xref: /netbsd-src/share/man/man9/callout.9 (revision e75b3ef75fe5ac4486f4d32df8998f3119477de2)
1.\"	$NetBSD: callout.9,v 1.29 2020/01/12 20:49:21 sevan Exp $
2.\"
3.\" Copyright (c) 2000, 2003, 2009 The NetBSD Foundation, Inc.
4.\" All rights reserved.
5.\"
6.\" This code is derived from software contributed to The NetBSD Foundation
7.\" by Jason R. Thorpe.
8.\"
9.\" Redistribution and use in source and binary forms, with or without
10.\" modification, are permitted provided that the following conditions
11.\" are met:
12.\" 1. Redistributions of source code must retain the above copyright
13.\"    notice, this list of conditions and the following disclaimer.
14.\" 2. Redistributions in binary form must reproduce the above copyright
15.\"    notice, this list of conditions and the following disclaimer in the
16.\"    documentation and/or other materials provided with the distribution.
17.\"
18.\" THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
19.\" ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
20.\" TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
21.\" PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
22.\" BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
23.\" CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
24.\" SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
25.\" INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
26.\" CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
27.\" ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
28.\" POSSIBILITY OF SUCH DAMAGE.
29.\"
30.Dd January 12, 2020
31.Dt CALLOUT 9
32.Os
33.Sh NAME
34.Nm callout_init ,
35.Nm callout_destroy ,
36.Nm callout_halt ,
37.Nm callout_reset ,
38.Nm callout_schedule ,
39.Nm callout_setfunc ,
40.Nm callout_stop ,
41.Nm callout_pending ,
42.Nm callout_expired ,
43.Nm callout_invoking ,
44.Nm callout_ack
45.Nd execute a function after a specified length of time
46.Sh SYNOPSIS
47.In sys/callout.h
48.Ft void
49.Fn "callout_init" "callout_t *c" "u_int flags"
50.Ft void
51.Fn "callout_destroy" "callout_t *c"
52.Ft void
53.Fn "callout_reset" "callout_t *c" "int ticks" \
54    "void (*func)(void *)" "void *arg"
55.Ft void
56.Fn "callout_schedule" "callout_t *c" "int ticks"
57.Ft void
58.Fn "callout_setfunc" "callout_t *c" "void (*func)(void *)" "void *arg"
59.Ft bool
60.Fn "callout_stop" "callout_t *c"
61.Ft bool
62.Fn "callout_halt" "callout_t *c" "void *interlock"
63.Ft bool
64.Fn "callout_pending" "callout_t *c"
65.Ft bool
66.Fn "callout_expired" "callout_t *c"
67.Ft bool
68.Fn "callout_active" "callout_t *c"
69.Ft bool
70.Fn "callout_invoking" "callout_t *c"
71.Ft void
72.Fn "callout_ack" "callout_t *c"
73.Sh DESCRIPTION
74The
75.Nm callout
76facility provides a mechanism to execute a function at a given time.
77The timer is based on the hardclock timer which ticks
78.Dv hz
79times per second.
80The function is called at softclock interrupt level.
81.Pp
82Clients of the
83.Nm callout
84facility are responsible for providing pre-allocated
85callout structures, or
86.Dq handles .
87The
88.Nm callout
89facility replaces the historic
90.Ux
91functions
92.Fn timeout
93and
94.Fn untimeout .
95.Sh FUNCTIONS
96The
97.Fn callout_init
98function initializes the callout handle
99.Fa c
100for use.
101No operations can be performed on the callout before it is initialized.
102If the
103.Fa flags
104argument is
105.Dv CALLOUT_MPSAFE ,
106the handler will be called without getting the global kernel lock.
107In this case it should only use functions that are multiprocessor
108safe.
109.Pp
110.Fn callout_destroy
111destroys the callout, preventing further use.
112It is provided as a diagnostic facility intended to catch bugs.
113To ensure future compatibility,
114.Fn callout_destroy
115should always be called when the callout is no longer required (for instance,
116when a device is being detached).
117The callout should be stopped before
118.Fn callout_destroy
119is called by calling
120.Fn callout_halt .
121Note that
122.Fn callout_stop
123shouldn't be used for this purpose.
124.Pp
125The
126.Fn callout_reset
127function resets and starts the timer associated with the callout handle
128.Fa c .
129When the timer expires after
130.Fa ticks Ns No /hz
131seconds, the function specified by
132.Fa func
133will be called with the argument
134.Fa arg .
135If the timer associated with the callout handle is already running,
136the callout will simply be rescheduled to execute at the newly specified
137time.
138Once the timer is started, the callout handle is marked as
139.Em PENDING .
140Once the timer expires,
141the handle is marked as
142.Em EXPIRED
143and
144.Em INVOKING ,
145and the
146.Em PENDING
147status is cleared.
148.Pp
149The
150.Fn callout_setfunc
151function sets the function and argument of the callout handle
152.Fa c
153to
154.Fa func
155and
156.Fa arg
157respectively.
158The callout handle must already be initialized.
159If a callout will always be used with the same function and argument,
160then
161.Fn callout_setfunc
162used in conjunction with
163.Fn callout_schedule
164is slightly more efficient than using
165.Fn callout_reset .
166.Pp
167The
168.Fn callout_stop
169function requests that the timer associated with the callout handle
170.Fa c
171be stopped.
172The
173.Em PENDING
174and
175.Em EXPIRED
176status for the callout handle is cleared.
177It is safe to call
178.Fn callout_stop
179on a callout handle that is not pending, so long as it is initialized.
180.Fn callout_stop
181will return a non-zero value if the callout was
182.Em EXPIRED .
183Note that
184.Fn callout_stop
185can return while the callout is running on a different CPU or at a
186different interrupt priority level on the current CPU.
187It can only be said to prevent the callout from firing in the future,
188unless explicitly re-scheduled.
189To stop a callout and wait for completion, use
190.Fn callout_halt .
191.Pp
192.Fn callout_halt
193acts much like
194.Fn callout_stop ,
195but waits for the callout to complete if it is currently in-flight.
196.Fn callout_halt
197may not be called from a hard interrupt handler as it will sleep if the
198callout is currently executing.
199If the callout can take locks (such as mutexes or RW locks), the caller of
200.Fn callout_halt
201must not hold any of those locks, otherwise the two could deadlock.
202To facilitate this,
203.Fn callout_halt
204can optionally release a single mutex specified by the
205.Fa interlock
206parameter.
207If
208.Fa interlock
209is not
210.Dv NULL
211and the calling thread must wait for the callout to complete,
212.Fa interlock
213will be released before waiting and re-acquired before returning.
214If no wait is required,
215.Fa interlock
216will not be released.
217However, to avoid race conditions the caller should always assume that
218.Fa interlock
219has been released and reacquired, and act accordingly.
220.Pp
221The
222.Fn callout_pending
223function tests the
224.Em PENDING
225status of the callout handle
226.Fa c .
227A
228.Em PENDING
229callout is one that has been started and whose function has not yet
230been called.
231Note that it is possible for a callout's timer to have expired without
232its function being called if interrupt level has not dropped low enough
233to let softclock interrupts through.
234Note that it is only safe to test
235.Em PENDING
236status when at softclock interrupt level or higher.
237.Pp
238The
239.Fn callout_expired
240function tests to see if the callout's timer has expired and its
241function called.
242.Pp
243The
244.Fn callout_active
245function returns true if a timer has been started but not explicitly stopped,
246even if it has already fired.
247.Fn callout_active foo
248is logically the same as
249.Fn callout_pending foo
250||
251.Fn callout_expired foo ;
252it is implemented as a separate function for compatibility with
253.Fx
254and for the special case of
255.Fn TCP_TIMER_ISARMED .
256Its use is not recommended.
257.Pp
258The
259.Fn callout_invoking
260function tests the
261.Em INVOKING
262status of the callout handle
263.Fa c .
264This flag is set just before a callout's function is being called.
265Since the priority level is lowered prior to invocation of the
266callout function, other pending higher-priority code may run before
267the callout function is allowed to run.
268This may create a race condition if this higher-priority code
269deallocates storage containing one or more callout structures whose
270callout functions are about to be run.
271In such cases, one technique to prevent references to deallocated
272storage would be to test whether any callout functions are in the
273.Em INVOKING
274state using
275.Fn callout_invoking ,
276and if so, to mark the data structure and defer storage
277deallocation until the callout function is allowed to run.
278For this handshake protocol to work, the callout function will
279have to use the
280.Fn callout_ack
281function to clear this flag.
282.Pp
283The
284.Fn callout_ack
285function clears the
286.Em INVOKING
287state in the callout handle
288.Fa c .
289This is used in situations where it is necessary to protect against
290the race condition described under
291.Fn callout_invoking .
292.Sh CONCURRENCY
293The callout facility performs locking internally in order to guarantee the
294atomicity of individual operations performed on callouts.
295It does not provide life cycle management of user-provided callout data
296structures, nor does it ensure that groups of operations (multiple function
297calls) are performed atomically.
298These aspects of callout management are the responsibility of the user of
299the callout facility.
300.Pp
301Scheduled callouts may be active concurrently in a context different to the
302user of the callout facility: on another CPU, or at a different interrupt
303priority level or thread on the current CPU.
304The callout facility provides only one guarantee in this regard: any given
305callout will never have multiple concurrent invocations.
306.Sh SEE ALSO
307.Xr condvar 9 ,
308.Xr hz 9 ,
309.Xr softint 9 ,
310.Xr workqueue 9
311.Sh HISTORY
312The
313.Nm callout
314facility was implemented by Artur Grabowski and Thomas Nordin, based
315on the work of G. Varghese and A. Lauck, described in the paper
316Hashed and Hierarchical Timing Wheels: Data Structures for the
317Efficient Implementation of a Timer Facility
318in the Proceedings of the 11th ACM Annual Symposium on Operating System
319Principles, Austin, Texas, November 1987.
320It was adapted to the
321.Nx
322kernel by Jason R. Thorpe.
323