rcutree_plugin.h 61.9 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/oom.h>
29
#include <linux/smpboot.h>
30

31 32 33 34 35 36 37 38
#define RCU_KTHREAD_PRIO 1

#ifdef CONFIG_RCU_BOOST
#define RCU_BOOST_PRIO CONFIG_RCU_BOOST_PRIO
#else
#define RCU_BOOST_PRIO RCU_KTHREAD_PRIO
#endif

39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
/*
 * 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
66
#if defined(CONFIG_TREE_PREEMPT_RCU) && !defined(CONFIG_RCU_CPU_STALL_VERBOSE)
67 68 69 70
	printk(KERN_INFO "\tDump stacks of tasks blocking RCU-preempt GP.\n");
#endif
#if defined(CONFIG_RCU_CPU_STALL_INFO)
	printk(KERN_INFO "\tAdditional per-CPU info printed with stalls.\n");
71 72
#endif
#if NUM_RCU_LVL_4 != 0
73
	printk(KERN_INFO "\tFour-level hierarchy is enabled.\n");
74
#endif
75 76
	if (rcu_fanout_leaf != CONFIG_RCU_FANOUT_LEAF)
		printk(KERN_INFO "\tExperimental boot-time adjustment of leaf fanout to %d.\n", rcu_fanout_leaf);
77 78
	if (nr_cpu_ids != NR_CPUS)
		printk(KERN_INFO "\tRCU restricting CPUs from NR_CPUS=%d to nr_cpu_ids=%d.\n", NR_CPUS, nr_cpu_ids);
79 80
}

81 82
#ifdef CONFIG_TREE_PREEMPT_RCU

83 84
struct rcu_state rcu_preempt_state =
	RCU_STATE_INITIALIZER(rcu_preempt, call_rcu);
85
DEFINE_PER_CPU(struct rcu_data, rcu_preempt_data);
86
static struct rcu_state *rcu_state = &rcu_preempt_state;
87

88 89
static int rcu_preempted_readers_exp(struct rcu_node *rnp);

90 91 92
/*
 * Tell them what RCU they are running.
 */
93
static void __init rcu_bootup_announce(void)
94
{
P
Paul E. McKenney 已提交
95
	printk(KERN_INFO "Preemptible hierarchical RCU implementation.\n");
96
	rcu_bootup_announce_oddness();
97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117
}

/*
 * 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);

118 119 120 121 122
/*
 * Force a quiescent state for preemptible RCU.
 */
void rcu_force_quiescent_state(void)
{
123
	force_quiescent_state(&rcu_preempt_state);
124 125 126
}
EXPORT_SYMBOL_GPL(rcu_force_quiescent_state);

127
/*
P
Paul E. McKenney 已提交
128
 * Record a preemptible-RCU quiescent state for the specified CPU.  Note
129 130 131
 * 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.
132 133 134 135
 *
 * Unlike the other rcu_*_qs() functions, callers to this function
 * must disable irqs in order to protect the assignment to
 * ->rcu_read_unlock_special.
136
 */
137
static void rcu_preempt_qs(int cpu)
138 139
{
	struct rcu_data *rdp = &per_cpu(rcu_preempt_data, cpu);
140

141
	if (rdp->passed_quiesce == 0)
142
		trace_rcu_grace_period("rcu_preempt", rdp->gpnum, "cpuqs");
143
	rdp->passed_quiesce = 1;
144
	current->rcu_read_unlock_special &= ~RCU_READ_UNLOCK_NEED_QS;
145 146 147
}

/*
148 149 150
 * 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
151 152 153 154 155 156
 * 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.
157 158
 *
 * Caller must disable preemption.
159
 */
160
static void rcu_preempt_note_context_switch(int cpu)
161 162
{
	struct task_struct *t = current;
163
	unsigned long flags;
164 165 166
	struct rcu_data *rdp;
	struct rcu_node *rnp;

167
	if (t->rcu_read_lock_nesting > 0 &&
168 169 170
	    (t->rcu_read_unlock_special & RCU_READ_UNLOCK_BLOCKED) == 0) {

		/* Possibly blocking in an RCU read-side critical section. */
171
		rdp = per_cpu_ptr(rcu_preempt_state.rda, cpu);
172
		rnp = rdp->mynode;
P
Paul E. McKenney 已提交
173
		raw_spin_lock_irqsave(&rnp->lock, flags);
174
		t->rcu_read_unlock_special |= RCU_READ_UNLOCK_BLOCKED;
175
		t->rcu_blocked_node = rnp;
176 177 178 179 180 181 182 183 184

		/*
		 * 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
185 186 187 188 189 190
		 * 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!!!
191 192 193
		 *
		 * But first, note that the current CPU must still be
		 * on line!
194
		 */
195
		WARN_ON_ONCE((rdp->grpmask & rnp->qsmaskinit) == 0);
196
		WARN_ON_ONCE(!list_empty(&t->rcu_node_entry));
197 198 199
		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;
200 201 202 203
#ifdef CONFIG_RCU_BOOST
			if (rnp->boost_tasks != NULL)
				rnp->boost_tasks = rnp->gp_tasks;
#endif /* #ifdef CONFIG_RCU_BOOST */
204 205 206 207 208
		} else {
			list_add(&t->rcu_node_entry, &rnp->blkd_tasks);
			if (rnp->qsmask & rdp->grpmask)
				rnp->gp_tasks = &t->rcu_node_entry;
		}
209 210 211 212 213
		trace_rcu_preempt_task(rdp->rsp->name,
				       t->pid,
				       (rnp->qsmask & rdp->grpmask)
				       ? rnp->gpnum
				       : rnp->gpnum + 1);
P
Paul E. McKenney 已提交
214
		raw_spin_unlock_irqrestore(&rnp->lock, flags);
215 216 217 218 219 220 221 222
	} 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);
223 224 225 226 227 228 229 230 231 232 233
	}

	/*
	 * 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.
	 */
234
	local_irq_save(flags);
235
	rcu_preempt_qs(cpu);
236
	local_irq_restore(flags);
237 238
}

239 240 241 242 243
/*
 * 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.
 */
244
static int rcu_preempt_blocked_readers_cgp(struct rcu_node *rnp)
245
{
246
	return rnp->gp_tasks != NULL;
247 248
}

249 250 251 252 253 254 255
/*
 * 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
256
static void rcu_report_unblock_qs_rnp(struct rcu_node *rnp, unsigned long flags)
257 258 259 260 261
	__releases(rnp->lock)
{
	unsigned long mask;
	struct rcu_node *rnp_p;

262
	if (rnp->qsmask != 0 || rcu_preempt_blocked_readers_cgp(rnp)) {
P
Paul E. McKenney 已提交
263
		raw_spin_unlock_irqrestore(&rnp->lock, flags);
264 265 266 267 268 269 270 271 272 273
		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
274
		rcu_report_qs_rsp(&rcu_preempt_state, flags);
275 276 277 278 279
		return;
	}

	/* Report up the rest of the hierarchy. */
	mask = rnp->grpmask;
P
Paul E. McKenney 已提交
280 281
	raw_spin_unlock(&rnp->lock);	/* irqs remain disabled. */
	raw_spin_lock(&rnp_p->lock);	/* irqs already disabled. */
P
Paul E. McKenney 已提交
282
	rcu_report_qs_rnp(mask, &rcu_preempt_state, rnp_p, flags);
283 284
}

285 286 287 288 289 290 291 292 293 294 295 296 297 298 299
/*
 * 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;
}

300 301 302 303 304
/*
 * 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.
 */
305
void rcu_read_unlock_special(struct task_struct *t)
306 307
{
	int empty;
308
	int empty_exp;
309
	int empty_exp_now;
310
	unsigned long flags;
311
	struct list_head *np;
312 313 314
#ifdef CONFIG_RCU_BOOST
	struct rt_mutex *rbmp = NULL;
#endif /* #ifdef CONFIG_RCU_BOOST */
315 316 317 318 319 320 321 322 323 324 325 326 327 328 329
	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) {
330
		rcu_preempt_qs(smp_processor_id());
331 332 333
	}

	/* Hardware IRQ handlers cannot block. */
334
	if (in_irq() || in_serving_softirq()) {
335 336 337 338 339 340 341 342
		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;

343 344 345 346 347 348
		/*
		 * 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 (;;) {
349
			rnp = t->rcu_blocked_node;
P
Paul E. McKenney 已提交
350
			raw_spin_lock(&rnp->lock);  /* irqs already disabled. */
351
			if (rnp == t->rcu_blocked_node)
352
				break;
P
Paul E. McKenney 已提交
353
			raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */
354
		}
355
		empty = !rcu_preempt_blocked_readers_cgp(rnp);
356 357
		empty_exp = !rcu_preempted_readers_exp(rnp);
		smp_mb(); /* ensure expedited fastpath sees end of RCU c-s. */
358
		np = rcu_next_node_entry(t, rnp);
359
		list_del_init(&t->rcu_node_entry);
360
		t->rcu_blocked_node = NULL;
361 362
		trace_rcu_unlock_preempted_task("rcu_preempt",
						rnp->gpnum, t->pid);
363 364 365 366
		if (&t->rcu_node_entry == rnp->gp_tasks)
			rnp->gp_tasks = np;
		if (&t->rcu_node_entry == rnp->exp_tasks)
			rnp->exp_tasks = np;
367 368 369
#ifdef CONFIG_RCU_BOOST
		if (&t->rcu_node_entry == rnp->boost_tasks)
			rnp->boost_tasks = np;
370 371 372 373
		/* Snapshot/clear ->rcu_boost_mutex with rcu_node lock held. */
		if (t->rcu_boost_mutex) {
			rbmp = t->rcu_boost_mutex;
			t->rcu_boost_mutex = NULL;
374
		}
375
#endif /* #ifdef CONFIG_RCU_BOOST */
376 377 378 379

		/*
		 * If this was the last task on the current list, and if
		 * we aren't waiting on any CPUs, report the quiescent state.
380 381
		 * Note that rcu_report_unblock_qs_rnp() releases rnp->lock,
		 * so we must take a snapshot of the expedited state.
382
		 */
383
		empty_exp_now = !rcu_preempted_readers_exp(rnp);
384 385 386 387 388 389 390 391
		if (!empty && !rcu_preempt_blocked_readers_cgp(rnp)) {
			trace_rcu_quiescent_state_report("preempt_rcu",
							 rnp->gpnum,
							 0, rnp->qsmask,
							 rnp->level,
							 rnp->grplo,
							 rnp->grphi,
							 !!rnp->gp_tasks);
P
Paul E. McKenney 已提交
392
			rcu_report_unblock_qs_rnp(rnp, flags);
393
		} else {
394
			raw_spin_unlock_irqrestore(&rnp->lock, flags);
395
		}
396

397 398
#ifdef CONFIG_RCU_BOOST
		/* Unboost if we were boosted. */
399 400
		if (rbmp)
			rt_mutex_unlock(rbmp);
401 402
#endif /* #ifdef CONFIG_RCU_BOOST */

403 404 405 406
		/*
		 * If this was the last task on the expedited lists,
		 * then we need to report up the rcu_node hierarchy.
		 */
407
		if (!empty_exp && empty_exp_now)
408
			rcu_report_exp_rnp(&rcu_preempt_state, rnp, true);
409 410
	} else {
		local_irq_restore(flags);
411 412 413
	}
}

414 415 416 417 418 419 420 421 422 423 424
#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;

425
	raw_spin_lock_irqsave(&rnp->lock, flags);
426 427 428 429
	if (!rcu_preempt_blocked_readers_cgp(rnp)) {
		raw_spin_unlock_irqrestore(&rnp->lock, flags);
		return;
	}
430 431 432 433 434
	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);
435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457
}

/*
 * 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 */

458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482
#ifdef CONFIG_RCU_CPU_STALL_INFO

static void rcu_print_task_stall_begin(struct rcu_node *rnp)
{
	printk(KERN_ERR "\tTasks blocked on level-%d rcu_node (CPUs %d-%d):",
	       rnp->level, rnp->grplo, rnp->grphi);
}

static void rcu_print_task_stall_end(void)
{
	printk(KERN_CONT "\n");
}

#else /* #ifdef CONFIG_RCU_CPU_STALL_INFO */

static void rcu_print_task_stall_begin(struct rcu_node *rnp)
{
}

static void rcu_print_task_stall_end(void)
{
}

#endif /* #else #ifdef CONFIG_RCU_CPU_STALL_INFO */

483 484 485 486
/*
 * Scan the current list of tasks blocked within RCU read-side critical
 * sections, printing out the tid of each.
 */
487
static int rcu_print_task_stall(struct rcu_node *rnp)
488 489
{
	struct task_struct *t;
490
	int ndetected = 0;
491

492
	if (!rcu_preempt_blocked_readers_cgp(rnp))
493
		return 0;
494
	rcu_print_task_stall_begin(rnp);
495 496
	t = list_entry(rnp->gp_tasks,
		       struct task_struct, rcu_node_entry);
497
	list_for_each_entry_continue(t, &rnp->blkd_tasks, rcu_node_entry) {
498
		printk(KERN_CONT " P%d", t->pid);
499 500
		ndetected++;
	}
501
	rcu_print_task_stall_end();
502
	return ndetected;
503 504
}

505 506 507 508 509 510
/*
 * 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.
511 512 513
 *
 * Also, if there are blocked tasks on the list, they automatically
 * block the newly created grace period, so set up ->gp_tasks accordingly.
514 515 516
 */
static void rcu_preempt_check_blocked_tasks(struct rcu_node *rnp)
{
517
	WARN_ON_ONCE(rcu_preempt_blocked_readers_cgp(rnp));
518 519
	if (!list_empty(&rnp->blkd_tasks))
		rnp->gp_tasks = rnp->blkd_tasks.next;
520
	WARN_ON_ONCE(rnp->qsmask);
521 522
}

523 524
#ifdef CONFIG_HOTPLUG_CPU

525 526 527 528 529 530
/*
 * 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.
531 532
 * Returns true if there were tasks blocking the current RCU grace
 * period.
533
 *
534 535 536
 * Returns 1 if there was previously a task blocking the current grace
 * period on the specified rcu_node structure.
 *
537 538
 * The caller must hold rnp->lock with irqs disabled.
 */
539 540 541
static int rcu_preempt_offline_tasks(struct rcu_state *rsp,
				     struct rcu_node *rnp,
				     struct rcu_data *rdp)
542 543 544
{
	struct list_head *lp;
	struct list_head *lp_root;
545
	int retval = 0;
546
	struct rcu_node *rnp_root = rcu_get_root(rsp);
547
	struct task_struct *t;
548

549 550
	if (rnp == rnp_root) {
		WARN_ONCE(1, "Last CPU thought to be offlined?");
551
		return 0;  /* Shouldn't happen: at least one CPU online. */
552
	}
553 554 555

	/* If we are on an internal node, complain bitterly. */
	WARN_ON_ONCE(rnp != rdp->mynode);
556 557

	/*
558 559 560 561 562 563 564
	 * 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.
565
	 */
566
	if (rcu_preempt_blocked_readers_cgp(rnp) && rnp->qsmask == 0)
567 568 569
		retval |= RCU_OFL_TASKS_NORM_GP;
	if (rcu_preempted_readers_exp(rnp))
		retval |= RCU_OFL_TASKS_EXP_GP;
570 571 572 573 574 575 576 577 578 579 580 581
	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;
582 583 584 585
#ifdef CONFIG_RCU_BOOST
		if (&t->rcu_node_entry == rnp->boost_tasks)
			rnp_root->boost_tasks = rnp->boost_tasks;
#endif /* #ifdef CONFIG_RCU_BOOST */
586
		raw_spin_unlock(&rnp_root->lock); /* irqs still disabled */
587
	}
588

589 590
	rnp->gp_tasks = NULL;
	rnp->exp_tasks = NULL;
591
#ifdef CONFIG_RCU_BOOST
592
	rnp->boost_tasks = NULL;
593 594 595 596 597
	/*
	 * In case root is being boosted and leaf was not.  Make sure
	 * that we boost the tasks blocking the current grace period
	 * in this case.
	 */
598 599
	raw_spin_lock(&rnp_root->lock); /* irqs already disabled */
	if (rnp_root->boost_tasks != NULL &&
600 601
	    rnp_root->boost_tasks != rnp_root->gp_tasks &&
	    rnp_root->boost_tasks != rnp_root->exp_tasks)
602 603 604 605
		rnp_root->boost_tasks = rnp_root->gp_tasks;
	raw_spin_unlock(&rnp_root->lock); /* irqs still disabled */
#endif /* #ifdef CONFIG_RCU_BOOST */

606
	return retval;
607 608
}

609 610
#endif /* #ifdef CONFIG_HOTPLUG_CPU */

611 612 613 614 615 616 617 618 619 620 621 622
/*
 * 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) {
623
		rcu_preempt_qs(cpu);
624 625
		return;
	}
626 627
	if (t->rcu_read_lock_nesting > 0 &&
	    per_cpu(rcu_preempt_data, cpu).qs_pending)
628
		t->rcu_read_unlock_special |= RCU_READ_UNLOCK_NEED_QS;
629 630
}

631 632
#ifdef CONFIG_RCU_BOOST

633 634 635 636 637
static void rcu_preempt_do_callbacks(void)
{
	rcu_do_batch(&rcu_preempt_state, &__get_cpu_var(rcu_preempt_data));
}

638 639
#endif /* #ifdef CONFIG_RCU_BOOST */

640
/*
P
Paul E. McKenney 已提交
641
 * Queue a preemptible-RCU callback for invocation after a grace period.
642 643 644
 */
void call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu))
{
645
	__call_rcu(head, func, &rcu_preempt_state, 0);
646 647 648
}
EXPORT_SYMBOL_GPL(call_rcu);

649 650 651 652 653 654 655 656 657 658 659 660 661 662
/*
 * Queue an RCU callback for lazy invocation after a grace period.
 * This will likely be later named something like "call_rcu_lazy()",
 * but this change will require some way of tagging the lazy RCU
 * callbacks in the list of pending callbacks.  Until then, this
 * function may only be called from __kfree_rcu().
 */
void kfree_call_rcu(struct rcu_head *head,
		    void (*func)(struct rcu_head *rcu))
{
	__call_rcu(head, func, &rcu_preempt_state, 1);
}
EXPORT_SYMBOL_GPL(kfree_call_rcu);

663 664 665 666 667
/**
 * 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
668 669 670 671 672
 * 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.
673 674 675
 */
void synchronize_rcu(void)
{
676 677 678 679
	rcu_lockdep_assert(!lock_is_held(&rcu_bh_lock_map) &&
			   !lock_is_held(&rcu_lock_map) &&
			   !lock_is_held(&rcu_sched_lock_map),
			   "Illegal synchronize_rcu() in RCU read-side critical section");
680 681
	if (!rcu_scheduler_active)
		return;
682
	wait_rcu_gp(call_rcu);
683 684 685
}
EXPORT_SYMBOL_GPL(synchronize_rcu);

686
static DECLARE_WAIT_QUEUE_HEAD(sync_rcu_preempt_exp_wq);
687
static unsigned long sync_rcu_preempt_exp_count;
688 689 690 691 692 693 694 695 696 697
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)
{
698
	return rnp->exp_tasks != NULL;
699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723
}

/*
 * 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!)
 *
724 725 726
 * Most callers will set the "wake" flag, but the task initiating the
 * expedited grace period need not wake itself.
 *
727 728
 * Caller must hold sync_rcu_preempt_exp_mutex.
 */
729 730
static void rcu_report_exp_rnp(struct rcu_state *rsp, struct rcu_node *rnp,
			       bool wake)
731 732 733 734
{
	unsigned long flags;
	unsigned long mask;

P
Paul E. McKenney 已提交
735
	raw_spin_lock_irqsave(&rnp->lock, flags);
736
	for (;;) {
737 738
		if (!sync_rcu_preempt_exp_done(rnp)) {
			raw_spin_unlock_irqrestore(&rnp->lock, flags);
739
			break;
740
		}
741
		if (rnp->parent == NULL) {
742
			raw_spin_unlock_irqrestore(&rnp->lock, flags);
743 744
			if (wake)
				wake_up(&sync_rcu_preempt_exp_wq);
745 746 747
			break;
		}
		mask = rnp->grpmask;
P
Paul E. McKenney 已提交
748
		raw_spin_unlock(&rnp->lock); /* irqs remain disabled */
749
		rnp = rnp->parent;
P
Paul E. McKenney 已提交
750
		raw_spin_lock(&rnp->lock); /* irqs already disabled */
751 752 753 754 755 756 757 758 759
		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.
 *
760 761
 * Caller must hold sync_rcu_preempt_exp_mutex and must exclude
 * CPU hotplug operations.
762 763 764 765
 */
static void
sync_rcu_preempt_exp_init(struct rcu_state *rsp, struct rcu_node *rnp)
{
766
	unsigned long flags;
767
	int must_wait = 0;
768

769
	raw_spin_lock_irqsave(&rnp->lock, flags);
770
	if (list_empty(&rnp->blkd_tasks)) {
771
		raw_spin_unlock_irqrestore(&rnp->lock, flags);
772
	} else {
773
		rnp->exp_tasks = rnp->blkd_tasks.next;
774
		rcu_initiate_boost(rnp, flags);  /* releases rnp->lock */
775 776
		must_wait = 1;
	}
777
	if (!must_wait)
778
		rcu_report_exp_rnp(rsp, rnp, false); /* Don't wake self. */
779 780
}

781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796
/**
 * synchronize_rcu_expedited - Brute-force RCU grace period
 *
 * 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
 * the ->blkd_tasks lists and wait for this list to drain.  This consumes
 * significant time on all CPUs and is unfriendly to real-time workloads,
 * so is thus not recommended for any sort of common-case code.
 * In fact, if you are using synchronize_rcu_expedited() in a loop,
 * please restructure your code to batch your updates, and then Use a
 * single synchronize_rcu() instead.
 *
 * Note that it is illegal to call this function while holding any lock
 * that is acquired by a CPU-hotplug notifier.  And yes, it is also illegal
 * to call this function from a CPU-hotplug notifier.  Failing to observe
 * these restriction will result in deadlock.
797 798 799
 */
void synchronize_rcu_expedited(void)
{
800 801 802
	unsigned long flags;
	struct rcu_node *rnp;
	struct rcu_state *rsp = &rcu_preempt_state;
803
	unsigned long snap;
804 805 806 807 808 809
	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. */

810 811 812 813 814 815 816 817 818 819
	/*
	 * Block CPU-hotplug operations.  This means that any CPU-hotplug
	 * operation that finds an rcu_node structure with tasks in the
	 * process of being boosted will know that all tasks blocking
	 * this expedited grace period will already be in the process of
	 * being boosted.  This simplifies the process of moving tasks
	 * from leaf to root rcu_node structures.
	 */
	get_online_cpus();

820 821 822 823 824 825
	/*
	 * 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)) {
826 827 828 829 830
		if (ULONG_CMP_LT(snap,
		    ACCESS_ONCE(sync_rcu_preempt_exp_count))) {
			put_online_cpus();
			goto mb_ret; /* Others did our work for us. */
		}
831
		if (trycount++ < 10) {
832
			udelay(trycount * num_online_cpus());
833
		} else {
834
			put_online_cpus();
835 836 837 838
			synchronize_rcu();
			return;
		}
	}
839 840
	if (ULONG_CMP_LT(snap, ACCESS_ONCE(sync_rcu_preempt_exp_count))) {
		put_online_cpus();
841
		goto unlock_mb_ret; /* Others did our work for us. */
842
	}
843

844
	/* force all RCU readers onto ->blkd_tasks lists. */
845 846 847 848
	synchronize_sched_expedited();

	/* Initialize ->expmask for all non-leaf rcu_node structures. */
	rcu_for_each_nonleaf_node_breadth_first(rsp, rnp) {
849
		raw_spin_lock_irqsave(&rnp->lock, flags);
850
		rnp->expmask = rnp->qsmaskinit;
851
		raw_spin_unlock_irqrestore(&rnp->lock, flags);
852 853
	}

854
	/* Snapshot current state of ->blkd_tasks lists. */
855 856 857 858 859
	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));

860
	put_online_cpus();
861

862
	/* Wait for snapshotted ->blkd_tasks lists to drain. */
863 864 865 866 867 868 869 870 871 872 873
	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. */
874 875 876
}
EXPORT_SYMBOL_GPL(synchronize_rcu_expedited);

877 878 879 880 881
/**
 * rcu_barrier - Wait until all in-flight call_rcu() callbacks complete.
 */
void rcu_barrier(void)
{
882
	_rcu_barrier(&rcu_preempt_state);
883 884 885
}
EXPORT_SYMBOL_GPL(rcu_barrier);

886
/*
P
Paul E. McKenney 已提交
887
 * Initialize preemptible RCU's state structures.
888 889 890
 */
static void __init __rcu_init_preempt(void)
{
891
	rcu_init_one(&rcu_preempt_state, &rcu_preempt_data);
892 893
}

894 895
#else /* #ifdef CONFIG_TREE_PREEMPT_RCU */

896 897
static struct rcu_state *rcu_state = &rcu_sched_state;

898 899 900
/*
 * Tell them what RCU they are running.
 */
901
static void __init rcu_bootup_announce(void)
902 903
{
	printk(KERN_INFO "Hierarchical RCU implementation.\n");
904
	rcu_bootup_announce_oddness();
905 906 907 908 909 910 911 912 913 914 915
}

/*
 * 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);

916 917 918 919 920 921 922 923 924 925
/*
 * 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);

926 927 928 929 930 931 932 933
/*
 * Because preemptible RCU does not exist, we never have to check for
 * CPUs being in quiescent states.
 */
static void rcu_preempt_note_context_switch(int cpu)
{
}

934
/*
P
Paul E. McKenney 已提交
935
 * Because preemptible RCU does not exist, there are never any preempted
936 937
 * RCU readers.
 */
938
static int rcu_preempt_blocked_readers_cgp(struct rcu_node *rnp)
939 940 941 942
{
	return 0;
}

943 944 945
#ifdef CONFIG_HOTPLUG_CPU

/* Because preemptible RCU does not exist, no quieting of tasks. */
P
Paul E. McKenney 已提交
946
static void rcu_report_unblock_qs_rnp(struct rcu_node *rnp, unsigned long flags)
947
{
P
Paul E. McKenney 已提交
948
	raw_spin_unlock_irqrestore(&rnp->lock, flags);
949 950 951 952
}

#endif /* #ifdef CONFIG_HOTPLUG_CPU */

953
/*
P
Paul E. McKenney 已提交
954
 * Because preemptible RCU does not exist, we never have to check for
955 956 957 958 959 960
 * tasks blocked within RCU read-side critical sections.
 */
static void rcu_print_detail_task_stall(struct rcu_state *rsp)
{
}

961
/*
P
Paul E. McKenney 已提交
962
 * Because preemptible RCU does not exist, we never have to check for
963 964
 * tasks blocked within RCU read-side critical sections.
 */
965
static int rcu_print_task_stall(struct rcu_node *rnp)
966
{
967
	return 0;
968 969
}

970
/*
P
Paul E. McKenney 已提交
971
 * Because there is no preemptible RCU, there can be no readers blocked,
972 973
 * so there is no need to check for blocked tasks.  So check only for
 * bogus qsmask values.
974 975 976
 */
static void rcu_preempt_check_blocked_tasks(struct rcu_node *rnp)
{
977
	WARN_ON_ONCE(rnp->qsmask);
978 979
}

980 981
#ifdef CONFIG_HOTPLUG_CPU

982
/*
P
Paul E. McKenney 已提交
983
 * Because preemptible RCU does not exist, it never needs to migrate
984 985 986
 * tasks that were blocked within RCU read-side critical sections, and
 * such non-existent tasks cannot possibly have been blocking the current
 * grace period.
987
 */
988 989 990
static int rcu_preempt_offline_tasks(struct rcu_state *rsp,
				     struct rcu_node *rnp,
				     struct rcu_data *rdp)
991
{
992
	return 0;
993 994
}

995 996
#endif /* #ifdef CONFIG_HOTPLUG_CPU */

997
/*
P
Paul E. McKenney 已提交
998
 * Because preemptible RCU does not exist, it never has any callbacks
999 1000
 * to check.
 */
1001
static void rcu_preempt_check_callbacks(int cpu)
1002 1003 1004
{
}

1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020
/*
 * Queue an RCU callback for lazy invocation after a grace period.
 * This will likely be later named something like "call_rcu_lazy()",
 * but this change will require some way of tagging the lazy RCU
 * callbacks in the list of pending callbacks.  Until then, this
 * function may only be called from __kfree_rcu().
 *
 * Because there is no preemptible RCU, we use RCU-sched instead.
 */
void kfree_call_rcu(struct rcu_head *head,
		    void (*func)(struct rcu_head *rcu))
{
	__call_rcu(head, func, &rcu_sched_state, 1);
}
EXPORT_SYMBOL_GPL(kfree_call_rcu);

1021 1022
/*
 * Wait for an rcu-preempt grace period, but make it happen quickly.
P
Paul E. McKenney 已提交
1023
 * But because preemptible RCU does not exist, map to rcu-sched.
1024 1025 1026 1027 1028 1029 1030
 */
void synchronize_rcu_expedited(void)
{
	synchronize_sched_expedited();
}
EXPORT_SYMBOL_GPL(synchronize_rcu_expedited);

1031 1032 1033
#ifdef CONFIG_HOTPLUG_CPU

/*
P
Paul E. McKenney 已提交
1034
 * Because preemptible RCU does not exist, there is never any need to
1035 1036 1037
 * report on tasks preempted in RCU read-side critical sections during
 * expedited RCU grace periods.
 */
1038 1039
static void rcu_report_exp_rnp(struct rcu_state *rsp, struct rcu_node *rnp,
			       bool wake)
1040 1041 1042 1043 1044
{
}

#endif /* #ifdef CONFIG_HOTPLUG_CPU */

1045
/*
P
Paul E. McKenney 已提交
1046
 * Because preemptible RCU does not exist, rcu_barrier() is just
1047 1048 1049 1050 1051 1052 1053 1054
 * another name for rcu_barrier_sched().
 */
void rcu_barrier(void)
{
	rcu_barrier_sched();
}
EXPORT_SYMBOL_GPL(rcu_barrier);

1055
/*
P
Paul E. McKenney 已提交
1056
 * Because preemptible RCU does not exist, it need not be initialized.
1057 1058 1059 1060 1061
 */
static void __init __rcu_init_preempt(void)
{
}

1062
#endif /* #else #ifdef CONFIG_TREE_PREEMPT_RCU */
1063

1064 1065 1066 1067
#ifdef CONFIG_RCU_BOOST

#include "rtmutex_common.h"

1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080
#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 &&
1081
		 ULONG_CMP_LT(jiffies, rnp->boost_time))
1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094
		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 */

T
Thomas Gleixner 已提交
1095 1096 1097 1098 1099 1100 1101 1102 1103 1104
static void rcu_wake_cond(struct task_struct *t, int status)
{
	/*
	 * If the thread is yielding, only wake it when this
	 * is invoked from idle
	 */
	if (status != RCU_KTHREAD_YIELDING || is_idle_task(current))
		wake_up_process(t);
}

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 1138 1139
/*
 * 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.
	 */
1140
	if (rnp->exp_tasks != NULL) {
1141
		tb = rnp->exp_tasks;
1142 1143
		rnp->n_exp_boosts++;
	} else {
1144
		tb = rnp->boost_tasks;
1145 1146 1147
		rnp->n_normal_boosts++;
	}
	rnp->n_tasks_boosted++;
1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171

	/*
	 * 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;
	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. */

1172 1173
	return ACCESS_ONCE(rnp->exp_tasks) != NULL ||
	       ACCESS_ONCE(rnp->boost_tasks) != NULL;
1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185
}

/*
 * 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;

1186
	trace_rcu_utilization("Start boost kthread@init");
1187
	for (;;) {
1188
		rnp->boost_kthread_status = RCU_KTHREAD_WAITING;
1189
		trace_rcu_utilization("End boost kthread@rcu_wait");
1190
		rcu_wait(rnp->boost_tasks || rnp->exp_tasks);
1191
		trace_rcu_utilization("Start boost kthread@rcu_wait");
1192
		rnp->boost_kthread_status = RCU_KTHREAD_RUNNING;
1193 1194 1195 1196 1197 1198
		more2boost = rcu_boost(rnp);
		if (more2boost)
			spincnt++;
		else
			spincnt = 0;
		if (spincnt > 10) {
T
Thomas Gleixner 已提交
1199
			rnp->boost_kthread_status = RCU_KTHREAD_YIELDING;
1200
			trace_rcu_utilization("End boost kthread@rcu_yield");
T
Thomas Gleixner 已提交
1201
			schedule_timeout_interruptible(2);
1202
			trace_rcu_utilization("Start boost kthread@rcu_yield");
1203 1204 1205
			spincnt = 0;
		}
	}
1206
	/* NOTREACHED */
1207
	trace_rcu_utilization("End boost kthread@notreached");
1208 1209 1210 1211 1212 1213 1214 1215 1216
	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.
 *
1217 1218 1219
 * The caller must hold rnp->lock, which this function releases.
 * The ->boost_kthread_task is immortal, so we don't need to worry
 * about it going away.
1220
 */
1221
static void rcu_initiate_boost(struct rcu_node *rnp, unsigned long flags)
1222 1223 1224
{
	struct task_struct *t;

1225 1226
	if (!rcu_preempt_blocked_readers_cgp(rnp) && rnp->exp_tasks == NULL) {
		rnp->n_balk_exp_gp_tasks++;
1227
		raw_spin_unlock_irqrestore(&rnp->lock, flags);
1228
		return;
1229
	}
1230 1231 1232 1233 1234 1235 1236
	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;
1237
		raw_spin_unlock_irqrestore(&rnp->lock, flags);
1238
		t = rnp->boost_kthread_task;
T
Thomas Gleixner 已提交
1239 1240
		if (t)
			rcu_wake_cond(t, rnp->boost_kthread_status);
1241
	} else {
1242
		rcu_initiate_boost_trace(rnp);
1243 1244
		raw_spin_unlock_irqrestore(&rnp->lock, flags);
	}
1245 1246
}

1247 1248 1249 1250 1251 1252 1253 1254 1255
/*
 * 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);
1256
	if (__this_cpu_read(rcu_cpu_kthread_task) != NULL &&
T
Thomas Gleixner 已提交
1257 1258 1259 1260
	    current != __this_cpu_read(rcu_cpu_kthread_task)) {
		rcu_wake_cond(__this_cpu_read(rcu_cpu_kthread_task),
			      __this_cpu_read(rcu_cpu_kthread_status));
	}
1261 1262 1263
	local_irq_restore(flags);
}

1264 1265 1266 1267 1268 1269 1270 1271 1272
/*
 * Is the current CPU running the RCU-callbacks kthread?
 * Caller must have preemption disabled.
 */
static bool rcu_is_callbacks_kthread(void)
{
	return __get_cpu_var(rcu_cpu_kthread_task) == current;
}

1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288
#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,
T
Thomas Gleixner 已提交
1289
						 struct rcu_node *rnp)
1290
{
T
Thomas Gleixner 已提交
1291
	int rnp_index = rnp - &rsp->node[0];
1292 1293 1294 1295 1296 1297
	unsigned long flags;
	struct sched_param sp;
	struct task_struct *t;

	if (&rcu_preempt_state != rsp)
		return 0;
T
Thomas Gleixner 已提交
1298 1299 1300 1301

	if (!rcu_scheduler_fully_active || rnp->qsmaskinit == 0)
		return 0;

1302
	rsp->boost = 1;
1303 1304 1305
	if (rnp->boost_kthread_task != NULL)
		return 0;
	t = kthread_create(rcu_boost_kthread, (void *)rnp,
1306
			   "rcub/%d", rnp_index);
1307 1308 1309 1310 1311
	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);
1312
	sp.sched_priority = RCU_BOOST_PRIO;
1313
	sched_setscheduler_nocheck(t, SCHED_FIFO, &sp);
1314
	wake_up_process(t); /* get to TASK_INTERRUPTIBLE quickly. */
1315 1316 1317
	return 0;
}

1318 1319 1320 1321 1322 1323 1324
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();
}

1325
static void rcu_cpu_kthread_setup(unsigned int cpu)
1326 1327 1328
{
	struct sched_param sp;

1329 1330
	sp.sched_priority = RCU_KTHREAD_PRIO;
	sched_setscheduler_nocheck(current, SCHED_FIFO, &sp);
1331 1332
}

1333
static void rcu_cpu_kthread_park(unsigned int cpu)
1334
{
1335
	per_cpu(rcu_cpu_kthread_status, cpu) = RCU_KTHREAD_OFFCPU;
1336 1337
}

1338
static int rcu_cpu_kthread_should_run(unsigned int cpu)
1339
{
1340
	return __get_cpu_var(rcu_cpu_has_work);
1341 1342 1343 1344
}

/*
 * Per-CPU kernel thread that invokes RCU callbacks.  This replaces the
1345 1346
 * RCU softirq used in flavors and configurations of RCU that do not
 * support RCU priority boosting.
1347
 */
1348
static void rcu_cpu_kthread(unsigned int cpu)
1349
{
1350 1351 1352
	unsigned int *statusp = &__get_cpu_var(rcu_cpu_kthread_status);
	char work, *workp = &__get_cpu_var(rcu_cpu_has_work);
	int spincnt;
1353

1354
	for (spincnt = 0; spincnt < 10; spincnt++) {
1355
		trace_rcu_utilization("Start CPU kthread@rcu_wait");
1356 1357
		local_bh_disable();
		*statusp = RCU_KTHREAD_RUNNING;
1358 1359
		this_cpu_inc(rcu_cpu_kthread_loops);
		local_irq_disable();
1360 1361
		work = *workp;
		*workp = 0;
1362
		local_irq_enable();
1363 1364 1365
		if (work)
			rcu_kthread_do_work();
		local_bh_enable();
1366 1367 1368 1369
		if (*workp == 0) {
			trace_rcu_utilization("End CPU kthread@rcu_wait");
			*statusp = RCU_KTHREAD_WAITING;
			return;
1370 1371
		}
	}
1372 1373 1374 1375 1376
	*statusp = RCU_KTHREAD_YIELDING;
	trace_rcu_utilization("Start CPU kthread@rcu_yield");
	schedule_timeout_interruptible(2);
	trace_rcu_utilization("End CPU kthread@rcu_yield");
	*statusp = RCU_KTHREAD_WAITING;
1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387
}

/*
 * 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.
 */
T
Thomas Gleixner 已提交
1388
static void rcu_boost_kthread_setaffinity(struct rcu_node *rnp, int outgoingcpu)
1389
{
T
Thomas Gleixner 已提交
1390 1391
	struct task_struct *t = rnp->boost_kthread_task;
	unsigned long mask = rnp->qsmaskinit;
1392 1393 1394
	cpumask_var_t cm;
	int cpu;

T
Thomas Gleixner 已提交
1395
	if (!t)
1396
		return;
T
Thomas Gleixner 已提交
1397
	if (!zalloc_cpumask_var(&cm, GFP_KERNEL))
1398 1399 1400 1401 1402 1403 1404 1405 1406 1407
		return;
	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);
	}
T
Thomas Gleixner 已提交
1408
	set_cpus_allowed_ptr(t, cm);
1409 1410 1411
	free_cpumask_var(cm);
}

1412 1413 1414 1415 1416 1417 1418 1419
static struct smp_hotplug_thread rcu_cpu_thread_spec = {
	.store			= &rcu_cpu_kthread_task,
	.thread_should_run	= rcu_cpu_kthread_should_run,
	.thread_fn		= rcu_cpu_kthread,
	.thread_comm		= "rcuc/%u",
	.setup			= rcu_cpu_kthread_setup,
	.park			= rcu_cpu_kthread_park,
};
1420 1421 1422 1423 1424 1425 1426

/*
 * Spawn all kthreads -- called as soon as the scheduler is running.
 */
static int __init rcu_spawn_kthreads(void)
{
	struct rcu_node *rnp;
T
Thomas Gleixner 已提交
1427
	int cpu;
1428

1429
	rcu_scheduler_fully_active = 1;
1430
	for_each_possible_cpu(cpu)
1431
		per_cpu(rcu_cpu_has_work, cpu) = 0;
1432
	BUG_ON(smpboot_register_percpu_thread(&rcu_cpu_thread_spec));
1433
	rnp = rcu_get_root(rcu_state);
T
Thomas Gleixner 已提交
1434
	(void)rcu_spawn_one_boost_kthread(rcu_state, rnp);
1435 1436
	if (NUM_RCU_NODES > 1) {
		rcu_for_each_leaf_node(rcu_state, rnp)
T
Thomas Gleixner 已提交
1437
			(void)rcu_spawn_one_boost_kthread(rcu_state, rnp);
1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448
	}
	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. */
1449
	if (rcu_scheduler_fully_active)
T
Thomas Gleixner 已提交
1450
		(void)rcu_spawn_one_boost_kthread(rcu_state, rnp);
1451 1452
}

1453 1454
#else /* #ifdef CONFIG_RCU_BOOST */

1455
static void rcu_initiate_boost(struct rcu_node *rnp, unsigned long flags)
1456
{
1457
	raw_spin_unlock_irqrestore(&rnp->lock, flags);
1458 1459
}

1460
static void invoke_rcu_callbacks_kthread(void)
1461
{
1462
	WARN_ON_ONCE(1);
1463 1464
}

1465 1466 1467 1468 1469
static bool rcu_is_callbacks_kthread(void)
{
	return false;
}

1470 1471 1472 1473
static void rcu_preempt_boost_start_gp(struct rcu_node *rnp)
{
}

T
Thomas Gleixner 已提交
1474
static void rcu_boost_kthread_setaffinity(struct rcu_node *rnp, int outgoingcpu)
1475 1476 1477
{
}

1478 1479 1480 1481 1482 1483 1484
static int __init rcu_scheduler_really_started(void)
{
	rcu_scheduler_fully_active = 1;
	return 0;
}
early_initcall(rcu_scheduler_really_started);

1485 1486 1487 1488
static void __cpuinit rcu_prepare_kthreads(int cpu)
{
}

1489 1490
#endif /* #else #ifdef CONFIG_RCU_BOOST */

1491 1492 1493 1494 1495 1496 1497 1498
#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.
 *
1499 1500
 * Because we not have RCU_FAST_NO_HZ, just check whether this CPU needs
 * any flavor of RCU.
1501
 */
1502
int rcu_needs_cpu(int cpu, unsigned long *delta_jiffies)
1503
{
1504
	*delta_jiffies = ULONG_MAX;
1505 1506 1507
	return rcu_cpu_has_callbacks(cpu);
}

1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522
/*
 * Because we do not have RCU_FAST_NO_HZ, don't bother initializing for it.
 */
static void rcu_prepare_for_idle_init(int cpu)
{
}

/*
 * Because we do not have RCU_FAST_NO_HZ, don't bother cleaning up
 * after it.
 */
static void rcu_cleanup_after_idle(int cpu)
{
}

1523
/*
1524
 * Do the idle-entry grace-period work, which, because CONFIG_RCU_FAST_NO_HZ=n,
1525 1526 1527 1528 1529 1530
 * is nothing.
 */
static void rcu_prepare_for_idle(int cpu)
{
}

1531 1532 1533 1534 1535 1536 1537 1538
/*
 * Don't bother keeping a running count of the number of RCU callbacks
 * posted because CONFIG_RCU_FAST_NO_HZ=n.
 */
static void rcu_idle_count_callbacks_posted(void)
{
}

1539 1540
#else /* #if !defined(CONFIG_RCU_FAST_NO_HZ) */

1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565
/*
 * This code is invoked when a CPU goes idle, at which point we want
 * to have the CPU do everything required for RCU so that it can enter
 * the energy-efficient dyntick-idle mode.  This is handled by a
 * state machine implemented by rcu_prepare_for_idle() below.
 *
 * The following three proprocessor symbols control this state machine:
 *
 * RCU_IDLE_FLUSHES gives the maximum number of times that we will attempt
 *	to satisfy RCU.  Beyond this point, it is better to incur a periodic
 *	scheduling-clock interrupt than to loop through the state machine
 *	at full power.
 * RCU_IDLE_OPT_FLUSHES gives the number of RCU_IDLE_FLUSHES that are
 *	optional if RCU does not need anything immediately from this
 *	CPU, even if this CPU still has RCU callbacks queued.  The first
 *	times through the state machine are mandatory: we need to give
 *	the state machine a chance to communicate a quiescent state
 *	to the RCU core.
 * RCU_IDLE_GP_DELAY gives the number of jiffies that a CPU is permitted
 *	to sleep in dyntick-idle mode with RCU callbacks pending.  This
 *	is sized to be roughly one RCU grace period.  Those energy-efficiency
 *	benchmarkers who might otherwise be tempted to set this to a large
 *	number, be warned: Setting RCU_IDLE_GP_DELAY too high can hang your
 *	system.  And if you are -that- concerned about energy efficiency,
 *	just power the system down and be done with it!
1566 1567 1568
 * RCU_IDLE_LAZY_GP_DELAY gives the number of jiffies that a CPU is
 *	permitted to sleep in dyntick-idle mode with only lazy RCU
 *	callbacks pending.  Setting this too high can OOM your system.
1569 1570 1571 1572 1573 1574 1575
 *
 * The values below work well in practice.  If future workloads require
 * adjustment, they can be converted into kernel config parameters, though
 * making the state machine smarter might be a better option.
 */
#define RCU_IDLE_FLUSHES 5		/* Number of dyntick-idle tries. */
#define RCU_IDLE_OPT_FLUSHES 3		/* Optional dyntick-idle tries. */
1576
#define RCU_IDLE_GP_DELAY 4		/* Roughly one grace period. */
1577
#define RCU_IDLE_LAZY_GP_DELAY (6 * HZ)	/* Roughly six seconds. */
1578

1579 1580
extern int tick_nohz_enabled;

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
/*
 * Does the specified flavor of RCU have non-lazy callbacks pending on
 * the specified CPU?  Both RCU flavor and CPU are specified by the
 * rcu_data structure.
 */
static bool __rcu_cpu_has_nonlazy_callbacks(struct rcu_data *rdp)
{
	return rdp->qlen != rdp->qlen_lazy;
}

#ifdef CONFIG_TREE_PREEMPT_RCU

/*
 * Are there non-lazy RCU-preempt callbacks?  (There cannot be if there
 * is no RCU-preempt in the kernel.)
 */
static bool rcu_preempt_cpu_has_nonlazy_callbacks(int cpu)
{
	struct rcu_data *rdp = &per_cpu(rcu_preempt_data, cpu);

	return __rcu_cpu_has_nonlazy_callbacks(rdp);
}

#else /* #ifdef CONFIG_TREE_PREEMPT_RCU */

static bool rcu_preempt_cpu_has_nonlazy_callbacks(int cpu)
{
	return 0;
}

#endif /* else #ifdef CONFIG_TREE_PREEMPT_RCU */

/*
 * Does any flavor of RCU have non-lazy callbacks on the specified CPU?
 */
static bool rcu_cpu_has_nonlazy_callbacks(int cpu)
{
	return __rcu_cpu_has_nonlazy_callbacks(&per_cpu(rcu_sched_data, cpu)) ||
	       __rcu_cpu_has_nonlazy_callbacks(&per_cpu(rcu_bh_data, cpu)) ||
	       rcu_preempt_cpu_has_nonlazy_callbacks(cpu);
}

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
/*
 * Allow the CPU to enter dyntick-idle mode if either: (1) There are no
 * callbacks on this CPU, (2) this CPU has not yet attempted to enter
 * dyntick-idle mode, or (3) this CPU is in the process of attempting to
 * enter dyntick-idle mode.  Otherwise, if we have recently tried and failed
 * to enter dyntick-idle mode, we refuse to try to enter it.  After all,
 * it is better to incur scheduling-clock interrupts than to spin
 * continuously for the same time duration!
 *
 * The delta_jiffies argument is used to store the time when RCU is
 * going to need the CPU again if it still has callbacks.  The reason
 * for this is that rcu_prepare_for_idle() might need to post a timer,
 * but if so, it will do so after tick_nohz_stop_sched_tick() has set
 * the wakeup time for this CPU.  This means that RCU's timer can be
 * delayed until the wakeup time, which defeats the purpose of posting
 * a timer.
 */
int rcu_needs_cpu(int cpu, unsigned long *delta_jiffies)
{
	struct rcu_dynticks *rdtp = &per_cpu(rcu_dynticks, cpu);

	/* Flag a new idle sojourn to the idle-entry state machine. */
	rdtp->idle_first_pass = 1;
	/* If no callbacks, RCU doesn't need the CPU. */
	if (!rcu_cpu_has_callbacks(cpu)) {
		*delta_jiffies = ULONG_MAX;
		return 0;
	}
	if (rdtp->dyntick_holdoff == jiffies) {
		/* RCU recently tried and failed, so don't try again. */
		*delta_jiffies = 1;
		return 1;
	}
	/* Set up for the possibility that RCU will post a timer. */
1657 1658 1659 1660 1661 1662 1663
	if (rcu_cpu_has_nonlazy_callbacks(cpu)) {
		*delta_jiffies = round_up(RCU_IDLE_GP_DELAY + jiffies,
					  RCU_IDLE_GP_DELAY) - jiffies;
	} else {
		*delta_jiffies = jiffies + RCU_IDLE_LAZY_GP_DELAY;
		*delta_jiffies = round_jiffies(*delta_jiffies) - jiffies;
	}
1664 1665 1666
	return 0;
}

1667 1668 1669 1670 1671 1672 1673 1674 1675
/*
 * Handler for smp_call_function_single().  The only point of this
 * handler is to wake the CPU up, so the handler does only tracing.
 */
void rcu_idle_demigrate(void *unused)
{
	trace_rcu_prep_idle("Demigrate");
}

1676 1677 1678 1679 1680 1681
/*
 * Timer handler used to force CPU to start pushing its remaining RCU
 * callbacks in the case where it entered dyntick-idle mode with callbacks
 * pending.  The hander doesn't really need to do anything because the
 * real work is done upon re-entry to idle, or by the next scheduling-clock
 * interrupt should idle not be re-entered.
1682 1683 1684 1685
 *
 * One special case: the timer gets migrated without awakening the CPU
 * on which the timer was scheduled on.  In this case, we must wake up
 * that CPU.  We do so with smp_call_function_single().
1686
 */
1687
static void rcu_idle_gp_timer_func(unsigned long cpu_in)
1688
{
1689 1690
	int cpu = (int)cpu_in;

1691
	trace_rcu_prep_idle("Timer");
1692 1693 1694 1695
	if (cpu != smp_processor_id())
		smp_call_function_single(cpu, rcu_idle_demigrate, NULL, 0);
	else
		WARN_ON_ONCE(1); /* Getting here can hang the system... */
1696 1697 1698 1699 1700 1701 1702
}

/*
 * Initialize the timer used to pull CPUs out of dyntick-idle mode.
 */
static void rcu_prepare_for_idle_init(int cpu)
{
1703 1704 1705 1706 1707 1708
	struct rcu_dynticks *rdtp = &per_cpu(rcu_dynticks, cpu);

	rdtp->dyntick_holdoff = jiffies - 1;
	setup_timer(&rdtp->idle_gp_timer, rcu_idle_gp_timer_func, cpu);
	rdtp->idle_gp_timer_expires = jiffies - 1;
	rdtp->idle_first_pass = 1;
1709 1710 1711 1712
}

/*
 * Clean up for exit from idle.  Because we are exiting from idle, there
1713
 * is no longer any point to ->idle_gp_timer, so cancel it.  This will
1714 1715 1716 1717
 * do nothing if this timer is not active, so just cancel it unconditionally.
 */
static void rcu_cleanup_after_idle(int cpu)
{
1718 1719 1720
	struct rcu_dynticks *rdtp = &per_cpu(rcu_dynticks, cpu);

	del_timer(&rdtp->idle_gp_timer);
1721
	trace_rcu_prep_idle("Cleanup after idle");
1722
	rdtp->tick_nohz_enabled_snap = ACCESS_ONCE(tick_nohz_enabled);
1723 1724
}

1725 1726 1727 1728
/*
 * Check to see if any RCU-related work can be done by the current CPU,
 * and if so, schedule a softirq to get it done.  This function is part
 * of the RCU implementation; it is -not- an exported member of the RCU API.
1729
 *
1730 1731 1732 1733 1734 1735
 * The idea is for the current CPU to clear out all work required by the
 * RCU core for the current grace period, so that this CPU can be permitted
 * to enter dyntick-idle mode.  In some cases, it will need to be awakened
 * at the end of the grace period by whatever CPU ends the grace period.
 * This allows CPUs to go dyntick-idle more quickly, and to reduce the
 * number of wakeups by a modest integer factor.
1736 1737 1738
 *
 * Because it is not legal to invoke rcu_process_callbacks() with irqs
 * disabled, we do one pass of force_quiescent_state(), then do a
1739
 * invoke_rcu_core() to cause rcu_process_callbacks() to be invoked
1740
 * later.  The ->dyntick_drain field controls the sequencing.
1741 1742
 *
 * The caller must have disabled interrupts.
1743
 */
1744
static void rcu_prepare_for_idle(int cpu)
1745
{
1746
	struct timer_list *tp;
1747
	struct rcu_dynticks *rdtp = &per_cpu(rcu_dynticks, cpu);
1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759
	int tne;

	/* Handle nohz enablement switches conservatively. */
	tne = ACCESS_ONCE(tick_nohz_enabled);
	if (tne != rdtp->tick_nohz_enabled_snap) {
		if (rcu_cpu_has_callbacks(cpu))
			invoke_rcu_core(); /* force nohz to see update. */
		rdtp->tick_nohz_enabled_snap = tne;
		return;
	}
	if (!tne)
		return;
1760

1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780
	/* Adaptive-tick mode, where usermode execution is idle to RCU. */
	if (!is_idle_task(current)) {
		rdtp->dyntick_holdoff = jiffies - 1;
		if (rcu_cpu_has_nonlazy_callbacks(cpu)) {
			trace_rcu_prep_idle("User dyntick with callbacks");
			rdtp->idle_gp_timer_expires =
				round_up(jiffies + RCU_IDLE_GP_DELAY,
					 RCU_IDLE_GP_DELAY);
		} else if (rcu_cpu_has_callbacks(cpu)) {
			rdtp->idle_gp_timer_expires =
				round_jiffies(jiffies + RCU_IDLE_LAZY_GP_DELAY);
			trace_rcu_prep_idle("User dyntick with lazy callbacks");
		} else {
			return;
		}
		tp = &rdtp->idle_gp_timer;
		mod_timer_pinned(tp, rdtp->idle_gp_timer_expires);
		return;
	}

1781 1782 1783 1784 1785
	/*
	 * If this is an idle re-entry, for example, due to use of
	 * RCU_NONIDLE() or the new idle-loop tracing API within the idle
	 * loop, then don't take any state-machine actions, unless the
	 * momentary exit from idle queued additional non-lazy callbacks.
1786
	 * Instead, repost the ->idle_gp_timer if this CPU has callbacks
1787 1788
	 * pending.
	 */
1789 1790
	if (!rdtp->idle_first_pass &&
	    (rdtp->nonlazy_posted == rdtp->nonlazy_posted_snap)) {
1791
		if (rcu_cpu_has_callbacks(cpu)) {
1792 1793
			tp = &rdtp->idle_gp_timer;
			mod_timer_pinned(tp, rdtp->idle_gp_timer_expires);
1794
		}
1795 1796
		return;
	}
1797 1798
	rdtp->idle_first_pass = 0;
	rdtp->nonlazy_posted_snap = rdtp->nonlazy_posted - 1;
1799

1800
	/*
1801 1802
	 * If there are no callbacks on this CPU, enter dyntick-idle mode.
	 * Also reset state to avoid prejudicing later attempts.
1803
	 */
1804
	if (!rcu_cpu_has_callbacks(cpu)) {
1805 1806
		rdtp->dyntick_holdoff = jiffies - 1;
		rdtp->dyntick_drain = 0;
1807
		trace_rcu_prep_idle("No callbacks");
1808
		return;
1809
	}
1810 1811 1812 1813 1814

	/*
	 * If in holdoff mode, just return.  We will presumably have
	 * refrained from disabling the scheduling-clock tick.
	 */
1815
	if (rdtp->dyntick_holdoff == jiffies) {
1816
		trace_rcu_prep_idle("In holdoff");
1817
		return;
1818
	}
1819

1820 1821
	/* Check and update the ->dyntick_drain sequencing. */
	if (rdtp->dyntick_drain <= 0) {
1822
		/* First time through, initialize the counter. */
1823 1824
		rdtp->dyntick_drain = RCU_IDLE_FLUSHES;
	} else if (rdtp->dyntick_drain <= RCU_IDLE_OPT_FLUSHES &&
1825 1826
		   !rcu_pending(cpu) &&
		   !local_softirq_pending()) {
1827
		/* Can we go dyntick-idle despite still having callbacks? */
1828 1829
		rdtp->dyntick_drain = 0;
		rdtp->dyntick_holdoff = jiffies;
1830 1831
		if (rcu_cpu_has_nonlazy_callbacks(cpu)) {
			trace_rcu_prep_idle("Dyntick with callbacks");
1832
			rdtp->idle_gp_timer_expires =
1833 1834
				round_up(jiffies + RCU_IDLE_GP_DELAY,
					 RCU_IDLE_GP_DELAY);
1835
		} else {
1836
			rdtp->idle_gp_timer_expires =
1837
				round_jiffies(jiffies + RCU_IDLE_LAZY_GP_DELAY);
1838 1839
			trace_rcu_prep_idle("Dyntick with lazy callbacks");
		}
1840 1841 1842
		tp = &rdtp->idle_gp_timer;
		mod_timer_pinned(tp, rdtp->idle_gp_timer_expires);
		rdtp->nonlazy_posted_snap = rdtp->nonlazy_posted;
1843
		return; /* Nothing more to do immediately. */
1844
	} else if (--(rdtp->dyntick_drain) <= 0) {
1845
		/* We have hit the limit, so time to give up. */
1846
		rdtp->dyntick_holdoff = jiffies;
1847
		trace_rcu_prep_idle("Begin holdoff");
1848 1849
		invoke_rcu_core();  /* Force the CPU out of dyntick-idle. */
		return;
1850 1851
	}

1852 1853 1854 1855 1856 1857 1858
	/*
	 * Do one step of pushing the remaining RCU callbacks through
	 * the RCU core state machine.
	 */
#ifdef CONFIG_TREE_PREEMPT_RCU
	if (per_cpu(rcu_preempt_data, cpu).nxtlist) {
		rcu_preempt_qs(cpu);
1859
		force_quiescent_state(&rcu_preempt_state);
1860 1861
	}
#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */
1862 1863
	if (per_cpu(rcu_sched_data, cpu).nxtlist) {
		rcu_sched_qs(cpu);
1864
		force_quiescent_state(&rcu_sched_state);
1865 1866 1867
	}
	if (per_cpu(rcu_bh_data, cpu).nxtlist) {
		rcu_bh_qs(cpu);
1868
		force_quiescent_state(&rcu_bh_state);
1869 1870
	}

1871 1872 1873 1874
	/*
	 * If RCU callbacks are still pending, RCU still needs this CPU.
	 * So try forcing the callbacks through the grace period.
	 */
1875
	if (rcu_cpu_has_callbacks(cpu)) {
1876
		trace_rcu_prep_idle("More callbacks");
1877
		invoke_rcu_core();
1878
	} else {
1879
		trace_rcu_prep_idle("Callbacks drained");
1880
	}
1881 1882
}

1883
/*
1884 1885 1886 1887 1888 1889
 * Keep a running count of the number of non-lazy callbacks posted
 * on this CPU.  This running counter (which is never decremented) allows
 * rcu_prepare_for_idle() to detect when something out of the idle loop
 * posts a callback, even if an equal number of callbacks are invoked.
 * Of course, callbacks should only be posted from within a trace event
 * designed to be called from idle or from within RCU_NONIDLE().
1890 1891 1892
 */
static void rcu_idle_count_callbacks_posted(void)
{
1893
	__this_cpu_add(rcu_dynticks.nonlazy_posted, 1);
1894 1895
}

1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977
/*
 * Data for flushing lazy RCU callbacks at OOM time.
 */
static atomic_t oom_callback_count;
static DECLARE_WAIT_QUEUE_HEAD(oom_callback_wq);

/*
 * RCU OOM callback -- decrement the outstanding count and deliver the
 * wake-up if we are the last one.
 */
static void rcu_oom_callback(struct rcu_head *rhp)
{
	if (atomic_dec_and_test(&oom_callback_count))
		wake_up(&oom_callback_wq);
}

/*
 * Post an rcu_oom_notify callback on the current CPU if it has at
 * least one lazy callback.  This will unnecessarily post callbacks
 * to CPUs that already have a non-lazy callback at the end of their
 * callback list, but this is an infrequent operation, so accept some
 * extra overhead to keep things simple.
 */
static void rcu_oom_notify_cpu(void *unused)
{
	struct rcu_state *rsp;
	struct rcu_data *rdp;

	for_each_rcu_flavor(rsp) {
		rdp = __this_cpu_ptr(rsp->rda);
		if (rdp->qlen_lazy != 0) {
			atomic_inc(&oom_callback_count);
			rsp->call(&rdp->oom_head, rcu_oom_callback);
		}
	}
}

/*
 * If low on memory, ensure that each CPU has a non-lazy callback.
 * This will wake up CPUs that have only lazy callbacks, in turn
 * ensuring that they free up the corresponding memory in a timely manner.
 * Because an uncertain amount of memory will be freed in some uncertain
 * timeframe, we do not claim to have freed anything.
 */
static int rcu_oom_notify(struct notifier_block *self,
			  unsigned long notused, void *nfreed)
{
	int cpu;

	/* Wait for callbacks from earlier instance to complete. */
	wait_event(oom_callback_wq, atomic_read(&oom_callback_count) == 0);

	/*
	 * Prevent premature wakeup: ensure that all increments happen
	 * before there is a chance of the counter reaching zero.
	 */
	atomic_set(&oom_callback_count, 1);

	get_online_cpus();
	for_each_online_cpu(cpu) {
		smp_call_function_single(cpu, rcu_oom_notify_cpu, NULL, 1);
		cond_resched();
	}
	put_online_cpus();

	/* Unconditionally decrement: no need to wake ourselves up. */
	atomic_dec(&oom_callback_count);

	return NOTIFY_OK;
}

static struct notifier_block rcu_oom_nb = {
	.notifier_call = rcu_oom_notify
};

static int __init rcu_register_oom_notifier(void)
{
	register_oom_notifier(&rcu_oom_nb);
	return 0;
}
early_initcall(rcu_register_oom_notifier);

1978
#endif /* #else #if !defined(CONFIG_RCU_FAST_NO_HZ) */
1979 1980 1981 1982 1983 1984 1985

#ifdef CONFIG_RCU_CPU_STALL_INFO

#ifdef CONFIG_RCU_FAST_NO_HZ

static void print_cpu_stall_fast_no_hz(char *cp, int cpu)
{
1986 1987
	struct rcu_dynticks *rdtp = &per_cpu(rcu_dynticks, cpu);
	struct timer_list *tltp = &rdtp->idle_gp_timer;
1988
	char c;
1989

1990 1991 1992 1993 1994 1995 1996
	c = rdtp->dyntick_holdoff == jiffies ? 'H' : '.';
	if (timer_pending(tltp))
		sprintf(cp, "drain=%d %c timer=%lu",
			rdtp->dyntick_drain, c, tltp->expires - jiffies);
	else
		sprintf(cp, "drain=%d %c timer not pending",
			rdtp->dyntick_drain, c);
1997 1998 1999 2000 2001 2002
}

#else /* #ifdef CONFIG_RCU_FAST_NO_HZ */

static void print_cpu_stall_fast_no_hz(char *cp, int cpu)
{
2003
	*cp = '\0';
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063
}

#endif /* #else #ifdef CONFIG_RCU_FAST_NO_HZ */

/* Initiate the stall-info list. */
static void print_cpu_stall_info_begin(void)
{
	printk(KERN_CONT "\n");
}

/*
 * Print out diagnostic information for the specified stalled CPU.
 *
 * If the specified CPU is aware of the current RCU grace period
 * (flavor specified by rsp), then print the number of scheduling
 * clock interrupts the CPU has taken during the time that it has
 * been aware.  Otherwise, print the number of RCU grace periods
 * that this CPU is ignorant of, for example, "1" if the CPU was
 * aware of the previous grace period.
 *
 * Also print out idle and (if CONFIG_RCU_FAST_NO_HZ) idle-entry info.
 */
static void print_cpu_stall_info(struct rcu_state *rsp, int cpu)
{
	char fast_no_hz[72];
	struct rcu_data *rdp = per_cpu_ptr(rsp->rda, cpu);
	struct rcu_dynticks *rdtp = rdp->dynticks;
	char *ticks_title;
	unsigned long ticks_value;

	if (rsp->gpnum == rdp->gpnum) {
		ticks_title = "ticks this GP";
		ticks_value = rdp->ticks_this_gp;
	} else {
		ticks_title = "GPs behind";
		ticks_value = rsp->gpnum - rdp->gpnum;
	}
	print_cpu_stall_fast_no_hz(fast_no_hz, cpu);
	printk(KERN_ERR "\t%d: (%lu %s) idle=%03x/%llx/%d %s\n",
	       cpu, ticks_value, ticks_title,
	       atomic_read(&rdtp->dynticks) & 0xfff,
	       rdtp->dynticks_nesting, rdtp->dynticks_nmi_nesting,
	       fast_no_hz);
}

/* Terminate the stall-info list. */
static void print_cpu_stall_info_end(void)
{
	printk(KERN_ERR "\t");
}

/* Zero ->ticks_this_gp for all flavors of RCU. */
static void zero_cpu_stall_ticks(struct rcu_data *rdp)
{
	rdp->ticks_this_gp = 0;
}

/* Increment ->ticks_this_gp for all flavors of RCU. */
static void increment_cpu_stall_ticks(void)
{
2064 2065 2066 2067
	struct rcu_state *rsp;

	for_each_rcu_flavor(rsp)
		__this_cpu_ptr(rsp->rda)->ticks_this_gp++;
2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095
}

#else /* #ifdef CONFIG_RCU_CPU_STALL_INFO */

static void print_cpu_stall_info_begin(void)
{
	printk(KERN_CONT " {");
}

static void print_cpu_stall_info(struct rcu_state *rsp, int cpu)
{
	printk(KERN_CONT " %d", cpu);
}

static void print_cpu_stall_info_end(void)
{
	printk(KERN_CONT "} ");
}

static void zero_cpu_stall_ticks(struct rcu_data *rdp)
{
}

static void increment_cpu_stall_ticks(void)
{
}

#endif /* #else #ifdef CONFIG_RCU_CPU_STALL_INFO */