freebsd-dev/sys/kern/kern_timeout.c

/*-
 * Copyright (c) 1982, 1986, 1991, 1993
 *	The Regents of the University of California.  All rights reserved.
 * (c) UNIX System Laboratories, Inc.
 * All or some portions of this file are derived from material licensed
 * to the University of California by American Telephone and Telegraph
 * Co. or Unix System Laboratories, Inc. and are reproduced herein with
 * the permission of UNIX System Laboratories, Inc.
 *
 * 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.
 * 3. All advertising materials mentioning features or use of this software
 *    must display the following acknowledgement:
 *	This product includes software developed by the University of
 *	California, Berkeley and its contributors.
 * 4. Neither the name of the University nor the names of its contributors
 *    may be used to endorse or promote products derived from this software
 *    without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``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 REGENTS 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.
 *
 *	From: @(#)kern_clock.c	8.5 (Berkeley) 1/21/94
 */

#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/callout.h>
#include <sys/kernel.h>
#include <sys/lock.h>
#include <sys/mutex.h>
#include <sys/sysctl.h>

static int avg_depth;
SYSCTL_INT(_debug, OID_AUTO, to_avg_depth, CTLFLAG_RD, &avg_depth, 0,
    "Average number of items examined per softclock call. Units = 1/1000");
static int avg_gcalls;
SYSCTL_INT(_debug, OID_AUTO, to_avg_gcalls, CTLFLAG_RD, &avg_gcalls, 0,
    "Average number of Giant callouts made per softclock call. Units = 1/1000");
static int avg_mpcalls;
SYSCTL_INT(_debug, OID_AUTO, to_avg_mpcalls, CTLFLAG_RD, &avg_mpcalls, 0,
    "Average number of MP callouts made per softclock call. Units = 1/1000");
/*
 * TODO:
 *	allocate more timeout table slots when table overflows.
 */

/* Exported to machdep.c and/or kern_clock.c.  */
struct callout *callout;
struct callout_list callfree;
int callwheelsize, callwheelbits, callwheelmask;
struct callout_tailq *callwheel;
int softticks;			/* Like ticks, but for softclock(). */
struct mtx callout_lock;

static struct callout *nextsoftcheck;	/* Next callout to be checked. */

/*
 * kern_timeout_callwheel_alloc() - kernel low level callwheel initialization 
 *
 *	This code is called very early in the kernel initialization sequence,
 *	and may be called more then once.
 */
caddr_t
kern_timeout_callwheel_alloc(caddr_t v)
{
	/*
	 * Calculate callout wheel size
	 */
	for (callwheelsize = 1, callwheelbits = 0;
	     callwheelsize < ncallout;
	     callwheelsize <<= 1, ++callwheelbits)
		;
	callwheelmask = callwheelsize - 1;

	callout = (struct callout *)v;
	v = (caddr_t)(callout + ncallout);
	callwheel = (struct callout_tailq *)v;
	v = (caddr_t)(callwheel + callwheelsize);
	return(v);
}

/*
 * kern_timeout_callwheel_init() - initialize previously reserved callwheel
 *				   space.
 *
 *	This code is called just once, after the space reserved for the
 *	callout wheel has been finalized.
 */
void
kern_timeout_callwheel_init(void)
{
	int i;

	SLIST_INIT(&callfree);
	for (i = 0; i < ncallout; i++) {
		callout_init(&callout[i], 0);
		callout[i].c_flags = CALLOUT_LOCAL_ALLOC;
		SLIST_INSERT_HEAD(&callfree, &callout[i], c_links.sle);
	}
	for (i = 0; i < callwheelsize; i++) {
		TAILQ_INIT(&callwheel[i]);
	}
	mtx_init(&callout_lock, "callout", NULL, MTX_SPIN | MTX_RECURSE);
}

/*
 * The callout mechanism is based on the work of Adam M. Costello and 
 * George Varghese, published in a technical report entitled "Redesigning
 * the BSD Callout and Timer Facilities" and modified slightly for inclusion
 * in FreeBSD by Justin T. Gibbs.  The original work on the data structures
 * used in this implementation was published by G.Varghese and A. Lauck in
 * the paper "Hashed and Hierarchical Timing Wheels: Data Structures for
 * the Efficient Implementation of a Timer Facility" in the Proceedings of
 * the 11th ACM Annual Symposium on Operating Systems Principles,
 * Austin, Texas Nov 1987.
 */

/*
 * Software (low priority) clock interrupt.
 * Run periodic events from timeout queue.
 */
void
softclock(void *dummy)
{
	struct callout *c;
	struct callout_tailq *bucket;
	int curticks;
	int steps;	/* #steps since we last allowed interrupts */
	int depth;
	int mpcalls;
	int gcalls;

#ifndef MAX_SOFTCLOCK_STEPS
#define MAX_SOFTCLOCK_STEPS 100 /* Maximum allowed value of steps. */
#endif /* MAX_SOFTCLOCK_STEPS */

	mpcalls = 0;
	gcalls = 0;
	depth = 0;
	steps = 0;
	mtx_lock_spin(&callout_lock);
	while (softticks != ticks) {
		softticks++;
		/*
		 * softticks may be modified by hard clock, so cache
		 * it while we work on a given bucket.
		 */
		curticks = softticks;
		bucket = &callwheel[curticks & callwheelmask];
		c = TAILQ_FIRST(bucket);
		while (c) {
			depth++;
			if (c->c_time != curticks) {
				c = TAILQ_NEXT(c, c_links.tqe);
				++steps;
				if (steps >= MAX_SOFTCLOCK_STEPS) {
					nextsoftcheck = c;
					/* Give interrupts a chance. */
					mtx_unlock_spin(&callout_lock);
					;	/* nothing */
					mtx_lock_spin(&callout_lock);
					c = nextsoftcheck;
					steps = 0;
				}
			} else {
				void (*c_func)(void *);
				void *c_arg;
				int c_flags;

				nextsoftcheck = TAILQ_NEXT(c, c_links.tqe);
				TAILQ_REMOVE(bucket, c, c_links.tqe);
				c_func = c->c_func;
				c_arg = c->c_arg;
				c_flags = c->c_flags;
				c->c_func = NULL;
				if (c->c_flags & CALLOUT_LOCAL_ALLOC) {
					c->c_flags = CALLOUT_LOCAL_ALLOC;
					SLIST_INSERT_HEAD(&callfree, c,
							  c_links.sle);
				} else {
					c->c_flags =
					    (c->c_flags & ~CALLOUT_PENDING);
				}
				mtx_unlock_spin(&callout_lock);
				if (!(c_flags & CALLOUT_MPSAFE)) {
					mtx_lock(&Giant);
					gcalls++;
				} else {
					mpcalls++;
				}
				c_func(c_arg);
				if (!(c_flags & CALLOUT_MPSAFE))
					mtx_unlock(&Giant);
				mtx_lock_spin(&callout_lock);
				steps = 0;
				c = nextsoftcheck;
			}
		}
	}
	avg_depth += (depth * 1000 - avg_depth) >> 8;
	avg_mpcalls += (mpcalls * 1000 - avg_mpcalls) >> 8;
	avg_gcalls += (gcalls * 1000 - avg_gcalls) >> 8;
	nextsoftcheck = NULL;
	mtx_unlock_spin(&callout_lock);
}

/*
 * timeout --
 *	Execute a function after a specified length of time.
 *
 * untimeout --
 *	Cancel previous timeout function call.
 *
 * callout_handle_init --
 *	Initialize a handle so that using it with untimeout is benign.
 *
 *	See AT&T BCI Driver Reference Manual for specification.  This
 *	implementation differs from that one in that although an 
 *	identification value is returned from timeout, the original
 *	arguments to timeout as well as the identifier are used to
 *	identify entries for untimeout.
 */
struct callout_handle
timeout(ftn, arg, to_ticks)
	timeout_t *ftn;
	void *arg;
	int to_ticks;
{
	struct callout *new;
	struct callout_handle handle;

	mtx_lock_spin(&callout_lock);

	/* Fill in the next free callout structure. */
	new = SLIST_FIRST(&callfree);
	if (new == NULL)
		/* XXX Attempt to malloc first */
		panic("timeout table full");
	SLIST_REMOVE_HEAD(&callfree, c_links.sle);
	
	callout_reset(new, to_ticks, ftn, arg);

	handle.callout = new;
	mtx_unlock_spin(&callout_lock);
	return (handle);
}

void
untimeout(ftn, arg, handle)
	timeout_t *ftn;
	void *arg;
	struct callout_handle handle;
{

	/*
	 * Check for a handle that was initialized
	 * by callout_handle_init, but never used
	 * for a real timeout.
	 */
	if (handle.callout == NULL)
		return;

	mtx_lock_spin(&callout_lock);
	if (handle.callout->c_func == ftn && handle.callout->c_arg == arg)
		callout_stop(handle.callout);
	mtx_unlock_spin(&callout_lock);
}

void
callout_handle_init(struct callout_handle *handle)
{
	handle->callout = NULL;
}

/*
 * New interface; clients allocate their own callout structures.
 *
 * callout_reset() - establish or change a timeout
 * callout_stop() - disestablish a timeout
 * callout_init() - initialize a callout structure so that it can
 *	safely be passed to callout_reset() and callout_stop()
 *
 * <sys/callout.h> defines three convenience macros:
 *
 * callout_active() - returns truth if callout has not been serviced
 * callout_pending() - returns truth if callout is still waiting for timeout
 * callout_deactivate() - marks the callout as having been serviced
 */
void
callout_reset(c, to_ticks, ftn, arg)
	struct	callout *c;
	int	to_ticks;
	void	(*ftn)(void *);
	void	*arg;
{

	mtx_lock_spin(&callout_lock);
	if (c->c_flags & CALLOUT_PENDING)
		callout_stop(c);

	/*
	 * We could unlock callout_lock here and lock it again before the
	 * TAILQ_INSERT_TAIL, but there's no point since doing this setup
	 * doesn't take much time.
	 */
	if (to_ticks <= 0)
		to_ticks = 1;

	c->c_arg = arg;
	c->c_flags |= (CALLOUT_ACTIVE | CALLOUT_PENDING);
	c->c_func = ftn;
	c->c_time = ticks + to_ticks;
	TAILQ_INSERT_TAIL(&callwheel[c->c_time & callwheelmask], 
			  c, c_links.tqe);
	mtx_unlock_spin(&callout_lock);
}

int
callout_stop(c)
	struct	callout *c;
{

	mtx_lock_spin(&callout_lock);
	/*
	 * Don't attempt to delete a callout that's not on the queue.
	 */
	if (!(c->c_flags & CALLOUT_PENDING)) {
		c->c_flags &= ~CALLOUT_ACTIVE;
		mtx_unlock_spin(&callout_lock);
		return (0);
	}
	c->c_flags &= ~(CALLOUT_ACTIVE | CALLOUT_PENDING);

	if (nextsoftcheck == c) {
		nextsoftcheck = TAILQ_NEXT(c, c_links.tqe);
	}
	TAILQ_REMOVE(&callwheel[c->c_time & callwheelmask], c, c_links.tqe);
	c->c_func = NULL;

	if (c->c_flags & CALLOUT_LOCAL_ALLOC) {
		SLIST_INSERT_HEAD(&callfree, c, c_links.sle);
	}
	mtx_unlock_spin(&callout_lock);
	return (1);
}

void
callout_init(c, mpsafe)
	struct	callout *c;
	int mpsafe;
{
	bzero(c, sizeof *c);
	if (mpsafe)
		c->c_flags |= CALLOUT_MPSAFE;
}

#ifdef APM_FIXUP_CALLTODO
/* 
 * Adjust the kernel calltodo timeout list.  This routine is used after 
 * an APM resume to recalculate the calltodo timer list values with the 
 * number of hz's we have been sleeping.  The next hardclock() will detect 
 * that there are fired timers and run softclock() to execute them.
 *
 * Please note, I have not done an exhaustive analysis of what code this
 * might break.  I am motivated to have my select()'s and alarm()'s that
 * have expired during suspend firing upon resume so that the applications
 * which set the timer can do the maintanence the timer was for as close
 * as possible to the originally intended time.  Testing this code for a 
 * week showed that resuming from a suspend resulted in 22 to 25 timers 
 * firing, which seemed independant on whether the suspend was 2 hours or
 * 2 days.  Your milage may vary.   - Ken Key <key@cs.utk.edu>
 */
void
adjust_timeout_calltodo(time_change)
    struct timeval *time_change;
{
	register struct callout *p;
	unsigned long delta_ticks;

	/* 
	 * How many ticks were we asleep?
	 * (stolen from tvtohz()).
	 */

	/* Don't do anything */
	if (time_change->tv_sec < 0)
		return;
	else if (time_change->tv_sec <= LONG_MAX / 1000000)
		delta_ticks = (time_change->tv_sec * 1000000 +
			       time_change->tv_usec + (tick - 1)) / tick + 1;
	else if (time_change->tv_sec <= LONG_MAX / hz)
		delta_ticks = time_change->tv_sec * hz +
			      (time_change->tv_usec + (tick - 1)) / tick + 1;
	else
		delta_ticks = LONG_MAX;

	if (delta_ticks > INT_MAX)
		delta_ticks = INT_MAX;

	/* 
	 * Now rip through the timer calltodo list looking for timers
	 * to expire.
	 */

	/* don't collide with softclock() */
	mtx_lock_spin(&callout_lock);
	for (p = calltodo.c_next; p != NULL; p = p->c_next) {
		p->c_time -= delta_ticks;

		/* Break if the timer had more time on it than delta_ticks */
		if (p->c_time > 0)
			break;

		/* take back the ticks the timer didn't use (p->c_time <= 0) */
		delta_ticks = -p->c_time;
	}
	mtx_unlock_spin(&callout_lock);

	return;
}
#endif /* APM_FIXUP_CALLTODO */