1.\" $NetBSD: callout.9,v 1.26 2013/02/03 08:19:58 jdc 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 February 2, 2013 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" "kmutex_t *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 bool 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. 120.Pp 121The 122.Fn callout_reset 123function resets and starts the timer associated with the callout handle 124.Fa c . 125When the timer expires after 126.Fa ticks Ns No /hz 127seconds, the function specified by 128.Fa func 129will be called with the argument 130.Fa arg . 131If the timer associated with the callout handle is already running, 132the callout will simply be rescheduled to execute at the newly specified 133time. 134Once the timer is started, the callout handle is marked as 135.Em PENDING . 136Once the timer expires, 137the handle is marked as 138.Em EXPIRED 139and 140.Em INVOKING , 141and the 142.Em PENDING 143status is cleared. 144.Pp 145The 146.Fn callout_setfunc 147function sets the function and argument of the callout handle 148.Fa c 149to 150.Fa func 151and 152.Fa arg 153respectively. 154The callout handle must already be initialized. 155If a callout will always be used with the same function and argument, 156then 157.Fn callout_setfunc 158used in conjunction with 159.Fn callout_schedule 160is slightly more efficient than using 161.Fn callout_reset . 162.Pp 163The 164.Fn callout_stop 165function requests that the timer associated with the callout handle 166.Fa c 167be stopped. 168The 169.Em PENDING 170and 171.Em EXPIRED 172status for the callout handle is cleared. 173It is safe to call 174.Fn callout_stop 175on a callout handle that is not pending, so long as it is initialized. 176.Fn callout_stop 177will return a non-zero value if the callout was 178.Em EXPIRED . 179Note that 180.Fn callout_stop 181can return while the callout is running on a different CPU or at a 182different interrupt priority level on the current CPU. 183It can only be said to prevent the callout from firing in the future, 184unless explicitly re-scheduled. 185To stop a callout and wait for completion, use 186.Fn callout_halt . 187.Pp 188.Fn callout_halt 189acts much like 190.Fn callout_stop , 191but waits for the callout to complete if it is currently in-flight. 192.Fn callout_halt 193may not be called from a hard interrupt handler as it will sleep if the 194callout is currently executing. 195If the callout can take locks (such as mutexes or RW locks), the caller of 196.Fn callout_halt 197must not hold any of those locks, otherwise the two could deadlock. 198To facilitate this, 199.Fn callout_halt 200can optionally release a single mutex specified by the 201.Fa interlock 202parameter. 203If 204.Fa interlock 205is not 206.Dv NULL 207and the calling thread must wait for the callout to complete, 208.Fa interlock 209will be released before waiting and re-acquired before returning. 210If no wait is required, 211.Fa interlock 212will not be released. 213However, to avoid race conditions the caller should always assume that 214.Fa interlock 215has been released and reacquired, and act accordingly. 216.Pp 217The 218.Fn callout_pending 219function tests the 220.Em PENDING 221status of the callout handle 222.Fa c . 223A 224.Em PENDING 225callout is one that has been started and whose function has not yet 226been called. 227Note that it is possible for a callout's timer to have expired without 228its function being called if interrupt level has not dropped low enough 229to let softclock interrupts through. 230Note that it is only safe to test 231.Em PENDING 232status when at softclock interrupt level or higher. 233.Pp 234The 235.Fn callout_expired 236function tests to see if the callout's timer has expired and its 237function called. 238.Pp 239The 240.Fn callout_active 241function returns true if a timer has been started but not explicitly stopped, 242even if it has already fired. 243.Fn callout_active foo 244is logically the same as 245.Fn callout_pending foo 246|| 247.Fn callout_expired foo ; 248it is implemented as a separate function for compatibility with 249.Fx 250and for the special case of 251.Fn TCP_TIMER_ISARMED . 252Its use is not recommended. 253.Pp 254The 255.Fn callout_invoking 256function tests the 257.Em INVOKING 258status of the callout handle 259.Fa c . 260This flag is set just before a callout's function is being called. 261Since the priority level is lowered prior to invocation of the 262callout function, other pending higher-priority code may run before 263the callout function is allowed to run. 264This may create a race condition if this higher-priority code 265deallocates storage containing one or more callout structures whose 266callout functions are about to be run. 267In such cases, one technique to prevent references to deallocated 268storage would be to test whether any callout functions are in the 269.Em INVOKING 270state using 271.Fn callout_invoking , 272and if so, to mark the data structure and defer storage 273deallocation until the callout function is allowed to run. 274For this handshake protocol to work, the callout function will 275have to use the 276.Fn callout_ack 277function to clear this flag. 278.Pp 279The 280.Fn callout_ack 281function clears the 282.Em INVOKING 283state in the callout handle 284.Fa c . 285This is used in situations where it is necessary to protect against 286the race condition described under 287.Fn callout_invoking . 288.Sh CONCURRENCY 289The callout facility performs locking internally in order to guarantee the 290atomicity of individual operations performed on callouts. 291It does not provide life cycle management of user-provided callout data 292structures, nor does it ensure that groups of operations (multiple function 293calls) are performed atomically. 294These aspects of callout management are the responsibility of the user of 295the callout facility. 296.Pp 297Scheduled callouts may be active concurrently in a context different to the 298user of the callout facility: on another CPU, or at a different interrupt 299priority level or thread on the current CPU. 300The callout facility provides only one guarantee in this regard: any given 301callout will never have multiple concurrent invocations. 302.Sh SEE ALSO 303.Xr condvar 9 , 304.Xr hz 9 , 305.Xr softint 9 , 306.Xr workqueue 9 307.Sh HISTORY 308The 309.Nm callout 310facility was implemented by Artur Grabowski and Thomas Nordin, based 311on the work of G. Varghese and A. Lauck, described in the paper 312Hashed and Hierarchical Timing Wheels: Data Structures for the 313Efficient Implementation of a Timer Facility 314in the Proceedings of the 11th ACM Annual Symposium on Operating System 315Principles, Austin, Texas, November 1987. 316It was adapted to the 317.Nx 318kernel by Jason R. Thorpe. 319