rcutree.c 49.8 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
/*
 * Read-Copy Update mechanism for mutual exclusion
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
 *
 * Copyright IBM Corporation, 2008
 *
 * Authors: Dipankar Sarma <dipankar@in.ibm.com>
 *	    Manfred Spraul <manfred@colorfullife.com>
 *	    Paul E. McKenney <paulmck@linux.vnet.ibm.com> Hierarchical version
 *
 * Based on the original work by Paul McKenney <paulmck@us.ibm.com>
 * and inputs from Rusty Russell, Andrea Arcangeli and Andi Kleen.
 *
 * For detailed explanation of Read-Copy Update mechanism see -
28
 *	Documentation/RCU
29 30 31 32 33 34 35 36 37
 */
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/init.h>
#include <linux/spinlock.h>
#include <linux/smp.h>
#include <linux/rcupdate.h>
#include <linux/interrupt.h>
#include <linux/sched.h>
38
#include <linux/nmi.h>
39 40 41 42 43 44 45 46 47 48 49
#include <asm/atomic.h>
#include <linux/bitops.h>
#include <linux/module.h>
#include <linux/completion.h>
#include <linux/moduleparam.h>
#include <linux/percpu.h>
#include <linux/notifier.h>
#include <linux/cpu.h>
#include <linux/mutex.h>
#include <linux/time.h>

50 51
#include "rcutree.h"

52 53
/* Data structures. */

54 55
static struct lock_class_key rcu_root_class;

56 57 58 59 60 61 62 63
#define RCU_STATE_INITIALIZER(name) { \
	.level = { &name.node[0] }, \
	.levelcnt = { \
		NUM_RCU_LVL_0,  /* root of hierarchy. */ \
		NUM_RCU_LVL_1, \
		NUM_RCU_LVL_2, \
		NUM_RCU_LVL_3, /* == MAX_RCU_LVLS */ \
	}, \
64
	.signaled = RCU_GP_IDLE, \
65 66 67
	.gpnum = -300, \
	.completed = -300, \
	.onofflock = __SPIN_LOCK_UNLOCKED(&name.onofflock), \
68 69 70
	.orphan_cbs_list = NULL, \
	.orphan_cbs_tail = &name.orphan_cbs_list, \
	.orphan_qlen = 0, \
71 72 73 74 75
	.fqslock = __SPIN_LOCK_UNLOCKED(&name.fqslock), \
	.n_force_qs = 0, \
	.n_force_qs_ngp = 0, \
}

76 77
struct rcu_state rcu_sched_state = RCU_STATE_INITIALIZER(rcu_sched_state);
DEFINE_PER_CPU(struct rcu_data, rcu_sched_data);
78

79 80
struct rcu_state rcu_bh_state = RCU_STATE_INITIALIZER(rcu_bh_state);
DEFINE_PER_CPU(struct rcu_data, rcu_bh_data);
81

82

83 84 85 86 87 88 89 90 91 92
/*
 * Return true if an RCU grace period is in progress.  The ACCESS_ONCE()s
 * permit this function to be invoked without holding the root rcu_node
 * structure's ->lock, but of course results can be subject to change.
 */
static int rcu_gp_in_progress(struct rcu_state *rsp)
{
	return ACCESS_ONCE(rsp->completed) != ACCESS_ONCE(rsp->gpnum);
}

93
/*
94
 * Note a quiescent state.  Because we do not need to know
95
 * how many quiescent states passed, just if there was at least
96
 * one since the start of the grace period, this just sets a flag.
97
 */
98
void rcu_sched_qs(int cpu)
99
{
100 101 102
	struct rcu_data *rdp;

	rdp = &per_cpu(rcu_sched_data, cpu);
103
	rdp->passed_quiesc_completed = rdp->completed;
104 105 106
	barrier();
	rdp->passed_quiesc = 1;
	rcu_preempt_note_context_switch(cpu);
107 108
}

109
void rcu_bh_qs(int cpu)
110
{
111 112 113
	struct rcu_data *rdp;

	rdp = &per_cpu(rcu_bh_data, cpu);
114
	rdp->passed_quiesc_completed = rdp->completed;
115 116
	barrier();
	rdp->passed_quiesc = 1;
117
}
118 119

#ifdef CONFIG_NO_HZ
120 121 122 123
DEFINE_PER_CPU(struct rcu_dynticks, rcu_dynticks) = {
	.dynticks_nesting = 1,
	.dynticks = 1,
};
124 125 126 127 128 129
#endif /* #ifdef CONFIG_NO_HZ */

static int blimit = 10;		/* Maximum callbacks per softirq. */
static int qhimark = 10000;	/* If this many pending, ignore blimit. */
static int qlowmark = 100;	/* Once only this many pending, use blimit. */

130 131 132 133
module_param(blimit, int, 0);
module_param(qhimark, int, 0);
module_param(qlowmark, int, 0);

134
static void force_quiescent_state(struct rcu_state *rsp, int relaxed);
135
static int rcu_pending(int cpu);
136 137

/*
138
 * Return the number of RCU-sched batches processed thus far for debug & stats.
139
 */
140
long rcu_batches_completed_sched(void)
141
{
142
	return rcu_sched_state.completed;
143
}
144
EXPORT_SYMBOL_GPL(rcu_batches_completed_sched);
145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169

/*
 * Return the number of RCU BH batches processed thus far for debug & stats.
 */
long rcu_batches_completed_bh(void)
{
	return rcu_bh_state.completed;
}
EXPORT_SYMBOL_GPL(rcu_batches_completed_bh);

/*
 * Does the CPU have callbacks ready to be invoked?
 */
static int
cpu_has_callbacks_ready_to_invoke(struct rcu_data *rdp)
{
	return &rdp->nxtlist != rdp->nxttail[RCU_DONE_TAIL];
}

/*
 * Does the current CPU require a yet-as-unscheduled grace period?
 */
static int
cpu_needs_another_gp(struct rcu_state *rsp, struct rcu_data *rdp)
{
170
	return *rdp->nxttail[RCU_DONE_TAIL] && !rcu_gp_in_progress(rsp);
171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204
}

/*
 * Return the root node of the specified rcu_state structure.
 */
static struct rcu_node *rcu_get_root(struct rcu_state *rsp)
{
	return &rsp->node[0];
}

#ifdef CONFIG_SMP

/*
 * If the specified CPU is offline, tell the caller that it is in
 * a quiescent state.  Otherwise, whack it with a reschedule IPI.
 * Grace periods can end up waiting on an offline CPU when that
 * CPU is in the process of coming online -- it will be added to the
 * rcu_node bitmasks before it actually makes it online.  The same thing
 * can happen while a CPU is in the process of coming online.  Because this
 * race is quite rare, we check for it after detecting that the grace
 * period has been delayed rather than checking each and every CPU
 * each and every time we start a new grace period.
 */
static int rcu_implicit_offline_qs(struct rcu_data *rdp)
{
	/*
	 * If the CPU is offline, it is in a quiescent state.  We can
	 * trust its state not to change because interrupts are disabled.
	 */
	if (cpu_is_offline(rdp->cpu)) {
		rdp->offline_fqs++;
		return 1;
	}

205 206 207 208
	/* If preemptable RCU, no point in sending reschedule IPI. */
	if (rdp->preemptable)
		return 0;

209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239
	/* The CPU is online, so send it a reschedule IPI. */
	if (rdp->cpu != smp_processor_id())
		smp_send_reschedule(rdp->cpu);
	else
		set_need_resched();
	rdp->resched_ipi++;
	return 0;
}

#endif /* #ifdef CONFIG_SMP */

#ifdef CONFIG_NO_HZ

/**
 * rcu_enter_nohz - inform RCU that current CPU is entering nohz
 *
 * Enter nohz mode, in other words, -leave- the mode in which RCU
 * read-side critical sections can occur.  (Though RCU read-side
 * critical sections can occur in irq handlers in nohz mode, a possibility
 * handled by rcu_irq_enter() and rcu_irq_exit()).
 */
void rcu_enter_nohz(void)
{
	unsigned long flags;
	struct rcu_dynticks *rdtp;

	smp_mb(); /* CPUs seeing ++ must see prior RCU read-side crit sects */
	local_irq_save(flags);
	rdtp = &__get_cpu_var(rcu_dynticks);
	rdtp->dynticks++;
	rdtp->dynticks_nesting--;
240
	WARN_ON_ONCE(rdtp->dynticks & 0x1);
241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258
	local_irq_restore(flags);
}

/*
 * rcu_exit_nohz - inform RCU that current CPU is leaving nohz
 *
 * Exit nohz mode, in other words, -enter- the mode in which RCU
 * read-side critical sections normally occur.
 */
void rcu_exit_nohz(void)
{
	unsigned long flags;
	struct rcu_dynticks *rdtp;

	local_irq_save(flags);
	rdtp = &__get_cpu_var(rcu_dynticks);
	rdtp->dynticks++;
	rdtp->dynticks_nesting++;
259
	WARN_ON_ONCE(!(rdtp->dynticks & 0x1));
260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277
	local_irq_restore(flags);
	smp_mb(); /* CPUs seeing ++ must see later RCU read-side crit sects */
}

/**
 * rcu_nmi_enter - inform RCU of entry to NMI context
 *
 * If the CPU was idle with dynamic ticks active, and there is no
 * irq handler running, this updates rdtp->dynticks_nmi to let the
 * RCU grace-period handling know that the CPU is active.
 */
void rcu_nmi_enter(void)
{
	struct rcu_dynticks *rdtp = &__get_cpu_var(rcu_dynticks);

	if (rdtp->dynticks & 0x1)
		return;
	rdtp->dynticks_nmi++;
278
	WARN_ON_ONCE(!(rdtp->dynticks_nmi & 0x1));
279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296
	smp_mb(); /* CPUs seeing ++ must see later RCU read-side crit sects */
}

/**
 * rcu_nmi_exit - inform RCU of exit from NMI context
 *
 * If the CPU was idle with dynamic ticks active, and there is no
 * irq handler running, this updates rdtp->dynticks_nmi to let the
 * RCU grace-period handling know that the CPU is no longer active.
 */
void rcu_nmi_exit(void)
{
	struct rcu_dynticks *rdtp = &__get_cpu_var(rcu_dynticks);

	if (rdtp->dynticks & 0x1)
		return;
	smp_mb(); /* CPUs seeing ++ must see prior RCU read-side crit sects */
	rdtp->dynticks_nmi++;
297
	WARN_ON_ONCE(rdtp->dynticks_nmi & 0x1);
298 299 300 301 302 303 304 305 306 307 308 309 310 311 312
}

/**
 * rcu_irq_enter - inform RCU of entry to hard irq context
 *
 * If the CPU was idle with dynamic ticks active, this updates the
 * rdtp->dynticks to let the RCU handling know that the CPU is active.
 */
void rcu_irq_enter(void)
{
	struct rcu_dynticks *rdtp = &__get_cpu_var(rcu_dynticks);

	if (rdtp->dynticks_nesting++)
		return;
	rdtp->dynticks++;
313
	WARN_ON_ONCE(!(rdtp->dynticks & 0x1));
314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331
	smp_mb(); /* CPUs seeing ++ must see later RCU read-side crit sects */
}

/**
 * rcu_irq_exit - inform RCU of exit from hard irq context
 *
 * If the CPU was idle with dynamic ticks active, update the rdp->dynticks
 * to put let the RCU handling be aware that the CPU is going back to idle
 * with no ticks.
 */
void rcu_irq_exit(void)
{
	struct rcu_dynticks *rdtp = &__get_cpu_var(rcu_dynticks);

	if (--rdtp->dynticks_nesting)
		return;
	smp_mb(); /* CPUs seeing ++ must see prior RCU read-side crit sects */
	rdtp->dynticks++;
332
	WARN_ON_ONCE(rdtp->dynticks & 0x1);
333 334

	/* If the interrupt queued a callback, get out of dyntick mode. */
335
	if (__get_cpu_var(rcu_sched_data).nxtlist ||
336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363
	    __get_cpu_var(rcu_bh_data).nxtlist)
		set_need_resched();
}

/*
 * Record the specified "completed" value, which is later used to validate
 * dynticks counter manipulations.  Specify "rsp->completed - 1" to
 * unconditionally invalidate any future dynticks manipulations (which is
 * useful at the beginning of a grace period).
 */
static void dyntick_record_completed(struct rcu_state *rsp, long comp)
{
	rsp->dynticks_completed = comp;
}

#ifdef CONFIG_SMP

/*
 * Recall the previously recorded value of the completion for dynticks.
 */
static long dyntick_recall_completed(struct rcu_state *rsp)
{
	return rsp->dynticks_completed;
}

/*
 * Snapshot the specified CPU's dynticks counter so that we can later
 * credit them with an implicit quiescent state.  Return 1 if this CPU
364
 * is in dynticks idle mode, which is an extended quiescent state.
365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474
 */
static int dyntick_save_progress_counter(struct rcu_data *rdp)
{
	int ret;
	int snap;
	int snap_nmi;

	snap = rdp->dynticks->dynticks;
	snap_nmi = rdp->dynticks->dynticks_nmi;
	smp_mb();	/* Order sampling of snap with end of grace period. */
	rdp->dynticks_snap = snap;
	rdp->dynticks_nmi_snap = snap_nmi;
	ret = ((snap & 0x1) == 0) && ((snap_nmi & 0x1) == 0);
	if (ret)
		rdp->dynticks_fqs++;
	return ret;
}

/*
 * Return true if the specified CPU has passed through a quiescent
 * state by virtue of being in or having passed through an dynticks
 * idle state since the last call to dyntick_save_progress_counter()
 * for this same CPU.
 */
static int rcu_implicit_dynticks_qs(struct rcu_data *rdp)
{
	long curr;
	long curr_nmi;
	long snap;
	long snap_nmi;

	curr = rdp->dynticks->dynticks;
	snap = rdp->dynticks_snap;
	curr_nmi = rdp->dynticks->dynticks_nmi;
	snap_nmi = rdp->dynticks_nmi_snap;
	smp_mb(); /* force ordering with cpu entering/leaving dynticks. */

	/*
	 * If the CPU passed through or entered a dynticks idle phase with
	 * no active irq/NMI handlers, then we can safely pretend that the CPU
	 * already acknowledged the request to pass through a quiescent
	 * state.  Either way, that CPU cannot possibly be in an RCU
	 * read-side critical section that started before the beginning
	 * of the current RCU grace period.
	 */
	if ((curr != snap || (curr & 0x1) == 0) &&
	    (curr_nmi != snap_nmi || (curr_nmi & 0x1) == 0)) {
		rdp->dynticks_fqs++;
		return 1;
	}

	/* Go check for the CPU being offline. */
	return rcu_implicit_offline_qs(rdp);
}

#endif /* #ifdef CONFIG_SMP */

#else /* #ifdef CONFIG_NO_HZ */

static void dyntick_record_completed(struct rcu_state *rsp, long comp)
{
}

#ifdef CONFIG_SMP

/*
 * If there are no dynticks, then the only way that a CPU can passively
 * be in a quiescent state is to be offline.  Unlike dynticks idle, which
 * is a point in time during the prior (already finished) grace period,
 * an offline CPU is always in a quiescent state, and thus can be
 * unconditionally applied.  So just return the current value of completed.
 */
static long dyntick_recall_completed(struct rcu_state *rsp)
{
	return rsp->completed;
}

static int dyntick_save_progress_counter(struct rcu_data *rdp)
{
	return 0;
}

static int rcu_implicit_dynticks_qs(struct rcu_data *rdp)
{
	return rcu_implicit_offline_qs(rdp);
}

#endif /* #ifdef CONFIG_SMP */

#endif /* #else #ifdef CONFIG_NO_HZ */

#ifdef CONFIG_RCU_CPU_STALL_DETECTOR

static void record_gp_stall_check_time(struct rcu_state *rsp)
{
	rsp->gp_start = jiffies;
	rsp->jiffies_stall = jiffies + RCU_SECONDS_TILL_STALL_CHECK;
}

static void print_other_cpu_stall(struct rcu_state *rsp)
{
	int cpu;
	long delta;
	unsigned long flags;
	struct rcu_node *rnp = rcu_get_root(rsp);

	/* Only let one CPU complain about others per time interval. */

	spin_lock_irqsave(&rnp->lock, flags);
	delta = jiffies - rsp->jiffies_stall;
475
	if (delta < RCU_STALL_RAT_DELAY || !rcu_gp_in_progress(rsp)) {
476 477 478 479
		spin_unlock_irqrestore(&rnp->lock, flags);
		return;
	}
	rsp->jiffies_stall = jiffies + RCU_SECONDS_TILL_STALL_RECHECK;
480 481 482 483 484 485

	/*
	 * Now rat on any tasks that got kicked up to the root rcu_node
	 * due to CPU offlining.
	 */
	rcu_print_task_stall(rnp);
486 487 488 489 490
	spin_unlock_irqrestore(&rnp->lock, flags);

	/* OK, time to rat on our buddy... */

	printk(KERN_ERR "INFO: RCU detected CPU stalls:");
491
	rcu_for_each_leaf_node(rsp, rnp) {
492
		rcu_print_task_stall(rnp);
493
		if (rnp->qsmask == 0)
494
			continue;
495 496 497
		for (cpu = 0; cpu <= rnp->grphi - rnp->grplo; cpu++)
			if (rnp->qsmask & (1UL << cpu))
				printk(" %d", rnp->grplo + cpu);
498 499 500
	}
	printk(" (detected by %d, t=%ld jiffies)\n",
	       smp_processor_id(), (long)(jiffies - rsp->gp_start));
501 502
	trigger_all_cpu_backtrace();

503 504 505 506 507 508 509 510 511 512
	force_quiescent_state(rsp, 0);  /* Kick them all. */
}

static void print_cpu_stall(struct rcu_state *rsp)
{
	unsigned long flags;
	struct rcu_node *rnp = rcu_get_root(rsp);

	printk(KERN_ERR "INFO: RCU detected CPU %d stall (t=%lu jiffies)\n",
			smp_processor_id(), jiffies - rsp->gp_start);
513 514
	trigger_all_cpu_backtrace();

515 516 517 518 519
	spin_lock_irqsave(&rnp->lock, flags);
	if ((long)(jiffies - rsp->jiffies_stall) >= 0)
		rsp->jiffies_stall =
			jiffies + RCU_SECONDS_TILL_STALL_RECHECK;
	spin_unlock_irqrestore(&rnp->lock, flags);
520

521 522 523 524 525 526 527 528 529 530 531 532 533 534 535
	set_need_resched();  /* kick ourselves to get things going. */
}

static void check_cpu_stall(struct rcu_state *rsp, struct rcu_data *rdp)
{
	long delta;
	struct rcu_node *rnp;

	delta = jiffies - rsp->jiffies_stall;
	rnp = rdp->mynode;
	if ((rnp->qsmask & rdp->grpmask) && delta >= 0) {

		/* We haven't checked in, so go dump stack. */
		print_cpu_stall(rsp);

536
	} else if (rcu_gp_in_progress(rsp) && delta >= RCU_STALL_RAT_DELAY) {
537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606

		/* They had two time units to dump stack, so complain. */
		print_other_cpu_stall(rsp);
	}
}

#else /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */

static void record_gp_stall_check_time(struct rcu_state *rsp)
{
}

static void check_cpu_stall(struct rcu_state *rsp, struct rcu_data *rdp)
{
}

#endif /* #else #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */

/*
 * Update CPU-local rcu_data state to record the newly noticed grace period.
 * This is used both when we started the grace period and when we notice
 * that someone else started the grace period.
 */
static void note_new_gpnum(struct rcu_state *rsp, struct rcu_data *rdp)
{
	rdp->qs_pending = 1;
	rdp->passed_quiesc = 0;
	rdp->gpnum = rsp->gpnum;
}

/*
 * Did someone else start a new RCU grace period start since we last
 * checked?  Update local state appropriately if so.  Must be called
 * on the CPU corresponding to rdp.
 */
static int
check_for_new_grace_period(struct rcu_state *rsp, struct rcu_data *rdp)
{
	unsigned long flags;
	int ret = 0;

	local_irq_save(flags);
	if (rdp->gpnum != rsp->gpnum) {
		note_new_gpnum(rsp, rdp);
		ret = 1;
	}
	local_irq_restore(flags);
	return ret;
}

/*
 * Start a new RCU grace period if warranted, re-initializing the hierarchy
 * in preparation for detecting the next grace period.  The caller must hold
 * the root node's ->lock, which is released before return.  Hard irqs must
 * be disabled.
 */
static void
rcu_start_gp(struct rcu_state *rsp, unsigned long flags)
	__releases(rcu_get_root(rsp)->lock)
{
	struct rcu_data *rdp = rsp->rda[smp_processor_id()];
	struct rcu_node *rnp = rcu_get_root(rsp);

	if (!cpu_needs_another_gp(rsp, rdp)) {
		spin_unlock_irqrestore(&rnp->lock, flags);
		return;
	}

	/* Advance to a new grace period and initialize state. */
	rsp->gpnum++;
607
	WARN_ON_ONCE(rsp->signaled == RCU_GP_INIT);
608 609 610 611 612 613 614
	rsp->signaled = RCU_GP_INIT; /* Hold off force_quiescent_state. */
	rsp->jiffies_force_qs = jiffies + RCU_JIFFIES_TILL_FORCE_QS;
	record_gp_stall_check_time(rsp);
	dyntick_record_completed(rsp, rsp->completed - 1);
	note_new_gpnum(rsp, rdp);

	/*
615 616 617 618 619 620 621 622 623
	 * Because this CPU just now started the new grace period, we know
	 * that all of its callbacks will be covered by this upcoming grace
	 * period, even the ones that were registered arbitrarily recently.
	 * Therefore, advance all outstanding callbacks to RCU_WAIT_TAIL.
	 *
	 * Other CPUs cannot be sure exactly when the grace period started.
	 * Therefore, their recently registered callbacks must pass through
	 * an additional RCU_NEXT_READY stage, so that they will be handled
	 * by the next RCU grace period.
624 625 626 627 628 629
	 */
	rdp->nxttail[RCU_NEXT_READY_TAIL] = rdp->nxttail[RCU_NEXT_TAIL];
	rdp->nxttail[RCU_WAIT_TAIL] = rdp->nxttail[RCU_NEXT_TAIL];

	/* Special-case the common single-level case. */
	if (NUM_RCU_NODES == 1) {
630
		rcu_preempt_check_blocked_tasks(rnp);
631
		rnp->qsmask = rnp->qsmaskinit;
632
		rnp->gpnum = rsp->gpnum;
633
		rsp->signaled = RCU_SIGNAL_INIT; /* force_quiescent_state OK. */
634 635 636 637 638 639 640 641 642 643 644
		spin_unlock_irqrestore(&rnp->lock, flags);
		return;
	}

	spin_unlock(&rnp->lock);  /* leave irqs disabled. */


	/* Exclude any concurrent CPU-hotplug operations. */
	spin_lock(&rsp->onofflock);  /* irqs already disabled. */

	/*
645 646 647 648 649 650 651 652 653
	 * Set the quiescent-state-needed bits in all the rcu_node
	 * structures for all currently online CPUs in breadth-first
	 * order, starting from the root rcu_node structure.  This
	 * operation relies on the layout of the hierarchy within the
	 * rsp->node[] array.  Note that other CPUs will access only
	 * the leaves of the hierarchy, which still indicate that no
	 * grace period is in progress, at least until the corresponding
	 * leaf node has been initialized.  In addition, we have excluded
	 * CPU-hotplug operations.
654 655 656 657
	 *
	 * Note that the grace period cannot complete until we finish
	 * the initialization process, as there will be at least one
	 * qsmask bit set in the root node until that time, namely the
658 659
	 * one corresponding to this CPU, due to the fact that we have
	 * irqs disabled.
660
	 */
661
	rcu_for_each_node_breadth_first(rsp, rnp) {
662
		spin_lock(&rnp->lock);		/* irqs already disabled. */
663
		rcu_preempt_check_blocked_tasks(rnp);
664
		rnp->qsmask = rnp->qsmaskinit;
665
		rnp->gpnum = rsp->gpnum;
666
		spin_unlock(&rnp->lock);	/* irqs remain disabled. */
667 668
	}

669 670
	rnp = rcu_get_root(rsp);
	spin_lock(&rnp->lock);			/* irqs already disabled. */
671
	rsp->signaled = RCU_SIGNAL_INIT; /* force_quiescent_state now OK. */
672
	spin_unlock(&rnp->lock);		/* irqs remain disabled. */
673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703
	spin_unlock_irqrestore(&rsp->onofflock, flags);
}

/*
 * Advance this CPU's callbacks, but only if the current grace period
 * has ended.  This may be called only from the CPU to whom the rdp
 * belongs.
 */
static void
rcu_process_gp_end(struct rcu_state *rsp, struct rcu_data *rdp)
{
	long completed_snap;
	unsigned long flags;

	local_irq_save(flags);
	completed_snap = ACCESS_ONCE(rsp->completed);  /* outside of lock. */

	/* Did another grace period end? */
	if (rdp->completed != completed_snap) {

		/* Advance callbacks.  No harm if list empty. */
		rdp->nxttail[RCU_DONE_TAIL] = rdp->nxttail[RCU_WAIT_TAIL];
		rdp->nxttail[RCU_WAIT_TAIL] = rdp->nxttail[RCU_NEXT_READY_TAIL];
		rdp->nxttail[RCU_NEXT_READY_TAIL] = rdp->nxttail[RCU_NEXT_TAIL];

		/* Remember that we saw this grace-period completion. */
		rdp->completed = completed_snap;
	}
	local_irq_restore(flags);
}

704 705 706 707 708 709
/*
 * Clean up after the prior grace period and let rcu_start_gp() start up
 * the next grace period if one is needed.  Note that the caller must
 * hold rnp->lock, as required by rcu_start_gp(), which will release it.
 */
static void cpu_quiet_msk_finish(struct rcu_state *rsp, unsigned long flags)
710
	__releases(rcu_get_root(rsp)->lock)
711
{
712
	WARN_ON_ONCE(!rcu_gp_in_progress(rsp));
713
	rsp->completed = rsp->gpnum;
714
	rsp->signaled = RCU_GP_IDLE;
715 716 717 718
	rcu_process_gp_end(rsp, rsp->rda[smp_processor_id()]);
	rcu_start_gp(rsp, flags);  /* releases root node's rnp->lock. */
}

719 720 721 722 723 724 725 726 727 728 729 730
/*
 * Similar to cpu_quiet(), for which it is a helper function.  Allows
 * a group of CPUs to be quieted at one go, though all the CPUs in the
 * group must be represented by the same leaf rcu_node structure.
 * That structure's lock must be held upon entry, and it is released
 * before return.
 */
static void
cpu_quiet_msk(unsigned long mask, struct rcu_state *rsp, struct rcu_node *rnp,
	      unsigned long flags)
	__releases(rnp->lock)
{
731 732
	struct rcu_node *rnp_c;

733 734 735 736 737 738 739 740 741
	/* Walk up the rcu_node hierarchy. */
	for (;;) {
		if (!(rnp->qsmask & mask)) {

			/* Our bit has already been cleared, so done. */
			spin_unlock_irqrestore(&rnp->lock, flags);
			return;
		}
		rnp->qsmask &= ~mask;
742
		if (rnp->qsmask != 0 || rcu_preempted_readers(rnp)) {
743 744 745 746 747 748 749 750 751 752 753 754 755

			/* Other bits still set at this level, so done. */
			spin_unlock_irqrestore(&rnp->lock, flags);
			return;
		}
		mask = rnp->grpmask;
		if (rnp->parent == NULL) {

			/* No more levels.  Exit loop holding root lock. */

			break;
		}
		spin_unlock_irqrestore(&rnp->lock, flags);
756
		rnp_c = rnp;
757 758
		rnp = rnp->parent;
		spin_lock_irqsave(&rnp->lock, flags);
759
		WARN_ON_ONCE(rnp_c->qsmask);
760 761 762 763
	}

	/*
	 * Get here if we are the last CPU to pass through a quiescent
764 765
	 * state for this grace period.  Invoke cpu_quiet_msk_finish()
	 * to clean up and start the next grace period if one is needed.
766
	 */
767
	cpu_quiet_msk_finish(rsp, flags); /* releases rnp->lock. */
768 769 770 771
}

/*
 * Record a quiescent state for the specified CPU, which must either be
772 773 774 775
 * the current CPU.  The lastcomp argument is used to make sure we are
 * still in the grace period of interest.  We don't want to end the current
 * grace period based on quiescent states detected in an earlier grace
 * period!
776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848
 */
static void
cpu_quiet(int cpu, struct rcu_state *rsp, struct rcu_data *rdp, long lastcomp)
{
	unsigned long flags;
	unsigned long mask;
	struct rcu_node *rnp;

	rnp = rdp->mynode;
	spin_lock_irqsave(&rnp->lock, flags);
	if (lastcomp != ACCESS_ONCE(rsp->completed)) {

		/*
		 * Someone beat us to it for this grace period, so leave.
		 * The race with GP start is resolved by the fact that we
		 * hold the leaf rcu_node lock, so that the per-CPU bits
		 * cannot yet be initialized -- so we would simply find our
		 * CPU's bit already cleared in cpu_quiet_msk() if this race
		 * occurred.
		 */
		rdp->passed_quiesc = 0;	/* try again later! */
		spin_unlock_irqrestore(&rnp->lock, flags);
		return;
	}
	mask = rdp->grpmask;
	if ((rnp->qsmask & mask) == 0) {
		spin_unlock_irqrestore(&rnp->lock, flags);
	} else {
		rdp->qs_pending = 0;

		/*
		 * This GP can't end until cpu checks in, so all of our
		 * callbacks can be processed during the next GP.
		 */
		rdp->nxttail[RCU_NEXT_READY_TAIL] = rdp->nxttail[RCU_NEXT_TAIL];

		cpu_quiet_msk(mask, rsp, rnp, flags); /* releases rnp->lock */
	}
}

/*
 * Check to see if there is a new grace period of which this CPU
 * is not yet aware, and if so, set up local rcu_data state for it.
 * Otherwise, see if this CPU has just passed through its first
 * quiescent state for this grace period, and record that fact if so.
 */
static void
rcu_check_quiescent_state(struct rcu_state *rsp, struct rcu_data *rdp)
{
	/* If there is now a new grace period, record and return. */
	if (check_for_new_grace_period(rsp, rdp))
		return;

	/*
	 * Does this CPU still need to do its part for current grace period?
	 * If no, return and let the other CPUs do their part as well.
	 */
	if (!rdp->qs_pending)
		return;

	/*
	 * Was there a quiescent state since the beginning of the grace
	 * period? If no, then exit and wait for the next call.
	 */
	if (!rdp->passed_quiesc)
		return;

	/* Tell RCU we are done (but cpu_quiet() will be the judge of that). */
	cpu_quiet(rdp->cpu, rsp, rdp, rdp->passed_quiesc_completed);
}

#ifdef CONFIG_HOTPLUG_CPU

849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896
/*
 * Move a dying CPU's RCU callbacks to the ->orphan_cbs_list for the
 * specified flavor of RCU.  The callbacks will be adopted by the next
 * _rcu_barrier() invocation or by the CPU_DEAD notifier, whichever
 * comes first.  Because this is invoked from the CPU_DYING notifier,
 * irqs are already disabled.
 */
static void rcu_send_cbs_to_orphanage(struct rcu_state *rsp)
{
	int i;
	struct rcu_data *rdp = rsp->rda[smp_processor_id()];

	if (rdp->nxtlist == NULL)
		return;  /* irqs disabled, so comparison is stable. */
	spin_lock(&rsp->onofflock);  /* irqs already disabled. */
	*rsp->orphan_cbs_tail = rdp->nxtlist;
	rsp->orphan_cbs_tail = rdp->nxttail[RCU_NEXT_TAIL];
	rdp->nxtlist = NULL;
	for (i = 0; i < RCU_NEXT_SIZE; i++)
		rdp->nxttail[i] = &rdp->nxtlist;
	rsp->orphan_qlen += rdp->qlen;
	rdp->qlen = 0;
	spin_unlock(&rsp->onofflock);  /* irqs remain disabled. */
}

/*
 * Adopt previously orphaned RCU callbacks.
 */
static void rcu_adopt_orphan_cbs(struct rcu_state *rsp)
{
	unsigned long flags;
	struct rcu_data *rdp;

	spin_lock_irqsave(&rsp->onofflock, flags);
	rdp = rsp->rda[smp_processor_id()];
	if (rsp->orphan_cbs_list == NULL) {
		spin_unlock_irqrestore(&rsp->onofflock, flags);
		return;
	}
	*rdp->nxttail[RCU_NEXT_TAIL] = rsp->orphan_cbs_list;
	rdp->nxttail[RCU_NEXT_TAIL] = rsp->orphan_cbs_tail;
	rdp->qlen += rsp->orphan_qlen;
	rsp->orphan_cbs_list = NULL;
	rsp->orphan_cbs_tail = &rsp->orphan_cbs_list;
	rsp->orphan_qlen = 0;
	spin_unlock_irqrestore(&rsp->onofflock, flags);
}

897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912
/*
 * Remove the outgoing CPU from the bitmasks in the rcu_node hierarchy
 * and move all callbacks from the outgoing CPU to the current one.
 */
static void __rcu_offline_cpu(int cpu, struct rcu_state *rsp)
{
	unsigned long flags;
	long lastcomp;
	unsigned long mask;
	struct rcu_data *rdp = rsp->rda[cpu];
	struct rcu_node *rnp;

	/* Exclude any attempts to start a new grace period. */
	spin_lock_irqsave(&rsp->onofflock, flags);

	/* Remove the outgoing CPU from the masks in the rcu_node hierarchy. */
913
	rnp = rdp->mynode;	/* this is the outgoing CPU's rnp. */
914 915 916 917 918
	mask = rdp->grpmask;	/* rnp->grplo is constant. */
	do {
		spin_lock(&rnp->lock);		/* irqs already disabled. */
		rnp->qsmaskinit &= ~mask;
		if (rnp->qsmaskinit != 0) {
919
			spin_unlock(&rnp->lock); /* irqs remain disabled. */
920 921
			break;
		}
922 923 924 925 926 927 928 929 930 931 932 933 934 935

		/*
		 * If there was a task blocking the current grace period,
		 * and if all CPUs have checked in, we need to propagate
		 * the quiescent state up the rcu_node hierarchy.  But that
		 * is inconvenient at the moment due to deadlock issues if
		 * this should end the current grace period.  So set the
		 * offlined CPU's bit in ->qsmask in order to force the
		 * next force_quiescent_state() invocation to clean up this
		 * mess in a deadlock-free manner.
		 */
		if (rcu_preempt_offline_tasks(rsp, rnp, rdp) && !rnp->qsmask)
			rnp->qsmask |= mask;

936
		mask = rnp->grpmask;
937
		spin_unlock(&rnp->lock);	/* irqs remain disabled. */
938 939 940 941
		rnp = rnp->parent;
	} while (rnp != NULL);
	lastcomp = rsp->completed;

942
	spin_unlock_irqrestore(&rsp->onofflock, flags);
943

944
	rcu_adopt_orphan_cbs(rsp);
945 946 947 948 949 950 951 952 953 954
}

/*
 * Remove the specified CPU from the RCU hierarchy and move any pending
 * callbacks that it might have to the current CPU.  This code assumes
 * that at least one CPU in the system will remain running at all times.
 * Any attempt to offline -all- CPUs is likely to strand RCU callbacks.
 */
static void rcu_offline_cpu(int cpu)
{
955
	__rcu_offline_cpu(cpu, &rcu_sched_state);
956
	__rcu_offline_cpu(cpu, &rcu_bh_state);
957
	rcu_preempt_offline_cpu(cpu);
958 959 960 961
}

#else /* #ifdef CONFIG_HOTPLUG_CPU */

962 963 964 965 966 967 968 969
static void rcu_send_cbs_to_orphanage(struct rcu_state *rsp)
{
}

static void rcu_adopt_orphan_cbs(struct rcu_state *rsp)
{
}

970 971 972 973 974 975 976 977 978 979
static void rcu_offline_cpu(int cpu)
{
}

#endif /* #else #ifdef CONFIG_HOTPLUG_CPU */

/*
 * Invoke any RCU callbacks that have made it to the end of their grace
 * period.  Thottle as specified by rdp->blimit.
 */
980
static void rcu_do_batch(struct rcu_state *rsp, struct rcu_data *rdp)
981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032
{
	unsigned long flags;
	struct rcu_head *next, *list, **tail;
	int count;

	/* If no callbacks are ready, just return.*/
	if (!cpu_has_callbacks_ready_to_invoke(rdp))
		return;

	/*
	 * Extract the list of ready callbacks, disabling to prevent
	 * races with call_rcu() from interrupt handlers.
	 */
	local_irq_save(flags);
	list = rdp->nxtlist;
	rdp->nxtlist = *rdp->nxttail[RCU_DONE_TAIL];
	*rdp->nxttail[RCU_DONE_TAIL] = NULL;
	tail = rdp->nxttail[RCU_DONE_TAIL];
	for (count = RCU_NEXT_SIZE - 1; count >= 0; count--)
		if (rdp->nxttail[count] == rdp->nxttail[RCU_DONE_TAIL])
			rdp->nxttail[count] = &rdp->nxtlist;
	local_irq_restore(flags);

	/* Invoke callbacks. */
	count = 0;
	while (list) {
		next = list->next;
		prefetch(next);
		list->func(list);
		list = next;
		if (++count >= rdp->blimit)
			break;
	}

	local_irq_save(flags);

	/* Update count, and requeue any remaining callbacks. */
	rdp->qlen -= count;
	if (list != NULL) {
		*tail = rdp->nxtlist;
		rdp->nxtlist = list;
		for (count = 0; count < RCU_NEXT_SIZE; count++)
			if (&rdp->nxtlist == rdp->nxttail[count])
				rdp->nxttail[count] = tail;
			else
				break;
	}

	/* Reinstate batch limit if we have worked down the excess. */
	if (rdp->blimit == LONG_MAX && rdp->qlen <= qlowmark)
		rdp->blimit = blimit;

1033 1034 1035 1036 1037 1038 1039
	/* Reset ->qlen_last_fqs_check trigger if enough CBs have drained. */
	if (rdp->qlen == 0 && rdp->qlen_last_fqs_check != 0) {
		rdp->qlen_last_fqs_check = 0;
		rdp->n_force_qs_snap = rsp->n_force_qs;
	} else if (rdp->qlen < rdp->qlen_last_fqs_check - qhimark)
		rdp->qlen_last_fqs_check = rdp->qlen;

1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057
	local_irq_restore(flags);

	/* Re-raise the RCU softirq if there are callbacks remaining. */
	if (cpu_has_callbacks_ready_to_invoke(rdp))
		raise_softirq(RCU_SOFTIRQ);
}

/*
 * Check to see if this CPU is in a non-context-switch quiescent state
 * (user mode or idle loop for rcu, non-softirq execution for rcu_bh).
 * Also schedule the RCU softirq handler.
 *
 * This function must be called with hardirqs disabled.  It is normally
 * invoked from the scheduling-clock interrupt.  If rcu_pending returns
 * false, there is no point in invoking rcu_check_callbacks().
 */
void rcu_check_callbacks(int cpu, int user)
{
1058 1059
	if (!rcu_pending(cpu))
		return; /* if nothing for RCU to do. */
1060
	if (user ||
1061 1062
	    (idle_cpu(cpu) && rcu_scheduler_active &&
	     !in_softirq() && hardirq_count() <= (1 << HARDIRQ_SHIFT))) {
1063 1064 1065 1066 1067

		/*
		 * Get here if this CPU took its interrupt from user
		 * mode or from the idle loop, and if this is not a
		 * nested interrupt.  In this case, the CPU is in
1068
		 * a quiescent state, so note it.
1069 1070
		 *
		 * No memory barrier is required here because both
1071 1072 1073
		 * rcu_sched_qs() and rcu_bh_qs() reference only CPU-local
		 * variables that other CPUs neither access nor modify,
		 * at least not while the corresponding CPU is online.
1074 1075
		 */

1076 1077
		rcu_sched_qs(cpu);
		rcu_bh_qs(cpu);
1078 1079 1080 1081 1082 1083 1084

	} else if (!in_softirq()) {

		/*
		 * Get here if this CPU did not take its interrupt from
		 * softirq, in other words, if it is not interrupting
		 * a rcu_bh read-side critical section.  This is an _bh
1085
		 * critical section, so note it.
1086 1087
		 */

1088
		rcu_bh_qs(cpu);
1089
	}
1090
	rcu_preempt_check_callbacks(cpu);
1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108
	raise_softirq(RCU_SOFTIRQ);
}

#ifdef CONFIG_SMP

/*
 * Scan the leaf rcu_node structures, processing dyntick state for any that
 * have not yet encountered a quiescent state, using the function specified.
 * Returns 1 if the current grace period ends while scanning (possibly
 * because we made it end).
 */
static int rcu_process_dyntick(struct rcu_state *rsp, long lastcomp,
			       int (*f)(struct rcu_data *))
{
	unsigned long bit;
	int cpu;
	unsigned long flags;
	unsigned long mask;
1109
	struct rcu_node *rnp;
1110

1111
	rcu_for_each_leaf_node(rsp, rnp) {
1112
		mask = 0;
1113
		spin_lock_irqsave(&rnp->lock, flags);
1114
		if (rsp->completed != lastcomp) {
1115
			spin_unlock_irqrestore(&rnp->lock, flags);
1116 1117
			return 1;
		}
1118 1119
		if (rnp->qsmask == 0) {
			spin_unlock_irqrestore(&rnp->lock, flags);
1120 1121
			continue;
		}
1122
		cpu = rnp->grplo;
1123
		bit = 1;
1124 1125
		for (; cpu <= rnp->grphi; cpu++, bit <<= 1) {
			if ((rnp->qsmask & bit) != 0 && f(rsp->rda[cpu]))
1126 1127 1128 1129
				mask |= bit;
		}
		if (mask != 0 && rsp->completed == lastcomp) {

1130 1131
			/* cpu_quiet_msk() releases rnp->lock. */
			cpu_quiet_msk(mask, rsp, rnp, flags);
1132 1133
			continue;
		}
1134
		spin_unlock_irqrestore(&rnp->lock, flags);
1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149
	}
	return 0;
}

/*
 * Force quiescent states on reluctant CPUs, and also detect which
 * CPUs are in dyntick-idle mode.
 */
static void force_quiescent_state(struct rcu_state *rsp, int relaxed)
{
	unsigned long flags;
	long lastcomp;
	struct rcu_node *rnp = rcu_get_root(rsp);
	u8 signaled;

1150
	if (!rcu_gp_in_progress(rsp))
1151 1152 1153 1154 1155 1156
		return;  /* No grace period in progress, nothing to force. */
	if (!spin_trylock_irqsave(&rsp->fqslock, flags)) {
		rsp->n_force_qs_lh++; /* Inexact, can lose counts.  Tough! */
		return;	/* Someone else is already on the job. */
	}
	if (relaxed &&
1157
	    (long)(rsp->jiffies_force_qs - jiffies) >= 0)
1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170
		goto unlock_ret; /* no emergency and done recently. */
	rsp->n_force_qs++;
	spin_lock(&rnp->lock);
	lastcomp = rsp->completed;
	signaled = rsp->signaled;
	rsp->jiffies_force_qs = jiffies + RCU_JIFFIES_TILL_FORCE_QS;
	if (lastcomp == rsp->gpnum) {
		rsp->n_force_qs_ngp++;
		spin_unlock(&rnp->lock);
		goto unlock_ret;  /* no GP in progress, time updated. */
	}
	spin_unlock(&rnp->lock);
	switch (signaled) {
1171
	case RCU_GP_IDLE:
1172 1173
	case RCU_GP_INIT:

1174
		break; /* grace period idle or initializing, ignore. */
1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187

	case RCU_SAVE_DYNTICK:

		if (RCU_SIGNAL_INIT != RCU_SAVE_DYNTICK)
			break; /* So gcc recognizes the dead code. */

		/* Record dyntick-idle state. */
		if (rcu_process_dyntick(rsp, lastcomp,
					dyntick_save_progress_counter))
			goto unlock_ret;

		/* Update state, record completion counter. */
		spin_lock(&rnp->lock);
1188 1189
		if (lastcomp == rsp->completed &&
		    rsp->signaled == RCU_SAVE_DYNTICK) {
1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229
			rsp->signaled = RCU_FORCE_QS;
			dyntick_record_completed(rsp, lastcomp);
		}
		spin_unlock(&rnp->lock);
		break;

	case RCU_FORCE_QS:

		/* Check dyntick-idle state, send IPI to laggarts. */
		if (rcu_process_dyntick(rsp, dyntick_recall_completed(rsp),
					rcu_implicit_dynticks_qs))
			goto unlock_ret;

		/* Leave state in case more forcing is required. */

		break;
	}
unlock_ret:
	spin_unlock_irqrestore(&rsp->fqslock, flags);
}

#else /* #ifdef CONFIG_SMP */

static void force_quiescent_state(struct rcu_state *rsp, int relaxed)
{
	set_need_resched();
}

#endif /* #else #ifdef CONFIG_SMP */

/*
 * This does the RCU processing work from softirq context for the
 * specified rcu_state and rcu_data structures.  This may be called
 * only from the CPU to whom the rdp belongs.
 */
static void
__rcu_process_callbacks(struct rcu_state *rsp, struct rcu_data *rdp)
{
	unsigned long flags;

1230 1231
	WARN_ON_ONCE(rdp->beenonline == 0);

1232 1233 1234 1235
	/*
	 * If an RCU GP has gone long enough, go check for dyntick
	 * idle CPUs and, if needed, send resched IPIs.
	 */
1236
	if ((long)(ACCESS_ONCE(rsp->jiffies_force_qs) - jiffies) < 0)
1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254
		force_quiescent_state(rsp, 1);

	/*
	 * Advance callbacks in response to end of earlier grace
	 * period that some other CPU ended.
	 */
	rcu_process_gp_end(rsp, rdp);

	/* Update RCU state based on any recent quiescent states. */
	rcu_check_quiescent_state(rsp, rdp);

	/* Does this CPU require a not-yet-started grace period? */
	if (cpu_needs_another_gp(rsp, rdp)) {
		spin_lock_irqsave(&rcu_get_root(rsp)->lock, flags);
		rcu_start_gp(rsp, flags);  /* releases above lock */
	}

	/* If there are callbacks ready, invoke them. */
1255
	rcu_do_batch(rsp, rdp);
1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269
}

/*
 * Do softirq processing for the current CPU.
 */
static void rcu_process_callbacks(struct softirq_action *unused)
{
	/*
	 * Memory references from any prior RCU read-side critical sections
	 * executed by the interrupted code must be seen before any RCU
	 * grace-period manipulations below.
	 */
	smp_mb(); /* See above block comment. */

1270 1271
	__rcu_process_callbacks(&rcu_sched_state,
				&__get_cpu_var(rcu_sched_data));
1272
	__rcu_process_callbacks(&rcu_bh_state, &__get_cpu_var(rcu_bh_data));
1273
	rcu_preempt_process_callbacks();
1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310

	/*
	 * Memory references from any later RCU read-side critical sections
	 * executed by the interrupted code must be seen after any RCU
	 * grace-period manipulations above.
	 */
	smp_mb(); /* See above block comment. */
}

static void
__call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu),
	   struct rcu_state *rsp)
{
	unsigned long flags;
	struct rcu_data *rdp;

	head->func = func;
	head->next = NULL;

	smp_mb(); /* Ensure RCU update seen before callback registry. */

	/*
	 * Opportunistically note grace-period endings and beginnings.
	 * Note that we might see a beginning right after we see an
	 * end, but never vice versa, since this CPU has to pass through
	 * a quiescent state betweentimes.
	 */
	local_irq_save(flags);
	rdp = rsp->rda[smp_processor_id()];
	rcu_process_gp_end(rsp, rdp);
	check_for_new_grace_period(rsp, rdp);

	/* Add the callback to our list. */
	*rdp->nxttail[RCU_NEXT_TAIL] = head;
	rdp->nxttail[RCU_NEXT_TAIL] = &head->next;

	/* Start a new grace period if one not already started. */
1311
	if (!rcu_gp_in_progress(rsp)) {
1312 1313 1314 1315 1316 1317 1318
		unsigned long nestflag;
		struct rcu_node *rnp_root = rcu_get_root(rsp);

		spin_lock_irqsave(&rnp_root->lock, nestflag);
		rcu_start_gp(rsp, nestflag);  /* releases rnp_root->lock. */
	}

1319 1320 1321 1322 1323 1324 1325 1326
	/*
	 * Force the grace period if too many callbacks or too long waiting.
	 * Enforce hysteresis, and don't invoke force_quiescent_state()
	 * if some other CPU has recently done so.  Also, don't bother
	 * invoking force_quiescent_state() if the newly enqueued callback
	 * is the only one waiting for a grace period to complete.
	 */
	if (unlikely(++rdp->qlen > rdp->qlen_last_fqs_check + qhimark)) {
1327
		rdp->blimit = LONG_MAX;
1328 1329 1330 1331 1332
		if (rsp->n_force_qs == rdp->n_force_qs_snap &&
		    *rdp->nxttail[RCU_DONE_TAIL] != head)
			force_quiescent_state(rsp, 0);
		rdp->n_force_qs_snap = rsp->n_force_qs;
		rdp->qlen_last_fqs_check = rdp->qlen;
1333
	} else if ((long)(ACCESS_ONCE(rsp->jiffies_force_qs) - jiffies) < 0)
1334 1335 1336 1337 1338
		force_quiescent_state(rsp, 1);
	local_irq_restore(flags);
}

/*
1339
 * Queue an RCU-sched callback for invocation after a grace period.
1340
 */
1341
void call_rcu_sched(struct rcu_head *head, void (*func)(struct rcu_head *rcu))
1342
{
1343
	__call_rcu(head, func, &rcu_sched_state);
1344
}
1345
EXPORT_SYMBOL_GPL(call_rcu_sched);
1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370

/*
 * Queue an RCU for invocation after a quicker grace period.
 */
void call_rcu_bh(struct rcu_head *head, void (*func)(struct rcu_head *rcu))
{
	__call_rcu(head, func, &rcu_bh_state);
}
EXPORT_SYMBOL_GPL(call_rcu_bh);

/*
 * Check to see if there is any immediate RCU-related work to be done
 * by the current CPU, for the specified type of RCU, returning 1 if so.
 * The checks are in order of increasing expense: checks that can be
 * carried out against CPU-local state are performed first.  However,
 * we must check for CPU stalls first, else we might not get a chance.
 */
static int __rcu_pending(struct rcu_state *rsp, struct rcu_data *rdp)
{
	rdp->n_rcu_pending++;

	/* Check for CPU stalls, if enabled. */
	check_cpu_stall(rsp, rdp);

	/* Is the RCU core waiting for a quiescent state from this CPU? */
1371 1372
	if (rdp->qs_pending) {
		rdp->n_rp_qs_pending++;
1373
		return 1;
1374
	}
1375 1376

	/* Does this CPU have callbacks ready to invoke? */
1377 1378
	if (cpu_has_callbacks_ready_to_invoke(rdp)) {
		rdp->n_rp_cb_ready++;
1379
		return 1;
1380
	}
1381 1382

	/* Has RCU gone idle with this CPU needing another grace period? */
1383 1384
	if (cpu_needs_another_gp(rsp, rdp)) {
		rdp->n_rp_cpu_needs_gp++;
1385
		return 1;
1386
	}
1387 1388

	/* Has another RCU grace period completed?  */
1389 1390
	if (ACCESS_ONCE(rsp->completed) != rdp->completed) { /* outside lock */
		rdp->n_rp_gp_completed++;
1391
		return 1;
1392
	}
1393 1394

	/* Has a new RCU grace period started? */
1395 1396
	if (ACCESS_ONCE(rsp->gpnum) != rdp->gpnum) { /* outside lock */
		rdp->n_rp_gp_started++;
1397
		return 1;
1398
	}
1399 1400

	/* Has an RCU GP gone long enough to send resched IPIs &c? */
1401
	if (rcu_gp_in_progress(rsp) &&
1402 1403
	    ((long)(ACCESS_ONCE(rsp->jiffies_force_qs) - jiffies) < 0)) {
		rdp->n_rp_need_fqs++;
1404
		return 1;
1405
	}
1406 1407

	/* nothing to do */
1408
	rdp->n_rp_need_nothing++;
1409 1410 1411 1412 1413 1414 1415 1416
	return 0;
}

/*
 * Check to see if there is any immediate RCU-related work to be done
 * by the current CPU, returning 1 if so.  This function is part of the
 * RCU implementation; it is -not- an exported member of the RCU API.
 */
1417
static int rcu_pending(int cpu)
1418
{
1419
	return __rcu_pending(&rcu_sched_state, &per_cpu(rcu_sched_data, cpu)) ||
1420 1421
	       __rcu_pending(&rcu_bh_state, &per_cpu(rcu_bh_data, cpu)) ||
	       rcu_preempt_pending(cpu);
1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432
}

/*
 * Check to see if any future RCU-related work will need to be done
 * by the current CPU, even if none need be done immediately, returning
 * 1 if so.  This function is part of the RCU implementation; it is -not-
 * an exported member of the RCU API.
 */
int rcu_needs_cpu(int cpu)
{
	/* RCU callbacks either ready or pending? */
1433
	return per_cpu(rcu_sched_data, cpu).nxtlist ||
1434 1435
	       per_cpu(rcu_bh_data, cpu).nxtlist ||
	       rcu_preempt_needs_cpu(cpu);
1436 1437
}

1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467
static DEFINE_PER_CPU(struct rcu_head, rcu_barrier_head) = {NULL};
static atomic_t rcu_barrier_cpu_count;
static DEFINE_MUTEX(rcu_barrier_mutex);
static struct completion rcu_barrier_completion;

static void rcu_barrier_callback(struct rcu_head *notused)
{
	if (atomic_dec_and_test(&rcu_barrier_cpu_count))
		complete(&rcu_barrier_completion);
}

/*
 * Called with preemption disabled, and from cross-cpu IRQ context.
 */
static void rcu_barrier_func(void *type)
{
	int cpu = smp_processor_id();
	struct rcu_head *head = &per_cpu(rcu_barrier_head, cpu);
	void (*call_rcu_func)(struct rcu_head *head,
			      void (*func)(struct rcu_head *head));

	atomic_inc(&rcu_barrier_cpu_count);
	call_rcu_func = type;
	call_rcu_func(head, rcu_barrier_callback);
}

/*
 * Orchestrate the specified type of RCU barrier, waiting for all
 * RCU callbacks of the specified type to complete.
 */
1468 1469
static void _rcu_barrier(struct rcu_state *rsp,
			 void (*call_rcu_func)(struct rcu_head *head,
1470 1471 1472
					       void (*func)(struct rcu_head *head)))
{
	BUG_ON(in_interrupt());
1473
	/* Take mutex to serialize concurrent rcu_barrier() requests. */
1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485
	mutex_lock(&rcu_barrier_mutex);
	init_completion(&rcu_barrier_completion);
	/*
	 * Initialize rcu_barrier_cpu_count to 1, then invoke
	 * rcu_barrier_func() on each CPU, so that each CPU also has
	 * incremented rcu_barrier_cpu_count.  Only then is it safe to
	 * decrement rcu_barrier_cpu_count -- otherwise the first CPU
	 * might complete its grace period before all of the other CPUs
	 * did their increment, causing this function to return too
	 * early.
	 */
	atomic_set(&rcu_barrier_cpu_count, 1);
1486 1487
	preempt_disable(); /* stop CPU_DYING from filling orphan_cbs_list */
	rcu_adopt_orphan_cbs(rsp);
1488
	on_each_cpu(rcu_barrier_func, (void *)call_rcu_func, 1);
1489
	preempt_enable(); /* CPU_DYING can again fill orphan_cbs_list */
1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500
	if (atomic_dec_and_test(&rcu_barrier_cpu_count))
		complete(&rcu_barrier_completion);
	wait_for_completion(&rcu_barrier_completion);
	mutex_unlock(&rcu_barrier_mutex);
}

/**
 * rcu_barrier_bh - Wait until all in-flight call_rcu_bh() callbacks complete.
 */
void rcu_barrier_bh(void)
{
1501
	_rcu_barrier(&rcu_bh_state, call_rcu_bh);
1502 1503 1504 1505 1506 1507 1508 1509
}
EXPORT_SYMBOL_GPL(rcu_barrier_bh);

/**
 * rcu_barrier_sched - Wait for in-flight call_rcu_sched() callbacks.
 */
void rcu_barrier_sched(void)
{
1510
	_rcu_barrier(&rcu_sched_state, call_rcu_sched);
1511 1512 1513
}
EXPORT_SYMBOL_GPL(rcu_barrier_sched);

1514
/*
1515
 * Do boot-time initialization of a CPU's per-CPU RCU data.
1516
 */
1517 1518
static void __init
rcu_boot_init_percpu_data(int cpu, struct rcu_state *rsp)
1519 1520 1521
{
	unsigned long flags;
	int i;
1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543
	struct rcu_data *rdp = rsp->rda[cpu];
	struct rcu_node *rnp = rcu_get_root(rsp);

	/* Set up local state, ensuring consistent view of global state. */
	spin_lock_irqsave(&rnp->lock, flags);
	rdp->grpmask = 1UL << (cpu - rdp->mynode->grplo);
	rdp->nxtlist = NULL;
	for (i = 0; i < RCU_NEXT_SIZE; i++)
		rdp->nxttail[i] = &rdp->nxtlist;
	rdp->qlen = 0;
#ifdef CONFIG_NO_HZ
	rdp->dynticks = &per_cpu(rcu_dynticks, cpu);
#endif /* #ifdef CONFIG_NO_HZ */
	rdp->cpu = cpu;
	spin_unlock_irqrestore(&rnp->lock, flags);
}

/*
 * Initialize a CPU's per-CPU RCU data.  Note that only one online or
 * offline event can be happening at a given time.  Note also that we
 * can accept some slop in the rsp->completed access due to the fact
 * that this CPU cannot possibly have any RCU callbacks in flight yet.
1544
 */
1545
static void __cpuinit
1546
rcu_init_percpu_data(int cpu, struct rcu_state *rsp, int preemptable)
1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561
{
	unsigned long flags;
	long lastcomp;
	unsigned long mask;
	struct rcu_data *rdp = rsp->rda[cpu];
	struct rcu_node *rnp = rcu_get_root(rsp);

	/* Set up local state, ensuring consistent view of global state. */
	spin_lock_irqsave(&rnp->lock, flags);
	lastcomp = rsp->completed;
	rdp->completed = lastcomp;
	rdp->gpnum = lastcomp;
	rdp->passed_quiesc = 0;  /* We could be racing with new GP, */
	rdp->qs_pending = 1;	 /*  so set up to respond to current GP. */
	rdp->beenonline = 1;	 /* We have now been online. */
1562
	rdp->preemptable = preemptable;
1563
	rdp->passed_quiesc_completed = lastcomp - 1;
1564 1565
	rdp->qlen_last_fqs_check = 0;
	rdp->n_force_qs_snap = rsp->n_force_qs;
1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588
	rdp->blimit = blimit;
	spin_unlock(&rnp->lock);		/* irqs remain disabled. */

	/*
	 * A new grace period might start here.  If so, we won't be part
	 * of it, but that is OK, as we are currently in a quiescent state.
	 */

	/* Exclude any attempts to start a new GP on large systems. */
	spin_lock(&rsp->onofflock);		/* irqs already disabled. */

	/* Add CPU to rcu_node bitmasks. */
	rnp = rdp->mynode;
	mask = rdp->grpmask;
	do {
		/* Exclude any attempts to start a new GP on small systems. */
		spin_lock(&rnp->lock);	/* irqs already disabled. */
		rnp->qsmaskinit |= mask;
		mask = rnp->grpmask;
		spin_unlock(&rnp->lock); /* irqs already disabled. */
		rnp = rnp->parent;
	} while (rnp != NULL && !(rnp->qsmaskinit & mask));

1589
	spin_unlock_irqrestore(&rsp->onofflock, flags);
1590 1591 1592 1593
}

static void __cpuinit rcu_online_cpu(int cpu)
{
1594 1595 1596
	rcu_init_percpu_data(cpu, &rcu_sched_state, 0);
	rcu_init_percpu_data(cpu, &rcu_bh_state, 0);
	rcu_preempt_init_percpu_data(cpu);
1597 1598 1599
}

/*
1600
 * Handle CPU online/offline notification events.
1601
 */
1602 1603
int __cpuinit rcu_cpu_notify(struct notifier_block *self,
			     unsigned long action, void *hcpu)
1604 1605 1606 1607 1608 1609 1610 1611
{
	long cpu = (long)hcpu;

	switch (action) {
	case CPU_UP_PREPARE:
	case CPU_UP_PREPARE_FROZEN:
		rcu_online_cpu(cpu);
		break;
1612 1613 1614
	case CPU_DYING:
	case CPU_DYING_FROZEN:
		/*
1615
		 * preempt_disable() in _rcu_barrier() prevents stop_machine(),
1616
		 * so when "on_each_cpu(rcu_barrier_func, (void *)type, 1);"
1617 1618 1619 1620 1621 1622
		 * returns, all online cpus have queued rcu_barrier_func().
		 * The dying CPU clears its cpu_online_mask bit and
		 * moves all of its RCU callbacks to ->orphan_cbs_list
		 * in the context of stop_machine(), so subsequent calls
		 * to _rcu_barrier() will adopt these callbacks and only
		 * then queue rcu_barrier_func() on all remaining CPUs.
1623
		 */
1624 1625 1626
		rcu_send_cbs_to_orphanage(&rcu_bh_state);
		rcu_send_cbs_to_orphanage(&rcu_sched_state);
		rcu_preempt_send_cbs_to_orphanage();
1627
		break;
1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689
	case CPU_DEAD:
	case CPU_DEAD_FROZEN:
	case CPU_UP_CANCELED:
	case CPU_UP_CANCELED_FROZEN:
		rcu_offline_cpu(cpu);
		break;
	default:
		break;
	}
	return NOTIFY_OK;
}

/*
 * Compute the per-level fanout, either using the exact fanout specified
 * or balancing the tree, depending on CONFIG_RCU_FANOUT_EXACT.
 */
#ifdef CONFIG_RCU_FANOUT_EXACT
static void __init rcu_init_levelspread(struct rcu_state *rsp)
{
	int i;

	for (i = NUM_RCU_LVLS - 1; i >= 0; i--)
		rsp->levelspread[i] = CONFIG_RCU_FANOUT;
}
#else /* #ifdef CONFIG_RCU_FANOUT_EXACT */
static void __init rcu_init_levelspread(struct rcu_state *rsp)
{
	int ccur;
	int cprv;
	int i;

	cprv = NR_CPUS;
	for (i = NUM_RCU_LVLS - 1; i >= 0; i--) {
		ccur = rsp->levelcnt[i];
		rsp->levelspread[i] = (cprv + ccur - 1) / ccur;
		cprv = ccur;
	}
}
#endif /* #else #ifdef CONFIG_RCU_FANOUT_EXACT */

/*
 * Helper function for rcu_init() that initializes one rcu_state structure.
 */
static void __init rcu_init_one(struct rcu_state *rsp)
{
	int cpustride = 1;
	int i;
	int j;
	struct rcu_node *rnp;

	/* Initialize the level-tracking arrays. */

	for (i = 1; i < NUM_RCU_LVLS; i++)
		rsp->level[i] = rsp->level[i - 1] + rsp->levelcnt[i - 1];
	rcu_init_levelspread(rsp);

	/* Initialize the elements themselves, starting from the leaves. */

	for (i = NUM_RCU_LVLS - 1; i >= 0; i--) {
		cpustride *= rsp->levelspread[i];
		rnp = rsp->level[i];
		for (j = 0; j < rsp->levelcnt[i]; j++, rnp++) {
1690
			spin_lock_init(&rnp->lock);
1691
			rnp->gpnum = 0;
1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708
			rnp->qsmask = 0;
			rnp->qsmaskinit = 0;
			rnp->grplo = j * cpustride;
			rnp->grphi = (j + 1) * cpustride - 1;
			if (rnp->grphi >= NR_CPUS)
				rnp->grphi = NR_CPUS - 1;
			if (i == 0) {
				rnp->grpnum = 0;
				rnp->grpmask = 0;
				rnp->parent = NULL;
			} else {
				rnp->grpnum = j % rsp->levelspread[i - 1];
				rnp->grpmask = 1UL << rnp->grpnum;
				rnp->parent = rsp->level[i - 1] +
					      j / rsp->levelspread[i - 1];
			}
			rnp->level = i;
1709 1710
			INIT_LIST_HEAD(&rnp->blocked_tasks[0]);
			INIT_LIST_HEAD(&rnp->blocked_tasks[1]);
1711 1712
		}
	}
1713
	lockdep_set_class(&rcu_get_root(rsp)->lock, &rcu_root_class);
1714 1715 1716
}

/*
1717 1718 1719
 * Helper macro for __rcu_init() and __rcu_init_preempt().  To be used
 * nowhere else!  Assigns leaf node pointers into each CPU's rcu_data
 * structure.
1720
 */
1721
#define RCU_INIT_FLAVOR(rsp, rcu_data) \
1722
do { \
1723 1724 1725 1726
	int i; \
	int j; \
	struct rcu_node *rnp; \
	\
1727
	rcu_init_one(rsp); \
1728 1729 1730 1731 1732 1733 1734
	rnp = (rsp)->level[NUM_RCU_LVLS - 1]; \
	j = 0; \
	for_each_possible_cpu(i) { \
		if (i > rnp[j].grphi) \
			j++; \
		per_cpu(rcu_data, i).mynode = &rnp[j]; \
		(rsp)->rda[i] = &per_cpu(rcu_data, i); \
1735
		rcu_boot_init_percpu_data(i, rsp); \
1736 1737 1738 1739 1740
	} \
} while (0)

void __init __rcu_init(void)
{
1741
	rcu_bootup_announce();
1742 1743 1744
#ifdef CONFIG_RCU_CPU_STALL_DETECTOR
	printk(KERN_INFO "RCU-based detection of stalled CPUs is enabled.\n");
#endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */
1745 1746
	RCU_INIT_FLAVOR(&rcu_sched_state, rcu_sched_data);
	RCU_INIT_FLAVOR(&rcu_bh_state, rcu_bh_data);
1747
	__rcu_init_preempt();
1748
	open_softirq(RCU_SOFTIRQ, rcu_process_callbacks);
1749 1750
}

1751
#include "rcutree_plugin.h"