1 .\" Copyright (c) 2000 Jonathan Lemon
2 .\" All rights reserved.
4 .\" Redistribution and use in source and binary forms, with or without
5 .\" modification, are permitted provided that the following conditions
7 .\" 1. Redistributions of source code must retain the above copyright
8 .\" notice, this list of conditions and the following disclaimer.
9 .\" 2. Redistributions in binary form must reproduce the above copyright
10 .\" notice, this list of conditions and the following disclaimer in the
11 .\" documentation and/or other materials provided with the distribution.
13 .\" THIS SOFTWARE IS PROVIDED ``AS IS'' AND
14 .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
15 .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
16 .\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
17 .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
18 .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
19 .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
20 .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
21 .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
22 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
27 .Dd September 15, 2009
33 .Nd kernel event notification mechanism
43 .Fn kevent "int kq" "const struct kevent *changelist" "int nchanges" "struct kevent *eventlist" "int nevents" "const struct timespec *timeout"
44 .Fn EV_SET "&kev" ident filter flags fflags data udata
49 provides a generic method of notifying the user when an event
50 happens or a condition holds, based on the results of small
51 pieces of kernel code termed filters.
52 A kevent is identified by the (ident, filter) pair; there may only
53 be one unique kevent per kqueue.
55 The filter is executed upon the initial registration of a kevent
56 in order to detect whether a preexisting condition is present, and is also
57 executed whenever an event is passed to the filter for evaluation.
58 If the filter determines that the condition should be reported,
59 then the kevent is placed on the kqueue for the user to retrieve.
61 The filter is also run when the user attempts to retrieve the kevent
63 If the filter indicates that the condition that triggered
64 the event no longer holds, the kevent is removed from the kqueue and
67 Multiple events which trigger the filter do not result in multiple
68 kevents being placed on the kqueue; instead, the filter will aggregate
69 the events into a single struct kevent.
72 on a file descriptor will remove any kevents that reference the descriptor.
77 creates a new kernel event queue and returns a descriptor.
78 The queue is not inherited by a child created with
84 flag, then the descriptor table is shared,
85 which will allow sharing of the kqueue between two processes.
90 is used to register events with the queue, and return any pending
95 is a pointer to an array of
97 structures, as defined in
99 All changes contained in the
101 are applied before any pending events are read from the queue.
110 is a pointer to an array of kevent structures.
114 determines the size of
120 will return immediately even if there is a
126 is a non-NULL pointer, it specifies a maximum interval to wait
127 for an event, which will be interpreted as a struct timespec.
133 To effect a poll, the
135 argument should be non-NULL, pointing to a zero-valued
138 The same array may be used for the
145 macro is provided for ease of initializing a
150 structure is defined as:
153 uintptr_t ident; /* identifier for this event */
154 short filter; /* filter for event */
155 u_short flags; /* action flags for kqueue */
156 u_int fflags; /* filter flag value */
157 intptr_t data; /* filter data value */
158 void *udata; /* opaque user data identifier */
165 .Bl -tag -width XXXfilter
167 Value used to identify this event.
168 The exact interpretation is determined by the attached filter,
169 but often is a file descriptor.
171 Identifies the kernel filter used to process this event.
173 system filters are described below.
175 Actions to perform on the event.
177 Filter-specific flags.
179 Filter-specific data value.
181 Opaque user-defined value passed through the kernel unchanged.
186 field can contain the following values:
187 .Bl -tag -width XXXEV_ONESHOT
189 Adds the event to the kqueue.
190 Re-adding an existing event
191 will modify the parameters of the original event, and not result
192 in a duplicate entry.
193 Adding an event automatically enables it,
194 unless overridden by the EV_DISABLE flag.
198 to return the event if it is triggered.
203 The filter itself is not disabled.
205 Disable the event source immediately after delivery of an event.
210 Removes the event from the kqueue.
211 Events which are attached to
212 file descriptors are automatically deleted on the last close of
215 This flag is useful for making bulk changes to a kqueue without draining
217 When passed as input, it forces
219 to always be returned.
220 When a filter is successfully added the
224 Causes the event to return only the first occurrence of the filter
226 After the user retrieves the event from the kqueue,
229 After the event is retrieved by the user, its state is reset.
230 This is useful for filters which report state transitions
231 instead of the current state.
232 Note that some filters may automatically
233 set this flag internally.
235 Filters may set this flag to indicate filter-specific EOF condition.
242 The predefined system filters are listed below.
243 Arguments may be passed to and from the filter via the
247 fields in the kevent structure.
248 .Bl -tag -width EVFILT_SIGNAL
250 Takes a descriptor as the identifier, and returns whenever
251 there is data available to read.
252 The behavior of the filter is slightly different depending
253 on the descriptor type.
257 Sockets which have previously been passed to
259 return when there is an incoming connection pending.
261 contains the size of the listen backlog.
263 Other socket descriptors return when there is data to be read,
266 value of the socket buffer.
267 This may be overridden with a per-filter low water mark at the
268 time the filter is added by setting the
272 and specifying the new low water mark in
276 contains the number of bytes of protocol data available to read.
278 If the read direction of the socket has shutdown, then the filter
281 and returns the socket error (if any) in
283 It is possible for EOF to be returned (indicating the connection is gone)
284 while there is still data pending in the socket buffer.
286 Returns when the file pointer is not at the end of file.
288 contains the offset from current position to end of file,
291 Returns when the there is data to read;
293 contains the number of bytes available.
295 When the last writer disconnects, the filter will set EV_EOF in
297 This may be cleared by passing in EV_CLEAR, at which point the
298 filter will resume waiting for data to become available before
301 Returns when the BPF buffer is full, the BPF timeout has expired, or
304 enabled and there is any data to read;
306 contains the number of bytes available.
309 Takes a descriptor as the identifier, and returns whenever
310 it is possible to write to the descriptor.
314 will contain the amount of space remaining in the write buffer.
315 The filter will set EV_EOF when the reader disconnects, and for
316 the fifo case, this may be cleared by use of EV_CLEAR.
317 Note that this filter is not supported for vnodes or BPF devices.
319 For sockets, the low water mark and socket error handling is
320 identical to the EVFILT_READ case.
322 The sigevent portion of the AIO request is filled in, with
323 .Va sigev_notify_kqueue
324 containing the descriptor of the kqueue that the event should
327 containing the udata value, and
332 system call is made, the event will be registered
333 with the specified kqueue, and the
340 The filter returns under the same conditions as aio_error.
342 Takes a file descriptor as the identifier and the events to watch for in
344 and returns when one or more of the requested events occurs on the descriptor.
345 The events to monitor are:
346 .Bl -tag -width XXNOTE_RENAME
351 was called on the file referenced by the descriptor.
353 A write occurred on the file referenced by the descriptor.
355 The file referenced by the descriptor was extended.
357 The file referenced by the descriptor had its attributes changed.
359 The link count on the file changed.
361 The file referenced by the descriptor was renamed.
363 Access to the file was revoked via
365 or the underlying file system was unmounted.
370 contains the events which triggered the filter.
372 Takes the process ID to monitor as the identifier and the events to watch for
375 and returns when the process performs one or more of the requested events.
376 If a process can normally see another process, it can attach an event to it.
377 The events to monitor are:
378 .Bl -tag -width XXNOTE_TRACKERR
380 The process has exited.
381 The exit status will be stored in
384 The process has called
387 The process has executed a new process via
391 Follow a process across
394 The parent process will return with NOTE_TRACK set in the
396 field, while the child process will return with NOTE_CHILD set in
398 and the parent PID in
401 This flag is returned if the system was unable to attach an event to
402 the child process, usually due to resource limitations.
407 contains the events which triggered the filter.
409 Takes the signal number to monitor as the identifier and returns
410 when the given signal is delivered to the process.
411 This coexists with the
415 facilities, and has a lower precedence.
416 The filter will record
417 all attempts to deliver a signal to a process, even if the signal has
418 been marked as SIG_IGN.
419 Event notification happens after normal
420 signal delivery processing.
422 returns the number of times the signal has occurred since the last call to
424 This filter automatically sets the EV_CLEAR flag internally.
426 Establishes an arbitrary timer identified by
430 specifies the timeout period in milliseconds.
431 The timer will be periodic unless EV_ONESHOT is specified.
434 contains the number of times the timeout has expired since the last call to
436 This filter automatically sets the EV_CLEAR flag internally.
437 There is a system wide limit on the number of timers
438 which is controlled by the
439 .Va kern.kq_calloutmax
444 contains the events which triggered the filter.
446 Establishes a user event identified by
448 which is not assosicated with any kernel mechanism but is triggered by
450 The lower 24 bits of the
452 may be used for user defined flags and manipulated using the following:
453 .Bl -tag -width XXNOTE_FFLAGSMASK
466 .It Dv NOTE_FFCTRLMASK
469 .It Dv NOTE_FFLAGSMASK
470 User defined flag mask for
474 A user event is triggered for output with the following:
475 .Bl -tag -width XXNOTE_FFLAGSMASK
477 Cause the event to be triggered.
482 contains the users defined flags in the lower 24 bits.
488 creates a new kernel event queue and returns a file descriptor.
489 If there was an error creating the kernel event queue, a value of -1 is
490 returned and errno set.
495 returns the number of events placed in the
497 up to the value given by
499 If an error occurs while processing an element of the
501 and there is enough room in the
503 then the event will be placed in the
509 and the system error in
513 will be returned, and
515 will be set to indicate the error condition.
516 If the time limit expires, then
522 system call fails if:
525 The kernel failed to allocate enough memory for the kernel queue.
527 The per-process descriptor table is full.
529 The system file table is full.
534 system call fails if:
537 The process does not have permission to register a filter.
539 There was an error reading or writing the
543 The specified descriptor is invalid.
545 A signal was delivered before the timeout expired and before any
546 events were placed on the kqueue for return.
548 The specified time limit or filter is invalid.
550 The event could not be found to be modified or deleted.
552 No memory was available to register the event
553 or, in the special case of a timer, the maximum number of
554 timers has been exceeded.
555 This maximum is configurable via the
556 .Va kern.kq_calloutmax
559 The specified process to attach to does not exist.
576 system calls first appeared in
581 system and this manual page were written by
582 .An Jonathan Lemon Aq jlemon@FreeBSD.org .
586 value is limited to 24 hours; longer timeouts will be silently
587 reinterpreted as 24 hours.