.\"
.\" Copyright (c) 2008-2015 Apple Inc. All rights reserved.
.\"
.\" @APPLE_LICENSE_HEADER_START@
.\"
.\" This file contains Original Code and/or Modifications of Original Code
.\" as defined in and that are subject to the Apple Public Source License
.\" Version 2.0 (the 'License'). You may not use this file except in
.\" compliance with the License. Please obtain a copy of the License at
.\" http://www.opensource.apple.com/apsl/ and read it before using this
.\" file.
.\"
.\" The Original Code and all software distributed under the License are
.\" distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
.\" EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
.\" INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
.\" FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
.\" Please see the License for the specific language governing rights and
.\" limitations under the License.
.\"
.\" @APPLE_LICENSE_HEADER_END@
.\"
.\"
.\" Copyright (c) 2000 Jonathan Lemon
.\" All rights reserved.
.\"
.\" 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 1.32 2002/12/19 09:40:25 ru Exp $
.\"
.Dd October 21, 2008
.Dt KQUEUE 2
.Os
.Sh NAME
.Nm kqueue ,
.Nm kevent ,
.Nm kevent64
and
.Nm kevent_qos
.Nd kernel event notification mechanism
.Sh LIBRARY
.Lb libc
.Sh SYNOPSIS
.In sys/types.h
.In sys/event.h
.In sys/time.h
.Ft int
.Fn kqueue "void"
.Ft int
.Fn kevent "int kq" "const struct kevent *changelist" "int nchanges" "struct kevent *eventlist" "int nevents" "const struct timespec *timeout"
.Ft int
.Fn kevent64 "int kq" "const struct kevent64_s *changelist" "int nchanges" "struct kevent64_s *eventlist" "int nevents" "unsigned int flags" "const struct timespec *timeout"
.Ft int
.Fn kevent_qos "int kq" "const struct kevent_qos_s *changelist" "int nchanges" "struct kevent_qos_s *eventlist" "int nevents" "void *data_out" "size_t *data_available" "unsigned int flags"
.Fn EV_SET "&kev" ident filter flags fflags data udata
.Fn EV_SET64 "&kev" ident filter flags fflags data udata "ext[0]" "ext[1]"
.Fn EV_SET_QOS "&kev" ident filter flags qos udata fflags xflags data "ext[0]" "ext[1]" "ext[2]" "ext[3]"
.Sh DESCRIPTION
The
.Fn kqueue
system call allocates a kqueue file descriptor. This file descriptor
provides a generic method of notifying the user when a kernel
event (kevent) happens or a condition holds, based on the results
of small pieces of kernel code termed filters.
.Pp
A kevent is identified by an (ident, filter, and optional udata value)
tuple. It specifies the interesting conditions to be notified about
for that tuple. An (ident, filter, and optional udata value) tuple can
only appear once in a given kqueue.
Subsequent attempts to register the same tuple for a given kqueue
will result in the replacement of the conditions being watched,
not an addition.
Whether the udata value is considered as part of the tuple is controlled
by the EV_UDATA_SPECIFIC flag on the kevent.
.Pp
The filter identified in a kevent is executed upon the initial
registration of that event 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
.Fn close
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.
The queue is not inherited by a child created with
.Xr fork 2 .
.Pp
The
.Fn kevent,
.Fn kevent64
and
.Fn kevent_qos
system calls
are 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,
.Va kevent64_s
or
.Va kevent_qos_s
structures, as defined in
.Aq Pa 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 out
.Va kevent,
.Va kevent64_s
or
.Va kevent_qos_s
structures.
The
.Fa nevents
argument determines the size of
.Fa eventlist .
If the KEVENT_FLAG_STACK_EVENTS flag is provided on the system call,
the eventlist array is filled in in stack order (starting in the
highest available index) instead of typical array order.
The
.Fa out_data
argument provides space for extra out data provided by specific filters.
The
.Fa data_available
argument's contents specified the space available in the data pool on input,
and contains the amount still remaining on output.
If the KEVENT_FLAG_STACK_DATA flag is specified on the system call,
the data is allocated from the pool in stack order instead of typical heap order.
If
.Fa timeout
is a non-NULL pointer, it specifies a maximum interval to wait
for an event, which will be interpreted as a struct timespec. If
.Fa timeout
is a NULL pointer, both
.Fn kevent
and
.Fn kevent64
wait indefinitely. To effect a poll, the
.Fa flags
argument to
.Fn kevent64
or
.Fn kevent_qos
can include the KEVENT_FLAG_IMMEDIATE value to indicate an
immediate timeout. Alternatively, the
.Fa timeout
argument should be non-NULL, pointing to a zero-valued
.Va timespec
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
.Va kevent
structure. Similarly,
.Fn EV_SET64
initializes a
.Va kevent64_s
structure and
.Fn EV_SET_QOS
initializes a
.Va kevent_qos_s
structure.
.Pp
The
.Va kevent,
.Va kevent64_s
and
.Va kevent_qos_s
structures are defined as:
.Bd -literal
struct kevent {
uintptr_t ident; /* identifier for this event */
int16_t filter; /* filter for event */
uint16_t flags; /* general flags */
uint32_t fflags; /* filter-specific flags */
intptr_t data; /* filter-specific data */
void *udata; /* opaque user data identifier */
};
struct kevent64_s {
uint64_t ident; /* identifier for this event */
int16_t filter; /* filter for event */
uint16_t flags; /* general flags */
uint32_t fflags; /* filter-specific flags */
int64_t data; /* filter-specific data */
uint64_t udata; /* opaque user data identifier */
uint64_t ext[2]; /* filter-specific extensions */
};
struct kevent_qos_s {
uint64_t ident; /* identifier for this event */
int16_t filter; /* filter for event */
uint16_t flags; /* general flags */
uint32_t qos; /* quality of service when servicing event */
uint64_t udata; /* opaque user data identifier */
uint32_t fflags; /* filter-specific flags */
uint32_t xflags; /* extra filter-specific flags */
int64_t data; /* filter-specific data */
uint64_t ext[4]; /* filter-specific extensions */
};
.Ed
.Pp
----
.Pp
The fields of
.Fa struct kevent,
.Fa struct kevent64_s
and
.Fa struct kevent_qos_s
are:
.Bl -tag -width XXXfilter
.It ident
Value used to identify the source of the event.
The exact interpretation is determined by the attached filter,
but often is a file descriptor.
.It filter
Identifies the kernel filter used to process this event. The pre-defined
system filters are described below.
.It flags
Actions to perform on the event.
.It fflags
Filter-specific flags.
.It data
Filter-specific data value.
.It udata
Opaque user-defined value passed through the kernel unchanged. It can
optionally be part of the uniquing decision of the kevent system
.El
.Pp
In addition,
.Fa struct kevent64_s
contains:
.Bl -tag -width XXXfilter
.It ext[2]
This field stores extensions for the event's filter. What type of extension depends on
what type of filter is being used.
.El
.Pp
In addition,
.Fa struct kevent_qos_s
contains:
.Bl -tag -width XXXfilter
.It xflags
Extra filter-specific flags.
.It ext[4]
The QoS variant provides twice as many extension values for filter-specific uses.
.El
.Pp
----
.Pp
The
.Va flags
field can contain the following values:
.Bl -tag -width XXXEV_ONESHOT
.It 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 EV_ENABLE
Permit
.Fn kevent,
.Fn kevent64
and
.Fn kevent_qos
to return the event if it is triggered.
.It EV_DISABLE
Disable the event so
.Fn kevent,
.Fn kevent64
and
.Fn kevent_qos
will not return it. The filter itself is not disabled.
.It 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 EV_RECEIPT
This flag is useful for making bulk changes to a kqueue without draining any
pending events. When passed as input, it forces EV_ERROR to always be returned.
When a filter is successfully added, the
.Va data
field will be zero.
.It 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 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 EV_EOF
Filters may set this flag to indicate filter-specific EOF condition.
.It EV_OOBAND
Read filter on socket may set this flag to indicate the presence of out of
band data on the descriptor.
.It EV_ERROR
See
.Sx RETURN VALUES
below.
.El
.Pp
----
.Pp
The predefined system filters are listed below.
Arguments may be passed to and from the filter via the
.Va data,
.Va fflags
and optionally
.Va xflags
fields in the
.Va kevent,
.Va kevent64_s
or
.Va kevent_qos_s
structure.
.Bl -tag -width EVFILT_MACHPORT
.It EVFILT_READ
Takes a file 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.
.Pp
.Bl -tag -width 2n
.It Sockets
Sockets which have previously been passed to
.Fn listen
return when there is an incoming connection pending.
.Va data
contains the size of the listen backlog.
.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 .
The derived per filter low water mark value is, however, bounded
by socket receive buffer's high and low water mark values.
On return,
.Va data
contains the number of bytes of protocol data available to read.
.Pp
The presence of EV_OOBAND in
.Va flags ,
indicates the presence of out of band data on the socket
.Va data
equal to the potential number of OOB bytes availble to read.
.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.
.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 "Device nodes"
Returns when there is data to read from the device;
.Va data
contains the number of bytes available. If the device does
not support returning number of bytes, it will not allow the
filter to be attached. However, if the NOTE_LOWAT flag is
specified and the
.Va data
field contains 1 on input, those devices will attach - but
cannot be relied upon to provide an accurate count of bytes
to be read on output.
.El
.It EVFILT_WRITE
Takes a file descriptor as the identifier, and returns whenever
it is possible to write to the descriptor. For sockets, pipes
and fifos,
.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 EVFILT_AIO
This filter is currently unsupported.
.\"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_KEVENT.
.\"When the
.\".Fn aio_*
.\"system call is made, the event will be registered
.\"with the specified kqueue, and the
.\".Va ident
.\"argument set to the
.\".Fa struct aiocb
.\"returned by the
.\".Fn aio_*
.\"system call.
.\"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 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 NOTE_DELETE
The
.Fn unlink
system call
was called on the file referenced by the descriptor.
.It NOTE_WRITE
A write occurred on the file referenced by the descriptor.
.It NOTE_EXTEND
The file referenced by the descriptor was extended.
.It NOTE_ATTRIB
The file referenced by the descriptor had its attributes changed.
.It NOTE_LINK
The link count on the file changed.
.It NOTE_RENAME
The file referenced by the descriptor was renamed.
.It NOTE_REVOKE
Access to the file was revoked via
.Xr revoke 2
or the underlying fileystem was unmounted.
.El
.Pp
On return,
.Va fflags
contains the filter-specific flags which are associated with
the triggered events seen by this filter.
.It 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 NOTE_SIGNAL
.It NOTE_EXIT
The process has exited.
.It NOTE_EXITSTATUS
The process has exited and its exit status is in filter specific data. Valid only on child processes and to be used along with NOTE_EXIT.
.It NOTE_FORK
The process created a child process via
.Xr fork 2
or similar call.
.It NOTE_EXEC
The process executed a new process via
.Xr execve 2
or similar call.
.It NOTE_SIGNAL
The process was sent a signal. Status can be checked via
.Xr waitpid 2
or similar call.
.It NOTE_REAP
The process was reaped by the parent via
.Xr wait 2
or similar call. Deprecated, use NOTE_EXIT.
.El
.Pp
On return,
.Va fflags
contains the events which triggered the filter.
.It EVFILT_SIGNAL
Takes the signal number to monitor as the identifier and returns
when the given signal is generated for the process.
This coexists with the
.Fn signal
and
.Fn sigaction
facilities, and has a lower precedence. Only signals sent to the process,
not to a particular thread, will trigger the filter. 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 before normal
signal delivery processing.
.Va data
returns the number of times the signal has been generated since the last call to
.Fn kevent .
This filter automatically sets the EV_CLEAR flag internally.
.It EVFILT_MACHPORT
Takes the name of a mach port, or port set, in
.Va ident
and waits until a message is enqueued on the port or port set. When a message
is detected, but not directly received by the kevent call, the name of the
specific port where the message is enqueued is returned in
.Va data .
If
.Va fflags
contains MACH_RCV_MSG, the ext[0] and ext[1] flags are assumed to contain
a pointer to the buffer where the message is to be received and the size
of the receive buffer, respectively. If MACH_RCV_MSG is specifed, yet the
buffer size in ext[1] is zero, The space for the buffer may be carved out
of the
.Va
data_out
area provided to
.Fn kevent_qos
if there is enough space remaining there.
.It EVFILT_TIMER
Establishes an interval timer with the data
timer identified by
.Va ident .
When adding a timer,
.Va data
specifies the timeout period and
.Va fflags
can be set to one of the following:
.Bl -tag -width NOTE_ABSOLUTE
.It NOTE_SECONDS
data is in seconds
.It NOTE_USECONDS
data is in microseconds
.It NOTE_NSECONDS
data is in nanoseconds
.It NOTE_ABSOLUTE
data is an absolute timeout
.It NOTE_CRITICAL
system makes a best effort to fire this timer as scheduled.
.It NOTE_BACKGROUND
system has extra leeway to coalesce this timer.
.It NOTE_LEEWAY
ext[1] holds user-supplied slop in deadline for timer coalescing.
.El
.Pp
If fflags is not set, the default is 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 ,
.Fn kevent64
or
.Fn kevent_qos .
This filter automatically sets the EV_CLEAR flag internally.
.El
.Pp
----
.Pp
In the
.Va ext[2]
field of the
.Va kevent64_s
struture,
.Va ext[0]
is only used with the EVFILT_MACHPORT filter.
With other filters,
.Va ext[0]
is passed through
.Fn kevent64
much like
.Va udata .
.Va ext[1]
can always be used like
.Va udata .
For the use of ext[0], see the EVFILT_MACHPORT filter above.
.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 errno set.
.Pp
The
.Fn kevent ,
.Fn kevent64
and
.Fn kevent_qos
system calls
return 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 ,
.Fn kevent64
and
.Fn kevent_qos
return 0.
.Sh ERRORS
The
.Fn kqueue
system call fails if:
.Bl -tag -width Er
.It Bq Er ENOMEM
The kernel failed to allocate enough memory for the kernel queue.
.It Bq Er EMFILE
The per-process descriptor table is full.
.It Bq Er ENFILE
The system file table is full.
.El
.Pp
The
.Fn kevent
and
.Fn kevent64
system calls fail if:
.Bl -tag -width Er
.It Bq Er EACCES
The process does not have permission to register a filter.
.It Bq Er EFAULT
There was an error reading or writing the
.Va kevent
or
.Va kevent64_s
structure.
.It Bq Er EBADF
The specified descriptor is invalid.
.It Bq Er EINTR
A signal was delivered before the timeout expired and before any
events were placed on the kqueue for return.
.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 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 read 2 ,
.Xr select 2 ,
.Xr sigaction 2 ,
.Xr write 2 ,
.Xr signal 3
.Sh HISTORY
The
.Fn kqueue
and
.Fn kevent
system calls first appeared in
.Fx 4.1 .
.Sh AUTHORS
The
.Fn kqueue
system and this manual page were written by
.An Jonathan Lemon Aq jlemon@FreeBSD.org .
.Sh BUGS
Not all filesystem types support kqueue-style notifications.
And even some that do, like some remote filesystems, may only
support a subset of the notification semantics described
here.