events.c 34.5 KB
Newer Older
J
Jeremy Fitzhardinge 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
/*
 * Xen event channels
 *
 * Xen models interrupts with abstract event channels.  Because each
 * domain gets 1024 event channels, but NR_IRQ is not that large, we
 * must dynamically map irqs<->event channels.  The event channels
 * interface with the rest of the kernel by defining a xen interrupt
 * chip.  When an event is recieved, it is mapped to an irq and sent
 * through the normal interrupt processing path.
 *
 * There are four kinds of events which can be mapped to an event
 * channel:
 *
 * 1. Inter-domain notifications.  This includes all the virtual
 *    device events, since they're driven by front-ends in another domain
 *    (typically dom0).
 * 2. VIRQs, typically used for timers.  These are per-cpu events.
 * 3. IPIs.
19
 * 4. PIRQs - Hardware interrupts.
J
Jeremy Fitzhardinge 已提交
20 21 22 23 24 25 26 27 28
 *
 * Jeremy Fitzhardinge <jeremy@xensource.com>, XenSource Inc, 2007
 */

#include <linux/linkage.h>
#include <linux/interrupt.h>
#include <linux/irq.h>
#include <linux/module.h>
#include <linux/string.h>
29
#include <linux/bootmem.h>
30
#include <linux/slab.h>
31
#include <linux/irqnr.h>
32
#include <linux/pci.h>
J
Jeremy Fitzhardinge 已提交
33

34
#include <asm/desc.h>
J
Jeremy Fitzhardinge 已提交
35 36
#include <asm/ptrace.h>
#include <asm/irq.h>
37
#include <asm/idle.h>
38
#include <asm/io_apic.h>
J
Jeremy Fitzhardinge 已提交
39
#include <asm/sync_bitops.h>
40
#include <asm/xen/pci.h>
J
Jeremy Fitzhardinge 已提交
41
#include <asm/xen/hypercall.h>
42
#include <asm/xen/hypervisor.h>
J
Jeremy Fitzhardinge 已提交
43

44 45
#include <xen/xen.h>
#include <xen/hvm.h>
46
#include <xen/xen-ops.h>
J
Jeremy Fitzhardinge 已提交
47 48 49
#include <xen/events.h>
#include <xen/interface/xen.h>
#include <xen/interface/event_channel.h>
50 51
#include <xen/interface/hvm/hvm_op.h>
#include <xen/interface/hvm/params.h>
J
Jeremy Fitzhardinge 已提交
52 53 54 55 56 57 58 59

/*
 * This lock protects updates to the following mapping and reference-count
 * arrays. The lock does not need to be acquired to read the mapping tables.
 */
static DEFINE_SPINLOCK(irq_mapping_update_lock);

/* IRQ <-> VIRQ mapping. */
60
static DEFINE_PER_CPU(int [NR_VIRQS], virq_to_irq) = {[0 ... NR_VIRQS-1] = -1};
J
Jeremy Fitzhardinge 已提交
61

J
Jeremy Fitzhardinge 已提交
62
/* IRQ <-> IPI mapping */
63
static DEFINE_PER_CPU(int [XEN_NR_IPIS], ipi_to_irq) = {[0 ... XEN_NR_IPIS-1] = -1};
J
Jeremy Fitzhardinge 已提交
64

65 66
/* Interrupt types. */
enum xen_irq_type {
J
Jeremy Fitzhardinge 已提交
67
	IRQT_UNBOUND = 0,
J
Jeremy Fitzhardinge 已提交
68 69 70 71 72
	IRQT_PIRQ,
	IRQT_VIRQ,
	IRQT_IPI,
	IRQT_EVTCHN
};
J
Jeremy Fitzhardinge 已提交
73

74 75 76 77 78 79
/*
 * Packed IRQ information:
 * type - enum xen_irq_type
 * event channel - irq->event channel mapping
 * cpu - cpu this event channel is bound to
 * index - type-specific information:
80 81
 *    PIRQ - vector, with MSB being "needs EIO", or physical IRQ of the HVM
 *           guest, or GSI (real passthrough IRQ) of the device.
82 83 84 85 86 87 88 89 90 91 92 93 94 95
 *    VIRQ - virq number
 *    IPI - IPI vector
 *    EVTCHN -
 */
struct irq_info
{
	enum xen_irq_type type;	/* type */
	unsigned short evtchn;	/* event channel */
	unsigned short cpu;	/* cpu bound */

	union {
		unsigned short virq;
		enum ipi_vector ipi;
		struct {
S
Stefano Stabellini 已提交
96
			unsigned short pirq;
97
			unsigned short gsi;
98 99
			unsigned char vector;
			unsigned char flags;
100 101 102
		} pirq;
	} u;
};
103
#define PIRQ_NEEDS_EOI	(1 << 0)
104
#define PIRQ_SHAREABLE	(1 << 1)
105

106
static struct irq_info *irq_info;
S
Stefano Stabellini 已提交
107
static int *pirq_to_irq;
J
Jeremy Fitzhardinge 已提交
108

109
static int *evtchn_to_irq;
110

111 112
static DEFINE_PER_CPU(unsigned long [NR_EVENT_CHANNELS/BITS_PER_LONG],
		      cpu_evtchn_mask);
J
Jeremy Fitzhardinge 已提交
113 114 115 116 117

/* Xen will never allocate port zero for any purpose. */
#define VALID_EVTCHN(chn)	((chn) != 0)

static struct irq_chip xen_dynamic_chip;
118
static struct irq_chip xen_percpu_chip;
119
static struct irq_chip xen_pirq_chip;
J
Jeremy Fitzhardinge 已提交
120

121 122
/* Get info for IRQ */
static struct irq_info *info_for_irq(unsigned irq)
123
{
124
	return &irq_info[irq];
125 126
}

127 128
/* Constructors for packed IRQ information. */
static void xen_irq_info_common_init(struct irq_info *info,
129
				     unsigned irq,
130 131 132
				     enum xen_irq_type type,
				     unsigned short evtchn,
				     unsigned short cpu)
133
{
134 135 136 137 138 139

	BUG_ON(info->type != IRQT_UNBOUND && info->type != type);

	info->type = type;
	info->evtchn = evtchn;
	info->cpu = cpu;
140 141

	evtchn_to_irq[evtchn] = irq;
142 143
}

144 145
static void xen_irq_info_evtchn_init(unsigned irq,
				     unsigned short evtchn)
J
Jeremy Fitzhardinge 已提交
146
{
147 148
	struct irq_info *info = info_for_irq(irq);

149
	xen_irq_info_common_init(info, irq, IRQT_EVTCHN, evtchn, 0);
150 151
}

152 153
static void xen_irq_info_ipi_init(unsigned cpu,
				  unsigned irq,
154 155
				  unsigned short evtchn,
				  enum ipi_vector ipi)
156
{
157 158
	struct irq_info *info = info_for_irq(irq);

159
	xen_irq_info_common_init(info, irq, IRQT_IPI, evtchn, 0);
160 161

	info->u.ipi = ipi;
162 163

	per_cpu(ipi_to_irq, cpu)[ipi] = irq;
164 165
}

166 167
static void xen_irq_info_virq_init(unsigned cpu,
				   unsigned irq,
168 169
				   unsigned short evtchn,
				   unsigned short virq)
170
{
171 172
	struct irq_info *info = info_for_irq(irq);

173
	xen_irq_info_common_init(info, irq, IRQT_VIRQ, evtchn, 0);
174 175

	info->u.virq = virq;
176 177

	per_cpu(virq_to_irq, cpu)[virq] = irq;
J
Jeremy Fitzhardinge 已提交
178 179
}

180 181 182 183 184 185
static void xen_irq_info_pirq_init(unsigned irq,
				   unsigned short evtchn,
				   unsigned short pirq,
				   unsigned short gsi,
				   unsigned short vector,
				   unsigned char flags)
J
Jeremy Fitzhardinge 已提交
186
{
187 188
	struct irq_info *info = info_for_irq(irq);

189
	xen_irq_info_common_init(info, irq, IRQT_PIRQ, evtchn, 0);
190 191 192 193 194

	info->u.pirq.pirq = pirq;
	info->u.pirq.gsi = gsi;
	info->u.pirq.vector = vector;
	info->u.pirq.flags = flags;
195 196

	pirq_to_irq[pirq] = irq;
J
Jeremy Fitzhardinge 已提交
197 198
}

199 200 201
/*
 * Accessors for packed IRQ information.
 */
202
static unsigned int evtchn_from_irq(unsigned irq)
J
Jeremy Fitzhardinge 已提交
203
{
204 205 206
	if (unlikely(WARN(irq < 0 || irq >= nr_irqs, "Invalid irq %d!\n", irq)))
		return 0;

207
	return info_for_irq(irq)->evtchn;
J
Jeremy Fitzhardinge 已提交
208 209
}

I
Ian Campbell 已提交
210 211 212 213 214 215
unsigned irq_from_evtchn(unsigned int evtchn)
{
	return evtchn_to_irq[evtchn];
}
EXPORT_SYMBOL_GPL(irq_from_evtchn);

216
static enum ipi_vector ipi_from_irq(unsigned irq)
J
Jeremy Fitzhardinge 已提交
217
{
218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235
	struct irq_info *info = info_for_irq(irq);

	BUG_ON(info == NULL);
	BUG_ON(info->type != IRQT_IPI);

	return info->u.ipi;
}

static unsigned virq_from_irq(unsigned irq)
{
	struct irq_info *info = info_for_irq(irq);

	BUG_ON(info == NULL);
	BUG_ON(info->type != IRQT_VIRQ);

	return info->u.virq;
}

S
Stefano Stabellini 已提交
236 237 238 239 240 241 242 243 244 245
static unsigned pirq_from_irq(unsigned irq)
{
	struct irq_info *info = info_for_irq(irq);

	BUG_ON(info == NULL);
	BUG_ON(info->type != IRQT_PIRQ);

	return info->u.pirq.pirq;
}

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
static unsigned gsi_from_irq(unsigned irq)
{
	struct irq_info *info = info_for_irq(irq);

	BUG_ON(info == NULL);
	BUG_ON(info->type != IRQT_PIRQ);

	return info->u.pirq.gsi;
}

static enum xen_irq_type type_from_irq(unsigned irq)
{
	return info_for_irq(irq)->type;
}

static unsigned cpu_from_irq(unsigned irq)
{
	return info_for_irq(irq)->cpu;
}

static unsigned int cpu_from_evtchn(unsigned int evtchn)
{
	int irq = evtchn_to_irq[evtchn];
	unsigned ret = 0;

	if (irq != -1)
		ret = cpu_from_irq(irq);

	return ret;
J
Jeremy Fitzhardinge 已提交
275 276
}

277 278 279 280 281 282 283 284 285
static bool pirq_needs_eoi(unsigned irq)
{
	struct irq_info *info = info_for_irq(irq);

	BUG_ON(info->type != IRQT_PIRQ);

	return info->u.pirq.flags & PIRQ_NEEDS_EOI;
}

J
Jeremy Fitzhardinge 已提交
286 287 288 289 290
static inline unsigned long active_evtchns(unsigned int cpu,
					   struct shared_info *sh,
					   unsigned int idx)
{
	return (sh->evtchn_pending[idx] &
291
		per_cpu(cpu_evtchn_mask, cpu)[idx] &
J
Jeremy Fitzhardinge 已提交
292 293 294 295 296 297 298 299 300
		~sh->evtchn_mask[idx]);
}

static void bind_evtchn_to_cpu(unsigned int chn, unsigned int cpu)
{
	int irq = evtchn_to_irq[chn];

	BUG_ON(irq == -1);
#ifdef CONFIG_SMP
301
	cpumask_copy(irq_to_desc(irq)->irq_data.affinity, cpumask_of(cpu));
J
Jeremy Fitzhardinge 已提交
302 303
#endif

304 305
	clear_bit(chn, per_cpu(cpu_evtchn_mask, cpu_from_irq(irq)));
	set_bit(chn, per_cpu(cpu_evtchn_mask, cpu));
J
Jeremy Fitzhardinge 已提交
306

307
	irq_info[irq].cpu = cpu;
J
Jeremy Fitzhardinge 已提交
308 309 310 311
}

static void init_evtchn_cpu_bindings(void)
{
312
	int i;
J
Jeremy Fitzhardinge 已提交
313
#ifdef CONFIG_SMP
314 315
	struct irq_desc *desc;

J
Jeremy Fitzhardinge 已提交
316
	/* By default all event channels notify CPU#0. */
317
	for_each_irq_desc(i, desc) {
318
		cpumask_copy(desc->irq_data.affinity, cpumask_of(0));
319
	}
J
Jeremy Fitzhardinge 已提交
320 321
#endif

322
	for_each_possible_cpu(i)
323 324
		memset(per_cpu(cpu_evtchn_mask, i),
		       (i == 0) ? ~0 : 0, sizeof(*per_cpu(cpu_evtchn_mask, i)));
325

J
Jeremy Fitzhardinge 已提交
326 327 328 329 330 331 332 333 334 335 336 337 338 339
}

static inline void clear_evtchn(int port)
{
	struct shared_info *s = HYPERVISOR_shared_info;
	sync_clear_bit(port, &s->evtchn_pending[0]);
}

static inline void set_evtchn(int port)
{
	struct shared_info *s = HYPERVISOR_shared_info;
	sync_set_bit(port, &s->evtchn_pending[0]);
}

340 341 342 343 344 345
static inline int test_evtchn(int port)
{
	struct shared_info *s = HYPERVISOR_shared_info;
	return sync_test_bit(port, &s->evtchn_pending[0]);
}

J
Jeremy Fitzhardinge 已提交
346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381

/**
 * notify_remote_via_irq - send event to remote end of event channel via irq
 * @irq: irq of event channel to send event to
 *
 * Unlike notify_remote_via_evtchn(), this is safe to use across
 * save/restore. Notifications on a broken connection are silently
 * dropped.
 */
void notify_remote_via_irq(int irq)
{
	int evtchn = evtchn_from_irq(irq);

	if (VALID_EVTCHN(evtchn))
		notify_remote_via_evtchn(evtchn);
}
EXPORT_SYMBOL_GPL(notify_remote_via_irq);

static void mask_evtchn(int port)
{
	struct shared_info *s = HYPERVISOR_shared_info;
	sync_set_bit(port, &s->evtchn_mask[0]);
}

static void unmask_evtchn(int port)
{
	struct shared_info *s = HYPERVISOR_shared_info;
	unsigned int cpu = get_cpu();

	BUG_ON(!irqs_disabled());

	/* Slow path (hypercall) if this is a non-local port. */
	if (unlikely(cpu != cpu_from_evtchn(port))) {
		struct evtchn_unmask unmask = { .port = port };
		(void)HYPERVISOR_event_channel_op(EVTCHNOP_unmask, &unmask);
	} else {
C
Christoph Lameter 已提交
382
		struct vcpu_info *vcpu_info = __this_cpu_read(xen_vcpu);
J
Jeremy Fitzhardinge 已提交
383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399

		sync_clear_bit(port, &s->evtchn_mask[0]);

		/*
		 * The following is basically the equivalent of
		 * 'hw_resend_irq'. Just like a real IO-APIC we 'lose
		 * the interrupt edge' if the channel is masked.
		 */
		if (sync_test_bit(port, &s->evtchn_pending[0]) &&
		    !sync_test_and_set_bit(port / BITS_PER_LONG,
					   &vcpu_info->evtchn_pending_sel))
			vcpu_info->evtchn_upcall_pending = 1;
	}

	put_cpu();
}

400
static int xen_allocate_irq_dynamic(void)
401
{
402 403
	int first = 0;
	int irq;
404 405

#ifdef CONFIG_X86_IO_APIC
406 407 408 409 410 411 412 413 414
	/*
	 * For an HVM guest or domain 0 which see "real" (emulated or
	 * actual repectively) GSIs we allocate dynamic IRQs
	 * e.g. those corresponding to event channels or MSIs
	 * etc. from the range above those "real" GSIs to avoid
	 * collisions.
	 */
	if (xen_initial_domain() || xen_hvm_domain())
		first = get_nr_irqs_gsi();
415 416
#endif

417 418
retry:
	irq = irq_alloc_desc_from(first, -1);
419

420 421 422 423
	if (irq == -ENOMEM && first > NR_IRQS_LEGACY) {
		printk(KERN_ERR "Out of dynamic IRQ space and eating into GSI space. You should increase nr_irqs\n");
		first = max(NR_IRQS_LEGACY, first - NR_IRQS_LEGACY);
		goto retry;
424
	}
J
Jeremy Fitzhardinge 已提交
425

426 427
	if (irq < 0)
		panic("No available IRQ to bind to: increase nr_irqs!\n");
428

J
Jeremy Fitzhardinge 已提交
429
	return irq;
430 431
}

432 433 434 435
static int xen_allocate_irq_gsi(unsigned gsi)
{
	int irq;

436 437 438 439 440 441 442
	/*
	 * A PV guest has no concept of a GSI (since it has no ACPI
	 * nor access to/knowledge of the physical APICs). Therefore
	 * all IRQs are dynamically allocated from the entire IRQ
	 * space.
	 */
	if (xen_pv_domain() && !xen_initial_domain())
443 444 445 446 447 448 449 450 451 452 453 454 455 456 457
		return xen_allocate_irq_dynamic();

	/* Legacy IRQ descriptors are already allocated by the arch. */
	if (gsi < NR_IRQS_LEGACY)
		return gsi;

	irq = irq_alloc_desc_at(gsi, -1);
	if (irq < 0)
		panic("Unable to allocate to IRQ%d (%d)\n", gsi, irq);

	return irq;
}

static void xen_free_irq(unsigned irq)
{
458 459
	irq_info[irq].type = IRQT_UNBOUND;

460 461 462 463
	/* Legacy IRQ descriptors are managed by the arch. */
	if (irq < NR_IRQS_LEGACY)
		return;

464 465 466
	irq_free_desc(irq);
}

467 468
static void pirq_unmask_notify(int irq)
{
S
Stefano Stabellini 已提交
469
	struct physdev_eoi eoi = { .irq = pirq_from_irq(irq) };
470 471 472 473 474 475 476 477 478 479 480 481 482 483

	if (unlikely(pirq_needs_eoi(irq))) {
		int rc = HYPERVISOR_physdev_op(PHYSDEVOP_eoi, &eoi);
		WARN_ON(rc);
	}
}

static void pirq_query_unmask(int irq)
{
	struct physdev_irq_status_query irq_status;
	struct irq_info *info = info_for_irq(irq);

	BUG_ON(info->type != IRQT_PIRQ);

S
Stefano Stabellini 已提交
484
	irq_status.irq = pirq_from_irq(irq);
485 486 487 488 489 490 491 492 493 494 495 496 497 498 499
	if (HYPERVISOR_physdev_op(PHYSDEVOP_irq_status_query, &irq_status))
		irq_status.flags = 0;

	info->u.pirq.flags &= ~PIRQ_NEEDS_EOI;
	if (irq_status.flags & XENIRQSTAT_needs_eoi)
		info->u.pirq.flags |= PIRQ_NEEDS_EOI;
}

static bool probing_irq(int irq)
{
	struct irq_desc *desc = irq_to_desc(irq);

	return desc && desc->action == NULL;
}

500
static unsigned int __startup_pirq(unsigned int irq)
501 502 503 504
{
	struct evtchn_bind_pirq bind_pirq;
	struct irq_info *info = info_for_irq(irq);
	int evtchn = evtchn_from_irq(irq);
505
	int rc;
506 507 508 509 510 511

	BUG_ON(info->type != IRQT_PIRQ);

	if (VALID_EVTCHN(evtchn))
		goto out;

S
Stefano Stabellini 已提交
512
	bind_pirq.pirq = pirq_from_irq(irq);
513
	/* NB. We are happy to share unless we are probing. */
514 515 516 517
	bind_pirq.flags = info->u.pirq.flags & PIRQ_SHAREABLE ?
					BIND_PIRQ__WILL_SHARE : 0;
	rc = HYPERVISOR_event_channel_op(EVTCHNOP_bind_pirq, &bind_pirq);
	if (rc != 0) {
518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537
		if (!probing_irq(irq))
			printk(KERN_INFO "Failed to obtain physical IRQ %d\n",
			       irq);
		return 0;
	}
	evtchn = bind_pirq.port;

	pirq_query_unmask(irq);

	evtchn_to_irq[evtchn] = irq;
	bind_evtchn_to_cpu(evtchn, 0);
	info->evtchn = evtchn;

out:
	unmask_evtchn(evtchn);
	pirq_unmask_notify(irq);

	return 0;
}

538 539 540 541 542 543
static unsigned int startup_pirq(struct irq_data *data)
{
	return __startup_pirq(data->irq);
}

static void shutdown_pirq(struct irq_data *data)
544 545
{
	struct evtchn_close close;
546
	unsigned int irq = data->irq;
547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565
	struct irq_info *info = info_for_irq(irq);
	int evtchn = evtchn_from_irq(irq);

	BUG_ON(info->type != IRQT_PIRQ);

	if (!VALID_EVTCHN(evtchn))
		return;

	mask_evtchn(evtchn);

	close.port = evtchn;
	if (HYPERVISOR_event_channel_op(EVTCHNOP_close, &close) != 0)
		BUG();

	bind_evtchn_to_cpu(evtchn, 0);
	evtchn_to_irq[evtchn] = -1;
	info->evtchn = 0;
}

566
static void enable_pirq(struct irq_data *data)
567
{
568
	startup_pirq(data);
569 570
}

571
static void disable_pirq(struct irq_data *data)
572 573 574
{
}

575
static void ack_pirq(struct irq_data *data)
576
{
577
	int evtchn = evtchn_from_irq(data->irq);
578

579
	move_native_irq(data->irq);
580 581 582 583 584 585 586 587 588 589 590

	if (VALID_EVTCHN(evtchn)) {
		mask_evtchn(evtchn);
		clear_evtchn(evtchn);
	}
}

static int find_irq_by_gsi(unsigned gsi)
{
	int irq;

591
	for (irq = 0; irq < nr_irqs; irq++) {
592 593 594 595 596 597 598 599 600 601 602 603
		struct irq_info *info = info_for_irq(irq);

		if (info == NULL || info->type != IRQT_PIRQ)
			continue;

		if (gsi_from_irq(irq) == gsi)
			return irq;
	}

	return -1;
}

604
int xen_allocate_pirq_gsi(unsigned gsi)
S
Stefano Stabellini 已提交
605
{
606
	return gsi;
S
Stefano Stabellini 已提交
607 608
}

I
Ian Campbell 已提交
609 610 611
/*
 * Do not make any assumptions regarding the relationship between the
 * IRQ number returned here and the Xen pirq argument.
S
Stefano Stabellini 已提交
612 613 614
 *
 * Note: We don't assign an event channel until the irq actually started
 * up.  Return an existing irq if we've already got one for the gsi.
615
 */
616 617
int xen_bind_pirq_gsi_to_irq(unsigned gsi,
			     unsigned pirq, int shareable, char *name)
618
{
619
	int irq = -1;
620 621 622 623
	struct physdev_irq irq_op;

	spin_lock(&irq_mapping_update_lock);

624
	if ((pirq > nr_irqs) || (gsi > nr_irqs)) {
625
		printk(KERN_WARNING "xen_map_pirq_gsi: %s %s is incorrect!\n",
626 627
			pirq > nr_irqs ? "pirq" :"",
			gsi > nr_irqs ? "gsi" : "");
628 629 630
		goto out;
	}

631 632
	irq = find_irq_by_gsi(gsi);
	if (irq != -1) {
S
Stefano Stabellini 已提交
633
		printk(KERN_INFO "xen_map_pirq_gsi: returning irq %d for gsi %u\n",
634 635 636 637
		       irq, gsi);
		goto out;	/* XXX need refcount? */
	}

638
	irq = xen_allocate_irq_gsi(gsi);
639 640

	set_irq_chip_and_handler_name(irq, &xen_pirq_chip,
641
				      handle_level_irq, name);
642 643

	irq_op.irq = irq;
644 645 646 647 648 649 650
	irq_op.vector = 0;

	/* Only the privileged domain can do this. For non-priv, the pcifront
	 * driver provides a PCI bus that does the call to do exactly
	 * this in the priv domain. */
	if (xen_initial_domain() &&
	    HYPERVISOR_physdev_op(PHYSDEVOP_alloc_irq_vector, &irq_op)) {
651
		xen_free_irq(irq);
652 653 654 655
		irq = -ENOSPC;
		goto out;
	}

656 657
	xen_irq_info_pirq_init(irq, 0, pirq, gsi, irq_op.vector,
			       shareable ? PIRQ_SHAREABLE : 0);
658 659 660 661 662 663 664

out:
	spin_unlock(&irq_mapping_update_lock);

	return irq;
}

665
#ifdef CONFIG_PCI_MSI
666
int xen_allocate_pirq_msi(struct pci_dev *dev, struct msi_desc *msidesc)
667
{
668
	int rc;
669 670
	struct physdev_get_free_pirq op_get_free_pirq;

671
	op_get_free_pirq.type = MAP_PIRQ_TYPE_MSI;
672 673
	rc = HYPERVISOR_physdev_op(PHYSDEVOP_get_free_pirq, &op_get_free_pirq);

674 675 676 677
	WARN_ONCE(rc == -ENOSYS,
		  "hypervisor does not support the PHYSDEVOP_get_free_pirq interface\n");

	return rc ? -1 : op_get_free_pirq.pirq;
678 679
}

680
int xen_bind_pirq_msi_to_irq(struct pci_dev *dev, struct msi_desc *msidesc,
681
			     int pirq, int vector, const char *name)
S
Stefano Stabellini 已提交
682
{
683
	int irq, ret;
684

S
Stefano Stabellini 已提交
685 686
	spin_lock(&irq_mapping_update_lock);

687 688
	irq = xen_allocate_irq_dynamic();
	if (irq == -1)
689
		goto out;
S
Stefano Stabellini 已提交
690

691
	set_irq_chip_and_handler_name(irq, &xen_pirq_chip,
S
Stefano Stabellini 已提交
692 693
				      handle_level_irq, name);

694
	xen_irq_info_pirq_init(irq, 0, pirq, 0, vector, 0);
695 696 697
	ret = set_irq_msi(irq, msidesc);
	if (ret < 0)
		goto error_irq;
S
Stefano Stabellini 已提交
698 699
out:
	spin_unlock(&irq_mapping_update_lock);
700
	return irq;
701 702 703 704
error_irq:
	spin_unlock(&irq_mapping_update_lock);
	xen_free_irq(irq);
	return -1;
S
Stefano Stabellini 已提交
705
}
706 707
#endif

708 709 710
int xen_destroy_irq(int irq)
{
	struct irq_desc *desc;
711 712
	struct physdev_unmap_pirq unmap_irq;
	struct irq_info *info = info_for_irq(irq);
713 714 715 716 717 718 719 720
	int rc = -ENOENT;

	spin_lock(&irq_mapping_update_lock);

	desc = irq_to_desc(irq);
	if (!desc)
		goto out;

721
	if (xen_initial_domain()) {
722
		unmap_irq.pirq = info->u.pirq.pirq;
723 724 725 726 727 728 729
		unmap_irq.domid = DOMID_SELF;
		rc = HYPERVISOR_physdev_op(PHYSDEVOP_unmap_pirq, &unmap_irq);
		if (rc) {
			printk(KERN_WARNING "unmap irq failed %d\n", rc);
			goto out;
		}
	}
730 731
	pirq_to_irq[info->u.pirq.pirq] = -1;

732
	xen_free_irq(irq);
733 734 735 736 737 738

out:
	spin_unlock(&irq_mapping_update_lock);
	return rc;
}

739 740 741 742 743
int xen_irq_from_pirq(unsigned pirq)
{
	return pirq_to_irq[pirq];
}

744
int bind_evtchn_to_irq(unsigned int evtchn)
J
Jeremy Fitzhardinge 已提交
745 746 747 748 749 750 751 752
{
	int irq;

	spin_lock(&irq_mapping_update_lock);

	irq = evtchn_to_irq[evtchn];

	if (irq == -1) {
753
		irq = xen_allocate_irq_dynamic();
J
Jeremy Fitzhardinge 已提交
754 755

		set_irq_chip_and_handler_name(irq, &xen_dynamic_chip,
756
					      handle_fasteoi_irq, "event");
J
Jeremy Fitzhardinge 已提交
757

758
		xen_irq_info_evtchn_init(irq, evtchn);
J
Jeremy Fitzhardinge 已提交
759 760 761 762 763 764
	}

	spin_unlock(&irq_mapping_update_lock);

	return irq;
}
765
EXPORT_SYMBOL_GPL(bind_evtchn_to_irq);
J
Jeremy Fitzhardinge 已提交
766

J
Jeremy Fitzhardinge 已提交
767 768 769 770 771 772 773 774
static int bind_ipi_to_irq(unsigned int ipi, unsigned int cpu)
{
	struct evtchn_bind_ipi bind_ipi;
	int evtchn, irq;

	spin_lock(&irq_mapping_update_lock);

	irq = per_cpu(ipi_to_irq, cpu)[ipi];
775

J
Jeremy Fitzhardinge 已提交
776
	if (irq == -1) {
777
		irq = xen_allocate_irq_dynamic();
J
Jeremy Fitzhardinge 已提交
778 779 780
		if (irq < 0)
			goto out;

781 782
		set_irq_chip_and_handler_name(irq, &xen_percpu_chip,
					      handle_percpu_irq, "ipi");
J
Jeremy Fitzhardinge 已提交
783 784 785 786 787 788 789

		bind_ipi.vcpu = cpu;
		if (HYPERVISOR_event_channel_op(EVTCHNOP_bind_ipi,
						&bind_ipi) != 0)
			BUG();
		evtchn = bind_ipi.port;

790
		xen_irq_info_ipi_init(cpu, irq, evtchn, ipi);
J
Jeremy Fitzhardinge 已提交
791 792 793 794 795 796 797 798 799 800

		bind_evtchn_to_cpu(evtchn, cpu);
	}

 out:
	spin_unlock(&irq_mapping_update_lock);
	return irq;
}


801
int bind_virq_to_irq(unsigned int virq, unsigned int cpu)
J
Jeremy Fitzhardinge 已提交
802 803 804 805 806 807 808 809 810
{
	struct evtchn_bind_virq bind_virq;
	int evtchn, irq;

	spin_lock(&irq_mapping_update_lock);

	irq = per_cpu(virq_to_irq, cpu)[virq];

	if (irq == -1) {
811
		irq = xen_allocate_irq_dynamic();
812 813 814 815

		set_irq_chip_and_handler_name(irq, &xen_percpu_chip,
					      handle_percpu_irq, "virq");

J
Jeremy Fitzhardinge 已提交
816 817 818 819 820 821 822
		bind_virq.virq = virq;
		bind_virq.vcpu = cpu;
		if (HYPERVISOR_event_channel_op(EVTCHNOP_bind_virq,
						&bind_virq) != 0)
			BUG();
		evtchn = bind_virq.port;

823
		xen_irq_info_virq_init(cpu, irq, evtchn, virq);
J
Jeremy Fitzhardinge 已提交
824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839

		bind_evtchn_to_cpu(evtchn, cpu);
	}

	spin_unlock(&irq_mapping_update_lock);

	return irq;
}

static void unbind_from_irq(unsigned int irq)
{
	struct evtchn_close close;
	int evtchn = evtchn_from_irq(irq);

	spin_lock(&irq_mapping_update_lock);

J
Jeremy Fitzhardinge 已提交
840
	if (VALID_EVTCHN(evtchn)) {
J
Jeremy Fitzhardinge 已提交
841 842 843 844 845 846 847
		close.port = evtchn;
		if (HYPERVISOR_event_channel_op(EVTCHNOP_close, &close) != 0)
			BUG();

		switch (type_from_irq(irq)) {
		case IRQT_VIRQ:
			per_cpu(virq_to_irq, cpu_from_evtchn(evtchn))
848
				[virq_from_irq(irq)] = -1;
J
Jeremy Fitzhardinge 已提交
849
			break;
A
Alex Nixon 已提交
850 851
		case IRQT_IPI:
			per_cpu(ipi_to_irq, cpu_from_evtchn(evtchn))
852
				[ipi_from_irq(irq)] = -1;
A
Alex Nixon 已提交
853
			break;
J
Jeremy Fitzhardinge 已提交
854 855 856 857 858 859 860 861
		default:
			break;
		}

		/* Closed ports are implicitly re-bound to VCPU0. */
		bind_evtchn_to_cpu(evtchn, 0);

		evtchn_to_irq[evtchn] = -1;
862 863
	}

864
	BUG_ON(irq_info[irq].type == IRQT_UNBOUND);
J
Jeremy Fitzhardinge 已提交
865

866
	xen_free_irq(irq);
J
Jeremy Fitzhardinge 已提交
867 868 869 870 871

	spin_unlock(&irq_mapping_update_lock);
}

int bind_evtchn_to_irqhandler(unsigned int evtchn,
872
			      irq_handler_t handler,
J
Jeremy Fitzhardinge 已提交
873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890
			      unsigned long irqflags,
			      const char *devname, void *dev_id)
{
	unsigned int irq;
	int retval;

	irq = bind_evtchn_to_irq(evtchn);
	retval = request_irq(irq, handler, irqflags, devname, dev_id);
	if (retval != 0) {
		unbind_from_irq(irq);
		return retval;
	}

	return irq;
}
EXPORT_SYMBOL_GPL(bind_evtchn_to_irqhandler);

int bind_virq_to_irqhandler(unsigned int virq, unsigned int cpu,
891
			    irq_handler_t handler,
J
Jeremy Fitzhardinge 已提交
892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907
			    unsigned long irqflags, const char *devname, void *dev_id)
{
	unsigned int irq;
	int retval;

	irq = bind_virq_to_irq(virq, cpu);
	retval = request_irq(irq, handler, irqflags, devname, dev_id);
	if (retval != 0) {
		unbind_from_irq(irq);
		return retval;
	}

	return irq;
}
EXPORT_SYMBOL_GPL(bind_virq_to_irqhandler);

J
Jeremy Fitzhardinge 已提交
908 909 910 911 912 913 914 915 916 917 918 919 920
int bind_ipi_to_irqhandler(enum ipi_vector ipi,
			   unsigned int cpu,
			   irq_handler_t handler,
			   unsigned long irqflags,
			   const char *devname,
			   void *dev_id)
{
	int irq, retval;

	irq = bind_ipi_to_irq(ipi, cpu);
	if (irq < 0)
		return irq;

T
Thomas Gleixner 已提交
921
	irqflags |= IRQF_NO_SUSPEND | IRQF_FORCE_RESUME;
J
Jeremy Fitzhardinge 已提交
922 923 924 925 926 927 928 929 930
	retval = request_irq(irq, handler, irqflags, devname, dev_id);
	if (retval != 0) {
		unbind_from_irq(irq);
		return retval;
	}

	return irq;
}

J
Jeremy Fitzhardinge 已提交
931 932 933 934 935 936 937
void unbind_from_irqhandler(unsigned int irq, void *dev_id)
{
	free_irq(irq, dev_id);
	unbind_from_irq(irq);
}
EXPORT_SYMBOL_GPL(unbind_from_irqhandler);

J
Jeremy Fitzhardinge 已提交
938 939 940 941 942 943 944
void xen_send_IPI_one(unsigned int cpu, enum ipi_vector vector)
{
	int irq = per_cpu(ipi_to_irq, cpu)[vector];
	BUG_ON(irq < 0);
	notify_remote_via_irq(irq);
}

945 946 947 948
irqreturn_t xen_debug_interrupt(int irq, void *dev_id)
{
	struct shared_info *sh = HYPERVISOR_shared_info;
	int cpu = smp_processor_id();
949
	unsigned long *cpu_evtchn = per_cpu(cpu_evtchn_mask, cpu);
950 951 952
	int i;
	unsigned long flags;
	static DEFINE_SPINLOCK(debug_lock);
953
	struct vcpu_info *v;
954 955 956

	spin_lock_irqsave(&debug_lock, flags);

957
	printk("\nvcpu %d\n  ", cpu);
958 959

	for_each_online_cpu(i) {
960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002
		int pending;
		v = per_cpu(xen_vcpu, i);
		pending = (get_irq_regs() && i == cpu)
			? xen_irqs_disabled(get_irq_regs())
			: v->evtchn_upcall_mask;
		printk("%d: masked=%d pending=%d event_sel %0*lx\n  ", i,
		       pending, v->evtchn_upcall_pending,
		       (int)(sizeof(v->evtchn_pending_sel)*2),
		       v->evtchn_pending_sel);
	}
	v = per_cpu(xen_vcpu, cpu);

	printk("\npending:\n   ");
	for (i = ARRAY_SIZE(sh->evtchn_pending)-1; i >= 0; i--)
		printk("%0*lx%s", (int)sizeof(sh->evtchn_pending[0])*2,
		       sh->evtchn_pending[i],
		       i % 8 == 0 ? "\n   " : " ");
	printk("\nglobal mask:\n   ");
	for (i = ARRAY_SIZE(sh->evtchn_mask)-1; i >= 0; i--)
		printk("%0*lx%s",
		       (int)(sizeof(sh->evtchn_mask[0])*2),
		       sh->evtchn_mask[i],
		       i % 8 == 0 ? "\n   " : " ");

	printk("\nglobally unmasked:\n   ");
	for (i = ARRAY_SIZE(sh->evtchn_mask)-1; i >= 0; i--)
		printk("%0*lx%s", (int)(sizeof(sh->evtchn_mask[0])*2),
		       sh->evtchn_pending[i] & ~sh->evtchn_mask[i],
		       i % 8 == 0 ? "\n   " : " ");

	printk("\nlocal cpu%d mask:\n   ", cpu);
	for (i = (NR_EVENT_CHANNELS/BITS_PER_LONG)-1; i >= 0; i--)
		printk("%0*lx%s", (int)(sizeof(cpu_evtchn[0])*2),
		       cpu_evtchn[i],
		       i % 8 == 0 ? "\n   " : " ");

	printk("\nlocally unmasked:\n   ");
	for (i = ARRAY_SIZE(sh->evtchn_mask)-1; i >= 0; i--) {
		unsigned long pending = sh->evtchn_pending[i]
			& ~sh->evtchn_mask[i]
			& cpu_evtchn[i];
		printk("%0*lx%s", (int)(sizeof(sh->evtchn_mask[0])*2),
		       pending, i % 8 == 0 ? "\n   " : " ");
1003 1004 1005
	}

	printk("\npending list:\n");
1006
	for (i = 0; i < NR_EVENT_CHANNELS; i++) {
1007
		if (sync_test_bit(i, sh->evtchn_pending)) {
1008 1009
			int word_idx = i / BITS_PER_LONG;
			printk("  %d: event %d -> irq %d%s%s%s\n",
1010
			       cpu_from_evtchn(i), i,
1011 1012 1013 1014 1015 1016 1017
			       evtchn_to_irq[i],
			       sync_test_bit(word_idx, &v->evtchn_pending_sel)
					     ? "" : " l2-clear",
			       !sync_test_bit(i, sh->evtchn_mask)
					     ? "" : " globally-masked",
			       sync_test_bit(i, cpu_evtchn)
					     ? "" : " locally-masked");
1018 1019 1020 1021 1022 1023 1024 1025
		}
	}

	spin_unlock_irqrestore(&debug_lock, flags);

	return IRQ_HANDLED;
}

1026 1027
static DEFINE_PER_CPU(unsigned, xed_nesting_count);

J
Jeremy Fitzhardinge 已提交
1028 1029 1030 1031 1032 1033 1034 1035 1036
/*
 * Search the CPUs pending events bitmasks.  For each one found, map
 * the event number to an irq, and feed it into do_IRQ() for
 * handling.
 *
 * Xen uses a two-level bitmap to speed searching.  The first level is
 * a bitset of words which contain pending event bits.  The second
 * level is a bitset of pending events themselves.
 */
1037
static void __xen_evtchn_do_upcall(void)
J
Jeremy Fitzhardinge 已提交
1038 1039 1040
{
	int cpu = get_cpu();
	struct shared_info *s = HYPERVISOR_shared_info;
C
Christoph Lameter 已提交
1041
	struct vcpu_info *vcpu_info = __this_cpu_read(xen_vcpu);
1042
 	unsigned count;
J
Jeremy Fitzhardinge 已提交
1043

1044 1045
	do {
		unsigned long pending_words;
J
Jeremy Fitzhardinge 已提交
1046

1047
		vcpu_info->evtchn_upcall_pending = 0;
J
Jeremy Fitzhardinge 已提交
1048

C
Christoph Lameter 已提交
1049
		if (__this_cpu_inc_return(xed_nesting_count) - 1)
1050
			goto out;
J
Jeremy Fitzhardinge 已提交
1051

1052 1053
#ifndef CONFIG_X86 /* No need for a barrier -- XCHG is a barrier on x86. */
		/* Clear master flag /before/ clearing selector flag. */
1054
		wmb();
1055
#endif
1056 1057 1058 1059 1060 1061 1062 1063 1064 1065
		pending_words = xchg(&vcpu_info->evtchn_pending_sel, 0);
		while (pending_words != 0) {
			unsigned long pending_bits;
			int word_idx = __ffs(pending_words);
			pending_words &= ~(1UL << word_idx);

			while ((pending_bits = active_evtchns(cpu, s, word_idx)) != 0) {
				int bit_idx = __ffs(pending_bits);
				int port = (word_idx * BITS_PER_LONG) + bit_idx;
				int irq = evtchn_to_irq[port];
1066
				struct irq_desc *desc;
1067

1068 1069 1070
				mask_evtchn(port);
				clear_evtchn(port);

1071 1072 1073 1074 1075
				if (irq != -1) {
					desc = irq_to_desc(irq);
					if (desc)
						generic_handle_irq_desc(irq, desc);
				}
J
Jeremy Fitzhardinge 已提交
1076 1077 1078
			}
		}

1079 1080
		BUG_ON(!irqs_disabled());

C
Christoph Lameter 已提交
1081 1082
		count = __this_cpu_read(xed_nesting_count);
		__this_cpu_write(xed_nesting_count, 0);
1083
	} while (count != 1 || vcpu_info->evtchn_upcall_pending);
1084 1085

out:
1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098

	put_cpu();
}

void xen_evtchn_do_upcall(struct pt_regs *regs)
{
	struct pt_regs *old_regs = set_irq_regs(regs);

	exit_idle();
	irq_enter();

	__xen_evtchn_do_upcall();

1099 1100
	irq_exit();
	set_irq_regs(old_regs);
1101
}
1102

1103 1104 1105
void xen_hvm_evtchn_do_upcall(void)
{
	__xen_evtchn_do_upcall();
J
Jeremy Fitzhardinge 已提交
1106
}
1107
EXPORT_SYMBOL_GPL(xen_hvm_evtchn_do_upcall);
J
Jeremy Fitzhardinge 已提交
1108

J
Jeremy Fitzhardinge 已提交
1109 1110 1111
/* Rebind a new event channel to an existing irq. */
void rebind_evtchn_irq(int evtchn, int irq)
{
J
Jeremy Fitzhardinge 已提交
1112 1113
	struct irq_info *info = info_for_irq(irq);

J
Jeremy Fitzhardinge 已提交
1114 1115 1116 1117 1118 1119 1120 1121 1122
	/* Make sure the irq is masked, since the new event channel
	   will also be masked. */
	disable_irq(irq);

	spin_lock(&irq_mapping_update_lock);

	/* After resume the irq<->evtchn mappings are all cleared out */
	BUG_ON(evtchn_to_irq[evtchn] != -1);
	/* Expect irq to have been bound before,
J
Jeremy Fitzhardinge 已提交
1123 1124
	   so there should be a proper type */
	BUG_ON(info->type == IRQT_UNBOUND);
J
Jeremy Fitzhardinge 已提交
1125

1126
	xen_irq_info_evtchn_init(irq, evtchn);
J
Jeremy Fitzhardinge 已提交
1127 1128 1129 1130

	spin_unlock(&irq_mapping_update_lock);

	/* new event channels are always bound to cpu 0 */
1131
	irq_set_affinity(irq, cpumask_of(0));
J
Jeremy Fitzhardinge 已提交
1132 1133 1134 1135 1136

	/* Unmask the event channel. */
	enable_irq(irq);
}

J
Jeremy Fitzhardinge 已提交
1137
/* Rebind an evtchn so that it gets delivered to a specific cpu */
1138
static int rebind_irq_to_cpu(unsigned irq, unsigned tcpu)
J
Jeremy Fitzhardinge 已提交
1139 1140 1141 1142
{
	struct evtchn_bind_vcpu bind_vcpu;
	int evtchn = evtchn_from_irq(irq);

1143 1144 1145 1146 1147 1148 1149 1150
	if (!VALID_EVTCHN(evtchn))
		return -1;

	/*
	 * Events delivered via platform PCI interrupts are always
	 * routed to vcpu 0 and hence cannot be rebound.
	 */
	if (xen_hvm_domain() && !xen_have_vector_callback)
1151
		return -1;
J
Jeremy Fitzhardinge 已提交
1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164

	/* Send future instances of this interrupt to other vcpu. */
	bind_vcpu.port = evtchn;
	bind_vcpu.vcpu = tcpu;

	/*
	 * If this fails, it usually just indicates that we're dealing with a
	 * virq or IPI channel, which don't actually need to be rebound. Ignore
	 * it, but don't do the xenlinux-level rebind in that case.
	 */
	if (HYPERVISOR_event_channel_op(EVTCHNOP_bind_vcpu, &bind_vcpu) >= 0)
		bind_evtchn_to_cpu(evtchn, tcpu);

1165 1166
	return 0;
}
J
Jeremy Fitzhardinge 已提交
1167

1168 1169
static int set_affinity_irq(struct irq_data *data, const struct cpumask *dest,
			    bool force)
J
Jeremy Fitzhardinge 已提交
1170
{
1171
	unsigned tcpu = cpumask_first(dest);
1172

1173
	return rebind_irq_to_cpu(data->irq, tcpu);
J
Jeremy Fitzhardinge 已提交
1174 1175
}

1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191
int resend_irq_on_evtchn(unsigned int irq)
{
	int masked, evtchn = evtchn_from_irq(irq);
	struct shared_info *s = HYPERVISOR_shared_info;

	if (!VALID_EVTCHN(evtchn))
		return 1;

	masked = sync_test_and_set_bit(evtchn, s->evtchn_mask);
	sync_set_bit(evtchn, s->evtchn_pending);
	if (!masked)
		unmask_evtchn(evtchn);

	return 1;
}

1192
static void enable_dynirq(struct irq_data *data)
J
Jeremy Fitzhardinge 已提交
1193
{
1194
	int evtchn = evtchn_from_irq(data->irq);
J
Jeremy Fitzhardinge 已提交
1195 1196 1197 1198 1199

	if (VALID_EVTCHN(evtchn))
		unmask_evtchn(evtchn);
}

1200
static void disable_dynirq(struct irq_data *data)
J
Jeremy Fitzhardinge 已提交
1201
{
1202
	int evtchn = evtchn_from_irq(data->irq);
J
Jeremy Fitzhardinge 已提交
1203 1204 1205 1206 1207

	if (VALID_EVTCHN(evtchn))
		mask_evtchn(evtchn);
}

1208
static void ack_dynirq(struct irq_data *data)
J
Jeremy Fitzhardinge 已提交
1209
{
1210
	int evtchn = evtchn_from_irq(data->irq);
J
Jeremy Fitzhardinge 已提交
1211

1212
	move_masked_irq(data->irq);
J
Jeremy Fitzhardinge 已提交
1213 1214

	if (VALID_EVTCHN(evtchn))
1215
		unmask_evtchn(evtchn);
J
Jeremy Fitzhardinge 已提交
1216 1217
}

1218
static int retrigger_dynirq(struct irq_data *data)
J
Jeremy Fitzhardinge 已提交
1219
{
1220
	int evtchn = evtchn_from_irq(data->irq);
1221
	struct shared_info *sh = HYPERVISOR_shared_info;
J
Jeremy Fitzhardinge 已提交
1222 1223 1224
	int ret = 0;

	if (VALID_EVTCHN(evtchn)) {
1225 1226 1227 1228 1229 1230
		int masked;

		masked = sync_test_and_set_bit(evtchn, sh->evtchn_mask);
		sync_set_bit(evtchn, sh->evtchn_pending);
		if (!masked)
			unmask_evtchn(evtchn);
J
Jeremy Fitzhardinge 已提交
1231 1232 1233 1234 1235 1236
		ret = 1;
	}

	return ret;
}

1237
static void restore_pirqs(void)
1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261
{
	int pirq, rc, irq, gsi;
	struct physdev_map_pirq map_irq;

	for (pirq = 0; pirq < nr_irqs; pirq++) {
		irq = pirq_to_irq[pirq];
		if (irq == -1)
			continue;

		/* save/restore of PT devices doesn't work, so at this point the
		 * only devices present are GSI based emulated devices */
		gsi = gsi_from_irq(irq);
		if (!gsi)
			continue;

		map_irq.domid = DOMID_SELF;
		map_irq.type = MAP_PIRQ_TYPE_GSI;
		map_irq.index = gsi;
		map_irq.pirq = pirq;

		rc = HYPERVISOR_physdev_op(PHYSDEVOP_map_pirq, &map_irq);
		if (rc) {
			printk(KERN_WARNING "xen map irq failed gsi=%d irq=%d pirq=%d rc=%d\n",
					gsi, irq, pirq, rc);
1262
			xen_free_irq(irq);
1263 1264 1265 1266 1267 1268
			pirq_to_irq[pirq] = -1;
			continue;
		}

		printk(KERN_DEBUG "xen: --> irq=%d, pirq=%d\n", irq, map_irq.pirq);

1269
		__startup_pirq(irq);
1270 1271 1272
	}
}

1273 1274 1275 1276 1277 1278 1279 1280 1281
static void restore_cpu_virqs(unsigned int cpu)
{
	struct evtchn_bind_virq bind_virq;
	int virq, irq, evtchn;

	for (virq = 0; virq < NR_VIRQS; virq++) {
		if ((irq = per_cpu(virq_to_irq, cpu)[virq]) == -1)
			continue;

1282
		BUG_ON(virq_from_irq(irq) != virq);
1283 1284 1285 1286 1287 1288 1289 1290 1291 1292

		/* Get a new binding from Xen. */
		bind_virq.virq = virq;
		bind_virq.vcpu = cpu;
		if (HYPERVISOR_event_channel_op(EVTCHNOP_bind_virq,
						&bind_virq) != 0)
			BUG();
		evtchn = bind_virq.port;

		/* Record the new mapping. */
1293
		xen_irq_info_virq_init(cpu, irq, evtchn, virq);
1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306
		bind_evtchn_to_cpu(evtchn, cpu);
	}
}

static void restore_cpu_ipis(unsigned int cpu)
{
	struct evtchn_bind_ipi bind_ipi;
	int ipi, irq, evtchn;

	for (ipi = 0; ipi < XEN_NR_IPIS; ipi++) {
		if ((irq = per_cpu(ipi_to_irq, cpu)[ipi]) == -1)
			continue;

1307
		BUG_ON(ipi_from_irq(irq) != ipi);
1308 1309 1310 1311 1312 1313 1314 1315 1316

		/* Get a new binding from Xen. */
		bind_ipi.vcpu = cpu;
		if (HYPERVISOR_event_channel_op(EVTCHNOP_bind_ipi,
						&bind_ipi) != 0)
			BUG();
		evtchn = bind_ipi.port;

		/* Record the new mapping. */
1317
		xen_irq_info_ipi_init(cpu, irq, evtchn, ipi);
1318 1319 1320 1321
		bind_evtchn_to_cpu(evtchn, cpu);
	}
}

1322 1323 1324 1325 1326 1327 1328 1329
/* Clear an irq's pending state, in preparation for polling on it */
void xen_clear_irq_pending(int irq)
{
	int evtchn = evtchn_from_irq(irq);

	if (VALID_EVTCHN(evtchn))
		clear_evtchn(evtchn);
}
1330
EXPORT_SYMBOL(xen_clear_irq_pending);
1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349
void xen_set_irq_pending(int irq)
{
	int evtchn = evtchn_from_irq(irq);

	if (VALID_EVTCHN(evtchn))
		set_evtchn(evtchn);
}

bool xen_test_irq_pending(int irq)
{
	int evtchn = evtchn_from_irq(irq);
	bool ret = false;

	if (VALID_EVTCHN(evtchn))
		ret = test_evtchn(evtchn);

	return ret;
}

1350 1351 1352
/* Poll waiting for an irq to become pending with timeout.  In the usual case,
 * the irq will be disabled so it won't deliver an interrupt. */
void xen_poll_irq_timeout(int irq, u64 timeout)
1353 1354 1355 1356 1357 1358 1359
{
	evtchn_port_t evtchn = evtchn_from_irq(irq);

	if (VALID_EVTCHN(evtchn)) {
		struct sched_poll poll;

		poll.nr_ports = 1;
1360
		poll.timeout = timeout;
1361
		set_xen_guest_handle(poll.ports, &evtchn);
1362 1363 1364 1365 1366

		if (HYPERVISOR_sched_op(SCHEDOP_poll, &poll) != 0)
			BUG();
	}
}
1367 1368 1369 1370 1371 1372 1373
EXPORT_SYMBOL(xen_poll_irq_timeout);
/* Poll waiting for an irq to become pending.  In the usual case, the
 * irq will be disabled so it won't deliver an interrupt. */
void xen_poll_irq(int irq)
{
	xen_poll_irq_timeout(irq, 0 /* no timeout */);
}
1374

1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385
void xen_irq_resume(void)
{
	unsigned int cpu, irq, evtchn;

	init_evtchn_cpu_bindings();

	/* New event-channel space is not 'live' yet. */
	for (evtchn = 0; evtchn < NR_EVENT_CHANNELS; evtchn++)
		mask_evtchn(evtchn);

	/* No IRQ <-> event-channel mappings. */
1386
	for (irq = 0; irq < nr_irqs; irq++)
1387 1388 1389 1390 1391 1392 1393 1394 1395
		irq_info[irq].evtchn = 0; /* zap event-channel binding */

	for (evtchn = 0; evtchn < NR_EVENT_CHANNELS; evtchn++)
		evtchn_to_irq[evtchn] = -1;

	for_each_possible_cpu(cpu) {
		restore_cpu_virqs(cpu);
		restore_cpu_ipis(cpu);
	}
1396

1397
	restore_pirqs();
1398 1399
}

J
Jeremy Fitzhardinge 已提交
1400
static struct irq_chip xen_dynamic_chip __read_mostly = {
1401
	.name			= "xen-dyn",
J
Jeremy Fitzhardinge 已提交
1402

1403 1404 1405
	.irq_disable		= disable_dynirq,
	.irq_mask		= disable_dynirq,
	.irq_unmask		= enable_dynirq,
J
Jeremy Fitzhardinge 已提交
1406

1407 1408 1409
	.irq_eoi		= ack_dynirq,
	.irq_set_affinity	= set_affinity_irq,
	.irq_retrigger		= retrigger_dynirq,
J
Jeremy Fitzhardinge 已提交
1410 1411
};

1412
static struct irq_chip xen_pirq_chip __read_mostly = {
1413
	.name			= "xen-pirq",
1414

1415 1416
	.irq_startup		= startup_pirq,
	.irq_shutdown		= shutdown_pirq,
1417

1418 1419
	.irq_enable		= enable_pirq,
	.irq_unmask		= enable_pirq,
1420

1421 1422
	.irq_disable		= disable_pirq,
	.irq_mask		= disable_pirq,
1423

1424
	.irq_ack		= ack_pirq,
1425

1426
	.irq_set_affinity	= set_affinity_irq,
1427

1428
	.irq_retrigger		= retrigger_dynirq,
1429 1430
};

1431
static struct irq_chip xen_percpu_chip __read_mostly = {
1432
	.name			= "xen-percpu",
1433

1434 1435 1436
	.irq_disable		= disable_dynirq,
	.irq_mask		= disable_dynirq,
	.irq_unmask		= enable_dynirq,
1437

1438
	.irq_ack		= ack_dynirq,
1439 1440
};

1441 1442 1443 1444 1445 1446 1447 1448 1449 1450
int xen_set_callback_via(uint64_t via)
{
	struct xen_hvm_param a;
	a.domid = DOMID_SELF;
	a.index = HVM_PARAM_CALLBACK_IRQ;
	a.value = via;
	return HYPERVISOR_hvm_op(HVMOP_set_param, &a);
}
EXPORT_SYMBOL_GPL(xen_set_callback_via);

1451
#ifdef CONFIG_XEN_PVHVM
1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474
/* Vector callbacks are better than PCI interrupts to receive event
 * channel notifications because we can receive vector callbacks on any
 * vcpu and we don't need PCI support or APIC interactions. */
void xen_callback_vector(void)
{
	int rc;
	uint64_t callback_via;
	if (xen_have_vector_callback) {
		callback_via = HVM_CALLBACK_VECTOR(XEN_HVM_EVTCHN_CALLBACK);
		rc = xen_set_callback_via(callback_via);
		if (rc) {
			printk(KERN_ERR "Request for Xen HVM callback vector"
					" failed.\n");
			xen_have_vector_callback = 0;
			return;
		}
		printk(KERN_INFO "Xen HVM callback vector for event delivery is "
				"enabled\n");
		/* in the restore case the vector has already been allocated */
		if (!test_bit(XEN_HVM_EVTCHN_CALLBACK, used_vectors))
			alloc_intr_gate(XEN_HVM_EVTCHN_CALLBACK, xen_hvm_callback_vector);
	}
}
1475 1476 1477
#else
void xen_callback_vector(void) {}
#endif
1478

J
Jeremy Fitzhardinge 已提交
1479 1480
void __init xen_init_IRQ(void)
{
1481
	int i;
1482

1483 1484
	irq_info = kcalloc(nr_irqs, sizeof(*irq_info), GFP_KERNEL);

1485 1486 1487 1488 1489
	/* We are using nr_irqs as the maximum number of pirq available but
	 * that number is actually chosen by Xen and we don't know exactly
	 * what it is. Be careful choosing high pirq numbers. */
	pirq_to_irq = kcalloc(nr_irqs, sizeof(*pirq_to_irq), GFP_KERNEL);
	for (i = 0; i < nr_irqs; i++)
S
Stefano Stabellini 已提交
1490 1491
		pirq_to_irq[i] = -1;

1492 1493 1494 1495
	evtchn_to_irq = kcalloc(NR_EVENT_CHANNELS, sizeof(*evtchn_to_irq),
				    GFP_KERNEL);
	for (i = 0; i < NR_EVENT_CHANNELS; i++)
		evtchn_to_irq[i] = -1;
J
Jeremy Fitzhardinge 已提交
1496 1497 1498 1499 1500 1501 1502

	init_evtchn_cpu_bindings();

	/* No event channels are 'live' right now. */
	for (i = 0; i < NR_EVENT_CHANNELS; i++)
		mask_evtchn(i);

1503 1504 1505
	if (xen_hvm_domain()) {
		xen_callback_vector();
		native_init_IRQ();
1506 1507 1508
		/* pci_xen_hvm_init must be called after native_init_IRQ so that
		 * __acpi_register_gsi can point at the right function */
		pci_xen_hvm_init();
1509 1510
	} else {
		irq_ctx_init(smp_processor_id());
1511 1512
		if (xen_initial_domain())
			xen_setup_pirqs();
1513
	}
J
Jeremy Fitzhardinge 已提交
1514
}