manage.c 43.4 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3
/*
 * linux/kernel/irq/manage.c
 *
I
Ingo Molnar 已提交
4 5
 * Copyright (C) 1992, 1998-2006 Linus Torvalds, Ingo Molnar
 * Copyright (C) 2005-2006 Thomas Gleixner
L
Linus Torvalds 已提交
6 7 8 9
 *
 * This file contains driver APIs to the irq subsystem.
 */

10 11
#define pr_fmt(fmt) "genirq: " fmt

L
Linus Torvalds 已提交
12
#include <linux/irq.h>
13
#include <linux/kthread.h>
L
Linus Torvalds 已提交
14 15 16
#include <linux/module.h>
#include <linux/random.h>
#include <linux/interrupt.h>
17
#include <linux/slab.h>
18
#include <linux/sched.h>
19
#include <linux/task_work.h>
L
Linus Torvalds 已提交
20 21 22

#include "internals.h"

23 24 25 26 27 28 29 30 31 32 33
#ifdef CONFIG_IRQ_FORCED_THREADING
__read_mostly bool force_irqthreads;

static int __init setup_forced_irqthreads(char *arg)
{
	force_irqthreads = true;
	return 0;
}
early_param("threadirqs", setup_forced_irqthreads);
#endif

L
Linus Torvalds 已提交
34 35
/**
 *	synchronize_irq - wait for pending IRQ handlers (on other CPUs)
36
 *	@irq: interrupt number to wait for
L
Linus Torvalds 已提交
37 38 39 40 41 42 43 44 45
 *
 *	This function waits for any pending IRQ handlers for this interrupt
 *	to complete before returning. If you use this function while
 *	holding a resource the IRQ handler may need you will deadlock.
 *
 *	This function may be called - with care - from IRQ context.
 */
void synchronize_irq(unsigned int irq)
{
46
	struct irq_desc *desc = irq_to_desc(irq);
47
	bool inprogress;
L
Linus Torvalds 已提交
48

49
	if (!desc)
50 51
		return;

52 53 54 55 56 57 58
	do {
		unsigned long flags;

		/*
		 * Wait until we're out of the critical section.  This might
		 * give the wrong answer due to the lack of memory barriers.
		 */
59
		while (irqd_irq_inprogress(&desc->irq_data))
60 61 62
			cpu_relax();

		/* Ok, that indicated we're done: double-check carefully. */
63
		raw_spin_lock_irqsave(&desc->lock, flags);
64
		inprogress = irqd_irq_inprogress(&desc->irq_data);
65
		raw_spin_unlock_irqrestore(&desc->lock, flags);
66 67

		/* Oops, that failed? */
68
	} while (inprogress);
69 70 71 72 73 74

	/*
	 * We made sure that no hardirq handler is running. Now verify
	 * that no threaded handlers are active.
	 */
	wait_event(desc->wait_for_threads, !atomic_read(&desc->threads_active));
L
Linus Torvalds 已提交
75 76 77
}
EXPORT_SYMBOL(synchronize_irq);

78 79 80
#ifdef CONFIG_SMP
cpumask_var_t irq_default_affinity;

81 82 83 84 85 86 87
/**
 *	irq_can_set_affinity - Check if the affinity of a given irq can be set
 *	@irq:		Interrupt to check
 *
 */
int irq_can_set_affinity(unsigned int irq)
{
88
	struct irq_desc *desc = irq_to_desc(irq);
89

90 91
	if (!desc || !irqd_can_balance(&desc->irq_data) ||
	    !desc->irq_data.chip || !desc->irq_data.chip->irq_set_affinity)
92 93 94 95 96
		return 0;

	return 1;
}

97 98 99 100 101 102 103 104 105 106
/**
 *	irq_set_thread_affinity - Notify irq threads to adjust affinity
 *	@desc:		irq descriptor which has affitnity changed
 *
 *	We just set IRQTF_AFFINITY and delegate the affinity setting
 *	to the interrupt thread itself. We can not call
 *	set_cpus_allowed_ptr() here as we hold desc->lock and this
 *	code can be called from hard interrupt context.
 */
void irq_set_thread_affinity(struct irq_desc *desc)
107 108 109 110 111
{
	struct irqaction *action = desc->action;

	while (action) {
		if (action->thread)
112
			set_bit(IRQTF_AFFINITY, &action->thread_flags);
113 114 115 116
		action = action->next;
	}
}

117
#ifdef CONFIG_GENERIC_PENDING_IRQ
T
Thomas Gleixner 已提交
118
static inline bool irq_can_move_pcntxt(struct irq_data *data)
119
{
T
Thomas Gleixner 已提交
120
	return irqd_can_move_in_process_context(data);
121
}
T
Thomas Gleixner 已提交
122
static inline bool irq_move_pending(struct irq_data *data)
123
{
T
Thomas Gleixner 已提交
124
	return irqd_is_setaffinity_pending(data);
125 126 127 128 129 130 131 132 133 134 135 136
}
static inline void
irq_copy_pending(struct irq_desc *desc, const struct cpumask *mask)
{
	cpumask_copy(desc->pending_mask, mask);
}
static inline void
irq_get_pending(struct cpumask *mask, struct irq_desc *desc)
{
	cpumask_copy(mask, desc->pending_mask);
}
#else
T
Thomas Gleixner 已提交
137
static inline bool irq_can_move_pcntxt(struct irq_data *data) { return true; }
T
Thomas Gleixner 已提交
138
static inline bool irq_move_pending(struct irq_data *data) { return false; }
139 140 141 142 143 144
static inline void
irq_copy_pending(struct irq_desc *desc, const struct cpumask *mask) { }
static inline void
irq_get_pending(struct cpumask *mask, struct irq_desc *desc) { }
#endif

145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163
int irq_do_set_affinity(struct irq_data *data, const struct cpumask *mask,
			bool force)
{
	struct irq_desc *desc = irq_data_to_desc(data);
	struct irq_chip *chip = irq_data_get_irq_chip(data);
	int ret;

	ret = chip->irq_set_affinity(data, mask, false);
	switch (ret) {
	case IRQ_SET_MASK_OK:
		cpumask_copy(data->affinity, mask);
	case IRQ_SET_MASK_OK_NOCOPY:
		irq_set_thread_affinity(desc);
		ret = 0;
	}

	return ret;
}

164
int __irq_set_affinity_locked(struct irq_data *data, const struct cpumask *mask)
165
{
166 167
	struct irq_chip *chip = irq_data_get_irq_chip(data);
	struct irq_desc *desc = irq_data_to_desc(data);
168
	int ret = 0;
169

170
	if (!chip || !chip->irq_set_affinity)
171 172
		return -EINVAL;

T
Thomas Gleixner 已提交
173
	if (irq_can_move_pcntxt(data)) {
174
		ret = irq_do_set_affinity(data, mask, false);
175
	} else {
176
		irqd_set_move_pending(data);
177
		irq_copy_pending(desc, mask);
178
	}
179

180 181 182 183
	if (desc->affinity_notify) {
		kref_get(&desc->affinity_notify->kref);
		schedule_work(&desc->affinity_notify->work);
	}
184 185 186 187 188 189 190 191
	irqd_set(data, IRQD_AFFINITY_SET);

	return ret;
}

/**
 *	irq_set_affinity - Set the irq affinity of a given irq
 *	@irq:		Interrupt to set affinity
192
 *	@mask:		cpumask
193 194 195 196 197 198 199 200 201 202 203 204 205
 *
 */
int irq_set_affinity(unsigned int irq, const struct cpumask *mask)
{
	struct irq_desc *desc = irq_to_desc(irq);
	unsigned long flags;
	int ret;

	if (!desc)
		return -EINVAL;

	raw_spin_lock_irqsave(&desc->lock, flags);
	ret =  __irq_set_affinity_locked(irq_desc_get_irq_data(desc), mask);
206
	raw_spin_unlock_irqrestore(&desc->lock, flags);
207
	return ret;
208 209
}

210 211 212
int irq_set_affinity_hint(unsigned int irq, const struct cpumask *m)
{
	unsigned long flags;
213
	struct irq_desc *desc = irq_get_desc_lock(irq, &flags, IRQ_GET_DESC_CHECK_GLOBAL);
214 215 216 217

	if (!desc)
		return -EINVAL;
	desc->affinity_hint = m;
218
	irq_put_desc_unlock(desc, flags);
219 220 221 222
	return 0;
}
EXPORT_SYMBOL_GPL(irq_set_affinity_hint);

223 224 225 226 227 228 229 230
static void irq_affinity_notify(struct work_struct *work)
{
	struct irq_affinity_notify *notify =
		container_of(work, struct irq_affinity_notify, work);
	struct irq_desc *desc = irq_to_desc(notify->irq);
	cpumask_var_t cpumask;
	unsigned long flags;

231
	if (!desc || !alloc_cpumask_var(&cpumask, GFP_KERNEL))
232 233 234
		goto out;

	raw_spin_lock_irqsave(&desc->lock, flags);
T
Thomas Gleixner 已提交
235
	if (irq_move_pending(&desc->irq_data))
236
		irq_get_pending(cpumask, desc);
237
	else
238
		cpumask_copy(cpumask, desc->irq_data.affinity);
239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290
	raw_spin_unlock_irqrestore(&desc->lock, flags);

	notify->notify(notify, cpumask);

	free_cpumask_var(cpumask);
out:
	kref_put(&notify->kref, notify->release);
}

/**
 *	irq_set_affinity_notifier - control notification of IRQ affinity changes
 *	@irq:		Interrupt for which to enable/disable notification
 *	@notify:	Context for notification, or %NULL to disable
 *			notification.  Function pointers must be initialised;
 *			the other fields will be initialised by this function.
 *
 *	Must be called in process context.  Notification may only be enabled
 *	after the IRQ is allocated and must be disabled before the IRQ is
 *	freed using free_irq().
 */
int
irq_set_affinity_notifier(unsigned int irq, struct irq_affinity_notify *notify)
{
	struct irq_desc *desc = irq_to_desc(irq);
	struct irq_affinity_notify *old_notify;
	unsigned long flags;

	/* The release function is promised process context */
	might_sleep();

	if (!desc)
		return -EINVAL;

	/* Complete initialisation of *notify */
	if (notify) {
		notify->irq = irq;
		kref_init(&notify->kref);
		INIT_WORK(&notify->work, irq_affinity_notify);
	}

	raw_spin_lock_irqsave(&desc->lock, flags);
	old_notify = desc->affinity_notify;
	desc->affinity_notify = notify;
	raw_spin_unlock_irqrestore(&desc->lock, flags);

	if (old_notify)
		kref_put(&old_notify->kref, old_notify->release);

	return 0;
}
EXPORT_SYMBOL_GPL(irq_set_affinity_notifier);

291 292 293 294
#ifndef CONFIG_AUTO_IRQ_AFFINITY
/*
 * Generic version of the affinity autoselector.
 */
295 296
static int
setup_affinity(unsigned int irq, struct irq_desc *desc, struct cpumask *mask)
297
{
298
	struct cpumask *set = irq_default_affinity;
299
	int node = desc->irq_data.node;
300

301
	/* Excludes PER_CPU and NO_BALANCE interrupts */
302 303 304
	if (!irq_can_set_affinity(irq))
		return 0;

305 306 307 308
	/*
	 * Preserve an userspace affinity setup, but make sure that
	 * one of the targets is online.
	 */
309
	if (irqd_has_set(&desc->irq_data, IRQD_AFFINITY_SET)) {
310 311 312
		if (cpumask_intersects(desc->irq_data.affinity,
				       cpu_online_mask))
			set = desc->irq_data.affinity;
T
Thomas Gleixner 已提交
313
		else
314
			irqd_clear(&desc->irq_data, IRQD_AFFINITY_SET);
315
	}
316

317
	cpumask_and(mask, cpu_online_mask, set);
318 319 320 321 322 323 324
	if (node != NUMA_NO_NODE) {
		const struct cpumask *nodemask = cpumask_of_node(node);

		/* make sure at least one of the cpus in nodemask is online */
		if (cpumask_intersects(mask, nodemask))
			cpumask_and(mask, mask, nodemask);
	}
325
	irq_do_set_affinity(&desc->irq_data, mask, false);
326 327
	return 0;
}
328
#else
329 330
static inline int
setup_affinity(unsigned int irq, struct irq_desc *d, struct cpumask *mask)
331 332 333
{
	return irq_select_affinity(irq);
}
334 335
#endif

336 337 338
/*
 * Called when affinity is set via /proc/irq
 */
339
int irq_select_affinity_usr(unsigned int irq, struct cpumask *mask)
340 341 342 343 344
{
	struct irq_desc *desc = irq_to_desc(irq);
	unsigned long flags;
	int ret;

345
	raw_spin_lock_irqsave(&desc->lock, flags);
346
	ret = setup_affinity(irq, desc, mask);
347
	raw_spin_unlock_irqrestore(&desc->lock, flags);
348 349 350 351
	return ret;
}

#else
352 353
static inline int
setup_affinity(unsigned int irq, struct irq_desc *desc, struct cpumask *mask)
354 355 356
{
	return 0;
}
L
Linus Torvalds 已提交
357 358
#endif

359 360 361
void __disable_irq(struct irq_desc *desc, unsigned int irq, bool suspend)
{
	if (suspend) {
362
		if (!desc->action || (desc->action->flags & IRQF_NO_SUSPEND))
363
			return;
364
		desc->istate |= IRQS_SUSPENDED;
365 366
	}

367
	if (!desc->depth++)
368
		irq_disable(desc);
369 370
}

371 372 373
static int __disable_irq_nosync(unsigned int irq)
{
	unsigned long flags;
374
	struct irq_desc *desc = irq_get_desc_buslock(irq, &flags, IRQ_GET_DESC_CHECK_GLOBAL);
375 376 377 378 379 380 381 382

	if (!desc)
		return -EINVAL;
	__disable_irq(desc, irq, false);
	irq_put_desc_busunlock(desc, flags);
	return 0;
}

L
Linus Torvalds 已提交
383 384 385 386 387 388 389 390 391 392 393 394 395
/**
 *	disable_irq_nosync - disable an irq without waiting
 *	@irq: Interrupt to disable
 *
 *	Disable the selected interrupt line.  Disables and Enables are
 *	nested.
 *	Unlike disable_irq(), this function does not ensure existing
 *	instances of the IRQ handler have completed before returning.
 *
 *	This function may be called from IRQ context.
 */
void disable_irq_nosync(unsigned int irq)
{
396
	__disable_irq_nosync(irq);
L
Linus Torvalds 已提交
397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413
}
EXPORT_SYMBOL(disable_irq_nosync);

/**
 *	disable_irq - disable an irq and wait for completion
 *	@irq: Interrupt to disable
 *
 *	Disable the selected interrupt line.  Enables and Disables are
 *	nested.
 *	This function waits for any pending IRQ handlers for this interrupt
 *	to complete before returning. If you use this function while
 *	holding a resource the IRQ handler may need you will deadlock.
 *
 *	This function may be called - with care - from IRQ context.
 */
void disable_irq(unsigned int irq)
{
414
	if (!__disable_irq_nosync(irq))
L
Linus Torvalds 已提交
415 416 417 418
		synchronize_irq(irq);
}
EXPORT_SYMBOL(disable_irq);

419
void __enable_irq(struct irq_desc *desc, unsigned int irq, bool resume)
420
{
T
Thomas Gleixner 已提交
421
	if (resume) {
422
		if (!(desc->istate & IRQS_SUSPENDED)) {
T
Thomas Gleixner 已提交
423 424 425 426 427 428 429
			if (!desc->action)
				return;
			if (!(desc->action->flags & IRQF_FORCE_RESUME))
				return;
			/* Pretend that it got disabled ! */
			desc->depth++;
		}
430
		desc->istate &= ~IRQS_SUSPENDED;
T
Thomas Gleixner 已提交
431
	}
432

433 434
	switch (desc->depth) {
	case 0:
435
 err_out:
436
		WARN(1, KERN_WARNING "Unbalanced enable for IRQ %d\n", irq);
437 438
		break;
	case 1: {
439
		if (desc->istate & IRQS_SUSPENDED)
440
			goto err_out;
441
		/* Prevent probing on this irq: */
442
		irq_settings_set_noprobe(desc);
443
		irq_enable(desc);
444 445 446 447 448 449 450 451
		check_irq_resend(desc, irq);
		/* fall-through */
	}
	default:
		desc->depth--;
	}
}

L
Linus Torvalds 已提交
452 453 454 455 456 457 458 459
/**
 *	enable_irq - enable handling of an irq
 *	@irq: Interrupt to enable
 *
 *	Undoes the effect of one call to disable_irq().  If this
 *	matches the last disable, processing of interrupts on this
 *	IRQ line is re-enabled.
 *
T
Thomas Gleixner 已提交
460
 *	This function may be called from IRQ context only when
461
 *	desc->irq_data.chip->bus_lock and desc->chip->bus_sync_unlock are NULL !
L
Linus Torvalds 已提交
462 463 464 465
 */
void enable_irq(unsigned int irq)
{
	unsigned long flags;
466
	struct irq_desc *desc = irq_get_desc_buslock(irq, &flags, IRQ_GET_DESC_CHECK_GLOBAL);
L
Linus Torvalds 已提交
467

468
	if (!desc)
469
		return;
T
Thomas Gleixner 已提交
470 471
	if (WARN(!desc->irq_data.chip,
		 KERN_ERR "enable_irq before setup/request_irq: irq %u\n", irq))
472
		goto out;
473

474
	__enable_irq(desc, irq, false);
475 476
out:
	irq_put_desc_busunlock(desc, flags);
L
Linus Torvalds 已提交
477 478 479
}
EXPORT_SYMBOL(enable_irq);

D
David Brownell 已提交
480
static int set_irq_wake_real(unsigned int irq, unsigned int on)
481
{
482
	struct irq_desc *desc = irq_to_desc(irq);
483 484
	int ret = -ENXIO;

485 486 487
	if (irq_desc_get_chip(desc)->flags &  IRQCHIP_SKIP_SET_WAKE)
		return 0;

488 489
	if (desc->irq_data.chip->irq_set_wake)
		ret = desc->irq_data.chip->irq_set_wake(&desc->irq_data, on);
490 491 492 493

	return ret;
}

494
/**
T
Thomas Gleixner 已提交
495
 *	irq_set_irq_wake - control irq power management wakeup
496 497 498
 *	@irq:	interrupt to control
 *	@on:	enable/disable power management wakeup
 *
499 500 501 502 503 504
 *	Enable/disable power management wakeup mode, which is
 *	disabled by default.  Enables and disables must match,
 *	just as they match for non-wakeup mode support.
 *
 *	Wakeup mode lets this IRQ wake the system from sleep
 *	states like "suspend to RAM".
505
 */
T
Thomas Gleixner 已提交
506
int irq_set_irq_wake(unsigned int irq, unsigned int on)
507 508
{
	unsigned long flags;
509
	struct irq_desc *desc = irq_get_desc_buslock(irq, &flags, IRQ_GET_DESC_CHECK_GLOBAL);
510
	int ret = 0;
511

512 513 514
	if (!desc)
		return -EINVAL;

515 516 517 518
	/* wakeup-capable irqs can be shared between drivers that
	 * don't need to have the same sleep mode behaviors.
	 */
	if (on) {
519 520 521 522 523
		if (desc->wake_depth++ == 0) {
			ret = set_irq_wake_real(irq, on);
			if (ret)
				desc->wake_depth = 0;
			else
524
				irqd_set(&desc->irq_data, IRQD_WAKEUP_STATE);
525
		}
526 527
	} else {
		if (desc->wake_depth == 0) {
528
			WARN(1, "Unbalanced IRQ %d wake disable\n", irq);
529 530 531 532 533
		} else if (--desc->wake_depth == 0) {
			ret = set_irq_wake_real(irq, on);
			if (ret)
				desc->wake_depth = 1;
			else
534
				irqd_clear(&desc->irq_data, IRQD_WAKEUP_STATE);
535
		}
536
	}
537
	irq_put_desc_busunlock(desc, flags);
538 539
	return ret;
}
T
Thomas Gleixner 已提交
540
EXPORT_SYMBOL(irq_set_irq_wake);
541

L
Linus Torvalds 已提交
542 543 544 545 546 547 548
/*
 * Internal function that tells the architecture code whether a
 * particular irq has been exclusively allocated or is available
 * for driver use.
 */
int can_request_irq(unsigned int irq, unsigned long irqflags)
{
549
	unsigned long flags;
550
	struct irq_desc *desc = irq_get_desc_lock(irq, &flags, 0);
551
	int canrequest = 0;
L
Linus Torvalds 已提交
552

553 554 555
	if (!desc)
		return 0;

556 557 558 559 560 561 562
	if (irq_settings_can_request(desc)) {
		if (desc->action)
			if (irqflags & desc->action->flags & IRQF_SHARED)
				canrequest =1;
	}
	irq_put_desc_unlock(desc, flags);
	return canrequest;
L
Linus Torvalds 已提交
563 564
}

D
David Brownell 已提交
565
int __irq_set_trigger(struct irq_desc *desc, unsigned int irq,
566
		      unsigned long flags)
567
{
568
	struct irq_chip *chip = desc->irq_data.chip;
569
	int ret, unmask = 0;
570

571
	if (!chip || !chip->irq_set_type) {
572 573 574 575
		/*
		 * IRQF_TRIGGER_* but the PIC does not support multiple
		 * flow-types?
		 */
576
		pr_debug("No set_type function for IRQ %d (%s)\n", irq,
577
			 chip ? (chip->name ? : "unknown") : "unknown");
578 579 580
		return 0;
	}

581
	flags &= IRQ_TYPE_SENSE_MASK;
582 583

	if (chip->flags & IRQCHIP_SET_TYPE_MASKED) {
584
		if (!irqd_irq_masked(&desc->irq_data))
585
			mask_irq(desc);
586
		if (!irqd_irq_disabled(&desc->irq_data))
587 588 589
			unmask = 1;
	}

590
	/* caller masked out all except trigger mode flags */
591
	ret = chip->irq_set_type(&desc->irq_data, flags);
592

593 594 595 596 597 598 599 600 601 602 603 604 605 606
	switch (ret) {
	case IRQ_SET_MASK_OK:
		irqd_clear(&desc->irq_data, IRQD_TRIGGER_MASK);
		irqd_set(&desc->irq_data, flags);

	case IRQ_SET_MASK_OK_NOCOPY:
		flags = irqd_get_trigger_type(&desc->irq_data);
		irq_settings_set_trigger_mask(desc, flags);
		irqd_clear(&desc->irq_data, IRQD_LEVEL);
		irq_settings_clr_level(desc);
		if (flags & IRQ_TYPE_LEVEL_MASK) {
			irq_settings_set_level(desc);
			irqd_set(&desc->irq_data, IRQD_LEVEL);
		}
607

608
		ret = 0;
609
		break;
610
	default:
611
		pr_err("Setting trigger mode %lu for irq %u failed (%pF)\n",
612
		       flags, irq, chip->irq_set_type);
D
David Brownell 已提交
613
	}
614 615
	if (unmask)
		unmask_irq(desc);
616 617 618
	return ret;
}

619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634
#ifdef CONFIG_HARDIRQS_SW_RESEND
int irq_set_parent(int irq, int parent_irq)
{
	unsigned long flags;
	struct irq_desc *desc = irq_get_desc_lock(irq, &flags, 0);

	if (!desc)
		return -EINVAL;

	desc->parent_irq = parent_irq;

	irq_put_desc_unlock(desc, flags);
	return 0;
}
#endif

T
Thomas Gleixner 已提交
635 636 637 638 639 640 641 642 643 644
/*
 * Default primary interrupt handler for threaded interrupts. Is
 * assigned as primary handler when request_threaded_irq is called
 * with handler == NULL. Useful for oneshot interrupts.
 */
static irqreturn_t irq_default_primary_handler(int irq, void *dev_id)
{
	return IRQ_WAKE_THREAD;
}

645 646 647 648 649 650 651 652 653 654
/*
 * Primary handler for nested threaded interrupts. Should never be
 * called.
 */
static irqreturn_t irq_nested_primary_handler(int irq, void *dev_id)
{
	WARN(1, "Primary handler called for nested irq %d\n", irq);
	return IRQ_NONE;
}

655 656
static int irq_wait_for_interrupt(struct irqaction *action)
{
657 658
	set_current_state(TASK_INTERRUPTIBLE);

659
	while (!kthread_should_stop()) {
660 661 662

		if (test_and_clear_bit(IRQTF_RUNTHREAD,
				       &action->thread_flags)) {
663 664
			__set_current_state(TASK_RUNNING);
			return 0;
665 666
		}
		schedule();
667
		set_current_state(TASK_INTERRUPTIBLE);
668
	}
669
	__set_current_state(TASK_RUNNING);
670 671 672
	return -1;
}

T
Thomas Gleixner 已提交
673 674 675 676 677
/*
 * Oneshot interrupts keep the irq line masked until the threaded
 * handler finished. unmask if the interrupt has not been disabled and
 * is marked MASKED.
 */
678
static void irq_finalize_oneshot(struct irq_desc *desc,
679
				 struct irqaction *action)
T
Thomas Gleixner 已提交
680
{
681 682
	if (!(desc->istate & IRQS_ONESHOT))
		return;
683
again:
684
	chip_bus_lock(desc);
685
	raw_spin_lock_irq(&desc->lock);
686 687 688 689 690 691 692 693

	/*
	 * Implausible though it may be we need to protect us against
	 * the following scenario:
	 *
	 * The thread is faster done than the hard interrupt handler
	 * on the other CPU. If we unmask the irq line then the
	 * interrupt can come in again and masks the line, leaves due
694
	 * to IRQS_INPROGRESS and the irq line is masked forever.
695 696 697 698 699
	 *
	 * This also serializes the state of shared oneshot handlers
	 * versus "desc->threads_onehsot |= action->thread_mask;" in
	 * irq_wake_thread(). See the comment there which explains the
	 * serialization.
700
	 */
701
	if (unlikely(irqd_irq_inprogress(&desc->irq_data))) {
702
		raw_spin_unlock_irq(&desc->lock);
703
		chip_bus_sync_unlock(desc);
704 705 706 707
		cpu_relax();
		goto again;
	}

708 709 710 711 712
	/*
	 * Now check again, whether the thread should run. Otherwise
	 * we would clear the threads_oneshot bit of this thread which
	 * was just set.
	 */
713
	if (test_bit(IRQTF_RUNTHREAD, &action->thread_flags))
714 715 716 717
		goto out_unlock;

	desc->threads_oneshot &= ~action->thread_mask;

718 719 720 721
	if (!desc->threads_oneshot && !irqd_irq_disabled(&desc->irq_data) &&
	    irqd_irq_masked(&desc->irq_data))
		unmask_irq(desc);

722
out_unlock:
723
	raw_spin_unlock_irq(&desc->lock);
724
	chip_bus_sync_unlock(desc);
T
Thomas Gleixner 已提交
725 726
}

727
#ifdef CONFIG_SMP
728
/*
729
 * Check whether we need to chasnge the affinity of the interrupt thread.
730 731 732 733 734
 */
static void
irq_thread_check_affinity(struct irq_desc *desc, struct irqaction *action)
{
	cpumask_var_t mask;
735
	bool valid = true;
736 737 738 739 740 741 742 743 744 745 746 747 748

	if (!test_and_clear_bit(IRQTF_AFFINITY, &action->thread_flags))
		return;

	/*
	 * In case we are out of memory we set IRQTF_AFFINITY again and
	 * try again next time
	 */
	if (!alloc_cpumask_var(&mask, GFP_KERNEL)) {
		set_bit(IRQTF_AFFINITY, &action->thread_flags);
		return;
	}

749
	raw_spin_lock_irq(&desc->lock);
750 751 752 753 754 755 756 757
	/*
	 * This code is triggered unconditionally. Check the affinity
	 * mask pointer. For CPU_MASK_OFFSTACK=n this is optimized out.
	 */
	if (desc->irq_data.affinity)
		cpumask_copy(mask, desc->irq_data.affinity);
	else
		valid = false;
758
	raw_spin_unlock_irq(&desc->lock);
759

760 761
	if (valid)
		set_cpus_allowed_ptr(current, mask);
762 763
	free_cpumask_var(mask);
}
764 765 766 767
#else
static inline void
irq_thread_check_affinity(struct irq_desc *desc, struct irqaction *action) { }
#endif
768

769 770 771 772 773 774
/*
 * Interrupts which are not explicitely requested as threaded
 * interrupts rely on the implicit bh/preempt disable of the hard irq
 * context. So we need to disable bh here to avoid deadlocks and other
 * side effects.
 */
775
static irqreturn_t
776 777
irq_forced_thread_fn(struct irq_desc *desc, struct irqaction *action)
{
778 779
	irqreturn_t ret;

780
	local_bh_disable();
781
	ret = action->thread_fn(action->irq, action->dev_id);
782
	irq_finalize_oneshot(desc, action);
783
	local_bh_enable();
784
	return ret;
785 786 787 788 789 790 791
}

/*
 * Interrupts explicitely requested as threaded interupts want to be
 * preemtible - many of them need to sleep and wait for slow busses to
 * complete.
 */
792 793
static irqreturn_t irq_thread_fn(struct irq_desc *desc,
		struct irqaction *action)
794
{
795 796 797
	irqreturn_t ret;

	ret = action->thread_fn(action->irq, action->dev_id);
798
	irq_finalize_oneshot(desc, action);
799
	return ret;
800 801
}

802 803 804 805 806 807 808
static void wake_threads_waitq(struct irq_desc *desc)
{
	if (atomic_dec_and_test(&desc->threads_active) &&
	    waitqueue_active(&desc->wait_for_threads))
		wake_up(&desc->wait_for_threads);
}

809
static void irq_thread_dtor(struct callback_head *unused)
810 811 812 813 814 815 816 817 818 819
{
	struct task_struct *tsk = current;
	struct irq_desc *desc;
	struct irqaction *action;

	if (WARN_ON_ONCE(!(current->flags & PF_EXITING)))
		return;

	action = kthread_data(tsk);

820
	pr_err("exiting task \"%s\" (%d) is an active IRQ thread (irq %d)\n",
A
Alan Cox 已提交
821
	       tsk->comm, tsk->pid, action->irq);
822 823 824 825 826 827 828 829 830 831 832 833 834 835


	desc = irq_to_desc(action->irq);
	/*
	 * If IRQTF_RUNTHREAD is set, we need to decrement
	 * desc->threads_active and wake possible waiters.
	 */
	if (test_and_clear_bit(IRQTF_RUNTHREAD, &action->thread_flags))
		wake_threads_waitq(desc);

	/* Prevent a stale desc->threads_oneshot */
	irq_finalize_oneshot(desc, action);
}

836 837 838 839 840
/*
 * Interrupt handler thread
 */
static int irq_thread(void *data)
{
841
	struct callback_head on_exit_work;
842
	static const struct sched_param param = {
843 844
		.sched_priority = MAX_USER_RT_PRIO/2,
	};
845 846
	struct irqaction *action = data;
	struct irq_desc *desc = irq_to_desc(action->irq);
847 848
	irqreturn_t (*handler_fn)(struct irq_desc *desc,
			struct irqaction *action);
849

850
	if (force_irqthreads && test_bit(IRQTF_FORCED_THREAD,
851 852 853 854 855
					&action->thread_flags))
		handler_fn = irq_forced_thread_fn;
	else
		handler_fn = irq_thread_fn;

856
	sched_setscheduler(current, SCHED_FIFO, &param);
857

A
Al Viro 已提交
858
	init_task_work(&on_exit_work, irq_thread_dtor);
859
	task_work_add(current, &on_exit_work, false);
860

861 862
	irq_thread_check_affinity(desc, action);

863
	while (!irq_wait_for_interrupt(action)) {
864
		irqreturn_t action_ret;
865

866 867
		irq_thread_check_affinity(desc, action);

868 869 870
		action_ret = handler_fn(desc, action);
		if (!noirqdebug)
			note_interrupt(action->irq, desc, action_ret);
871

872
		wake_threads_waitq(desc);
873 874
	}

875 876 877 878
	/*
	 * This is the regular exit path. __free_irq() is stopping the
	 * thread via kthread_stop() after calling
	 * synchronize_irq(). So neither IRQTF_RUNTHREAD nor the
879 880 881 882
	 * oneshot mask bit can be set. We cannot verify that as we
	 * cannot touch the oneshot mask at this point anymore as
	 * __setup_irq() might have given out currents thread_mask
	 * again.
883
	 */
884
	task_work_cancel(current, irq_thread_dtor);
885 886 887
	return 0;
}

888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903
static void irq_setup_forced_threading(struct irqaction *new)
{
	if (!force_irqthreads)
		return;
	if (new->flags & (IRQF_NO_THREAD | IRQF_PERCPU | IRQF_ONESHOT))
		return;

	new->flags |= IRQF_ONESHOT;

	if (!new->thread_fn) {
		set_bit(IRQTF_FORCED_THREAD, &new->thread_flags);
		new->thread_fn = new->handler;
		new->handler = irq_default_primary_handler;
	}
}

L
Linus Torvalds 已提交
904 905 906 907
/*
 * Internal function to register an irqaction - typically used to
 * allocate special interrupts that are part of the architecture.
 */
908
static int
I
Ingo Molnar 已提交
909
__setup_irq(unsigned int irq, struct irq_desc *desc, struct irqaction *new)
L
Linus Torvalds 已提交
910
{
911
	struct irqaction *old, **old_ptr;
912
	unsigned long flags, thread_mask = 0;
913 914
	int ret, nested, shared = 0;
	cpumask_var_t mask;
L
Linus Torvalds 已提交
915

916
	if (!desc)
917 918
		return -EINVAL;

919
	if (desc->irq_data.chip == &no_irq_chip)
L
Linus Torvalds 已提交
920
		return -ENOSYS;
921 922
	if (!try_module_get(desc->owner))
		return -ENODEV;
L
Linus Torvalds 已提交
923

924
	/*
925 926 927
	 * Check whether the interrupt nests into another interrupt
	 * thread.
	 */
928
	nested = irq_settings_is_nested_thread(desc);
929
	if (nested) {
930 931 932 933
		if (!new->thread_fn) {
			ret = -EINVAL;
			goto out_mput;
		}
934 935 936 937 938 939
		/*
		 * Replace the primary handler which was provided from
		 * the driver for non nested interrupt handling by the
		 * dummy function which warns when called.
		 */
		new->handler = irq_nested_primary_handler;
940
	} else {
941 942
		if (irq_settings_can_thread(desc))
			irq_setup_forced_threading(new);
943 944
	}

945
	/*
946 947 948
	 * Create a handler thread when a thread function is supplied
	 * and the interrupt does not nest into another interrupt
	 * thread.
949
	 */
950
	if (new->thread_fn && !nested) {
951 952 953 954
		struct task_struct *t;

		t = kthread_create(irq_thread, new, "irq/%d-%s", irq,
				   new->name);
955 956 957 958
		if (IS_ERR(t)) {
			ret = PTR_ERR(t);
			goto out_mput;
		}
959 960 961 962 963 964 965
		/*
		 * We keep the reference to the task struct even if
		 * the thread dies to avoid that the interrupt code
		 * references an already freed task_struct.
		 */
		get_task_struct(t);
		new->thread = t;
966 967 968 969 970 971 972 973 974 975
		/*
		 * Tell the thread to set its affinity. This is
		 * important for shared interrupt handlers as we do
		 * not invoke setup_affinity() for the secondary
		 * handlers as everything is already set up. Even for
		 * interrupts marked with IRQF_NO_BALANCE this is
		 * correct as we want the thread to move to the cpu(s)
		 * on which the requesting code placed the interrupt.
		 */
		set_bit(IRQTF_AFFINITY, &new->thread_flags);
976 977
	}

978 979 980 981 982
	if (!alloc_cpumask_var(&mask, GFP_KERNEL)) {
		ret = -ENOMEM;
		goto out_thread;
	}

983 984 985 986 987 988 989 990 991 992 993 994
	/*
	 * Drivers are often written to work w/o knowledge about the
	 * underlying irq chip implementation, so a request for a
	 * threaded irq without a primary hard irq context handler
	 * requires the ONESHOT flag to be set. Some irq chips like
	 * MSI based interrupts are per se one shot safe. Check the
	 * chip flags, so we can avoid the unmask dance at the end of
	 * the threaded handler for those.
	 */
	if (desc->irq_data.chip->flags & IRQCHIP_ONESHOT_SAFE)
		new->flags &= ~IRQF_ONESHOT;

L
Linus Torvalds 已提交
995 996 997
	/*
	 * The following block of code has to be executed atomically
	 */
998
	raw_spin_lock_irqsave(&desc->lock, flags);
999 1000
	old_ptr = &desc->action;
	old = *old_ptr;
1001
	if (old) {
1002 1003 1004
		/*
		 * Can't share interrupts unless both agree to and are
		 * the same type (level, edge, polarity). So both flag
1005
		 * fields must have IRQF_SHARED set and the bits which
1006 1007
		 * set the trigger type must match. Also all must
		 * agree on ONESHOT.
1008
		 */
1009
		if (!((old->flags & new->flags) & IRQF_SHARED) ||
1010
		    ((old->flags ^ new->flags) & IRQF_TRIGGER_MASK) ||
1011
		    ((old->flags ^ new->flags) & IRQF_ONESHOT))
1012 1013 1014
			goto mismatch;

		/* All handlers must agree on per-cpuness */
1015 1016
		if ((old->flags & IRQF_PERCPU) !=
		    (new->flags & IRQF_PERCPU))
1017
			goto mismatch;
L
Linus Torvalds 已提交
1018 1019 1020

		/* add new interrupt at end of irq queue */
		do {
1021 1022 1023 1024 1025
			/*
			 * Or all existing action->thread_mask bits,
			 * so we can find the next zero bit for this
			 * new action.
			 */
1026
			thread_mask |= old->thread_mask;
1027 1028
			old_ptr = &old->next;
			old = *old_ptr;
L
Linus Torvalds 已提交
1029 1030 1031 1032
		} while (old);
		shared = 1;
	}

1033
	/*
1034 1035 1036
	 * Setup the thread mask for this irqaction for ONESHOT. For
	 * !ONESHOT irqs the thread mask is 0 so we can avoid a
	 * conditional in irq_wake_thread().
1037
	 */
1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067
	if (new->flags & IRQF_ONESHOT) {
		/*
		 * Unlikely to have 32 resp 64 irqs sharing one line,
		 * but who knows.
		 */
		if (thread_mask == ~0UL) {
			ret = -EBUSY;
			goto out_mask;
		}
		/*
		 * The thread_mask for the action is or'ed to
		 * desc->thread_active to indicate that the
		 * IRQF_ONESHOT thread handler has been woken, but not
		 * yet finished. The bit is cleared when a thread
		 * completes. When all threads of a shared interrupt
		 * line have completed desc->threads_active becomes
		 * zero and the interrupt line is unmasked. See
		 * handle.c:irq_wake_thread() for further information.
		 *
		 * If no thread is woken by primary (hard irq context)
		 * interrupt handlers, then desc->threads_active is
		 * also checked for zero to unmask the irq line in the
		 * affected hard irq flow handlers
		 * (handle_[fasteoi|level]_irq).
		 *
		 * The new action gets the first zero bit of
		 * thread_mask assigned. See the loop above which or's
		 * all existing action->thread_mask bits.
		 */
		new->thread_mask = 1 << ffz(thread_mask);
1068

1069 1070
	} else if (new->handler == irq_default_primary_handler &&
		   !(desc->irq_data.chip->flags & IRQCHIP_ONESHOT_SAFE)) {
1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085
		/*
		 * The interrupt was requested with handler = NULL, so
		 * we use the default primary handler for it. But it
		 * does not have the oneshot flag set. In combination
		 * with level interrupts this is deadly, because the
		 * default primary handler just wakes the thread, then
		 * the irq lines is reenabled, but the device still
		 * has the level irq asserted. Rinse and repeat....
		 *
		 * While this works for edge type interrupts, we play
		 * it safe and reject unconditionally because we can't
		 * say for sure which type this interrupt really
		 * has. The type flags are unreliable as the
		 * underlying chip implementation can override them.
		 */
1086
		pr_err("Threaded irq requested with handler=NULL and !ONESHOT for irq %d\n",
1087 1088 1089
		       irq);
		ret = -EINVAL;
		goto out_mask;
1090 1091
	}

L
Linus Torvalds 已提交
1092
	if (!shared) {
1093 1094
		init_waitqueue_head(&desc->wait_for_threads);

1095
		/* Setup the type (level, edge polarity) if configured: */
1096
		if (new->flags & IRQF_TRIGGER_MASK) {
1097 1098
			ret = __irq_set_trigger(desc, irq,
					new->flags & IRQF_TRIGGER_MASK);
1099

1100
			if (ret)
1101
				goto out_mask;
1102
		}
T
Thomas Gleixner 已提交
1103

1104
		desc->istate &= ~(IRQS_AUTODETECT | IRQS_SPURIOUS_DISABLED | \
1105 1106
				  IRQS_ONESHOT | IRQS_WAITING);
		irqd_clear(&desc->irq_data, IRQD_IRQ_INPROGRESS);
1107

1108 1109 1110 1111
		if (new->flags & IRQF_PERCPU) {
			irqd_set(&desc->irq_data, IRQD_PER_CPU);
			irq_settings_set_per_cpu(desc);
		}
1112

T
Thomas Gleixner 已提交
1113
		if (new->flags & IRQF_ONESHOT)
1114
			desc->istate |= IRQS_ONESHOT;
T
Thomas Gleixner 已提交
1115

1116
		if (irq_settings_can_autoenable(desc))
1117
			irq_startup(desc, true);
1118
		else
1119 1120
			/* Undo nested disables: */
			desc->depth = 1;
1121

1122
		/* Exclude IRQ from balancing if requested */
1123 1124 1125 1126
		if (new->flags & IRQF_NOBALANCING) {
			irq_settings_set_no_balancing(desc);
			irqd_set(&desc->irq_data, IRQD_NO_BALANCING);
		}
1127

1128
		/* Set default affinity mask once everything is setup */
1129
		setup_affinity(irq, desc, mask);
D
David Brownell 已提交
1130

1131 1132 1133 1134 1135 1136
	} else if (new->flags & IRQF_TRIGGER_MASK) {
		unsigned int nmsk = new->flags & IRQF_TRIGGER_MASK;
		unsigned int omsk = irq_settings_get_trigger_mask(desc);

		if (nmsk != omsk)
			/* hope the handler works with current  trigger mode */
1137
			pr_warning("irq %d uses trigger mode %u; requested %u\n",
1138
				   irq, nmsk, omsk);
L
Linus Torvalds 已提交
1139
	}
1140

1141
	new->irq = irq;
1142
	*old_ptr = new;
1143

1144 1145 1146
	/* Reset broken irq detection when installing new handler */
	desc->irq_count = 0;
	desc->irqs_unhandled = 0;
1147 1148 1149 1150 1151

	/*
	 * Check whether we disabled the irq via the spurious handler
	 * before. Reenable it and give it another chance.
	 */
1152 1153
	if (shared && (desc->istate & IRQS_SPURIOUS_DISABLED)) {
		desc->istate &= ~IRQS_SPURIOUS_DISABLED;
1154
		__enable_irq(desc, irq, false);
1155 1156
	}

1157
	raw_spin_unlock_irqrestore(&desc->lock, flags);
L
Linus Torvalds 已提交
1158

1159 1160 1161 1162 1163 1164 1165
	/*
	 * Strictly no need to wake it up, but hung_task complains
	 * when no hard interrupt wakes the thread up.
	 */
	if (new->thread)
		wake_up_process(new->thread);

1166
	register_irq_proc(irq, desc);
L
Linus Torvalds 已提交
1167 1168
	new->dir = NULL;
	register_handler_proc(irq, new);
1169
	free_cpumask_var(mask);
L
Linus Torvalds 已提交
1170 1171

	return 0;
1172 1173

mismatch:
1174
	if (!(new->flags & IRQF_PROBE_SHARED)) {
1175
		pr_err("Flags mismatch irq %d. %08x (%s) vs. %08x (%s)\n",
1176 1177
		       irq, new->flags, new->name, old->flags, old->name);
#ifdef CONFIG_DEBUG_SHIRQ
1178
		dump_stack();
1179
#endif
1180
	}
1181 1182
	ret = -EBUSY;

1183
out_mask:
1184
	raw_spin_unlock_irqrestore(&desc->lock, flags);
1185 1186
	free_cpumask_var(mask);

1187 1188 1189 1190 1191
out_thread:
	if (new->thread) {
		struct task_struct *t = new->thread;

		new->thread = NULL;
1192
		kthread_stop(t);
1193 1194
		put_task_struct(t);
	}
1195 1196
out_mput:
	module_put(desc->owner);
1197
	return ret;
L
Linus Torvalds 已提交
1198 1199
}

1200 1201 1202 1203 1204 1205 1206 1207 1208
/**
 *	setup_irq - setup an interrupt
 *	@irq: Interrupt line to setup
 *	@act: irqaction for the interrupt
 *
 * Used to statically setup interrupts in the early boot process.
 */
int setup_irq(unsigned int irq, struct irqaction *act)
{
1209
	int retval;
1210 1211
	struct irq_desc *desc = irq_to_desc(irq);

1212 1213
	if (WARN_ON(irq_settings_is_per_cpu_devid(desc)))
		return -EINVAL;
1214 1215 1216 1217 1218
	chip_bus_lock(desc);
	retval = __setup_irq(irq, desc, act);
	chip_bus_sync_unlock(desc);

	return retval;
1219
}
1220
EXPORT_SYMBOL_GPL(setup_irq);
1221

1222
/*
1223 1224
 * Internal function to unregister an irqaction - used to free
 * regular and special interrupts that are part of the architecture.
L
Linus Torvalds 已提交
1225
 */
1226
static struct irqaction *__free_irq(unsigned int irq, void *dev_id)
L
Linus Torvalds 已提交
1227
{
1228
	struct irq_desc *desc = irq_to_desc(irq);
1229
	struct irqaction *action, **action_ptr;
L
Linus Torvalds 已提交
1230 1231
	unsigned long flags;

1232
	WARN(in_interrupt(), "Trying to free IRQ %d from IRQ context!\n", irq);
1233 1234

	if (!desc)
1235
		return NULL;
L
Linus Torvalds 已提交
1236

1237
	raw_spin_lock_irqsave(&desc->lock, flags);
1238 1239 1240 1241 1242

	/*
	 * There can be multiple actions per IRQ descriptor, find the right
	 * one based on the dev_id:
	 */
1243
	action_ptr = &desc->action;
L
Linus Torvalds 已提交
1244
	for (;;) {
1245
		action = *action_ptr;
L
Linus Torvalds 已提交
1246

1247 1248
		if (!action) {
			WARN(1, "Trying to free already-free IRQ %d\n", irq);
1249
			raw_spin_unlock_irqrestore(&desc->lock, flags);
L
Linus Torvalds 已提交
1250

1251
			return NULL;
1252
		}
L
Linus Torvalds 已提交
1253

1254 1255
		if (action->dev_id == dev_id)
			break;
1256
		action_ptr = &action->next;
1257
	}
1258

1259
	/* Found it - now remove it from the list of entries: */
1260
	*action_ptr = action->next;
1261 1262

	/* If this was the last handler, shut down the IRQ line: */
1263 1264
	if (!desc->action)
		irq_shutdown(desc);
1265

1266 1267 1268 1269 1270 1271
#ifdef CONFIG_SMP
	/* make sure affinity_hint is cleaned up */
	if (WARN_ON_ONCE(desc->affinity_hint))
		desc->affinity_hint = NULL;
#endif

1272
	raw_spin_unlock_irqrestore(&desc->lock, flags);
1273 1274 1275 1276 1277

	unregister_handler_proc(irq, action);

	/* Make sure it's not being used on another CPU: */
	synchronize_irq(irq);
L
Linus Torvalds 已提交
1278

1279
#ifdef CONFIG_DEBUG_SHIRQ
1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291
	/*
	 * It's a shared IRQ -- the driver ought to be prepared for an IRQ
	 * event to happen even now it's being freed, so let's make sure that
	 * is so by doing an extra call to the handler ....
	 *
	 * ( We do this after actually deregistering it, to make sure that a
	 *   'real' IRQ doesn't run in * parallel with our fake. )
	 */
	if (action->flags & IRQF_SHARED) {
		local_irq_save(flags);
		action->handler(irq, dev_id);
		local_irq_restore(flags);
L
Linus Torvalds 已提交
1292
	}
1293
#endif
1294 1295

	if (action->thread) {
1296
		kthread_stop(action->thread);
1297 1298 1299
		put_task_struct(action->thread);
	}

1300
	module_put(desc->owner);
1301 1302 1303
	return action;
}

1304 1305 1306 1307 1308 1309 1310 1311 1312
/**
 *	remove_irq - free an interrupt
 *	@irq: Interrupt line to free
 *	@act: irqaction for the interrupt
 *
 * Used to remove interrupts statically setup by the early boot process.
 */
void remove_irq(unsigned int irq, struct irqaction *act)
{
1313 1314 1315 1316
	struct irq_desc *desc = irq_to_desc(irq);

	if (desc && !WARN_ON(irq_settings_is_per_cpu_devid(desc)))
	    __free_irq(irq, act->dev_id);
1317
}
1318
EXPORT_SYMBOL_GPL(remove_irq);
1319

1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335
/**
 *	free_irq - free an interrupt allocated with request_irq
 *	@irq: Interrupt line to free
 *	@dev_id: Device identity to free
 *
 *	Remove an interrupt handler. The handler is removed and if the
 *	interrupt line is no longer in use by any driver it is disabled.
 *	On a shared IRQ the caller must ensure the interrupt is disabled
 *	on the card it drives before calling this function. The function
 *	does not return until any executing interrupts for this IRQ
 *	have completed.
 *
 *	This function must not be called from interrupt context.
 */
void free_irq(unsigned int irq, void *dev_id)
{
T
Thomas Gleixner 已提交
1336 1337
	struct irq_desc *desc = irq_to_desc(irq);

1338
	if (!desc || WARN_ON(irq_settings_is_per_cpu_devid(desc)))
T
Thomas Gleixner 已提交
1339 1340
		return;

1341 1342 1343 1344 1345
#ifdef CONFIG_SMP
	if (WARN_ON(desc->affinity_notify))
		desc->affinity_notify = NULL;
#endif

1346
	chip_bus_lock(desc);
1347
	kfree(__free_irq(irq, dev_id));
1348
	chip_bus_sync_unlock(desc);
L
Linus Torvalds 已提交
1349 1350 1351 1352
}
EXPORT_SYMBOL(free_irq);

/**
1353
 *	request_threaded_irq - allocate an interrupt line
L
Linus Torvalds 已提交
1354
 *	@irq: Interrupt line to allocate
1355 1356
 *	@handler: Function to be called when the IRQ occurs.
 *		  Primary handler for threaded interrupts
T
Thomas Gleixner 已提交
1357 1358
 *		  If NULL and thread_fn != NULL the default
 *		  primary handler is installed
1359 1360
 *	@thread_fn: Function called from the irq handler thread
 *		    If NULL, no irq thread is created
L
Linus Torvalds 已提交
1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371
 *	@irqflags: Interrupt type flags
 *	@devname: An ascii name for the claiming device
 *	@dev_id: A cookie passed back to the handler function
 *
 *	This call allocates interrupt resources and enables the
 *	interrupt line and IRQ handling. From the point this
 *	call is made your handler function may be invoked. Since
 *	your handler function must clear any interrupt the board
 *	raises, you must take care both to initialise your hardware
 *	and to set up the interrupt handler in the right order.
 *
1372
 *	If you want to set up a threaded irq handler for your device
J
Javi Merino 已提交
1373
 *	then you need to supply @handler and @thread_fn. @handler is
1374 1375 1376
 *	still called in hard interrupt context and has to check
 *	whether the interrupt originates from the device. If yes it
 *	needs to disable the interrupt on the device and return
1377
 *	IRQ_WAKE_THREAD which will wake up the handler thread and run
1378 1379 1380
 *	@thread_fn. This split handler design is necessary to support
 *	shared interrupts.
 *
L
Linus Torvalds 已提交
1381 1382 1383 1384 1385 1386 1387 1388 1389
 *	Dev_id must be globally unique. Normally the address of the
 *	device data structure is used as the cookie. Since the handler
 *	receives this value it makes sense to use it.
 *
 *	If your interrupt is shared you must pass a non NULL dev_id
 *	as this is required when freeing the interrupt.
 *
 *	Flags:
 *
1390
 *	IRQF_SHARED		Interrupt is shared
D
David Brownell 已提交
1391
 *	IRQF_TRIGGER_*		Specify active edge(s) or level
L
Linus Torvalds 已提交
1392 1393
 *
 */
1394 1395 1396
int request_threaded_irq(unsigned int irq, irq_handler_t handler,
			 irq_handler_t thread_fn, unsigned long irqflags,
			 const char *devname, void *dev_id)
L
Linus Torvalds 已提交
1397
{
1398
	struct irqaction *action;
1399
	struct irq_desc *desc;
1400
	int retval;
L
Linus Torvalds 已提交
1401 1402 1403 1404 1405 1406 1407

	/*
	 * Sanity-check: shared interrupts must pass in a real dev-ID,
	 * otherwise we'll have trouble later trying to figure out
	 * which interrupt is which (messes up the interrupt freeing
	 * logic etc).
	 */
1408
	if ((irqflags & IRQF_SHARED) && !dev_id)
L
Linus Torvalds 已提交
1409
		return -EINVAL;
1410

1411
	desc = irq_to_desc(irq);
1412
	if (!desc)
L
Linus Torvalds 已提交
1413
		return -EINVAL;
1414

1415 1416
	if (!irq_settings_can_request(desc) ||
	    WARN_ON(irq_settings_is_per_cpu_devid(desc)))
1417
		return -EINVAL;
T
Thomas Gleixner 已提交
1418 1419 1420 1421 1422 1423

	if (!handler) {
		if (!thread_fn)
			return -EINVAL;
		handler = irq_default_primary_handler;
	}
L
Linus Torvalds 已提交
1424

1425
	action = kzalloc(sizeof(struct irqaction), GFP_KERNEL);
L
Linus Torvalds 已提交
1426 1427 1428 1429
	if (!action)
		return -ENOMEM;

	action->handler = handler;
1430
	action->thread_fn = thread_fn;
L
Linus Torvalds 已提交
1431 1432 1433 1434
	action->flags = irqflags;
	action->name = devname;
	action->dev_id = dev_id;

1435
	chip_bus_lock(desc);
1436
	retval = __setup_irq(irq, desc, action);
1437
	chip_bus_sync_unlock(desc);
T
Thomas Gleixner 已提交
1438

1439 1440 1441
	if (retval)
		kfree(action);

1442
#ifdef CONFIG_DEBUG_SHIRQ_FIXME
1443
	if (!retval && (irqflags & IRQF_SHARED)) {
D
David Woodhouse 已提交
1444 1445 1446
		/*
		 * It's a shared IRQ -- the driver ought to be prepared for it
		 * to happen immediately, so let's make sure....
1447 1448
		 * We disable the irq to make sure that a 'real' IRQ doesn't
		 * run in parallel with our fake.
D
David Woodhouse 已提交
1449
		 */
1450
		unsigned long flags;
D
David Woodhouse 已提交
1451

1452
		disable_irq(irq);
1453
		local_irq_save(flags);
1454

1455
		handler(irq, dev_id);
1456

1457
		local_irq_restore(flags);
1458
		enable_irq(irq);
D
David Woodhouse 已提交
1459 1460
	}
#endif
L
Linus Torvalds 已提交
1461 1462
	return retval;
}
1463
EXPORT_SYMBOL(request_threaded_irq);
1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490

/**
 *	request_any_context_irq - allocate an interrupt line
 *	@irq: Interrupt line to allocate
 *	@handler: Function to be called when the IRQ occurs.
 *		  Threaded handler for threaded interrupts.
 *	@flags: Interrupt type flags
 *	@name: An ascii name for the claiming device
 *	@dev_id: A cookie passed back to the handler function
 *
 *	This call allocates interrupt resources and enables the
 *	interrupt line and IRQ handling. It selects either a
 *	hardirq or threaded handling method depending on the
 *	context.
 *
 *	On failure, it returns a negative value. On success,
 *	it returns either IRQC_IS_HARDIRQ or IRQC_IS_NESTED.
 */
int request_any_context_irq(unsigned int irq, irq_handler_t handler,
			    unsigned long flags, const char *name, void *dev_id)
{
	struct irq_desc *desc = irq_to_desc(irq);
	int ret;

	if (!desc)
		return -EINVAL;

1491
	if (irq_settings_is_nested_thread(desc)) {
1492 1493 1494 1495 1496 1497 1498 1499 1500
		ret = request_threaded_irq(irq, NULL, handler,
					   flags, name, dev_id);
		return !ret ? IRQC_IS_NESTED : ret;
	}

	ret = request_irq(irq, handler, flags, name, dev_id);
	return !ret ? IRQC_IS_HARDIRQ : ret;
}
EXPORT_SYMBOL_GPL(request_any_context_irq);
1501

1502
void enable_percpu_irq(unsigned int irq, unsigned int type)
1503 1504 1505 1506 1507 1508 1509 1510
{
	unsigned int cpu = smp_processor_id();
	unsigned long flags;
	struct irq_desc *desc = irq_get_desc_lock(irq, &flags, IRQ_GET_DESC_CHECK_PERCPU);

	if (!desc)
		return;

1511 1512 1513 1514 1515 1516 1517
	type &= IRQ_TYPE_SENSE_MASK;
	if (type != IRQ_TYPE_NONE) {
		int ret;

		ret = __irq_set_trigger(desc, irq, type);

		if (ret) {
1518
			WARN(1, "failed to set type for IRQ%d\n", irq);
1519 1520 1521 1522
			goto out;
		}
	}

1523
	irq_percpu_enable(desc, cpu);
1524
out:
1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 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 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678
	irq_put_desc_unlock(desc, flags);
}

void disable_percpu_irq(unsigned int irq)
{
	unsigned int cpu = smp_processor_id();
	unsigned long flags;
	struct irq_desc *desc = irq_get_desc_lock(irq, &flags, IRQ_GET_DESC_CHECK_PERCPU);

	if (!desc)
		return;

	irq_percpu_disable(desc, cpu);
	irq_put_desc_unlock(desc, flags);
}

/*
 * Internal function to unregister a percpu irqaction.
 */
static struct irqaction *__free_percpu_irq(unsigned int irq, void __percpu *dev_id)
{
	struct irq_desc *desc = irq_to_desc(irq);
	struct irqaction *action;
	unsigned long flags;

	WARN(in_interrupt(), "Trying to free IRQ %d from IRQ context!\n", irq);

	if (!desc)
		return NULL;

	raw_spin_lock_irqsave(&desc->lock, flags);

	action = desc->action;
	if (!action || action->percpu_dev_id != dev_id) {
		WARN(1, "Trying to free already-free IRQ %d\n", irq);
		goto bad;
	}

	if (!cpumask_empty(desc->percpu_enabled)) {
		WARN(1, "percpu IRQ %d still enabled on CPU%d!\n",
		     irq, cpumask_first(desc->percpu_enabled));
		goto bad;
	}

	/* Found it - now remove it from the list of entries: */
	desc->action = NULL;

	raw_spin_unlock_irqrestore(&desc->lock, flags);

	unregister_handler_proc(irq, action);

	module_put(desc->owner);
	return action;

bad:
	raw_spin_unlock_irqrestore(&desc->lock, flags);
	return NULL;
}

/**
 *	remove_percpu_irq - free a per-cpu interrupt
 *	@irq: Interrupt line to free
 *	@act: irqaction for the interrupt
 *
 * Used to remove interrupts statically setup by the early boot process.
 */
void remove_percpu_irq(unsigned int irq, struct irqaction *act)
{
	struct irq_desc *desc = irq_to_desc(irq);

	if (desc && irq_settings_is_per_cpu_devid(desc))
	    __free_percpu_irq(irq, act->percpu_dev_id);
}

/**
 *	free_percpu_irq - free an interrupt allocated with request_percpu_irq
 *	@irq: Interrupt line to free
 *	@dev_id: Device identity to free
 *
 *	Remove a percpu interrupt handler. The handler is removed, but
 *	the interrupt line is not disabled. This must be done on each
 *	CPU before calling this function. The function does not return
 *	until any executing interrupts for this IRQ have completed.
 *
 *	This function must not be called from interrupt context.
 */
void free_percpu_irq(unsigned int irq, void __percpu *dev_id)
{
	struct irq_desc *desc = irq_to_desc(irq);

	if (!desc || !irq_settings_is_per_cpu_devid(desc))
		return;

	chip_bus_lock(desc);
	kfree(__free_percpu_irq(irq, dev_id));
	chip_bus_sync_unlock(desc);
}

/**
 *	setup_percpu_irq - setup a per-cpu interrupt
 *	@irq: Interrupt line to setup
 *	@act: irqaction for the interrupt
 *
 * Used to statically setup per-cpu interrupts in the early boot process.
 */
int setup_percpu_irq(unsigned int irq, struct irqaction *act)
{
	struct irq_desc *desc = irq_to_desc(irq);
	int retval;

	if (!desc || !irq_settings_is_per_cpu_devid(desc))
		return -EINVAL;
	chip_bus_lock(desc);
	retval = __setup_irq(irq, desc, act);
	chip_bus_sync_unlock(desc);

	return retval;
}

/**
 *	request_percpu_irq - allocate a percpu interrupt line
 *	@irq: Interrupt line to allocate
 *	@handler: Function to be called when the IRQ occurs.
 *	@devname: An ascii name for the claiming device
 *	@dev_id: A percpu cookie passed back to the handler function
 *
 *	This call allocates interrupt resources, but doesn't
 *	automatically enable the interrupt. It has to be done on each
 *	CPU using enable_percpu_irq().
 *
 *	Dev_id must be globally unique. It is a per-cpu variable, and
 *	the handler gets called with the interrupted CPU's instance of
 *	that variable.
 */
int request_percpu_irq(unsigned int irq, irq_handler_t handler,
		       const char *devname, void __percpu *dev_id)
{
	struct irqaction *action;
	struct irq_desc *desc;
	int retval;

	if (!dev_id)
		return -EINVAL;

	desc = irq_to_desc(irq);
	if (!desc || !irq_settings_can_request(desc) ||
	    !irq_settings_is_per_cpu_devid(desc))
		return -EINVAL;

	action = kzalloc(sizeof(struct irqaction), GFP_KERNEL);
	if (!action)
		return -ENOMEM;

	action->handler = handler;
1679
	action->flags = IRQF_PERCPU | IRQF_NO_SUSPEND;
1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691
	action->name = devname;
	action->percpu_dev_id = dev_id;

	chip_bus_lock(desc);
	retval = __setup_irq(irq, desc, action);
	chip_bus_sync_unlock(desc);

	if (retval)
		kfree(action);

	return retval;
}