rcutree_plugin.h 57.9 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
static void rcu_read_unlock_special(struct task_struct *t);
72 73
static int rcu_preempted_readers_exp(struct rcu_node *rnp);

74 75 76
/*
 * Tell them what RCU they are running.
 */
77
static void __init rcu_bootup_announce(void)
78
{
P
Paul E. McKenney 已提交
79
	printk(KERN_INFO "Preemptible hierarchical RCU implementation.\n");
80
	rcu_bootup_announce_oddness();
81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101
}

/*
 * 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);

102 103 104 105 106 107 108 109 110
/*
 * 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);

111
/*
P
Paul E. McKenney 已提交
112
 * Record a preemptible-RCU quiescent state for the specified CPU.  Note
113 114 115
 * 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.
116 117 118 119
 *
 * Unlike the other rcu_*_qs() functions, callers to this function
 * must disable irqs in order to protect the assignment to
 * ->rcu_read_unlock_special.
120
 */
121
static void rcu_preempt_qs(int cpu)
122 123
{
	struct rcu_data *rdp = &per_cpu(rcu_preempt_data, cpu);
124

125
	rdp->passed_quiesc_completed = rdp->gpnum - 1;
126 127
	barrier();
	rdp->passed_quiesc = 1;
128
	current->rcu_read_unlock_special &= ~RCU_READ_UNLOCK_NEED_QS;
129 130 131
}

/*
132 133 134
 * 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
135 136 137 138 139 140
 * 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.
141 142
 *
 * Caller must disable preemption.
143
 */
144
static void rcu_preempt_note_context_switch(int cpu)
145 146
{
	struct task_struct *t = current;
147
	unsigned long flags;
148 149 150
	struct rcu_data *rdp;
	struct rcu_node *rnp;

151
	if (t->rcu_read_lock_nesting > 0 &&
152 153 154
	    (t->rcu_read_unlock_special & RCU_READ_UNLOCK_BLOCKED) == 0) {

		/* Possibly blocking in an RCU read-side critical section. */
155
		rdp = per_cpu_ptr(rcu_preempt_state.rda, cpu);
156
		rnp = rdp->mynode;
P
Paul E. McKenney 已提交
157
		raw_spin_lock_irqsave(&rnp->lock, flags);
158
		t->rcu_read_unlock_special |= RCU_READ_UNLOCK_BLOCKED;
159
		t->rcu_blocked_node = rnp;
160 161 162 163 164 165 166 167 168

		/*
		 * 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
169 170 171 172 173 174
		 * 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!!!
175 176 177
		 *
		 * But first, note that the current CPU must still be
		 * on line!
178
		 */
179
		WARN_ON_ONCE((rdp->grpmask & rnp->qsmaskinit) == 0);
180
		WARN_ON_ONCE(!list_empty(&t->rcu_node_entry));
181 182 183
		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;
184 185 186 187
#ifdef CONFIG_RCU_BOOST
			if (rnp->boost_tasks != NULL)
				rnp->boost_tasks = rnp->gp_tasks;
#endif /* #ifdef CONFIG_RCU_BOOST */
188 189 190 191 192
		} 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
193
		raw_spin_unlock_irqrestore(&rnp->lock, flags);
194 195 196 197 198 199 200 201
	} else if (t->rcu_read_lock_nesting < 0 &&
		   t->rcu_read_unlock_special) {

		/*
		 * Complete exit from RCU read-side critical section on
		 * behalf of preempted instance of __rcu_read_unlock().
		 */
		rcu_read_unlock_special(t);
202 203 204 205 206 207 208 209 210 211 212
	}

	/*
	 * 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.
	 */
213
	local_irq_save(flags);
214
	rcu_preempt_qs(cpu);
215
	local_irq_restore(flags);
216 217 218
}

/*
P
Paul E. McKenney 已提交
219
 * Tree-preemptible RCU implementation for rcu_read_lock().
220 221 222 223 224
 * Just increment ->rcu_read_lock_nesting, shared state will be updated
 * if we block.
 */
void __rcu_read_lock(void)
{
225
	current->rcu_read_lock_nesting++;
226 227 228 229
	barrier();  /* needed if we ever invoke rcu_read_lock in rcutree.c */
}
EXPORT_SYMBOL_GPL(__rcu_read_lock);

230 231 232 233 234
/*
 * 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.
 */
235
static int rcu_preempt_blocked_readers_cgp(struct rcu_node *rnp)
236
{
237
	return rnp->gp_tasks != NULL;
238 239
}

240 241 242 243 244 245 246
/*
 * 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
247
static void rcu_report_unblock_qs_rnp(struct rcu_node *rnp, unsigned long flags)
248 249 250 251 252
	__releases(rnp->lock)
{
	unsigned long mask;
	struct rcu_node *rnp_p;

253
	if (rnp->qsmask != 0 || rcu_preempt_blocked_readers_cgp(rnp)) {
P
Paul E. McKenney 已提交
254
		raw_spin_unlock_irqrestore(&rnp->lock, flags);
255 256 257 258 259 260 261 262 263 264
		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
265
		rcu_report_qs_rsp(&rcu_preempt_state, flags);
266 267 268 269 270
		return;
	}

	/* Report up the rest of the hierarchy. */
	mask = rnp->grpmask;
P
Paul E. McKenney 已提交
271 272
	raw_spin_unlock(&rnp->lock);	/* irqs remain disabled. */
	raw_spin_lock(&rnp_p->lock);	/* irqs already disabled. */
P
Paul E. McKenney 已提交
273
	rcu_report_qs_rnp(mask, &rcu_preempt_state, rnp_p, flags);
274 275
}

276 277 278 279 280 281 282 283 284 285 286 287 288 289 290
/*
 * 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;
}

291 292 293 294 295
/*
 * 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.
 */
296
static noinline void rcu_read_unlock_special(struct task_struct *t)
297 298
{
	int empty;
299
	int empty_exp;
300
	unsigned long flags;
301
	struct list_head *np;
302 303 304 305 306 307 308 309 310 311 312 313 314 315 316
	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) {
317
		rcu_preempt_qs(smp_processor_id());
318 319 320
	}

	/* Hardware IRQ handlers cannot block. */
321
	if (in_irq() || in_serving_softirq()) {
322 323 324 325 326 327 328 329
		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;

330 331 332 333 334 335
		/*
		 * 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 (;;) {
336
			rnp = t->rcu_blocked_node;
P
Paul E. McKenney 已提交
337
			raw_spin_lock(&rnp->lock);  /* irqs already disabled. */
338
			if (rnp == t->rcu_blocked_node)
339
				break;
P
Paul E. McKenney 已提交
340
			raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */
341
		}
342
		empty = !rcu_preempt_blocked_readers_cgp(rnp);
343 344
		empty_exp = !rcu_preempted_readers_exp(rnp);
		smp_mb(); /* ensure expedited fastpath sees end of RCU c-s. */
345
		np = rcu_next_node_entry(t, rnp);
346
		list_del_init(&t->rcu_node_entry);
347 348 349 350
		if (&t->rcu_node_entry == rnp->gp_tasks)
			rnp->gp_tasks = np;
		if (&t->rcu_node_entry == rnp->exp_tasks)
			rnp->exp_tasks = np;
351 352 353
#ifdef CONFIG_RCU_BOOST
		if (&t->rcu_node_entry == rnp->boost_tasks)
			rnp->boost_tasks = np;
354 355 356 357 358
		/* Snapshot and clear ->rcu_boosted with rcu_node lock held. */
		if (t->rcu_boosted) {
			special |= RCU_READ_UNLOCK_BOOSTED;
			t->rcu_boosted = 0;
		}
359
#endif /* #ifdef CONFIG_RCU_BOOST */
360
		t->rcu_blocked_node = NULL;
361 362 363 364

		/*
		 * 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
365
		 * Note that rcu_report_unblock_qs_rnp() releases rnp->lock.
366
		 */
367
		if (empty)
P
Paul E. McKenney 已提交
368
			raw_spin_unlock_irqrestore(&rnp->lock, flags);
369
		else
P
Paul E. McKenney 已提交
370
			rcu_report_unblock_qs_rnp(rnp, flags);
371

372 373 374 375 376 377 378 379
#ifdef CONFIG_RCU_BOOST
		/* Unboost if we were boosted. */
		if (special & RCU_READ_UNLOCK_BOOSTED) {
			rt_mutex_unlock(t->rcu_boost_mutex);
			t->rcu_boost_mutex = NULL;
		}
#endif /* #ifdef CONFIG_RCU_BOOST */

380 381 382 383 384 385
		/*
		 * 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);
386 387
	} else {
		local_irq_restore(flags);
388 389 390 391
	}
}

/*
P
Paul E. McKenney 已提交
392
 * Tree-preemptible RCU implementation for rcu_read_unlock().
393 394 395 396 397 398 399 400 401 402
 * 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 */
403 404 405 406 407
	if (t->rcu_read_lock_nesting != 1)
		--t->rcu_read_lock_nesting;
	else {
		t->rcu_read_lock_nesting = INT_MIN;
		barrier();  /* assign before ->rcu_read_unlock_special load */
408 409
		if (unlikely(ACCESS_ONCE(t->rcu_read_unlock_special)))
			rcu_read_unlock_special(t);
410 411
		barrier();  /* ->rcu_read_unlock_special load before assign */
		t->rcu_read_lock_nesting = 0;
412
	}
413
#ifdef CONFIG_PROVE_LOCKING
414 415 416 417 418
	{
		int rrln = ACCESS_ONCE(t->rcu_read_lock_nesting);

		WARN_ON_ONCE(rrln < 0 && rrln > INT_MIN / 2);
	}
419
#endif /* #ifdef CONFIG_PROVE_LOCKING */
420 421 422
}
EXPORT_SYMBOL_GPL(__rcu_read_unlock);

423 424 425 426 427 428 429 430 431 432 433
#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;

434
	if (!rcu_preempt_blocked_readers_cgp(rnp))
435 436 437 438 439 440 441
		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);
442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464
}

/*
 * 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 */

465 466 467 468 469 470 471 472
/*
 * 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;

473
	if (!rcu_preempt_blocked_readers_cgp(rnp))
474 475 476 477 478
		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);
479 480
}

481 482 483 484 485 486 487 488 489 490
/*
 * 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;
}

491 492 493 494 495 496
/*
 * 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.
497 498 499
 *
 * Also, if there are blocked tasks on the list, they automatically
 * block the newly created grace period, so set up ->gp_tasks accordingly.
500 501 502
 */
static void rcu_preempt_check_blocked_tasks(struct rcu_node *rnp)
{
503
	WARN_ON_ONCE(rcu_preempt_blocked_readers_cgp(rnp));
504 505
	if (!list_empty(&rnp->blkd_tasks))
		rnp->gp_tasks = rnp->blkd_tasks.next;
506
	WARN_ON_ONCE(rnp->qsmask);
507 508
}

509 510
#ifdef CONFIG_HOTPLUG_CPU

511 512 513 514 515 516
/*
 * 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.
517 518
 * Returns true if there were tasks blocking the current RCU grace
 * period.
519
 *
520 521 522
 * Returns 1 if there was previously a task blocking the current grace
 * period on the specified rcu_node structure.
 *
523 524
 * The caller must hold rnp->lock with irqs disabled.
 */
525 526 527
static int rcu_preempt_offline_tasks(struct rcu_state *rsp,
				     struct rcu_node *rnp,
				     struct rcu_data *rdp)
528 529 530
{
	struct list_head *lp;
	struct list_head *lp_root;
531
	int retval = 0;
532
	struct rcu_node *rnp_root = rcu_get_root(rsp);
533
	struct task_struct *t;
534

535 536
	if (rnp == rnp_root) {
		WARN_ONCE(1, "Last CPU thought to be offlined?");
537
		return 0;  /* Shouldn't happen: at least one CPU online. */
538
	}
539 540 541

	/* If we are on an internal node, complain bitterly. */
	WARN_ON_ONCE(rnp != rdp->mynode);
542 543

	/*
544 545 546 547 548 549 550
	 * 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.
551
	 */
552
	if (rcu_preempt_blocked_readers_cgp(rnp))
553 554 555
		retval |= RCU_OFL_TASKS_NORM_GP;
	if (rcu_preempted_readers_exp(rnp))
		retval |= RCU_OFL_TASKS_EXP_GP;
556 557 558 559 560 561 562 563 564 565 566 567
	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;
568 569 570 571
#ifdef CONFIG_RCU_BOOST
		if (&t->rcu_node_entry == rnp->boost_tasks)
			rnp_root->boost_tasks = rnp->boost_tasks;
#endif /* #ifdef CONFIG_RCU_BOOST */
572
		raw_spin_unlock(&rnp_root->lock); /* irqs still disabled */
573
	}
574 575 576 577 578 579 580 581 582 583

#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 */

584 585
	rnp->gp_tasks = NULL;
	rnp->exp_tasks = NULL;
586
	return retval;
587 588
}

589
/*
P
Paul E. McKenney 已提交
590
 * Do CPU-offline processing for preemptible RCU.
591 592 593 594 595 596 597 598
 */
static void rcu_preempt_offline_cpu(int cpu)
{
	__rcu_offline_cpu(cpu, &rcu_preempt_state);
}

#endif /* #ifdef CONFIG_HOTPLUG_CPU */

599 600 601 602 603 604 605 606 607 608 609 610
/*
 * 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) {
611
		rcu_preempt_qs(cpu);
612 613
		return;
	}
614 615
	if (t->rcu_read_lock_nesting > 0 &&
	    per_cpu(rcu_preempt_data, cpu).qs_pending)
616
		t->rcu_read_unlock_special |= RCU_READ_UNLOCK_NEED_QS;
617 618 619
}

/*
P
Paul E. McKenney 已提交
620
 * Process callbacks for preemptible RCU.
621 622 623 624 625 626 627
 */
static void rcu_preempt_process_callbacks(void)
{
	__rcu_process_callbacks(&rcu_preempt_state,
				&__get_cpu_var(rcu_preempt_data));
}

628 629
#ifdef CONFIG_RCU_BOOST

630 631 632 633 634
static void rcu_preempt_do_callbacks(void)
{
	rcu_do_batch(&rcu_preempt_state, &__get_cpu_var(rcu_preempt_data));
}

635 636
#endif /* #ifdef CONFIG_RCU_BOOST */

637
/*
P
Paul E. McKenney 已提交
638
 * Queue a preemptible-RCU callback for invocation after a grace period.
639 640 641 642 643 644 645
 */
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);

646 647 648 649 650
/**
 * 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
651 652 653 654 655
 * 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.
656 657 658 659 660 661 662 663
 */
void synchronize_rcu(void)
{
	struct rcu_synchronize rcu;

	if (!rcu_scheduler_active)
		return;

664
	init_rcu_head_on_stack(&rcu.head);
665 666 667 668 669
	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);
670
	destroy_rcu_head_on_stack(&rcu.head);
671 672 673
}
EXPORT_SYMBOL_GPL(synchronize_rcu);

674 675 676 677 678 679 680 681 682 683 684 685
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)
{
686
	return rnp->exp_tasks != NULL;
687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718
}

/*
 * 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
719
	raw_spin_lock_irqsave(&rnp->lock, flags);
720
	for (;;) {
721 722
		if (!sync_rcu_preempt_exp_done(rnp)) {
			raw_spin_unlock_irqrestore(&rnp->lock, flags);
723
			break;
724
		}
725
		if (rnp->parent == NULL) {
726
			raw_spin_unlock_irqrestore(&rnp->lock, flags);
727 728 729 730
			wake_up(&sync_rcu_preempt_exp_wq);
			break;
		}
		mask = rnp->grpmask;
P
Paul E. McKenney 已提交
731
		raw_spin_unlock(&rnp->lock); /* irqs remain disabled */
732
		rnp = rnp->parent;
P
Paul E. McKenney 已提交
733
		raw_spin_lock(&rnp->lock); /* irqs already disabled */
734 735 736 737 738 739 740 741 742 743 744 745 746 747
		rnp->expmask &= ~mask;
	}
}

/*
 * 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)
{
748
	unsigned long flags;
749
	int must_wait = 0;
750

751 752 753 754
	raw_spin_lock_irqsave(&rnp->lock, flags);
	if (list_empty(&rnp->blkd_tasks))
		raw_spin_unlock_irqrestore(&rnp->lock, flags);
	else {
755
		rnp->exp_tasks = rnp->blkd_tasks.next;
756
		rcu_initiate_boost(rnp, flags);  /* releases rnp->lock */
757 758
		must_wait = 1;
	}
759 760 761 762
	if (!must_wait)
		rcu_report_exp_rnp(rsp, rnp);
}

763
/*
764 765
 * 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
766
 * the ->blkd_tasks lists and wait for this list to drain.
767 768 769
 */
void synchronize_rcu_expedited(void)
{
770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797
	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. */

798
	/* force all RCU readers onto ->blkd_tasks lists. */
799 800
	synchronize_sched_expedited();

P
Paul E. McKenney 已提交
801
	raw_spin_lock_irqsave(&rsp->onofflock, flags);
802 803 804

	/* 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
805
		raw_spin_lock(&rnp->lock); /* irqs already disabled. */
806
		rnp->expmask = rnp->qsmaskinit;
P
Paul E. McKenney 已提交
807
		raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */
808 809
	}

810
	/* Snapshot current state of ->blkd_tasks lists. */
811 812 813 814 815
	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
816
	raw_spin_unlock_irqrestore(&rsp->onofflock, flags);
817

818
	/* Wait for snapshotted ->blkd_tasks lists to drain. */
819 820 821 822 823 824 825 826 827 828 829
	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. */
830 831 832
}
EXPORT_SYMBOL_GPL(synchronize_rcu_expedited);

833
/*
P
Paul E. McKenney 已提交
834
 * Check to see if there is any immediate preemptible-RCU-related work
835 836 837 838 839 840 841 842 843
 * 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
844
 * Does preemptible RCU need the CPU to stay out of dynticks mode?
845 846 847 848 849 850
 */
static int rcu_preempt_needs_cpu(int cpu)
{
	return !!per_cpu(rcu_preempt_data, cpu).nxtlist;
}

851 852 853 854 855 856 857 858 859
/**
 * 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);

860
/*
P
Paul E. McKenney 已提交
861
 * Initialize preemptible RCU's per-CPU data.
862 863 864 865 866 867
 */
static void __cpuinit rcu_preempt_init_percpu_data(int cpu)
{
	rcu_init_percpu_data(cpu, &rcu_preempt_state, 1);
}

868
/*
P
Paul E. McKenney 已提交
869
 * Move preemptible RCU's callbacks from dying CPU to other online CPU.
870
 */
871
static void rcu_preempt_send_cbs_to_online(void)
872
{
873
	rcu_send_cbs_to_online(&rcu_preempt_state);
874 875
}

876
/*
P
Paul E. McKenney 已提交
877
 * Initialize preemptible RCU's state structures.
878 879 880
 */
static void __init __rcu_init_preempt(void)
{
881
	rcu_init_one(&rcu_preempt_state, &rcu_preempt_data);
882 883
}

884
/*
P
Paul E. McKenney 已提交
885
 * Check for a task exiting while in a preemptible-RCU read-side
886 887 888 889 890 891 892 893 894 895 896
 * 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;
897
	__rcu_read_unlock();
898 899 900 901
}

#else /* #ifdef CONFIG_TREE_PREEMPT_RCU */

902 903
static struct rcu_state *rcu_state = &rcu_sched_state;

904 905 906
/*
 * Tell them what RCU they are running.
 */
907
static void __init rcu_bootup_announce(void)
908 909
{
	printk(KERN_INFO "Hierarchical RCU implementation.\n");
910
	rcu_bootup_announce_oddness();
911 912 913 914 915 916 917 918 919 920 921
}

/*
 * 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);

922 923 924 925 926 927 928 929 930 931
/*
 * 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);

932
/*
P
Paul E. McKenney 已提交
933
 * Because preemptible RCU does not exist, we never have to check for
934 935
 * CPUs being in quiescent states.
 */
936
static void rcu_preempt_note_context_switch(int cpu)
937 938 939
{
}

940
/*
P
Paul E. McKenney 已提交
941
 * Because preemptible RCU does not exist, there are never any preempted
942 943
 * RCU readers.
 */
944
static int rcu_preempt_blocked_readers_cgp(struct rcu_node *rnp)
945 946 947 948
{
	return 0;
}

949 950 951
#ifdef CONFIG_HOTPLUG_CPU

/* Because preemptible RCU does not exist, no quieting of tasks. */
P
Paul E. McKenney 已提交
952
static void rcu_report_unblock_qs_rnp(struct rcu_node *rnp, unsigned long flags)
953
{
P
Paul E. McKenney 已提交
954
	raw_spin_unlock_irqrestore(&rnp->lock, flags);
955 956 957 958
}

#endif /* #ifdef CONFIG_HOTPLUG_CPU */

959
/*
P
Paul E. McKenney 已提交
960
 * Because preemptible RCU does not exist, we never have to check for
961 962 963 964 965 966
 * tasks blocked within RCU read-side critical sections.
 */
static void rcu_print_detail_task_stall(struct rcu_state *rsp)
{
}

967
/*
P
Paul E. McKenney 已提交
968
 * Because preemptible RCU does not exist, we never have to check for
969 970 971 972 973 974
 * tasks blocked within RCU read-side critical sections.
 */
static void rcu_print_task_stall(struct rcu_node *rnp)
{
}

975 976 977 978 979 980 981 982
/*
 * Because preemptible RCU does not exist, there is no need to suppress
 * its CPU stall warnings.
 */
static void rcu_preempt_stall_reset(void)
{
}

983
/*
P
Paul E. McKenney 已提交
984
 * Because there is no preemptible RCU, there can be no readers blocked,
985 986
 * so there is no need to check for blocked tasks.  So check only for
 * bogus qsmask values.
987 988 989
 */
static void rcu_preempt_check_blocked_tasks(struct rcu_node *rnp)
{
990
	WARN_ON_ONCE(rnp->qsmask);
991 992
}

993 994
#ifdef CONFIG_HOTPLUG_CPU

995
/*
P
Paul E. McKenney 已提交
996
 * Because preemptible RCU does not exist, it never needs to migrate
997 998 999
 * tasks that were blocked within RCU read-side critical sections, and
 * such non-existent tasks cannot possibly have been blocking the current
 * grace period.
1000
 */
1001 1002 1003
static int rcu_preempt_offline_tasks(struct rcu_state *rsp,
				     struct rcu_node *rnp,
				     struct rcu_data *rdp)
1004
{
1005
	return 0;
1006 1007
}

1008
/*
P
Paul E. McKenney 已提交
1009
 * Because preemptible RCU does not exist, it never needs CPU-offline
1010 1011 1012 1013 1014 1015 1016 1017
 * processing.
 */
static void rcu_preempt_offline_cpu(int cpu)
{
}

#endif /* #ifdef CONFIG_HOTPLUG_CPU */

1018
/*
P
Paul E. McKenney 已提交
1019
 * Because preemptible RCU does not exist, it never has any callbacks
1020 1021
 * to check.
 */
1022
static void rcu_preempt_check_callbacks(int cpu)
1023 1024 1025 1026
{
}

/*
P
Paul E. McKenney 已提交
1027
 * Because preemptible RCU does not exist, it never has any callbacks
1028 1029
 * to process.
 */
1030
static void rcu_preempt_process_callbacks(void)
1031 1032 1033
{
}

1034 1035
/*
 * Wait for an rcu-preempt grace period, but make it happen quickly.
P
Paul E. McKenney 已提交
1036
 * But because preemptible RCU does not exist, map to rcu-sched.
1037 1038 1039 1040 1041 1042 1043
 */
void synchronize_rcu_expedited(void)
{
	synchronize_sched_expedited();
}
EXPORT_SYMBOL_GPL(synchronize_rcu_expedited);

1044 1045 1046
#ifdef CONFIG_HOTPLUG_CPU

/*
P
Paul E. McKenney 已提交
1047
 * Because preemptible RCU does not exist, there is never any need to
1048 1049 1050 1051 1052 1053 1054 1055 1056 1057
 * 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 */

1058
/*
P
Paul E. McKenney 已提交
1059
 * Because preemptible RCU does not exist, it never has any work to do.
1060 1061 1062 1063 1064 1065 1066
 */
static int rcu_preempt_pending(int cpu)
{
	return 0;
}

/*
P
Paul E. McKenney 已提交
1067
 * Because preemptible RCU does not exist, it never needs any CPU.
1068 1069 1070 1071 1072 1073
 */
static int rcu_preempt_needs_cpu(int cpu)
{
	return 0;
}

1074
/*
P
Paul E. McKenney 已提交
1075
 * Because preemptible RCU does not exist, rcu_barrier() is just
1076 1077 1078 1079 1080 1081 1082 1083
 * another name for rcu_barrier_sched().
 */
void rcu_barrier(void)
{
	rcu_barrier_sched();
}
EXPORT_SYMBOL_GPL(rcu_barrier);

1084
/*
P
Paul E. McKenney 已提交
1085
 * Because preemptible RCU does not exist, there is no per-CPU
1086 1087 1088 1089 1090 1091
 * data to initialize.
 */
static void __cpuinit rcu_preempt_init_percpu_data(int cpu)
{
}

1092
/*
P
Paul E. McKenney 已提交
1093
 * Because there is no preemptible RCU, there are no callbacks to move.
1094
 */
1095
static void rcu_preempt_send_cbs_to_online(void)
1096 1097 1098
{
}

1099
/*
P
Paul E. McKenney 已提交
1100
 * Because preemptible RCU does not exist, it need not be initialized.
1101 1102 1103 1104 1105
 */
static void __init __rcu_init_preempt(void)
{
}

1106
#endif /* #else #ifdef CONFIG_TREE_PREEMPT_RCU */
1107

1108 1109 1110 1111
#ifdef CONFIG_RCU_BOOST

#include "rtmutex_common.h"

1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124
#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 &&
1125
		 ULONG_CMP_LT(jiffies, rnp->boost_time))
1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138
		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 */

1139 1140 1141 1142 1143 1144 1145 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
/*
 * 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.
	 */
1174
	if (rnp->exp_tasks != NULL) {
1175
		tb = rnp->exp_tasks;
1176 1177
		rnp->n_exp_boosts++;
	} else {
1178
		tb = rnp->boost_tasks;
1179 1180 1181
		rnp->n_normal_boosts++;
	}
	rnp->n_tasks_boosted++;
1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201

	/*
	 * 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;
1202
	t->rcu_boosted = 1;
1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217
	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)
{
1218
	invoke_rcu_node_kthread((struct rcu_node *)arg);
1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231
}

/*
 * 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 (;;) {
1232
		rnp->boost_kthread_status = RCU_KTHREAD_WAITING;
1233
		rcu_wait(rnp->boost_tasks || rnp->exp_tasks);
1234
		rnp->boost_kthread_status = RCU_KTHREAD_RUNNING;
1235 1236 1237 1238 1239 1240 1241 1242 1243 1244
		more2boost = rcu_boost(rnp);
		if (more2boost)
			spincnt++;
		else
			spincnt = 0;
		if (spincnt > 10) {
			rcu_yield(rcu_boost_kthread_timer, (unsigned long)rnp);
			spincnt = 0;
		}
	}
1245
	/* NOTREACHED */
1246 1247 1248 1249 1250 1251 1252 1253 1254
	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.
 *
1255 1256 1257
 * The caller must hold rnp->lock, which this function releases,
 * but irqs remain disabled.  The ->boost_kthread_task is immortal,
 * so we don't need to worry about it going away.
1258
 */
1259
static void rcu_initiate_boost(struct rcu_node *rnp, unsigned long flags)
1260 1261 1262
{
	struct task_struct *t;

1263 1264
	if (!rcu_preempt_blocked_readers_cgp(rnp) && rnp->exp_tasks == NULL) {
		rnp->n_balk_exp_gp_tasks++;
1265
		raw_spin_unlock_irqrestore(&rnp->lock, flags);
1266
		return;
1267
	}
1268 1269 1270 1271 1272 1273 1274
	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;
1275
		raw_spin_unlock_irqrestore(&rnp->lock, flags);
1276 1277 1278
		t = rnp->boost_kthread_task;
		if (t != NULL)
			wake_up_process(t);
1279
	} else {
1280
		rcu_initiate_boost_trace(rnp);
1281 1282
		raw_spin_unlock_irqrestore(&rnp->lock, flags);
	}
1283 1284
}

1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301
/*
 * Wake up the per-CPU kthread to invoke RCU callbacks.
 */
static void invoke_rcu_callbacks_kthread(void)
{
	unsigned long flags;

	local_irq_save(flags);
	__this_cpu_write(rcu_cpu_has_work, 1);
	if (__this_cpu_read(rcu_cpu_kthread_task) == NULL) {
		local_irq_restore(flags);
		return;
	}
	wake_up_process(__this_cpu_read(rcu_cpu_kthread_task));
	local_irq_restore(flags);
}

1302 1303 1304 1305 1306
/*
 * 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.
 */
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
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;
}

/*
 * 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;
1342
	rsp->boost = 1;
1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353
	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);
	sp.sched_priority = RCU_KTHREAD_PRIO;
	sched_setscheduler_nocheck(t, SCHED_FIFO, &sp);
1354
	wake_up_process(t); /* get to TASK_INTERRUPTIBLE quickly. */
1355 1356 1357
	return 0;
}

1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 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 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 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 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559
#ifdef CONFIG_HOTPLUG_CPU

/*
 * Stop the RCU's per-CPU kthread when its CPU goes offline,.
 */
static void rcu_stop_cpu_kthread(int cpu)
{
	struct task_struct *t;

	/* Stop the CPU's kthread. */
	t = per_cpu(rcu_cpu_kthread_task, cpu);
	if (t != NULL) {
		per_cpu(rcu_cpu_kthread_task, cpu) = NULL;
		kthread_stop(t);
	}
}

#endif /* #ifdef CONFIG_HOTPLUG_CPU */

static void rcu_kthread_do_work(void)
{
	rcu_do_batch(&rcu_sched_state, &__get_cpu_var(rcu_sched_data));
	rcu_do_batch(&rcu_bh_state, &__get_cpu_var(rcu_bh_data));
	rcu_preempt_do_callbacks();
}

/*
 * Wake up the specified per-rcu_node-structure kthread.
 * Because the per-rcu_node kthreads are immortal, we don't need
 * to do anything to keep them alive.
 */
static void invoke_rcu_node_kthread(struct rcu_node *rnp)
{
	struct task_struct *t;

	t = rnp->node_kthread_task;
	if (t != NULL)
		wake_up_process(t);
}

/*
 * Set the specified CPU's kthread to run RT or not, as specified by
 * the to_rt argument.  The CPU-hotplug locks are held, so the task
 * is not going away.
 */
static void rcu_cpu_kthread_setrt(int cpu, int to_rt)
{
	int policy;
	struct sched_param sp;
	struct task_struct *t;

	t = per_cpu(rcu_cpu_kthread_task, cpu);
	if (t == NULL)
		return;
	if (to_rt) {
		policy = SCHED_FIFO;
		sp.sched_priority = RCU_KTHREAD_PRIO;
	} else {
		policy = SCHED_NORMAL;
		sp.sched_priority = 0;
	}
	sched_setscheduler_nocheck(t, policy, &sp);
}

/*
 * Timer handler to initiate the waking up of per-CPU kthreads that
 * have yielded the CPU due to excess numbers of RCU callbacks.
 * We wake up the per-rcu_node kthread, which in turn will wake up
 * the booster kthread.
 */
static void rcu_cpu_kthread_timer(unsigned long arg)
{
	struct rcu_data *rdp = per_cpu_ptr(rcu_state->rda, arg);
	struct rcu_node *rnp = rdp->mynode;

	atomic_or(rdp->grpmask, &rnp->wakemask);
	invoke_rcu_node_kthread(rnp);
}

/*
 * Drop to non-real-time priority and yield, but only after posting a
 * timer that will cause us to regain our real-time priority if we
 * remain preempted.  Either way, we restore our real-time priority
 * before returning.
 */
static void rcu_yield(void (*f)(unsigned long), unsigned long arg)
{
	struct sched_param sp;
	struct timer_list yield_timer;

	setup_timer_on_stack(&yield_timer, f, arg);
	mod_timer(&yield_timer, jiffies + 2);
	sp.sched_priority = 0;
	sched_setscheduler_nocheck(current, SCHED_NORMAL, &sp);
	set_user_nice(current, 19);
	schedule();
	sp.sched_priority = RCU_KTHREAD_PRIO;
	sched_setscheduler_nocheck(current, SCHED_FIFO, &sp);
	del_timer(&yield_timer);
}

/*
 * Handle cases where the rcu_cpu_kthread() ends up on the wrong CPU.
 * This can happen while the corresponding CPU is either coming online
 * or going offline.  We cannot wait until the CPU is fully online
 * before starting the kthread, because the various notifier functions
 * can wait for RCU grace periods.  So we park rcu_cpu_kthread() until
 * the corresponding CPU is online.
 *
 * Return 1 if the kthread needs to stop, 0 otherwise.
 *
 * Caller must disable bh.  This function can momentarily enable it.
 */
static int rcu_cpu_kthread_should_stop(int cpu)
{
	while (cpu_is_offline(cpu) ||
	       !cpumask_equal(&current->cpus_allowed, cpumask_of(cpu)) ||
	       smp_processor_id() != cpu) {
		if (kthread_should_stop())
			return 1;
		per_cpu(rcu_cpu_kthread_status, cpu) = RCU_KTHREAD_OFFCPU;
		per_cpu(rcu_cpu_kthread_cpu, cpu) = raw_smp_processor_id();
		local_bh_enable();
		schedule_timeout_uninterruptible(1);
		if (!cpumask_equal(&current->cpus_allowed, cpumask_of(cpu)))
			set_cpus_allowed_ptr(current, cpumask_of(cpu));
		local_bh_disable();
	}
	per_cpu(rcu_cpu_kthread_cpu, cpu) = cpu;
	return 0;
}

/*
 * Per-CPU kernel thread that invokes RCU callbacks.  This replaces the
 * earlier RCU softirq.
 */
static int rcu_cpu_kthread(void *arg)
{
	int cpu = (int)(long)arg;
	unsigned long flags;
	int spincnt = 0;
	unsigned int *statusp = &per_cpu(rcu_cpu_kthread_status, cpu);
	char work;
	char *workp = &per_cpu(rcu_cpu_has_work, cpu);

	for (;;) {
		*statusp = RCU_KTHREAD_WAITING;
		rcu_wait(*workp != 0 || kthread_should_stop());
		local_bh_disable();
		if (rcu_cpu_kthread_should_stop(cpu)) {
			local_bh_enable();
			break;
		}
		*statusp = RCU_KTHREAD_RUNNING;
		per_cpu(rcu_cpu_kthread_loops, cpu)++;
		local_irq_save(flags);
		work = *workp;
		*workp = 0;
		local_irq_restore(flags);
		if (work)
			rcu_kthread_do_work();
		local_bh_enable();
		if (*workp != 0)
			spincnt++;
		else
			spincnt = 0;
		if (spincnt > 10) {
			*statusp = RCU_KTHREAD_YIELDING;
			rcu_yield(rcu_cpu_kthread_timer, (unsigned long)cpu);
			spincnt = 0;
		}
	}
	*statusp = RCU_KTHREAD_STOPPED;
	return 0;
}

/*
 * Spawn a per-CPU kthread, setting up affinity and priority.
 * Because the CPU hotplug lock is held, no other CPU will be attempting
 * to manipulate rcu_cpu_kthread_task.  There might be another CPU
 * attempting to access it during boot, but the locking in kthread_bind()
 * will enforce sufficient ordering.
 *
 * Please note that we cannot simply refuse to wake up the per-CPU
 * kthread because kthreads are created in TASK_UNINTERRUPTIBLE state,
 * which can result in softlockup complaints if the task ends up being
 * idle for more than a couple of minutes.
 *
 * However, please note also that we cannot bind the per-CPU kthread to its
 * CPU until that CPU is fully online.  We also cannot wait until the
 * CPU is fully online before we create its per-CPU kthread, as this would
 * deadlock the system when CPU notifiers tried waiting for grace
 * periods.  So we bind the per-CPU kthread to its CPU only if the CPU
 * is online.  If its CPU is not yet fully online, then the code in
 * rcu_cpu_kthread() will wait until it is fully online, and then do
 * the binding.
 */
static int __cpuinit rcu_spawn_one_cpu_kthread(int cpu)
{
	struct sched_param sp;
	struct task_struct *t;

1560
	if (!rcu_scheduler_fully_active ||
1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 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
	    per_cpu(rcu_cpu_kthread_task, cpu) != NULL)
		return 0;
	t = kthread_create(rcu_cpu_kthread, (void *)(long)cpu, "rcuc%d", cpu);
	if (IS_ERR(t))
		return PTR_ERR(t);
	if (cpu_online(cpu))
		kthread_bind(t, cpu);
	per_cpu(rcu_cpu_kthread_cpu, cpu) = cpu;
	WARN_ON_ONCE(per_cpu(rcu_cpu_kthread_task, cpu) != NULL);
	sp.sched_priority = RCU_KTHREAD_PRIO;
	sched_setscheduler_nocheck(t, SCHED_FIFO, &sp);
	per_cpu(rcu_cpu_kthread_task, cpu) = t;
	wake_up_process(t); /* Get to TASK_INTERRUPTIBLE quickly. */
	return 0;
}

/*
 * Per-rcu_node kthread, which is in charge of waking up the per-CPU
 * kthreads when needed.  We ignore requests to wake up kthreads
 * for offline CPUs, which is OK because force_quiescent_state()
 * takes care of this case.
 */
static int rcu_node_kthread(void *arg)
{
	int cpu;
	unsigned long flags;
	unsigned long mask;
	struct rcu_node *rnp = (struct rcu_node *)arg;
	struct sched_param sp;
	struct task_struct *t;

	for (;;) {
		rnp->node_kthread_status = RCU_KTHREAD_WAITING;
		rcu_wait(atomic_read(&rnp->wakemask) != 0);
		rnp->node_kthread_status = RCU_KTHREAD_RUNNING;
		raw_spin_lock_irqsave(&rnp->lock, flags);
		mask = atomic_xchg(&rnp->wakemask, 0);
		rcu_initiate_boost(rnp, flags); /* releases rnp->lock. */
		for (cpu = rnp->grplo; cpu <= rnp->grphi; cpu++, mask >>= 1) {
			if ((mask & 0x1) == 0)
				continue;
			preempt_disable();
			t = per_cpu(rcu_cpu_kthread_task, cpu);
			if (!cpu_online(cpu) || t == NULL) {
				preempt_enable();
				continue;
			}
			per_cpu(rcu_cpu_has_work, cpu) = 1;
			sp.sched_priority = RCU_KTHREAD_PRIO;
			sched_setscheduler_nocheck(t, SCHED_FIFO, &sp);
			preempt_enable();
		}
	}
	/* NOTREACHED */
	rnp->node_kthread_status = RCU_KTHREAD_STOPPED;
	return 0;
}

/*
 * Set the per-rcu_node kthread's affinity to cover all CPUs that are
 * served by the rcu_node in question.  The CPU hotplug lock is still
 * held, so the value of rnp->qsmaskinit will be stable.
 *
 * We don't include outgoingcpu in the affinity set, use -1 if there is
 * no outgoing CPU.  If there are no CPUs left in the affinity set,
 * this function allows the kthread to execute on any CPU.
 */
static void rcu_node_kthread_setaffinity(struct rcu_node *rnp, int outgoingcpu)
{
	cpumask_var_t cm;
	int cpu;
	unsigned long mask = rnp->qsmaskinit;

	if (rnp->node_kthread_task == NULL)
		return;
	if (!alloc_cpumask_var(&cm, GFP_KERNEL))
		return;
	cpumask_clear(cm);
	for (cpu = rnp->grplo; cpu <= rnp->grphi; cpu++, mask >>= 1)
		if ((mask & 0x1) && cpu != outgoingcpu)
			cpumask_set_cpu(cpu, cm);
	if (cpumask_weight(cm) == 0) {
		cpumask_setall(cm);
		for (cpu = rnp->grplo; cpu <= rnp->grphi; cpu++)
			cpumask_clear_cpu(cpu, cm);
		WARN_ON_ONCE(cpumask_weight(cm) == 0);
	}
	set_cpus_allowed_ptr(rnp->node_kthread_task, cm);
	rcu_boost_kthread_setaffinity(rnp, cm);
	free_cpumask_var(cm);
}

/*
 * Spawn a per-rcu_node kthread, setting priority and affinity.
 * Called during boot before online/offline can happen, or, if
 * during runtime, with the main CPU-hotplug locks held.  So only
 * one of these can be executing at a time.
 */
static int __cpuinit rcu_spawn_one_node_kthread(struct rcu_state *rsp,
						struct rcu_node *rnp)
{
	unsigned long flags;
	int rnp_index = rnp - &rsp->node[0];
	struct sched_param sp;
	struct task_struct *t;

1667
	if (!rcu_scheduler_fully_active ||
1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692
	    rnp->qsmaskinit == 0)
		return 0;
	if (rnp->node_kthread_task == NULL) {
		t = kthread_create(rcu_node_kthread, (void *)rnp,
				   "rcun%d", rnp_index);
		if (IS_ERR(t))
			return PTR_ERR(t);
		raw_spin_lock_irqsave(&rnp->lock, flags);
		rnp->node_kthread_task = t;
		raw_spin_unlock_irqrestore(&rnp->lock, flags);
		sp.sched_priority = 99;
		sched_setscheduler_nocheck(t, SCHED_FIFO, &sp);
		wake_up_process(t); /* get to TASK_INTERRUPTIBLE quickly. */
	}
	return rcu_spawn_one_boost_kthread(rsp, rnp, rnp_index);
}

/*
 * Spawn all kthreads -- called as soon as the scheduler is running.
 */
static int __init rcu_spawn_kthreads(void)
{
	int cpu;
	struct rcu_node *rnp;

1693
	rcu_scheduler_fully_active = 1;
1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714
	for_each_possible_cpu(cpu) {
		per_cpu(rcu_cpu_has_work, cpu) = 0;
		if (cpu_online(cpu))
			(void)rcu_spawn_one_cpu_kthread(cpu);
	}
	rnp = rcu_get_root(rcu_state);
	(void)rcu_spawn_one_node_kthread(rcu_state, rnp);
	if (NUM_RCU_NODES > 1) {
		rcu_for_each_leaf_node(rcu_state, rnp)
			(void)rcu_spawn_one_node_kthread(rcu_state, rnp);
	}
	return 0;
}
early_initcall(rcu_spawn_kthreads);

static void __cpuinit rcu_prepare_kthreads(int cpu)
{
	struct rcu_data *rdp = per_cpu_ptr(rcu_state->rda, cpu);
	struct rcu_node *rnp = rdp->mynode;

	/* Fire up the incoming CPU's kthread and leaf rcu_node kthread. */
1715
	if (rcu_scheduler_fully_active) {
1716 1717 1718 1719 1720 1721
		(void)rcu_spawn_one_cpu_kthread(cpu);
		if (rnp->node_kthread_task == NULL)
			(void)rcu_spawn_one_node_kthread(rcu_state, rnp);
	}
}

1722 1723
#else /* #ifdef CONFIG_RCU_BOOST */

1724
static void rcu_initiate_boost(struct rcu_node *rnp, unsigned long flags)
1725
{
1726
	raw_spin_unlock_irqrestore(&rnp->lock, flags);
1727 1728
}

1729
static void invoke_rcu_callbacks_kthread(void)
1730
{
1731
	WARN_ON_ONCE(1);
1732 1733 1734 1735 1736 1737
}

static void rcu_preempt_boost_start_gp(struct rcu_node *rnp)
{
}

1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753
#ifdef CONFIG_HOTPLUG_CPU

static void rcu_stop_cpu_kthread(int cpu)
{
}

#endif /* #ifdef CONFIG_HOTPLUG_CPU */

static void rcu_node_kthread_setaffinity(struct rcu_node *rnp, int outgoingcpu)
{
}

static void rcu_cpu_kthread_setrt(int cpu, int to_rt)
{
}

1754 1755 1756 1757 1758 1759 1760
static int __init rcu_scheduler_really_started(void)
{
	rcu_scheduler_fully_active = 1;
	return 0;
}
early_initcall(rcu_scheduler_really_started);

1761 1762 1763 1764
static void __cpuinit rcu_prepare_kthreads(int cpu)
{
}

1765 1766
#endif /* #else #ifdef CONFIG_RCU_BOOST */

1767 1768 1769 1770 1771 1772 1773 1774 1775 1776
#ifndef CONFIG_SMP

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

#else /* #ifndef CONFIG_SMP */

1777 1778
static atomic_t sync_sched_expedited_started = ATOMIC_INIT(0);
static atomic_t sync_sched_expedited_done = ATOMIC_INIT(0);
1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805

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.
1806
 *
1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826
 * 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().
1827 1828 1829
 */
void synchronize_sched_expedited(void)
{
1830
	int firstsnap, s, snap, trycount = 0;
1831

1832 1833
	/* Note that atomic_inc_return() implies full memory barrier. */
	firstsnap = snap = atomic_inc_return(&sync_sched_expedited_started);
1834
	get_online_cpus();
1835 1836 1837 1838 1839

	/*
	 * Each pass through the following loop attempts to force a
	 * context switch on each CPU.
	 */
1840 1841 1842 1843
	while (try_stop_cpus(cpu_online_mask,
			     synchronize_sched_expedited_cpu_stop,
			     NULL) == -EAGAIN) {
		put_online_cpus();
1844 1845

		/* No joy, try again later.  Or just synchronize_sched(). */
1846 1847 1848 1849 1850 1851
		if (trycount++ < 10)
			udelay(trycount * num_online_cpus());
		else {
			synchronize_sched();
			return;
		}
1852 1853 1854 1855

		/* 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)) {
1856 1857 1858
			smp_mb(); /* ensure test happens before caller kfree */
			return;
		}
1859 1860 1861 1862 1863 1864 1865 1866 1867

		/*
		 * 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.
		 */
1868
		get_online_cpus();
1869 1870
		snap = atomic_read(&sync_sched_expedited_started) - 1;
		smp_mb(); /* ensure read is before try_stop_cpus(). */
1871
	}
1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886

	/*
	 * 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);

1887 1888 1889 1890 1891 1892
	put_online_cpus();
}
EXPORT_SYMBOL_GPL(synchronize_sched_expedited);

#endif /* #else #ifndef CONFIG_SMP */

1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909
#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);
}

1910 1911 1912 1913 1914 1915 1916 1917 1918
/*
 * 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)
{
}

1919 1920 1921
#else /* #if !defined(CONFIG_RCU_FAST_NO_HZ) */

#define RCU_NEEDS_CPU_FLUSHES 5
1922
static DEFINE_PER_CPU(int, rcu_dyntick_drain);
1923
static DEFINE_PER_CPU(unsigned long, rcu_dyntick_holdoff);
1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935

/*
 * 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.
1936 1937 1938
 *
 * Because it is not legal to invoke rcu_process_callbacks() with irqs
 * disabled, we do one pass of force_quiescent_state(), then do a
1939
 * invoke_rcu_core() to cause rcu_process_callbacks() to be invoked
1940
 * later.  The per-cpu rcu_dyntick_drain variable controls the sequencing.
1941 1942 1943
 */
int rcu_needs_cpu(int cpu)
{
1944
	int c = 0;
1945
	int snap;
1946 1947
	int thatcpu;

1948 1949 1950 1951
	/* Check for being in the holdoff period. */
	if (per_cpu(rcu_dyntick_holdoff, cpu) == jiffies)
		return rcu_needs_cpu_quick_check(cpu);

1952
	/* Don't bother unless we are the last non-dyntick-idle CPU. */
1953 1954 1955
	for_each_online_cpu(thatcpu) {
		if (thatcpu == cpu)
			continue;
1956 1957
		snap = atomic_add_return(0, &per_cpu(rcu_dynticks,
						     thatcpu).dynticks);
1958
		smp_mb(); /* Order sampling of snap with end of grace period. */
1959
		if ((snap & 0x1) != 0) {
1960
			per_cpu(rcu_dyntick_drain, cpu) = 0;
1961
			per_cpu(rcu_dyntick_holdoff, cpu) = jiffies - 1;
1962 1963
			return rcu_needs_cpu_quick_check(cpu);
		}
1964
	}
1965 1966 1967 1968 1969 1970 1971

	/* 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. */
1972
		per_cpu(rcu_dyntick_holdoff, cpu) = jiffies;
1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985
		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;
1986 1987 1988
	}

	/* If RCU callbacks are still pending, RCU still needs this CPU. */
1989
	if (c)
1990
		invoke_rcu_core();
1991 1992 1993
	return c;
}

1994 1995 1996 1997 1998 1999 2000
/*
 * 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();
2001
	unsigned long flags;
2002 2003 2004

	if (per_cpu(rcu_dyntick_drain, cpu) <= 0)
		return;
2005
	local_irq_save(flags);
2006
	(void)rcu_needs_cpu(cpu);
2007
	local_irq_restore(flags);
2008 2009
}

2010
#endif /* #else #if !defined(CONFIG_RCU_FAST_NO_HZ) */