be returned (indicating the connection is gone)
while there is still data pending in the socket buffer.
.It Vnodes
Returns when the file pointer is not at the end of file.
.Va data
contains the offset from current position to end of file,
and may be negative.
.\" .Pp
.\" This behavior is different from
.\" .Xr poll 2 ,
.\" where read events are triggered for regular files unconditionally.
.\" This event can be triggered unconditionally by setting the
.\" .Dv NOTE_FILE_POLL
.\" flag in
.\" .Va fflags .
.It "Fifos, Pipes"
Returns when there is data to read;
.Va data
contains the number of bytes available.
.Pp
When the last writer disconnects, the filter will set EV_EOF in
.Va flags .
This may be cleared by passing in EV_CLEAR, at which point the
filter will resume waiting for data to become available before
returning.
.It "BPF devices"
Returns when the BPF buffer is full, the BPF timeout has expired, or
when the BPF has
.Dq immediate mode
enabled and there is any data to read;
.Va data
contains the number of bytes available.
.El
.It Dv EVFILT_WRITE
Takes a descriptor as the identifier, and returns whenever
it is possible to write to the descriptor.
For sockets, pipes, fifos, and ttys,
.Va data
will contain the amount of space remaining in the write buffer.
The filter will set EV_EOF when the reader disconnects, and for
the fifo case, this may be cleared by use of EV_CLEAR.
Note that this filter is not supported for vnodes.
.Pp
For sockets, the low water mark and socket error handling is
identical to the EVFILT_READ case.
.It Dv EVFILT_AIO
This is not implemented in
.Nx .
.ig
The sigevent portion of the AIO request is filled in, with
.Va sigev_notify_kqueue
containing the descriptor of the kqueue that the event should
be attached to,
.Va sigev_value
containing the udata value, and
.Va sigev_notify
set to SIGEV_EVENT.
When the aio_* function is called, the event will be registered
with the specified kqueue, and the
.Va ident
argument set to the
.Fa struct aiocb
returned by the aio_* function.
The filter returns under the same conditions as aio_error.
.Pp
Alternatively, a kevent structure may be initialized, with
.Va ident
containing the descriptor of the kqueue, and the
address of the kevent structure placed in the
.Va aio_lio_opcode
field of the AIO request.
However, this approach will not work on
architectures with 64-bit pointers, and should be considered deprecated.
..
.It Dv EVFILT_VNODE
Takes a file descriptor as the identifier and the events to watch for in
.Va fflags ,
and returns when one or more of the requested events occurs on the descriptor.
The events to monitor are:
.Bl -tag -width XXNOTE_RENAME
.It Dv NOTE_DELETE
.Xr unlink 2
was called on the file referenced by the descriptor.
.It Dv NOTE_WRITE
A write occurred on the file referenced by the descriptor.
.It Dv NOTE_EXTEND
The file referenced by the descriptor was extended.
.It Dv NOTE_ATTRIB
The file referenced by the descriptor had its attributes changed.
.It Dv NOTE_LINK
The link count on the file changed.
.It Dv NOTE_RENAME
The file referenced by the descriptor was renamed.
.It Dv NOTE_REVOKE
Access to the file was revoked via
.Xr revoke 2
or the underlying file system was unmounted.
.El
.Pp
On return,
.Va fflags
contains the events which triggered the filter.
.It Dv EVFILT_PROC
Takes the process ID to monitor as the identifier and the events to watch for
in
.Va fflags ,
and returns when the process performs one or more of the requested events.
If a process can normally see another process, it can attach an event to it.
The events to monitor are:
.Bl -tag -width XXNOTE_TRACKERR
.It Dv NOTE_EXIT
The process has exited.
The exit code of the process is stored in
.Va data .
.It Dv NOTE_FORK
The process has called
.Xr fork 2 .
.It Dv NOTE_EXEC
The process has executed a new process via
.Xr execve 2
or similar call.
.It Dv NOTE_TRACK
Follow a process across
.Xr fork 2
calls.
The parent process will return with NOTE_TRACK set in the
.Va fflags
field, while the child process will return with NOTE_CHILD set in
.Va fflags
and the parent PID in
.Va data .
.It Dv NOTE_TRACKERR
This flag is returned if the system was unable to attach an event to
the child process, usually due to resource limitations.
.El
.Pp
On return,
.Va fflags
contains the events which triggered the filter.
.It Dv EVFILT_SIGNAL
Takes the signal number to monitor as the identifier and returns
when the given signal is delivered to the current process.
This coexists with the
.Xr signal 3
and
.Xr sigaction 2
facilities, and has a lower precedence.
The filter will record
all attempts to deliver a signal to a process, even if the signal has
been marked as SIG_IGN.
Event notification happens after normal signal delivery processing.
.Va data
returns the number of times the signal has occurred since the last call to
.Fn kevent .
This filter automatically sets the EV_CLEAR flag internally.
.It Dv EVFILT_TIMER
Establishes an arbitrary timer identified by
.Va ident .
When adding a timer,
.Va data
specifies the timeout period in milliseconds.
The timer will be periodic unless EV_ONESHOT is specified.
On return,
.Va data
contains the number of times the timeout has expired since the last call to
.Fn kevent .
This filter automatically sets the EV_CLEAR flag internally.
.It Dv EVFILT_FS
Establishes a file system monitor.
Currently it only monitors file system mount and unmount actions.
.It Dv EVFILT_USER
Establishes a user event identified by
.Va ident
which is not associated with any kernel mechanism but is triggered by
user level code.
The lower 24 bits of the
.Va fflags
may be used for user defined flags and manipulated using the following:
.Bl -tag -width "Dv NOTE_FFLAGSMASK"
.It Dv NOTE_FFNOP
Ignore the input
.Va fflags .
.It Dv NOTE_FFAND
Bitwise AND
.Va fflags .
.It Dv NOTE_FFOR
Bitwise OR
.Va fflags .
.It Dv NOTE_FFCOPY
Copy
.Va fflags .
.It Dv NOTE_FFCTRLMASK
Control mask for
.Va fflags .
.It Dv NOTE_FFLAGSMASK
User defined flag mask for
.Va fflags .
.El
.Pp
A user event is triggered for output with the following:
.Bl -tag -width "Dv NOTE_FFLAGSMASK"
.It Dv NOTE_TRIGGER
Cause the event to be triggered.
.El
.Pp
On return,
.Va fflags
contains the users defined flags in the lower 24 bits.
.El
.Sh CANCELLATION BEHAVIOUR
If
.Fa nevents
is non-zero, i.e., the function is potentially blocking, the call
is a cancellation point.
Otherwise, i.e., if
.Fa nevents
is zero, the call is not cancellable.
Cancellation can only occur before any changes are made to the kqueue,
or when the call was blocked and no changes to the queue were requested.
.Sh RETURN VALUES
The
.Fn kqueue
system call
creates a new kernel event queue and returns a file descriptor.
If there was an error creating the kernel event queue, a value of \-1 is
returned and
.Dv errno
is set.
.Pp
The
.Fn kevent
system call
returns the number of events placed in the
.Fa eventlist ,
up to the value given by
.Fa nevents .
If an error occurs while processing an element of the
.Fa changelist
and there is enough room in the
.Fa eventlist ,
then the event will be placed in the
.Fa eventlist
with
.Dv EV_ERROR
set in
.Va flags
and the system error in
.Va data .
Otherwise,
.Dv \-1
will be returned, and
.Dv errno
will be set to indicate the error condition.
If the time limit expires, then
.Fn kevent
returns 0.
.Sh EXAMPLES
The following example program monitors a file (provided to it as the first
argument) and prints information about some common events it receives
notifications for:
.Bd -literal -offset indent
#include <sys/types.h>
#include <sys/event.h>
#include <sys/time.h>
#include <stdio.h>
#include <unistd.h>
#include <stdlib.h>
#include <fcntl.h>
#include <err.h>

int
main(int argc, char *argv[])
{
        int fd, kq, nev;
        struct kevent ev;
        static const struct timespec tout = { 1, 0 };

        if ((fd = open(argv[1], O_RDONLY)) == -1)
                err(1, "Cannot open `%s'", argv[1]);

        if ((kq = kqueue()) == -1)
                err(1, "Cannot create kqueue");

        EV_SET(&ev, fd, EVFILT_VNODE, EV_ADD | EV_ENABLE | EV_CLEAR,
            NOTE_DELETE|NOTE_WRITE|NOTE_EXTEND|NOTE_ATTRIB|NOTE_LINK|
            NOTE_RENAME|NOTE_REVOKE, 0, 0);
        if (kevent(kq, &ev, 1, NULL, 0, &tout) == -1)
                err(1, "kevent");
        for (;;) {
                nev = kevent(kq, NULL, 0, &ev, 1, &tout);
                if (nev == -1)
                        err(1, "kevent");
                if (nev == 0)
                        continue;
                if (ev.fflags & NOTE_DELETE) {
                        printf("deleted ");
                        ev.fflags &= ~NOTE_DELETE;
                }
                if (ev.fflags & NOTE_WRITE) {
                        printf("written ");
                        ev.fflags &= ~NOTE_WRITE;
                }
                if (ev.fflags & NOTE_EXTEND) {
                        printf("extended ");
                        ev.fflags &= ~NOTE_EXTEND;
                }
                if (ev.fflags & NOTE_ATTRIB) {
                        printf("chmod/chown/utimes ");
                        ev.fflags &= ~NOTE_ATTRIB;
                }
                if (ev.fflags & NOTE_LINK) {
                        printf("hardlinked ");
                        ev.fflags &= ~NOTE_LINK;
                }
                if (ev.fflags & NOTE_RENAME) {
                        printf("renamed ");
                        ev.fflags &= ~NOTE_RENAME;
                }
                if (ev.fflags & NOTE_REVOKE) {
                        printf("revoked ");
                        ev.fflags &= ~NOTE_REVOKE;
                }
                printf("\\n");
                if (ev.fflags)
                        warnx("unknown event 0x%x\\n", ev.fflags);
        }
}
.Ed
.Sh ERRORS
The
.Fn kqueue
function fails if:
.Bl -tag -width Er
.It Bq Er EMFILE
The per-process descriptor table is full.
.It Bq Er ENFILE
The system file table is full.
.It Bq Er ENOMEM
The kernel failed to allocate enough memory for the kernel queue.
.El
.Pp
The
.Fn kevent
function fails if:
.Bl -tag -width Er
.It Bq Er EACCES
The process does not have permission to register a filter.
.It Bq Er EBADF
The specified descriptor is invalid.
.It Bq Er EFAULT
There was an error reading or writing the
.Va kevent
structure.
.It Bq Er EINTR
A signal was delivered before the timeout expired and before any
events were placed on the kqueue for return.
All changes contained in the
.Fa changelist
are applied before returning this error.
.It Bq Er EINVAL
The specified time limit or filter is invalid.
.It Bq Er ENOENT
The event could not be found to be modified or deleted.
.It Bq Er ENOMEM
No memory was available to register the event.
.It Bq Er EOPNOTSUPP
This type of file descriptor is not supported for
.Fn kevent
operations.
.It Bq Er ESRCH
The specified process to attach to does not exist.
.El
.Sh SEE ALSO
.\" .Xr aio_error 2 ,
.\" .Xr aio_read 2 ,
.\" .Xr aio_return 2 ,
.Xr fork 2 ,
.Xr ioctl 2 ,
.Xr listen 2 ,
.Xr poll 2 ,
.Xr read 2 ,
.Xr select 2 ,
.Xr sigaction 2 ,
.Xr unlink 2 ,
.Xr write 2 ,
.Xr signal 3 ,
.Xr timespec 3 ,
.Xr kfilter_register 9 ,
.Xr knote 9
.Rs
.%A Jonathan Lemon
.%T "Kqueue: A Generic and Scalable Event Notification Facility"
.%I USENIX Association
.%B Proceedings of the FREENIX Track: 2001 USENIX Annual Technical Conference
.%D June 25-30, 2001
.%U http://www.usenix.org/event/usenix01/freenix01/full_papers/lemon/lemon.pdf
.Re
.Sh HISTORY
The
.Fn kqueue
and
.Fn kevent
functions first appeared in
.Fx 4.1 ,
and then in
.Nx 2.0 .
The
.Fn kqueue1
function first appeared in
.Nx 6.0 .
.Pp
The
.Fn EV_SET
macro was protected from evaluating multiple times the first argument in
.Nx 8.0 .
.Pp
The
.Va udata
type was changed from intptr_t to void * in
.Nx 10.0 .