rcutree_plugin.h 46.4 KB
Newer Older
1 2 3
/*
 * Read-Copy Update mechanism for mutual exclusion (tree-based version)
 * Internal non-public definitions that provide either classic
P
Paul E. McKenney 已提交
4
 * or preemptible semantics.
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
 *
 * 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 Red Hat, 2009
 * Copyright IBM Corporation, 2009
 *
 * Author: Ingo Molnar <mingo@elte.hu>
 *	   Paul E. McKenney <paulmck@linux.vnet.ibm.com>
 */

27
#include <linux/delay.h>
28
#include <linux/stop_machine.h>
29

30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56
/*
 * Check the RCU kernel configuration parameters and print informative
 * messages about anything out of the ordinary.  If you like #ifdef, you
 * will love this function.
 */
static void __init rcu_bootup_announce_oddness(void)
{
#ifdef CONFIG_RCU_TRACE
	printk(KERN_INFO "\tRCU debugfs-based tracing is enabled.\n");
#endif
#if (defined(CONFIG_64BIT) && CONFIG_RCU_FANOUT != 64) || (!defined(CONFIG_64BIT) && CONFIG_RCU_FANOUT != 32)
	printk(KERN_INFO "\tCONFIG_RCU_FANOUT set to non-default value of %d\n",
	       CONFIG_RCU_FANOUT);
#endif
#ifdef CONFIG_RCU_FANOUT_EXACT
	printk(KERN_INFO "\tHierarchical RCU autobalancing is disabled.\n");
#endif
#ifdef CONFIG_RCU_FAST_NO_HZ
	printk(KERN_INFO
	       "\tRCU dyntick-idle grace-period acceleration is enabled.\n");
#endif
#ifdef CONFIG_PROVE_RCU
	printk(KERN_INFO "\tRCU lockdep checking is enabled.\n");
#endif
#ifdef CONFIG_RCU_TORTURE_TEST_RUNNABLE
	printk(KERN_INFO "\tRCU torture testing starts during boot.\n");
#endif
57
#if defined(CONFIG_TREE_PREEMPT_RCU) && !defined(CONFIG_RCU_CPU_STALL_VERBOSE)
58 59 60 61 62 63 64
	printk(KERN_INFO "\tVerbose stalled-CPUs detection is disabled.\n");
#endif
#if NUM_RCU_LVL_4 != 0
	printk(KERN_INFO "\tExperimental four-level hierarchy is enabled.\n");
#endif
}

65 66 67 68
#ifdef CONFIG_TREE_PREEMPT_RCU

struct rcu_state rcu_preempt_state = RCU_STATE_INITIALIZER(rcu_preempt_state);
DEFINE_PER_CPU(struct rcu_data, rcu_preempt_data);
69
static struct rcu_state *rcu_state = &rcu_preempt_state;
70

71 72
static int rcu_preempted_readers_exp(struct rcu_node *rnp);

73 74 75
/*
 * Tell them what RCU they are running.
 */
76
static void __init rcu_bootup_announce(void)
77
{
P
Paul E. McKenney 已提交
78
	printk(KERN_INFO "Preemptible hierarchical RCU implementation.\n");
79
	rcu_bootup_announce_oddness();
80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100
}

/*
 * Return the number of RCU-preempt batches processed thus far
 * for debug and statistics.
 */
long rcu_batches_completed_preempt(void)
{
	return rcu_preempt_state.completed;
}
EXPORT_SYMBOL_GPL(rcu_batches_completed_preempt);

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

101 102 103 104 105 106 107 108 109
/*
 * Force a quiescent state for preemptible RCU.
 */
void rcu_force_quiescent_state(void)
{
	force_quiescent_state(&rcu_preempt_state, 0);
}
EXPORT_SYMBOL_GPL(rcu_force_quiescent_state);

110
/*
P
Paul E. McKenney 已提交
111
 * Record a preemptible-RCU quiescent state for the specified CPU.  Note
112 113 114
 * that this just means that the task currently running on the CPU is
 * not in a quiescent state.  There might be any number of tasks blocked
 * while in an RCU read-side critical section.
115 116 117 118
 *
 * Unlike the other rcu_*_qs() functions, callers to this function
 * must disable irqs in order to protect the assignment to
 * ->rcu_read_unlock_special.
119
 */
120
static void rcu_preempt_qs(int cpu)
121 122
{
	struct rcu_data *rdp = &per_cpu(rcu_preempt_data, cpu);
123

124
	rdp->passed_quiesc_completed = rdp->gpnum - 1;
125 126
	barrier();
	rdp->passed_quiesc = 1;
127
	current->rcu_read_unlock_special &= ~RCU_READ_UNLOCK_NEED_QS;
128 129 130
}

/*
131 132 133
 * We have entered the scheduler, and the current task might soon be
 * context-switched away from.  If this task is in an RCU read-side
 * critical section, we will no longer be able to rely on the CPU to
134 135 136 137 138 139
 * record that fact, so we enqueue the task on the blkd_tasks list.
 * The task will dequeue itself when it exits the outermost enclosing
 * RCU read-side critical section.  Therefore, the current grace period
 * cannot be permitted to complete until the blkd_tasks list entries
 * predating the current grace period drain, in other words, until
 * rnp->gp_tasks becomes NULL.
140 141
 *
 * Caller must disable preemption.
142
 */
143
static void rcu_preempt_note_context_switch(int cpu)
144 145
{
	struct task_struct *t = current;
146
	unsigned long flags;
147 148 149 150 151 152 153
	struct rcu_data *rdp;
	struct rcu_node *rnp;

	if (t->rcu_read_lock_nesting &&
	    (t->rcu_read_unlock_special & RCU_READ_UNLOCK_BLOCKED) == 0) {

		/* Possibly blocking in an RCU read-side critical section. */
154
		rdp = per_cpu_ptr(rcu_preempt_state.rda, cpu);
155
		rnp = rdp->mynode;
P
Paul E. McKenney 已提交
156
		raw_spin_lock_irqsave(&rnp->lock, flags);
157
		t->rcu_read_unlock_special |= RCU_READ_UNLOCK_BLOCKED;
158
		t->rcu_blocked_node = rnp;
159 160 161 162 163 164 165 166 167

		/*
		 * If this CPU has already checked in, then this task
		 * will hold up the next grace period rather than the
		 * current grace period.  Queue the task accordingly.
		 * If the task is queued for the current grace period
		 * (i.e., this CPU has not yet passed through a quiescent
		 * state for the current grace period), then as long
		 * as that task remains queued, the current grace period
168 169 170 171 172 173
		 * cannot end.  Note that there is some uncertainty as
		 * to exactly when the current grace period started.
		 * We take a conservative approach, which can result
		 * in unnecessarily waiting on tasks that started very
		 * slightly after the current grace period began.  C'est
		 * la vie!!!
174 175 176
		 *
		 * But first, note that the current CPU must still be
		 * on line!
177
		 */
178
		WARN_ON_ONCE((rdp->grpmask & rnp->qsmaskinit) == 0);
179
		WARN_ON_ONCE(!list_empty(&t->rcu_node_entry));
180 181 182
		if ((rnp->qsmask & rdp->grpmask) && rnp->gp_tasks != NULL) {
			list_add(&t->rcu_node_entry, rnp->gp_tasks->prev);
			rnp->gp_tasks = &t->rcu_node_entry;
183 184 185 186
#ifdef CONFIG_RCU_BOOST
			if (rnp->boost_tasks != NULL)
				rnp->boost_tasks = rnp->gp_tasks;
#endif /* #ifdef CONFIG_RCU_BOOST */
187 188 189 190 191
		} else {
			list_add(&t->rcu_node_entry, &rnp->blkd_tasks);
			if (rnp->qsmask & rdp->grpmask)
				rnp->gp_tasks = &t->rcu_node_entry;
		}
P
Paul E. McKenney 已提交
192
		raw_spin_unlock_irqrestore(&rnp->lock, flags);
193 194 195 196 197 198 199 200 201 202 203
	}

	/*
	 * Either we were not in an RCU read-side critical section to
	 * begin with, or we have now recorded that critical section
	 * globally.  Either way, we can now note a quiescent state
	 * for this CPU.  Again, if we were in an RCU read-side critical
	 * section, and if that critical section was blocking the current
	 * grace period, then the fact that the task has been enqueued
	 * means that we continue to block the current grace period.
	 */
204
	local_irq_save(flags);
205
	rcu_preempt_qs(cpu);
206
	local_irq_restore(flags);
207 208 209
}

/*
P
Paul E. McKenney 已提交
210
 * Tree-preemptible RCU implementation for rcu_read_lock().
211 212 213 214 215
 * Just increment ->rcu_read_lock_nesting, shared state will be updated
 * if we block.
 */
void __rcu_read_lock(void)
{
216
	current->rcu_read_lock_nesting++;
217 218 219 220
	barrier();  /* needed if we ever invoke rcu_read_lock in rcutree.c */
}
EXPORT_SYMBOL_GPL(__rcu_read_lock);

221 222 223 224 225
/*
 * Check for preempted RCU readers blocking the current grace period
 * for the specified rcu_node structure.  If the caller needs a reliable
 * answer, it must hold the rcu_node's ->lock.
 */
226
static int rcu_preempt_blocked_readers_cgp(struct rcu_node *rnp)
227
{
228
	return rnp->gp_tasks != NULL;
229 230
}

231 232 233 234 235 236 237
/*
 * Record a quiescent state for all tasks that were previously queued
 * on the specified rcu_node structure and that were blocking the current
 * RCU grace period.  The caller must hold the specified rnp->lock with
 * irqs disabled, and this lock is released upon return, but irqs remain
 * disabled.
 */
P
Paul E. McKenney 已提交
238
static void rcu_report_unblock_qs_rnp(struct rcu_node *rnp, unsigned long flags)
239 240 241 242 243
	__releases(rnp->lock)
{
	unsigned long mask;
	struct rcu_node *rnp_p;

244
	if (rnp->qsmask != 0 || rcu_preempt_blocked_readers_cgp(rnp)) {
P
Paul E. McKenney 已提交
245
		raw_spin_unlock_irqrestore(&rnp->lock, flags);
246 247 248 249 250 251 252 253 254 255
		return;  /* Still need more quiescent states! */
	}

	rnp_p = rnp->parent;
	if (rnp_p == NULL) {
		/*
		 * Either there is only one rcu_node in the tree,
		 * or tasks were kicked up to root rcu_node due to
		 * CPUs going offline.
		 */
P
Paul E. McKenney 已提交
256
		rcu_report_qs_rsp(&rcu_preempt_state, flags);
257 258 259 260 261
		return;
	}

	/* Report up the rest of the hierarchy. */
	mask = rnp->grpmask;
P
Paul E. McKenney 已提交
262 263
	raw_spin_unlock(&rnp->lock);	/* irqs remain disabled. */
	raw_spin_lock(&rnp_p->lock);	/* irqs already disabled. */
P
Paul E. McKenney 已提交
264
	rcu_report_qs_rnp(mask, &rcu_preempt_state, rnp_p, flags);
265 266
}

267 268 269 270 271 272 273 274 275 276 277 278 279 280 281
/*
 * Advance a ->blkd_tasks-list pointer to the next entry, instead
 * returning NULL if at the end of the list.
 */
static struct list_head *rcu_next_node_entry(struct task_struct *t,
					     struct rcu_node *rnp)
{
	struct list_head *np;

	np = t->rcu_node_entry.next;
	if (np == &rnp->blkd_tasks)
		np = NULL;
	return np;
}

282 283 284 285 286
/*
 * Handle special cases during rcu_read_unlock(), such as needing to
 * notify RCU core processing or task having blocked during the RCU
 * read-side critical section.
 */
287 288 289
static void rcu_read_unlock_special(struct task_struct *t)
{
	int empty;
290
	int empty_exp;
291
	unsigned long flags;
292
	struct list_head *np;
293 294 295 296 297 298 299 300 301 302 303 304 305 306 307
	struct rcu_node *rnp;
	int special;

	/* NMI handlers cannot block and cannot safely manipulate state. */
	if (in_nmi())
		return;

	local_irq_save(flags);

	/*
	 * If RCU core is waiting for this CPU to exit critical section,
	 * let it know that we have done so.
	 */
	special = t->rcu_read_unlock_special;
	if (special & RCU_READ_UNLOCK_NEED_QS) {
308
		rcu_preempt_qs(smp_processor_id());
309 310 311 312 313 314 315 316 317 318 319 320
	}

	/* Hardware IRQ handlers cannot block. */
	if (in_irq()) {
		local_irq_restore(flags);
		return;
	}

	/* Clean up if blocked during RCU read-side critical section. */
	if (special & RCU_READ_UNLOCK_BLOCKED) {
		t->rcu_read_unlock_special &= ~RCU_READ_UNLOCK_BLOCKED;

321 322 323 324 325 326
		/*
		 * Remove this task from the list it blocked on.  The
		 * task can migrate while we acquire the lock, but at
		 * most one time.  So at most two passes through loop.
		 */
		for (;;) {
327
			rnp = t->rcu_blocked_node;
P
Paul E. McKenney 已提交
328
			raw_spin_lock(&rnp->lock);  /* irqs already disabled. */
329
			if (rnp == t->rcu_blocked_node)
330
				break;
P
Paul E. McKenney 已提交
331
			raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */
332
		}
333
		empty = !rcu_preempt_blocked_readers_cgp(rnp);
334 335
		empty_exp = !rcu_preempted_readers_exp(rnp);
		smp_mb(); /* ensure expedited fastpath sees end of RCU c-s. */
336
		np = rcu_next_node_entry(t, rnp);
337
		list_del_init(&t->rcu_node_entry);
338 339 340 341
		if (&t->rcu_node_entry == rnp->gp_tasks)
			rnp->gp_tasks = np;
		if (&t->rcu_node_entry == rnp->exp_tasks)
			rnp->exp_tasks = np;
342 343 344 345
#ifdef CONFIG_RCU_BOOST
		if (&t->rcu_node_entry == rnp->boost_tasks)
			rnp->boost_tasks = np;
#endif /* #ifdef CONFIG_RCU_BOOST */
346
		t->rcu_blocked_node = NULL;
347 348 349 350

		/*
		 * If this was the last task on the current list, and if
		 * we aren't waiting on any CPUs, report the quiescent state.
P
Paul E. McKenney 已提交
351
		 * Note that rcu_report_unblock_qs_rnp() releases rnp->lock.
352
		 */
353
		if (empty)
P
Paul E. McKenney 已提交
354
			raw_spin_unlock_irqrestore(&rnp->lock, flags);
355
		else
P
Paul E. McKenney 已提交
356
			rcu_report_unblock_qs_rnp(rnp, flags);
357

358 359 360 361 362 363 364 365 366
#ifdef CONFIG_RCU_BOOST
		/* Unboost if we were boosted. */
		if (special & RCU_READ_UNLOCK_BOOSTED) {
			t->rcu_read_unlock_special &= ~RCU_READ_UNLOCK_BOOSTED;
			rt_mutex_unlock(t->rcu_boost_mutex);
			t->rcu_boost_mutex = NULL;
		}
#endif /* #ifdef CONFIG_RCU_BOOST */

367 368 369 370 371 372
		/*
		 * If this was the last task on the expedited lists,
		 * then we need to report up the rcu_node hierarchy.
		 */
		if (!empty_exp && !rcu_preempted_readers_exp(rnp))
			rcu_report_exp_rnp(&rcu_preempt_state, rnp);
373 374
	} else {
		local_irq_restore(flags);
375 376 377 378
	}
}

/*
P
Paul E. McKenney 已提交
379
 * Tree-preemptible RCU implementation for rcu_read_unlock().
380 381 382 383 384 385 386 387 388 389
 * Decrement ->rcu_read_lock_nesting.  If the result is zero (outermost
 * rcu_read_unlock()) and ->rcu_read_unlock_special is non-zero, then
 * invoke rcu_read_unlock_special() to clean up after a context switch
 * in an RCU read-side critical section and other special cases.
 */
void __rcu_read_unlock(void)
{
	struct task_struct *t = current;

	barrier();  /* needed if we ever invoke rcu_read_unlock in rcutree.c */
390 391 392
	--t->rcu_read_lock_nesting;
	barrier();  /* decrement before load of ->rcu_read_unlock_special */
	if (t->rcu_read_lock_nesting == 0 &&
393 394
	    unlikely(ACCESS_ONCE(t->rcu_read_unlock_special)))
		rcu_read_unlock_special(t);
395 396 397
#ifdef CONFIG_PROVE_LOCKING
	WARN_ON_ONCE(ACCESS_ONCE(t->rcu_read_lock_nesting) < 0);
#endif /* #ifdef CONFIG_PROVE_LOCKING */
398 399 400
}
EXPORT_SYMBOL_GPL(__rcu_read_unlock);

401 402 403 404 405 406 407 408 409 410 411
#ifdef CONFIG_RCU_CPU_STALL_VERBOSE

/*
 * Dump detailed information for all tasks blocking the current RCU
 * grace period on the specified rcu_node structure.
 */
static void rcu_print_detail_task_stall_rnp(struct rcu_node *rnp)
{
	unsigned long flags;
	struct task_struct *t;

412
	if (!rcu_preempt_blocked_readers_cgp(rnp))
413 414 415 416 417 418 419
		return;
	raw_spin_lock_irqsave(&rnp->lock, flags);
	t = list_entry(rnp->gp_tasks,
		       struct task_struct, rcu_node_entry);
	list_for_each_entry_continue(t, &rnp->blkd_tasks, rcu_node_entry)
		sched_show_task(t);
	raw_spin_unlock_irqrestore(&rnp->lock, flags);
420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442
}

/*
 * Dump detailed information for all tasks blocking the current RCU
 * grace period.
 */
static void rcu_print_detail_task_stall(struct rcu_state *rsp)
{
	struct rcu_node *rnp = rcu_get_root(rsp);

	rcu_print_detail_task_stall_rnp(rnp);
	rcu_for_each_leaf_node(rsp, rnp)
		rcu_print_detail_task_stall_rnp(rnp);
}

#else /* #ifdef CONFIG_RCU_CPU_STALL_VERBOSE */

static void rcu_print_detail_task_stall(struct rcu_state *rsp)
{
}

#endif /* #else #ifdef CONFIG_RCU_CPU_STALL_VERBOSE */

443 444 445 446 447 448 449 450
/*
 * Scan the current list of tasks blocked within RCU read-side critical
 * sections, printing out the tid of each.
 */
static void rcu_print_task_stall(struct rcu_node *rnp)
{
	struct task_struct *t;

451
	if (!rcu_preempt_blocked_readers_cgp(rnp))
452 453 454 455 456
		return;
	t = list_entry(rnp->gp_tasks,
		       struct task_struct, rcu_node_entry);
	list_for_each_entry_continue(t, &rnp->blkd_tasks, rcu_node_entry)
		printk(" P%d", t->pid);
457 458
}

459 460 461 462 463 464 465 466 467 468
/*
 * Suppress preemptible RCU's CPU stall warnings by pushing the
 * time of the next stall-warning message comfortably far into the
 * future.
 */
static void rcu_preempt_stall_reset(void)
{
	rcu_preempt_state.jiffies_stall = jiffies + ULONG_MAX / 2;
}

469 470 471 472 473 474
/*
 * Check that the list of blocked tasks for the newly completed grace
 * period is in fact empty.  It is a serious bug to complete a grace
 * period that still has RCU readers blocked!  This function must be
 * invoked -before- updating this rnp's ->gpnum, and the rnp's ->lock
 * must be held by the caller.
475 476 477
 *
 * Also, if there are blocked tasks on the list, they automatically
 * block the newly created grace period, so set up ->gp_tasks accordingly.
478 479 480
 */
static void rcu_preempt_check_blocked_tasks(struct rcu_node *rnp)
{
481
	WARN_ON_ONCE(rcu_preempt_blocked_readers_cgp(rnp));
482 483
	if (!list_empty(&rnp->blkd_tasks))
		rnp->gp_tasks = rnp->blkd_tasks.next;
484
	WARN_ON_ONCE(rnp->qsmask);
485 486
}

487 488
#ifdef CONFIG_HOTPLUG_CPU

489 490 491 492 493 494
/*
 * Handle tasklist migration for case in which all CPUs covered by the
 * specified rcu_node have gone offline.  Move them up to the root
 * rcu_node.  The reason for not just moving them to the immediate
 * parent is to remove the need for rcu_read_unlock_special() to
 * make more than two attempts to acquire the target rcu_node's lock.
495 496
 * Returns true if there were tasks blocking the current RCU grace
 * period.
497
 *
498 499 500
 * Returns 1 if there was previously a task blocking the current grace
 * period on the specified rcu_node structure.
 *
501 502
 * The caller must hold rnp->lock with irqs disabled.
 */
503 504 505
static int rcu_preempt_offline_tasks(struct rcu_state *rsp,
				     struct rcu_node *rnp,
				     struct rcu_data *rdp)
506 507 508
{
	struct list_head *lp;
	struct list_head *lp_root;
509
	int retval = 0;
510
	struct rcu_node *rnp_root = rcu_get_root(rsp);
511
	struct task_struct *t;
512

513 514
	if (rnp == rnp_root) {
		WARN_ONCE(1, "Last CPU thought to be offlined?");
515
		return 0;  /* Shouldn't happen: at least one CPU online. */
516
	}
517 518 519

	/* If we are on an internal node, complain bitterly. */
	WARN_ON_ONCE(rnp != rdp->mynode);
520 521

	/*
522 523 524 525 526 527 528
	 * Move tasks up to root rcu_node.  Don't try to get fancy for
	 * this corner-case operation -- just put this node's tasks
	 * at the head of the root node's list, and update the root node's
	 * ->gp_tasks and ->exp_tasks pointers to those of this node's,
	 * if non-NULL.  This might result in waiting for more tasks than
	 * absolutely necessary, but this is a good performance/complexity
	 * tradeoff.
529
	 */
530
	if (rcu_preempt_blocked_readers_cgp(rnp))
531 532 533
		retval |= RCU_OFL_TASKS_NORM_GP;
	if (rcu_preempted_readers_exp(rnp))
		retval |= RCU_OFL_TASKS_EXP_GP;
534 535 536 537 538 539 540 541 542 543 544 545
	lp = &rnp->blkd_tasks;
	lp_root = &rnp_root->blkd_tasks;
	while (!list_empty(lp)) {
		t = list_entry(lp->next, typeof(*t), rcu_node_entry);
		raw_spin_lock(&rnp_root->lock); /* irqs already disabled */
		list_del(&t->rcu_node_entry);
		t->rcu_blocked_node = rnp_root;
		list_add(&t->rcu_node_entry, lp_root);
		if (&t->rcu_node_entry == rnp->gp_tasks)
			rnp_root->gp_tasks = rnp->gp_tasks;
		if (&t->rcu_node_entry == rnp->exp_tasks)
			rnp_root->exp_tasks = rnp->exp_tasks;
546 547 548 549
#ifdef CONFIG_RCU_BOOST
		if (&t->rcu_node_entry == rnp->boost_tasks)
			rnp_root->boost_tasks = rnp->boost_tasks;
#endif /* #ifdef CONFIG_RCU_BOOST */
550
		raw_spin_unlock(&rnp_root->lock); /* irqs still disabled */
551
	}
552 553 554 555 556 557 558 559 560 561

#ifdef CONFIG_RCU_BOOST
	/* In case root is being boosted and leaf is not. */
	raw_spin_lock(&rnp_root->lock); /* irqs already disabled */
	if (rnp_root->boost_tasks != NULL &&
	    rnp_root->boost_tasks != rnp_root->gp_tasks)
		rnp_root->boost_tasks = rnp_root->gp_tasks;
	raw_spin_unlock(&rnp_root->lock); /* irqs still disabled */
#endif /* #ifdef CONFIG_RCU_BOOST */

562 563
	rnp->gp_tasks = NULL;
	rnp->exp_tasks = NULL;
564
	return retval;
565 566
}

567
/*
P
Paul E. McKenney 已提交
568
 * Do CPU-offline processing for preemptible RCU.
569 570 571 572 573 574 575 576
 */
static void rcu_preempt_offline_cpu(int cpu)
{
	__rcu_offline_cpu(cpu, &rcu_preempt_state);
}

#endif /* #ifdef CONFIG_HOTPLUG_CPU */

577 578 579 580 581 582 583 584 585 586 587 588
/*
 * Check for a quiescent state from the current CPU.  When a task blocks,
 * the task is recorded in the corresponding CPU's rcu_node structure,
 * which is checked elsewhere.
 *
 * Caller must disable hard irqs.
 */
static void rcu_preempt_check_callbacks(int cpu)
{
	struct task_struct *t = current;

	if (t->rcu_read_lock_nesting == 0) {
589
		rcu_preempt_qs(cpu);
590 591
		return;
	}
592
	if (per_cpu(rcu_preempt_data, cpu).qs_pending)
593
		t->rcu_read_unlock_special |= RCU_READ_UNLOCK_NEED_QS;
594 595 596
}

/*
P
Paul E. McKenney 已提交
597
 * Process callbacks for preemptible RCU.
598 599 600 601 602 603 604 605
 */
static void rcu_preempt_process_callbacks(void)
{
	__rcu_process_callbacks(&rcu_preempt_state,
				&__get_cpu_var(rcu_preempt_data));
}

/*
P
Paul E. McKenney 已提交
606
 * Queue a preemptible-RCU callback for invocation after a grace period.
607 608 609 610 611 612 613
 */
void call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu))
{
	__call_rcu(head, func, &rcu_preempt_state);
}
EXPORT_SYMBOL_GPL(call_rcu);

614 615 616 617 618
/**
 * synchronize_rcu - wait until a grace period has elapsed.
 *
 * Control will return to the caller some time after a full grace
 * period has elapsed, in other words after all currently executing RCU
619 620 621 622 623
 * read-side critical sections have completed.  Note, however, that
 * upon return from synchronize_rcu(), the caller might well be executing
 * concurrently with new RCU read-side critical sections that began while
 * synchronize_rcu() was waiting.  RCU read-side critical sections are
 * delimited by rcu_read_lock() and rcu_read_unlock(), and may be nested.
624 625 626 627 628 629 630 631
 */
void synchronize_rcu(void)
{
	struct rcu_synchronize rcu;

	if (!rcu_scheduler_active)
		return;

632
	init_rcu_head_on_stack(&rcu.head);
633 634 635 636 637
	init_completion(&rcu.completion);
	/* Will wake me after RCU finished. */
	call_rcu(&rcu.head, wakeme_after_rcu);
	/* Wait for it. */
	wait_for_completion(&rcu.completion);
638
	destroy_rcu_head_on_stack(&rcu.head);
639 640 641
}
EXPORT_SYMBOL_GPL(synchronize_rcu);

642 643 644 645 646 647 648 649 650 651 652 653
static DECLARE_WAIT_QUEUE_HEAD(sync_rcu_preempt_exp_wq);
static long sync_rcu_preempt_exp_count;
static DEFINE_MUTEX(sync_rcu_preempt_exp_mutex);

/*
 * Return non-zero if there are any tasks in RCU read-side critical
 * sections blocking the current preemptible-RCU expedited grace period.
 * If there is no preemptible-RCU expedited grace period currently in
 * progress, returns zero unconditionally.
 */
static int rcu_preempted_readers_exp(struct rcu_node *rnp)
{
654
	return rnp->exp_tasks != NULL;
655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686
}

/*
 * return non-zero if there is no RCU expedited grace period in progress
 * for the specified rcu_node structure, in other words, if all CPUs and
 * tasks covered by the specified rcu_node structure have done their bit
 * for the current expedited grace period.  Works only for preemptible
 * RCU -- other RCU implementation use other means.
 *
 * Caller must hold sync_rcu_preempt_exp_mutex.
 */
static int sync_rcu_preempt_exp_done(struct rcu_node *rnp)
{
	return !rcu_preempted_readers_exp(rnp) &&
	       ACCESS_ONCE(rnp->expmask) == 0;
}

/*
 * Report the exit from RCU read-side critical section for the last task
 * that queued itself during or before the current expedited preemptible-RCU
 * grace period.  This event is reported either to the rcu_node structure on
 * which the task was queued or to one of that rcu_node structure's ancestors,
 * recursively up the tree.  (Calm down, calm down, we do the recursion
 * iteratively!)
 *
 * Caller must hold sync_rcu_preempt_exp_mutex.
 */
static void rcu_report_exp_rnp(struct rcu_state *rsp, struct rcu_node *rnp)
{
	unsigned long flags;
	unsigned long mask;

P
Paul E. McKenney 已提交
687
	raw_spin_lock_irqsave(&rnp->lock, flags);
688 689 690 691 692 693 694 695
	for (;;) {
		if (!sync_rcu_preempt_exp_done(rnp))
			break;
		if (rnp->parent == NULL) {
			wake_up(&sync_rcu_preempt_exp_wq);
			break;
		}
		mask = rnp->grpmask;
P
Paul E. McKenney 已提交
696
		raw_spin_unlock(&rnp->lock); /* irqs remain disabled */
697
		rnp = rnp->parent;
P
Paul E. McKenney 已提交
698
		raw_spin_lock(&rnp->lock); /* irqs already disabled */
699 700
		rnp->expmask &= ~mask;
	}
P
Paul E. McKenney 已提交
701
	raw_spin_unlock_irqrestore(&rnp->lock, flags);
702 703 704 705 706 707 708 709 710 711 712 713
}

/*
 * Snapshot the tasks blocking the newly started preemptible-RCU expedited
 * grace period for the specified rcu_node structure.  If there are no such
 * tasks, report it up the rcu_node hierarchy.
 *
 * Caller must hold sync_rcu_preempt_exp_mutex and rsp->onofflock.
 */
static void
sync_rcu_preempt_exp_init(struct rcu_state *rsp, struct rcu_node *rnp)
{
714
	int must_wait = 0;
715

P
Paul E. McKenney 已提交
716
	raw_spin_lock(&rnp->lock); /* irqs already disabled */
717 718
	if (!list_empty(&rnp->blkd_tasks)) {
		rnp->exp_tasks = rnp->blkd_tasks.next;
719
		rcu_initiate_boost(rnp);
720 721
		must_wait = 1;
	}
P
Paul E. McKenney 已提交
722
	raw_spin_unlock(&rnp->lock); /* irqs remain disabled */
723 724 725 726
	if (!must_wait)
		rcu_report_exp_rnp(rsp, rnp);
}

727
/*
728 729
 * Wait for an rcu-preempt grace period, but expedite it.  The basic idea
 * is to invoke synchronize_sched_expedited() to push all the tasks to
730
 * the ->blkd_tasks lists and wait for this list to drain.
731 732 733
 */
void synchronize_rcu_expedited(void)
{
734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761
	unsigned long flags;
	struct rcu_node *rnp;
	struct rcu_state *rsp = &rcu_preempt_state;
	long snap;
	int trycount = 0;

	smp_mb(); /* Caller's modifications seen first by other CPUs. */
	snap = ACCESS_ONCE(sync_rcu_preempt_exp_count) + 1;
	smp_mb(); /* Above access cannot bleed into critical section. */

	/*
	 * Acquire lock, falling back to synchronize_rcu() if too many
	 * lock-acquisition failures.  Of course, if someone does the
	 * expedited grace period for us, just leave.
	 */
	while (!mutex_trylock(&sync_rcu_preempt_exp_mutex)) {
		if (trycount++ < 10)
			udelay(trycount * num_online_cpus());
		else {
			synchronize_rcu();
			return;
		}
		if ((ACCESS_ONCE(sync_rcu_preempt_exp_count) - snap) > 0)
			goto mb_ret; /* Others did our work for us. */
	}
	if ((ACCESS_ONCE(sync_rcu_preempt_exp_count) - snap) > 0)
		goto unlock_mb_ret; /* Others did our work for us. */

762
	/* force all RCU readers onto ->blkd_tasks lists. */
763 764
	synchronize_sched_expedited();

P
Paul E. McKenney 已提交
765
	raw_spin_lock_irqsave(&rsp->onofflock, flags);
766 767 768

	/* Initialize ->expmask for all non-leaf rcu_node structures. */
	rcu_for_each_nonleaf_node_breadth_first(rsp, rnp) {
P
Paul E. McKenney 已提交
769
		raw_spin_lock(&rnp->lock); /* irqs already disabled. */
770
		rnp->expmask = rnp->qsmaskinit;
P
Paul E. McKenney 已提交
771
		raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */
772 773
	}

774
	/* Snapshot current state of ->blkd_tasks lists. */
775 776 777 778 779
	rcu_for_each_leaf_node(rsp, rnp)
		sync_rcu_preempt_exp_init(rsp, rnp);
	if (NUM_RCU_NODES > 1)
		sync_rcu_preempt_exp_init(rsp, rcu_get_root(rsp));

P
Paul E. McKenney 已提交
780
	raw_spin_unlock_irqrestore(&rsp->onofflock, flags);
781

782
	/* Wait for snapshotted ->blkd_tasks lists to drain. */
783 784 785 786 787 788 789 790 791 792 793
	rnp = rcu_get_root(rsp);
	wait_event(sync_rcu_preempt_exp_wq,
		   sync_rcu_preempt_exp_done(rnp));

	/* Clean up and exit. */
	smp_mb(); /* ensure expedited GP seen before counter increment. */
	ACCESS_ONCE(sync_rcu_preempt_exp_count)++;
unlock_mb_ret:
	mutex_unlock(&sync_rcu_preempt_exp_mutex);
mb_ret:
	smp_mb(); /* ensure subsequent action seen after grace period. */
794 795 796
}
EXPORT_SYMBOL_GPL(synchronize_rcu_expedited);

797
/*
P
Paul E. McKenney 已提交
798
 * Check to see if there is any immediate preemptible-RCU-related work
799 800 801 802 803 804 805 806 807
 * to be done.
 */
static int rcu_preempt_pending(int cpu)
{
	return __rcu_pending(&rcu_preempt_state,
			     &per_cpu(rcu_preempt_data, cpu));
}

/*
P
Paul E. McKenney 已提交
808
 * Does preemptible RCU need the CPU to stay out of dynticks mode?
809 810 811 812 813 814
 */
static int rcu_preempt_needs_cpu(int cpu)
{
	return !!per_cpu(rcu_preempt_data, cpu).nxtlist;
}

815 816 817 818 819 820 821 822 823
/**
 * rcu_barrier - Wait until all in-flight call_rcu() callbacks complete.
 */
void rcu_barrier(void)
{
	_rcu_barrier(&rcu_preempt_state, call_rcu);
}
EXPORT_SYMBOL_GPL(rcu_barrier);

824
/*
P
Paul E. McKenney 已提交
825
 * Initialize preemptible RCU's per-CPU data.
826 827 828 829 830 831
 */
static void __cpuinit rcu_preempt_init_percpu_data(int cpu)
{
	rcu_init_percpu_data(cpu, &rcu_preempt_state, 1);
}

832
/*
P
Paul E. McKenney 已提交
833
 * Move preemptible RCU's callbacks from dying CPU to other online CPU.
834
 */
835
static void rcu_preempt_send_cbs_to_online(void)
836
{
837
	rcu_send_cbs_to_online(&rcu_preempt_state);
838 839
}

840
/*
P
Paul E. McKenney 已提交
841
 * Initialize preemptible RCU's state structures.
842 843 844
 */
static void __init __rcu_init_preempt(void)
{
845
	rcu_init_one(&rcu_preempt_state, &rcu_preempt_data);
846 847
}

848
/*
P
Paul E. McKenney 已提交
849
 * Check for a task exiting while in a preemptible-RCU read-side
850 851 852 853 854 855 856 857 858 859 860
 * critical section, clean up if so.  No need to issue warnings,
 * as debug_check_no_locks_held() already does this if lockdep
 * is enabled.
 */
void exit_rcu(void)
{
	struct task_struct *t = current;

	if (t->rcu_read_lock_nesting == 0)
		return;
	t->rcu_read_lock_nesting = 1;
861
	__rcu_read_unlock();
862 863 864 865
}

#else /* #ifdef CONFIG_TREE_PREEMPT_RCU */

866 867
static struct rcu_state *rcu_state = &rcu_sched_state;

868 869 870
/*
 * Tell them what RCU they are running.
 */
871
static void __init rcu_bootup_announce(void)
872 873
{
	printk(KERN_INFO "Hierarchical RCU implementation.\n");
874
	rcu_bootup_announce_oddness();
875 876 877 878 879 880 881 882 883 884 885
}

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

886 887 888 889 890 891 892 893 894 895
/*
 * Force a quiescent state for RCU, which, because there is no preemptible
 * RCU, becomes the same as rcu-sched.
 */
void rcu_force_quiescent_state(void)
{
	rcu_sched_force_quiescent_state();
}
EXPORT_SYMBOL_GPL(rcu_force_quiescent_state);

896
/*
P
Paul E. McKenney 已提交
897
 * Because preemptible RCU does not exist, we never have to check for
898 899
 * CPUs being in quiescent states.
 */
900
static void rcu_preempt_note_context_switch(int cpu)
901 902 903
{
}

904
/*
P
Paul E. McKenney 已提交
905
 * Because preemptible RCU does not exist, there are never any preempted
906 907
 * RCU readers.
 */
908
static int rcu_preempt_blocked_readers_cgp(struct rcu_node *rnp)
909 910 911 912
{
	return 0;
}

913 914 915
#ifdef CONFIG_HOTPLUG_CPU

/* Because preemptible RCU does not exist, no quieting of tasks. */
P
Paul E. McKenney 已提交
916
static void rcu_report_unblock_qs_rnp(struct rcu_node *rnp, unsigned long flags)
917
{
P
Paul E. McKenney 已提交
918
	raw_spin_unlock_irqrestore(&rnp->lock, flags);
919 920 921 922
}

#endif /* #ifdef CONFIG_HOTPLUG_CPU */

923
/*
P
Paul E. McKenney 已提交
924
 * Because preemptible RCU does not exist, we never have to check for
925 926 927 928 929 930
 * tasks blocked within RCU read-side critical sections.
 */
static void rcu_print_detail_task_stall(struct rcu_state *rsp)
{
}

931
/*
P
Paul E. McKenney 已提交
932
 * Because preemptible RCU does not exist, we never have to check for
933 934 935 936 937 938
 * tasks blocked within RCU read-side critical sections.
 */
static void rcu_print_task_stall(struct rcu_node *rnp)
{
}

939 940 941 942 943 944 945 946
/*
 * Because preemptible RCU does not exist, there is no need to suppress
 * its CPU stall warnings.
 */
static void rcu_preempt_stall_reset(void)
{
}

947
/*
P
Paul E. McKenney 已提交
948
 * Because there is no preemptible RCU, there can be no readers blocked,
949 950
 * so there is no need to check for blocked tasks.  So check only for
 * bogus qsmask values.
951 952 953
 */
static void rcu_preempt_check_blocked_tasks(struct rcu_node *rnp)
{
954
	WARN_ON_ONCE(rnp->qsmask);
955 956
}

957 958
#ifdef CONFIG_HOTPLUG_CPU

959
/*
P
Paul E. McKenney 已提交
960
 * Because preemptible RCU does not exist, it never needs to migrate
961 962 963
 * tasks that were blocked within RCU read-side critical sections, and
 * such non-existent tasks cannot possibly have been blocking the current
 * grace period.
964
 */
965 966 967
static int rcu_preempt_offline_tasks(struct rcu_state *rsp,
				     struct rcu_node *rnp,
				     struct rcu_data *rdp)
968
{
969
	return 0;
970 971
}

972
/*
P
Paul E. McKenney 已提交
973
 * Because preemptible RCU does not exist, it never needs CPU-offline
974 975 976 977 978 979 980 981
 * processing.
 */
static void rcu_preempt_offline_cpu(int cpu)
{
}

#endif /* #ifdef CONFIG_HOTPLUG_CPU */

982
/*
P
Paul E. McKenney 已提交
983
 * Because preemptible RCU does not exist, it never has any callbacks
984 985
 * to check.
 */
986
static void rcu_preempt_check_callbacks(int cpu)
987 988 989 990
{
}

/*
P
Paul E. McKenney 已提交
991
 * Because preemptible RCU does not exist, it never has any callbacks
992 993
 * to process.
 */
994
static void rcu_preempt_process_callbacks(void)
995 996 997
{
}

998 999
/*
 * Wait for an rcu-preempt grace period, but make it happen quickly.
P
Paul E. McKenney 已提交
1000
 * But because preemptible RCU does not exist, map to rcu-sched.
1001 1002 1003 1004 1005 1006 1007
 */
void synchronize_rcu_expedited(void)
{
	synchronize_sched_expedited();
}
EXPORT_SYMBOL_GPL(synchronize_rcu_expedited);

1008 1009 1010
#ifdef CONFIG_HOTPLUG_CPU

/*
P
Paul E. McKenney 已提交
1011
 * Because preemptible RCU does not exist, there is never any need to
1012 1013 1014 1015 1016 1017 1018 1019 1020 1021
 * report on tasks preempted in RCU read-side critical sections during
 * expedited RCU grace periods.
 */
static void rcu_report_exp_rnp(struct rcu_state *rsp, struct rcu_node *rnp)
{
	return;
}

#endif /* #ifdef CONFIG_HOTPLUG_CPU */

1022
/*
P
Paul E. McKenney 已提交
1023
 * Because preemptible RCU does not exist, it never has any work to do.
1024 1025 1026 1027 1028 1029 1030
 */
static int rcu_preempt_pending(int cpu)
{
	return 0;
}

/*
P
Paul E. McKenney 已提交
1031
 * Because preemptible RCU does not exist, it never needs any CPU.
1032 1033 1034 1035 1036 1037
 */
static int rcu_preempt_needs_cpu(int cpu)
{
	return 0;
}

1038
/*
P
Paul E. McKenney 已提交
1039
 * Because preemptible RCU does not exist, rcu_barrier() is just
1040 1041 1042 1043 1044 1045 1046 1047
 * another name for rcu_barrier_sched().
 */
void rcu_barrier(void)
{
	rcu_barrier_sched();
}
EXPORT_SYMBOL_GPL(rcu_barrier);

1048
/*
P
Paul E. McKenney 已提交
1049
 * Because preemptible RCU does not exist, there is no per-CPU
1050 1051 1052 1053 1054 1055
 * data to initialize.
 */
static void __cpuinit rcu_preempt_init_percpu_data(int cpu)
{
}

1056
/*
P
Paul E. McKenney 已提交
1057
 * Because there is no preemptible RCU, there are no callbacks to move.
1058
 */
1059
static void rcu_preempt_send_cbs_to_online(void)
1060 1061 1062
{
}

1063
/*
P
Paul E. McKenney 已提交
1064
 * Because preemptible RCU does not exist, it need not be initialized.
1065 1066 1067 1068 1069
 */
static void __init __rcu_init_preempt(void)
{
}

1070
#endif /* #else #ifdef CONFIG_TREE_PREEMPT_RCU */
1071

1072 1073 1074 1075
#ifdef CONFIG_RCU_BOOST

#include "rtmutex_common.h"

1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088
#ifdef CONFIG_RCU_TRACE

static void rcu_initiate_boost_trace(struct rcu_node *rnp)
{
	if (list_empty(&rnp->blkd_tasks))
		rnp->n_balk_blkd_tasks++;
	else if (rnp->exp_tasks == NULL && rnp->gp_tasks == NULL)
		rnp->n_balk_exp_gp_tasks++;
	else if (rnp->gp_tasks != NULL && rnp->boost_tasks != NULL)
		rnp->n_balk_boost_tasks++;
	else if (rnp->gp_tasks != NULL && rnp->qsmask != 0)
		rnp->n_balk_notblocked++;
	else if (rnp->gp_tasks != NULL &&
1089
		 ULONG_CMP_LT(jiffies, rnp->boost_time))
1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102
		rnp->n_balk_notyet++;
	else
		rnp->n_balk_nos++;
}

#else /* #ifdef CONFIG_RCU_TRACE */

static void rcu_initiate_boost_trace(struct rcu_node *rnp)
{
}

#endif /* #else #ifdef CONFIG_RCU_TRACE */

1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137
/*
 * Carry out RCU priority boosting on the task indicated by ->exp_tasks
 * or ->boost_tasks, advancing the pointer to the next task in the
 * ->blkd_tasks list.
 *
 * Note that irqs must be enabled: boosting the task can block.
 * Returns 1 if there are more tasks needing to be boosted.
 */
static int rcu_boost(struct rcu_node *rnp)
{
	unsigned long flags;
	struct rt_mutex mtx;
	struct task_struct *t;
	struct list_head *tb;

	if (rnp->exp_tasks == NULL && rnp->boost_tasks == NULL)
		return 0;  /* Nothing left to boost. */

	raw_spin_lock_irqsave(&rnp->lock, flags);

	/*
	 * Recheck under the lock: all tasks in need of boosting
	 * might exit their RCU read-side critical sections on their own.
	 */
	if (rnp->exp_tasks == NULL && rnp->boost_tasks == NULL) {
		raw_spin_unlock_irqrestore(&rnp->lock, flags);
		return 0;
	}

	/*
	 * Preferentially boost tasks blocking expedited grace periods.
	 * This cannot starve the normal grace periods because a second
	 * expedited grace period must boost all blocked tasks, including
	 * those blocking the pre-existing normal grace period.
	 */
1138
	if (rnp->exp_tasks != NULL) {
1139
		tb = rnp->exp_tasks;
1140 1141
		rnp->n_exp_boosts++;
	} else {
1142
		tb = rnp->boost_tasks;
1143 1144 1145
		rnp->n_normal_boosts++;
	}
	rnp->n_tasks_boosted++;
1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200

	/*
	 * We boost task t by manufacturing an rt_mutex that appears to
	 * be held by task t.  We leave a pointer to that rt_mutex where
	 * task t can find it, and task t will release the mutex when it
	 * exits its outermost RCU read-side critical section.  Then
	 * simply acquiring this artificial rt_mutex will boost task
	 * t's priority.  (Thanks to tglx for suggesting this approach!)
	 *
	 * Note that task t must acquire rnp->lock to remove itself from
	 * the ->blkd_tasks list, which it will do from exit() if from
	 * nowhere else.  We therefore are guaranteed that task t will
	 * stay around at least until we drop rnp->lock.  Note that
	 * rnp->lock also resolves races between our priority boosting
	 * and task t's exiting its outermost RCU read-side critical
	 * section.
	 */
	t = container_of(tb, struct task_struct, rcu_node_entry);
	rt_mutex_init_proxy_locked(&mtx, t);
	t->rcu_boost_mutex = &mtx;
	t->rcu_read_unlock_special |= RCU_READ_UNLOCK_BOOSTED;
	raw_spin_unlock_irqrestore(&rnp->lock, flags);
	rt_mutex_lock(&mtx);  /* Side effect: boosts task t's priority. */
	rt_mutex_unlock(&mtx);  /* Keep lockdep happy. */

	return rnp->exp_tasks != NULL || rnp->boost_tasks != NULL;
}

/*
 * Timer handler to initiate waking up of boost kthreads that
 * have yielded the CPU due to excessive numbers of tasks to
 * boost.  We wake up the per-rcu_node kthread, which in turn
 * will wake up the booster kthread.
 */
static void rcu_boost_kthread_timer(unsigned long arg)
{
	unsigned long flags;
	struct rcu_node *rnp = (struct rcu_node *)arg;

	raw_spin_lock_irqsave(&rnp->lock, flags);
	invoke_rcu_node_kthread(rnp);
	raw_spin_unlock_irqrestore(&rnp->lock, flags);
}

/*
 * Priority-boosting kthread.  One per leaf rcu_node and one for the
 * root rcu_node.
 */
static int rcu_boost_kthread(void *arg)
{
	struct rcu_node *rnp = (struct rcu_node *)arg;
	int spincnt = 0;
	int more2boost;

	for (;;) {
1201
		rnp->boost_kthread_status = RCU_KTHREAD_WAITING;
1202 1203 1204 1205 1206
		wait_event_interruptible(rnp->boost_wq, rnp->boost_tasks ||
							rnp->exp_tasks ||
							kthread_should_stop());
		if (kthread_should_stop())
			break;
1207
		rnp->boost_kthread_status = RCU_KTHREAD_RUNNING;
1208 1209 1210 1211 1212 1213 1214 1215 1216 1217
		more2boost = rcu_boost(rnp);
		if (more2boost)
			spincnt++;
		else
			spincnt = 0;
		if (spincnt > 10) {
			rcu_yield(rcu_boost_kthread_timer, (unsigned long)rnp);
			spincnt = 0;
		}
	}
1218
	rnp->boost_kthread_status = RCU_KTHREAD_STOPPED;
1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233
	return 0;
}

/*
 * Check to see if it is time to start boosting RCU readers that are
 * blocking the current grace period, and, if so, tell the per-rcu_node
 * kthread to start boosting them.  If there is an expedited grace
 * period in progress, it is always time to boost.
 *
 * The caller must hold rnp->lock.
 */
static void rcu_initiate_boost(struct rcu_node *rnp)
{
	struct task_struct *t;

1234 1235
	if (!rcu_preempt_blocked_readers_cgp(rnp) && rnp->exp_tasks == NULL) {
		rnp->n_balk_exp_gp_tasks++;
1236
		return;
1237
	}
1238 1239 1240 1241 1242 1243 1244 1245 1246 1247
	if (rnp->exp_tasks != NULL ||
	    (rnp->gp_tasks != NULL &&
	     rnp->boost_tasks == NULL &&
	     rnp->qsmask == 0 &&
	     ULONG_CMP_GE(jiffies, rnp->boost_time))) {
		if (rnp->exp_tasks == NULL)
			rnp->boost_tasks = rnp->gp_tasks;
		t = rnp->boost_kthread_task;
		if (t != NULL)
			wake_up_process(t);
1248 1249
	} else
		rcu_initiate_boost_trace(rnp);
1250 1251
}

1252 1253 1254 1255 1256
/*
 * Set the affinity of the boost kthread.  The CPU-hotplug locks are
 * held, so no one should be messing with the existence of the boost
 * kthread.
 */
1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 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 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367
static void rcu_boost_kthread_setaffinity(struct rcu_node *rnp,
					  cpumask_var_t cm)
{
	struct task_struct *t;

	t = rnp->boost_kthread_task;
	if (t != NULL)
		set_cpus_allowed_ptr(rnp->boost_kthread_task, cm);
}

#define RCU_BOOST_DELAY_JIFFIES DIV_ROUND_UP(CONFIG_RCU_BOOST_DELAY * HZ, 1000)

/*
 * Do priority-boost accounting for the start of a new grace period.
 */
static void rcu_preempt_boost_start_gp(struct rcu_node *rnp)
{
	rnp->boost_time = jiffies + RCU_BOOST_DELAY_JIFFIES;
}

/*
 * Initialize the RCU-boost waitqueue.
 */
static void __init rcu_init_boost_waitqueue(struct rcu_node *rnp)
{
	init_waitqueue_head(&rnp->boost_wq);
}

/*
 * Create an RCU-boost kthread for the specified node if one does not
 * already exist.  We only create this kthread for preemptible RCU.
 * Returns zero if all is well, a negated errno otherwise.
 */
static int __cpuinit rcu_spawn_one_boost_kthread(struct rcu_state *rsp,
						 struct rcu_node *rnp,
						 int rnp_index)
{
	unsigned long flags;
	struct sched_param sp;
	struct task_struct *t;

	if (&rcu_preempt_state != rsp)
		return 0;
	if (rnp->boost_kthread_task != NULL)
		return 0;
	t = kthread_create(rcu_boost_kthread, (void *)rnp,
			   "rcub%d", rnp_index);
	if (IS_ERR(t))
		return PTR_ERR(t);
	raw_spin_lock_irqsave(&rnp->lock, flags);
	rnp->boost_kthread_task = t;
	raw_spin_unlock_irqrestore(&rnp->lock, flags);
	wake_up_process(t);
	sp.sched_priority = RCU_KTHREAD_PRIO;
	sched_setscheduler_nocheck(t, SCHED_FIFO, &sp);
	return 0;
}

#ifdef CONFIG_HOTPLUG_CPU

static void rcu_stop_boost_kthread(struct rcu_node *rnp)
{
	unsigned long flags;
	struct task_struct *t;

	raw_spin_lock_irqsave(&rnp->lock, flags);
	t = rnp->boost_kthread_task;
	rnp->boost_kthread_task = NULL;
	raw_spin_unlock_irqrestore(&rnp->lock, flags);
	if (t != NULL)
		kthread_stop(t);
}

#endif /* #ifdef CONFIG_HOTPLUG_CPU */

#else /* #ifdef CONFIG_RCU_BOOST */

static void rcu_initiate_boost(struct rcu_node *rnp)
{
}

static void rcu_boost_kthread_setaffinity(struct rcu_node *rnp,
					  cpumask_var_t cm)
{
}

static void rcu_preempt_boost_start_gp(struct rcu_node *rnp)
{
}

static void __init rcu_init_boost_waitqueue(struct rcu_node *rnp)
{
}

static int __cpuinit rcu_spawn_one_boost_kthread(struct rcu_state *rsp,
						 struct rcu_node *rnp,
						 int rnp_index)
{
	return 0;
}

#ifdef CONFIG_HOTPLUG_CPU

static void rcu_stop_boost_kthread(struct rcu_node *rnp)
{
}

#endif /* #ifdef CONFIG_HOTPLUG_CPU */

#endif /* #else #ifdef CONFIG_RCU_BOOST */

1368 1369 1370 1371 1372 1373 1374 1375 1376 1377
#ifndef CONFIG_SMP

void synchronize_sched_expedited(void)
{
	cond_resched();
}
EXPORT_SYMBOL_GPL(synchronize_sched_expedited);

#else /* #ifndef CONFIG_SMP */

1378 1379
static atomic_t sync_sched_expedited_started = ATOMIC_INIT(0);
static atomic_t sync_sched_expedited_done = ATOMIC_INIT(0);
1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406

static int synchronize_sched_expedited_cpu_stop(void *data)
{
	/*
	 * There must be a full memory barrier on each affected CPU
	 * between the time that try_stop_cpus() is called and the
	 * time that it returns.
	 *
	 * In the current initial implementation of cpu_stop, the
	 * above condition is already met when the control reaches
	 * this point and the following smp_mb() is not strictly
	 * necessary.  Do smp_mb() anyway for documentation and
	 * robustness against future implementation changes.
	 */
	smp_mb(); /* See above comment block. */
	return 0;
}

/*
 * Wait for an rcu-sched grace period to elapse, but use "big hammer"
 * approach to force grace period to end quickly.  This consumes
 * significant time on all CPUs, and is thus not recommended for
 * any sort of common-case code.
 *
 * Note that it is illegal to call this function while holding any
 * lock that is acquired by a CPU-hotplug notifier.  Failing to
 * observe this restriction will result in deadlock.
1407
 *
1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427
 * This implementation can be thought of as an application of ticket
 * locking to RCU, with sync_sched_expedited_started and
 * sync_sched_expedited_done taking on the roles of the halves
 * of the ticket-lock word.  Each task atomically increments
 * sync_sched_expedited_started upon entry, snapshotting the old value,
 * then attempts to stop all the CPUs.  If this succeeds, then each
 * CPU will have executed a context switch, resulting in an RCU-sched
 * grace period.  We are then done, so we use atomic_cmpxchg() to
 * update sync_sched_expedited_done to match our snapshot -- but
 * only if someone else has not already advanced past our snapshot.
 *
 * On the other hand, if try_stop_cpus() fails, we check the value
 * of sync_sched_expedited_done.  If it has advanced past our
 * initial snapshot, then someone else must have forced a grace period
 * some time after we took our snapshot.  In this case, our work is
 * done for us, and we can simply return.  Otherwise, we try again,
 * but keep our initial snapshot for purposes of checking for someone
 * doing our work for us.
 *
 * If we fail too many times in a row, we fall back to synchronize_sched().
1428 1429 1430
 */
void synchronize_sched_expedited(void)
{
1431
	int firstsnap, s, snap, trycount = 0;
1432

1433 1434
	/* Note that atomic_inc_return() implies full memory barrier. */
	firstsnap = snap = atomic_inc_return(&sync_sched_expedited_started);
1435
	get_online_cpus();
1436 1437 1438 1439 1440

	/*
	 * Each pass through the following loop attempts to force a
	 * context switch on each CPU.
	 */
1441 1442 1443 1444
	while (try_stop_cpus(cpu_online_mask,
			     synchronize_sched_expedited_cpu_stop,
			     NULL) == -EAGAIN) {
		put_online_cpus();
1445 1446

		/* No joy, try again later.  Or just synchronize_sched(). */
1447 1448 1449 1450 1451 1452
		if (trycount++ < 10)
			udelay(trycount * num_online_cpus());
		else {
			synchronize_sched();
			return;
		}
1453 1454 1455 1456

		/* Check to see if someone else did our work for us. */
		s = atomic_read(&sync_sched_expedited_done);
		if (UINT_CMP_GE((unsigned)s, (unsigned)firstsnap)) {
1457 1458 1459
			smp_mb(); /* ensure test happens before caller kfree */
			return;
		}
1460 1461 1462 1463 1464 1465 1466 1467 1468

		/*
		 * Refetching sync_sched_expedited_started allows later
		 * callers to piggyback on our grace period.  We subtract
		 * 1 to get the same token that the last incrementer got.
		 * We retry after they started, so our grace period works
		 * for them, and they started after our first try, so their
		 * grace period works for us.
		 */
1469
		get_online_cpus();
1470 1471
		snap = atomic_read(&sync_sched_expedited_started) - 1;
		smp_mb(); /* ensure read is before try_stop_cpus(). */
1472
	}
1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487

	/*
	 * Everyone up to our most recent fetch is covered by our grace
	 * period.  Update the counter, but only if our work is still
	 * relevant -- which it won't be if someone who started later
	 * than we did beat us to the punch.
	 */
	do {
		s = atomic_read(&sync_sched_expedited_done);
		if (UINT_CMP_GE((unsigned)s, (unsigned)snap)) {
			smp_mb(); /* ensure test happens before caller kfree */
			break;
		}
	} while (atomic_cmpxchg(&sync_sched_expedited_done, s, snap) != s);

1488 1489 1490 1491 1492 1493
	put_online_cpus();
}
EXPORT_SYMBOL_GPL(synchronize_sched_expedited);

#endif /* #else #ifndef CONFIG_SMP */

1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510
#if !defined(CONFIG_RCU_FAST_NO_HZ)

/*
 * 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.
 *
 * Because we have preemptible RCU, just check whether this CPU needs
 * any flavor of RCU.  Do not chew up lots of CPU cycles with preemption
 * disabled in a most-likely vain attempt to cause RCU not to need this CPU.
 */
int rcu_needs_cpu(int cpu)
{
	return rcu_needs_cpu_quick_check(cpu);
}

1511 1512 1513 1514 1515 1516 1517 1518 1519
/*
 * Check to see if we need to continue a callback-flush operations to
 * allow the last CPU to enter dyntick-idle mode.  But fast dyntick-idle
 * entry is not configured, so we never do need to.
 */
static void rcu_needs_cpu_flush(void)
{
}

1520 1521 1522
#else /* #if !defined(CONFIG_RCU_FAST_NO_HZ) */

#define RCU_NEEDS_CPU_FLUSHES 5
1523
static DEFINE_PER_CPU(int, rcu_dyntick_drain);
1524
static DEFINE_PER_CPU(unsigned long, rcu_dyntick_holdoff);
1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536

/*
 * 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.
 *
 * Because we are not supporting preemptible RCU, attempt to accelerate
 * any current grace periods so that RCU no longer needs this CPU, but
 * only if all other CPUs are already in dynticks-idle mode.  This will
 * allow the CPU cores to be powered down immediately, as opposed to after
 * waiting many milliseconds for grace periods to elapse.
1537 1538 1539
 *
 * Because it is not legal to invoke rcu_process_callbacks() with irqs
 * disabled, we do one pass of force_quiescent_state(), then do a
1540 1541
 * invoke_rcu_cpu_kthread() to cause rcu_process_callbacks() to be invoked
 * later.  The per-cpu rcu_dyntick_drain variable controls the sequencing.
1542 1543 1544
 */
int rcu_needs_cpu(int cpu)
{
1545
	int c = 0;
1546
	int snap;
1547 1548
	int thatcpu;

1549 1550 1551 1552
	/* Check for being in the holdoff period. */
	if (per_cpu(rcu_dyntick_holdoff, cpu) == jiffies)
		return rcu_needs_cpu_quick_check(cpu);

1553
	/* Don't bother unless we are the last non-dyntick-idle CPU. */
1554 1555 1556
	for_each_online_cpu(thatcpu) {
		if (thatcpu == cpu)
			continue;
1557 1558
		snap = atomic_add_return(0, &per_cpu(rcu_dynticks,
						     thatcpu).dynticks);
1559
		smp_mb(); /* Order sampling of snap with end of grace period. */
1560
		if ((snap & 0x1) != 0) {
1561
			per_cpu(rcu_dyntick_drain, cpu) = 0;
1562
			per_cpu(rcu_dyntick_holdoff, cpu) = jiffies - 1;
1563 1564
			return rcu_needs_cpu_quick_check(cpu);
		}
1565
	}
1566 1567 1568 1569 1570 1571 1572

	/* Check and update the rcu_dyntick_drain sequencing. */
	if (per_cpu(rcu_dyntick_drain, cpu) <= 0) {
		/* First time through, initialize the counter. */
		per_cpu(rcu_dyntick_drain, cpu) = RCU_NEEDS_CPU_FLUSHES;
	} else if (--per_cpu(rcu_dyntick_drain, cpu) <= 0) {
		/* We have hit the limit, so time to give up. */
1573
		per_cpu(rcu_dyntick_holdoff, cpu) = jiffies;
1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586
		return rcu_needs_cpu_quick_check(cpu);
	}

	/* Do one step pushing remaining RCU callbacks through. */
	if (per_cpu(rcu_sched_data, cpu).nxtlist) {
		rcu_sched_qs(cpu);
		force_quiescent_state(&rcu_sched_state, 0);
		c = c || per_cpu(rcu_sched_data, cpu).nxtlist;
	}
	if (per_cpu(rcu_bh_data, cpu).nxtlist) {
		rcu_bh_qs(cpu);
		force_quiescent_state(&rcu_bh_state, 0);
		c = c || per_cpu(rcu_bh_data, cpu).nxtlist;
1587 1588 1589
	}

	/* If RCU callbacks are still pending, RCU still needs this CPU. */
1590
	if (c)
1591
		invoke_rcu_cpu_kthread();
1592 1593 1594
	return c;
}

1595 1596 1597 1598 1599 1600 1601
/*
 * Check to see if we need to continue a callback-flush operations to
 * allow the last CPU to enter dyntick-idle mode.
 */
static void rcu_needs_cpu_flush(void)
{
	int cpu = smp_processor_id();
1602
	unsigned long flags;
1603 1604 1605

	if (per_cpu(rcu_dyntick_drain, cpu) <= 0)
		return;
1606
	local_irq_save(flags);
1607
	(void)rcu_needs_cpu(cpu);
1608
	local_irq_restore(flags);
1609 1610
}

1611
#endif /* #else #if !defined(CONFIG_RCU_FAST_NO_HZ) */