msi.c 38.6 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6
/*
 * File:	msi.c
 * Purpose:	PCI Message Signaled Interrupt (MSI)
 *
 * Copyright (C) 2003-2004 Intel
 * Copyright (C) Tom Long Nguyen (tom.l.nguyen@intel.com)
7
 * Copyright (C) 2016 Christoph Hellwig.
L
Linus Torvalds 已提交
8 9
 */

10
#include <linux/err.h>
L
Linus Torvalds 已提交
11 12 13
#include <linux/mm.h>
#include <linux/irq.h>
#include <linux/interrupt.h>
14
#include <linux/export.h>
L
Linus Torvalds 已提交
15 16 17
#include <linux/ioport.h>
#include <linux/pci.h>
#include <linux/proc_fs.h>
18
#include <linux/msi.h>
D
Dan Williams 已提交
19
#include <linux/smp.h>
H
Hidetoshi Seto 已提交
20 21
#include <linux/errno.h>
#include <linux/io.h>
22
#include <linux/acpi_iort.h>
23
#include <linux/slab.h>
24
#include <linux/irqdomain.h>
25
#include <linux/of_irq.h>
L
Linus Torvalds 已提交
26 27 28 29

#include "pci.h"

static int pci_msi_enable = 1;
30
int pci_msi_ignore_mask;
L
Linus Torvalds 已提交
31

32 33
#define msix_table_size(flags)	((flags & PCI_MSIX_FLAGS_QSIZE) + 1)

34 35 36 37 38
#ifdef CONFIG_PCI_MSI_IRQ_DOMAIN
static int pci_msi_setup_msi_irqs(struct pci_dev *dev, int nvec, int type)
{
	struct irq_domain *domain;

39
	domain = dev_get_msi_domain(&dev->dev);
40
	if (domain && irq_domain_is_hierarchy(domain))
41
		return msi_domain_alloc_irqs(domain, &dev->dev, nvec);
42 43 44 45 46 47 48 49

	return arch_setup_msi_irqs(dev, nvec, type);
}

static void pci_msi_teardown_msi_irqs(struct pci_dev *dev)
{
	struct irq_domain *domain;

50
	domain = dev_get_msi_domain(&dev->dev);
51
	if (domain && irq_domain_is_hierarchy(domain))
52
		msi_domain_free_irqs(domain, &dev->dev);
53 54 55 56 57 58 59
	else
		arch_teardown_msi_irqs(dev);
}
#else
#define pci_msi_setup_msi_irqs		arch_setup_msi_irqs
#define pci_msi_teardown_msi_irqs	arch_teardown_msi_irqs
#endif
60

61 62
/* Arch hooks */

63 64
int __weak arch_setup_msi_irq(struct pci_dev *dev, struct msi_desc *desc)
{
65
	struct msi_controller *chip = dev->bus->msi;
66 67 68 69 70 71 72 73 74 75 76 77
	int err;

	if (!chip || !chip->setup_irq)
		return -EINVAL;

	err = chip->setup_irq(chip, dev, desc);
	if (err < 0)
		return err;

	irq_set_chip_data(desc->irq, chip);

	return 0;
78 79 80
}

void __weak arch_teardown_msi_irq(unsigned int irq)
81
{
82
	struct msi_controller *chip = irq_get_chip_data(irq);
83 84 85 86 87

	if (!chip || !chip->teardown_irq)
		return;

	chip->teardown_irq(chip, irq);
88 89
}

90
int __weak arch_setup_msi_irqs(struct pci_dev *dev, int nvec, int type)
91
{
92
	struct msi_controller *chip = dev->bus->msi;
93 94 95
	struct msi_desc *entry;
	int ret;

96 97
	if (chip && chip->setup_irqs)
		return chip->setup_irqs(chip, dev, nvec, type);
98 99 100 101 102 103 104
	/*
	 * If an architecture wants to support multiple MSI, it needs to
	 * override arch_setup_msi_irqs()
	 */
	if (type == PCI_CAP_ID_MSI && nvec > 1)
		return 1;

105
	for_each_pci_msi_entry(entry, dev) {
106
		ret = arch_setup_msi_irq(dev, entry);
107
		if (ret < 0)
108
			return ret;
109 110
		if (ret > 0)
			return -ENOSPC;
111 112 113 114
	}

	return 0;
}
115

116 117 118 119
/*
 * We have a default implementation available as a separate non-weak
 * function, as it is used by the Xen x86 PCI code
 */
120
void default_teardown_msi_irqs(struct pci_dev *dev)
121
{
122
	int i;
123 124
	struct msi_desc *entry;

125
	for_each_pci_msi_entry(entry, dev)
126 127 128
		if (entry->irq)
			for (i = 0; i < entry->nvec_used; i++)
				arch_teardown_msi_irq(entry->irq + i);
129 130
}

131 132 133 134
void __weak arch_teardown_msi_irqs(struct pci_dev *dev)
{
	return default_teardown_msi_irqs(dev);
}
135

136
static void default_restore_msi_irq(struct pci_dev *dev, int irq)
137 138 139 140 141
{
	struct msi_desc *entry;

	entry = NULL;
	if (dev->msix_enabled) {
142
		for_each_pci_msi_entry(entry, dev) {
143 144 145 146 147 148 149 150
			if (irq == entry->irq)
				break;
		}
	} else if (dev->msi_enabled)  {
		entry = irq_get_msi_desc(irq);
	}

	if (entry)
151
		__pci_write_msi_msg(entry, &entry->msg);
152
}
153

154
void __weak arch_restore_msi_irqs(struct pci_dev *dev)
155
{
156
	return default_restore_msi_irqs(dev);
157
}
158

159 160
static inline __attribute_const__ u32 msi_mask(unsigned x)
{
161 162 163 164
	/* Don't shift by >= width of type */
	if (x >= 5)
		return 0xffffffff;
	return (1 << (1 << x)) - 1;
165 166
}

167 168 169 170 171 172
/*
 * PCI 2.3 does not specify mask bits for each MSI interrupt.  Attempting to
 * mask all MSI interrupts by clearing the MSI enable bit does not work
 * reliably as devices without an INTx disable bit will then generate a
 * level IRQ which will never be cleared.
 */
173
u32 __pci_msi_desc_mask_irq(struct msi_desc *desc, u32 mask, u32 flag)
L
Linus Torvalds 已提交
174
{
175
	u32 mask_bits = desc->masked;
L
Linus Torvalds 已提交
176

177
	if (pci_msi_ignore_mask || !desc->msi_attrib.maskbit)
178
		return 0;
179 180 181

	mask_bits &= ~mask;
	mask_bits |= flag;
182 183
	pci_write_config_dword(msi_desc_to_pci_dev(desc), desc->mask_pos,
			       mask_bits);
184 185 186 187 188 189

	return mask_bits;
}

static void msi_mask_irq(struct msi_desc *desc, u32 mask, u32 flag)
{
190
	desc->masked = __pci_msi_desc_mask_irq(desc, mask, flag);
191 192
}

193 194 195 196 197 198
static void __iomem *pci_msix_desc_addr(struct msi_desc *desc)
{
	return desc->mask_base +
		desc->msi_attrib.entry_nr * PCI_MSIX_ENTRY_SIZE;
}

199 200 201 202 203 204 205
/*
 * This internal function does not flush PCI writes to the device.
 * All users must ensure that they read from the device before either
 * assuming that the device state is up to date, or returning out of this
 * file.  This saves a few milliseconds when initialising devices with lots
 * of MSI-X interrupts.
 */
206
u32 __pci_msix_desc_mask_irq(struct msi_desc *desc, u32 flag)
207 208
{
	u32 mask_bits = desc->masked;
209 210 211 212

	if (pci_msi_ignore_mask)
		return 0;

213 214 215
	mask_bits &= ~PCI_MSIX_ENTRY_CTRL_MASKBIT;
	if (flag)
		mask_bits |= PCI_MSIX_ENTRY_CTRL_MASKBIT;
216
	writel(mask_bits, pci_msix_desc_addr(desc) + PCI_MSIX_ENTRY_VECTOR_CTRL);
217 218 219 220 221 222

	return mask_bits;
}

static void msix_mask_irq(struct msi_desc *desc, u32 flag)
{
223
	desc->masked = __pci_msix_desc_mask_irq(desc, flag);
224
}
225

226
static void msi_set_mask_bit(struct irq_data *data, u32 flag)
227
{
228
	struct msi_desc *desc = irq_data_get_msi_desc(data);
229

230 231 232 233
	if (desc->msi_attrib.is_msix) {
		msix_mask_irq(desc, flag);
		readl(desc->mask_base);		/* Flush write to device */
	} else {
234
		unsigned offset = data->irq - desc->irq;
235
		msi_mask_irq(desc, 1 << offset, flag << offset);
L
Linus Torvalds 已提交
236
	}
237 238
}

239 240 241 242 243
/**
 * pci_msi_mask_irq - Generic irq chip callback to mask PCI/MSI interrupts
 * @data:	pointer to irqdata associated to that interrupt
 */
void pci_msi_mask_irq(struct irq_data *data)
244
{
245
	msi_set_mask_bit(data, 1);
246
}
247
EXPORT_SYMBOL_GPL(pci_msi_mask_irq);
248

249 250 251 252 253
/**
 * pci_msi_unmask_irq - Generic irq chip callback to unmask PCI/MSI interrupts
 * @data:	pointer to irqdata associated to that interrupt
 */
void pci_msi_unmask_irq(struct irq_data *data)
254
{
255
	msi_set_mask_bit(data, 0);
L
Linus Torvalds 已提交
256
}
257
EXPORT_SYMBOL_GPL(pci_msi_unmask_irq);
L
Linus Torvalds 已提交
258

259 260 261 262
void default_restore_msi_irqs(struct pci_dev *dev)
{
	struct msi_desc *entry;

263
	for_each_pci_msi_entry(entry, dev)
264 265 266
		default_restore_msi_irq(dev, entry->irq);
}

267
void __pci_read_msi_msg(struct msi_desc *entry, struct msi_msg *msg)
L
Linus Torvalds 已提交
268
{
269 270 271
	struct pci_dev *dev = msi_desc_to_pci_dev(entry);

	BUG_ON(dev->current_state != PCI_D0);
272 273

	if (entry->msi_attrib.is_msix) {
274
		void __iomem *base = pci_msix_desc_addr(entry);
275 276 277 278 279

		msg->address_lo = readl(base + PCI_MSIX_ENTRY_LOWER_ADDR);
		msg->address_hi = readl(base + PCI_MSIX_ENTRY_UPPER_ADDR);
		msg->data = readl(base + PCI_MSIX_ENTRY_DATA);
	} else {
280
		int pos = dev->msi_cap;
281 282
		u16 data;

283 284
		pci_read_config_dword(dev, pos + PCI_MSI_ADDRESS_LO,
				      &msg->address_lo);
285
		if (entry->msi_attrib.is_64) {
286 287
			pci_read_config_dword(dev, pos + PCI_MSI_ADDRESS_HI,
					      &msg->address_hi);
B
Bjorn Helgaas 已提交
288
			pci_read_config_word(dev, pos + PCI_MSI_DATA_64, &data);
289 290
		} else {
			msg->address_hi = 0;
B
Bjorn Helgaas 已提交
291
			pci_read_config_word(dev, pos + PCI_MSI_DATA_32, &data);
292 293 294 295 296
		}
		msg->data = data;
	}
}

297
void __pci_write_msi_msg(struct msi_desc *entry, struct msi_msg *msg)
Y
Yinghai Lu 已提交
298
{
299 300 301
	struct pci_dev *dev = msi_desc_to_pci_dev(entry);

	if (dev->current_state != PCI_D0) {
302 303
		/* Don't touch the hardware now */
	} else if (entry->msi_attrib.is_msix) {
304
		void __iomem *base = pci_msix_desc_addr(entry);
305

306 307 308
		writel(msg->address_lo, base + PCI_MSIX_ENTRY_LOWER_ADDR);
		writel(msg->address_hi, base + PCI_MSIX_ENTRY_UPPER_ADDR);
		writel(msg->data, base + PCI_MSIX_ENTRY_DATA);
309
	} else {
310
		int pos = dev->msi_cap;
311 312
		u16 msgctl;

313
		pci_read_config_word(dev, pos + PCI_MSI_FLAGS, &msgctl);
314 315
		msgctl &= ~PCI_MSI_FLAGS_QSIZE;
		msgctl |= entry->msi_attrib.multiple << 4;
316
		pci_write_config_word(dev, pos + PCI_MSI_FLAGS, msgctl);
317

318 319
		pci_write_config_dword(dev, pos + PCI_MSI_ADDRESS_LO,
				       msg->address_lo);
320
		if (entry->msi_attrib.is_64) {
321 322
			pci_write_config_dword(dev, pos + PCI_MSI_ADDRESS_HI,
					       msg->address_hi);
B
Bjorn Helgaas 已提交
323 324
			pci_write_config_word(dev, pos + PCI_MSI_DATA_64,
					      msg->data);
325
		} else {
B
Bjorn Helgaas 已提交
326 327
			pci_write_config_word(dev, pos + PCI_MSI_DATA_32,
					      msg->data);
328
		}
L
Linus Torvalds 已提交
329
	}
330
	entry->msg = *msg;
L
Linus Torvalds 已提交
331
}
332

333
void pci_write_msi_msg(unsigned int irq, struct msi_msg *msg)
Y
Yinghai Lu 已提交
334
{
335
	struct msi_desc *entry = irq_get_msi_desc(irq);
Y
Yinghai Lu 已提交
336

337
	__pci_write_msi_msg(entry, msg);
Y
Yinghai Lu 已提交
338
}
339
EXPORT_SYMBOL_GPL(pci_write_msi_msg);
Y
Yinghai Lu 已提交
340

341 342
static void free_msi_irqs(struct pci_dev *dev)
{
343
	struct list_head *msi_list = dev_to_msi_list(&dev->dev);
344
	struct msi_desc *entry, *tmp;
345 346
	struct attribute **msi_attrs;
	struct device_attribute *dev_attr;
347
	int i, count = 0;
348

349
	for_each_pci_msi_entry(entry, dev)
350 351 352
		if (entry->irq)
			for (i = 0; i < entry->nvec_used; i++)
				BUG_ON(irq_has_action(entry->irq + i));
353

354
	pci_msi_teardown_msi_irqs(dev);
355

356
	list_for_each_entry_safe(entry, tmp, msi_list, list) {
357
		if (entry->msi_attrib.is_msix) {
358
			if (list_is_last(&entry->list, msi_list))
359 360
				iounmap(entry->mask_base);
		}
361

362
		list_del(&entry->list);
363
		free_msi_entry(entry);
364
	}
365 366 367 368

	if (dev->msi_irq_groups) {
		sysfs_remove_groups(&dev->dev.kobj, dev->msi_irq_groups);
		msi_attrs = dev->msi_irq_groups[0]->attrs;
369
		while (msi_attrs[count]) {
370 371 372 373 374 375 376 377 378 379 380
			dev_attr = container_of(msi_attrs[count],
						struct device_attribute, attr);
			kfree(dev_attr->attr.name);
			kfree(dev_attr);
			++count;
		}
		kfree(msi_attrs);
		kfree(dev->msi_irq_groups[0]);
		kfree(dev->msi_irq_groups);
		dev->msi_irq_groups = NULL;
	}
381
}
S
Satoru Takeuchi 已提交
382

383 384 385 386 387 388
static void pci_intx_for_msi(struct pci_dev *dev, int enable)
{
	if (!(dev->dev_flags & PCI_DEV_FLAGS_MSI_INTX_DISABLE_BUG))
		pci_intx(dev, enable);
}

389
static void __pci_restore_msi_state(struct pci_dev *dev)
390 391
{
	u16 control;
392
	struct msi_desc *entry;
393

394 395 396
	if (!dev->msi_enabled)
		return;

397
	entry = irq_get_msi_desc(dev->irq);
398

399
	pci_intx_for_msi(dev, 0);
400
	pci_msi_set_enable(dev, 0);
401
	arch_restore_msi_irqs(dev);
402

403
	pci_read_config_word(dev, dev->msi_cap + PCI_MSI_FLAGS, &control);
404 405
	msi_mask_irq(entry, msi_mask(entry->msi_attrib.multi_cap),
		     entry->masked);
406
	control &= ~PCI_MSI_FLAGS_QSIZE;
407
	control |= (entry->msi_attrib.multiple << 4) | PCI_MSI_FLAGS_ENABLE;
408
	pci_write_config_word(dev, dev->msi_cap + PCI_MSI_FLAGS, control);
409 410 411
}

static void __pci_restore_msix_state(struct pci_dev *dev)
412 413 414
{
	struct msi_desc *entry;

E
Eric W. Biederman 已提交
415 416
	if (!dev->msix_enabled)
		return;
417
	BUG_ON(list_empty(dev_to_msi_list(&dev->dev)));
E
Eric W. Biederman 已提交
418

419
	/* route the table */
420
	pci_intx_for_msi(dev, 0);
421
	pci_msix_clear_and_set_ctrl(dev, 0,
422
				PCI_MSIX_FLAGS_ENABLE | PCI_MSIX_FLAGS_MASKALL);
423

424
	arch_restore_msi_irqs(dev);
425
	for_each_pci_msi_entry(entry, dev)
426
		msix_mask_irq(entry, entry->masked);
427

428
	pci_msix_clear_and_set_ctrl(dev, PCI_MSIX_FLAGS_MASKALL, 0);
429
}
430 431 432 433 434 435

void pci_restore_msi_state(struct pci_dev *dev)
{
	__pci_restore_msi_state(dev);
	__pci_restore_msix_state(dev);
}
436
EXPORT_SYMBOL_GPL(pci_restore_msi_state);
437

438
static ssize_t msi_mode_show(struct device *dev, struct device_attribute *attr,
439 440
			     char *buf)
{
441 442 443
	struct msi_desc *entry;
	unsigned long irq;
	int retval;
444

445 446 447
	retval = kstrtoul(attr->attr.name, 10, &irq);
	if (retval)
		return retval;
448

449 450 451 452 453
	entry = irq_get_msi_desc(irq);
	if (entry)
		return sprintf(buf, "%s\n",
				entry->msi_attrib.is_msix ? "msix" : "msi");

454
	return -ENODEV;
455 456 457 458
}

static int populate_msi_sysfs(struct pci_dev *pdev)
{
459 460 461 462 463
	struct attribute **msi_attrs;
	struct attribute *msi_attr;
	struct device_attribute *msi_dev_attr;
	struct attribute_group *msi_irq_group;
	const struct attribute_group **msi_irq_groups;
464
	struct msi_desc *entry;
465 466
	int ret = -ENOMEM;
	int num_msi = 0;
467
	int count = 0;
468
	int i;
469

470
	/* Determine how many msi entries we have */
471
	for_each_pci_msi_entry(entry, pdev)
472
		num_msi += entry->nvec_used;
473 474
	if (!num_msi)
		return 0;
475

476 477 478 479
	/* Dynamically create the MSI attributes for the PCI device */
	msi_attrs = kzalloc(sizeof(void *) * (num_msi + 1), GFP_KERNEL);
	if (!msi_attrs)
		return -ENOMEM;
480
	for_each_pci_msi_entry(entry, pdev) {
481 482 483 484 485 486 487 488 489 490 491 492 493 494 495
		for (i = 0; i < entry->nvec_used; i++) {
			msi_dev_attr = kzalloc(sizeof(*msi_dev_attr), GFP_KERNEL);
			if (!msi_dev_attr)
				goto error_attrs;
			msi_attrs[count] = &msi_dev_attr->attr;

			sysfs_attr_init(&msi_dev_attr->attr);
			msi_dev_attr->attr.name = kasprintf(GFP_KERNEL, "%d",
							    entry->irq + i);
			if (!msi_dev_attr->attr.name)
				goto error_attrs;
			msi_dev_attr->attr.mode = S_IRUGO;
			msi_dev_attr->show = msi_mode_show;
			++count;
		}
496 497
	}

498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513
	msi_irq_group = kzalloc(sizeof(*msi_irq_group), GFP_KERNEL);
	if (!msi_irq_group)
		goto error_attrs;
	msi_irq_group->name = "msi_irqs";
	msi_irq_group->attrs = msi_attrs;

	msi_irq_groups = kzalloc(sizeof(void *) * 2, GFP_KERNEL);
	if (!msi_irq_groups)
		goto error_irq_group;
	msi_irq_groups[0] = msi_irq_group;

	ret = sysfs_create_groups(&pdev->dev.kobj, msi_irq_groups);
	if (ret)
		goto error_irq_groups;
	pdev->msi_irq_groups = msi_irq_groups;

514 515
	return 0;

516 517 518 519 520 521 522 523 524 525 526 527 528
error_irq_groups:
	kfree(msi_irq_groups);
error_irq_group:
	kfree(msi_irq_group);
error_attrs:
	count = 0;
	msi_attr = msi_attrs[count];
	while (msi_attr) {
		msi_dev_attr = container_of(msi_attr, struct device_attribute, attr);
		kfree(msi_attr->name);
		kfree(msi_dev_attr);
		++count;
		msi_attr = msi_attrs[count];
529
	}
530
	kfree(msi_attrs);
531 532 533
	return ret;
}

534
static struct msi_desc *
535
msi_setup_entry(struct pci_dev *dev, int nvec, const struct irq_affinity *affd)
536
{
537
	struct cpumask *masks = NULL;
538
	struct msi_desc *entry;
539 540
	u16 control;

541 542
	if (affd) {
		masks = irq_create_affinity_masks(nvec, affd);
543 544 545
		if (!masks)
			pr_err("Unable to allocate affinity masks, ignoring\n");
	}
546 547

	/* MSI Entry Initialization */
548
	entry = alloc_msi_entry(&dev->dev, nvec, masks);
549
	if (!entry)
550
		goto out;
551 552 553 554 555 556 557 558 559

	pci_read_config_word(dev, dev->msi_cap + PCI_MSI_FLAGS, &control);

	entry->msi_attrib.is_msix	= 0;
	entry->msi_attrib.is_64		= !!(control & PCI_MSI_FLAGS_64BIT);
	entry->msi_attrib.entry_nr	= 0;
	entry->msi_attrib.maskbit	= !!(control & PCI_MSI_FLAGS_MASKBIT);
	entry->msi_attrib.default_irq	= dev->irq;	/* Save IOAPIC IRQ */
	entry->msi_attrib.multi_cap	= (control & PCI_MSI_FLAGS_QMASK) >> 1;
560
	entry->msi_attrib.multiple	= ilog2(__roundup_pow_of_two(nvec));
561 562 563 564 565 566 567 568 569 570

	if (control & PCI_MSI_FLAGS_64BIT)
		entry->mask_pos = dev->msi_cap + PCI_MSI_MASK_64;
	else
		entry->mask_pos = dev->msi_cap + PCI_MSI_MASK_32;

	/* Save the initial mask status */
	if (entry->msi_attrib.maskbit)
		pci_read_config_dword(dev, entry->mask_pos, &entry->masked);

571 572
out:
	kfree(masks);
573 574 575
	return entry;
}

576 577 578 579
static int msi_verify_entries(struct pci_dev *dev)
{
	struct msi_desc *entry;

580
	for_each_pci_msi_entry(entry, dev) {
581 582 583 584 585 586 587 588 589
		if (!dev->no_64bit_msi || !entry->msg.address_hi)
			continue;
		dev_err(&dev->dev, "Device has broken 64-bit MSI but arch"
			" tried to assign one above 4G\n");
		return -EIO;
	}
	return 0;
}

L
Linus Torvalds 已提交
590 591 592
/**
 * msi_capability_init - configure device's MSI capability structure
 * @dev: pointer to the pci_dev data structure of MSI device function
593
 * @nvec: number of interrupts to allocate
594
 * @affd: description of automatic irq affinity assignments (may be %NULL)
L
Linus Torvalds 已提交
595
 *
596 597 598 599 600 601
 * Setup the MSI capability structure of the device with the requested
 * number of interrupts.  A return value of zero indicates the successful
 * setup of an entry with the new MSI irq.  A negative return value indicates
 * an error, and a positive return value indicates the number of interrupts
 * which could have been allocated.
 */
602 603
static int msi_capability_init(struct pci_dev *dev, int nvec,
			       const struct irq_affinity *affd)
L
Linus Torvalds 已提交
604 605
{
	struct msi_desc *entry;
606
	int ret;
607
	unsigned mask;
L
Linus Torvalds 已提交
608

609
	pci_msi_set_enable(dev, 0);	/* Disable MSI during set up */
610

611
	entry = msi_setup_entry(dev, nvec, affd);
612 613
	if (!entry)
		return -ENOMEM;
614

615
	/* All MSIs are unmasked by default, Mask them all */
616
	mask = msi_mask(entry->msi_attrib.multi_cap);
617 618
	msi_mask_irq(entry, mask, mask);

619
	list_add_tail(&entry->list, dev_to_msi_list(&dev->dev));
620

L
Linus Torvalds 已提交
621
	/* Configure MSI capability structure */
622
	ret = pci_msi_setup_msi_irqs(dev, nvec, PCI_CAP_ID_MSI);
623
	if (ret) {
624
		msi_mask_irq(entry, mask, ~mask);
625
		free_msi_irqs(dev);
626
		return ret;
627
	}
628

629 630 631 632 633 634 635
	ret = msi_verify_entries(dev);
	if (ret) {
		msi_mask_irq(entry, mask, ~mask);
		free_msi_irqs(dev);
		return ret;
	}

636 637 638 639 640 641 642
	ret = populate_msi_sysfs(dev);
	if (ret) {
		msi_mask_irq(entry, mask, ~mask);
		free_msi_irqs(dev);
		return ret;
	}

L
Linus Torvalds 已提交
643
	/* Set MSI enabled bits	 */
644
	pci_intx_for_msi(dev, 0);
645
	pci_msi_set_enable(dev, 1);
646
	dev->msi_enabled = 1;
L
Linus Torvalds 已提交
647

648
	pcibios_free_irq(dev);
649
	dev->irq = entry->irq;
L
Linus Torvalds 已提交
650 651 652
	return 0;
}

653
static void __iomem *msix_map_region(struct pci_dev *dev, unsigned nr_entries)
654
{
655
	resource_size_t phys_addr;
656
	u32 table_offset;
657
	unsigned long flags;
658 659
	u8 bir;

660 661
	pci_read_config_dword(dev, dev->msix_cap + PCI_MSIX_TABLE,
			      &table_offset);
662
	bir = (u8)(table_offset & PCI_MSIX_TABLE_BIR);
663 664 665 666
	flags = pci_resource_flags(dev, bir);
	if (!flags || (flags & IORESOURCE_UNSET))
		return NULL;

667
	table_offset &= PCI_MSIX_TABLE_OFFSET;
668 669 670 671 672
	phys_addr = pci_resource_start(dev, bir) + table_offset;

	return ioremap_nocache(phys_addr, nr_entries * PCI_MSIX_ENTRY_SIZE);
}

673
static int msix_setup_entries(struct pci_dev *dev, void __iomem *base,
674
			      struct msix_entry *entries, int nvec,
675
			      const struct irq_affinity *affd)
676
{
677
	struct cpumask *curmsk, *masks = NULL;
678
	struct msi_desc *entry;
679
	int ret, i;
680

681 682
	if (affd) {
		masks = irq_create_affinity_masks(nvec, affd);
683 684 685
		if (!masks)
			pr_err("Unable to allocate affinity masks, ignoring\n");
	}
686

687 688
	for (i = 0, curmsk = masks; i < nvec; i++) {
		entry = alloc_msi_entry(&dev->dev, 1, curmsk);
689 690 691 692 693 694
		if (!entry) {
			if (!i)
				iounmap(base);
			else
				free_msi_irqs(dev);
			/* No enough memory. Don't try again */
695 696
			ret = -ENOMEM;
			goto out;
697 698 699 700
		}

		entry->msi_attrib.is_msix	= 1;
		entry->msi_attrib.is_64		= 1;
701 702 703 704
		if (entries)
			entry->msi_attrib.entry_nr = entries[i].entry;
		else
			entry->msi_attrib.entry_nr = i;
705 706 707
		entry->msi_attrib.default_irq	= dev->irq;
		entry->mask_base		= base;

708
		list_add_tail(&entry->list, dev_to_msi_list(&dev->dev));
709 710
		if (masks)
			curmsk++;
711
	}
712 713 714
	ret = 0;
out:
	kfree(masks);
715
	return ret;
716 717
}

718
static void msix_program_entries(struct pci_dev *dev,
719
				 struct msix_entry *entries)
720 721 722 723
{
	struct msi_desc *entry;
	int i = 0;

724
	for_each_pci_msi_entry(entry, dev) {
725 726
		if (entries)
			entries[i++].vector = entry->irq;
727 728
		entry->masked = readl(pci_msix_desc_addr(entry) +
				PCI_MSIX_ENTRY_VECTOR_CTRL);
729 730 731 732
		msix_mask_irq(entry, 1);
	}
}

L
Linus Torvalds 已提交
733 734 735
/**
 * msix_capability_init - configure device's MSI-X capability
 * @dev: pointer to the pci_dev data structure of MSI-X device function
R
Randy Dunlap 已提交
736 737
 * @entries: pointer to an array of struct msix_entry entries
 * @nvec: number of @entries
738
 * @affd: Optional pointer to enable automatic affinity assignement
L
Linus Torvalds 已提交
739
 *
740
 * Setup the MSI-X capability structure of device function with a
741 742
 * single MSI-X irq. A return of zero indicates the successful setup of
 * requested MSI-X entries with allocated irqs or non-zero for otherwise.
L
Linus Torvalds 已提交
743
 **/
744
static int msix_capability_init(struct pci_dev *dev, struct msix_entry *entries,
745
				int nvec, const struct irq_affinity *affd)
L
Linus Torvalds 已提交
746
{
747
	int ret;
748
	u16 control;
L
Linus Torvalds 已提交
749 750
	void __iomem *base;

751
	/* Ensure MSI-X is disabled while it is set up */
752
	pci_msix_clear_and_set_ctrl(dev, PCI_MSIX_FLAGS_ENABLE, 0);
753

754
	pci_read_config_word(dev, dev->msix_cap + PCI_MSIX_FLAGS, &control);
L
Linus Torvalds 已提交
755
	/* Request & Map MSI-X table region */
756
	base = msix_map_region(dev, msix_table_size(control));
757
	if (!base)
L
Linus Torvalds 已提交
758 759
		return -ENOMEM;

760
	ret = msix_setup_entries(dev, base, entries, nvec, affd);
761 762
	if (ret)
		return ret;
763

764
	ret = pci_msi_setup_msi_irqs(dev, nvec, PCI_CAP_ID_MSIX);
765
	if (ret)
766
		goto out_avail;
767

768 769 770 771 772
	/* Check if all MSI entries honor device restrictions */
	ret = msi_verify_entries(dev);
	if (ret)
		goto out_free;

773 774 775 776 777
	/*
	 * Some devices require MSI-X to be enabled before we can touch the
	 * MSI-X registers.  We need to mask all the vectors to prevent
	 * interrupts coming in before they're fully set up.
	 */
778
	pci_msix_clear_and_set_ctrl(dev, 0,
779
				PCI_MSIX_FLAGS_MASKALL | PCI_MSIX_FLAGS_ENABLE);
780

781
	msix_program_entries(dev, entries);
782

783
	ret = populate_msi_sysfs(dev);
784 785
	if (ret)
		goto out_free;
786

787
	/* Set MSI-X enabled bits and unmask the function */
788
	pci_intx_for_msi(dev, 0);
789
	dev->msix_enabled = 1;
790
	pci_msix_clear_and_set_ctrl(dev, PCI_MSIX_FLAGS_MASKALL, 0);
791

792
	pcibios_free_irq(dev);
L
Linus Torvalds 已提交
793
	return 0;
794

795
out_avail:
796 797 798 799 800
	if (ret < 0) {
		/*
		 * If we had some success, report the number of irqs
		 * we succeeded in setting up.
		 */
801
		struct msi_desc *entry;
802 803
		int avail = 0;

804
		for_each_pci_msi_entry(entry, dev) {
805 806 807 808 809 810 811
			if (entry->irq != 0)
				avail++;
		}
		if (avail != 0)
			ret = avail;
	}

812
out_free:
813 814 815
	free_msi_irqs(dev);

	return ret;
L
Linus Torvalds 已提交
816 817
}

818
/**
819
 * pci_msi_supported - check whether MSI may be enabled on a device
820
 * @dev: pointer to the pci_dev data structure of MSI device function
821
 * @nvec: how many MSIs have been requested ?
822
 *
823
 * Look at global flags, the device itself, and its parent buses
824
 * to determine if MSI/-X are supported for the device. If MSI/-X is
825
 * supported return 1, else return 0.
826
 **/
827
static int pci_msi_supported(struct pci_dev *dev, int nvec)
828 829 830
{
	struct pci_bus *bus;

831
	/* MSI must be globally enabled and supported by the device */
832
	if (!pci_msi_enable)
833
		return 0;
834 835

	if (!dev || dev->no_msi || dev->current_state != PCI_D0)
836
		return 0;
837

838 839 840 841 842 843
	/*
	 * You can't ask to have 0 or less MSIs configured.
	 *  a) it's stupid ..
	 *  b) the list manipulation code assumes nvec >= 1.
	 */
	if (nvec < 1)
844
		return 0;
845

H
Hidetoshi Seto 已提交
846 847 848
	/*
	 * Any bridge which does NOT route MSI transactions from its
	 * secondary bus to its primary bus must set NO_MSI flag on
849 850 851 852
	 * the secondary pci_bus.
	 * We expect only arch-specific PCI host bus controller driver
	 * or quirks for specific PCI bridges to be setting NO_MSI.
	 */
853 854
	for (bus = dev->bus; bus; bus = bus->parent)
		if (bus->bus_flags & PCI_BUS_FLAGS_NO_MSI)
855
			return 0;
856

857
	return 1;
858 859
}

860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884
/**
 * pci_msi_vec_count - Return the number of MSI vectors a device can send
 * @dev: device to report about
 *
 * This function returns the number of MSI vectors a device requested via
 * Multiple Message Capable register. It returns a negative errno if the
 * device is not capable sending MSI interrupts. Otherwise, the call succeeds
 * and returns a power of two, up to a maximum of 2^5 (32), according to the
 * MSI specification.
 **/
int pci_msi_vec_count(struct pci_dev *dev)
{
	int ret;
	u16 msgctl;

	if (!dev->msi_cap)
		return -EINVAL;

	pci_read_config_word(dev, dev->msi_cap + PCI_MSI_FLAGS, &msgctl);
	ret = 1 << ((msgctl & PCI_MSI_FLAGS_QMASK) >> 1);

	return ret;
}
EXPORT_SYMBOL(pci_msi_vec_count);

885
void pci_msi_shutdown(struct pci_dev *dev)
L
Linus Torvalds 已提交
886
{
887 888
	struct msi_desc *desc;
	u32 mask;
L
Linus Torvalds 已提交
889

890
	if (!pci_msi_enable || !dev || !dev->msi_enabled)
E
Eric W. Biederman 已提交
891 892
		return;

893
	BUG_ON(list_empty(dev_to_msi_list(&dev->dev)));
894
	desc = first_pci_msi_entry(dev);
895

896
	pci_msi_set_enable(dev, 0);
897
	pci_intx_for_msi(dev, 1);
898
	dev->msi_enabled = 0;
899

900
	/* Return the device with MSI unmasked as initial states */
901
	mask = msi_mask(desc->msi_attrib.multi_cap);
902
	/* Keep cached state to be restored */
903
	__pci_msi_desc_mask_irq(desc, mask, ~mask);
904 905

	/* Restore dev->irq to its default pin-assertion irq */
906
	dev->irq = desc->msi_attrib.default_irq;
907
	pcibios_alloc_irq(dev);
908
}
909

H
Hidetoshi Seto 已提交
910
void pci_disable_msi(struct pci_dev *dev)
911 912 913 914 915
{
	if (!pci_msi_enable || !dev || !dev->msi_enabled)
		return;

	pci_msi_shutdown(dev);
916
	free_msi_irqs(dev);
L
Linus Torvalds 已提交
917
}
918
EXPORT_SYMBOL(pci_disable_msi);
L
Linus Torvalds 已提交
919

920
/**
921
 * pci_msix_vec_count - return the number of device's MSI-X table entries
922
 * @dev: pointer to the pci_dev data structure of MSI-X device function
923 924 925 926 927 928
 * This function returns the number of device's MSI-X table entries and
 * therefore the number of MSI-X vectors device is capable of sending.
 * It returns a negative errno if the device is not capable of sending MSI-X
 * interrupts.
 **/
int pci_msix_vec_count(struct pci_dev *dev)
929 930 931
{
	u16 control;

932
	if (!dev->msix_cap)
933
		return -EINVAL;
934

935
	pci_read_config_word(dev, dev->msix_cap + PCI_MSIX_FLAGS, &control);
936
	return msix_table_size(control);
937
}
938
EXPORT_SYMBOL(pci_msix_vec_count);
939

940
static int __pci_enable_msix(struct pci_dev *dev, struct msix_entry *entries,
941
			     int nvec, const struct irq_affinity *affd)
L
Linus Torvalds 已提交
942
{
943
	int nr_entries;
E
Eric W. Biederman 已提交
944
	int i, j;
L
Linus Torvalds 已提交
945

946 947
	if (!pci_msi_supported(dev, nvec))
		return -EINVAL;
948

949 950 951
	nr_entries = pci_msix_vec_count(dev);
	if (nr_entries < 0)
		return nr_entries;
L
Linus Torvalds 已提交
952
	if (nvec > nr_entries)
953
		return nr_entries;
L
Linus Torvalds 已提交
954

955 956 957 958 959 960 961 962 963
	if (entries) {
		/* Check for any invalid entries */
		for (i = 0; i < nvec; i++) {
			if (entries[i].entry >= nr_entries)
				return -EINVAL;		/* invalid entry */
			for (j = i + 1; j < nvec; j++) {
				if (entries[i].entry == entries[j].entry)
					return -EINVAL;	/* duplicate entry */
			}
L
Linus Torvalds 已提交
964 965
		}
	}
E
Eric W. Biederman 已提交
966
	WARN_ON(!!dev->msix_enabled);
967

968
	/* Check whether driver already requested for MSI irq */
H
Hidetoshi Seto 已提交
969
	if (dev->msi_enabled) {
970
		dev_info(&dev->dev, "can't enable MSI-X (MSI IRQ already assigned)\n");
L
Linus Torvalds 已提交
971 972
		return -EINVAL;
	}
973
	return msix_capability_init(dev, entries, nvec, affd);
974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992
}

/**
 * pci_enable_msix - configure device's MSI-X capability structure
 * @dev: pointer to the pci_dev data structure of MSI-X device function
 * @entries: pointer to an array of MSI-X entries (optional)
 * @nvec: number of MSI-X irqs requested for allocation by device driver
 *
 * Setup the MSI-X capability structure of device function with the number
 * of requested irqs upon its software driver call to request for
 * MSI-X mode enabled on its hardware device function. A return of zero
 * indicates the successful configuration of MSI-X capability structure
 * with new allocated MSI-X irqs. A return of < 0 indicates a failure.
 * Or a return of > 0 indicates that driver request is exceeding the number
 * of irqs or MSI-X vectors available. Driver should use the returned value to
 * re-send its request.
 **/
int pci_enable_msix(struct pci_dev *dev, struct msix_entry *entries, int nvec)
{
993
	return __pci_enable_msix(dev, entries, nvec, NULL);
L
Linus Torvalds 已提交
994
}
995
EXPORT_SYMBOL(pci_enable_msix);
L
Linus Torvalds 已提交
996

H
Hidetoshi Seto 已提交
997
void pci_msix_shutdown(struct pci_dev *dev)
998
{
999 1000
	struct msi_desc *entry;

1001
	if (!pci_msi_enable || !dev || !dev->msix_enabled)
E
Eric W. Biederman 已提交
1002 1003
		return;

1004
	/* Return the device with MSI-X masked as initial states */
1005
	for_each_pci_msi_entry(entry, dev) {
1006
		/* Keep cached states to be restored */
1007
		__pci_msix_desc_mask_irq(entry, 1);
1008 1009
	}

1010
	pci_msix_clear_and_set_ctrl(dev, PCI_MSIX_FLAGS_ENABLE, 0);
1011
	pci_intx_for_msi(dev, 1);
1012
	dev->msix_enabled = 0;
1013
	pcibios_alloc_irq(dev);
1014
}
1015

H
Hidetoshi Seto 已提交
1016
void pci_disable_msix(struct pci_dev *dev)
1017 1018 1019 1020 1021
{
	if (!pci_msi_enable || !dev || !dev->msix_enabled)
		return;

	pci_msix_shutdown(dev);
1022
	free_msi_irqs(dev);
L
Linus Torvalds 已提交
1023
}
1024
EXPORT_SYMBOL(pci_disable_msix);
L
Linus Torvalds 已提交
1025

1026 1027 1028 1029
void pci_no_msi(void)
{
	pci_msi_enable = 0;
}
1030

1031 1032 1033 1034 1035 1036 1037
/**
 * pci_msi_enabled - is MSI enabled?
 *
 * Returns true if MSI has not been disabled by the command-line option
 * pci=nomsi.
 **/
int pci_msi_enabled(void)
1038
{
1039
	return pci_msi_enable;
1040
}
1041
EXPORT_SYMBOL(pci_msi_enabled);
1042

1043
static int __pci_enable_msi_range(struct pci_dev *dev, int minvec, int maxvec,
1044
				  const struct irq_affinity *affd)
1045
{
1046
	int nvec;
1047 1048
	int rc;

1049 1050
	if (!pci_msi_supported(dev, minvec))
		return -EINVAL;
1051 1052 1053 1054 1055 1056 1057 1058 1059 1060

	WARN_ON(!!dev->msi_enabled);

	/* Check whether driver already requested MSI-X irqs */
	if (dev->msix_enabled) {
		dev_info(&dev->dev,
			 "can't enable MSI (MSI-X already enabled)\n");
		return -EINVAL;
	}

1061 1062 1063
	if (maxvec < minvec)
		return -ERANGE;

1064 1065 1066
	nvec = pci_msi_vec_count(dev);
	if (nvec < 0)
		return nvec;
1067
	if (nvec < minvec)
1068
		return -ENOSPC;
1069 1070

	if (nvec > maxvec)
1071 1072
		nvec = maxvec;

1073
	for (;;) {
1074 1075
		if (affd) {
			nvec = irq_calc_affinity_vectors(nvec, affd);
1076 1077 1078 1079
			if (nvec < minvec)
				return -ENOSPC;
		}

1080
		rc = msi_capability_init(dev, nvec, affd);
1081 1082 1083 1084
		if (rc == 0)
			return nvec;

		if (rc < 0)
1085
			return rc;
1086 1087 1088 1089 1090 1091 1092
		if (rc < minvec)
			return -ENOSPC;

		nvec = rc;
	}
}

1093 1094
/* deprecated, don't use */
int pci_enable_msi(struct pci_dev *dev)
1095
{
1096 1097 1098 1099
	int rc = __pci_enable_msi_range(dev, 1, 1, NULL);
	if (rc < 0)
		return rc;
	return 0;
1100
}
1101
EXPORT_SYMBOL(pci_enable_msi);
1102 1103

static int __pci_enable_msix_range(struct pci_dev *dev,
1104 1105
				   struct msix_entry *entries, int minvec,
				   int maxvec, const struct irq_affinity *affd)
1106
{
1107
	int rc, nvec = maxvec;
1108 1109 1110 1111 1112

	if (maxvec < minvec)
		return -ERANGE;

	for (;;) {
1113 1114
		if (affd) {
			nvec = irq_calc_affinity_vectors(nvec, affd);
1115
			if (nvec < minvec)
1116 1117 1118
				return -ENOSPC;
		}

1119
		rc = __pci_enable_msix(dev, entries, nvec, affd);
1120 1121 1122 1123 1124 1125 1126 1127 1128 1129
		if (rc == 0)
			return nvec;

		if (rc < 0)
			return rc;
		if (rc < minvec)
			return -ENOSPC;

		nvec = rc;
	}
1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147
}

/**
 * pci_enable_msix_range - configure device's MSI-X capability structure
 * @dev: pointer to the pci_dev data structure of MSI-X device function
 * @entries: pointer to an array of MSI-X entries
 * @minvec: minimum number of MSI-X irqs requested
 * @maxvec: maximum number of MSI-X irqs requested
 *
 * Setup the MSI-X capability structure of device function with a maximum
 * possible number of interrupts in the range between @minvec and @maxvec
 * upon its software driver call to request for MSI-X mode enabled on its
 * hardware device function. It returns a negative errno if an error occurs.
 * If it succeeds, it returns the actual number of interrupts allocated and
 * indicates the successful configuration of MSI-X capability structure
 * with new allocated MSI-X interrupts.
 **/
int pci_enable_msix_range(struct pci_dev *dev, struct msix_entry *entries,
1148
		int minvec, int maxvec)
1149
{
1150
	return __pci_enable_msix_range(dev, entries, minvec, maxvec, NULL);
1151 1152
}
EXPORT_SYMBOL(pci_enable_msix_range);
1153

1154
/**
1155
 * pci_alloc_irq_vectors_affinity - allocate multiple IRQs for a device
1156 1157 1158 1159
 * @dev:		PCI device to operate on
 * @min_vecs:		minimum number of vectors required (must be >= 1)
 * @max_vecs:		maximum (desired) number of vectors
 * @flags:		flags or quirks for the allocation
1160
 * @affd:		optional description of the affinity requirements
1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171
 *
 * Allocate up to @max_vecs interrupt vectors for @dev, using MSI-X or MSI
 * vectors if available, and fall back to a single legacy vector
 * if neither is available.  Return the number of vectors allocated,
 * (which might be smaller than @max_vecs) if successful, or a negative
 * error code on error. If less than @min_vecs interrupt vectors are
 * available for @dev the function will fail with -ENOSPC.
 *
 * To get the Linux IRQ number used for a vector that can be passed to
 * request_irq() use the pci_irq_vector() helper.
 */
1172 1173 1174
int pci_alloc_irq_vectors_affinity(struct pci_dev *dev, unsigned int min_vecs,
				   unsigned int max_vecs, unsigned int flags,
				   const struct irq_affinity *affd)
1175
{
1176
	static const struct irq_affinity msi_default_affd;
1177 1178
	int vecs = -ENOSPC;

1179 1180 1181
	if (flags & PCI_IRQ_AFFINITY) {
		if (!affd)
			affd = &msi_default_affd;
1182 1183 1184 1185 1186 1187 1188 1189 1190 1191

		if (affd->pre_vectors + affd->post_vectors > min_vecs)
			return -EINVAL;

		/*
		 * If there aren't any vectors left after applying the pre/post
		 * vectors don't bother with assigning affinity.
		 */
		if (affd->pre_vectors + affd->post_vectors == min_vecs)
			affd = NULL;
1192 1193 1194 1195
	} else {
		if (WARN_ON(affd))
			affd = NULL;
	}
1196

1197
	if (flags & PCI_IRQ_MSIX) {
1198
		vecs = __pci_enable_msix_range(dev, NULL, min_vecs, max_vecs,
1199
				affd);
1200 1201 1202 1203
		if (vecs > 0)
			return vecs;
	}

1204
	if (flags & PCI_IRQ_MSI) {
1205
		vecs = __pci_enable_msi_range(dev, min_vecs, max_vecs, affd);
1206 1207 1208 1209 1210
		if (vecs > 0)
			return vecs;
	}

	/* use legacy irq if allowed */
1211 1212 1213 1214 1215
	if (flags & PCI_IRQ_LEGACY) {
		if (min_vecs == 1 && dev->irq) {
			pci_intx(dev, 1);
			return 1;
		}
1216 1217
	}

1218 1219
	return vecs;
}
1220
EXPORT_SYMBOL(pci_alloc_irq_vectors_affinity);
1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268

/**
 * pci_free_irq_vectors - free previously allocated IRQs for a device
 * @dev:		PCI device to operate on
 *
 * Undoes the allocations and enabling in pci_alloc_irq_vectors().
 */
void pci_free_irq_vectors(struct pci_dev *dev)
{
	pci_disable_msix(dev);
	pci_disable_msi(dev);
}
EXPORT_SYMBOL(pci_free_irq_vectors);

/**
 * pci_irq_vector - return Linux IRQ number of a device vector
 * @dev: PCI device to operate on
 * @nr: device-relative interrupt vector index (0-based).
 */
int pci_irq_vector(struct pci_dev *dev, unsigned int nr)
{
	if (dev->msix_enabled) {
		struct msi_desc *entry;
		int i = 0;

		for_each_pci_msi_entry(entry, dev) {
			if (i == nr)
				return entry->irq;
			i++;
		}
		WARN_ON_ONCE(1);
		return -EINVAL;
	}

	if (dev->msi_enabled) {
		struct msi_desc *entry = first_pci_msi_entry(dev);

		if (WARN_ON_ONCE(nr >= entry->nvec_used))
			return -EINVAL;
	} else {
		if (WARN_ON_ONCE(nr > 0))
			return -EINVAL;
	}

	return dev->irq + nr;
}
EXPORT_SYMBOL(pci_irq_vector);

1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289
/**
 * pci_irq_get_affinity - return the affinity of a particular msi vector
 * @dev:	PCI device to operate on
 * @nr:		device-relative interrupt vector index (0-based).
 */
const struct cpumask *pci_irq_get_affinity(struct pci_dev *dev, int nr)
{
	if (dev->msix_enabled) {
		struct msi_desc *entry;
		int i = 0;

		for_each_pci_msi_entry(entry, dev) {
			if (i == nr)
				return entry->affinity;
			i++;
		}
		WARN_ON_ONCE(1);
		return NULL;
	} else if (dev->msi_enabled) {
		struct msi_desc *entry = first_pci_msi_entry(dev);

1290 1291
		if (WARN_ON_ONCE(!entry || !entry->affinity ||
				 nr >= entry->nvec_used))
1292 1293 1294 1295 1296 1297 1298 1299 1300
			return NULL;

		return &entry->affinity[nr];
	} else {
		return cpu_possible_mask;
	}
}
EXPORT_SYMBOL(pci_irq_get_affinity);

1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316
/**
 * pci_irq_get_node - return the numa node of a particular msi vector
 * @pdev:	PCI device to operate on
 * @vec:	device-relative interrupt vector index (0-based).
 */
int pci_irq_get_node(struct pci_dev *pdev, int vec)
{
	const struct cpumask *mask;

	mask = pci_irq_get_affinity(pdev, vec);
	if (mask)
		return local_memory_node(cpu_to_node(cpumask_first(mask)));
	return dev_to_node(&pdev->dev);
}
EXPORT_SYMBOL(pci_irq_get_node);

1317 1318 1319 1320
struct pci_dev *msi_desc_to_pci_dev(struct msi_desc *desc)
{
	return to_pci_dev(desc->dev);
}
1321
EXPORT_SYMBOL(msi_desc_to_pci_dev);
1322

1323 1324 1325 1326 1327 1328 1329 1330
void *msi_desc_to_pci_sysdata(struct msi_desc *desc)
{
	struct pci_dev *dev = msi_desc_to_pci_dev(desc);

	return dev->bus->sysdata;
}
EXPORT_SYMBOL_GPL(msi_desc_to_pci_sysdata);

1331 1332 1333 1334 1335 1336 1337 1338
#ifdef CONFIG_PCI_MSI_IRQ_DOMAIN
/**
 * pci_msi_domain_write_msg - Helper to write MSI message to PCI config space
 * @irq_data:	Pointer to interrupt data of the MSI interrupt
 * @msg:	Pointer to the message
 */
void pci_msi_domain_write_msg(struct irq_data *irq_data, struct msi_msg *msg)
{
1339
	struct msi_desc *desc = irq_data_get_msi_desc(irq_data);
1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384

	/*
	 * For MSI-X desc->irq is always equal to irq_data->irq. For
	 * MSI only the first interrupt of MULTI MSI passes the test.
	 */
	if (desc->irq == irq_data->irq)
		__pci_write_msi_msg(desc, msg);
}

/**
 * pci_msi_domain_calc_hwirq - Generate a unique ID for an MSI source
 * @dev:	Pointer to the PCI device
 * @desc:	Pointer to the msi descriptor
 *
 * The ID number is only used within the irqdomain.
 */
irq_hw_number_t pci_msi_domain_calc_hwirq(struct pci_dev *dev,
					  struct msi_desc *desc)
{
	return (irq_hw_number_t)desc->msi_attrib.entry_nr |
		PCI_DEVID(dev->bus->number, dev->devfn) << 11 |
		(pci_domain_nr(dev->bus) & 0xFFFFFFFF) << 27;
}

static inline bool pci_msi_desc_is_multi_msi(struct msi_desc *desc)
{
	return !desc->msi_attrib.is_msix && desc->nvec_used > 1;
}

/**
 * pci_msi_domain_check_cap - Verify that @domain supports the capabilities for @dev
 * @domain:	The interrupt domain to check
 * @info:	The domain info for verification
 * @dev:	The device to check
 *
 * Returns:
 *  0 if the functionality is supported
 *  1 if Multi MSI is requested, but the domain does not support it
 *  -ENOTSUPP otherwise
 */
int pci_msi_domain_check_cap(struct irq_domain *domain,
			     struct msi_domain_info *info, struct device *dev)
{
	struct msi_desc *desc = first_pci_msi_entry(to_pci_dev(dev));

1385
	/* Special handling to support __pci_enable_msi_range() */
1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397
	if (pci_msi_desc_is_multi_msi(desc) &&
	    !(info->flags & MSI_FLAG_MULTI_PCI_MSI))
		return 1;
	else if (desc->msi_attrib.is_msix && !(info->flags & MSI_FLAG_PCI_MSIX))
		return -ENOTSUPP;

	return 0;
}

static int pci_msi_domain_handle_error(struct irq_domain *domain,
				       struct msi_desc *desc, int error)
{
1398
	/* Special handling to support __pci_enable_msi_range() */
1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445
	if (pci_msi_desc_is_multi_msi(desc) && error == -ENOSPC)
		return 1;

	return error;
}

#ifdef GENERIC_MSI_DOMAIN_OPS
static void pci_msi_domain_set_desc(msi_alloc_info_t *arg,
				    struct msi_desc *desc)
{
	arg->desc = desc;
	arg->hwirq = pci_msi_domain_calc_hwirq(msi_desc_to_pci_dev(desc),
					       desc);
}
#else
#define pci_msi_domain_set_desc		NULL
#endif

static struct msi_domain_ops pci_msi_domain_ops_default = {
	.set_desc	= pci_msi_domain_set_desc,
	.msi_check	= pci_msi_domain_check_cap,
	.handle_error	= pci_msi_domain_handle_error,
};

static void pci_msi_domain_update_dom_ops(struct msi_domain_info *info)
{
	struct msi_domain_ops *ops = info->ops;

	if (ops == NULL) {
		info->ops = &pci_msi_domain_ops_default;
	} else {
		if (ops->set_desc == NULL)
			ops->set_desc = pci_msi_domain_set_desc;
		if (ops->msi_check == NULL)
			ops->msi_check = pci_msi_domain_check_cap;
		if (ops->handle_error == NULL)
			ops->handle_error = pci_msi_domain_handle_error;
	}
}

static void pci_msi_domain_update_chip_ops(struct msi_domain_info *info)
{
	struct irq_chip *chip = info->chip;

	BUG_ON(!chip);
	if (!chip->irq_write_msi_msg)
		chip->irq_write_msi_msg = pci_msi_domain_write_msg;
1446 1447 1448 1449
	if (!chip->irq_mask)
		chip->irq_mask = pci_msi_mask_irq;
	if (!chip->irq_unmask)
		chip->irq_unmask = pci_msi_unmask_irq;
1450 1451 1452
}

/**
1453 1454
 * pci_msi_create_irq_domain - Create a MSI interrupt domain
 * @fwnode:	Optional fwnode of the interrupt controller
1455 1456 1457 1458 1459 1460 1461 1462
 * @info:	MSI domain info
 * @parent:	Parent irq domain
 *
 * Updates the domain and chip ops and creates a MSI interrupt domain.
 *
 * Returns:
 * A domain pointer or NULL in case of failure.
 */
1463
struct irq_domain *pci_msi_create_irq_domain(struct fwnode_handle *fwnode,
1464 1465 1466
					     struct msi_domain_info *info,
					     struct irq_domain *parent)
{
1467 1468
	struct irq_domain *domain;

1469 1470 1471 1472 1473
	if (info->flags & MSI_FLAG_USE_DEF_DOM_OPS)
		pci_msi_domain_update_dom_ops(info);
	if (info->flags & MSI_FLAG_USE_DEF_CHIP_OPS)
		pci_msi_domain_update_chip_ops(info);

1474 1475
	info->flags |= MSI_FLAG_ACTIVATE_EARLY;

1476
	domain = msi_create_irq_domain(fwnode, info, parent);
1477 1478 1479 1480 1481
	if (!domain)
		return NULL;

	domain->bus_token = DOMAIN_BUS_PCI_MSI;
	return domain;
1482
}
1483
EXPORT_SYMBOL_GPL(pci_msi_create_irq_domain);
1484

1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509
static int get_msi_id_cb(struct pci_dev *pdev, u16 alias, void *data)
{
	u32 *pa = data;

	*pa = alias;
	return 0;
}
/**
 * pci_msi_domain_get_msi_rid - Get the MSI requester id (RID)
 * @domain:	The interrupt domain
 * @pdev:	The PCI device.
 *
 * The RID for a device is formed from the alias, with a firmware
 * supplied mapping applied
 *
 * Returns: The RID.
 */
u32 pci_msi_domain_get_msi_rid(struct irq_domain *domain, struct pci_dev *pdev)
{
	struct device_node *of_node;
	u32 rid = 0;

	pci_for_each_dma_alias(pdev, get_msi_id_cb, &rid);

	of_node = irq_domain_get_of_node(domain);
1510 1511
	rid = of_node ? of_msi_map_rid(&pdev->dev, of_node, rid) :
			iort_msi_map_rid(&pdev->dev, rid);
1512 1513 1514

	return rid;
}
1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526

/**
 * pci_msi_get_device_domain - Get the MSI domain for a given PCI device
 * @pdev:	The PCI device
 *
 * Use the firmware data to find a device-specific MSI domain
 * (i.e. not one that is ste as a default).
 *
 * Returns: The coresponding MSI domain or NULL if none has been found.
 */
struct irq_domain *pci_msi_get_device_domain(struct pci_dev *pdev)
{
1527
	struct irq_domain *dom;
1528 1529 1530
	u32 rid = 0;

	pci_for_each_dma_alias(pdev, get_msi_id_cb, &rid);
1531 1532 1533 1534
	dom = of_msi_map_get_device_domain(&pdev->dev, rid);
	if (!dom)
		dom = iort_get_device_domain(&pdev->dev, rid);
	return dom;
1535
}
1536
#endif /* CONFIG_PCI_MSI_IRQ_DOMAIN */