probe.c 75.8 KB
Newer Older
1
// SPDX-License-Identifier: GPL-2.0
L
Linus Torvalds 已提交
2 3 4 5 6 7 8 9
/*
 * probe.c - PCI detection and setup code
 */

#include <linux/kernel.h>
#include <linux/delay.h>
#include <linux/init.h>
#include <linux/pci.h>
10
#include <linux/of_device.h>
11
#include <linux/of_pci.h>
12
#include <linux/pci_hotplug.h>
L
Linus Torvalds 已提交
13 14 15
#include <linux/slab.h>
#include <linux/module.h>
#include <linux/cpumask.h>
S
Shaohua Li 已提交
16
#include <linux/pci-aspm.h>
17
#include <linux/aer.h>
18
#include <linux/acpi.h>
19
#include <linux/irqdomain.h>
20
#include <linux/pm_runtime.h>
21
#include "pci.h"
L
Linus Torvalds 已提交
22 23 24 25

#define CARDBUS_LATENCY_TIMER	176	/* secondary latency timer */
#define CARDBUS_RESERVE_BUSNR	3

26
static struct resource busn_resource = {
Y
Yinghai Lu 已提交
27 28 29 30 31 32
	.name	= "PCI busn",
	.start	= 0,
	.end	= 255,
	.flags	= IORESOURCE_BUS,
};

L
Linus Torvalds 已提交
33 34 35 36
/* Ugh.  Need to stop exporting this to modules. */
LIST_HEAD(pci_root_buses);
EXPORT_SYMBOL(pci_root_buses);

37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
static LIST_HEAD(pci_domain_busn_res_list);

struct pci_domain_busn_res {
	struct list_head list;
	struct resource res;
	int domain_nr;
};

static struct resource *get_pci_domain_busn_res(int domain_nr)
{
	struct pci_domain_busn_res *r;

	list_for_each_entry(r, &pci_domain_busn_res_list, list)
		if (r->domain_nr == domain_nr)
			return &r->res;

	r = kzalloc(sizeof(*r), GFP_KERNEL);
	if (!r)
		return NULL;

	r->domain_nr = domain_nr;
	r->res.start = 0;
	r->res.end = 0xff;
	r->res.flags = IORESOURCE_BUS | IORESOURCE_PCI_FIXED;

	list_add_tail(&r->list, &pci_domain_busn_res_list);

	return &r->res;
}

67 68 69 70
static int find_anything(struct device *dev, void *data)
{
	return 1;
}
L
Linus Torvalds 已提交
71

Z
Zhang, Yanmin 已提交
72
/*
B
Bjorn Helgaas 已提交
73 74
 * Some device drivers need know if PCI is initiated.
 * Basically, we think PCI is not initiated when there
75
 * is no device to be found on the pci_bus_type.
Z
Zhang, Yanmin 已提交
76 77 78
 */
int no_pci_devices(void)
{
79 80
	struct device *dev;
	int no_devices;
Z
Zhang, Yanmin 已提交
81

82 83 84 85 86
	dev = bus_find_device(&pci_bus_type, NULL, NULL, find_anything);
	no_devices = (dev == NULL);
	put_device(dev);
	return no_devices;
}
Z
Zhang, Yanmin 已提交
87 88
EXPORT_SYMBOL(no_pci_devices);

L
Linus Torvalds 已提交
89 90 91
/*
 * PCI Bus Class
 */
92
static void release_pcibus_dev(struct device *dev)
L
Linus Torvalds 已提交
93
{
94
	struct pci_bus *pci_bus = to_pci_bus(dev);
L
Linus Torvalds 已提交
95

96
	put_device(pci_bus->bridge);
97
	pci_bus_remove_resources(pci_bus);
98
	pci_release_bus_of_node(pci_bus);
L
Linus Torvalds 已提交
99 100 101 102 103
	kfree(pci_bus);
}

static struct class pcibus_class = {
	.name		= "pci_bus",
104
	.dev_release	= &release_pcibus_dev,
105
	.dev_groups	= pcibus_groups,
L
Linus Torvalds 已提交
106 107 108 109 110 111 112 113
};

static int __init pcibus_class_init(void)
{
	return class_register(&pcibus_class);
}
postcore_initcall(pcibus_class_init);

114
static u64 pci_size(u64 base, u64 maxbase, u64 mask)
L
Linus Torvalds 已提交
115
{
116
	u64 size = mask & maxbase;	/* Find the significant bits */
L
Linus Torvalds 已提交
117 118 119
	if (!size)
		return 0;

B
Bjorn Helgaas 已提交
120 121 122 123
	/*
	 * Get the lowest of them to find the decode size, and from that
	 * the extent.
	 */
L
Linus Torvalds 已提交
124 125
	size = (size & ~(size-1)) - 1;

B
Bjorn Helgaas 已提交
126 127 128 129
	/*
	 * base == maxbase can be valid only if the BAR has already been
	 * programmed with all 1s.
	 */
L
Linus Torvalds 已提交
130 131 132 133 134 135
	if (base == maxbase && ((base | size) & mask) != mask)
		return 0;

	return size;
}

136
static inline unsigned long decode_bar(struct pci_dev *dev, u32 bar)
137
{
138
	u32 mem_type;
139
	unsigned long flags;
140

141
	if ((bar & PCI_BASE_ADDRESS_SPACE) == PCI_BASE_ADDRESS_SPACE_IO) {
142 143 144
		flags = bar & ~PCI_BASE_ADDRESS_IO_MASK;
		flags |= IORESOURCE_IO;
		return flags;
145
	}
146

147 148 149 150
	flags = bar & ~PCI_BASE_ADDRESS_MEM_MASK;
	flags |= IORESOURCE_MEM;
	if (flags & PCI_BASE_ADDRESS_MEM_PREFETCH)
		flags |= IORESOURCE_PREFETCH;
151

152 153 154 155 156
	mem_type = bar & PCI_BASE_ADDRESS_MEM_TYPE_MASK;
	switch (mem_type) {
	case PCI_BASE_ADDRESS_MEM_TYPE_32:
		break;
	case PCI_BASE_ADDRESS_MEM_TYPE_1M:
157
		/* 1M mem BAR treated as 32-bit BAR */
158 159
		break;
	case PCI_BASE_ADDRESS_MEM_TYPE_64:
160 161
		flags |= IORESOURCE_MEM_64;
		break;
162
	default:
163
		/* mem unknown type treated as 32-bit BAR */
164 165
		break;
	}
166
	return flags;
167 168
}

169 170
#define PCI_COMMAND_DECODE_ENABLE	(PCI_COMMAND_MEMORY | PCI_COMMAND_IO)

Y
Yu Zhao 已提交
171
/**
B
Bjorn Helgaas 已提交
172
 * pci_read_base - Read a PCI BAR
Y
Yu Zhao 已提交
173 174 175 176 177 178
 * @dev: the PCI device
 * @type: type of the BAR
 * @res: resource buffer to be filled in
 * @pos: BAR position in the config space
 *
 * Returns 1 if the BAR is 64-bit, or 0 if 32-bit.
179
 */
Y
Yu Zhao 已提交
180
int __pci_read_base(struct pci_dev *dev, enum pci_bar_type type,
R
Ryan Desfosses 已提交
181
		    struct resource *res, unsigned int pos)
182
{
183
	u32 l = 0, sz = 0, mask;
184
	u64 l64, sz64, mask64;
J
Jacob Pan 已提交
185
	u16 orig_cmd;
186
	struct pci_bus_region region, inverted_region;
187

188
	mask = type ? PCI_ROM_ADDRESS_MASK : ~0;
189

190
	/* No printks while decoding is disabled! */
J
Jacob Pan 已提交
191 192
	if (!dev->mmio_always_on) {
		pci_read_config_word(dev, PCI_COMMAND, &orig_cmd);
193 194 195 196
		if (orig_cmd & PCI_COMMAND_DECODE_ENABLE) {
			pci_write_config_word(dev, PCI_COMMAND,
				orig_cmd & ~PCI_COMMAND_DECODE_ENABLE);
		}
J
Jacob Pan 已提交
197 198
	}

199 200 201
	res->name = pci_name(dev);

	pci_read_config_dword(dev, pos, &l);
202
	pci_write_config_dword(dev, pos, l | mask);
203 204 205 206 207
	pci_read_config_dword(dev, pos, &sz);
	pci_write_config_dword(dev, pos, l);

	/*
	 * All bits set in sz means the device isn't working properly.
208 209 210
	 * If the BAR isn't implemented, all bits must be 0.  If it's a
	 * memory BAR or a ROM, bit 0 must be clear; if it's an io BAR, bit
	 * 1 must be clear.
211
	 */
212 213
	if (sz == 0xffffffff)
		sz = 0;
214 215 216 217 218 219 220 221 222

	/*
	 * I don't know how l can have all bits set.  Copied from old code.
	 * Maybe it fixes a bug on some ancient platform.
	 */
	if (l == 0xffffffff)
		l = 0;

	if (type == pci_bar_unknown) {
223 224 225
		res->flags = decode_bar(dev, l);
		res->flags |= IORESOURCE_SIZEALIGN;
		if (res->flags & IORESOURCE_IO) {
226 227 228
			l64 = l & PCI_BASE_ADDRESS_IO_MASK;
			sz64 = sz & PCI_BASE_ADDRESS_IO_MASK;
			mask64 = PCI_BASE_ADDRESS_IO_MASK & (u32)IO_SPACE_LIMIT;
229
		} else {
230 231 232
			l64 = l & PCI_BASE_ADDRESS_MEM_MASK;
			sz64 = sz & PCI_BASE_ADDRESS_MEM_MASK;
			mask64 = (u32)PCI_BASE_ADDRESS_MEM_MASK;
233 234
		}
	} else {
235 236
		if (l & PCI_ROM_ADDRESS_ENABLE)
			res->flags |= IORESOURCE_ROM_ENABLE;
237 238
		l64 = l & PCI_ROM_ADDRESS_MASK;
		sz64 = sz & PCI_ROM_ADDRESS_MASK;
239
		mask64 = PCI_ROM_ADDRESS_MASK;
240 241
	}

242
	if (res->flags & IORESOURCE_MEM_64) {
243 244 245 246 247 248 249
		pci_read_config_dword(dev, pos + 4, &l);
		pci_write_config_dword(dev, pos + 4, ~0);
		pci_read_config_dword(dev, pos + 4, &sz);
		pci_write_config_dword(dev, pos + 4, l);

		l64 |= ((u64)l << 32);
		sz64 |= ((u64)sz << 32);
250 251
		mask64 |= ((u64)~0 << 32);
	}
252

253 254
	if (!dev->mmio_always_on && (orig_cmd & PCI_COMMAND_DECODE_ENABLE))
		pci_write_config_word(dev, PCI_COMMAND, orig_cmd);
255

256 257
	if (!sz64)
		goto fail;
258

259
	sz64 = pci_size(l64, sz64, mask64);
260
	if (!sz64) {
261
		pci_info(dev, FW_BUG "reg 0x%x: invalid BAR (can't size)\n",
262
			 pos);
263
		goto fail;
264
	}
265 266

	if (res->flags & IORESOURCE_MEM_64) {
Y
Yinghai Lu 已提交
267 268
		if ((sizeof(pci_bus_addr_t) < 8 || sizeof(resource_size_t) < 8)
		    && sz64 > 0x100000000ULL) {
269 270 271
			res->flags |= IORESOURCE_UNSET | IORESOURCE_DISABLED;
			res->start = 0;
			res->end = 0;
272
			pci_err(dev, "reg 0x%x: can't handle BAR larger than 4GB (size %#010llx)\n",
273
				pos, (unsigned long long)sz64);
274
			goto out;
275 276
		}

Y
Yinghai Lu 已提交
277
		if ((sizeof(pci_bus_addr_t) < 8) && l) {
278
			/* Above 32-bit boundary; try to reallocate */
279
			res->flags |= IORESOURCE_UNSET;
280 281
			res->start = 0;
			res->end = sz64;
282
			pci_info(dev, "reg 0x%x: can't handle BAR above 4GB (bus address %#010llx)\n",
283
				 pos, (unsigned long long)l64);
284
			goto out;
285 286 287
		}
	}

288 289 290
	region.start = l64;
	region.end = l64 + sz64;

291 292
	pcibios_bus_to_resource(dev->bus, res, &region);
	pcibios_resource_to_bus(dev->bus, &inverted_region, res);
293 294 295 296 297 298 299 300 301 302 303 304 305 306 307

	/*
	 * If "A" is a BAR value (a bus address), "bus_to_resource(A)" is
	 * the corresponding resource address (the physical address used by
	 * the CPU.  Converting that resource address back to a bus address
	 * should yield the original BAR value:
	 *
	 *     resource_to_bus(bus_to_resource(A)) == A
	 *
	 * If it doesn't, CPU accesses to "bus_to_resource(A)" will not
	 * be claimed by the device.
	 */
	if (inverted_region.start != region.start) {
		res->flags |= IORESOURCE_UNSET;
		res->start = 0;
308
		res->end = region.end - region.start;
309
		pci_info(dev, "reg 0x%x: initial BAR value %#010llx invalid\n",
310
			 pos, (unsigned long long)region.start);
311
	}
312

313 314 315 316 317 318
	goto out;


fail:
	res->flags = 0;
out:
319
	if (res->flags)
320
		pci_printk(KERN_DEBUG, dev, "reg 0x%x: %pR\n", pos, res);
321

322
	return (res->flags & IORESOURCE_MEM_64) ? 1 : 0;
323 324
}

L
Linus Torvalds 已提交
325 326
static void pci_read_bases(struct pci_dev *dev, unsigned int howmany, int rom)
{
327
	unsigned int pos, reg;
328

329 330 331
	if (dev->non_compliant_bars)
		return;

332 333
	for (pos = 0; pos < howmany; pos++) {
		struct resource *res = &dev->resource[pos];
L
Linus Torvalds 已提交
334
		reg = PCI_BASE_ADDRESS_0 + (pos << 2);
335
		pos += __pci_read_base(dev, pci_bar_unknown, res, reg);
L
Linus Torvalds 已提交
336
	}
337

L
Linus Torvalds 已提交
338
	if (rom) {
339
		struct resource *res = &dev->resource[PCI_ROM_RESOURCE];
L
Linus Torvalds 已提交
340
		dev->rom_base_reg = rom;
341
		res->flags = IORESOURCE_MEM | IORESOURCE_PREFETCH |
342
				IORESOURCE_READONLY | IORESOURCE_SIZEALIGN;
343
		__pci_read_base(dev, pci_bar_mem32, res, rom);
L
Linus Torvalds 已提交
344 345 346
	}
}

B
Bill Pemberton 已提交
347
static void pci_read_bridge_io(struct pci_bus *child)
L
Linus Torvalds 已提交
348 349 350
{
	struct pci_dev *dev = child->self;
	u8 io_base_lo, io_limit_lo;
351
	unsigned long io_mask, io_granularity, base, limit;
352
	struct pci_bus_region region;
353 354 355 356 357 358 359 360 361
	struct resource *res;

	io_mask = PCI_IO_RANGE_MASK;
	io_granularity = 0x1000;
	if (dev->io_window_1k) {
		/* Support 1K I/O space granularity */
		io_mask = PCI_IO_1K_RANGE_MASK;
		io_granularity = 0x400;
	}
L
Linus Torvalds 已提交
362 363 364 365

	res = child->resource[0];
	pci_read_config_byte(dev, PCI_IO_BASE, &io_base_lo);
	pci_read_config_byte(dev, PCI_IO_LIMIT, &io_limit_lo);
366 367
	base = (io_base_lo & io_mask) << 8;
	limit = (io_limit_lo & io_mask) << 8;
L
Linus Torvalds 已提交
368 369 370

	if ((io_base_lo & PCI_IO_RANGE_TYPE_MASK) == PCI_IO_RANGE_TYPE_32) {
		u16 io_base_hi, io_limit_hi;
371

L
Linus Torvalds 已提交
372 373
		pci_read_config_word(dev, PCI_IO_BASE_UPPER16, &io_base_hi);
		pci_read_config_word(dev, PCI_IO_LIMIT_UPPER16, &io_limit_hi);
374 375
		base |= ((unsigned long) io_base_hi << 16);
		limit |= ((unsigned long) io_limit_hi << 16);
L
Linus Torvalds 已提交
376 377
	}

378
	if (base <= limit) {
L
Linus Torvalds 已提交
379
		res->flags = (io_base_lo & PCI_IO_RANGE_TYPE_MASK) | IORESOURCE_IO;
380
		region.start = base;
381
		region.end = limit + io_granularity - 1;
382
		pcibios_bus_to_resource(dev->bus, res, &region);
383
		pci_printk(KERN_DEBUG, dev, "  bridge window %pR\n", res);
L
Linus Torvalds 已提交
384
	}
385 386
}

B
Bill Pemberton 已提交
387
static void pci_read_bridge_mmio(struct pci_bus *child)
388 389 390 391
{
	struct pci_dev *dev = child->self;
	u16 mem_base_lo, mem_limit_lo;
	unsigned long base, limit;
392
	struct pci_bus_region region;
393
	struct resource *res;
L
Linus Torvalds 已提交
394 395 396 397

	res = child->resource[1];
	pci_read_config_word(dev, PCI_MEMORY_BASE, &mem_base_lo);
	pci_read_config_word(dev, PCI_MEMORY_LIMIT, &mem_limit_lo);
398 399
	base = ((unsigned long) mem_base_lo & PCI_MEMORY_RANGE_MASK) << 16;
	limit = ((unsigned long) mem_limit_lo & PCI_MEMORY_RANGE_MASK) << 16;
400
	if (base <= limit) {
L
Linus Torvalds 已提交
401
		res->flags = (mem_base_lo & PCI_MEMORY_RANGE_TYPE_MASK) | IORESOURCE_MEM;
402 403
		region.start = base;
		region.end = limit + 0xfffff;
404
		pcibios_bus_to_resource(dev->bus, res, &region);
405
		pci_printk(KERN_DEBUG, dev, "  bridge window %pR\n", res);
L
Linus Torvalds 已提交
406
	}
407 408
}

B
Bill Pemberton 已提交
409
static void pci_read_bridge_mmio_pref(struct pci_bus *child)
410 411 412
{
	struct pci_dev *dev = child->self;
	u16 mem_base_lo, mem_limit_lo;
413
	u64 base64, limit64;
Y
Yinghai Lu 已提交
414
	pci_bus_addr_t base, limit;
415
	struct pci_bus_region region;
416
	struct resource *res;
L
Linus Torvalds 已提交
417 418 419 420

	res = child->resource[2];
	pci_read_config_word(dev, PCI_PREF_MEMORY_BASE, &mem_base_lo);
	pci_read_config_word(dev, PCI_PREF_MEMORY_LIMIT, &mem_limit_lo);
421 422
	base64 = (mem_base_lo & PCI_PREF_RANGE_MASK) << 16;
	limit64 = (mem_limit_lo & PCI_PREF_RANGE_MASK) << 16;
L
Linus Torvalds 已提交
423 424 425

	if ((mem_base_lo & PCI_PREF_RANGE_TYPE_MASK) == PCI_PREF_RANGE_TYPE_64) {
		u32 mem_base_hi, mem_limit_hi;
426

L
Linus Torvalds 已提交
427 428 429 430 431 432 433 434 435
		pci_read_config_dword(dev, PCI_PREF_BASE_UPPER32, &mem_base_hi);
		pci_read_config_dword(dev, PCI_PREF_LIMIT_UPPER32, &mem_limit_hi);

		/*
		 * Some bridges set the base > limit by default, and some
		 * (broken) BIOSes do not initialize them.  If we find
		 * this, just assume they are not being used.
		 */
		if (mem_base_hi <= mem_limit_hi) {
436 437
			base64 |= (u64) mem_base_hi << 32;
			limit64 |= (u64) mem_limit_hi << 32;
L
Linus Torvalds 已提交
438 439
		}
	}
440

Y
Yinghai Lu 已提交
441 442
	base = (pci_bus_addr_t) base64;
	limit = (pci_bus_addr_t) limit64;
443 444

	if (base != base64) {
445
		pci_err(dev, "can't handle bridge window above 4GB (bus address %#010llx)\n",
446 447 448 449
			(unsigned long long) base64);
		return;
	}

450
	if (base <= limit) {
451 452 453 454
		res->flags = (mem_base_lo & PCI_PREF_RANGE_TYPE_MASK) |
					 IORESOURCE_MEM | IORESOURCE_PREFETCH;
		if (res->flags & PCI_PREF_RANGE_TYPE_64)
			res->flags |= IORESOURCE_MEM_64;
455 456
		region.start = base;
		region.end = limit + 0xfffff;
457
		pcibios_bus_to_resource(dev->bus, res, &region);
458
		pci_printk(KERN_DEBUG, dev, "  bridge window %pR\n", res);
L
Linus Torvalds 已提交
459 460 461
	}
}

B
Bill Pemberton 已提交
462
void pci_read_bridge_bases(struct pci_bus *child)
463 464
{
	struct pci_dev *dev = child->self;
465
	struct resource *res;
466 467 468 469 470
	int i;

	if (pci_is_root_bus(child))	/* It's a host bus, nothing to read */
		return;

471
	pci_info(dev, "PCI bridge to %pR%s\n",
472
		 &child->busn_res,
473 474
		 dev->transparent ? " (subtractive decode)" : "");

475 476 477 478
	pci_bus_remove_resources(child);
	for (i = 0; i < PCI_BRIDGE_RESOURCE_NUM; i++)
		child->resource[i] = &dev->resource[PCI_BRIDGE_RESOURCES+i];

479 480 481
	pci_read_bridge_io(child);
	pci_read_bridge_mmio(child);
	pci_read_bridge_mmio_pref(child);
482 483

	if (dev->transparent) {
484
		pci_bus_for_each_resource(child->parent, res, i) {
485
			if (res && res->flags) {
486 487
				pci_bus_add_resource(child, res,
						     PCI_SUBTRACTIVE_DECODE);
488
				pci_printk(KERN_DEBUG, dev,
489
					   "  bridge window %pR (subtractive decode)\n",
490 491
					   res);
			}
492 493
		}
	}
494 495
}

496
static struct pci_bus *pci_alloc_bus(struct pci_bus *parent)
L
Linus Torvalds 已提交
497 498 499
{
	struct pci_bus *b;

500
	b = kzalloc(sizeof(*b), GFP_KERNEL);
501 502 503 504 505 506 507 508 509 510
	if (!b)
		return NULL;

	INIT_LIST_HEAD(&b->node);
	INIT_LIST_HEAD(&b->children);
	INIT_LIST_HEAD(&b->devices);
	INIT_LIST_HEAD(&b->slots);
	INIT_LIST_HEAD(&b->resources);
	b->max_bus_speed = PCI_SPEED_UNKNOWN;
	b->cur_bus_speed = PCI_SPEED_UNKNOWN;
511 512 513 514
#ifdef CONFIG_PCI_DOMAINS_GENERIC
	if (parent)
		b->domain_nr = parent->domain_nr;
#endif
L
Linus Torvalds 已提交
515 516 517
	return b;
}

518
static void devm_pci_release_host_bridge_dev(struct device *dev)
519 520 521 522 523
{
	struct pci_host_bridge *bridge = to_pci_host_bridge(dev);

	if (bridge->release_fn)
		bridge->release_fn(bridge);
524
}
525

526 527 528 529
static void pci_release_host_bridge_dev(struct device *dev)
{
	devm_pci_release_host_bridge_dev(dev);
	pci_free_host_bridge(to_pci_host_bridge(dev));
530 531
}

532
struct pci_host_bridge *pci_alloc_host_bridge(size_t priv)
533 534 535
{
	struct pci_host_bridge *bridge;

536
	bridge = kzalloc(sizeof(*bridge) + priv, GFP_KERNEL);
537 538
	if (!bridge)
		return NULL;
539

540
	INIT_LIST_HEAD(&bridge->windows);
541
	bridge->dev.release = pci_release_host_bridge_dev;
542

543 544
	return bridge;
}
545
EXPORT_SYMBOL(pci_alloc_host_bridge);
546

547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562
struct pci_host_bridge *devm_pci_alloc_host_bridge(struct device *dev,
						   size_t priv)
{
	struct pci_host_bridge *bridge;

	bridge = devm_kzalloc(dev, sizeof(*bridge) + priv, GFP_KERNEL);
	if (!bridge)
		return NULL;

	INIT_LIST_HEAD(&bridge->windows);
	bridge->dev.release = devm_pci_release_host_bridge_dev;

	return bridge;
}
EXPORT_SYMBOL(devm_pci_alloc_host_bridge);

563 564 565 566 567 568 569 570
void pci_free_host_bridge(struct pci_host_bridge *bridge)
{
	pci_free_resource_list(&bridge->windows);

	kfree(bridge);
}
EXPORT_SYMBOL(pci_free_host_bridge);

571
static const unsigned char pcix_bus_speed[] = {
572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589
	PCI_SPEED_UNKNOWN,		/* 0 */
	PCI_SPEED_66MHz_PCIX,		/* 1 */
	PCI_SPEED_100MHz_PCIX,		/* 2 */
	PCI_SPEED_133MHz_PCIX,		/* 3 */
	PCI_SPEED_UNKNOWN,		/* 4 */
	PCI_SPEED_66MHz_PCIX_ECC,	/* 5 */
	PCI_SPEED_100MHz_PCIX_ECC,	/* 6 */
	PCI_SPEED_133MHz_PCIX_ECC,	/* 7 */
	PCI_SPEED_UNKNOWN,		/* 8 */
	PCI_SPEED_66MHz_PCIX_266,	/* 9 */
	PCI_SPEED_100MHz_PCIX_266,	/* A */
	PCI_SPEED_133MHz_PCIX_266,	/* B */
	PCI_SPEED_UNKNOWN,		/* C */
	PCI_SPEED_66MHz_PCIX_533,	/* D */
	PCI_SPEED_100MHz_PCIX_533,	/* E */
	PCI_SPEED_133MHz_PCIX_533	/* F */
};

590
const unsigned char pcie_link_speed[] = {
591 592 593
	PCI_SPEED_UNKNOWN,		/* 0 */
	PCIE_SPEED_2_5GT,		/* 1 */
	PCIE_SPEED_5_0GT,		/* 2 */
594
	PCIE_SPEED_8_0GT,		/* 3 */
595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610
	PCI_SPEED_UNKNOWN,		/* 4 */
	PCI_SPEED_UNKNOWN,		/* 5 */
	PCI_SPEED_UNKNOWN,		/* 6 */
	PCI_SPEED_UNKNOWN,		/* 7 */
	PCI_SPEED_UNKNOWN,		/* 8 */
	PCI_SPEED_UNKNOWN,		/* 9 */
	PCI_SPEED_UNKNOWN,		/* A */
	PCI_SPEED_UNKNOWN,		/* B */
	PCI_SPEED_UNKNOWN,		/* C */
	PCI_SPEED_UNKNOWN,		/* D */
	PCI_SPEED_UNKNOWN,		/* E */
	PCI_SPEED_UNKNOWN		/* F */
};

void pcie_update_link_speed(struct pci_bus *bus, u16 linksta)
{
611
	bus->cur_bus_speed = pcie_link_speed[linksta & PCI_EXP_LNKSTA_CLS];
612 613 614
}
EXPORT_SYMBOL_GPL(pcie_update_link_speed);

615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634
static unsigned char agp_speeds[] = {
	AGP_UNKNOWN,
	AGP_1X,
	AGP_2X,
	AGP_4X,
	AGP_8X
};

static enum pci_bus_speed agp_speed(int agp3, int agpstat)
{
	int index = 0;

	if (agpstat & 4)
		index = 3;
	else if (agpstat & 2)
		index = 2;
	else if (agpstat & 1)
		index = 1;
	else
		goto out;
635

636 637 638 639 640 641 642 643 644 645
	if (agp3) {
		index += 2;
		if (index == 5)
			index = 0;
	}

 out:
	return agp_speeds[index];
}

646 647 648 649 650
static void pci_set_bus_speed(struct pci_bus *bus)
{
	struct pci_dev *bridge = bus->self;
	int pos;

651 652 653 654 655 656 657 658 659 660 661 662 663
	pos = pci_find_capability(bridge, PCI_CAP_ID_AGP);
	if (!pos)
		pos = pci_find_capability(bridge, PCI_CAP_ID_AGP3);
	if (pos) {
		u32 agpstat, agpcmd;

		pci_read_config_dword(bridge, pos + PCI_AGP_STATUS, &agpstat);
		bus->max_bus_speed = agp_speed(agpstat & 8, agpstat & 7);

		pci_read_config_dword(bridge, pos + PCI_AGP_COMMAND, &agpcmd);
		bus->cur_bus_speed = agp_speed(agpstat & 8, agpcmd & 7);
	}

664 665 666 667 668
	pos = pci_find_capability(bridge, PCI_CAP_ID_PCIX);
	if (pos) {
		u16 status;
		enum pci_bus_speed max;

669 670 671 672
		pci_read_config_word(bridge, pos + PCI_X_BRIDGE_SSTATUS,
				     &status);

		if (status & PCI_X_SSTATUS_533MHZ) {
673
			max = PCI_SPEED_133MHz_PCIX_533;
674
		} else if (status & PCI_X_SSTATUS_266MHZ) {
675
			max = PCI_SPEED_133MHz_PCIX_266;
676
		} else if (status & PCI_X_SSTATUS_133MHZ) {
R
Ryan Desfosses 已提交
677
			if ((status & PCI_X_SSTATUS_VERS) == PCI_X_SSTATUS_V2)
678
				max = PCI_SPEED_133MHz_PCIX_ECC;
R
Ryan Desfosses 已提交
679
			else
680 681 682 683 684 685
				max = PCI_SPEED_133MHz_PCIX;
		} else {
			max = PCI_SPEED_66MHz_PCIX;
		}

		bus->max_bus_speed = max;
686 687
		bus->cur_bus_speed = pcix_bus_speed[
			(status & PCI_X_SSTATUS_FREQ) >> 6];
688 689 690 691

		return;
	}

692
	if (pci_is_pcie(bridge)) {
693 694 695
		u32 linkcap;
		u16 linksta;

696
		pcie_capability_read_dword(bridge, PCI_EXP_LNKCAP, &linkcap);
697
		bus->max_bus_speed = pcie_link_speed[linkcap & PCI_EXP_LNKCAP_SLS];
698

699
		pcie_capability_read_word(bridge, PCI_EXP_LNKSTA, &linksta);
700 701 702 703
		pcie_update_link_speed(bus, linksta);
	}
}

704 705
static struct irq_domain *pci_host_bridge_msi_domain(struct pci_bus *bus)
{
706 707
	struct irq_domain *d;

708 709 710 711
	/*
	 * Any firmware interface that can resolve the msi_domain
	 * should be called from here.
	 */
712
	d = pci_host_bridge_of_msi_domain(bus);
713 714
	if (!d)
		d = pci_host_bridge_acpi_msi_domain(bus);
715

716 717 718 719 720 721 722 723 724 725 726 727 728 729
#ifdef CONFIG_PCI_MSI_IRQ_DOMAIN
	/*
	 * If no IRQ domain was found via the OF tree, try looking it up
	 * directly through the fwnode_handle.
	 */
	if (!d) {
		struct fwnode_handle *fwnode = pci_root_bus_fwnode(bus);

		if (fwnode)
			d = irq_find_matching_fwnode(fwnode,
						     DOMAIN_BUS_PCI_MSI);
	}
#endif

730
	return d;
731 732 733 734 735
}

static void pci_set_bus_msi_domain(struct pci_bus *bus)
{
	struct irq_domain *d;
736
	struct pci_bus *b;
737 738

	/*
739 740 741
	 * The bus can be a root bus, a subordinate bus, or a virtual bus
	 * created by an SR-IOV device.  Walk up to the first bridge device
	 * found or derive the domain from the host bridge.
742
	 */
743 744 745 746 747 748 749
	for (b = bus, d = NULL; !d && !pci_is_root_bus(b); b = b->parent) {
		if (b->self)
			d = dev_get_msi_domain(&b->self->dev);
	}

	if (!d)
		d = pci_host_bridge_msi_domain(b);
750 751 752 753

	dev_set_msi_domain(&bus->dev, d);
}

754
static int pci_register_host_bridge(struct pci_host_bridge *bridge)
755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771
{
	struct device *parent = bridge->dev.parent;
	struct resource_entry *window, *n;
	struct pci_bus *bus, *b;
	resource_size_t offset;
	LIST_HEAD(resources);
	struct resource *res;
	char addr[64], *fmt;
	const char *name;
	int err;

	bus = pci_alloc_bus(NULL);
	if (!bus)
		return -ENOMEM;

	bridge->bus = bus;

B
Bjorn Helgaas 已提交
772
	/* Temporarily move resources off the list */
773 774 775 776 777 778 779 780 781 782 783
	list_splice_init(&bridge->windows, &resources);
	bus->sysdata = bridge->sysdata;
	bus->msi = bridge->msi;
	bus->ops = bridge->ops;
	bus->number = bus->busn_res.start = bridge->busnr;
#ifdef CONFIG_PCI_DOMAINS_GENERIC
	bus->domain_nr = pci_bus_find_domain_nr(bus, parent);
#endif

	b = pci_find_bus(pci_domain_nr(bus), bridge->busnr);
	if (b) {
B
Bjorn Helgaas 已提交
784
		/* Ignore it if we already got here via a different bridge */
785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869
		dev_dbg(&b->dev, "bus already known\n");
		err = -EEXIST;
		goto free;
	}

	dev_set_name(&bridge->dev, "pci%04x:%02x", pci_domain_nr(bus),
		     bridge->busnr);

	err = pcibios_root_bridge_prepare(bridge);
	if (err)
		goto free;

	err = device_register(&bridge->dev);
	if (err)
		put_device(&bridge->dev);

	bus->bridge = get_device(&bridge->dev);
	device_enable_async_suspend(bus->bridge);
	pci_set_bus_of_node(bus);
	pci_set_bus_msi_domain(bus);

	if (!parent)
		set_dev_node(bus->bridge, pcibus_to_node(bus));

	bus->dev.class = &pcibus_class;
	bus->dev.parent = bus->bridge;

	dev_set_name(&bus->dev, "%04x:%02x", pci_domain_nr(bus), bus->number);
	name = dev_name(&bus->dev);

	err = device_register(&bus->dev);
	if (err)
		goto unregister;

	pcibios_add_bus(bus);

	/* Create legacy_io and legacy_mem files for this bus */
	pci_create_legacy_files(bus);

	if (parent)
		dev_info(parent, "PCI host bridge to bus %s\n", name);
	else
		pr_info("PCI host bridge to bus %s\n", name);

	/* Add initial resources to the bus */
	resource_list_for_each_entry_safe(window, n, &resources) {
		list_move_tail(&window->node, &bridge->windows);
		offset = window->offset;
		res = window->res;

		if (res->flags & IORESOURCE_BUS)
			pci_bus_insert_busn_res(bus, bus->number, res->end);
		else
			pci_bus_add_resource(bus, res, 0);

		if (offset) {
			if (resource_type(res) == IORESOURCE_IO)
				fmt = " (bus address [%#06llx-%#06llx])";
			else
				fmt = " (bus address [%#010llx-%#010llx])";

			snprintf(addr, sizeof(addr), fmt,
				 (unsigned long long)(res->start - offset),
				 (unsigned long long)(res->end - offset));
		} else
			addr[0] = '\0';

		dev_info(&bus->dev, "root bus resource %pR%s\n", res, addr);
	}

	down_write(&pci_bus_sem);
	list_add_tail(&bus->node, &pci_root_buses);
	up_write(&pci_bus_sem);

	return 0;

unregister:
	put_device(&bridge->dev);
	device_unregister(&bridge->dev);

free:
	kfree(bus);
	return err;
}

870 871
static struct pci_bus *pci_alloc_child_bus(struct pci_bus *parent,
					   struct pci_dev *bridge, int busnr)
L
Linus Torvalds 已提交
872 873 874
{
	struct pci_bus *child;
	int i;
875
	int ret;
L
Linus Torvalds 已提交
876

B
Bjorn Helgaas 已提交
877
	/* Allocate a new bus and inherit stuff from the parent */
878
	child = pci_alloc_bus(parent);
L
Linus Torvalds 已提交
879 880 881 882 883
	if (!child)
		return NULL;

	child->parent = parent;
	child->ops = parent->ops;
884
	child->msi = parent->msi;
L
Linus Torvalds 已提交
885
	child->sysdata = parent->sysdata;
886
	child->bus_flags = parent->bus_flags;
L
Linus Torvalds 已提交
887

B
Bjorn Helgaas 已提交
888 889 890
	/*
	 * Initialize some portions of the bus device, but don't register
	 * it now as the parent is not properly set up yet.
891 892
	 */
	child->dev.class = &pcibus_class;
893
	dev_set_name(&child->dev, "%04x:%02x", pci_domain_nr(child), busnr);
L
Linus Torvalds 已提交
894

B
Bjorn Helgaas 已提交
895
	/* Set up the primary, secondary and subordinate bus numbers */
896 897 898
	child->number = child->busn_res.start = busnr;
	child->primary = parent->busn_res.start;
	child->busn_res.end = 0xff;
L
Linus Torvalds 已提交
899

900 901 902 903
	if (!bridge) {
		child->dev.parent = parent->bridge;
		goto add_dev;
	}
904 905 906

	child->self = bridge;
	child->bridge = get_device(&bridge->dev);
907
	child->dev.parent = child->bridge;
908
	pci_set_bus_of_node(child);
909 910
	pci_set_bus_speed(child);

B
Bjorn Helgaas 已提交
911
	/* Set up default resource pointers and names */
912
	for (i = 0; i < PCI_BRIDGE_RESOURCE_NUM; i++) {
L
Linus Torvalds 已提交
913 914 915 916 917
		child->resource[i] = &bridge->resource[PCI_BRIDGE_RESOURCES+i];
		child->resource[i]->name = child->name;
	}
	bridge->subordinate = child;

918
add_dev:
919
	pci_set_bus_msi_domain(child);
920 921 922
	ret = device_register(&child->dev);
	WARN_ON(ret < 0);

923 924
	pcibios_add_bus(child);

925 926 927 928 929 930
	if (child->ops->add_bus) {
		ret = child->ops->add_bus(child);
		if (WARN_ON(ret < 0))
			dev_err(&child->dev, "failed to add bus: %d\n", ret);
	}

931 932 933
	/* Create legacy_io and legacy_mem files for this bus */
	pci_create_legacy_files(child);

L
Linus Torvalds 已提交
934 935 936
	return child;
}

R
Ryan Desfosses 已提交
937 938
struct pci_bus *pci_add_new_bus(struct pci_bus *parent, struct pci_dev *dev,
				int busnr)
L
Linus Torvalds 已提交
939 940 941 942
{
	struct pci_bus *child;

	child = pci_alloc_child_bus(parent, dev, busnr);
943
	if (child) {
944
		down_write(&pci_bus_sem);
L
Linus Torvalds 已提交
945
		list_add_tail(&child->node, &parent->children);
946
		up_write(&pci_bus_sem);
947
	}
L
Linus Torvalds 已提交
948 949
	return child;
}
950
EXPORT_SYMBOL(pci_add_new_bus);
L
Linus Torvalds 已提交
951

952 953 954 955 956 957 958 959 960 961 962
static void pci_enable_crs(struct pci_dev *pdev)
{
	u16 root_cap = 0;

	/* Enable CRS Software Visibility if supported */
	pcie_capability_read_word(pdev, PCI_EXP_RTCAP, &root_cap);
	if (root_cap & PCI_EXP_RTCAP_CRSVIS)
		pcie_capability_set_word(pdev, PCI_EXP_RTCTL,
					 PCI_EXP_RTCTL_CRSSVE);
}

963 964 965
static unsigned int pci_scan_child_bus_extend(struct pci_bus *bus,
					      unsigned int available_buses);

L
Linus Torvalds 已提交
966
/*
967 968 969 970 971 972 973 974 975 976 977
 * pci_scan_bridge_extend() - Scan buses behind a bridge
 * @bus: Parent bus the bridge is on
 * @dev: Bridge itself
 * @max: Starting subordinate number of buses behind this bridge
 * @available_buses: Total number of buses available for this bridge and
 *		     the devices below. After the minimal bus space has
 *		     been allocated the remaining buses will be
 *		     distributed equally between hotplug-capable bridges.
 * @pass: Either %0 (scan already configured bridges) or %1 (scan bridges
 *        that need to be reconfigured.
 *
L
Linus Torvalds 已提交
978 979 980 981 982 983 984 985 986
 * If it's a bridge, configure it and scan the bus behind it.
 * For CardBus bridges, we don't scan behind as the devices will
 * be handled by the bridge driver itself.
 *
 * We need to process bridges in two passes -- first we scan those
 * already configured by the BIOS and after we are done with all of
 * them, we proceed to assigning numbers to the remaining buses in
 * order to avoid overlaps between old and new bus numbers.
 */
987 988 989
static int pci_scan_bridge_extend(struct pci_bus *bus, struct pci_dev *dev,
				  int max, unsigned int available_buses,
				  int pass)
L
Linus Torvalds 已提交
990 991 992
{
	struct pci_bus *child;
	int is_cardbus = (dev->hdr_type == PCI_HEADER_TYPE_CARDBUS);
993
	u32 buses, i, j = 0;
L
Linus Torvalds 已提交
994
	u16 bctl;
995
	u8 primary, secondary, subordinate;
996
	int broken = 0;
L
Linus Torvalds 已提交
997

998 999 1000 1001 1002 1003
	/*
	 * Make sure the bridge is powered on to be able to access config
	 * space of devices below it.
	 */
	pm_runtime_get_sync(&dev->dev);

L
Linus Torvalds 已提交
1004
	pci_read_config_dword(dev, PCI_PRIMARY_BUS, &buses);
1005 1006 1007
	primary = buses & 0xFF;
	secondary = (buses >> 8) & 0xFF;
	subordinate = (buses >> 16) & 0xFF;
L
Linus Torvalds 已提交
1008

1009
	pci_dbg(dev, "scanning [bus %02x-%02x] behind bridge, pass %d\n",
1010
		secondary, subordinate, pass);
L
Linus Torvalds 已提交
1011

1012
	if (!primary && (primary != bus->number) && secondary && subordinate) {
1013
		pci_warn(dev, "Primary bus is hard wired to 0\n");
1014 1015 1016
		primary = bus->number;
	}

1017 1018
	/* Check if setup is sensible at all */
	if (!pass &&
1019
	    (primary != bus->number || secondary <= bus->number ||
1020
	     secondary > subordinate)) {
1021
		pci_info(dev, "bridge configuration invalid ([bus %02x-%02x]), reconfiguring\n",
1022
			 secondary, subordinate);
1023 1024 1025
		broken = 1;
	}

B
Bjorn Helgaas 已提交
1026 1027 1028 1029
	/*
	 * Disable Master-Abort Mode during probing to avoid reporting of
	 * bus errors in some architectures.
	 */
L
Linus Torvalds 已提交
1030 1031 1032 1033
	pci_read_config_word(dev, PCI_BRIDGE_CONTROL, &bctl);
	pci_write_config_word(dev, PCI_BRIDGE_CONTROL,
			      bctl & ~PCI_BRIDGE_CTL_MASTER_ABORT);

1034 1035
	pci_enable_crs(dev);

1036 1037 1038
	if ((secondary || subordinate) && !pcibios_assign_all_busses() &&
	    !is_cardbus && !broken) {
		unsigned int cmax;
B
Bjorn Helgaas 已提交
1039

L
Linus Torvalds 已提交
1040
		/*
B
Bjorn Helgaas 已提交
1041 1042
		 * Bus already configured by firmware, process it in the
		 * first pass and just note the configuration.
L
Linus Torvalds 已提交
1043 1044
		 */
		if (pass)
1045
			goto out;
L
Linus Torvalds 已提交
1046 1047

		/*
B
Bjorn Helgaas 已提交
1048 1049 1050 1051
		 * The bus might already exist for two reasons: Either we
		 * are rescanning the bus or the bus is reachable through
		 * more than one bridge. The second case can happen with
		 * the i450NX chipset.
L
Linus Torvalds 已提交
1052
		 */
1053
		child = pci_find_bus(pci_domain_nr(bus), secondary);
A
Alex Chiang 已提交
1054
		if (!child) {
1055
			child = pci_add_new_bus(bus, dev, secondary);
A
Alex Chiang 已提交
1056 1057
			if (!child)
				goto out;
1058
			child->primary = primary;
Y
Yinghai Lu 已提交
1059
			pci_bus_insert_busn_res(child, secondary, subordinate);
A
Alex Chiang 已提交
1060
			child->bridge_ctl = bctl;
L
Linus Torvalds 已提交
1061 1062 1063
		}

		cmax = pci_scan_child_bus(child);
1064
		if (cmax > subordinate)
1065
			pci_warn(dev, "bridge has subordinate %02x but max busn %02x\n",
1066
				 subordinate, cmax);
B
Bjorn Helgaas 已提交
1067 1068

		/* Subordinate should equal child->busn_res.end */
1069 1070
		if (subordinate > max)
			max = subordinate;
L
Linus Torvalds 已提交
1071
	} else {
B
Bjorn Helgaas 已提交
1072

L
Linus Torvalds 已提交
1073 1074 1075 1076
		/*
		 * We need to assign a number to this bus which we always
		 * do in the second pass.
		 */
1077
		if (!pass) {
1078
			if (pcibios_assign_all_busses() || broken || is_cardbus)
B
Bjorn Helgaas 已提交
1079 1080 1081 1082 1083 1084 1085 1086 1087

				/*
				 * Temporarily disable forwarding of the
				 * configuration cycles on all bridges in
				 * this bus segment to avoid possible
				 * conflicts in the second pass between two
				 * bridges programmed with overlapping bus
				 * ranges.
				 */
1088 1089
				pci_write_config_dword(dev, PCI_PRIMARY_BUS,
						       buses & ~0xffffff);
1090
			goto out;
1091
		}
L
Linus Torvalds 已提交
1092 1093 1094 1095

		/* Clear errors */
		pci_write_config_word(dev, PCI_STATUS, 0xffff);

B
Bjorn Helgaas 已提交
1096 1097 1098 1099 1100
		/*
		 * Prevent assigning a bus number that already exists.
		 * This can happen when a bridge is hot-plugged, so in this
		 * case we only re-scan this bus.
		 */
1101 1102
		child = pci_find_bus(pci_domain_nr(bus), max+1);
		if (!child) {
1103
			child = pci_add_new_bus(bus, dev, max+1);
1104 1105
			if (!child)
				goto out;
1106 1107
			pci_bus_insert_busn_res(child, max+1,
						bus->busn_res.end);
1108
		}
1109
		max++;
1110 1111 1112
		if (available_buses)
			available_buses--;

L
Linus Torvalds 已提交
1113 1114
		buses = (buses & 0xff000000)
		      | ((unsigned int)(child->primary)     <<  0)
1115 1116
		      | ((unsigned int)(child->busn_res.start)   <<  8)
		      | ((unsigned int)(child->busn_res.end) << 16);
L
Linus Torvalds 已提交
1117 1118 1119 1120 1121 1122 1123 1124 1125

		/*
		 * yenta.c forces a secondary latency timer of 176.
		 * Copy that behaviour here.
		 */
		if (is_cardbus) {
			buses &= ~0xff000000;
			buses |= CARDBUS_LATENCY_TIMER << 24;
		}
1126

B
Bjorn Helgaas 已提交
1127
		/* We need to blast all three values with a single write */
L
Linus Torvalds 已提交
1128 1129 1130
		pci_write_config_dword(dev, PCI_PRIMARY_BUS, buses);

		if (!is_cardbus) {
1131
			child->bridge_ctl = bctl;
1132
			max = pci_scan_child_bus_extend(child, available_buses);
L
Linus Torvalds 已提交
1133
		} else {
B
Bjorn Helgaas 已提交
1134

L
Linus Torvalds 已提交
1135
			/*
B
Bjorn Helgaas 已提交
1136 1137 1138
			 * For CardBus bridges, we leave 4 bus numbers as
			 * cards with a PCI-to-PCI bridge can be inserted
			 * later.
L
Linus Torvalds 已提交
1139
			 */
R
Ryan Desfosses 已提交
1140
			for (i = 0; i < CARDBUS_RESERVE_BUSNR; i++) {
1141
				struct pci_bus *parent = bus;
1142 1143 1144
				if (pci_find_bus(pci_domain_nr(bus),
							max+i+1))
					break;
1145 1146
				while (parent->parent) {
					if ((!pcibios_assign_all_busses()) &&
1147 1148
					    (parent->busn_res.end > max) &&
					    (parent->busn_res.end <= max+i)) {
1149 1150 1151 1152 1153
						j = 1;
					}
					parent = parent->parent;
				}
				if (j) {
B
Bjorn Helgaas 已提交
1154

1155
					/*
B
Bjorn Helgaas 已提交
1156 1157 1158
					 * Often, there are two CardBus
					 * bridges -- try to leave one
					 * valid bus number for each one.
1159 1160 1161 1162 1163
					 */
					i /= 2;
					break;
				}
			}
1164
			max += i;
L
Linus Torvalds 已提交
1165
		}
B
Bjorn Helgaas 已提交
1166 1167

		/* Set subordinate bus number to its real value */
Y
Yinghai Lu 已提交
1168
		pci_bus_update_busn_res_end(child, max);
L
Linus Torvalds 已提交
1169 1170 1171
		pci_write_config_byte(dev, PCI_SUBORDINATE_BUS, max);
	}

1172 1173 1174
	sprintf(child->name,
		(is_cardbus ? "PCI CardBus %04x:%02x" : "PCI Bus %04x:%02x"),
		pci_domain_nr(bus), child->number);
L
Linus Torvalds 已提交
1175

1176
	/* Has only triggered on CardBus, fixup is in yenta_socket */
1177
	while (bus->parent) {
1178 1179
		if ((child->busn_res.end > bus->busn_res.end) ||
		    (child->number > bus->busn_res.end) ||
1180
		    (child->number < bus->number) ||
1181
		    (child->busn_res.end < bus->number)) {
1182
			dev_info(&child->dev, "%pR %s hidden behind%s bridge %s %pR\n",
1183 1184 1185
				&child->busn_res,
				(bus->number > child->busn_res.end &&
				 bus->busn_res.end < child->number) ?
1186 1187
					"wholly" : "partially",
				bus->self->transparent ? " transparent" : "",
1188
				dev_name(&bus->dev),
1189
				&bus->busn_res);
1190 1191 1192 1193
		}
		bus = bus->parent;
	}

1194 1195 1196
out:
	pci_write_config_word(dev, PCI_BRIDGE_CONTROL, bctl);

1197 1198
	pm_runtime_put(&dev->dev);

L
Linus Torvalds 已提交
1199 1200
	return max;
}
1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222

/*
 * pci_scan_bridge() - Scan buses behind a bridge
 * @bus: Parent bus the bridge is on
 * @dev: Bridge itself
 * @max: Starting subordinate number of buses behind this bridge
 * @pass: Either %0 (scan already configured bridges) or %1 (scan bridges
 *        that need to be reconfigured.
 *
 * If it's a bridge, configure it and scan the bus behind it.
 * For CardBus bridges, we don't scan behind as the devices will
 * be handled by the bridge driver itself.
 *
 * We need to process bridges in two passes -- first we scan those
 * already configured by the BIOS and after we are done with all of
 * them, we proceed to assigning numbers to the remaining buses in
 * order to avoid overlaps between old and new bus numbers.
 */
int pci_scan_bridge(struct pci_bus *bus, struct pci_dev *dev, int max, int pass)
{
	return pci_scan_bridge_extend(bus, dev, max, 0, pass);
}
1223
EXPORT_SYMBOL(pci_scan_bridge);
L
Linus Torvalds 已提交
1224 1225 1226 1227 1228 1229 1230 1231 1232

/*
 * Read interrupt line and base address registers.
 * The architecture-dependent code can tweak these, of course.
 */
static void pci_read_irq(struct pci_dev *dev)
{
	unsigned char irq;

1233 1234 1235 1236 1237 1238 1239
	/* VFs are not allowed to use INTx, so skip the config reads */
	if (dev->is_virtfn) {
		dev->pin = 0;
		dev->irq = 0;
		return;
	}

L
Linus Torvalds 已提交
1240
	pci_read_config_byte(dev, PCI_INTERRUPT_PIN, &irq);
1241
	dev->pin = irq;
L
Linus Torvalds 已提交
1242 1243 1244 1245 1246
	if (irq)
		pci_read_config_byte(dev, PCI_INTERRUPT_LINE, &irq);
	dev->irq = irq;
}

1247
void set_pcie_port_type(struct pci_dev *pdev)
Y
Yu Zhao 已提交
1248 1249 1250
{
	int pos;
	u16 reg16;
1251 1252
	int type;
	struct pci_dev *parent;
Y
Yu Zhao 已提交
1253 1254 1255 1256

	pos = pci_find_capability(pdev, PCI_CAP_ID_EXP);
	if (!pos)
		return;
1257

1258
	pdev->pcie_cap = pos;
Y
Yu Zhao 已提交
1259
	pci_read_config_word(pdev, pos + PCI_EXP_FLAGS, &reg16);
1260
	pdev->pcie_flags_reg = reg16;
1261 1262
	pci_read_config_word(pdev, pos + PCI_EXP_DEVCAP, &reg16);
	pdev->pcie_mpss = reg16 & PCI_EXP_DEVCAP_PAYLOAD;
1263 1264

	/*
1265 1266 1267 1268
	 * A Root Port or a PCI-to-PCIe bridge is always the upstream end
	 * of a Link.  No PCIe component has two Links.  Two Links are
	 * connected by a Switch that has a Port on each Link and internal
	 * logic to connect the two Ports.
1269 1270
	 */
	type = pci_pcie_type(pdev);
1271 1272
	if (type == PCI_EXP_TYPE_ROOT_PORT ||
	    type == PCI_EXP_TYPE_PCIE_BRIDGE)
1273 1274 1275 1276
		pdev->has_secondary_link = 1;
	else if (type == PCI_EXP_TYPE_UPSTREAM ||
		 type == PCI_EXP_TYPE_DOWNSTREAM) {
		parent = pci_upstream_bridge(pdev);
1277 1278 1279 1280 1281 1282

		/*
		 * Usually there's an upstream device (Root Port or Switch
		 * Downstream Port), but we can't assume one exists.
		 */
		if (parent && !parent->has_secondary_link)
1283 1284
			pdev->has_secondary_link = 1;
	}
Y
Yu Zhao 已提交
1285 1286
}

1287
void set_pcie_hotplug_bridge(struct pci_dev *pdev)
1288 1289 1290
{
	u32 reg32;

1291
	pcie_capability_read_dword(pdev, PCI_EXP_SLTCAP, &reg32);
1292 1293 1294 1295
	if (reg32 & PCI_EXP_SLTCAP_HPC)
		pdev->is_hotplug_bridge = 1;
}

1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313
static void set_pcie_thunderbolt(struct pci_dev *dev)
{
	int vsec = 0;
	u32 header;

	while ((vsec = pci_find_next_ext_capability(dev, vsec,
						    PCI_EXT_CAP_ID_VNDR))) {
		pci_read_config_dword(dev, vsec + PCI_VNDR_HEADER, &header);

		/* Is the device part of a Thunderbolt controller? */
		if (dev->vendor == PCI_VENDOR_ID_INTEL &&
		    PCI_VNDR_HEADER_ID(header) == PCI_VSEC_ID_INTEL_TBT) {
			dev->is_thunderbolt = 1;
			return;
		}
	}
}

1314
/**
B
Bjorn Helgaas 已提交
1315
 * pci_ext_cfg_is_aliased - Is ext config space just an alias of std config?
1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350
 * @dev: PCI device
 *
 * PCI Express to PCI/PCI-X Bridge Specification, rev 1.0, 4.1.4 says that
 * when forwarding a type1 configuration request the bridge must check that
 * the extended register address field is zero.  The bridge is not permitted
 * to forward the transactions and must handle it as an Unsupported Request.
 * Some bridges do not follow this rule and simply drop the extended register
 * bits, resulting in the standard config space being aliased, every 256
 * bytes across the entire configuration space.  Test for this condition by
 * comparing the first dword of each potential alias to the vendor/device ID.
 * Known offenders:
 *   ASM1083/1085 PCIe-to-PCI Reversible Bridge (1b21:1080, rev 01 & 03)
 *   AMD/ATI SBx00 PCI to PCI Bridge (1002:4384, rev 40)
 */
static bool pci_ext_cfg_is_aliased(struct pci_dev *dev)
{
#ifdef CONFIG_PCI_QUIRKS
	int pos;
	u32 header, tmp;

	pci_read_config_dword(dev, PCI_VENDOR_ID, &header);

	for (pos = PCI_CFG_SPACE_SIZE;
	     pos < PCI_CFG_SPACE_EXP_SIZE; pos += PCI_CFG_SPACE_SIZE) {
		if (pci_read_config_dword(dev, pos, &tmp) != PCIBIOS_SUCCESSFUL
		    || header != tmp)
			return false;
	}

	return true;
#else
	return false;
#endif
}

1351
/**
B
Bjorn Helgaas 已提交
1352
 * pci_cfg_space_size - Get the configuration space size of the PCI device
1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367
 * @dev: PCI device
 *
 * Regular PCI devices have 256 bytes, but PCI-X 2 and PCI Express devices
 * have 4096 bytes.  Even if the device is capable, that doesn't mean we can
 * access it.  Maybe we don't have a way to generate extended config space
 * accesses, or the device is behind a reverse Express bridge.  So we try
 * reading the dword at 0x100 which must either be 0 or a valid extended
 * capability header.
 */
static int pci_cfg_space_size_ext(struct pci_dev *dev)
{
	u32 status;
	int pos = PCI_CFG_SPACE_SIZE;

	if (pci_read_config_dword(dev, pos, &status) != PCIBIOS_SUCCESSFUL)
1368
		return PCI_CFG_SPACE_SIZE;
1369
	if (status == 0xffffffff || pci_ext_cfg_is_aliased(dev))
1370
		return PCI_CFG_SPACE_SIZE;
1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384

	return PCI_CFG_SPACE_EXP_SIZE;
}

int pci_cfg_space_size(struct pci_dev *dev)
{
	int pos;
	u32 status;
	u16 class;

	class = dev->class >> 8;
	if (class == PCI_CLASS_BRIDGE_HOST)
		return pci_cfg_space_size_ext(dev);

1385 1386
	if (pci_is_pcie(dev))
		return pci_cfg_space_size_ext(dev);
1387

1388 1389 1390
	pos = pci_find_capability(dev, PCI_CAP_ID_PCIX);
	if (!pos)
		return PCI_CFG_SPACE_SIZE;
1391

1392 1393 1394
	pci_read_config_dword(dev, pos + PCI_X_STATUS, &status);
	if (status & (PCI_X_STATUS_266MHZ | PCI_X_STATUS_533MHZ))
		return pci_cfg_space_size_ext(dev);
1395 1396 1397 1398

	return PCI_CFG_SPACE_SIZE;
}

1399
#define LEGACY_IO_RESOURCE	(IORESOURCE_IO | IORESOURCE_PCI_FIXED)
1400

1401
static void pci_msi_setup_pci_dev(struct pci_dev *dev)
1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416
{
	/*
	 * Disable the MSI hardware to avoid screaming interrupts
	 * during boot.  This is the power on reset default so
	 * usually this should be a noop.
	 */
	dev->msi_cap = pci_find_capability(dev, PCI_CAP_ID_MSI);
	if (dev->msi_cap)
		pci_msi_set_enable(dev, 0);

	dev->msix_cap = pci_find_capability(dev, PCI_CAP_ID_MSIX);
	if (dev->msix_cap)
		pci_msix_clear_and_set_ctrl(dev, PCI_MSIX_FLAGS_ENABLE, 0);
}

1417
/**
B
Bjorn Helgaas 已提交
1418
 * pci_intx_mask_broken - Test PCI_COMMAND_INTX_DISABLE writability
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
 * @dev: PCI device
 *
 * Test whether PCI_COMMAND_INTX_DISABLE is writable for @dev.  Check this
 * at enumeration-time to avoid modifying PCI_COMMAND at run-time.
 */
static int pci_intx_mask_broken(struct pci_dev *dev)
{
	u16 orig, toggle, new;

	pci_read_config_word(dev, PCI_COMMAND, &orig);
	toggle = orig ^ PCI_COMMAND_INTX_DISABLE;
	pci_write_config_word(dev, PCI_COMMAND, toggle);
	pci_read_config_word(dev, PCI_COMMAND, &new);

	pci_write_config_word(dev, PCI_COMMAND, orig);

	/*
	 * PCI_COMMAND_INTX_DISABLE was reserved and read-only prior to PCI
	 * r2.3, so strictly speaking, a device is not *broken* if it's not
	 * writable.  But we'll live with the misnomer for now.
	 */
	if (new != toggle)
		return 1;
	return 0;
}

L
Linus Torvalds 已提交
1445
/**
B
Bjorn Helgaas 已提交
1446
 * pci_setup_device - Fill in class and map information of a device
L
Linus Torvalds 已提交
1447 1448
 * @dev: the device structure to fill
 *
1449
 * Initialize the device structure with information about the device's
B
Bjorn Helgaas 已提交
1450
 * vendor,class,memory and IO-space addresses, IRQ lines etc.
L
Linus Torvalds 已提交
1451
 * Called at initialisation of the PCI subsystem and by CardBus services.
Y
Yu Zhao 已提交
1452 1453
 * Returns 0 on success and negative if unknown type of device (not normal,
 * bridge or CardBus).
L
Linus Torvalds 已提交
1454
 */
Y
Yu Zhao 已提交
1455
int pci_setup_device(struct pci_dev *dev)
L
Linus Torvalds 已提交
1456 1457
{
	u32 class;
1458
	u16 cmd;
Y
Yu Zhao 已提交
1459
	u8 hdr_type;
1460
	int pos = 0;
1461 1462
	struct pci_bus_region region;
	struct resource *res;
Y
Yu Zhao 已提交
1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474

	if (pci_read_config_byte(dev, PCI_HEADER_TYPE, &hdr_type))
		return -EIO;

	dev->sysdata = dev->bus->sysdata;
	dev->dev.parent = dev->bus->bridge;
	dev->dev.bus = &pci_bus_type;
	dev->hdr_type = hdr_type & 0x7f;
	dev->multifunction = !!(hdr_type & 0x80);
	dev->error_state = pci_channel_io_normal;
	set_pcie_port_type(dev);

1475
	pci_dev_assign_slot(dev);
B
Bjorn Helgaas 已提交
1476 1477 1478 1479 1480

	/*
	 * Assume 32-bit PCI; let 64-bit PCI cards (which are far rarer)
	 * set this higher, assuming the system even supports it.
	 */
Y
Yu Zhao 已提交
1481
	dev->dma_mask = 0xffffffff;
L
Linus Torvalds 已提交
1482

1483 1484 1485
	dev_set_name(&dev->dev, "%04x:%02x:%02x.%d", pci_domain_nr(dev->bus),
		     dev->bus->number, PCI_SLOT(dev->devfn),
		     PCI_FUNC(dev->devfn));
L
Linus Torvalds 已提交
1486 1487

	pci_read_config_dword(dev, PCI_CLASS_REVISION, &class);
A
Auke Kok 已提交
1488
	dev->revision = class & 0xff;
Y
Yinghai Lu 已提交
1489
	dev->class = class >> 8;		    /* upper 3 bytes */
L
Linus Torvalds 已提交
1490

1491
	pci_printk(KERN_DEBUG, dev, "[%04x:%04x] type %02x class %#08x\n",
Y
Yinghai Lu 已提交
1492
		   dev->vendor, dev->device, dev->hdr_type, dev->class);
L
Linus Torvalds 已提交
1493

B
Bjorn Helgaas 已提交
1494
	/* Need to have dev->class ready */
1495 1496
	dev->cfg_size = pci_cfg_space_size(dev);

B
Bjorn Helgaas 已提交
1497
	/* Need to have dev->cfg_size ready */
1498 1499
	set_pcie_thunderbolt(dev);

L
Linus Torvalds 已提交
1500
	/* "Unknown power state" */
1501
	dev->current_state = PCI_UNKNOWN;
L
Linus Torvalds 已提交
1502 1503 1504

	/* Early fixups, before probing the BARs */
	pci_fixup_device(pci_fixup_early, dev);
B
Bjorn Helgaas 已提交
1505 1506

	/* Device class may be changed after fixup */
1507
	class = dev->class >> 8;
L
Linus Torvalds 已提交
1508

1509 1510 1511
	if (dev->non_compliant_bars) {
		pci_read_config_word(dev, PCI_COMMAND, &cmd);
		if (cmd & (PCI_COMMAND_IO | PCI_COMMAND_MEMORY)) {
1512
			pci_info(dev, "device has non-compliant BARs; disabling IO/MEM decoding\n");
1513 1514 1515 1516 1517 1518
			cmd &= ~PCI_COMMAND_IO;
			cmd &= ~PCI_COMMAND_MEMORY;
			pci_write_config_word(dev, PCI_COMMAND, cmd);
		}
	}

1519 1520
	dev->broken_intx_masking = pci_intx_mask_broken(dev);

L
Linus Torvalds 已提交
1521 1522 1523 1524 1525 1526 1527 1528
	switch (dev->hdr_type) {		    /* header type */
	case PCI_HEADER_TYPE_NORMAL:		    /* standard header */
		if (class == PCI_CLASS_BRIDGE_PCI)
			goto bad;
		pci_read_irq(dev);
		pci_read_bases(dev, 6, PCI_ROM_ADDRESS);
		pci_read_config_word(dev, PCI_SUBSYSTEM_VENDOR_ID, &dev->subsystem_vendor);
		pci_read_config_word(dev, PCI_SUBSYSTEM_ID, &dev->subsystem_device);
1529 1530

		/*
1531 1532 1533 1534
		 * Do the ugly legacy mode stuff here rather than broken chip
		 * quirk code. Legacy mode ATA controllers have fixed
		 * addresses. These are not always echoed in BAR0-3, and
		 * BAR0-3 in a few cases contain junk!
1535 1536 1537 1538 1539
		 */
		if (class == PCI_CLASS_STORAGE_IDE) {
			u8 progif;
			pci_read_config_byte(dev, PCI_CLASS_PROG, &progif);
			if ((progif & 1) == 0) {
1540 1541 1542 1543
				region.start = 0x1F0;
				region.end = 0x1F7;
				res = &dev->resource[0];
				res->flags = LEGACY_IO_RESOURCE;
1544
				pcibios_bus_to_resource(dev->bus, res, &region);
1545
				pci_info(dev, "legacy IDE quirk: reg 0x10: %pR\n",
1546
					 res);
1547 1548 1549 1550
				region.start = 0x3F6;
				region.end = 0x3F6;
				res = &dev->resource[1];
				res->flags = LEGACY_IO_RESOURCE;
1551
				pcibios_bus_to_resource(dev->bus, res, &region);
1552
				pci_info(dev, "legacy IDE quirk: reg 0x14: %pR\n",
1553
					 res);
1554 1555
			}
			if ((progif & 4) == 0) {
1556 1557 1558 1559
				region.start = 0x170;
				region.end = 0x177;
				res = &dev->resource[2];
				res->flags = LEGACY_IO_RESOURCE;
1560
				pcibios_bus_to_resource(dev->bus, res, &region);
1561
				pci_info(dev, "legacy IDE quirk: reg 0x18: %pR\n",
1562
					 res);
1563 1564 1565 1566
				region.start = 0x376;
				region.end = 0x376;
				res = &dev->resource[3];
				res->flags = LEGACY_IO_RESOURCE;
1567
				pcibios_bus_to_resource(dev->bus, res, &region);
1568
				pci_info(dev, "legacy IDE quirk: reg 0x1c: %pR\n",
1569
					 res);
1570 1571
			}
		}
L
Linus Torvalds 已提交
1572 1573 1574 1575 1576
		break;

	case PCI_HEADER_TYPE_BRIDGE:		    /* bridge header */
		if (class != PCI_CLASS_BRIDGE_PCI)
			goto bad;
B
Bjorn Helgaas 已提交
1577 1578 1579 1580 1581 1582

		/*
		 * The PCI-to-PCI bridge spec requires that subtractive
		 * decoding (i.e. transparent) bridge must have programming
		 * interface code of 0x01.
		 */
1583
		pci_read_irq(dev);
L
Linus Torvalds 已提交
1584 1585
		dev->transparent = ((dev->class & 0xff) == 1);
		pci_read_bases(dev, 2, PCI_ROM_ADDRESS1);
1586
		set_pcie_hotplug_bridge(dev);
1587 1588 1589 1590 1591
		pos = pci_find_capability(dev, PCI_CAP_ID_SSVID);
		if (pos) {
			pci_read_config_word(dev, pos + PCI_SSVID_VENDOR_ID, &dev->subsystem_vendor);
			pci_read_config_word(dev, pos + PCI_SSVID_DEVICE_ID, &dev->subsystem_device);
		}
L
Linus Torvalds 已提交
1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603
		break;

	case PCI_HEADER_TYPE_CARDBUS:		    /* CardBus bridge header */
		if (class != PCI_CLASS_BRIDGE_CARDBUS)
			goto bad;
		pci_read_irq(dev);
		pci_read_bases(dev, 1, 0);
		pci_read_config_word(dev, PCI_CB_SUBSYSTEM_VENDOR_ID, &dev->subsystem_vendor);
		pci_read_config_word(dev, PCI_CB_SUBSYSTEM_ID, &dev->subsystem_device);
		break;

	default:				    /* unknown header */
1604
		pci_err(dev, "unknown header type %02x, ignoring device\n",
1605
			dev->hdr_type);
Y
Yu Zhao 已提交
1606
		return -EIO;
L
Linus Torvalds 已提交
1607 1608

	bad:
1609
		pci_err(dev, "ignoring class %#08x (doesn't match header type %02x)\n",
1610
			dev->class, dev->hdr_type);
1611
		dev->class = PCI_CLASS_NOT_DEFINED << 8;
L
Linus Torvalds 已提交
1612 1613 1614 1615 1616 1617
	}

	/* We found a fine healthy device, go go go... */
	return 0;
}

1618 1619 1620
static void pci_configure_mps(struct pci_dev *dev)
{
	struct pci_dev *bridge = pci_upstream_bridge(dev);
1621
	int mps, p_mps, rc;
1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632

	if (!pci_is_pcie(dev) || !bridge || !pci_is_pcie(bridge))
		return;

	mps = pcie_get_mps(dev);
	p_mps = pcie_get_mps(bridge);

	if (mps == p_mps)
		return;

	if (pcie_bus_config == PCIE_BUS_TUNE_OFF) {
1633
		pci_warn(dev, "Max Payload Size %d, but upstream %s set to %d; if necessary, use \"pci=pcie_bus_safe\" and report a bug\n",
1634 1635 1636
			 mps, pci_name(bridge), p_mps);
		return;
	}
1637 1638 1639 1640 1641 1642 1643 1644 1645 1646

	/*
	 * Fancier MPS configuration is done later by
	 * pcie_bus_configure_settings()
	 */
	if (pcie_bus_config != PCIE_BUS_DEFAULT)
		return;

	rc = pcie_set_mps(dev, p_mps);
	if (rc) {
1647
		pci_warn(dev, "can't set Max Payload Size to %d; if necessary, use \"pci=pcie_bus_safe\" and report a bug\n",
1648 1649 1650 1651
			 p_mps);
		return;
	}

1652
	pci_info(dev, "Max Payload Size set to %d (was %d, max %d)\n",
1653
		 p_mps, mps, 128 << dev->pcie_mpss);
1654 1655
}

1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667
static struct hpp_type0 pci_default_type0 = {
	.revision = 1,
	.cache_line_size = 8,
	.latency_timer = 0x40,
	.enable_serr = 0,
	.enable_perr = 0,
};

static void program_hpp_type0(struct pci_dev *dev, struct hpp_type0 *hpp)
{
	u16 pci_cmd, pci_bctl;

1668
	if (!hpp)
1669 1670 1671
		hpp = &pci_default_type0;

	if (hpp->revision > 1) {
1672
		pci_warn(dev, "PCI settings rev %d not supported; using defaults\n",
1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700
			 hpp->revision);
		hpp = &pci_default_type0;
	}

	pci_write_config_byte(dev, PCI_CACHE_LINE_SIZE, hpp->cache_line_size);
	pci_write_config_byte(dev, PCI_LATENCY_TIMER, hpp->latency_timer);
	pci_read_config_word(dev, PCI_COMMAND, &pci_cmd);
	if (hpp->enable_serr)
		pci_cmd |= PCI_COMMAND_SERR;
	if (hpp->enable_perr)
		pci_cmd |= PCI_COMMAND_PARITY;
	pci_write_config_word(dev, PCI_COMMAND, pci_cmd);

	/* Program bridge control value */
	if ((dev->class >> 8) == PCI_CLASS_BRIDGE_PCI) {
		pci_write_config_byte(dev, PCI_SEC_LATENCY_TIMER,
				      hpp->latency_timer);
		pci_read_config_word(dev, PCI_BRIDGE_CONTROL, &pci_bctl);
		if (hpp->enable_serr)
			pci_bctl |= PCI_BRIDGE_CTL_SERR;
		if (hpp->enable_perr)
			pci_bctl |= PCI_BRIDGE_CTL_PARITY;
		pci_write_config_word(dev, PCI_BRIDGE_CONTROL, pci_bctl);
	}
}

static void program_hpp_type1(struct pci_dev *dev, struct hpp_type1 *hpp)
{
1701 1702 1703 1704 1705 1706 1707 1708 1709
	int pos;

	if (!hpp)
		return;

	pos = pci_find_capability(dev, PCI_CAP_ID_PCIX);
	if (!pos)
		return;

1710
	pci_warn(dev, "PCI-X settings not supported\n");
1711 1712
}

1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727
static bool pcie_root_rcb_set(struct pci_dev *dev)
{
	struct pci_dev *rp = pcie_find_root_port(dev);
	u16 lnkctl;

	if (!rp)
		return false;

	pcie_capability_read_word(rp, PCI_EXP_LNKCTL, &lnkctl);
	if (lnkctl & PCI_EXP_LNKCTL_RCB)
		return true;

	return false;
}

1728 1729 1730 1731 1732 1733 1734 1735
static void program_hpp_type2(struct pci_dev *dev, struct hpp_type2 *hpp)
{
	int pos;
	u32 reg32;

	if (!hpp)
		return;

1736 1737 1738
	if (!pci_is_pcie(dev))
		return;

1739
	if (hpp->revision > 1) {
1740
		pci_warn(dev, "PCIe settings rev %d not supported\n",
1741 1742 1743 1744
			 hpp->revision);
		return;
	}

1745 1746 1747 1748 1749 1750 1751 1752 1753 1754
	/*
	 * Don't allow _HPX to change MPS or MRRS settings.  We manage
	 * those to make sure they're consistent with the rest of the
	 * platform.
	 */
	hpp->pci_exp_devctl_and |= PCI_EXP_DEVCTL_PAYLOAD |
				    PCI_EXP_DEVCTL_READRQ;
	hpp->pci_exp_devctl_or &= ~(PCI_EXP_DEVCTL_PAYLOAD |
				    PCI_EXP_DEVCTL_READRQ);

1755 1756 1757 1758 1759
	/* Initialize Device Control Register */
	pcie_capability_clear_and_set_word(dev, PCI_EXP_DEVCTL,
			~hpp->pci_exp_devctl_and, hpp->pci_exp_devctl_or);

	/* Initialize Link Control Register */
1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770
	if (pcie_cap_has_lnkctl(dev)) {

		/*
		 * If the Root Port supports Read Completion Boundary of
		 * 128, set RCB to 128.  Otherwise, clear it.
		 */
		hpp->pci_exp_lnkctl_and |= PCI_EXP_LNKCTL_RCB;
		hpp->pci_exp_lnkctl_or &= ~PCI_EXP_LNKCTL_RCB;
		if (pcie_root_rcb_set(dev))
			hpp->pci_exp_lnkctl_or |= PCI_EXP_LNKCTL_RCB;

1771 1772
		pcie_capability_clear_and_set_word(dev, PCI_EXP_LNKCTL,
			~hpp->pci_exp_lnkctl_and, hpp->pci_exp_lnkctl_or);
1773
	}
1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797

	/* Find Advanced Error Reporting Enhanced Capability */
	pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ERR);
	if (!pos)
		return;

	/* Initialize Uncorrectable Error Mask Register */
	pci_read_config_dword(dev, pos + PCI_ERR_UNCOR_MASK, &reg32);
	reg32 = (reg32 & hpp->unc_err_mask_and) | hpp->unc_err_mask_or;
	pci_write_config_dword(dev, pos + PCI_ERR_UNCOR_MASK, reg32);

	/* Initialize Uncorrectable Error Severity Register */
	pci_read_config_dword(dev, pos + PCI_ERR_UNCOR_SEVER, &reg32);
	reg32 = (reg32 & hpp->unc_err_sever_and) | hpp->unc_err_sever_or;
	pci_write_config_dword(dev, pos + PCI_ERR_UNCOR_SEVER, reg32);

	/* Initialize Correctable Error Mask Register */
	pci_read_config_dword(dev, pos + PCI_ERR_COR_MASK, &reg32);
	reg32 = (reg32 & hpp->cor_err_mask_and) | hpp->cor_err_mask_or;
	pci_write_config_dword(dev, pos + PCI_ERR_COR_MASK, reg32);

	/* Initialize Advanced Error Capabilities and Control Register */
	pci_read_config_dword(dev, pos + PCI_ERR_CAP, &reg32);
	reg32 = (reg32 & hpp->adv_err_cap_and) | hpp->adv_err_cap_or;
B
Bjorn Helgaas 已提交
1798

1799 1800 1801 1802 1803
	/* Don't enable ECRC generation or checking if unsupported */
	if (!(reg32 & PCI_ERR_CAP_ECRC_GENC))
		reg32 &= ~PCI_ERR_CAP_ECRC_GENE;
	if (!(reg32 & PCI_ERR_CAP_ECRC_CHKC))
		reg32 &= ~PCI_ERR_CAP_ECRC_CHKE;
1804 1805 1806 1807 1808 1809 1810 1811 1812 1813
	pci_write_config_dword(dev, pos + PCI_ERR_CAP, reg32);

	/*
	 * FIXME: The following two registers are not supported yet.
	 *
	 *   o Secondary Uncorrectable Error Severity Register
	 *   o Secondary Uncorrectable Error Mask Register
	 */
}

1814
int pci_configure_extended_tags(struct pci_dev *dev, void *ign)
1815
{
1816 1817 1818
	struct pci_host_bridge *host;
	u32 cap;
	u16 ctl;
1819 1820 1821
	int ret;

	if (!pci_is_pcie(dev))
1822
		return 0;
1823

1824
	ret = pcie_capability_read_dword(dev, PCI_EXP_DEVCAP, &cap);
1825
	if (ret)
1826 1827 1828 1829
		return 0;

	if (!(cap & PCI_EXP_DEVCAP_EXT_TAG))
		return 0;
1830

1831 1832 1833 1834 1835 1836 1837
	ret = pcie_capability_read_word(dev, PCI_EXP_DEVCTL, &ctl);
	if (ret)
		return 0;

	host = pci_find_host_bridge(dev->bus);
	if (!host)
		return 0;
1838

1839 1840 1841 1842 1843 1844
	/*
	 * If some device in the hierarchy doesn't handle Extended Tags
	 * correctly, make sure they're disabled.
	 */
	if (host->no_ext_tags) {
		if (ctl & PCI_EXP_DEVCTL_EXT_TAG) {
1845
			pci_info(dev, "disabling Extended Tags\n");
1846 1847 1848 1849 1850 1851 1852
			pcie_capability_clear_word(dev, PCI_EXP_DEVCTL,
						   PCI_EXP_DEVCTL_EXT_TAG);
		}
		return 0;
	}

	if (!(ctl & PCI_EXP_DEVCTL_EXT_TAG)) {
1853
		pci_info(dev, "enabling Extended Tags\n");
1854 1855
		pcie_capability_set_word(dev, PCI_EXP_DEVCTL,
					 PCI_EXP_DEVCTL_EXT_TAG);
1856 1857
	}
	return 0;
1858 1859
}

1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897
/**
 * pcie_relaxed_ordering_enabled - Probe for PCIe relaxed ordering enable
 * @dev: PCI device to query
 *
 * Returns true if the device has enabled relaxed ordering attribute.
 */
bool pcie_relaxed_ordering_enabled(struct pci_dev *dev)
{
	u16 v;

	pcie_capability_read_word(dev, PCI_EXP_DEVCTL, &v);

	return !!(v & PCI_EXP_DEVCTL_RELAX_EN);
}
EXPORT_SYMBOL(pcie_relaxed_ordering_enabled);

static void pci_configure_relaxed_ordering(struct pci_dev *dev)
{
	struct pci_dev *root;

	/* PCI_EXP_DEVICE_RELAX_EN is RsvdP in VFs */
	if (dev->is_virtfn)
		return;

	if (!pcie_relaxed_ordering_enabled(dev))
		return;

	/*
	 * For now, we only deal with Relaxed Ordering issues with Root
	 * Ports. Peer-to-Peer DMA is another can of worms.
	 */
	root = pci_find_pcie_root_port(dev);
	if (!root)
		return;

	if (root->dev_flags & PCI_DEV_FLAGS_NO_RELAXED_ORDERING) {
		pcie_capability_clear_word(dev, PCI_EXP_DEVCTL,
					   PCI_EXP_DEVCTL_RELAX_EN);
1898
		pci_info(dev, "Relaxed Ordering disabled because the Root Port didn't support it\n");
1899 1900 1901
	}
}

1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933
static void pci_configure_ltr(struct pci_dev *dev)
{
#ifdef CONFIG_PCIEASPM
	u32 cap;
	struct pci_dev *bridge;

	if (!pci_is_pcie(dev))
		return;

	pcie_capability_read_dword(dev, PCI_EXP_DEVCAP2, &cap);
	if (!(cap & PCI_EXP_DEVCAP2_LTR))
		return;

	/*
	 * Software must not enable LTR in an Endpoint unless the Root
	 * Complex and all intermediate Switches indicate support for LTR.
	 * PCIe r3.1, sec 6.18.
	 */
	if (pci_pcie_type(dev) == PCI_EXP_TYPE_ROOT_PORT)
		dev->ltr_path = 1;
	else {
		bridge = pci_upstream_bridge(dev);
		if (bridge && bridge->ltr_path)
			dev->ltr_path = 1;
	}

	if (dev->ltr_path)
		pcie_capability_set_word(dev, PCI_EXP_DEVCTL2,
					 PCI_EXP_DEVCTL2_LTR_EN);
#endif
}

1934 1935 1936 1937 1938
static void pci_configure_device(struct pci_dev *dev)
{
	struct hotplug_params hpp;
	int ret;

1939
	pci_configure_mps(dev);
1940
	pci_configure_extended_tags(dev, NULL);
1941
	pci_configure_relaxed_ordering(dev);
1942
	pci_configure_ltr(dev);
1943

1944 1945 1946 1947 1948 1949 1950 1951 1952 1953
	memset(&hpp, 0, sizeof(hpp));
	ret = pci_get_hp_params(dev, &hpp);
	if (ret)
		return;

	program_hpp_type2(dev, hpp.t2);
	program_hpp_type1(dev, hpp.t1);
	program_hpp_type0(dev, hpp.t0);
}

1954 1955 1956
static void pci_release_capabilities(struct pci_dev *dev)
{
	pci_vpd_release(dev);
1957
	pci_iov_release(dev);
1958
	pci_free_cap_save_buffers(dev);
1959 1960
}

L
Linus Torvalds 已提交
1961
/**
B
Bjorn Helgaas 已提交
1962 1963
 * pci_release_dev - Free a PCI device structure when all users of it are
 *		     finished
L
Linus Torvalds 已提交
1964 1965
 * @dev: device that's been disconnected
 *
B
Bjorn Helgaas 已提交
1966
 * Will be called only by the device core when all users of this PCI device are
L
Linus Torvalds 已提交
1967 1968 1969 1970
 * done.
 */
static void pci_release_dev(struct device *dev)
{
1971
	struct pci_dev *pci_dev;
L
Linus Torvalds 已提交
1972

1973
	pci_dev = to_pci_dev(dev);
1974
	pci_release_capabilities(pci_dev);
1975
	pci_release_of_node(pci_dev);
1976
	pcibios_release_device(pci_dev);
1977
	pci_bus_put(pci_dev->bus);
1978
	kfree(pci_dev->driver_override);
1979
	kfree(pci_dev->dma_alias_mask);
L
Linus Torvalds 已提交
1980 1981 1982
	kfree(pci_dev);
}

1983
struct pci_dev *pci_alloc_dev(struct pci_bus *bus)
1984 1985 1986 1987 1988 1989 1990 1991
{
	struct pci_dev *dev;

	dev = kzalloc(sizeof(struct pci_dev), GFP_KERNEL);
	if (!dev)
		return NULL;

	INIT_LIST_HEAD(&dev->bus_list);
1992
	dev->dev.type = &pci_dev_type;
1993
	dev->bus = pci_bus_get(bus);
1994 1995 1996

	return dev;
}
1997 1998
EXPORT_SYMBOL(pci_alloc_dev);

1999 2000 2001 2002 2003
static bool pci_bus_crs_vendor_id(u32 l)
{
	return (l & 0xffff) == 0x0001;
}

S
Sinan Kaya 已提交
2004 2005
static bool pci_bus_wait_crs(struct pci_bus *bus, int devfn, u32 *l,
			     int timeout)
L
Linus Torvalds 已提交
2006 2007 2008
{
	int delay = 1;

S
Sinan Kaya 已提交
2009 2010
	if (!pci_bus_crs_vendor_id(*l))
		return true;	/* not a CRS completion */
L
Linus Torvalds 已提交
2011

S
Sinan Kaya 已提交
2012 2013
	if (!timeout)
		return false;	/* CRS, but caller doesn't want to wait */
L
Linus Torvalds 已提交
2014

2015
	/*
S
Sinan Kaya 已提交
2016 2017 2018
	 * We got the reserved Vendor ID that indicates a completion with
	 * Configuration Request Retry Status (CRS).  Retry until we get a
	 * valid Vendor ID or we time out.
2019
	 */
2020
	while (pci_bus_crs_vendor_id(*l)) {
S
Sinan Kaya 已提交
2021
		if (delay > timeout) {
2022 2023 2024 2025
			pr_warn("pci %04x:%02x:%02x.%d: not ready after %dms; giving up\n",
				pci_domain_nr(bus), bus->number,
				PCI_SLOT(devfn), PCI_FUNC(devfn), delay - 1);

2026
			return false;
L
Linus Torvalds 已提交
2027
		}
2028 2029 2030 2031
		if (delay >= 1000)
			pr_info("pci %04x:%02x:%02x.%d: not ready after %dms; waiting\n",
				pci_domain_nr(bus), bus->number,
				PCI_SLOT(devfn), PCI_FUNC(devfn), delay - 1);
2032

L
Linus Torvalds 已提交
2033 2034
		msleep(delay);
		delay *= 2;
2035

2036 2037
		if (pci_bus_read_config_dword(bus, devfn, PCI_VENDOR_ID, l))
			return false;
L
Linus Torvalds 已提交
2038 2039
	}

2040 2041 2042 2043 2044
	if (delay >= 1000)
		pr_info("pci %04x:%02x:%02x.%d: ready after %dms\n",
			pci_domain_nr(bus), bus->number,
			PCI_SLOT(devfn), PCI_FUNC(devfn), delay - 1);

2045 2046
	return true;
}
S
Sinan Kaya 已提交
2047 2048 2049 2050 2051 2052 2053

bool pci_bus_read_dev_vendor_id(struct pci_bus *bus, int devfn, u32 *l,
				int timeout)
{
	if (pci_bus_read_config_dword(bus, devfn, PCI_VENDOR_ID, l))
		return false;

B
Bjorn Helgaas 已提交
2054
	/* Some broken boards return 0 or ~0 if a slot is empty: */
S
Sinan Kaya 已提交
2055 2056 2057 2058 2059 2060 2061
	if (*l == 0xffffffff || *l == 0x00000000 ||
	    *l == 0x0000ffff || *l == 0xffff0000)
		return false;

	if (pci_bus_crs_vendor_id(*l))
		return pci_bus_wait_crs(bus, devfn, l, timeout);

2062 2063 2064 2065 2066
	return true;
}
EXPORT_SYMBOL(pci_bus_read_dev_vendor_id);

/*
B
Bjorn Helgaas 已提交
2067 2068
 * Read the config data for a PCI device, sanity-check it,
 * and fill in the dev structure.
2069 2070 2071 2072 2073 2074 2075 2076 2077
 */
static struct pci_dev *pci_scan_device(struct pci_bus *bus, int devfn)
{
	struct pci_dev *dev;
	u32 l;

	if (!pci_bus_read_dev_vendor_id(bus, devfn, &l, 60*1000))
		return NULL;

2078
	dev = pci_alloc_dev(bus);
L
Linus Torvalds 已提交
2079 2080 2081 2082 2083 2084
	if (!dev)
		return NULL;

	dev->devfn = devfn;
	dev->vendor = l & 0xffff;
	dev->device = (l >> 16) & 0xffff;
2085

2086 2087
	pci_set_of_node(dev);

Y
Yu Zhao 已提交
2088
	if (pci_setup_device(dev)) {
2089
		pci_bus_put(dev->bus);
L
Linus Torvalds 已提交
2090 2091 2092 2093 2094 2095 2096
		kfree(dev);
		return NULL;
	}

	return dev;
}

2097 2098
static void pci_init_capabilities(struct pci_dev *dev)
{
2099 2100 2101
	/* Enhanced Allocation */
	pci_ea_init(dev);

2102 2103
	/* Setup MSI caps & disable MSI/MSI-X interrupts */
	pci_msi_setup_pci_dev(dev);
2104

2105 2106 2107
	/* Buffers for saving PCIe and PCI-X capabilities */
	pci_allocate_cap_save_buffers(dev);

2108 2109 2110 2111
	/* Power Management */
	pci_pm_init(dev);

	/* Vital Product Data */
2112
	pci_vpd_init(dev);
Y
Yu Zhao 已提交
2113 2114

	/* Alternative Routing-ID Forwarding */
2115
	pci_configure_ari(dev);
2116 2117 2118

	/* Single Root I/O Virtualization */
	pci_iov_init(dev);
2119

2120 2121 2122
	/* Address Translation Services */
	pci_ats_init(dev);

2123
	/* Enable ACS P2P upstream forwarding */
C
Chris Wright 已提交
2124
	pci_enable_acs(dev);
2125

2126 2127
	/* Precision Time Measurement */
	pci_ptm_init(dev);
B
Bjorn Helgaas 已提交
2128

K
Keith Busch 已提交
2129 2130
	/* Advanced Error Reporting */
	pci_aer_init(dev);
2131 2132 2133

	if (pci_probe_reset_function(dev) == 0)
		dev->reset_fn = 1;
2134 2135
}

M
Marc Zyngier 已提交
2136
/*
B
Bjorn Helgaas 已提交
2137
 * This is the equivalent of pci_host_bridge_msi_domain() that acts on
M
Marc Zyngier 已提交
2138 2139 2140 2141 2142 2143 2144 2145
 * devices. Firmware interfaces that can select the MSI domain on a
 * per-device basis should be called from here.
 */
static struct irq_domain *pci_dev_msi_domain(struct pci_dev *dev)
{
	struct irq_domain *d;

	/*
B
Bjorn Helgaas 已提交
2146
	 * If a domain has been set through the pcibios_add_device()
M
Marc Zyngier 已提交
2147 2148 2149 2150 2151 2152
	 * callback, then this is the one (platform code knows best).
	 */
	d = dev_get_msi_domain(&dev->dev);
	if (d)
		return d;

2153 2154 2155 2156 2157 2158 2159 2160
	/*
	 * Let's see if we have a firmware interface able to provide
	 * the domain.
	 */
	d = pci_msi_get_device_domain(dev);
	if (d)
		return d;

M
Marc Zyngier 已提交
2161 2162 2163
	return NULL;
}

2164 2165
static void pci_set_msi_domain(struct pci_dev *dev)
{
M
Marc Zyngier 已提交
2166 2167
	struct irq_domain *d;

2168
	/*
M
Marc Zyngier 已提交
2169 2170 2171
	 * If the platform or firmware interfaces cannot supply a
	 * device-specific MSI domain, then inherit the default domain
	 * from the host bridge itself.
2172
	 */
M
Marc Zyngier 已提交
2173 2174 2175 2176 2177
	d = pci_dev_msi_domain(dev);
	if (!d)
		d = dev_get_msi_domain(&dev->bus->dev);

	dev_set_msi_domain(&dev->dev, d);
2178 2179
}

2180
void pci_device_add(struct pci_dev *dev, struct pci_bus *bus)
L
Linus Torvalds 已提交
2181
{
2182 2183
	int ret;

2184 2185
	pci_configure_device(dev);

2186 2187
	device_initialize(&dev->dev);
	dev->dev.release = pci_release_dev;
L
Linus Torvalds 已提交
2188

2189
	set_dev_node(&dev->dev, pcibus_to_node(bus));
2190
	dev->dev.dma_mask = &dev->dma_mask;
2191
	dev->dev.dma_parms = &dev->dma_parms;
2192
	dev->dev.coherent_dma_mask = 0xffffffffull;
L
Linus Torvalds 已提交
2193

2194
	pci_set_dma_max_seg_size(dev, 65536);
2195
	pci_set_dma_seg_boundary(dev, 0xffffffff);
2196

L
Linus Torvalds 已提交
2197 2198 2199
	/* Fix up broken headers */
	pci_fixup_device(pci_fixup_header, dev);

B
Bjorn Helgaas 已提交
2200
	/* Moved out from quirk header fixup code */
2201 2202
	pci_reassigndev_resource_alignment(dev);

B
Bjorn Helgaas 已提交
2203
	/* Clear the state_saved flag */
2204 2205
	dev->state_saved = false;

2206 2207
	/* Initialize various capabilities */
	pci_init_capabilities(dev);
2208

L
Linus Torvalds 已提交
2209 2210 2211 2212
	/*
	 * Add the device to our list of discovered devices
	 * and the bus list for fixup functions, etc.
	 */
2213
	down_write(&pci_bus_sem);
L
Linus Torvalds 已提交
2214
	list_add_tail(&dev->bus_list, &bus->devices);
2215
	up_write(&pci_bus_sem);
2216 2217 2218 2219

	ret = pcibios_add_device(dev);
	WARN_ON(ret < 0);

B
Bjorn Helgaas 已提交
2220
	/* Set up MSI IRQ domain */
2221 2222
	pci_set_msi_domain(dev);

2223 2224 2225 2226
	/* Notifier could use PCI capabilities */
	dev->match_driver = false;
	ret = device_add(&dev->dev);
	WARN_ON(ret < 0);
2227 2228
}

2229
struct pci_dev *pci_scan_single_device(struct pci_bus *bus, int devfn)
2230 2231 2232
{
	struct pci_dev *dev;

T
Trent Piepho 已提交
2233 2234 2235 2236 2237 2238
	dev = pci_get_slot(bus, devfn);
	if (dev) {
		pci_dev_put(dev);
		return dev;
	}

2239 2240 2241 2242 2243
	dev = pci_scan_device(bus, devfn);
	if (!dev)
		return NULL;

	pci_device_add(dev, bus);
L
Linus Torvalds 已提交
2244 2245 2246

	return dev;
}
2247
EXPORT_SYMBOL(pci_scan_single_device);
L
Linus Torvalds 已提交
2248

2249
static unsigned next_fn(struct pci_bus *bus, struct pci_dev *dev, unsigned fn)
M
Matthew Wilcox 已提交
2250
{
2251 2252 2253
	int pos;
	u16 cap = 0;
	unsigned next_fn;
2254

2255 2256 2257 2258 2259 2260
	if (pci_ari_enabled(bus)) {
		if (!dev)
			return 0;
		pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ARI);
		if (!pos)
			return 0;
2261

2262 2263 2264 2265
		pci_read_config_word(dev, pos + PCI_ARI_CAP, &cap);
		next_fn = PCI_ARI_CAP_NFN(cap);
		if (next_fn <= fn)
			return 0;	/* protect against malformed list */
M
Matthew Wilcox 已提交
2266

2267 2268 2269 2270 2271 2272
		return next_fn;
	}

	/* dev may be NULL for non-contiguous multifunction devices */
	if (!dev || dev->multifunction)
		return (fn + 1) % 8;
M
Matthew Wilcox 已提交
2273 2274 2275 2276 2277 2278

	return 0;
}

static int only_one_child(struct pci_bus *bus)
{
2279
	struct pci_dev *bridge = bus->self;
2280

2281 2282 2283 2284 2285
	/*
	 * Systems with unusual topologies set PCI_SCAN_ALL_PCIE_DEVS so
	 * we scan for all possible devices, not just Device 0.
	 */
	if (pci_has_flag(PCI_SCAN_ALL_PCIE_DEVS))
M
Matthew Wilcox 已提交
2286
		return 0;
2287 2288

	/*
2289 2290 2291 2292 2293 2294 2295
	 * A PCIe Downstream Port normally leads to a Link with only Device
	 * 0 on it (PCIe spec r3.1, sec 7.3.1).  As an optimization, scan
	 * only for Device 0 in that situation.
	 *
	 * Checking has_secondary_link is a hack to identify Downstream
	 * Ports because sometimes Switches are configured such that the
	 * PCIe Port Type labels are backwards.
2296
	 */
2297
	if (bridge && pci_is_pcie(bridge) && bridge->has_secondary_link)
M
Matthew Wilcox 已提交
2298
		return 1;
2299

M
Matthew Wilcox 已提交
2300 2301 2302
	return 0;
}

L
Linus Torvalds 已提交
2303
/**
B
Bjorn Helgaas 已提交
2304
 * pci_scan_slot - Scan a PCI slot on a bus for devices
L
Linus Torvalds 已提交
2305
 * @bus: PCI bus to scan
B
Bjorn Helgaas 已提交
2306
 * @devfn: slot number to scan (must have zero function)
L
Linus Torvalds 已提交
2307 2308 2309
 *
 * Scan a PCI slot on the specified PCI bus for devices, adding
 * discovered devices to the @bus->devices list.  New devices
2310
 * will not have is_added set.
2311 2312
 *
 * Returns the number of new devices found.
L
Linus Torvalds 已提交
2313
 */
2314
int pci_scan_slot(struct pci_bus *bus, int devfn)
L
Linus Torvalds 已提交
2315
{
M
Matthew Wilcox 已提交
2316
	unsigned fn, nr = 0;
2317
	struct pci_dev *dev;
M
Matthew Wilcox 已提交
2318 2319 2320

	if (only_one_child(bus) && (devfn > 0))
		return 0; /* Already scanned the entire slot */
L
Linus Torvalds 已提交
2321

2322
	dev = pci_scan_single_device(bus, devfn);
2323 2324 2325
	if (!dev)
		return 0;
	if (!dev->is_added)
2326 2327
		nr++;

2328
	for (fn = next_fn(bus, dev, 0); fn > 0; fn = next_fn(bus, dev, fn)) {
M
Matthew Wilcox 已提交
2329 2330 2331 2332 2333
		dev = pci_scan_single_device(bus, devfn + fn);
		if (dev) {
			if (!dev->is_added)
				nr++;
			dev->multifunction = 1;
L
Linus Torvalds 已提交
2334 2335
		}
	}
S
Shaohua Li 已提交
2336

B
Bjorn Helgaas 已提交
2337
	/* Only one slot has PCIe device */
2338
	if (bus->self && nr)
S
Shaohua Li 已提交
2339 2340
		pcie_aspm_init_link_state(bus->self);

L
Linus Torvalds 已提交
2341 2342
	return nr;
}
2343
EXPORT_SYMBOL(pci_scan_slot);
L
Linus Torvalds 已提交
2344

2345 2346 2347 2348 2349 2350 2351
static int pcie_find_smpss(struct pci_dev *dev, void *data)
{
	u8 *smpss = data;

	if (!pci_is_pcie(dev))
		return 0;

2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365
	/*
	 * We don't have a way to change MPS settings on devices that have
	 * drivers attached.  A hot-added device might support only the minimum
	 * MPS setting (MPS=128).  Therefore, if the fabric contains a bridge
	 * where devices may be hot-added, we limit the fabric MPS to 128 so
	 * hot-added devices will work correctly.
	 *
	 * However, if we hot-add a device to a slot directly below a Root
	 * Port, it's impossible for there to be other existing devices below
	 * the port.  We don't limit the MPS in this case because we can
	 * reconfigure MPS on both the Root Port and the hot-added device,
	 * and there are no other devices involved.
	 *
	 * Note that this PCIE_BUS_SAFE path assumes no peer-to-peer DMA.
2366
	 */
2367 2368
	if (dev->is_hotplug_bridge &&
	    pci_pcie_type(dev) != PCI_EXP_TYPE_ROOT_PORT)
2369 2370 2371 2372 2373 2374 2375 2376 2377 2378
		*smpss = 0;

	if (*smpss > dev->pcie_mpss)
		*smpss = dev->pcie_mpss;

	return 0;
}

static void pcie_write_mps(struct pci_dev *dev, int mps)
{
2379
	int rc;
2380 2381

	if (pcie_bus_config == PCIE_BUS_PERFORMANCE) {
2382
		mps = 128 << dev->pcie_mpss;
2383

2384 2385
		if (pci_pcie_type(dev) != PCI_EXP_TYPE_ROOT_PORT &&
		    dev->bus->self)
B
Bjorn Helgaas 已提交
2386 2387 2388

			/*
			 * For "Performance", the assumption is made that
2389 2390 2391 2392 2393
			 * downstream communication will never be larger than
			 * the MRRS.  So, the MPS only needs to be configured
			 * for the upstream communication.  This being the case,
			 * walk from the top down and set the MPS of the child
			 * to that of the parent bus.
2394 2395 2396 2397 2398
			 *
			 * Configure the device MPS with the smaller of the
			 * device MPSS or the bridge MPS (which is assumed to be
			 * properly configured at this point to the largest
			 * allowable MPS based on its parent bus).
2399
			 */
2400
			mps = min(mps, pcie_get_mps(dev->bus->self));
2401 2402 2403 2404
	}

	rc = pcie_set_mps(dev, mps);
	if (rc)
2405
		pci_err(dev, "Failed attempting to set the MPS\n");
2406 2407
}

2408
static void pcie_write_mrrs(struct pci_dev *dev)
2409
{
2410
	int rc, mrrs;
2411

B
Bjorn Helgaas 已提交
2412 2413
	/*
	 * In the "safe" case, do not configure the MRRS.  There appear to be
2414 2415 2416 2417 2418
	 * issues with setting MRRS to 0 on a number of devices.
	 */
	if (pcie_bus_config != PCIE_BUS_PERFORMANCE)
		return;

B
Bjorn Helgaas 已提交
2419 2420
	/*
	 * For max performance, the MRRS must be set to the largest supported
2421
	 * value.  However, it cannot be configured larger than the MPS the
2422
	 * device or the bus can support.  This should already be properly
B
Bjorn Helgaas 已提交
2423
	 * configured by a prior call to pcie_write_mps().
2424
	 */
2425
	mrrs = pcie_get_mps(dev);
2426

B
Bjorn Helgaas 已提交
2427 2428
	/*
	 * MRRS is a R/W register.  Invalid values can be written, but a
2429
	 * subsequent read will verify if the value is acceptable or not.
2430 2431
	 * If the MRRS value provided is not acceptable (e.g., too large),
	 * shrink the value until it is acceptable to the HW.
2432
	 */
2433 2434
	while (mrrs != pcie_get_readrq(dev) && mrrs >= 128) {
		rc = pcie_set_readrq(dev, mrrs);
2435 2436
		if (!rc)
			break;
2437

2438
		pci_warn(dev, "Failed attempting to set the MRRS\n");
2439 2440
		mrrs /= 2;
	}
2441 2442

	if (mrrs < 128)
2443
		pci_err(dev, "MRRS was unable to be configured with a safe value.  If problems are experienced, try running with pci=pcie_bus_safe\n");
2444 2445 2446 2447
}

static int pcie_bus_configure_set(struct pci_dev *dev, void *data)
{
J
Jon Mason 已提交
2448
	int mps, orig_mps;
2449 2450 2451 2452

	if (!pci_is_pcie(dev))
		return 0;

2453 2454
	if (pcie_bus_config == PCIE_BUS_TUNE_OFF ||
	    pcie_bus_config == PCIE_BUS_DEFAULT)
2455 2456
		return 0;

J
Jon Mason 已提交
2457 2458
	mps = 128 << *(u8 *)data;
	orig_mps = pcie_get_mps(dev);
2459 2460

	pcie_write_mps(dev, mps);
2461
	pcie_write_mrrs(dev);
2462

2463
	pci_info(dev, "Max Payload Size set to %4d/%4d (was %4d), Max Read Rq %4d\n",
2464
		 pcie_get_mps(dev), 128 << dev->pcie_mpss,
J
Jon Mason 已提交
2465
		 orig_mps, pcie_get_readrq(dev));
2466 2467 2468 2469

	return 0;
}

B
Bjorn Helgaas 已提交
2470 2471
/*
 * pcie_bus_configure_settings() requires that pci_walk_bus work in a top-down,
2472 2473 2474
 * parents then children fashion.  If this changes, then this code will not
 * work as designed.
 */
2475
void pcie_bus_configure_settings(struct pci_bus *bus)
2476
{
2477
	u8 smpss = 0;
2478

2479
	if (!bus->self)
2480 2481 2482
		return;

	if (!pci_is_pcie(bus->self))
2483 2484
		return;

B
Bjorn Helgaas 已提交
2485 2486
	/*
	 * FIXME - Peer to peer DMA is possible, though the endpoint would need
2487
	 * to be aware of the MPS of the destination.  To work around this,
2488 2489 2490 2491 2492
	 * simply force the MPS of the entire system to the smallest possible.
	 */
	if (pcie_bus_config == PCIE_BUS_PEER2PEER)
		smpss = 0;

2493
	if (pcie_bus_config == PCIE_BUS_SAFE) {
2494
		smpss = bus->self->pcie_mpss;
2495

2496 2497 2498 2499 2500 2501 2502
		pcie_find_smpss(bus->self, &smpss);
		pci_walk_bus(bus, pcie_find_smpss, &smpss);
	}

	pcie_bus_configure_set(bus->self, &smpss);
	pci_walk_bus(bus, pcie_bus_configure_set, &smpss);
}
2503
EXPORT_SYMBOL_GPL(pcie_bus_configure_settings);
2504

2505 2506 2507 2508 2509 2510 2511 2512 2513
/*
 * Called after each bus is probed, but before its children are examined.  This
 * is marked as __weak because multiple architectures define it.
 */
void __weak pcibios_fixup_bus(struct pci_bus *bus)
{
       /* nothing to do, expected to be removed in the future */
}

2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527
/**
 * pci_scan_child_bus_extend() - Scan devices below a bus
 * @bus: Bus to scan for devices
 * @available_buses: Total number of buses available (%0 does not try to
 *		     extend beyond the minimal)
 *
 * Scans devices below @bus including subordinate buses. Returns new
 * subordinate number including all the found devices. Passing
 * @available_buses causes the remaining bus space to be distributed
 * equally between hotplug-capable bridges to allow future extension of the
 * hierarchy.
 */
static unsigned int pci_scan_child_bus_extend(struct pci_bus *bus,
					      unsigned int available_buses)
L
Linus Torvalds 已提交
2528
{
2529 2530 2531
	unsigned int used_buses, normal_bridges = 0, hotplug_bridges = 0;
	unsigned int start = bus->busn_res.start;
	unsigned int devfn, cmax, max = start;
L
Linus Torvalds 已提交
2532 2533
	struct pci_dev *dev;

B
Bjorn Helgaas 已提交
2534
	dev_dbg(&bus->dev, "scanning bus\n");
L
Linus Torvalds 已提交
2535 2536 2537 2538 2539

	/* Go find them, Rover! */
	for (devfn = 0; devfn < 0x100; devfn += 8)
		pci_scan_slot(bus, devfn);

B
Bjorn Helgaas 已提交
2540
	/* Reserve buses for SR-IOV capability */
2541 2542
	used_buses = pci_iov_bus_range(bus);
	max += used_buses;
2543

L
Linus Torvalds 已提交
2544 2545 2546 2547
	/*
	 * After performing arch-dependent fixup of the bus, look behind
	 * all PCI-to-PCI bridges on this bus.
	 */
A
Alex Chiang 已提交
2548
	if (!bus->is_added) {
B
Bjorn Helgaas 已提交
2549
		dev_dbg(&bus->dev, "fixups for bus\n");
A
Alex Chiang 已提交
2550
		pcibios_fixup_bus(bus);
2551
		bus->is_added = 1;
A
Alex Chiang 已提交
2552 2553
	}

2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565
	/*
	 * Calculate how many hotplug bridges and normal bridges there
	 * are on this bus. We will distribute the additional available
	 * buses between hotplug bridges.
	 */
	for_each_pci_bridge(dev, bus) {
		if (dev->is_hotplug_bridge)
			hotplug_bridges++;
		else
			normal_bridges++;
	}

2566 2567 2568 2569 2570
	/*
	 * Scan bridges that are already configured. We don't touch them
	 * unless they are misconfigured (which will be done in the second
	 * scan below).
	 */
2571 2572 2573 2574 2575
	for_each_pci_bridge(dev, bus) {
		cmax = max;
		max = pci_scan_bridge_extend(bus, dev, max, 0, 0);
		used_buses += cmax - max;
	}
2576 2577

	/* Scan bridges that need to be reconfigured */
2578 2579 2580 2581
	for_each_pci_bridge(dev, bus) {
		unsigned int buses = 0;

		if (!hotplug_bridges && normal_bridges == 1) {
B
Bjorn Helgaas 已提交
2582

2583 2584 2585 2586 2587 2588 2589 2590
			/*
			 * There is only one bridge on the bus (upstream
			 * port) so it gets all available buses which it
			 * can then distribute to the possible hotplug
			 * bridges below.
			 */
			buses = available_buses;
		} else if (dev->is_hotplug_bridge) {
B
Bjorn Helgaas 已提交
2591

2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603
			/*
			 * Distribute the extra buses between hotplug
			 * bridges if any.
			 */
			buses = available_buses / hotplug_bridges;
			buses = min(buses, available_buses - used_buses);
		}

		cmax = max;
		max = pci_scan_bridge_extend(bus, dev, cmax, buses, 1);
		used_buses += max - cmax;
	}
L
Linus Torvalds 已提交
2604

2605 2606
	/*
	 * Make sure a hotplug bridge has at least the minimum requested
2607 2608
	 * number of buses but allow it to grow up to the maximum available
	 * bus number of there is room.
2609
	 */
2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622
	if (bus->self && bus->self->is_hotplug_bridge) {
		used_buses = max_t(unsigned int, available_buses,
				   pci_hotplug_bus_size - 1);
		if (max - start < used_buses) {
			max = start + used_buses;

			/* Do not allocate more buses than we have room left */
			if (max > bus->busn_res.end)
				max = bus->busn_res.end;

			dev_dbg(&bus->dev, "%pR extended by %#02x\n",
				&bus->busn_res, max - start);
		}
2623 2624
	}

L
Linus Torvalds 已提交
2625 2626 2627 2628 2629 2630 2631
	/*
	 * We've scanned the bus and so we know all about what's on
	 * the other side of any bridges that may be on this bus plus
	 * any devices.
	 *
	 * Return how far we've got finding sub-buses.
	 */
B
Bjorn Helgaas 已提交
2632
	dev_dbg(&bus->dev, "bus scan returning with max=%02x\n", max);
L
Linus Torvalds 已提交
2633 2634
	return max;
}
2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646

/**
 * pci_scan_child_bus() - Scan devices below a bus
 * @bus: Bus to scan for devices
 *
 * Scans devices below @bus including subordinate buses. Returns new
 * subordinate number including all the found devices.
 */
unsigned int pci_scan_child_bus(struct pci_bus *bus)
{
	return pci_scan_child_bus_extend(bus, 0);
}
2647
EXPORT_SYMBOL_GPL(pci_scan_child_bus);
L
Linus Torvalds 已提交
2648

2649
/**
B
Bjorn Helgaas 已提交
2650 2651
 * pcibios_root_bridge_prepare - Platform-specific host bridge setup
 * @bridge: Host bridge to set up
2652 2653 2654 2655 2656 2657 2658 2659 2660
 *
 * Default empty implementation.  Replace with an architecture-specific setup
 * routine, if necessary.
 */
int __weak pcibios_root_bridge_prepare(struct pci_host_bridge *bridge)
{
	return 0;
}

2661 2662 2663 2664 2665 2666 2667 2668
void __weak pcibios_add_bus(struct pci_bus *bus)
{
}

void __weak pcibios_remove_bus(struct pci_bus *bus)
{
}

2669 2670
struct pci_bus *pci_create_root_bus(struct device *parent, int bus,
		struct pci_ops *ops, void *sysdata, struct list_head *resources)
L
Linus Torvalds 已提交
2671
{
2672
	int error;
2673
	struct pci_host_bridge *bridge;
L
Linus Torvalds 已提交
2674

2675
	bridge = pci_alloc_host_bridge(0);
2676
	if (!bridge)
2677
		return NULL;
2678 2679

	bridge->dev.parent = parent;
2680

2681 2682 2683 2684
	list_splice_init(resources, &bridge->windows);
	bridge->sysdata = sysdata;
	bridge->busnr = bus;
	bridge->ops = ops;
2685

2686 2687 2688
	error = pci_register_host_bridge(bridge);
	if (error < 0)
		goto err_out;
2689

2690
	return bridge->bus;
L
Linus Torvalds 已提交
2691 2692

err_out:
2693
	kfree(bridge);
L
Linus Torvalds 已提交
2694 2695
	return NULL;
}
2696
EXPORT_SYMBOL_GPL(pci_create_root_bus);
2697

2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730
int pci_host_probe(struct pci_host_bridge *bridge)
{
	struct pci_bus *bus, *child;
	int ret;

	ret = pci_scan_root_bus_bridge(bridge);
	if (ret < 0) {
		dev_err(bridge->dev.parent, "Scanning root bridge failed");
		return ret;
	}

	bus = bridge->bus;

	/*
	 * We insert PCI resources into the iomem_resource and
	 * ioport_resource trees in either pci_bus_claim_resources()
	 * or pci_bus_assign_resources().
	 */
	if (pci_has_flag(PCI_PROBE_ONLY)) {
		pci_bus_claim_resources(bus);
	} else {
		pci_bus_size_bridges(bus);
		pci_bus_assign_resources(bus);

		list_for_each_entry(child, &bus->children, node)
			pcie_bus_configure_settings(child);
	}

	pci_bus_add_devices(bus);
	return 0;
}
EXPORT_SYMBOL_GPL(pci_host_probe);

2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746
int pci_bus_insert_busn_res(struct pci_bus *b, int bus, int bus_max)
{
	struct resource *res = &b->busn_res;
	struct resource *parent_res, *conflict;

	res->start = bus;
	res->end = bus_max;
	res->flags = IORESOURCE_BUS;

	if (!pci_is_root_bus(b))
		parent_res = &b->parent->busn_res;
	else {
		parent_res = get_pci_domain_busn_res(pci_domain_nr(b));
		res->flags |= IORESOURCE_PCI_FIXED;
	}

2747
	conflict = request_resource_conflict(parent_res, res);
2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793

	if (conflict)
		dev_printk(KERN_DEBUG, &b->dev,
			   "busn_res: can not insert %pR under %s%pR (conflicts with %s %pR)\n",
			    res, pci_is_root_bus(b) ? "domain " : "",
			    parent_res, conflict->name, conflict);

	return conflict == NULL;
}

int pci_bus_update_busn_res_end(struct pci_bus *b, int bus_max)
{
	struct resource *res = &b->busn_res;
	struct resource old_res = *res;
	resource_size_t size;
	int ret;

	if (res->start > bus_max)
		return -EINVAL;

	size = bus_max - res->start + 1;
	ret = adjust_resource(res, res->start, size);
	dev_printk(KERN_DEBUG, &b->dev,
			"busn_res: %pR end %s updated to %02x\n",
			&old_res, ret ? "can not be" : "is", bus_max);

	if (!ret && !res->parent)
		pci_bus_insert_busn_res(b, res->start, res->end);

	return ret;
}

void pci_bus_release_busn_res(struct pci_bus *b)
{
	struct resource *res = &b->busn_res;
	int ret;

	if (!res->flags || !res->parent)
		return;

	ret = release_resource(res);
	dev_printk(KERN_DEBUG, &b->dev,
			"busn_res: %pR %s released\n",
			res, ret ? "can not be" : "is");
}

2794
int pci_scan_root_bus_bridge(struct pci_host_bridge *bridge)
2795
{
2796
	struct resource_entry *window;
2797
	bool found = false;
2798
	struct pci_bus *b;
2799
	int max, bus, ret;
2800

2801 2802 2803 2804
	if (!bridge)
		return -EINVAL;

	resource_list_for_each_entry(window, &bridge->windows)
2805 2806 2807 2808
		if (window->res->flags & IORESOURCE_BUS) {
			found = true;
			break;
		}
2809

2810 2811 2812 2813 2814 2815
	ret = pci_register_host_bridge(bridge);
	if (ret < 0)
		return ret;

	b = bridge->bus;
	bus = bridge->busnr;
2816

2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828
	if (!found) {
		dev_info(&b->dev,
		 "No busn resource found for root bus, will use [bus %02x-ff]\n",
			bus);
		pci_bus_insert_busn_res(b, bus, 255);
	}

	max = pci_scan_child_bus(b);

	if (!found)
		pci_bus_update_busn_res_end(b, max);

2829
	return 0;
2830
}
2831
EXPORT_SYMBOL(pci_scan_root_bus_bridge);
2832 2833 2834 2835

struct pci_bus *pci_scan_root_bus(struct device *parent, int bus,
		struct pci_ops *ops, void *sysdata, struct list_head *resources)
{
2836
	struct resource_entry *window;
2837
	bool found = false;
2838
	struct pci_bus *b;
2839 2840
	int max;

2841
	resource_list_for_each_entry(window, resources)
2842 2843 2844 2845
		if (window->res->flags & IORESOURCE_BUS) {
			found = true;
			break;
		}
2846

2847
	b = pci_create_root_bus(parent, bus, ops, sysdata, resources);
2848 2849 2850
	if (!b)
		return NULL;

2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862
	if (!found) {
		dev_info(&b->dev,
		 "No busn resource found for root bus, will use [bus %02x-ff]\n",
			bus);
		pci_bus_insert_busn_res(b, bus, 255);
	}

	max = pci_scan_child_bus(b);

	if (!found)
		pci_bus_update_busn_res_end(b, max);

2863
	return b;
2864
}
2865 2866
EXPORT_SYMBOL(pci_scan_root_bus);

B
Bill Pemberton 已提交
2867
struct pci_bus *pci_scan_bus(int bus, struct pci_ops *ops,
2868 2869 2870 2871 2872 2873 2874
					void *sysdata)
{
	LIST_HEAD(resources);
	struct pci_bus *b;

	pci_add_resource(&resources, &ioport_resource);
	pci_add_resource(&resources, &iomem_resource);
2875
	pci_add_resource(&resources, &busn_resource);
2876 2877
	b = pci_create_root_bus(NULL, bus, ops, sysdata, &resources);
	if (b) {
2878
		pci_scan_child_bus(b);
2879 2880 2881 2882 2883 2884 2885
	} else {
		pci_free_resource_list(&resources);
	}
	return b;
}
EXPORT_SYMBOL(pci_scan_bus);

2886
/**
B
Bjorn Helgaas 已提交
2887
 * pci_rescan_bus_bridge_resize - Scan a PCI bus for devices
2888 2889 2890 2891 2892 2893 2894 2895 2896
 * @bridge: PCI bridge for the bus to scan
 *
 * Scan a PCI bus and child buses for new devices, add them,
 * and enable them, resizing bridge mmio/io resource if necessary
 * and possible.  The caller must ensure the child devices are already
 * removed for resizing to occur.
 *
 * Returns the max number of subordinate bus discovered.
 */
2897
unsigned int pci_rescan_bus_bridge_resize(struct pci_dev *bridge)
2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910
{
	unsigned int max;
	struct pci_bus *bus = bridge->subordinate;

	max = pci_scan_child_bus(bus);

	pci_assign_unassigned_bridge_resources(bridge);

	pci_bus_add_devices(bus);

	return max;
}

2911
/**
B
Bjorn Helgaas 已提交
2912
 * pci_rescan_bus - Scan a PCI bus for devices
2913 2914
 * @bus: PCI bus to scan
 *
B
Bjorn Helgaas 已提交
2915 2916
 * Scan a PCI bus and child buses for new devices, add them,
 * and enable them.
2917 2918 2919
 *
 * Returns the max number of subordinate bus discovered.
 */
2920
unsigned int pci_rescan_bus(struct pci_bus *bus)
2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931
{
	unsigned int max;

	max = pci_scan_child_bus(bus);
	pci_assign_unassigned_bus_resources(bus);
	pci_bus_add_devices(bus);

	return max;
}
EXPORT_SYMBOL_GPL(pci_rescan_bus);

2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949
/*
 * pci_rescan_bus(), pci_rescan_bus_bridge_resize() and PCI device removal
 * routines should always be executed under this mutex.
 */
static DEFINE_MUTEX(pci_rescan_remove_lock);

void pci_lock_rescan_remove(void)
{
	mutex_lock(&pci_rescan_remove_lock);
}
EXPORT_SYMBOL_GPL(pci_lock_rescan_remove);

void pci_unlock_rescan_remove(void)
{
	mutex_unlock(&pci_rescan_remove_lock);
}
EXPORT_SYMBOL_GPL(pci_unlock_rescan_remove);

R
Ryan Desfosses 已提交
2950 2951
static int __init pci_sort_bf_cmp(const struct device *d_a,
				  const struct device *d_b)
2952
{
2953 2954 2955
	const struct pci_dev *a = to_pci_dev(d_a);
	const struct pci_dev *b = to_pci_dev(d_b);

2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967
	if      (pci_domain_nr(a->bus) < pci_domain_nr(b->bus)) return -1;
	else if (pci_domain_nr(a->bus) > pci_domain_nr(b->bus)) return  1;

	if      (a->bus->number < b->bus->number) return -1;
	else if (a->bus->number > b->bus->number) return  1;

	if      (a->devfn < b->devfn) return -1;
	else if (a->devfn > b->devfn) return  1;

	return 0;
}

2968
void __init pci_sort_breadthfirst(void)
2969
{
2970
	bus_sort_breadthfirst(&pci_bus_type, &pci_sort_bf_cmp);
2971
}
2972 2973 2974 2975

int pci_hp_add_bridge(struct pci_dev *dev)
{
	struct pci_bus *parent = dev->bus;
2976
	int busnr, start = parent->busn_res.start;
2977
	unsigned int available_buses = 0;
2978 2979 2980 2981 2982 2983 2984
	int end = parent->busn_res.end;

	for (busnr = start; busnr <= end; busnr++) {
		if (!pci_find_bus(pci_domain_nr(parent), busnr))
			break;
	}
	if (busnr-- > end) {
2985
		pci_err(dev, "No bus number available for hot-added bridge\n");
2986 2987
		return -1;
	}
2988 2989 2990 2991

	/* Scan bridges that are already configured */
	busnr = pci_scan_bridge(parent, dev, busnr, 0);

2992 2993 2994 2995 2996 2997
	/*
	 * Distribute the available bus numbers between hotplug-capable
	 * bridges to make extending the chain later possible.
	 */
	available_buses = end - busnr;

2998
	/* Scan bridges that need to be reconfigured */
2999
	pci_scan_bridge_extend(parent, dev, busnr, available_buses, 1);
3000

3001 3002 3003 3004 3005 3006
	if (!dev->subordinate)
		return -1;

	return 0;
}
EXPORT_SYMBOL_GPL(pci_hp_add_bridge);