vector.c 24.1 KB
Newer Older
1 2 3 4 5
/*
 * Local APIC related interfaces to support IOAPIC, MSI, HT_IRQ etc.
 *
 * Copyright (C) 1997, 1998, 1999, 2000, 2009 Ingo Molnar, Hajnalka Szabo
 *	Moved from arch/x86/kernel/apic/io_apic.c.
6 7
 * Jiang Liu <jiang.liu@linux.intel.com>
 *	Enable support of hierarchical irqdomains
8 9 10 11 12 13
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 */
#include <linux/interrupt.h>
14
#include <linux/seq_file.h>
15 16 17
#include <linux/init.h>
#include <linux/compiler.h>
#include <linux/slab.h>
18
#include <asm/irqdomain.h>
19 20 21 22 23 24
#include <asm/hw_irq.h>
#include <asm/apic.h>
#include <asm/i8259.h>
#include <asm/desc.h>
#include <asm/irq_remapping.h>

25 26
#include <asm/trace/irq_vectors.h>

27
struct apic_chip_data {
28 29 30
	struct irq_cfg		hw_irq_cfg;
	unsigned int		vector;
	unsigned int		prev_vector;
31 32
	unsigned int		cpu;
	unsigned int		prev_cpu;
33
	unsigned int		irq;
34
	struct hlist_node	clist;
35 36 37
	u8			move_in_progress : 1;
};

38
struct irq_domain *x86_vector_domain;
39
EXPORT_SYMBOL_GPL(x86_vector_domain);
40
static DEFINE_RAW_SPINLOCK(vector_lock);
41
static cpumask_var_t vector_searchmask;
42
static struct irq_chip lapic_controller;
43
static struct irq_matrix *vector_matrix;
44 45 46
#ifdef CONFIG_SMP
static DEFINE_PER_CPU(struct hlist_head, cleanup_list);
#endif
47 48 49 50 51 52 53 54 55 56 57 58 59 60

void lock_vector_lock(void)
{
	/* Used to the online set of cpus does not change
	 * during assign_irq_vector.
	 */
	raw_spin_lock(&vector_lock);
}

void unlock_vector_lock(void)
{
	raw_spin_unlock(&vector_lock);
}

61 62 63 64 65 66 67 68 69 70 71 72 73 74 75
void init_irq_alloc_info(struct irq_alloc_info *info,
			 const struct cpumask *mask)
{
	memset(info, 0, sizeof(*info));
	info->mask = mask;
}

void copy_irq_alloc_info(struct irq_alloc_info *dst, struct irq_alloc_info *src)
{
	if (src)
		*dst = *src;
	else
		memset(dst, 0, sizeof(*dst));
}

76
static struct apic_chip_data *apic_chip_data(struct irq_data *irqd)
77
{
78
	if (!irqd)
79 80
		return NULL;

81 82
	while (irqd->parent_data)
		irqd = irqd->parent_data;
83

84
	return irqd->chip_data;
85 86
}

87
struct irq_cfg *irqd_cfg(struct irq_data *irqd)
88
{
89
	struct apic_chip_data *apicd = apic_chip_data(irqd);
90

91
	return apicd ? &apicd->hw_irq_cfg : NULL;
92
}
93
EXPORT_SYMBOL_GPL(irqd_cfg);
94 95

struct irq_cfg *irq_cfg(unsigned int irq)
96
{
97 98
	return irqd_cfg(irq_get_irq_data(irq));
}
99

100 101
static struct apic_chip_data *alloc_apic_chip_data(int node)
{
102
	struct apic_chip_data *apicd;
103

104
	apicd = kzalloc_node(sizeof(*apicd), GFP_KERNEL, node);
105 106
	if (apicd)
		INIT_HLIST_NODE(&apicd->clist);
107
	return apicd;
108 109
}

110
static void free_apic_chip_data(struct apic_chip_data *apicd)
111
{
112
	kfree(apicd);
113 114
}

115 116
static void apic_update_irq_cfg(struct irq_data *irqd, unsigned int vector,
				unsigned int cpu)
117
{
118
	struct apic_chip_data *apicd = apic_chip_data(irqd);
119

120
	lockdep_assert_held(&vector_lock);
121

122 123 124 125 126
	apicd->hw_irq_cfg.vector = vector;
	apicd->hw_irq_cfg.dest_apicid = apic->calc_dest_apicid(cpu);
	irq_data_update_effective_affinity(irqd, cpumask_of(cpu));
	trace_vector_config(irqd->irq, vector, cpu,
			    apicd->hw_irq_cfg.dest_apicid);
127
}
128

129 130 131 132 133
static void apic_update_vector(struct irq_data *irqd, unsigned int newvec,
			       unsigned int newcpu)
{
	struct apic_chip_data *apicd = apic_chip_data(irqd);
	struct irq_desc *desc = irq_data_to_desc(irqd);
134

135
	lockdep_assert_held(&vector_lock);
136

137
	trace_vector_update(irqd->irq, newvec, newcpu, apicd->vector,
138
			    apicd->cpu);
139

140
	/* Setup the vector move, if required  */
141
	if (apicd->vector && cpu_online(apicd->cpu)) {
142
		apicd->move_in_progress = true;
143
		apicd->prev_vector = apicd->vector;
144 145
		apicd->prev_cpu = apicd->cpu;
	} else {
146
		apicd->prev_vector = 0;
147
	}
148

149
	apicd->vector = newvec;
150 151 152 153
	apicd->cpu = newcpu;
	BUG_ON(!IS_ERR_OR_NULL(per_cpu(vector_irq, newcpu)[newvec]));
	per_cpu(vector_irq, newcpu)[newvec] = desc;
}
154

155 156 157 158
static int allocate_vector(struct irq_data *irqd, const struct cpumask *dest)
{
	struct apic_chip_data *apicd = apic_chip_data(irqd);
	unsigned int cpu = apicd->cpu;
159 160 161
	int vector = apicd->vector;

	lockdep_assert_held(&vector_lock);
162

163
	/*
164 165 166
	 * If the current target CPU is online and in the new requested
	 * affinity mask, there is no point in moving the interrupt from
	 * one CPU to another.
167
	 */
168 169 170 171 172 173 174 175 176 177 178 179 180
	if (vector && cpu_online(cpu) && cpumask_test_cpu(cpu, dest))
		return 0;

	vector = irq_matrix_alloc(vector_matrix, dest, false, &cpu);
	if (vector > 0)
		apic_update_vector(irqd, vector, cpu);
	trace_vector_alloc(irqd->irq, vector, false, vector);
	return vector;
}

static int assign_vector_locked(struct irq_data *irqd,
				const struct cpumask *dest)
{
181
	struct apic_chip_data *apicd = apic_chip_data(irqd);
182 183 184 185 186
	int vector = allocate_vector(irqd, dest);

	if (vector < 0)
		return vector;

187
	apic_update_irq_cfg(irqd, apicd->vector, apicd->cpu);
188
	return 0;
189 190
}

191
static int assign_irq_vector(struct irq_data *irqd, const struct cpumask *dest)
192 193
{
	unsigned long flags;
194
	int ret;
195 196

	raw_spin_lock_irqsave(&vector_lock, flags);
197 198
	cpumask_and(vector_searchmask, dest, cpu_online_mask);
	ret = assign_vector_locked(irqd, vector_searchmask);
199
	raw_spin_unlock_irqrestore(&vector_lock, flags);
200
	return ret;
201 202
}

203 204
static int assign_irq_vector_policy(struct irq_data *irqd,
				    struct irq_alloc_info *info, int node)
205
{
206
	if (info->mask)
207
		return assign_irq_vector(irqd, info->mask);
208
	if (node != NUMA_NO_NODE &&
209
	    !assign_irq_vector(irqd, cpumask_of_node(node)))
210
		return 0;
211
	return assign_irq_vector(irqd, cpu_online_mask);
212 213
}

214
static void clear_irq_vector(struct irq_data *irqd)
215
{
216
	struct apic_chip_data *apicd = apic_chip_data(irqd);
217
	unsigned int vector = apicd->vector;
218

219
	lockdep_assert_held(&vector_lock);
220

221
	if (!vector)
222
		return;
223

224
	trace_vector_clear(irqd->irq, vector, apicd->cpu, apicd->prev_vector,
225 226
			   apicd->prev_cpu);

227
	per_cpu(vector_irq, apicd->cpu)[vector] = VECTOR_UNUSED;
228
	irq_matrix_free(vector_matrix, apicd->cpu, vector, false);
229
	apicd->vector = 0;
230

231
	/* Clean up move in progress */
232
	vector = apicd->prev_vector;
233
	if (!vector)
234 235
		return;

236
	per_cpu(vector_irq, apicd->prev_cpu)[vector] = VECTOR_UNUSED;
237
	irq_matrix_free(vector_matrix, apicd->prev_cpu, vector, false);
238
	apicd->prev_vector = 0;
239
	apicd->move_in_progress = 0;
240
	hlist_del_init(&apicd->clist);
241 242
}

243 244 245
static void x86_vector_free_irqs(struct irq_domain *domain,
				 unsigned int virq, unsigned int nr_irqs)
{
246 247
	struct apic_chip_data *apicd;
	struct irq_data *irqd;
248
	unsigned long flags;
249 250 251
	int i;

	for (i = 0; i < nr_irqs; i++) {
252 253
		irqd = irq_domain_get_irq_data(x86_vector_domain, virq + i);
		if (irqd && irqd->chip_data) {
254
			raw_spin_lock_irqsave(&vector_lock, flags);
255
			clear_irq_vector(irqd);
256 257
			apicd = irqd->chip_data;
			irq_domain_reset_irq_data(irqd);
258
			raw_spin_unlock_irqrestore(&vector_lock, flags);
259
			free_apic_chip_data(apicd);
260 261 262 263 264 265 266 267
		}
	}
}

static int x86_vector_alloc_irqs(struct irq_domain *domain, unsigned int virq,
				 unsigned int nr_irqs, void *arg)
{
	struct irq_alloc_info *info = arg;
268 269
	struct apic_chip_data *apicd;
	struct irq_data *irqd;
270
	unsigned long flags;
271
	int i, err, node;
272 273 274 275 276 277 278 279 280

	if (disable_apic)
		return -ENXIO;

	/* Currently vector allocator can't guarantee contiguous allocations */
	if ((info->flags & X86_IRQ_ALLOC_CONTIGUOUS_VECTORS) && nr_irqs > 1)
		return -ENOSYS;

	for (i = 0; i < nr_irqs; i++) {
281 282 283
		irqd = irq_domain_get_irq_data(domain, virq + i);
		BUG_ON(!irqd);
		node = irq_data_get_node(irqd);
284 285
		WARN_ON_ONCE(irqd->chip_data);
		apicd = alloc_apic_chip_data(node);
286
		if (!apicd) {
287 288 289 290
			err = -ENOMEM;
			goto error;
		}

291
		apicd->irq = virq + i;
292 293 294 295
		irqd->chip = &lapic_controller;
		irqd->chip_data = apicd;
		irqd->hwirq = virq + i;
		irqd_set_single_target(irqd);
296
		/*
297 298 299 300 301
		 * Legacy vectors are already assigned when the IOAPIC
		 * takes them over. They stay on the same vector. This is
		 * required for check_timer() to work correctly as it might
		 * switch back to legacy mode. Only update the hardware
		 * config.
302 303
		 */
		if (info->flags & X86_IRQ_ALLOC_LEGACY) {
304
			apicd->vector = ISA_IRQ_VECTOR(virq + i);
305
			apicd->cpu = 0;
306 307
			trace_vector_setup(virq + i, true, 0);
			raw_spin_lock_irqsave(&vector_lock, flags);
308
			apic_update_irq_cfg(irqd, apicd->vector, apicd->cpu);
309 310
			raw_spin_unlock_irqrestore(&vector_lock, flags);
			continue;
311 312
		}

313 314
		err = assign_irq_vector_policy(irqd, info, node);
		trace_vector_setup(virq + i, false, err);
315 316 317 318 319 320 321 322 323 324 325
		if (err)
			goto error;
	}

	return 0;

error:
	x86_vector_free_irqs(domain, virq, i + 1);
	return err;
}

326 327 328 329
#ifdef CONFIG_GENERIC_IRQ_DEBUGFS
void x86_vector_debug_show(struct seq_file *m, struct irq_domain *d,
			   struct irq_data *irqd, int ind)
{
330
	unsigned int cpu, vector, prev_cpu, prev_vector;
331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354
	struct apic_chip_data *apicd;
	unsigned long flags;
	int irq;

	if (!irqd) {
		irq_matrix_debug_show(m, vector_matrix, ind);
		return;
	}

	irq = irqd->irq;
	if (irq < nr_legacy_irqs() && !test_bit(irq, &io_apic_irqs)) {
		seq_printf(m, "%*sVector: %5d\n", ind, "", ISA_IRQ_VECTOR(irq));
		seq_printf(m, "%*sTarget: Legacy PIC all CPUs\n", ind, "");
		return;
	}

	apicd = irqd->chip_data;
	if (!apicd) {
		seq_printf(m, "%*sVector: Not assigned\n", ind, "");
		return;
	}

	raw_spin_lock_irqsave(&vector_lock, flags);
	cpu = apicd->cpu;
355
	vector = apicd->vector;
356
	prev_cpu = apicd->prev_cpu;
357
	prev_vector = apicd->prev_vector;
358
	raw_spin_unlock_irqrestore(&vector_lock, flags);
359
	seq_printf(m, "%*sVector: %5u\n", ind, "", vector);
360
	seq_printf(m, "%*sTarget: %5u\n", ind, "", cpu);
361 362
	if (prev_vector) {
		seq_printf(m, "%*sPrevious vector: %5u\n", ind, "", prev_vector);
363 364 365 366 367
		seq_printf(m, "%*sPrevious target: %5u\n", ind, "", prev_cpu);
	}
}
#endif

T
Thomas Gleixner 已提交
368
static const struct irq_domain_ops x86_vector_domain_ops = {
369 370 371 372 373
	.alloc		= x86_vector_alloc_irqs,
	.free		= x86_vector_free_irqs,
#ifdef CONFIG_GENERIC_IRQ_DEBUGFS
	.debug_show	= x86_vector_debug_show,
#endif
374 375
};

376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395
int __init arch_probe_nr_irqs(void)
{
	int nr;

	if (nr_irqs > (NR_VECTORS * nr_cpu_ids))
		nr_irqs = NR_VECTORS * nr_cpu_ids;

	nr = (gsi_top + nr_legacy_irqs()) + 8 * nr_cpu_ids;
#if defined(CONFIG_PCI_MSI) || defined(CONFIG_HT_IRQ)
	/*
	 * for MSI and HT dyn irq
	 */
	if (gsi_top <= NR_IRQS_LEGACY)
		nr +=  8 * nr_cpu_ids;
	else
		nr += gsi_top * 16;
#endif
	if (nr < nr_irqs)
		nr_irqs = nr;

396 397 398 399 400
	/*
	 * We don't know if PIC is present at this point so we need to do
	 * probe() to get the right number of legacy IRQs.
	 */
	return legacy_pic->probe();
401 402
}

403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432
void lapic_assign_legacy_vector(unsigned int irq, bool replace)
{
	/*
	 * Use assign system here so it wont get accounted as allocated
	 * and moveable in the cpu hotplug check and it prevents managed
	 * irq reservation from touching it.
	 */
	irq_matrix_assign_system(vector_matrix, ISA_IRQ_VECTOR(irq), replace);
}

void __init lapic_assign_system_vectors(void)
{
	unsigned int i, vector = 0;

	for_each_set_bit_from(vector, system_vectors, NR_VECTORS)
		irq_matrix_assign_system(vector_matrix, vector, false);

	if (nr_legacy_irqs() > 1)
		lapic_assign_legacy_vector(PIC_CASCADE_IR, false);

	/* System vectors are reserved, online it */
	irq_matrix_online(vector_matrix);

	/* Mark the preallocated legacy interrupts */
	for (i = 0; i < nr_legacy_irqs(); i++) {
		if (i != PIC_CASCADE_IR)
			irq_matrix_assign(vector_matrix, ISA_IRQ_VECTOR(i));
	}
}

433 434
int __init arch_early_irq_init(void)
{
435 436 437 438 439 440
	struct fwnode_handle *fn;

	fn = irq_domain_alloc_named_fwnode("VECTOR");
	BUG_ON(!fn);
	x86_vector_domain = irq_domain_create_tree(fn, &x86_vector_domain_ops,
						   NULL);
441
	BUG_ON(x86_vector_domain == NULL);
442
	irq_domain_free_fwnode(fn);
443 444
	irq_set_default_host(x86_vector_domain);

445
	arch_init_msi_domain(x86_vector_domain);
446
	arch_init_htirq_domain(x86_vector_domain);
447

448
	BUG_ON(!alloc_cpumask_var(&vector_searchmask, GFP_KERNEL));
449

450 451 452 453 454 455 456 457
	/*
	 * Allocate the vector matrix allocator data structure and limit the
	 * search area.
	 */
	vector_matrix = irq_alloc_matrix(NR_VECTORS, FIRST_EXTERNAL_VECTOR,
					 FIRST_SYSTEM_VECTOR);
	BUG_ON(!vector_matrix);

458 459 460
	return arch_early_ioapic_init();
}

461
#ifdef CONFIG_SMP
462 463
/* Temporary hack to keep things working */
static void vector_update_shutdown_irqs(void)
464
{
465
	struct irq_desc *desc;
466
	int irq;
467

468
	for_each_irq_desc(irq, desc) {
469 470
		struct irq_data *irqd = irq_desc_get_irq_data(desc);
		struct apic_chip_data *ad = apic_chip_data(irqd);
471

472
		if (!ad || !ad->vector || ad->cpu != smp_processor_id())
473
			continue;
474 475
		this_cpu_write(vector_irq[ad->vector], desc);
		irq_matrix_assign(vector_matrix, ad->vector);
476 477 478
	}
}

479 480 481 482 483 484 485 486 487 488 489 490 491
static struct irq_desc *__setup_vector_irq(int vector)
{
	int isairq = vector - ISA_IRQ_VECTOR(0);

	/* Check whether the irq is in the legacy space */
	if (isairq < 0 || isairq >= nr_legacy_irqs())
		return VECTOR_UNUSED;
	/* Check whether the irq is handled by the IOAPIC */
	if (test_bit(isairq, &io_apic_irqs))
		return VECTOR_UNUSED;
	return irq_to_desc(isairq);
}

492 493
/* Online the local APIC infrastructure and initialize the vectors */
void lapic_online(void)
494
{
495
	unsigned int vector;
496

497
	lockdep_assert_held(&vector_lock);
498 499 500 501

	/* Online the vector matrix array for this CPU */
	irq_matrix_online(vector_matrix);

502
	/*
503 504 505 506 507 508 509
	 * The interrupt affinity logic never targets interrupts to offline
	 * CPUs. The exception are the legacy PIC interrupts. In general
	 * they are only targeted to CPU0, but depending on the platform
	 * they can be distributed to any online CPU in hardware. The
	 * kernel has no influence on that. So all active legacy vectors
	 * must be installed on all CPUs. All non legacy interrupts can be
	 * cleared.
510
	 */
511 512
	for (vector = 0; vector < NR_VECTORS; vector++)
		this_cpu_write(vector_irq[vector], __setup_vector_irq(vector));
513

514 515 516 517 518 519
	/*
	 * Until the rewrite of the managed interrupt management is in
	 * place it's necessary to walk the irq descriptors and check for
	 * interrupts which are targeted at this CPU.
	 */
	vector_update_shutdown_irqs();
520 521
}

522 523 524 525 526 527 528
void lapic_offline(void)
{
	lock_vector_lock();
	irq_matrix_offline(vector_matrix);
	unlock_vector_lock();
}

529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547
static int apic_set_affinity(struct irq_data *irqd,
			     const struct cpumask *dest, bool force)
{
	int err;

	if (!IS_ENABLED(CONFIG_SMP))
		return -EPERM;

	if (!cpumask_intersects(dest, cpu_online_mask))
		return -EINVAL;

	err = assign_irq_vector(irqd, dest);
	return err ? err : IRQ_SET_MASK_OK;
}

#else
# define apic_set_affinity	NULL
#endif

548
static int apic_retrigger_irq(struct irq_data *irqd)
549
{
550
	struct apic_chip_data *apicd = apic_chip_data(irqd);
551 552 553
	unsigned long flags;

	raw_spin_lock_irqsave(&vector_lock, flags);
554
	apic->send_IPI(apicd->cpu, apicd->vector);
555 556 557 558 559
	raw_spin_unlock_irqrestore(&vector_lock, flags);

	return 1;
}

560
void apic_ack_edge(struct irq_data *irqd)
561
{
562 563
	irq_complete_move(irqd_cfg(irqd));
	irq_move_irq(irqd);
564 565 566
	ack_APIC_irq();
}

567
static struct irq_chip lapic_controller = {
T
Thomas Gleixner 已提交
568
	.name			= "APIC",
569
	.irq_ack		= apic_ack_edge,
570
	.irq_set_affinity	= apic_set_affinity,
571 572 573
	.irq_retrigger		= apic_retrigger_irq,
};

574
#ifdef CONFIG_SMP
575

576 577
static void free_moved_vector(struct apic_chip_data *apicd)
{
578
	unsigned int vector = apicd->prev_vector;
579 580 581 582 583 584
	unsigned int cpu = apicd->prev_cpu;

	trace_vector_free_moved(apicd->irq, vector, false);
	irq_matrix_free(vector_matrix, cpu, vector, false);
	__this_cpu_write(vector_irq[vector], VECTOR_UNUSED);
	hlist_del_init(&apicd->clist);
585
	apicd->prev_vector = 0;
586 587 588
	apicd->move_in_progress = 0;
}

589
asmlinkage __visible void __irq_entry smp_irq_move_cleanup_interrupt(void)
590
{
591 592 593
	struct hlist_head *clhead = this_cpu_ptr(&cleanup_list);
	struct apic_chip_data *apicd;
	struct hlist_node *tmp;
594

595
	entering_ack_irq();
596 597 598
	/* Prevent vectors vanishing under us */
	raw_spin_lock(&vector_lock);

599
	hlist_for_each_entry_safe(apicd, tmp, clhead, clist) {
600
		unsigned int irr, vector = apicd->prev_vector;
601 602

		/*
603 604 605
		 * Paranoia: Check if the vector that needs to be cleaned
		 * up is registered at the APICs IRR. If so, then this is
		 * not the best time to clean it up. Clean it up in the
606
		 * next attempt by sending another IRQ_MOVE_CLEANUP_VECTOR
607 608 609
		 * to this CPU. IRQ_MOVE_CLEANUP_VECTOR is the lowest
		 * priority external vector, so on return from this
		 * interrupt the device interrupt will happen first.
610
		 */
611 612
		irr = apic_read(APIC_IRR + (vector / 32 * 0x10));
		if (irr & (1U << (vector % 32))) {
613
			apic->send_IPI_self(IRQ_MOVE_CLEANUP_VECTOR);
614
			continue;
615
		}
616
		free_moved_vector(apicd);
617 618
	}

619
	raw_spin_unlock(&vector_lock);
620
	exiting_irq();
621 622
}

623 624 625 626 627 628 629 630 631 632 633
static void __send_cleanup_vector(struct apic_chip_data *apicd)
{
	unsigned int cpu;

	raw_spin_lock(&vector_lock);
	apicd->move_in_progress = 0;
	cpu = apicd->prev_cpu;
	if (cpu_online(cpu)) {
		hlist_add_head(&apicd->clist, per_cpu_ptr(&cleanup_list, cpu));
		apic->send_IPI(cpu, IRQ_MOVE_CLEANUP_VECTOR);
	} else {
634
		apicd->prev_vector = 0;
635 636 637 638 639 640 641 642
	}
	raw_spin_unlock(&vector_lock);
}

void send_cleanup_vector(struct irq_cfg *cfg)
{
	struct apic_chip_data *apicd;

643
	apicd = container_of(cfg, struct apic_chip_data, hw_irq_cfg);
644 645 646 647
	if (apicd->move_in_progress)
		__send_cleanup_vector(apicd);
}

648 649
static void __irq_complete_move(struct irq_cfg *cfg, unsigned vector)
{
650
	struct apic_chip_data *apicd;
651

652
	apicd = container_of(cfg, struct apic_chip_data, hw_irq_cfg);
653
	if (likely(!apicd->move_in_progress))
654 655
		return;

656
	if (vector == apicd->vector && apicd->cpu == smp_processor_id())
657
		__send_cleanup_vector(apicd);
658 659 660 661 662 663 664
}

void irq_complete_move(struct irq_cfg *cfg)
{
	__irq_complete_move(cfg, ~get_irq_regs()->orig_ax);
}

665
/*
666
 * Called from fixup_irqs() with @desc->lock held and interrupts disabled.
667 668
 */
void irq_force_complete_move(struct irq_desc *desc)
669
{
670
	struct apic_chip_data *apicd;
671 672
	struct irq_data *irqd;
	unsigned int vector;
673

674 675 676 677 678 679 680 681 682
	/*
	 * The function is called for all descriptors regardless of which
	 * irqdomain they belong to. For example if an IRQ is provided by
	 * an irq_chip as part of a GPIO driver, the chip data for that
	 * descriptor is specific to the irq_chip in question.
	 *
	 * Check first that the chip_data is what we expect
	 * (apic_chip_data) before touching it any further.
	 */
683
	irqd = irq_domain_get_irq_data(x86_vector_domain,
684
				       irq_desc_get_irq(desc));
685
	if (!irqd)
686 687
		return;

688
	raw_spin_lock(&vector_lock);
689
	apicd = apic_chip_data(irqd);
690 691
	if (!apicd)
		goto unlock;
692

693
	/*
694
	 * If prev_vector is empty, no action required.
695
	 */
696
	vector = apicd->prev_vector;
697 698
	if (!vector)
		goto unlock;
699

700
	/*
701
	 * This is tricky. If the cleanup of the old vector has not been
702 703 704
	 * done yet, then the following setaffinity call will fail with
	 * -EBUSY. This can leave the interrupt in a stale state.
	 *
705 706
	 * All CPUs are stuck in stop machine with interrupts disabled so
	 * calling __irq_complete_move() would be completely pointless.
707
	 *
708 709 710 711 712 713 714
	 * 1) The interrupt is in move_in_progress state. That means that we
	 *    have not seen an interrupt since the io_apic was reprogrammed to
	 *    the new vector.
	 *
	 * 2) The interrupt has fired on the new vector, but the cleanup IPIs
	 *    have not been processed yet.
	 */
715
	if (apicd->move_in_progress) {
716
		/*
717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746
		 * In theory there is a race:
		 *
		 * set_ioapic(new_vector) <-- Interrupt is raised before update
		 *			      is effective, i.e. it's raised on
		 *			      the old vector.
		 *
		 * So if the target cpu cannot handle that interrupt before
		 * the old vector is cleaned up, we get a spurious interrupt
		 * and in the worst case the ioapic irq line becomes stale.
		 *
		 * But in case of cpu hotplug this should be a non issue
		 * because if the affinity update happens right before all
		 * cpus rendevouz in stop machine, there is no way that the
		 * interrupt can be blocked on the target cpu because all cpus
		 * loops first with interrupts enabled in stop machine, so the
		 * old vector is not yet cleaned up when the interrupt fires.
		 *
		 * So the only way to run into this issue is if the delivery
		 * of the interrupt on the apic/system bus would be delayed
		 * beyond the point where the target cpu disables interrupts
		 * in stop machine. I doubt that it can happen, but at least
		 * there is a theroretical chance. Virtualization might be
		 * able to expose this, but AFAICT the IOAPIC emulation is not
		 * as stupid as the real hardware.
		 *
		 * Anyway, there is nothing we can do about that at this point
		 * w/o refactoring the whole fixup_irq() business completely.
		 * We print at least the irq number and the old vector number,
		 * so we have the necessary information when a problem in that
		 * area arises.
747
		 */
748
		pr_warn("IRQ fixup: irq %d move in progress, old vector %d\n",
749
			irqd->irq, vector);
750
	}
751
	free_moved_vector(apicd);
752
unlock:
753
	raw_spin_unlock(&vector_lock);
754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773
}
#endif

static void __init print_APIC_field(int base)
{
	int i;

	printk(KERN_DEBUG);

	for (i = 0; i < 8; i++)
		pr_cont("%08x", apic_read(base + i*0x10));

	pr_cont("\n");
}

static void __init print_local_APIC(void *dummy)
{
	unsigned int i, v, ver, maxlvt;
	u64 icr;

774 775
	pr_debug("printing local APIC contents on CPU#%d/%d:\n",
		 smp_processor_id(), hard_smp_processor_id());
776
	v = apic_read(APIC_ID);
777
	pr_info("... APIC ID:      %08x (%01x)\n", v, read_apic_id());
778
	v = apic_read(APIC_LVR);
779
	pr_info("... APIC VERSION: %08x\n", v);
780 781 782 783
	ver = GET_APIC_VERSION(v);
	maxlvt = lapic_get_maxlvt();

	v = apic_read(APIC_TASKPRI);
784
	pr_debug("... APIC TASKPRI: %08x (%02x)\n", v, v & APIC_TPRI_MASK);
785 786 787 788 789

	/* !82489DX */
	if (APIC_INTEGRATED(ver)) {
		if (!APIC_XAPIC(ver)) {
			v = apic_read(APIC_ARBPRI);
790 791
			pr_debug("... APIC ARBPRI: %08x (%02x)\n",
				 v, v & APIC_ARBPRI_MASK);
792 793
		}
		v = apic_read(APIC_PROCPRI);
794
		pr_debug("... APIC PROCPRI: %08x\n", v);
795 796 797 798 799 800 801 802
	}

	/*
	 * Remote read supported only in the 82489DX and local APIC for
	 * Pentium processors.
	 */
	if (!APIC_INTEGRATED(ver) || maxlvt == 3) {
		v = apic_read(APIC_RRR);
803
		pr_debug("... APIC RRR: %08x\n", v);
804 805 806
	}

	v = apic_read(APIC_LDR);
807
	pr_debug("... APIC LDR: %08x\n", v);
808 809
	if (!x2apic_enabled()) {
		v = apic_read(APIC_DFR);
810
		pr_debug("... APIC DFR: %08x\n", v);
811 812
	}
	v = apic_read(APIC_SPIV);
813
	pr_debug("... APIC SPIV: %08x\n", v);
814

815
	pr_debug("... APIC ISR field:\n");
816
	print_APIC_field(APIC_ISR);
817
	pr_debug("... APIC TMR field:\n");
818
	print_APIC_field(APIC_TMR);
819
	pr_debug("... APIC IRR field:\n");
820 821 822 823 824 825 826 827 828
	print_APIC_field(APIC_IRR);

	/* !82489DX */
	if (APIC_INTEGRATED(ver)) {
		/* Due to the Pentium erratum 3AP. */
		if (maxlvt > 3)
			apic_write(APIC_ESR, 0);

		v = apic_read(APIC_ESR);
829
		pr_debug("... APIC ESR: %08x\n", v);
830 831 832
	}

	icr = apic_icr_read();
833 834
	pr_debug("... APIC ICR: %08x\n", (u32)icr);
	pr_debug("... APIC ICR2: %08x\n", (u32)(icr >> 32));
835 836

	v = apic_read(APIC_LVTT);
837
	pr_debug("... APIC LVTT: %08x\n", v);
838 839 840 841

	if (maxlvt > 3) {
		/* PC is LVT#4. */
		v = apic_read(APIC_LVTPC);
842
		pr_debug("... APIC LVTPC: %08x\n", v);
843 844
	}
	v = apic_read(APIC_LVT0);
845
	pr_debug("... APIC LVT0: %08x\n", v);
846
	v = apic_read(APIC_LVT1);
847
	pr_debug("... APIC LVT1: %08x\n", v);
848 849 850 851

	if (maxlvt > 2) {
		/* ERR is LVT#3. */
		v = apic_read(APIC_LVTERR);
852
		pr_debug("... APIC LVTERR: %08x\n", v);
853 854 855
	}

	v = apic_read(APIC_TMICT);
856
	pr_debug("... APIC TMICT: %08x\n", v);
857
	v = apic_read(APIC_TMCCT);
858
	pr_debug("... APIC TMCCT: %08x\n", v);
859
	v = apic_read(APIC_TDCR);
860
	pr_debug("... APIC TDCR: %08x\n", v);
861 862 863 864

	if (boot_cpu_has(X86_FEATURE_EXTAPIC)) {
		v = apic_read(APIC_EFEAT);
		maxlvt = (v >> 16) & 0xff;
865
		pr_debug("... APIC EFEAT: %08x\n", v);
866
		v = apic_read(APIC_ECTRL);
867
		pr_debug("... APIC ECTRL: %08x\n", v);
868 869
		for (i = 0; i < maxlvt; i++) {
			v = apic_read(APIC_EILVTn(i));
870
			pr_debug("... APIC EILVT%d: %08x\n", i, v);
871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899
		}
	}
	pr_cont("\n");
}

static void __init print_local_APICs(int maxcpu)
{
	int cpu;

	if (!maxcpu)
		return;

	preempt_disable();
	for_each_online_cpu(cpu) {
		if (cpu >= maxcpu)
			break;
		smp_call_function_single(cpu, print_local_APIC, NULL, 1);
	}
	preempt_enable();
}

static void __init print_PIC(void)
{
	unsigned int v;
	unsigned long flags;

	if (!nr_legacy_irqs())
		return;

900
	pr_debug("\nprinting PIC contents\n");
901 902 903 904

	raw_spin_lock_irqsave(&i8259A_lock, flags);

	v = inb(0xa1) << 8 | inb(0x21);
905
	pr_debug("... PIC  IMR: %04x\n", v);
906 907

	v = inb(0xa0) << 8 | inb(0x20);
908
	pr_debug("... PIC  IRR: %04x\n", v);
909 910 911 912 913 914 915 916 917

	outb(0x0b, 0xa0);
	outb(0x0b, 0x20);
	v = inb(0xa0) << 8 | inb(0x20);
	outb(0x0a, 0xa0);
	outb(0x0a, 0x20);

	raw_spin_unlock_irqrestore(&i8259A_lock, flags);

918
	pr_debug("... PIC  ISR: %04x\n", v);
919 920

	v = inb(0x4d1) << 8 | inb(0x4d0);
921
	pr_debug("... PIC ELCR: %04x\n", v);
922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948
}

static int show_lapic __initdata = 1;
static __init int setup_show_lapic(char *arg)
{
	int num = -1;

	if (strcmp(arg, "all") == 0) {
		show_lapic = CONFIG_NR_CPUS;
	} else {
		get_option(&arg, &num);
		if (num >= 0)
			show_lapic = num;
	}

	return 1;
}
__setup("show_lapic=", setup_show_lapic);

static int __init print_ICs(void)
{
	if (apic_verbosity == APIC_QUIET)
		return 0;

	print_PIC();

	/* don't print out if apic is not there */
949
	if (!boot_cpu_has(X86_FEATURE_APIC) && !apic_from_smp_config())
950 951 952 953 954 955 956 957 958
		return 0;

	print_local_APICs(show_lapic);
	print_IO_APICs();

	return 0;
}

late_initcall(print_ICs);