tree_plugin.h 80.5 KB
Newer Older
1
/* SPDX-License-Identifier: GPL-2.0+ */
2 3 4
/*
 * Read-Copy Update mechanism for mutual exclusion (tree-based version)
 * Internal non-public definitions that provide either classic
Paul E. McKenney's avatar
Paul E. McKenney committed
5
 * or preemptible semantics.
6 7 8 9 10
 *
 * Copyright Red Hat, 2009
 * Copyright IBM Corporation, 2009
 *
 * Author: Ingo Molnar <mingo@elte.hu>
11
 *	   Paul E. McKenney <paulmck@linux.ibm.com>
12 13
 */

14
#include "../locking/rtmutex_common.h"
15

16 17
#ifdef CONFIG_RCU_NOCB_CPU
static cpumask_var_t rcu_nocb_mask; /* CPUs to have callbacks offloaded. */
18
static bool __read_mostly rcu_nocb_poll;    /* Offload kthread are to poll. */
19 20
#endif /* #ifdef CONFIG_RCU_NOCB_CPU */

21 22
/*
 * Check the RCU kernel configuration parameters and print informative
23
 * messages about anything out of the ordinary.
24 25 26
 */
static void __init rcu_bootup_announce_oddness(void)
{
27
	if (IS_ENABLED(CONFIG_RCU_TRACE))
28
		pr_info("\tRCU event tracing is enabled.\n");
29 30
	if ((IS_ENABLED(CONFIG_64BIT) && RCU_FANOUT != 64) ||
	    (!IS_ENABLED(CONFIG_64BIT) && RCU_FANOUT != 32))
31 32
		pr_info("\tCONFIG_RCU_FANOUT set to non-default value of %d.\n",
			RCU_FANOUT);
33
	if (rcu_fanout_exact)
34 35 36
		pr_info("\tHierarchical RCU autobalancing is disabled.\n");
	if (IS_ENABLED(CONFIG_RCU_FAST_NO_HZ))
		pr_info("\tRCU dyntick-idle grace-period acceleration is enabled.\n");
37
	if (IS_ENABLED(CONFIG_PROVE_RCU))
38
		pr_info("\tRCU lockdep checking is enabled.\n");
39 40
	if (RCU_NUM_LVLS >= 4)
		pr_info("\tFour(or more)-level hierarchy is enabled.\n");
41
	if (RCU_FANOUT_LEAF != 16)
42
		pr_info("\tBuild-time adjustment of leaf fanout to %d.\n",
43 44
			RCU_FANOUT_LEAF);
	if (rcu_fanout_leaf != RCU_FANOUT_LEAF)
45 46
		pr_info("\tBoot-time adjustment of leaf fanout to %d.\n",
			rcu_fanout_leaf);
47
	if (nr_cpu_ids != NR_CPUS)
48
		pr_info("\tRCU restricting CPUs from NR_CPUS=%d to nr_cpu_ids=%u.\n", NR_CPUS, nr_cpu_ids);
49
#ifdef CONFIG_RCU_BOOST
50 51
	pr_info("\tRCU priority boosting: priority %d delay %d ms.\n",
		kthread_prio, CONFIG_RCU_BOOST_DELAY);
52 53 54 55 56 57 58 59 60 61 62
#endif
	if (blimit != DEFAULT_RCU_BLIMIT)
		pr_info("\tBoot-time adjustment of callback invocation limit to %ld.\n", blimit);
	if (qhimark != DEFAULT_RCU_QHIMARK)
		pr_info("\tBoot-time adjustment of callback high-water mark to %ld.\n", qhimark);
	if (qlowmark != DEFAULT_RCU_QLOMARK)
		pr_info("\tBoot-time adjustment of callback low-water mark to %ld.\n", qlowmark);
	if (jiffies_till_first_fqs != ULONG_MAX)
		pr_info("\tBoot-time adjustment of first FQS scan delay to %ld jiffies.\n", jiffies_till_first_fqs);
	if (jiffies_till_next_fqs != ULONG_MAX)
		pr_info("\tBoot-time adjustment of subsequent FQS scan delay to %ld jiffies.\n", jiffies_till_next_fqs);
63 64
	if (jiffies_till_sched_qs != ULONG_MAX)
		pr_info("\tBoot-time adjustment of scheduler-enlistment delay to %ld jiffies.\n", jiffies_till_sched_qs);
65 66 67 68
	if (rcu_kick_kthreads)
		pr_info("\tKick kthreads if too-long grace period.\n");
	if (IS_ENABLED(CONFIG_DEBUG_OBJECTS_RCU_HEAD))
		pr_info("\tRCU callback double-/use-after-free debug enabled.\n");
69
	if (gp_preinit_delay)
70
		pr_info("\tRCU debug GP pre-init slowdown %d jiffies.\n", gp_preinit_delay);
71
	if (gp_init_delay)
72
		pr_info("\tRCU debug GP init slowdown %d jiffies.\n", gp_init_delay);
73
	if (gp_cleanup_delay)
74
		pr_info("\tRCU debug GP init slowdown %d jiffies.\n", gp_cleanup_delay);
75 76
	if (!use_softirq)
		pr_info("\tRCU_SOFTIRQ processing moved to rcuc kthreads.\n");
77 78
	if (IS_ENABLED(CONFIG_RCU_EQS_DEBUG))
		pr_info("\tRCU debug extended QS entry/exit.\n");
79
	rcupdate_announce_bootup_oddness();
80 81
}

82
#ifdef CONFIG_PREEMPT_RCU
83

84
static void rcu_report_exp_rnp(struct rcu_node *rnp, bool wake);
85
static void rcu_read_unlock_special(struct task_struct *t);
86

87 88 89
/*
 * Tell them what RCU they are running.
 */
90
static void __init rcu_bootup_announce(void)
91
{
92
	pr_info("Preemptible hierarchical RCU implementation.\n");
93
	rcu_bootup_announce_oddness();
94 95
}

96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129
/* Flags for rcu_preempt_ctxt_queue() decision table. */
#define RCU_GP_TASKS	0x8
#define RCU_EXP_TASKS	0x4
#define RCU_GP_BLKD	0x2
#define RCU_EXP_BLKD	0x1

/*
 * Queues a task preempted within an RCU-preempt read-side critical
 * section into the appropriate location within the ->blkd_tasks list,
 * depending on the states of any ongoing normal and expedited grace
 * periods.  The ->gp_tasks pointer indicates which element the normal
 * grace period is waiting on (NULL if none), and the ->exp_tasks pointer
 * indicates which element the expedited grace period is waiting on (again,
 * NULL if none).  If a grace period is waiting on a given element in the
 * ->blkd_tasks list, it also waits on all subsequent elements.  Thus,
 * adding a task to the tail of the list blocks any grace period that is
 * already waiting on one of the elements.  In contrast, adding a task
 * to the head of the list won't block any grace period that is already
 * waiting on one of the elements.
 *
 * This queuing is imprecise, and can sometimes make an ongoing grace
 * period wait for a task that is not strictly speaking blocking it.
 * Given the choice, we needlessly block a normal grace period rather than
 * blocking an expedited grace period.
 *
 * Note that an endless sequence of expedited grace periods still cannot
 * indefinitely postpone a normal grace period.  Eventually, all of the
 * fixed number of preempted tasks blocking the normal grace period that are
 * not also blocking the expedited grace period will resume and complete
 * their RCU read-side critical sections.  At that point, the ->gp_tasks
 * pointer will equal the ->exp_tasks pointer, at which point the end of
 * the corresponding expedited grace period will also be the end of the
 * normal grace period.
 */
130 131
static void rcu_preempt_ctxt_queue(struct rcu_node *rnp, struct rcu_data *rdp)
	__releases(rnp->lock) /* But leaves rrupts disabled. */
132 133 134 135 136 137 138
{
	int blkd_state = (rnp->gp_tasks ? RCU_GP_TASKS : 0) +
			 (rnp->exp_tasks ? RCU_EXP_TASKS : 0) +
			 (rnp->qsmask & rdp->grpmask ? RCU_GP_BLKD : 0) +
			 (rnp->expmask & rdp->grpmask ? RCU_EXP_BLKD : 0);
	struct task_struct *t = current;

139
	raw_lockdep_assert_held_rcu_node(rnp);
140
	WARN_ON_ONCE(rdp->mynode != rnp);
141
	WARN_ON_ONCE(!rcu_is_leaf_node(rnp));
142 143 144
	/* RCU better not be waiting on newly onlined CPUs! */
	WARN_ON_ONCE(rnp->qsmaskinitnext & ~rnp->qsmaskinit & rnp->qsmask &
		     rdp->grpmask);
145

146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221
	/*
	 * Decide where to queue the newly blocked task.  In theory,
	 * this could be an if-statement.  In practice, when I tried
	 * that, it was quite messy.
	 */
	switch (blkd_state) {
	case 0:
	case                RCU_EXP_TASKS:
	case                RCU_EXP_TASKS + RCU_GP_BLKD:
	case RCU_GP_TASKS:
	case RCU_GP_TASKS + RCU_EXP_TASKS:

		/*
		 * Blocking neither GP, or first task blocking the normal
		 * GP but not blocking the already-waiting expedited GP.
		 * Queue at the head of the list to avoid unnecessarily
		 * blocking the already-waiting GPs.
		 */
		list_add(&t->rcu_node_entry, &rnp->blkd_tasks);
		break;

	case                                              RCU_EXP_BLKD:
	case                                RCU_GP_BLKD:
	case                                RCU_GP_BLKD + RCU_EXP_BLKD:
	case RCU_GP_TASKS +                               RCU_EXP_BLKD:
	case RCU_GP_TASKS +                 RCU_GP_BLKD + RCU_EXP_BLKD:
	case RCU_GP_TASKS + RCU_EXP_TASKS + RCU_GP_BLKD + RCU_EXP_BLKD:

		/*
		 * First task arriving that blocks either GP, or first task
		 * arriving that blocks the expedited GP (with the normal
		 * GP already waiting), or a task arriving that blocks
		 * both GPs with both GPs already waiting.  Queue at the
		 * tail of the list to avoid any GP waiting on any of the
		 * already queued tasks that are not blocking it.
		 */
		list_add_tail(&t->rcu_node_entry, &rnp->blkd_tasks);
		break;

	case                RCU_EXP_TASKS +               RCU_EXP_BLKD:
	case                RCU_EXP_TASKS + RCU_GP_BLKD + RCU_EXP_BLKD:
	case RCU_GP_TASKS + RCU_EXP_TASKS +               RCU_EXP_BLKD:

		/*
		 * Second or subsequent task blocking the expedited GP.
		 * The task either does not block the normal GP, or is the
		 * first task blocking the normal GP.  Queue just after
		 * the first task blocking the expedited GP.
		 */
		list_add(&t->rcu_node_entry, rnp->exp_tasks);
		break;

	case RCU_GP_TASKS +                 RCU_GP_BLKD:
	case RCU_GP_TASKS + RCU_EXP_TASKS + RCU_GP_BLKD:

		/*
		 * Second or subsequent task blocking the normal GP.
		 * The task does not block the expedited GP. Queue just
		 * after the first task blocking the normal GP.
		 */
		list_add(&t->rcu_node_entry, rnp->gp_tasks);
		break;

	default:

		/* Yet another exercise in excessive paranoia. */
		WARN_ON_ONCE(1);
		break;
	}

	/*
	 * We have now queued the task.  If it was the first one to
	 * block either grace period, update the ->gp_tasks and/or
	 * ->exp_tasks pointers, respectively, to reference the newly
	 * blocked tasks.
	 */
222
	if (!rnp->gp_tasks && (blkd_state & RCU_GP_BLKD)) {
223
		WRITE_ONCE(rnp->gp_tasks, &t->rcu_node_entry);
224
		WARN_ON_ONCE(rnp->completedqs == rnp->gp_seq);
225
	}
226 227
	if (!rnp->exp_tasks && (blkd_state & RCU_EXP_BLKD))
		rnp->exp_tasks = &t->rcu_node_entry;
228 229 230 231
	WARN_ON_ONCE(!(blkd_state & RCU_GP_BLKD) !=
		     !(rnp->qsmask & rdp->grpmask));
	WARN_ON_ONCE(!(blkd_state & RCU_EXP_BLKD) !=
		     !(rnp->expmask & rdp->grpmask));
Boqun Feng's avatar
Boqun Feng committed
232
	raw_spin_unlock_rcu_node(rnp); /* interrupts remain disabled. */
233 234 235 236 237 238 239

	/*
	 * Report the quiescent state for the expedited GP.  This expedited
	 * GP should not be able to end until we report, so there should be
	 * no need to check for a subsequent expedited GP.  (Though we are
	 * still in a quiescent state in any case.)
	 */
240
	if (blkd_state & RCU_EXP_BLKD && rdp->exp_deferred_qs)
241
		rcu_report_exp_rdp(rdp);
242
	else
243
		WARN_ON_ONCE(rdp->exp_deferred_qs);
244 245
}

246
/*
247 248 249 250 251 252 253 254 255
 * Record a preemptible-RCU quiescent state for the specified CPU.
 * Note that this does not necessarily mean that the task currently running
 * on the CPU is in a quiescent state:  Instead, it means that the current
 * grace period need not wait on any RCU read-side critical section that
 * starts later on this CPU.  It also means that if the current task is
 * in an RCU read-side critical section, it has already added itself to
 * some leaf rcu_node structure's ->blkd_tasks list.  In addition to the
 * current task, there might be any number of other tasks blocked while
 * in an RCU read-side critical section.
256
 *
257
 * Callers to this function must disable preemption.
258
 */
259
static void rcu_qs(void)
260
{
261
	RCU_LOCKDEP_WARN(preemptible(), "rcu_qs() invoked with preemption enabled!!!\n");
262
	if (__this_cpu_read(rcu_data.cpu_no_qs.s)) {
263
		trace_rcu_grace_period(TPS("rcu_preempt"),
264
				       __this_cpu_read(rcu_data.gp_seq),
265
				       TPS("cpuqs"));
266
		__this_cpu_write(rcu_data.cpu_no_qs.b.norm, false);
267
		barrier(); /* Coordinate with rcu_flavor_sched_clock_irq(). */
268
		WRITE_ONCE(current->rcu_read_unlock_special.b.need_qs, false);
269
	}
270 271 272
}

/*
273 274 275
 * 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
276 277 278 279 280 281
 * 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.
282
 *
283
 * Caller must disable interrupts.
284
 */
285
void rcu_note_context_switch(bool preempt)
286 287
{
	struct task_struct *t = current;
288
	struct rcu_data *rdp = this_cpu_ptr(&rcu_data);
289 290
	struct rcu_node *rnp;

291
	trace_rcu_utilization(TPS("Start context switch"));
292
	lockdep_assert_irqs_disabled();
293
	WARN_ON_ONCE(!preempt && t->rcu_read_lock_nesting > 0);
294
	if (t->rcu_read_lock_nesting > 0 &&
295
	    !t->rcu_read_unlock_special.b.blocked) {
296 297 298

		/* Possibly blocking in an RCU read-side critical section. */
		rnp = rdp->mynode;
299
		raw_spin_lock_rcu_node(rnp);
300
		t->rcu_read_unlock_special.b.blocked = true;
301
		t->rcu_blocked_node = rnp;
302 303

		/*
304 305 306
		 * Verify the CPU's sanity, trace the preemption, and
		 * then queue the task as required based on the states
		 * of any ongoing and expedited grace periods.
307
		 */
308
		WARN_ON_ONCE((rdp->grpmask & rcu_rnp_online_cpus(rnp)) == 0);
309
		WARN_ON_ONCE(!list_empty(&t->rcu_node_entry));
310
		trace_rcu_preempt_task(rcu_state.name,
311 312
				       t->pid,
				       (rnp->qsmask & rdp->grpmask)
313 314
				       ? rnp->gp_seq
				       : rcu_seq_snap(&rnp->gp_seq));
315
		rcu_preempt_ctxt_queue(rnp, rdp);
316 317
	} else {
		rcu_preempt_deferred_qs(t);
318 319 320 321 322 323 324 325 326 327 328
	}

	/*
	 * 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.
	 */
329
	rcu_qs();
330
	if (rdp->exp_deferred_qs)
331
		rcu_report_exp_rdp(rdp);
332
	trace_rcu_utilization(TPS("End context switch"));
333
}
334
EXPORT_SYMBOL_GPL(rcu_note_context_switch);
335

336 337 338 339 340
/*
 * 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.
 */
341
static int rcu_preempt_blocked_readers_cgp(struct rcu_node *rnp)
342
{
343
	return READ_ONCE(rnp->gp_tasks) != NULL;
344 345
}

346 347 348 349 350
/* Bias and limit values for ->rcu_read_lock_nesting. */
#define RCU_NEST_BIAS INT_MAX
#define RCU_NEST_NMAX (-INT_MAX / 2)
#define RCU_NEST_PMAX (INT_MAX / 2)

351 352 353 354 355 356 357 358
/*
 * Preemptible RCU implementation for rcu_read_lock().
 * Just increment ->rcu_read_lock_nesting, shared state will be updated
 * if we block.
 */
void __rcu_read_lock(void)
{
	current->rcu_read_lock_nesting++;
359 360
	if (IS_ENABLED(CONFIG_PROVE_LOCKING))
		WARN_ON_ONCE(current->rcu_read_lock_nesting > RCU_NEST_PMAX);
361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379
	barrier();  /* critical section after entry code. */
}
EXPORT_SYMBOL_GPL(__rcu_read_lock);

/*
 * Preemptible RCU implementation for rcu_read_unlock().
 * 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;

	if (t->rcu_read_lock_nesting != 1) {
		--t->rcu_read_lock_nesting;
	} else {
		barrier();  /* critical section before exit code. */
380
		t->rcu_read_lock_nesting = -RCU_NEST_BIAS;
381 382 383 384 385 386
		barrier();  /* assign before ->rcu_read_unlock_special load */
		if (unlikely(READ_ONCE(t->rcu_read_unlock_special.s)))
			rcu_read_unlock_special(t);
		barrier();  /* ->rcu_read_unlock_special load before assign */
		t->rcu_read_lock_nesting = 0;
	}
387 388
	if (IS_ENABLED(CONFIG_PROVE_LOCKING)) {
		int rrln = t->rcu_read_lock_nesting;
389

390
		WARN_ON_ONCE(rrln < 0 && rrln > RCU_NEST_NMAX);
391 392 393 394
	}
}
EXPORT_SYMBOL_GPL(__rcu_read_unlock);

395 396 397 398 399 400 401 402 403 404 405 406 407 408 409
/*
 * 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;
}

410 411 412 413 414 415 416 417 418
/*
 * Return true if the specified rcu_node structure has tasks that were
 * preempted within an RCU read-side critical section.
 */
static bool rcu_preempt_has_tasks(struct rcu_node *rnp)
{
	return !list_empty(&rnp->blkd_tasks);
}

419
/*
420 421 422
 * Report deferred quiescent states.  The deferral time can
 * be quite short, for example, in the case of the call from
 * rcu_read_unlock_special().
423
 */
424 425
static void
rcu_preempt_deferred_qs_irqrestore(struct task_struct *t, unsigned long flags)
426
{
427 428 429
	bool empty_exp;
	bool empty_norm;
	bool empty_exp_now;
430
	struct list_head *np;
431
	bool drop_boost_mutex = false;
432
	struct rcu_data *rdp;
433
	struct rcu_node *rnp;
434
	union rcu_special special;
435 436

	/*
437 438
	 * If RCU core is waiting for this CPU to exit its critical section,
	 * report the fact that it has exited.  Because irqs are disabled,
439
	 * t->rcu_read_unlock_special cannot change.
440 441
	 */
	special = t->rcu_read_unlock_special;
442
	rdp = this_cpu_ptr(&rcu_data);
443
	if (!special.s && !rdp->exp_deferred_qs) {
444 445 446
		local_irq_restore(flags);
		return;
	}
447 448
	t->rcu_read_unlock_special.s = 0;
	if (special.b.need_qs)
449
		rcu_qs();
450

451
	/*
452 453 454 455
	 * Respond to a request by an expedited grace period for a
	 * quiescent state from this CPU.  Note that requests from
	 * tasks are handled when removing the task from the
	 * blocked-tasks list below.
456
	 */
457
	if (rdp->exp_deferred_qs)
458
		rcu_report_exp_rdp(rdp);
459

460
	/* Clean up if blocked during RCU read-side critical section. */
461
	if (special.b.blocked) {
462

463
		/*
464
		 * Remove this task from the list it blocked on.  The task
465 466 467
		 * now remains queued on the rcu_node corresponding to the
		 * CPU it first blocked on, so there is no longer any need
		 * to loop.  Retain a WARN_ON_ONCE() out of sheer paranoia.
468
		 */
469 470 471
		rnp = t->rcu_blocked_node;
		raw_spin_lock_rcu_node(rnp); /* irqs already disabled. */
		WARN_ON_ONCE(rnp != t->rcu_blocked_node);
472
		WARN_ON_ONCE(!rcu_is_leaf_node(rnp));
473
		empty_norm = !rcu_preempt_blocked_readers_cgp(rnp);
474
		WARN_ON_ONCE(rnp->completedqs == rnp->gp_seq &&
475
			     (!empty_norm || rnp->qsmask));
476
		empty_exp = sync_rcu_preempt_exp_done(rnp);
477
		smp_mb(); /* ensure expedited fastpath sees end of RCU c-s. */
478
		np = rcu_next_node_entry(t, rnp);
479
		list_del_init(&t->rcu_node_entry);
480
		t->rcu_blocked_node = NULL;
481
		trace_rcu_unlock_preempted_task(TPS("rcu_preempt"),
482
						rnp->gp_seq, t->pid);
483
		if (&t->rcu_node_entry == rnp->gp_tasks)
484
			WRITE_ONCE(rnp->gp_tasks, np);
485 486
		if (&t->rcu_node_entry == rnp->exp_tasks)
			rnp->exp_tasks = np;
487 488 489
		if (IS_ENABLED(CONFIG_RCU_BOOST)) {
			/* Snapshot ->boost_mtx ownership w/rnp->lock held. */
			drop_boost_mutex = rt_mutex_owner(&rnp->boost_mtx) == t;
490 491
			if (&t->rcu_node_entry == rnp->boost_tasks)
				rnp->boost_tasks = np;
492
		}
493 494 495 496

		/*
		 * If this was the last task on the current list, and if
		 * we aren't waiting on any CPUs, report the quiescent state.
497 498
		 * Note that rcu_report_unblock_qs_rnp() releases rnp->lock,
		 * so we must take a snapshot of the expedited state.
499
		 */
500
		empty_exp_now = sync_rcu_preempt_exp_done(rnp);
501
		if (!empty_norm && !rcu_preempt_blocked_readers_cgp(rnp)) {
502
			trace_rcu_quiescent_state_report(TPS("preempt_rcu"),
503
							 rnp->gp_seq,
504 505 506 507 508
							 0, rnp->qsmask,
							 rnp->level,
							 rnp->grplo,
							 rnp->grphi,
							 !!rnp->gp_tasks);
509
			rcu_report_unblock_qs_rnp(rnp, flags);
510
		} else {
Boqun Feng's avatar
Boqun Feng committed
511
			raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
512
		}
513

514
		/* Unboost if we were boosted. */
515
		if (IS_ENABLED(CONFIG_RCU_BOOST) && drop_boost_mutex)
516
			rt_mutex_futex_unlock(&rnp->boost_mtx);
517

518 519 520 521
		/*
		 * If this was the last task on the expedited lists,
		 * then we need to report up the rcu_node hierarchy.
		 */
522
		if (!empty_exp && empty_exp_now)
523
			rcu_report_exp_rnp(rnp, true);
524 525
	} else {
		local_irq_restore(flags);
526 527 528
	}
}

529 530 531 532 533 534 535 536 537 538 539
/*
 * Is a deferred quiescent-state pending, and are we also not in
 * an RCU read-side critical section?  It is the caller's responsibility
 * to ensure it is otherwise safe to report any deferred quiescent
 * states.  The reason for this is that it is safe to report a
 * quiescent state during context switch even though preemption
 * is disabled.  This function cannot be expected to understand these
 * nuances, so the caller must handle them.
 */
static bool rcu_preempt_need_deferred_qs(struct task_struct *t)
{
540
	return (__this_cpu_read(rcu_data.exp_deferred_qs) ||
541
		READ_ONCE(t->rcu_read_unlock_special.s)) &&
542
	       t->rcu_read_lock_nesting <= 0;
543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559
}

/*
 * Report a deferred quiescent state if needed and safe to do so.
 * As with rcu_preempt_need_deferred_qs(), "safe" involves only
 * not being in an RCU read-side critical section.  The caller must
 * evaluate safety in terms of interrupt, softirq, and preemption
 * disabling.
 */
static void rcu_preempt_deferred_qs(struct task_struct *t)
{
	unsigned long flags;
	bool couldrecurse = t->rcu_read_lock_nesting >= 0;

	if (!rcu_preempt_need_deferred_qs(t))
		return;
	if (couldrecurse)
560
		t->rcu_read_lock_nesting -= RCU_NEST_BIAS;
561 562 563
	local_irq_save(flags);
	rcu_preempt_deferred_qs_irqrestore(t, flags);
	if (couldrecurse)
564
		t->rcu_read_lock_nesting += RCU_NEST_BIAS;
565 566
}

567 568 569 570 571 572 573 574 575 576 577
/*
 * Minimal handler to give the scheduler a chance to re-evaluate.
 */
static void rcu_preempt_deferred_qs_handler(struct irq_work *iwp)
{
	struct rcu_data *rdp;

	rdp = container_of(iwp, struct rcu_data, defer_qs_iw);
	rdp->defer_qs_iw_pending = false;
}

578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595
/*
 * 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.
 */
static void rcu_read_unlock_special(struct task_struct *t)
{
	unsigned long flags;
	bool preempt_bh_were_disabled =
			!!(preempt_count() & (PREEMPT_MASK | SOFTIRQ_MASK));
	bool irqs_were_disabled;

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

	local_irq_save(flags);
	irqs_were_disabled = irqs_disabled_flags(flags);
596
	if (preempt_bh_were_disabled || irqs_were_disabled) {
597 598 599 600 601 602 603
		bool exp;
		struct rcu_data *rdp = this_cpu_ptr(&rcu_data);
		struct rcu_node *rnp = rdp->mynode;

		exp = (t->rcu_blocked_node && t->rcu_blocked_node->exp_tasks) ||
		      (rdp->grpmask & rnp->expmask) ||
		      tick_nohz_full_cpu(rdp->cpu);
604
		// Need to defer quiescent state until everything is enabled.
605 606 607
		if (irqs_were_disabled && use_softirq &&
		    (in_interrupt() ||
		     (exp && !t->rcu_read_unlock_special.b.deferred_qs))) {
608 609
			// Using softirq, safe to awaken, and we get
			// no help from enabling irqs, unlike bh/preempt.
610 611
			raise_softirq_irqoff(RCU_SOFTIRQ);
		} else {
612
			// Enabling BH or preempt does reschedule, so...
613
			// Also if no expediting or NO_HZ_FULL, slow is OK.
614 615
			set_tsk_need_resched(current);
			set_preempt_need_resched();
616
			if (IS_ENABLED(CONFIG_IRQ_WORK) && irqs_were_disabled &&
617 618 619 620 621 622 623 624
			    !rdp->defer_qs_iw_pending && exp) {
				// Get scheduler to re-evaluate and call hooks.
				// If !IRQ_WORK, FQS scan will eventually IPI.
				init_irq_work(&rdp->defer_qs_iw,
					      rcu_preempt_deferred_qs_handler);
				rdp->defer_qs_iw_pending = true;
				irq_work_queue_on(&rdp->defer_qs_iw, rdp->cpu);
			}
625
		}
626
		t->rcu_read_unlock_special.b.deferred_qs = true;
627 628 629 630 631 632
		local_irq_restore(flags);
		return;
	}
	rcu_preempt_deferred_qs_irqrestore(t, flags);
}

633 634 635 636
/*
 * 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
637
 * invoked -before- updating this rnp's ->gp_seq.
638 639 640
 *
 * Also, if there are blocked tasks on the list, they automatically
 * block the newly created grace period, so set up ->gp_tasks accordingly.
641
 */
642
static void rcu_preempt_check_blocked_tasks(struct rcu_node *rnp)
643
{
644 645
	struct task_struct *t;

646
	RCU_LOCKDEP_WARN(preemptible(), "rcu_preempt_check_blocked_tasks() invoked with preemption enabled!!!\n");
647
	raw_lockdep_assert_held_rcu_node(rnp);
648
	if (WARN_ON_ONCE(rcu_preempt_blocked_readers_cgp(rnp)))
649
		dump_blkd_tasks(rnp, 10);
650 651
	if (rcu_preempt_has_tasks(rnp) &&
	    (rnp->qsmaskinit || rnp->wait_blkd_tasks)) {
652
		WRITE_ONCE(rnp->gp_tasks, rnp->blkd_tasks.next);
653 654 655
		t = container_of(rnp->gp_tasks, struct task_struct,
				 rcu_node_entry);
		trace_rcu_unlock_preempted_task(TPS("rcu_preempt-GPS"),
656
						rnp->gp_seq, t->pid);
657
	}
658
	WARN_ON_ONCE(rnp->qsmask);
659 660
}

661
/*
662 663 664 665 666
 * Check for a quiescent state from the current CPU, including voluntary
 * context switches for Tasks RCU.  When a task blocks, the task is
 * recorded in the corresponding CPU's rcu_node structure, which is checked
 * elsewhere, hence this function need only check for quiescent states
 * related to the current CPU, not to those related to tasks.
667
 */
668
static void rcu_flavor_sched_clock_irq(int user)
669 670 671
{
	struct task_struct *t = current;

672 673 674
	if (user || rcu_is_cpu_rrupt_from_idle()) {
		rcu_note_voluntary_context_switch(current);
	}
675 676 677
	if (t->rcu_read_lock_nesting > 0 ||
	    (preempt_count() & (PREEMPT_MASK | SOFTIRQ_MASK))) {
		/* No QS, force context switch if deferred. */
678 679 680 681
		if (rcu_preempt_need_deferred_qs(t)) {
			set_tsk_need_resched(t);
			set_preempt_need_resched();
		}
682 683 684 685
	} else if (rcu_preempt_need_deferred_qs(t)) {
		rcu_preempt_deferred_qs(t); /* Report deferred QS. */
		return;
	} else if (!t->rcu_read_lock_nesting) {
686
		rcu_qs(); /* Report immediate QS. */
687 688
		return;
	}
689 690

	/* If GP is oldish, ask for help from rcu_read_unlock_special(). */
691
	if (t->rcu_read_lock_nesting > 0 &&
692 693
	    __this_cpu_read(rcu_data.core_needs_qs) &&
	    __this_cpu_read(rcu_data.cpu_no_qs.b.norm) &&
694
	    !t->rcu_read_unlock_special.b.need_qs &&
695
	    time_after(jiffies, rcu_state.gp_start + HZ))
696
		t->rcu_read_unlock_special.b.need_qs = true;
697 698
}

699 700
/*
 * Check for a task exiting while in a preemptible-RCU read-side
701 702 703 704 705
 * critical section, clean up if so.  No need to issue warnings, as
 * debug_check_no_locks_held() already does this if lockdep is enabled.
 * Besides, if this function does anything other than just immediately
 * return, there was a bug of some sort.  Spewing warnings from this
 * function is like as not to simply obscure important prior warnings.
706 707 708 709 710
 */
void exit_rcu(void)
{
	struct task_struct *t = current;

711 712 713
	if (unlikely(!list_empty(&current->rcu_node_entry))) {
		t->rcu_read_lock_nesting = 1;
		barrier();
714
		WRITE_ONCE(t->rcu_read_unlock_special.b.blocked, true);
715 716 717
	} else if (unlikely(t->rcu_read_lock_nesting)) {
		t->rcu_read_lock_nesting = 1;
	} else {
718
		return;
719
	}
720
	__rcu_read_unlock();
721
	rcu_preempt_deferred_qs(current);
722 723
}

724 725 726 727
/*
 * Dump the blocked-tasks state, but limit the list dump to the
 * specified number of elements.
 */
728
static void
729
dump_blkd_tasks(struct rcu_node *rnp, int ncheck)
730
{
731
	int cpu;
732 733
	int i;
	struct list_head *lhp;
734 735
	bool onl;
	struct rcu_data *rdp;
736
	struct rcu_node *rnp1;
737

738
	raw_lockdep_assert_held_rcu_node(rnp);
739
	pr_info("%s: grp: %d-%d level: %d ->gp_seq %ld ->completedqs %ld\n",
740
		__func__, rnp->grplo, rnp->grphi, rnp->level,
741 742 743 744
		(long)rnp->gp_seq, (long)rnp->completedqs);
	for (rnp1 = rnp; rnp1; rnp1 = rnp1->parent)
		pr_info("%s: %d:%d ->qsmask %#lx ->qsmaskinit %#lx ->qsmaskinitnext %#lx\n",
			__func__, rnp1->grplo, rnp1->grphi, rnp1->qsmask, rnp1->qsmaskinit, rnp1->qsmaskinitnext);
745
	pr_info("%s: ->gp_tasks %p ->boost_tasks %p ->exp_tasks %p\n",
746 747
		__func__, READ_ONCE(rnp->gp_tasks), rnp->boost_tasks,
		rnp->exp_tasks);
748
	pr_info("%s: ->blkd_tasks", __func__);
749 750 751
	i = 0;
	list_for_each(lhp, &rnp->blkd_tasks) {
		pr_cont(" %p", lhp);
752
		if (++i >= ncheck)
753 754 755
			break;
	}
	pr_cont("\n");
756
	for (cpu = rnp->grplo; cpu <= rnp->grphi; cpu++) {
757
		rdp = per_cpu_ptr(&rcu_data, cpu);
758 759 760 761 762 763
		onl = !!(rdp->grpmask & rcu_rnp_online_cpus(rnp));
		pr_info("\t%d: %c online: %ld(%d) offline: %ld(%d)\n",
			cpu, ".o"[onl],
			(long)rdp->rcu_onl_gp_seq, rdp->rcu_onl_gp_flags,
			(long)rdp->rcu_ofl_gp_seq, rdp->rcu_ofl_gp_flags);
	}
764 765
}

766
#else /* #ifdef CONFIG_PREEMPT_RCU */
767 768 769 770

/*
 * Tell them what RCU they are running.
 */
771
static void __init rcu_bootup_announce(void)
772
{
773
	pr_info("Hierarchical RCU implementation.\n");
774
	rcu_bootup_announce_oddness();
775 776
}

777
/*
778
 * Note a quiescent state for PREEMPTION=n.  Because we do not need to know
779 780 781 782 783
 * how many quiescent states passed, just if there was at least one since
 * the start of the grace period, this just sets a flag.  The caller must
 * have disabled preemption.
 */
static void rcu_qs(void)
784
{
785 786 787 788 789 790 791 792 793
	RCU_LOCKDEP_WARN(preemptible(), "rcu_qs() invoked with preemption enabled!!!");
	if (!__this_cpu_read(rcu_data.cpu_no_qs.s))
		return;
	trace_rcu_grace_period(TPS("rcu_sched"),
			       __this_cpu_read(rcu_data.gp_seq), TPS("cpuqs"));
	__this_cpu_write(rcu_data.cpu_no_qs.b.norm, false);
	if (!__this_cpu_read(rcu_data.cpu_no_qs.b.exp))
		return;
	__this_cpu_write(rcu_data.cpu_no_qs.b.exp, false);
794
	rcu_report_exp_rdp(this_cpu_ptr(&rcu_data));
795 796
}

797 798 799 800 801 802 803 804 805 806 807
/*
 * Register an urgently needed quiescent state.  If there is an
 * emergency, invoke rcu_momentary_dyntick_idle() to do a heavy-weight
 * dyntick-idle quiescent state visible to other CPUs, which will in
 * some cases serve for expedited as well as normal grace periods.
 * Either way, register a lightweight quiescent state.
 */
void rcu_all_qs(void)
{
	unsigned long flags;

808
	if (!raw_cpu_read(rcu_data.rcu_urgent_qs))
809 810 811
		return;
	preempt_disable();
	/* Load rcu_urgent_qs before other flags. */
812
	if (!smp_load_acquire(this_cpu_ptr(&rcu_data.rcu_urgent_qs))) {
813 814 815
		preempt_enable();
		return;
	}
816 817
	this_cpu_write(rcu_data.rcu_urgent_qs, false);
	if (unlikely(raw_cpu_read(rcu_data.rcu_need_heavy_qs))) {
818 819 820 821
		local_irq_save(flags);
		rcu_momentary_dyntick_idle();
		local_irq_restore(flags);
	}
822
	rcu_qs();
823 824 825 826
	preempt_enable();
}
EXPORT_SYMBOL_GPL(rcu_all_qs);

827
/*
828
 * Note a PREEMPTION=n context switch. The caller must have disabled interrupts.
829
 */
830
void rcu_note_context_switch(bool preempt)
831
{
832 833 834
	trace_rcu_utilization(TPS("Start context switch"));
	rcu_qs();
	/* Load rcu_urgent_qs before other flags. */
835
	if (!smp_load_acquire(this_cpu_ptr(&rcu_data.rcu_urgent_qs)))
836
		goto out;
837 838
	this_cpu_write(rcu_data.rcu_urgent_qs, false);
	if (unlikely(raw_cpu_read(rcu_data.rcu_need_heavy_qs)))
839 840 841 842 843
		rcu_momentary_dyntick_idle();
	if (!preempt)
		rcu_tasks_qs(current);
out:
	trace_rcu_utilization(TPS("End context switch"));
844
}
845
EXPORT_SYMBOL_GPL(rcu_note_context_switch);
846

847
/*
Paul E. McKenney's avatar
Paul E. McKenney committed
848
 * Because preemptible RCU does not exist, there are never any preempted
849 850
 * RCU readers.
 */
851
static int rcu_preempt_blocked_readers_cgp(struct rcu_node *rnp)
852 853 854 855
{
	return 0;
}

856 857 858 859
/*
 * Because there is no preemptible RCU, there can be no readers blocked.
 */
static bool rcu_preempt_has_tasks(struct rcu_node *rnp)
860
{
861
	return false;
862 863
}

864 865 866 867 868 869 870 871 872 873
/*
 * Because there is no preemptible RCU, there can be no deferred quiescent
 * states.
 */
static bool rcu_preempt_need_deferred_qs(struct task_struct *t)
{
	return false;
}
static void rcu_preempt_deferred_qs(struct task_struct *t) { }

874
/*
Paul E. McKenney's avatar
Paul E. McKenney committed
875
 * Because there is no preemptible RCU, there can be no readers blocked,
876 877
 * so there is no need to check for blocked tasks.  So check only for
 * bogus qsmask values.
878
 */
879
static void rcu_preempt_check_blocked_tasks(struct rcu_node *rnp)
880
{
881
	WARN_ON_ONCE(rnp->qsmask);
882 883
}

884
/*
885 886
 * Check to see if this CPU is in a non-context-switch quiescent state,
 * namely user mode and idle loop.
887
 */
888
static void rcu_flavor_sched_clock_irq(int user)
889
{
890
	if (user || rcu_is_cpu_rrupt_from_idle()) {
891

892 893 894 895 896 897 898 899 900 901 902 903 904 905
		/*
		 * Get here if this CPU took its interrupt from user
		 * mode or from the idle loop, and if this is not a
		 * nested interrupt.  In this case, the CPU is in
		 * a quiescent state, so note it.
		 *
		 * No memory barrier is required here because rcu_qs()
		 * references only CPU-local variables that other CPUs
		 * neither access nor modify, at least not while the
		 * corresponding CPU is online.
		 */

		rcu_qs();
	}
906 907
}

908 909 910 911 912 913 914 915
/*
 * Because preemptible RCU does not exist, tasks cannot possibly exit
 * while in preemptible RCU read-side critical sections.
 */
void exit_rcu(void)
{
}

916 917 918
/*
 * Dump the guaranteed-empty blocked-tasks state.  Trust but verify.
 */
919
static void
920
dump_blkd_tasks(struct rcu_node *rnp, int ncheck)
921 922 923 924
{
	WARN_ON_ONCE(!list_empty(&rnp->blkd_tasks));
}

925
#endif /* #else #ifdef CONFIG_PREEMPT_RCU */
926

927 928 929 930 931
/*
 * If boosting, set rcuc kthreads to realtime priority.
 */
static void rcu_cpu_kthread_setup(unsigned int cpu)
{
932
#ifdef CONFIG_RCU_BOOST
933
	struct sched_param sp;
934

935 936 937
	sp.sched_priority = kthread_prio;
	sched_setscheduler_nocheck(current, SCHED_FIFO, &sp);
#endif /* #ifdef CONFIG_RCU_BOOST */
Thomas Gleixner's avatar
Thomas Gleixner committed
938 939
}

940 941
#ifdef CONFIG_RCU_BOOST

942 943 944 945 946 947 948 949 950 951 952 953 954 955
/*
 * 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 task_struct *t;
	struct list_head *tb;

956 957
	if (READ_ONCE(rnp->exp_tasks) == NULL &&
	    READ_ONCE(rnp->boost_tasks) == NULL)
958 959
		return 0;  /* Nothing left to boost. */

960
	raw_spin_lock_irqsave_rcu_node(rnp, flags);
961 962 963 964 965 966

	/*
	 * 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) {
Boqun Feng's avatar
Boqun Feng committed
967
		raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
968 969 970 971 972 973 974 975 976
		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.
	 */
977
	if (rnp->exp_tasks != NULL)
978
		tb = rnp->exp_tasks;
979
	else
980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998
		tb = rnp->boost_tasks;

	/*
	 * 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);
999
	rt_mutex_init_proxy_locked(&rnp->boost_mtx, t);
Boqun Feng's avatar
Boqun Feng committed
1000
	raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
1001 1002 1003
	/* Lock only for side effect: boosts task t's priority. */
	rt_mutex_lock(&rnp->boost_mtx);
	rt_mutex_unlock(&rnp->boost_mtx);  /* Then keep lockdep happy. */
1004

1005 1006
	return READ_ONCE(rnp->exp_tasks) != NULL ||
	       READ_ONCE(rnp->boost_tasks) != NULL;
1007 1008 1009
}

/*
1010
 * Priority-boosting kthread, one per leaf rcu_node.
1011 1012 1013 1014 1015 1016 1017
 */
static int rcu_boost_kthread(void *arg)
{
	struct rcu_node *rnp = (struct rcu_node *)arg;
	int spincnt = 0;
	int more2boost;

1018
	trace_rcu_utilization(TPS("Start boost kthread@init"));
1019
	for (;;) {
1020
		rnp->boost_kthread_status = RCU_KTHREAD_WAITING;
1021
		trace_rcu_utilization(TPS("End boost kthread@rcu_wait"));
1022
		rcu_wait(rnp->boost_tasks || rnp->exp_tasks);
1023
		trace_rcu_utilization(TPS("Start boost kthread@rcu_wait"));
1024
		rnp->boost_kthread_status = RCU_KTHREAD_RUNNING;
1025 1026 1027 1028 1029 1030
		more2boost = rcu_boost(rnp);
		if (more2boost)
			spincnt++;
		else
			spincnt = 0;
		if (spincnt > 10) {
Thomas Gleixner's avatar
Thomas Gleixner committed
1031
			rnp->boost_kthread_status = RCU_KTHREAD_YIELDING;
1032
			trace_rcu_utilization(TPS("End boost kthread@rcu_yield"));
Thomas Gleixner's avatar
Thomas Gleixner committed
1033
			schedule_timeout_interruptible(2);
1034
			trace_rcu_utilization(TPS("Start boost kthread@rcu_yield"));
1035 1036 1037
			spincnt = 0;
		}
	}
1038
	/* NOTREACHED */
1039
	trace_rcu_utilization(TPS("End boost kthread@notreached"));
1040 1041 1042 1043 1044 1045 1046 1047 1048
	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.
 *
1049 1050 1051
 * 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.
1052
 */
1053
static void rcu_initiate_boost(struct rcu_node *rnp, unsigned long flags)
1054
	__releases(rnp->lock)
1055
{
1056
	raw_lockdep_assert_held_rcu_node(rnp);
1057
	if (!rcu_preempt_blocked_readers_cgp(rnp) && rnp->exp_tasks == NULL) {
Boqun Feng's avatar
Boqun Feng committed
1058
		raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
1059
		return;
1060
	}
1061 1062 1063 1064 1065 1066 1067
	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;
Boqun Feng's avatar
Boqun Feng committed
1068
		raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
1069 1070
		rcu_wake_cond(rnp->boost_kthread_task,
			      rnp->boost_kthread_status);
1071
	} else {
Boqun Feng's avatar
Boqun Feng committed
1072
		raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
1073
	}
1074 1075
}

1076 1077 1078 1079 1080 1081
/*
 * Is the current CPU running the RCU-callbacks kthread?
 * Caller must have preemption disabled.
 */
static bool rcu_is_callbacks_kthread(void)
{
1082
	return __this_cpu_read(rcu_data.rcu_cpu_kthread_task) == current;
1083 1084
}

1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099
#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.
 */
1100
static void rcu_spawn_one_boost_kthread(struct rcu_node *rnp)
1101
{
1102
	int rnp_index = rnp - rcu_get_root();
1103 1104 1105 1106
	unsigned long flags;
	struct sched_param sp;
	struct task_struct *t;

1107
	if (!IS_ENABLED(CONFIG_PREEMPT_RCU))
1108
		return;
Thomas Gleixner's avatar
Thomas Gleixner committed
1109

1110
	if (!rcu_scheduler_fully_active || rcu_rnp_online_cpus(rnp) == 0)
1111
		return;
Thomas Gleixner's avatar
Thomas Gleixner committed
1112

1113
	rcu_state.boost = 1;
1114

1115
	if (rnp->boost_kthread_task != NULL)
1116 1117
		return;

1118
	t = kthread_create(rcu_boost_kthread, (void *)rnp,
1119
			   "rcub/%d", rnp_index);
1120 1121 1122
	if (WARN_ON_ONCE(IS_ERR(t)))
		return;

1123
	raw_spin_lock_irqsave_rcu_node(rnp, flags);
1124
	rnp->boost_kthread_task = t;
Boqun Feng's avatar
Boqun Feng committed
1125
	raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
1126
	sp.sched_priority = kthread_prio;
1127
	sched_setscheduler_nocheck(t, SCHED_FIFO, &sp);
1128
	wake_up_process(t); /* get to TASK_INTERRUPTIBLE quickly. */
1129 1130
}

1131 1132 1133 1134 1135 1136 1137 1138 1139
/*
 * 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.
 */
Thomas Gleixner's avatar
Thomas Gleixner committed
1140
static void rcu_boost_kthread_setaffinity(struct rcu_node *rnp, int outgoingcpu)
1141
{
Thomas Gleixner's avatar
Thomas Gleixner committed
1142
	struct task_struct *t = rnp->boost_kthread_task;
1143
	unsigned long mask = rcu_rnp_online_cpus(rnp);
1144 1145 1146
	cpumask_var_t cm;
	int cpu;

Thomas Gleixner's avatar
Thomas Gleixner committed
1147
	if (!t)
1148
		return;
Thomas Gleixner's avatar
Thomas Gleixner committed
1149
	if (!zalloc_cpumask_var(&cm, GFP_KERNEL))
1150
		return;
1151 1152 1153
	for_each_leaf_node_possible_cpu(rnp, cpu)
		if ((mask & leaf_node_cpu_bit(rnp, cpu)) &&
		    cpu != outgoingcpu)
1154
			cpumask_set_cpu(cpu, cm);
1155
	if (cpumask_weight(cm) == 0)
1156
		cpumask_setall(cm);
Thomas Gleixner's avatar
Thomas Gleixner committed
1157
	set_cpus_allowed_ptr(t, cm);
1158 1159 1160 1161
	free_cpumask_var(cm);
}

/*
1162
 * Spawn boost kthreads -- called as soon as the scheduler is running.
1163
 */
1164
static void __init rcu_spawn_boost_kthreads(void)
1165 1166 1167
{
	struct rcu_node *rnp;

1168
	rcu_for_each_leaf_node(rnp)
1169
		rcu_spawn_one_boost_kthread(rnp);
1170 1171
}

1172
static void rcu_prepare_kthreads(int cpu)
1173
{
1174
	struct rcu_data *rdp = per_cpu_ptr(&rcu_data, cpu);
1175 1176 1177
	struct rcu_node *rnp = rdp->mynode;

	/* Fire up the incoming CPU's kthread and leaf rcu_node kthread. */
1178
	if (rcu_scheduler_fully_active)
1179
		rcu_spawn_one_boost_kthread(rnp);
1180 1181
}

1182 1183
#else /* #ifdef CONFIG_RCU_BOOST */

1184
static void rcu_initiate_boost(struct rcu_node *rnp, unsigned long flags)
1185
	__releases(rnp->lock)
1186
{
Boqun Feng's avatar
Boqun Feng committed
1187
	raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
1188 1189
}

1190 1191 1192 1193 1194
static bool rcu_is_callbacks_kthread(void)
{
	return false;
}

1195 1196 1197 1198
static void rcu_preempt_boost_start_gp(struct rcu_node *rnp)
{
}

Thomas Gleixner's avatar
Thomas Gleixner committed
1199
static void rcu_boost_kthread_setaffinity(struct rcu_node *rnp, int outgoingcpu)
1200 1201 1202
{
}

1203
static void __init rcu_spawn_boost_kthreads(void)
1204 1205 1206
{
}

1207
static void rcu_prepare_kthreads(int cpu)
1208 1209 1210
{
}

1211 1212
#endif /* #else #ifdef CONFIG_RCU_BOOST */

1213 1214 1215
#if !defined(CONFIG_RCU_FAST_NO_HZ)

/*
1216 1217 1218 1219
 * Check to see if any future non-offloaded 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.
1220
 *
1221 1222
 * Because we not have RCU_FAST_NO_HZ, just check whether or not this
 * CPU has RCU callbacks queued.
1223
 */
1224
int rcu_needs_cpu(u64 basemono, u64 *nextevt)
1225
{
1226
	*nextevt = KTIME_MAX;
1227 1228
	return !rcu_segcblist_empty(&this_cpu_ptr(&rcu_data)->cblist) &&
	       !rcu_segcblist_is_offloaded(&this_cpu_ptr(&rcu_data)->cblist);
1229 1230 1231 1232 1233 1234
}

/*
 * Because we do not have RCU_FAST_NO_HZ, don't bother cleaning up
 * after it.
 */
1235
static void rcu_cleanup_after_idle(void)
1236 1237 1238
{
}

1239
/*
1240
 * Do the idle-entry grace-period work, which, because CONFIG_RCU_FAST_NO_HZ=n,
1241 1242
 * is nothing.
 */
1243
static void rcu_prepare_for_idle(void)
1244 1245 1246
{
}

1247 1248
#else /* #if !defined(CONFIG_RCU_FAST_NO_HZ) */

1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263
/*
 * 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_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!
1264 1265 1266
 * 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.
1267 1268 1269 1270 1271
 *
 * 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.
 */
1272
#define RCU_IDLE_GP_DELAY 4		/* Roughly one grace period. */
1273
#define RCU_IDLE_LAZY_GP_DELAY (6 * HZ)	/* Roughly six seconds. */
1274

1275 1276 1277 1278
static int rcu_idle_gp_delay = RCU_IDLE_GP_DELAY;
module_param(rcu_idle_gp_delay, int, 0644);
static int rcu_idle_lazy_gp_delay = RCU_IDLE_LAZY_GP_DELAY;
module_param(rcu_idle_lazy_gp_delay, int, 0644);
1279 1280

/*
1281 1282 1283
 * Try to advance callbacks on the current CPU, but only if it has been
 * awhile since the last time we did so.  Afterwards, if there are any
 * callbacks ready for immediate invocation, return true.
1284
 */
1285
static bool __maybe_unused rcu_try_advance_all_cbs(void)
1286
{
1287
	bool cbs_ready = false;
1288
	struct rcu_data *rdp = this_cpu_ptr(&rcu_data);
1289
	struct rcu_node *rnp;
1290

1291
	/* Exit early if we advanced recently. */
1292
	if (jiffies == rdp->last_advance_all)
1293
		return false;
1294
	rdp->last_advance_all = jiffies;
1295

1296
	rnp = rdp->mynode;
1297

1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310
	/*
	 * Don't bother checking unless a grace period has
	 * completed since we last checked and there are
	 * callbacks not yet ready to invoke.
	 */
	if ((rcu_seq_completed_gp(rdp->gp_seq,
				  rcu_seq_current(&rnp->gp_seq)) ||
	     unlikely(READ_ONCE(rdp->gpwrap))) &&
	    rcu_segcblist_pend_cbs(&rdp->cblist))
		note_gp_changes(rdp);

	if (rcu_segcblist_ready_cbs(&rdp->cblist))
		cbs_ready = true;
1311
	return cbs_ready;
1312 1313
}

1314
/*
1315 1316 1317 1318
 * Allow the CPU to enter dyntick-idle mode unless it has callbacks ready
 * to invoke.  If the CPU has callbacks, try to advance them.  Tell the
 * caller to set the timeout based on whether or not there are non-lazy
 * callbacks.
1319
 *
1320
 * The caller must have disabled interrupts.
1321
 */
1322
int rcu_needs_cpu(u64 basemono, u64 *nextevt)
1323
{
1324
	struct rcu_data *rdp = this_cpu_ptr(&rcu_data);
1325
	unsigned long dj;
1326

1327
	lockdep_assert_irqs_disabled();
1328

1329 1330 1331
	/* If no non-offloaded callbacks, RCU doesn't need the CPU. */
	if (rcu_segcblist_empty(&rdp->cblist) ||
	    rcu_segcblist_is_offloaded(&this_cpu_ptr(&rcu_data)->cblist)) {
1332
		*nextevt = KTIME_MAX;
1333 1334
		return 0;
	}
1335 1336 1337 1338 1339

	/* Attempt to advance callbacks. */
	if (rcu_try_advance_all_cbs()) {
		/* Some ready to invoke, so initiate later invocation. */
		invoke_rcu_core();
1340 1341
		return 1;
	}
1342
	rdp->last_accelerate = jiffies;
1343 1344

	/* Request timer delay depending on laziness, and round. */
1345 1346 1347 1348
	rdp->all_lazy = !rcu_segcblist_n_nonlazy_cbs(&rdp->cblist);
	if (rdp->all_lazy) {
		dj = round_jiffies(rcu_idle_lazy_gp_delay + jiffies) - jiffies;
	} else {
1349
		dj = round_up(rcu_idle_gp_delay + jiffies,
1350
			       rcu_idle_gp_delay) - jiffies;
1351
	}
1352
	*nextevt = basemono + dj * TICK_NSEC;
1353 1354 1355
	return 0;
}

1356
/*
1357 1358 1359 1360 1361 1362
 * Prepare a CPU for idle from an RCU perspective.  The first major task
 * is to sense whether nohz mode has been enabled or disabled via sysfs.
 * The second major task is to check to see if a non-lazy callback has
 * arrived at a CPU that previously had only lazy callbacks.  The third
 * major task is to accelerate (that is, assign grace-period numbers to)
 * any recently arrived callbacks.
1363 1364
 *
 * The caller must have disabled interrupts.
1365
 */
1366
static void rcu_prepare_for_idle(void)
1367
{
1368
	bool needwake;
1369
	struct rcu_data *rdp = this_cpu_ptr(&rcu_data);
1370
	struct rcu_node *rnp;
1371 1372
	int tne;

1373
	lockdep_assert_irqs_disabled();
1374
	if (rcu_segcblist_is_offloaded(&rdp->cblist))
1375 1376
		return;

1377
	/* Handle nohz enablement switches conservatively. */
1378
	tne = READ_ONCE(tick_nohz_active);
1379
	if (tne != rdp->tick_nohz_enabled_snap) {
1380
		if (!rcu_segcblist_empty(&rdp->cblist))
1381
			invoke_rcu_core(); /* force nohz to see update. */
1382
		rdp->tick_nohz_enabled_snap = tne;
1383 1384 1385 1386
		return;
	}
	if (!tne)
		return;
1387

1388
	/*
1389 1390 1391
	 * If a non-lazy callback arrived at a CPU having only lazy
	 * callbacks, invoke RCU core for the side-effect of recalculating
	 * idle duration on re-entry to idle.
1392
	 */
1393
	if (rdp->all_lazy && rcu_segcblist_n_nonlazy_cbs(&rdp->cblist)) {
1394
		rdp->all_lazy = false;
1395
		invoke_rcu_core();
1396 1397 1398
		return;
	}

1399
	/*
1400 1401
	 * If we have not yet accelerated this jiffy, accelerate all
	 * callbacks on this CPU.
1402
	 */
1403
	if (rdp->last_accelerate == jiffies)
1404
		return;
1405
	rdp->last_accelerate = jiffies;
1406
	if (rcu_segcblist_pend_cbs(&rdp->cblist)) {
1407
		rnp = rdp->mynode;
1408
		raw_spin_lock_rcu_node(rnp); /* irqs already disabled. */
1409
		needwake = rcu_accelerate_cbs(rnp, rdp);
Boqun Feng's avatar
Boqun Feng committed
1410
		raw_spin_unlock_rcu_node(rnp); /* irqs remain disabled. */
1411
		if (needwake)
1412
			rcu_gp_kthread_wake();
1413
	}
1414
}
1415

1416 1417 1418 1419 1420
/*
 * Clean up for exit from idle.  Attempt to advance callbacks based on
 * any grace periods that elapsed while the CPU was idle, and if any
 * callbacks are now ready to invoke, initiate invocation.
 */
1421
static void rcu_cleanup_after_idle(void)
1422
{
1423 1424
	struct rcu_data *rdp = this_cpu_ptr(&rcu_data);

1425
	lockdep_assert_irqs_disabled();
1426
	if (rcu_segcblist_is_offloaded(&rdp->cblist))
1427
		return;
1428 1429
	if (rcu_try_advance_all_cbs())
		invoke_rcu_core();
1430 1431 1432
}

#endif /* #else #if !defined(CONFIG_RCU_FAST_NO_HZ) */
1433

1434 1435 1436 1437
#ifdef CONFIG_RCU_NOCB_CPU

/*
 * Offload callback processing from the boot-time-specified set of CPUs
1438 1439 1440
 * specified by rcu_nocb_mask.  For the CPUs in the set, there are kthreads
 * created that pull the callbacks from the corresponding CPU, wait for
 * a grace period to elapse, and invoke the callbacks.  These kthreads
1441 1442 1443 1444
 * are organized into GP kthreads, which manage incoming callbacks, wait for
 * grace periods, and awaken CB kthreads, and the CB kthreads, which only
 * invoke callbacks.  Each GP kthread invokes its own CBs.  The no-CBs CPUs
 * do a wake_up() on their GP kthread when they insert a callback into any
1445 1446 1447
 * empty list, unless the rcu_nocb_poll boot parameter has been specified,
 * in which case each kthread actively polls its CPU.  (Which isn't so great
 * for energy efficiency, but which does reduce RCU's overhead on that CPU.)
1448 1449 1450 1451 1452
 *
 * This is intended to be used in conjunction with Frederic Weisbecker's
 * adaptive-idle work, which would seriously reduce OS jitter on CPUs
 * running CPU-bound user-mode computations.
 *
1453 1454 1455
 * Offloading of callbacks can also be used as an energy-efficiency
 * measure because CPUs with no RCU callbacks queued are more aggressive
 * about entering dyntick-idle mode.
1456 1457 1458
 */


1459 1460 1461 1462 1463 1464
/*
 * Parse the boot-time rcu_nocb_mask CPU list from the kernel parameters.
 * The string after the "rcu_nocbs=" is either "all" for all CPUs, or a
 * comma-separated list of CPUs and/or CPU ranges.  If an invalid list is
 * given, a warning is emitted and all CPUs are offloaded.
 */
1465 1466 1467
static int __init rcu_nocb_setup(char *str)
{
	alloc_bootmem_cpumask_var(&rcu_nocb_mask);
1468 1469 1470
	if (!strcasecmp(str, "all"))
		cpumask_setall(rcu_nocb_mask);
	else
1471 1472 1473 1474
		if (cpulist_parse(str, rcu_nocb_mask)) {
			pr_warn("rcu_nocbs= bad CPU range, all CPUs set\n");
			cpumask_setall(rcu_nocb_mask);
		}
1475 1476 1477 1478
	return 1;
}
__setup("rcu_nocbs=", rcu_nocb_setup);

1479 1480
static int __init parse_rcu_nocb_poll(char *arg)
{
1481
	rcu_nocb_poll = true;
1482 1483 1484 1485
	return 0;
}
early_param("rcu_nocb_poll", parse_rcu_nocb_poll);

1486
/*
1487 1488 1489
 * Don't bother bypassing ->cblist if the call_rcu() rate is low.
 * After all, the main point of bypassing is to avoid lock contention
 * on ->nocb_lock, which only can happen at high call_rcu() rates.
1490
 */
1491 1492 1493 1494 1495 1496 1497 1498 1499
int nocb_nobypass_lim_per_jiffy = 16 * 1000 / HZ;
module_param(nocb_nobypass_lim_per_jiffy, int, 0);

/*
 * Acquire the specified rcu_data structure's ->nocb_bypass_lock.  If the
 * lock isn't immediately available, increment ->nocb_lock_contended to
 * flag the contention.
 */
static void rcu_nocb_bypass_lock(struct rcu_data *rdp)
1500
{
1501
	lockdep_assert_irqs_disabled();
1502
	if (raw_spin_trylock(&rdp->nocb_bypass_lock))
1503 1504
		return;
	atomic_inc(&rdp->nocb_lock_contended);
1505
	WARN_ON_ONCE(smp_processor_id() != rdp->cpu);
1506
	smp_mb__after_atomic(); /* atomic_inc() before lock. */
1507
	raw_spin_lock(&rdp->nocb_bypass_lock);
1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523
	smp_mb__before_atomic(); /* atomic_dec() after lock. */
	atomic_dec(&rdp->nocb_lock_contended);
}

/*
 * Spinwait until the specified rcu_data structure's ->nocb_lock is
 * not contended.  Please note that this is extremely special-purpose,
 * relying on the fact that at most two kthreads and one CPU contend for
 * this lock, and also that the two kthreads are guaranteed to have frequent
 * grace-period-duration time intervals between successive acquisitions
 * of the lock.  This allows us to use an extremely simple throttling
 * mechanism, and further to apply it only to the CPU doing floods of
 * call_rcu() invocations.  Don't try this at home!
 */
static void rcu_nocb_wait_contended(struct rcu_data *rdp)
{
1524 1525
	WARN_ON_ONCE(smp_processor_id() != rdp->cpu);
	while (WARN_ON_ONCE(atomic_read(&rdp->nocb_lock_contended)))
1526
		cpu_relax();
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
/*
 * Conditionally acquire the specified rcu_data structure's
 * ->nocb_bypass_lock.
 */
static bool rcu_nocb_bypass_trylock(struct rcu_data *rdp)
{
	lockdep_assert_irqs_disabled();
	return raw_spin_trylock(&rdp->nocb_bypass_lock);
}

/*
 * Release the specified rcu_data structure's ->nocb_bypass_lock.
 */
static void rcu_nocb_bypass_unlock(struct rcu_data *rdp)
{
	lockdep_assert_irqs_disabled();
	raw_spin_unlock(&rdp->nocb_bypass_lock);
}

/*
 * Acquire the specified rcu_data structure's ->nocb_lock, but only
 * if it corresponds to a no-CBs CPU.
 */
static void rcu_nocb_lock(struct rcu_data *rdp)
{
	lockdep_assert_irqs_disabled();
	if (!rcu_segcblist_is_offloaded(&rdp->cblist))
		return;
	raw_spin_lock(&rdp->nocb_lock);
}

1560 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
/*
 * Release the specified rcu_data structure's ->nocb_lock, but only
 * if it corresponds to a no-CBs CPU.
 */
static void rcu_nocb_unlock(struct rcu_data *rdp)
{
	if (rcu_segcblist_is_offloaded(&rdp->cblist)) {
		lockdep_assert_irqs_disabled();
		raw_spin_unlock(&rdp->nocb_lock);
	}
}

/*
 * Release the specified rcu_data structure's ->nocb_lock and restore
 * interrupts, but only if it corresponds to a no-CBs CPU.
 */
static void rcu_nocb_unlock_irqrestore(struct rcu_data *rdp,
				       unsigned long flags)
{
	if (rcu_segcblist_is_offloaded(&rdp->cblist)) {
		lockdep_assert_irqs_disabled();
		raw_spin_unlock_irqrestore(&rdp->nocb_lock, flags);
	} else {
		local_irq_restore(flags);
	}
}

1587 1588 1589 1590 1591 1592 1593 1594 1595
/* Lockdep check that ->cblist may be safely accessed. */
static void rcu_lockdep_assert_cblist_protected(struct rcu_data *rdp)
{
	lockdep_assert_irqs_disabled();
	if (rcu_segcblist_is_offloaded(&rdp->cblist) &&
	    cpu_online(rdp->cpu))
		lockdep_assert_held(&rdp->nocb_lock);
}

1596
/*
1597 1598
 * Wake up any no-CBs CPUs' kthreads that were waiting on the just-ended
 * grace period.
1599
 */
1600
static void rcu_nocb_gp_cleanup(struct swait_queue_head *sq)
1601
{
1602
	swake_up_all(sq);
1603 1604
}

1605
static struct swait_queue_head *rcu_nocb_gp_get(struct rcu_node *rnp)
1606
{
1607
	return &rnp->nocb_gp_wq[rcu_seq_ctr(rnp->gp_seq) & 0x1];
1608 1609
}

1610
static void rcu_init_one_nocb(struct rcu_node *rnp)
1611
{
1612 1613
	init_swait_queue_head(&rnp->nocb_gp_wq[0]);
	init_swait_queue_head(&rnp->nocb_gp_wq[1]);
1614 1615
}

1616
/* Is the specified CPU a no-CBs CPU? */
1617
bool rcu_is_nocb_cpu(int cpu)
1618
{
1619
	if (cpumask_available(rcu_nocb_mask))
1620 1621 1622 1623
		return cpumask_test_cpu(cpu, rcu_nocb_mask);
	return false;
}

1624
/*
1625
 * Kick the GP kthread for this NOCB group.  Caller holds ->nocb_lock
1626
 * and this function releases it.
1627
 */
1628
static void wake_nocb_gp(struct rcu_data *rdp, bool force,
1629
			   unsigned long flags)
1630
	__releases(rdp->nocb_lock)
1631
{
1632
	bool needwake = false;
1633
	struct rcu_data *rdp_gp = rdp->nocb_gp_rdp;
1634

1635
	lockdep_assert_held(&rdp->nocb_lock);
1636
	if (!READ_ONCE(rdp_gp->nocb_gp_kthread)) {
1637 1638
		trace_rcu_nocb_wake(rcu_state.name, rdp->cpu,
				    TPS("AlreadyAwake"));
1639
		rcu_nocb_unlock_irqrestore(rdp, flags);
1640
		return;
1641
	}
1642 1643 1644 1645
	del_timer(&rdp->nocb_timer);
	rcu_nocb_unlock_irqrestore(rdp, flags);
	raw_spin_lock_irqsave(&rdp_gp->nocb_gp_lock, flags);
	if (force || READ_ONCE(rdp_gp->nocb_gp_sleep)) {
1646
		WRITE_ONCE(rdp_gp->nocb_gp_sleep, false);
1647 1648
		needwake = true;
		trace_rcu_nocb_wake(rcu_state.name, rdp->cpu, TPS("DoWake"));
1649
	}
1650 1651 1652
	raw_spin_unlock_irqrestore(&rdp_gp->nocb_gp_lock, flags);
	if (needwake)
		wake_up_process(rdp_gp->nocb_gp_kthread);
1653 1654
}

1655
/*
1656 1657
 * Arrange to wake the GP kthread for this NOCB group at some future
 * time when it is safe to do so.
1658
 */
1659 1660
static void wake_nocb_gp_defer(struct rcu_data *rdp, int waketype,
			       const char *reason)
1661 1662 1663
{
	if (rdp->nocb_defer_wakeup == RCU_NOCB_WAKE_NOT)
		mod_timer(&rdp->nocb_timer, jiffies + 1);
1664 1665
	if (rdp->nocb_defer_wakeup < waketype)
		WRITE_ONCE(rdp->nocb_defer_wakeup, waketype);
1666
	trace_rcu_nocb_wake(rcu_state.name, rdp->cpu, reason);
1667 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 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 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 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851
/*
 * Flush the ->nocb_bypass queue into ->cblist, enqueuing rhp if non-NULL.
 * However, if there is a callback to be enqueued and if ->nocb_bypass
 * proves to be initially empty, just return false because the no-CB GP
 * kthread may need to be awakened in this case.
 *
 * Note that this function always returns true if rhp is NULL.
 */
static bool rcu_nocb_do_flush_bypass(struct rcu_data *rdp, struct rcu_head *rhp,
				     unsigned long j)
{
	struct rcu_cblist rcl;

	WARN_ON_ONCE(!rcu_segcblist_is_offloaded(&rdp->cblist));
	rcu_lockdep_assert_cblist_protected(rdp);
	lockdep_assert_held(&rdp->nocb_bypass_lock);
	if (rhp && !rcu_cblist_n_cbs(&rdp->nocb_bypass)) {
		raw_spin_unlock(&rdp->nocb_bypass_lock);
		return false;
	}
	/* Note: ->cblist.len already accounts for ->nocb_bypass contents. */
	if (rhp)
		rcu_segcblist_inc_len(&rdp->cblist); /* Must precede enqueue. */
	rcu_cblist_flush_enqueue(&rcl, &rdp->nocb_bypass, rhp);
	rcu_segcblist_insert_pend_cbs(&rdp->cblist, &rcl);
	WRITE_ONCE(rdp->nocb_bypass_first, j);
	rcu_nocb_bypass_unlock(rdp);
	return true;
}

/*
 * Flush the ->nocb_bypass queue into ->cblist, enqueuing rhp if non-NULL.
 * However, if there is a callback to be enqueued and if ->nocb_bypass
 * proves to be initially empty, just return false because the no-CB GP
 * kthread may need to be awakened in this case.
 *
 * Note that this function always returns true if rhp is NULL.
 */
static bool rcu_nocb_flush_bypass(struct rcu_data *rdp, struct rcu_head *rhp,
				  unsigned long j)
{
	if (!rcu_segcblist_is_offloaded(&rdp->cblist))
		return true;
	rcu_lockdep_assert_cblist_protected(rdp);
	rcu_nocb_bypass_lock(rdp);
	return rcu_nocb_do_flush_bypass(rdp, rhp, j);
}

/*
 * If the ->nocb_bypass_lock is immediately available, flush the
 * ->nocb_bypass queue into ->cblist.
 */
static void rcu_nocb_try_flush_bypass(struct rcu_data *rdp, unsigned long j)
{
	rcu_lockdep_assert_cblist_protected(rdp);
	if (!rcu_segcblist_is_offloaded(&rdp->cblist) ||
	    !rcu_nocb_bypass_trylock(rdp))
		return;
	WARN_ON_ONCE(!rcu_nocb_do_flush_bypass(rdp, NULL, j));
}

/*
 * See whether it is appropriate to use the ->nocb_bypass list in order
 * to control contention on ->nocb_lock.  A limited number of direct
 * enqueues are permitted into ->cblist per jiffy.  If ->nocb_bypass
 * is non-empty, further callbacks must be placed into ->nocb_bypass,
 * otherwise rcu_barrier() breaks.  Use rcu_nocb_flush_bypass() to switch
 * back to direct use of ->cblist.  However, ->nocb_bypass should not be
 * used if ->cblist is empty, because otherwise callbacks can be stranded
 * on ->nocb_bypass because we cannot count on the current CPU ever again
 * invoking call_rcu().  The general rule is that if ->nocb_bypass is
 * non-empty, the corresponding no-CBs grace-period kthread must not be
 * in an indefinite sleep state.
 *
 * Finally, it is not permitted to use the bypass during early boot,
 * as doing so would confuse the auto-initialization code.  Besides
 * which, there is no point in worrying about lock contention while
 * there is only one CPU in operation.
 */
static bool rcu_nocb_try_bypass(struct rcu_data *rdp, struct rcu_head *rhp,
				bool *was_alldone, unsigned long flags)
{
	unsigned long c;
	unsigned long cur_gp_seq;
	unsigned long j = jiffies;
	long ncbs = rcu_cblist_n_cbs(&rdp->nocb_bypass);

	if (!rcu_segcblist_is_offloaded(&rdp->cblist)) {
		*was_alldone = !rcu_segcblist_pend_cbs(&rdp->cblist);
		return false; /* Not offloaded, no bypassing. */
	}
	lockdep_assert_irqs_disabled();

	// Don't use ->nocb_bypass during early boot.
	if (rcu_scheduler_active != RCU_SCHEDULER_RUNNING) {
		rcu_nocb_lock(rdp);
		WARN_ON_ONCE(rcu_cblist_n_cbs(&rdp->nocb_bypass));
		*was_alldone = !rcu_segcblist_pend_cbs(&rdp->cblist);
		return false;
	}

	// If we have advanced to a new jiffy, reset counts to allow
	// moving back from ->nocb_bypass to ->cblist.
	if (j == rdp->nocb_nobypass_last) {
		c = rdp->nocb_nobypass_count + 1;
	} else {
		WRITE_ONCE(rdp->nocb_nobypass_last, j);
		c = rdp->nocb_nobypass_count - nocb_nobypass_lim_per_jiffy;
		if (ULONG_CMP_LT(rdp->nocb_nobypass_count,
				 nocb_nobypass_lim_per_jiffy))
			c = 0;
		else if (c > nocb_nobypass_lim_per_jiffy)
			c = nocb_nobypass_lim_per_jiffy;
	}
	WRITE_ONCE(rdp->nocb_nobypass_count, c);

	// If there hasn't yet been all that many ->cblist enqueues
	// this jiffy, tell the caller to enqueue onto ->cblist.  But flush
	// ->nocb_bypass first.
	if (rdp->nocb_nobypass_count < nocb_nobypass_lim_per_jiffy) {
		rcu_nocb_lock(rdp);
		*was_alldone = !rcu_segcblist_pend_cbs(&rdp->cblist);
		if (*was_alldone)
			trace_rcu_nocb_wake(rcu_state.name, rdp->cpu,
					    TPS("FirstQ"));
		WARN_ON_ONCE(!rcu_nocb_flush_bypass(rdp, NULL, j));
		WARN_ON_ONCE(rcu_cblist_n_cbs(&rdp->nocb_bypass));
		return false; // Caller must enqueue the callback.
	}

	// If ->nocb_bypass has been used too long or is too full,
	// flush ->nocb_bypass to ->cblist.
	if ((ncbs && j != READ_ONCE(rdp->nocb_bypass_first)) ||
	    ncbs >= qhimark) {
		rcu_nocb_lock(rdp);
		if (!rcu_nocb_flush_bypass(rdp, rhp, j)) {
			*was_alldone = !rcu_segcblist_pend_cbs(&rdp->cblist);
			if (*was_alldone)
				trace_rcu_nocb_wake(rcu_state.name, rdp->cpu,
						    TPS("FirstQ"));
			WARN_ON_ONCE(rcu_cblist_n_cbs(&rdp->nocb_bypass));
			return false; // Caller must enqueue the callback.
		}
		if (j != rdp->nocb_gp_adv_time &&
		    rcu_segcblist_nextgp(&rdp->cblist, &cur_gp_seq) &&
		    rcu_seq_done(&rdp->mynode->gp_seq, cur_gp_seq)) {
			rcu_advance_cbs_nowake(rdp->mynode, rdp);
			rdp->nocb_gp_adv_time = j;
		}
		rcu_nocb_unlock_irqrestore(rdp, flags);
		return true; // Callback already enqueued.
	}

	// We need to use the bypass.
	rcu_nocb_wait_contended(rdp);
	rcu_nocb_bypass_lock(rdp);
	ncbs = rcu_cblist_n_cbs(&rdp->nocb_bypass);
	rcu_segcblist_inc_len(&rdp->cblist); /* Must precede enqueue. */
	rcu_cblist_enqueue(&rdp->nocb_bypass, rhp);
	if (!ncbs) {
		WRITE_ONCE(rdp->nocb_bypass_first, j);
		trace_rcu_nocb_wake(rcu_state.name, rdp->cpu, TPS("FirstBQ"));
	}
	rcu_nocb_bypass_unlock(rdp);
	smp_mb(); /* Order enqueue before wake. */
	if (ncbs) {
		local_irq_restore(flags);
	} else {
		// No-CBs GP kthread might be indefinitely asleep, if so, wake.
		rcu_nocb_lock(rdp); // Rare during call_rcu() flood.
		if (!rcu_segcblist_pend_cbs(&rdp->cblist)) {
			trace_rcu_nocb_wake(rcu_state.name, rdp->cpu,
					    TPS("FirstBQwake"));
			__call_rcu_nocb_wake(rdp, true, flags);
		} else {
			trace_rcu_nocb_wake(rcu_state.name, rdp->cpu,
					    TPS("FirstBQnoWake"));
			rcu_nocb_unlock_irqrestore(rdp, flags);
		}
	}
	return true; // Callback already enqueued.
}

1852
/*
1853 1854
 * Awaken the no-CBs grace-period kthead if needed, either due to it
 * legitimately being asleep or due to overload conditions.
1855 1856 1857
 *
 * If warranted, also wake up the kthread servicing this CPUs queues.
 */
1858 1859 1860
static void __call_rcu_nocb_wake(struct rcu_data *rdp, bool was_alldone,
				 unsigned long flags)
				 __releases(rdp->nocb_lock)
1861
{
1862 1863
	unsigned long cur_gp_seq;
	unsigned long j;
1864
	long len;
1865 1866
	struct task_struct *t;

1867
	// If we are being polled or there is no kthread, just leave.
1868
	t = READ_ONCE(rdp->nocb_gp_kthread);
1869
	if (rcu_nocb_poll || !t) {
1870
		trace_rcu_nocb_wake(rcu_state.name, rdp->cpu,
1871
				    TPS("WakeNotPoll"));
1872
		rcu_nocb_unlock_irqrestore(rdp, flags);
1873
		return;
1874
	}
1875 1876 1877
	// Need to actually to a wakeup.
	len = rcu_segcblist_n_cbs(&rdp->cblist);
	if (was_alldone) {
1878
		rdp->qlen_last_fqs_check = len;
1879
		if (!irqs_disabled_flags(flags)) {
1880
			/* ... if queue was empty ... */
1881
			wake_nocb_gp(rdp, false, flags);
1882
			trace_rcu_nocb_wake(rcu_state.name, rdp->cpu,
1883 1884
					    TPS("WakeEmpty"));
		} else {
1885 1886
			wake_nocb_gp_defer(rdp, RCU_NOCB_WAKE,
					   TPS("WakeEmptyIsDeferred"));
1887
			rcu_nocb_unlock_irqrestore(rdp, flags);
1888
		}
1889
	} else if (len > rdp->qlen_last_fqs_check + qhimark) {
1890
		/* ... or if many callbacks queued. */
1891
		rdp->qlen_last_fqs_check = len;
1892 1893 1894 1895
		j = jiffies;
		if (j != rdp->nocb_gp_adv_time &&
		    rcu_segcblist_nextgp(&rdp->cblist, &cur_gp_seq) &&
		    rcu_seq_done(&rdp->mynode->gp_seq, cur_gp_seq)) {
1896
			rcu_advance_cbs_nowake(rdp->mynode, rdp);
1897 1898
			rdp->nocb_gp_adv_time = j;
		}
1899 1900 1901 1902
		smp_mb(); /* Enqueue before timer_pending(). */
		if ((rdp->nocb_cb_sleep ||
		     !rcu_segcblist_ready_cbs(&rdp->cblist)) &&
		    !timer_pending(&rdp->nocb_bypass_timer))
1903 1904 1905
			wake_nocb_gp_defer(rdp, RCU_NOCB_WAKE_FORCE,
					   TPS("WakeOvfIsDeferred"));
		rcu_nocb_unlock_irqrestore(rdp, flags);
1906
	} else {
1907
		trace_rcu_nocb_wake(rcu_state.name, rdp->cpu, TPS("WakeNot"));
1908
		rcu_nocb_unlock_irqrestore(rdp, flags);
1909 1910 1911 1912
	}
	return;
}

1913 1914 1915 1916 1917 1918 1919 1920
/* Wake up the no-CBs GP kthread to flush ->nocb_bypass. */
static void do_nocb_bypass_wakeup_timer(struct timer_list *t)
{
	unsigned long flags;
	struct rcu_data *rdp = from_timer(rdp, t, nocb_bypass_timer);

	trace_rcu_nocb_wake(rcu_state.name, rdp->cpu, TPS("Timer"));
	rcu_nocb_lock_irqsave(rdp, flags);
1921
	smp_mb__after_spinlock(); /* Timer expire before wakeup. */
1922 1923 1924
	__call_rcu_nocb_wake(rdp, true, flags);
}

1925
/*
1926 1927
 * No-CBs GP kthreads come here to wait for additional callbacks to show up
 * or for grace periods to end.
1928
 */
1929
static void nocb_gp_wait(struct rcu_data *my_rdp)
1930
{
1931 1932
	bool bypass = false;
	long bypass_ncbs;
1933 1934
	int __maybe_unused cpu = my_rdp->cpu;
	unsigned long cur_gp_seq;
1935
	unsigned long flags;
1936
	bool gotcbs = false;
1937
	unsigned long j = jiffies;
1938
	bool needwait_gp = false; // This prevents actual uninitialized use.
1939 1940
	bool needwake;
	bool needwake_gp;
1941
	struct rcu_data *rdp;
1942
	struct rcu_node *rnp;
1943
	unsigned long wait_gp_seq = 0; // Suppress "use uninitialized" warning.
1944 1945

	/*
1946 1947 1948
	 * Each pass through the following loop checks for CBs and for the
	 * nearest grace period (if any) to wait for next.  The CB kthreads
	 * and the global grace-period kthread are awakened if needed.
1949
	 */
1950
	for (rdp = my_rdp; rdp; rdp = rdp->nocb_next_cb_rdp) {
1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961
		trace_rcu_nocb_wake(rcu_state.name, rdp->cpu, TPS("Check"));
		rcu_nocb_lock_irqsave(rdp, flags);
		bypass_ncbs = rcu_cblist_n_cbs(&rdp->nocb_bypass);
		if (bypass_ncbs &&
		    (time_after(j, READ_ONCE(rdp->nocb_bypass_first) + 1) ||
		     bypass_ncbs > 2 * qhimark)) {
			// Bypass full or old, so flush it.
			(void)rcu_nocb_try_flush_bypass(rdp, j);
			bypass_ncbs = rcu_cblist_n_cbs(&rdp->nocb_bypass);
		} else if (!bypass_ncbs && rcu_segcblist_empty(&rdp->cblist)) {
			rcu_nocb_unlock_irqrestore(rdp, flags);
1962
			continue; /* No callbacks here, try next. */
1963 1964 1965 1966 1967 1968
		}
		if (bypass_ncbs) {
			trace_rcu_nocb_wake(rcu_state.name, rdp->cpu,
					    TPS("Bypass"));
			bypass = true;
		}
1969
		rnp = rdp->mynode;
1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984
		if (bypass) {  // Avoid race with first bypass CB.
			WRITE_ONCE(my_rdp->nocb_defer_wakeup,
				   RCU_NOCB_WAKE_NOT);
			del_timer(&my_rdp->nocb_timer);
		}
		// Advance callbacks if helpful and low contention.
		needwake_gp = false;
		if (!rcu_segcblist_restempty(&rdp->cblist,
					     RCU_NEXT_READY_TAIL) ||
		    (rcu_segcblist_nextgp(&rdp->cblist, &cur_gp_seq) &&
		     rcu_seq_done(&rnp->gp_seq, cur_gp_seq))) {
			raw_spin_lock_rcu_node(rnp); /* irqs disabled. */
			needwake_gp = rcu_advance_cbs(rnp, rdp);
			raw_spin_unlock_rcu_node(rnp); /* irqs disabled. */
		}
1985
		// Need to wait on some grace period?
1986 1987
		WARN_ON_ONCE(!rcu_segcblist_restempty(&rdp->cblist,
						      RCU_NEXT_READY_TAIL));
1988 1989 1990 1991 1992
		if (rcu_segcblist_nextgp(&rdp->cblist, &cur_gp_seq)) {
			if (!needwait_gp ||
			    ULONG_CMP_LT(cur_gp_seq, wait_gp_seq))
				wait_gp_seq = cur_gp_seq;
			needwait_gp = true;
1993 1994
			trace_rcu_nocb_wake(rcu_state.name, rdp->cpu,
					    TPS("NeedWaitGP"));
1995
		}
1996 1997 1998 1999 2000 2001
		if (rcu_segcblist_ready_cbs(&rdp->cblist)) {
			needwake = rdp->nocb_cb_sleep;
			WRITE_ONCE(rdp->nocb_cb_sleep, false);
			smp_mb(); /* CB invocation -after- GP end. */
		} else {
			needwake = false;
2002
		}
2003
		rcu_nocb_unlock_irqrestore(rdp, flags);
2004
		if (needwake) {
2005
			swake_up_one(&rdp->nocb_cb_wq);
2006
			gotcbs = true;
2007
		}
2008 2009 2010 2011
		if (needwake_gp)
			rcu_gp_kthread_wake();
	}

2012 2013 2014
	my_rdp->nocb_gp_bypass = bypass;
	my_rdp->nocb_gp_gp = needwait_gp;
	my_rdp->nocb_gp_seq = needwait_gp ? wait_gp_seq : 0;
2015 2016 2017 2018 2019 2020 2021
	if (bypass && !rcu_nocb_poll) {
		// At least one child with non-empty ->nocb_bypass, so set
		// timer in order to avoid stranding its callbacks.
		raw_spin_lock_irqsave(&my_rdp->nocb_gp_lock, flags);
		mod_timer(&my_rdp->nocb_bypass_timer, j + 2);
		raw_spin_unlock_irqrestore(&my_rdp->nocb_gp_lock, flags);
	}
2022 2023 2024 2025 2026 2027 2028 2029 2030 2031
	if (rcu_nocb_poll) {
		/* Polling, so trace if first poll in the series. */
		if (gotcbs)
			trace_rcu_nocb_wake(rcu_state.name, cpu, TPS("Poll"));
		schedule_timeout_interruptible(1);
	} else if (!needwait_gp) {
		/* Wait for callbacks to appear. */
		trace_rcu_nocb_wake(rcu_state.name, cpu, TPS("Sleep"));
		swait_event_interruptible_exclusive(my_rdp->nocb_gp_wq,
				!READ_ONCE(my_rdp->nocb_gp_sleep));
2032
		trace_rcu_nocb_wake(rcu_state.name, cpu, TPS("EndSleep"));
2033 2034 2035 2036 2037 2038 2039 2040 2041 2042
	} else {
		rnp = my_rdp->mynode;
		trace_rcu_this_gp(rnp, my_rdp, wait_gp_seq, TPS("StartWait"));
		swait_event_interruptible_exclusive(
			rnp->nocb_gp_wq[rcu_seq_ctr(wait_gp_seq) & 0x1],
			rcu_seq_done(&rnp->gp_seq, wait_gp_seq) ||
			!READ_ONCE(my_rdp->nocb_gp_sleep));
		trace_rcu_this_gp(rnp, my_rdp, wait_gp_seq, TPS("EndWait"));
	}
	if (!rcu_nocb_poll) {
2043
		raw_spin_lock_irqsave(&my_rdp->nocb_gp_lock, flags);
2044 2045
		if (bypass)
			del_timer(&my_rdp->nocb_bypass_timer);
2046
		WRITE_ONCE(my_rdp->nocb_gp_sleep, true);
2047
		raw_spin_unlock_irqrestore(&my_rdp->nocb_gp_lock, flags);
2048
	}
2049
	my_rdp->nocb_gp_seq = -1;
2050
	WARN_ON(signal_pending(current));
2051
}
2052

2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064
/*
 * No-CBs grace-period-wait kthread.  There is one of these per group
 * of CPUs, but only once at least one CPU in that group has come online
 * at least once since boot.  This kthread checks for newly posted
 * callbacks from any of the CPUs it is responsible for, waits for a
 * grace period, then awakens all of the rcu_nocb_cb_kthread() instances
 * that then have callback-invocation work to do.
 */
static int rcu_nocb_gp_kthread(void *arg)
{
	struct rcu_data *rdp = arg;

2065
	for (;;) {
2066
		WRITE_ONCE(rdp->nocb_gp_loops, rdp->nocb_gp_loops + 1);
2067
		nocb_gp_wait(rdp);
2068 2069
		cond_resched_tasks_rcu_qs();
	}
2070
	return 0;
2071 2072 2073
}

/*
2074 2075
 * Invoke any ready callbacks from the corresponding no-CBs CPU,
 * then, if there are no more, wait for more to appear.
2076
 */
2077
static void nocb_cb_wait(struct rcu_data *rdp)
2078
{
2079
	unsigned long cur_gp_seq;
2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090
	unsigned long flags;
	bool needwake_gp = false;
	struct rcu_node *rnp = rdp->mynode;

	local_irq_save(flags);
	rcu_momentary_dyntick_idle();
	local_irq_restore(flags);
	local_bh_disable();
	rcu_do_batch(rdp);
	local_bh_enable();
	lockdep_assert_irqs_enabled();
2091
	rcu_nocb_lock_irqsave(rdp, flags);
2092 2093 2094
	if (rcu_segcblist_nextgp(&rdp->cblist, &cur_gp_seq) &&
	    rcu_seq_done(&rnp->gp_seq, cur_gp_seq) &&
	    raw_spin_trylock_rcu_node(rnp)) { /* irqs already disabled. */
2095 2096 2097
		needwake_gp = rcu_advance_cbs(rdp->mynode, rdp);
		raw_spin_unlock_rcu_node(rnp); /* irqs remain disabled. */
	}
2098
	if (rcu_segcblist_ready_cbs(&rdp->cblist)) {
2099
		rcu_nocb_unlock_irqrestore(rdp, flags);
2100 2101 2102 2103 2104
		if (needwake_gp)
			rcu_gp_kthread_wake();
		return;
	}

2105
	trace_rcu_nocb_wake(rcu_state.name, rdp->cpu, TPS("CBSleep"));
2106
	WRITE_ONCE(rdp->nocb_cb_sleep, true);
2107
	rcu_nocb_unlock_irqrestore(rdp, flags);
2108 2109
	if (needwake_gp)
		rcu_gp_kthread_wake();
2110
	swait_event_interruptible_exclusive(rdp->nocb_cb_wq,
2111 2112 2113 2114
				 !READ_ONCE(rdp->nocb_cb_sleep));
	if (!smp_load_acquire(&rdp->nocb_cb_sleep)) { /* VVV */
		/* ^^^ Ensure CB invocation follows _sleep test. */
		return;
2115
	}
2116 2117
	WARN_ON(signal_pending(current));
	trace_rcu_nocb_wake(rcu_state.name, rdp->cpu, TPS("WokeEmpty"));
2118 2119
}

2120
/*
2121 2122
 * Per-rcu_data kthread, but only for no-CBs CPUs.  Repeatedly invoke
 * nocb_cb_wait() to do the dirty work.
2123
 */
2124
static int rcu_nocb_cb_kthread(void *arg)
2125 2126 2127
{
	struct rcu_data *rdp = arg;

2128 2129
	// Each pass through this loop does one callback batch, and,
	// if there are no more ready callbacks, waits for them.
2130
	for (;;) {
2131 2132
		nocb_cb_wait(rdp);
		cond_resched_tasks_rcu_qs();
2133 2134 2135 2136
	}
	return 0;
}

2137
/* Is a deferred wakeup of rcu_nocb_kthread() required? */
2138
static int rcu_nocb_need_deferred_wakeup(struct rcu_data *rdp)
2139
{
2140
	return READ_ONCE(rdp->nocb_defer_wakeup);
2141 2142 2143
}

/* Do a deferred wakeup of rcu_nocb_kthread(). */
2144
static void do_nocb_deferred_wakeup_common(struct rcu_data *rdp)
2145
{
2146
	unsigned long flags;
2147 2148
	int ndw;

2149
	rcu_nocb_lock_irqsave(rdp, flags);
2150
	if (!rcu_nocb_need_deferred_wakeup(rdp)) {
2151
		rcu_nocb_unlock_irqrestore(rdp, flags);
2152
		return;
2153
	}
2154
	ndw = READ_ONCE(rdp->nocb_defer_wakeup);
2155
	WRITE_ONCE(rdp->nocb_defer_wakeup, RCU_NOCB_WAKE_NOT);
2156
	wake_nocb_gp(rdp, ndw == RCU_NOCB_WAKE_FORCE, flags);
2157
	trace_rcu_nocb_wake(rcu_state.name, rdp->cpu, TPS("DeferredWake"));
2158 2159
}

2160
/* Do a deferred wakeup of rcu_nocb_kthread() from a timer handler. */
2161
static void do_nocb_deferred_wakeup_timer(struct timer_list *t)
2162
{
2163 2164 2165
	struct rcu_data *rdp = from_timer(rdp, t, nocb_timer);

	do_nocb_deferred_wakeup_common(rdp);
2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178
}

/*
 * Do a deferred wakeup of rcu_nocb_kthread() from fastpath.
 * This means we do an inexact common-case check.  Note that if
 * we miss, ->nocb_timer will eventually clean things up.
 */
static void do_nocb_deferred_wakeup(struct rcu_data *rdp)
{
	if (rcu_nocb_need_deferred_wakeup(rdp))
		do_nocb_deferred_wakeup_common(rdp);
}

2179 2180 2181
void __init rcu_init_nohz(void)
{
	int cpu;
2182
	bool need_rcu_nocb_mask = false;
2183
	struct rcu_data *rdp;
2184 2185 2186 2187 2188 2189

#if defined(CONFIG_NO_HZ_FULL)
	if (tick_nohz_full_running && cpumask_weight(tick_nohz_full_mask))
		need_rcu_nocb_mask = true;
#endif /* #if defined(CONFIG_NO_HZ_FULL) */

2190
	if (!cpumask_available(rcu_nocb_mask) && need_rcu_nocb_mask) {
2191 2192 2193 2194
		if (!zalloc_cpumask_var(&rcu_nocb_mask, GFP_KERNEL)) {
			pr_info("rcu_nocb_mask allocation failed, callback offloading disabled.\n");
			return;
		}
2195
	}
2196
	if (!cpumask_available(rcu_nocb_mask))
2197 2198 2199 2200 2201 2202 2203 2204
		return;

#if defined(CONFIG_NO_HZ_FULL)
	if (tick_nohz_full_running)
		cpumask_or(rcu_nocb_mask, rcu_nocb_mask, tick_nohz_full_mask);
#endif /* #if defined(CONFIG_NO_HZ_FULL) */

	if (!cpumask_subset(rcu_nocb_mask, cpu_possible_mask)) {
2205
		pr_info("\tNote: kernel parameter 'rcu_nocbs=', 'nohz_full', or 'isolcpus=' contains nonexistent CPUs.\n");
2206 2207 2208
		cpumask_and(rcu_nocb_mask, cpu_possible_mask,
			    rcu_nocb_mask);
	}
2209 2210 2211 2212 2213
	if (cpumask_empty(rcu_nocb_mask))
		pr_info("\tOffload RCU callbacks from CPUs: (none).\n");
	else
		pr_info("\tOffload RCU callbacks from CPUs: %*pbl.\n",
			cpumask_pr_args(rcu_nocb_mask));
2214 2215 2216
	if (rcu_nocb_poll)
		pr_info("\tPoll for callbacks from no-CBs CPUs.\n");

2217 2218 2219 2220 2221 2222
	for_each_cpu(cpu, rcu_nocb_mask) {
		rdp = per_cpu_ptr(&rcu_data, cpu);
		if (rcu_segcblist_empty(&rdp->cblist))
			rcu_segcblist_init(&rdp->cblist);
		rcu_segcblist_offload(&rdp->cblist);
	}
2223
	rcu_organize_nocb_kthreads();
2224 2225
}

2226 2227 2228
/* Initialize per-rcu_data variables for no-CBs CPUs. */
static void __init rcu_boot_init_nocb_percpu_data(struct rcu_data *rdp)
{
2229 2230
	init_swait_queue_head(&rdp->nocb_cb_wq);
	init_swait_queue_head(&rdp->nocb_gp_wq);
2231
	raw_spin_lock_init(&rdp->nocb_lock);
2232
	raw_spin_lock_init(&rdp->nocb_bypass_lock);
2233
	raw_spin_lock_init(&rdp->nocb_gp_lock);
2234
	timer_setup(&rdp->nocb_timer, do_nocb_deferred_wakeup_timer, 0);
2235 2236
	timer_setup(&rdp->nocb_bypass_timer, do_nocb_bypass_wakeup_timer, 0);
	rcu_cblist_init(&rdp->nocb_bypass);
2237 2238
}

2239 2240
/*
 * If the specified CPU is a no-CBs CPU that does not already have its
2241 2242
 * rcuo CB kthread, spawn it.  Additionally, if the rcuo GP kthread
 * for this CPU's group has not yet been created, spawn it as well.
2243
 */
2244
static void rcu_spawn_one_nocb_kthread(int cpu)
2245
{
2246 2247
	struct rcu_data *rdp = per_cpu_ptr(&rcu_data, cpu);
	struct rcu_data *rdp_gp;
2248 2249 2250 2251 2252 2253
	struct task_struct *t;

	/*
	 * If this isn't a no-CBs CPU or if it already has an rcuo kthread,
	 * then nothing to do.
	 */
2254
	if (!rcu_is_nocb_cpu(cpu) || rdp->nocb_cb_kthread)
2255 2256
		return;

2257
	/* If we didn't spawn the GP kthread first, reorganize! */
2258 2259 2260 2261 2262 2263 2264
	rdp_gp = rdp->nocb_gp_rdp;
	if (!rdp_gp->nocb_gp_kthread) {
		t = kthread_run(rcu_nocb_gp_kthread, rdp_gp,
				"rcuog/%d", rdp_gp->cpu);
		if (WARN_ONCE(IS_ERR(t), "%s: Could not start rcuo GP kthread, OOM is now expected behavior\n", __func__))
			return;
		WRITE_ONCE(rdp_gp->nocb_gp_kthread, t);
2265 2266
	}

2267
	/* Spawn the kthread for this CPU. */
2268
	t = kthread_run(rcu_nocb_cb_kthread, rdp,
2269
			"rcuo%c/%d", rcu_state.abbr, cpu);
2270
	if (WARN_ONCE(IS_ERR(t), "%s: Could not start rcuo CB kthread, OOM is now expected behavior\n", __func__))
2271
		return;
2272 2273
	WRITE_ONCE(rdp->nocb_cb_kthread, t);
	WRITE_ONCE(rdp->nocb_gp_kthread, rdp_gp->nocb_gp_kthread);
2274 2275 2276 2277
}

/*
 * If the specified CPU is a no-CBs CPU that does not already have its
2278
 * rcuo kthread, spawn it.
2279
 */
2280
static void rcu_spawn_cpu_nocb_kthread(int cpu)
2281 2282
{
	if (rcu_scheduler_fully_active)
2283
		rcu_spawn_one_nocb_kthread(cpu);
2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296
}

/*
 * Once the scheduler is running, spawn rcuo kthreads for all online
 * no-CBs CPUs.  This assumes that the early_initcall()s happen before
 * non-boot CPUs come online -- if this changes, we will need to add
 * some mutual exclusion.
 */
static void __init rcu_spawn_nocb_kthreads(void)
{
	int cpu;

	for_each_online_cpu(cpu)
2297
		rcu_spawn_cpu_nocb_kthread(cpu);
2298 2299
}

2300
/* How many CB CPU IDs per GP kthread?  Default of -1 for sqrt(nr_cpu_ids). */
2301 2302
static int rcu_nocb_gp_stride = -1;
module_param(rcu_nocb_gp_stride, int, 0444);
2303 2304

/*
2305
 * Initialize GP-CB relationships for all no-CBs CPU.
2306
 */
2307
static void __init rcu_organize_nocb_kthreads(void)
2308 2309
{
	int cpu;
2310
	bool firsttime = true;
2311 2312
	bool gotnocbs = false;
	bool gotnocbscbs = true;
2313
	int ls = rcu_nocb_gp_stride;
2314
	int nl = 0;  /* Next GP kthread. */
2315
	struct rcu_data *rdp;
2316
	struct rcu_data *rdp_gp = NULL;  /* Suppress misguided gcc warn. */
2317
	struct rcu_data *rdp_prev = NULL;
2318

2319
	if (!cpumask_available(rcu_nocb_mask))
2320
		return;
2321
	if (ls == -1) {
2322
		ls = nr_cpu_ids / int_sqrt(nr_cpu_ids);
2323
		rcu_nocb_gp_stride = ls;
2324 2325 2326
	}

	/*
2327 2328 2329
	 * Each pass through this loop sets up one rcu_data structure.
	 * Should the corresponding CPU come online in the future, then
	 * we will spawn the needed set of rcu_nocb_kthread() kthreads.
2330
	 */
2331
	for_each_cpu(cpu, rcu_nocb_mask) {
2332
		rdp = per_cpu_ptr(&rcu_data, cpu);
2333
		if (rdp->cpu >= nl) {
2334
			/* New GP kthread, set up for CBs & next GP. */
2335
			gotnocbs = true;
2336
			nl = DIV_ROUND_UP(rdp->cpu + 1, ls) * ls;
2337
			rdp->nocb_gp_rdp = rdp;
2338
			rdp_gp = rdp;
2339 2340 2341 2342 2343 2344 2345 2346 2347
			if (dump_tree) {
				if (!firsttime)
					pr_cont("%s\n", gotnocbscbs
							? "" : " (self only)");
				gotnocbscbs = false;
				firsttime = false;
				pr_alert("%s: No-CB GP kthread CPU %d:",
					 __func__, cpu);
			}
2348
		} else {
2349
			/* Another CB kthread, link to previous GP kthread. */
2350
			gotnocbscbs = true;
2351
			rdp->nocb_gp_rdp = rdp_gp;
2352
			rdp_prev->nocb_next_cb_rdp = rdp;
2353 2354
			if (dump_tree)
				pr_cont(" %d", cpu);
2355 2356
		}
		rdp_prev = rdp;
2357
	}
2358 2359
	if (gotnocbs && dump_tree)
		pr_cont("%s\n", gotnocbscbs ? "" : " (self only)");
2360 2361
}

2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372
/*
 * Bind the current task to the offloaded CPUs.  If there are no offloaded
 * CPUs, leave the task unbound.  Splat if the bind attempt fails.
 */
void rcu_bind_current_to_nocb(void)
{
	if (cpumask_available(rcu_nocb_mask) && cpumask_weight(rcu_nocb_mask))
		WARN_ON(sched_setaffinity(current->pid, rcu_nocb_mask));
}
EXPORT_SYMBOL_GPL(rcu_bind_current_to_nocb);

2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445
/*
 * Dump out nocb grace-period kthread state for the specified rcu_data
 * structure.
 */
static void show_rcu_nocb_gp_state(struct rcu_data *rdp)
{
	struct rcu_node *rnp = rdp->mynode;

	pr_info("nocb GP %d %c%c%c%c%c%c %c[%c%c] %c%c:%ld rnp %d:%d %lu\n",
		rdp->cpu,
		"kK"[!!rdp->nocb_gp_kthread],
		"lL"[raw_spin_is_locked(&rdp->nocb_gp_lock)],
		"dD"[!!rdp->nocb_defer_wakeup],
		"tT"[timer_pending(&rdp->nocb_timer)],
		"bB"[timer_pending(&rdp->nocb_bypass_timer)],
		"sS"[!!rdp->nocb_gp_sleep],
		".W"[swait_active(&rdp->nocb_gp_wq)],
		".W"[swait_active(&rnp->nocb_gp_wq[0])],
		".W"[swait_active(&rnp->nocb_gp_wq[1])],
		".B"[!!rdp->nocb_gp_bypass],
		".G"[!!rdp->nocb_gp_gp],
		(long)rdp->nocb_gp_seq,
		rnp->grplo, rnp->grphi, READ_ONCE(rdp->nocb_gp_loops));
}

/* Dump out nocb kthread state for the specified rcu_data structure. */
static void show_rcu_nocb_state(struct rcu_data *rdp)
{
	struct rcu_segcblist *rsclp = &rdp->cblist;
	bool waslocked;
	bool wastimer;
	bool wassleep;

	if (rdp->nocb_gp_rdp == rdp)
		show_rcu_nocb_gp_state(rdp);

	pr_info("   CB %d->%d %c%c%c%c%c%c F%ld L%ld C%d %c%c%c%c%c q%ld\n",
		rdp->cpu, rdp->nocb_gp_rdp->cpu,
		"kK"[!!rdp->nocb_cb_kthread],
		"bB"[raw_spin_is_locked(&rdp->nocb_bypass_lock)],
		"cC"[!!atomic_read(&rdp->nocb_lock_contended)],
		"lL"[raw_spin_is_locked(&rdp->nocb_lock)],
		"sS"[!!rdp->nocb_cb_sleep],
		".W"[swait_active(&rdp->nocb_cb_wq)],
		jiffies - rdp->nocb_bypass_first,
		jiffies - rdp->nocb_nobypass_last,
		rdp->nocb_nobypass_count,
		".D"[rcu_segcblist_ready_cbs(rsclp)],
		".W"[!rcu_segcblist_restempty(rsclp, RCU_DONE_TAIL)],
		".R"[!rcu_segcblist_restempty(rsclp, RCU_WAIT_TAIL)],
		".N"[!rcu_segcblist_restempty(rsclp, RCU_NEXT_READY_TAIL)],
		".B"[!!rcu_cblist_n_cbs(&rdp->nocb_bypass)],
		rcu_segcblist_n_cbs(&rdp->cblist));

	/* It is OK for GP kthreads to have GP state. */
	if (rdp->nocb_gp_rdp == rdp)
		return;

	waslocked = raw_spin_is_locked(&rdp->nocb_gp_lock);
	wastimer = timer_pending(&rdp->nocb_timer);
	wassleep = swait_active(&rdp->nocb_gp_wq);
	if (!rdp->nocb_defer_wakeup && !rdp->nocb_gp_sleep &&
	    !waslocked && !wastimer && !wassleep)
		return;  /* Nothing untowards. */

	pr_info("   !!! %c%c%c%c %c\n",
		"lL"[waslocked],
		"dD"[!!rdp->nocb_defer_wakeup],
		"tT"[wastimer],
		"sS"[!!rdp->nocb_gp_sleep],
		".W"[wassleep]);
}

2446 2447
#else /* #ifdef CONFIG_RCU_NOCB_CPU */

2448 2449
/* No ->nocb_lock to acquire.  */
static void rcu_nocb_lock(struct rcu_data *rdp)
2450
{
2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462
}

/* No ->nocb_lock to release.  */
static void rcu_nocb_unlock(struct rcu_data *rdp)
{
}

/* No ->nocb_lock to release.  */
static void rcu_nocb_unlock_irqrestore(struct rcu_data *rdp,
				       unsigned long flags)
{
	local_irq_restore(flags);
2463 2464
}

2465 2466 2467 2468 2469 2470
/* Lockdep check that ->cblist may be safely accessed. */
static void rcu_lockdep_assert_cblist_protected(struct rcu_data *rdp)
{
	lockdep_assert_irqs_disabled();
}

2471
static void rcu_nocb_gp_cleanup(struct swait_queue_head *sq)
2472 2473 2474
{
}

2475
static struct swait_queue_head *rcu_nocb_gp_get(struct rcu_node *rnp)
2476 2477 2478 2479
{
	return NULL;
}

2480 2481 2482
static void rcu_init_one_nocb(struct rcu_node *rnp)
{
}
2483

2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495
static bool rcu_nocb_flush_bypass(struct rcu_data *rdp, struct rcu_head *rhp,
				  unsigned long j)
{
	return true;
}

static bool rcu_nocb_try_bypass(struct rcu_data *rdp, struct rcu_head *rhp,
				bool *was_alldone, unsigned long flags)
{
	return false;
}

2496 2497
static void __call_rcu_nocb_wake(struct rcu_data *rdp, bool was_empty,
				 unsigned long flags)
2498
{
2499
	WARN_ON_ONCE(1);  /* Should be dead code! */
2500 2501 2502 2503 2504 2505
}

static void __init rcu_boot_init_nocb_percpu_data(struct rcu_data *rdp)
{
}

2506
static int rcu_nocb_need_deferred_wakeup(struct rcu_data *rdp)
2507 2508 2509 2510 2511 2512 2513 2514
{
	return false;
}

static void do_nocb_deferred_wakeup(struct rcu_data *rdp)
{
}

2515
static void rcu_spawn_cpu_nocb_kthread(int cpu)
2516 2517 2518 2519
{
}

static void __init rcu_spawn_nocb_kthreads(void)
2520 2521 2522
{
}

2523 2524 2525 2526
static void show_rcu_nocb_state(struct rcu_data *rdp)
{
}

2527
#endif /* #else #ifdef CONFIG_RCU_NOCB_CPU */
2528

2529 2530 2531 2532 2533 2534 2535
/*
 * Is this CPU a NO_HZ_FULL CPU that should ignore RCU so that the
 * grace-period kthread will do force_quiescent_state() processing?
 * The idea is to avoid waking up RCU core processing on such a
 * CPU unless the grace period has extended for too long.
 *
 * This code relies on the fact that all NO_HZ_FULL CPUs are also
2536
 * CONFIG_RCU_NOCB_CPU CPUs.
2537
 */
2538
static bool rcu_nohz_full_cpu(void)
2539 2540 2541
{
#ifdef CONFIG_NO_HZ_FULL
	if (tick_nohz_full_cpu(smp_processor_id()) &&
2542
	    (!rcu_gp_in_progress() ||
2543
	     ULONG_CMP_LT(jiffies, READ_ONCE(rcu_state.gp_start) + HZ)))
2544
		return true;
2545
#endif /* #ifdef CONFIG_NO_HZ_FULL */
2546
	return false;
2547
}
2548 2549

/*
2550
 * Bind the RCU grace-period kthreads to the housekeeping CPU.
2551 2552 2553
 */
static void rcu_bind_gp_kthread(void)
{
2554
	if (!tick_nohz_full_enabled())
2555
		return;
2556
	housekeeping_affine(current, HK_FLAG_RCU);
2557
}
2558 2559 2560 2561 2562

/* Record the current task on dyntick-idle entry. */
static void rcu_dynticks_task_enter(void)
{
#if defined(CONFIG_TASKS_RCU) && defined(CONFIG_NO_HZ_FULL)
2563
	WRITE_ONCE(current->rcu_tasks_idle_cpu, smp_processor_id());
2564 2565 2566 2567 2568 2569 2570
#endif /* #if defined(CONFIG_TASKS_RCU) && defined(CONFIG_NO_HZ_FULL) */
}

/* Record no current task on dyntick-idle exit. */
static void rcu_dynticks_task_exit(void)
{
#if defined(CONFIG_TASKS_RCU) && defined(CONFIG_NO_HZ_FULL)
2571
	WRITE_ONCE(current->rcu_tasks_idle_cpu, -1);
2572 2573
#endif /* #if defined(CONFIG_TASKS_RCU) && defined(CONFIG_NO_HZ_FULL) */
}