probe.c 82.3 KB
Newer Older
1
// SPDX-License-Identifier: GPL-2.0
L
Linus Torvalds 已提交
2
/*
B
Bjorn Helgaas 已提交
3
 * PCI detection and setup code
L
Linus Torvalds 已提交
4 5 6 7 8 9
 */

#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>
16
#include <linux/aer.h>
17
#include <linux/acpi.h>
18
#include <linux/hypervisor.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;

332 333 334 335
	/* Per PCIe r4.0, sec 9.3.4.1.11, the VF BARs are all RO Zero */
	if (dev->is_virtfn)
		return;

336 337
	for (pos = 0; pos < howmany; pos++) {
		struct resource *res = &dev->resource[pos];
L
Linus Torvalds 已提交
338
		reg = PCI_BASE_ADDRESS_0 + (pos << 2);
339
		pos += __pci_read_base(dev, pci_bar_unknown, res, reg);
L
Linus Torvalds 已提交
340
	}
341

L
Linus Torvalds 已提交
342
	if (rom) {
343
		struct resource *res = &dev->resource[PCI_ROM_RESOURCE];
L
Linus Torvalds 已提交
344
		dev->rom_base_reg = rom;
345
		res->flags = IORESOURCE_MEM | IORESOURCE_PREFETCH |
346
				IORESOURCE_READONLY | IORESOURCE_SIZEALIGN;
347
		__pci_read_base(dev, pci_bar_mem32, res, rom);
L
Linus Torvalds 已提交
348 349 350
	}
}

B
Bill Pemberton 已提交
351
static void pci_read_bridge_io(struct pci_bus *child)
L
Linus Torvalds 已提交
352 353 354
{
	struct pci_dev *dev = child->self;
	u8 io_base_lo, io_limit_lo;
355
	unsigned long io_mask, io_granularity, base, limit;
356
	struct pci_bus_region region;
357 358 359 360 361 362 363 364 365
	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
366 367 368 369

	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);
370 371
	base = (io_base_lo & io_mask) << 8;
	limit = (io_limit_lo & io_mask) << 8;
L
Linus Torvalds 已提交
372 373 374

	if ((io_base_lo & PCI_IO_RANGE_TYPE_MASK) == PCI_IO_RANGE_TYPE_32) {
		u16 io_base_hi, io_limit_hi;
375

L
Linus Torvalds 已提交
376 377
		pci_read_config_word(dev, PCI_IO_BASE_UPPER16, &io_base_hi);
		pci_read_config_word(dev, PCI_IO_LIMIT_UPPER16, &io_limit_hi);
378 379
		base |= ((unsigned long) io_base_hi << 16);
		limit |= ((unsigned long) io_limit_hi << 16);
L
Linus Torvalds 已提交
380 381
	}

382
	if (base <= limit) {
L
Linus Torvalds 已提交
383
		res->flags = (io_base_lo & PCI_IO_RANGE_TYPE_MASK) | IORESOURCE_IO;
384
		region.start = base;
385
		region.end = limit + io_granularity - 1;
386
		pcibios_bus_to_resource(dev->bus, res, &region);
387
		pci_printk(KERN_DEBUG, dev, "  bridge window %pR\n", res);
L
Linus Torvalds 已提交
388
	}
389 390
}

B
Bill Pemberton 已提交
391
static void pci_read_bridge_mmio(struct pci_bus *child)
392 393 394 395
{
	struct pci_dev *dev = child->self;
	u16 mem_base_lo, mem_limit_lo;
	unsigned long base, limit;
396
	struct pci_bus_region region;
397
	struct resource *res;
L
Linus Torvalds 已提交
398 399 400 401

	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);
402 403
	base = ((unsigned long) mem_base_lo & PCI_MEMORY_RANGE_MASK) << 16;
	limit = ((unsigned long) mem_limit_lo & PCI_MEMORY_RANGE_MASK) << 16;
404
	if (base <= limit) {
L
Linus Torvalds 已提交
405
		res->flags = (mem_base_lo & PCI_MEMORY_RANGE_TYPE_MASK) | IORESOURCE_MEM;
406 407
		region.start = base;
		region.end = limit + 0xfffff;
408
		pcibios_bus_to_resource(dev->bus, res, &region);
409
		pci_printk(KERN_DEBUG, dev, "  bridge window %pR\n", res);
L
Linus Torvalds 已提交
410
	}
411 412
}

B
Bill Pemberton 已提交
413
static void pci_read_bridge_mmio_pref(struct pci_bus *child)
414 415 416
{
	struct pci_dev *dev = child->self;
	u16 mem_base_lo, mem_limit_lo;
417
	u64 base64, limit64;
Y
Yinghai Lu 已提交
418
	pci_bus_addr_t base, limit;
419
	struct pci_bus_region region;
420
	struct resource *res;
L
Linus Torvalds 已提交
421 422 423 424

	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);
425 426
	base64 = (mem_base_lo & PCI_PREF_RANGE_MASK) << 16;
	limit64 = (mem_limit_lo & PCI_PREF_RANGE_MASK) << 16;
L
Linus Torvalds 已提交
427 428 429

	if ((mem_base_lo & PCI_PREF_RANGE_TYPE_MASK) == PCI_PREF_RANGE_TYPE_64) {
		u32 mem_base_hi, mem_limit_hi;
430

L
Linus Torvalds 已提交
431 432 433 434 435 436 437 438 439
		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) {
440 441
			base64 |= (u64) mem_base_hi << 32;
			limit64 |= (u64) mem_limit_hi << 32;
L
Linus Torvalds 已提交
442 443
		}
	}
444

Y
Yinghai Lu 已提交
445 446
	base = (pci_bus_addr_t) base64;
	limit = (pci_bus_addr_t) limit64;
447 448

	if (base != base64) {
449
		pci_err(dev, "can't handle bridge window above 4GB (bus address %#010llx)\n",
450 451 452 453
			(unsigned long long) base64);
		return;
	}

454
	if (base <= limit) {
455 456 457 458
		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;
459 460
		region.start = base;
		region.end = limit + 0xfffff;
461
		pcibios_bus_to_resource(dev->bus, res, &region);
462
		pci_printk(KERN_DEBUG, dev, "  bridge window %pR\n", res);
L
Linus Torvalds 已提交
463 464 465
	}
}

B
Bill Pemberton 已提交
466
void pci_read_bridge_bases(struct pci_bus *child)
467 468
{
	struct pci_dev *dev = child->self;
469
	struct resource *res;
470 471 472 473 474
	int i;

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

475
	pci_info(dev, "PCI bridge to %pR%s\n",
476
		 &child->busn_res,
477 478
		 dev->transparent ? " (subtractive decode)" : "");

479 480 481 482
	pci_bus_remove_resources(child);
	for (i = 0; i < PCI_BRIDGE_RESOURCE_NUM; i++)
		child->resource[i] = &dev->resource[PCI_BRIDGE_RESOURCES+i];

483 484 485
	pci_read_bridge_io(child);
	pci_read_bridge_mmio(child);
	pci_read_bridge_mmio_pref(child);
486 487

	if (dev->transparent) {
488
		pci_bus_for_each_resource(child->parent, res, i) {
489
			if (res && res->flags) {
490 491
				pci_bus_add_resource(child, res,
						     PCI_SUBTRACTIVE_DECODE);
492
				pci_printk(KERN_DEBUG, dev,
493
					   "  bridge window %pR (subtractive decode)\n",
494 495
					   res);
			}
496 497
		}
	}
498 499
}

500
static struct pci_bus *pci_alloc_bus(struct pci_bus *parent)
L
Linus Torvalds 已提交
501 502 503
{
	struct pci_bus *b;

504
	b = kzalloc(sizeof(*b), GFP_KERNEL);
505 506 507 508 509 510 511 512 513 514
	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;
515 516 517 518
#ifdef CONFIG_PCI_DOMAINS_GENERIC
	if (parent)
		b->domain_nr = parent->domain_nr;
#endif
L
Linus Torvalds 已提交
519 520 521
	return b;
}

522
static void devm_pci_release_host_bridge_dev(struct device *dev)
523 524 525 526 527
{
	struct pci_host_bridge *bridge = to_pci_host_bridge(dev);

	if (bridge->release_fn)
		bridge->release_fn(bridge);
528 529

	pci_free_resource_list(&bridge->windows);
530
}
531

532 533 534
static void pci_release_host_bridge_dev(struct device *dev)
{
	devm_pci_release_host_bridge_dev(dev);
535
	kfree(to_pci_host_bridge(dev));
536 537
}

538
struct pci_host_bridge *pci_alloc_host_bridge(size_t priv)
539 540 541
{
	struct pci_host_bridge *bridge;

542
	bridge = kzalloc(sizeof(*bridge) + priv, GFP_KERNEL);
543 544
	if (!bridge)
		return NULL;
545

546
	INIT_LIST_HEAD(&bridge->windows);
547
	bridge->dev.release = pci_release_host_bridge_dev;
548

549 550 551 552 553 554 555
	/*
	 * We assume we can manage these PCIe features.  Some systems may
	 * reserve these for use by the platform itself, e.g., an ACPI BIOS
	 * may implement its own AER handling and use _OSC to prevent the
	 * OS from interfering.
	 */
	bridge->native_aer = 1;
556
	bridge->native_pcie_hotplug = 1;
557
	bridge->native_shpc_hotplug = 1;
558
	bridge->native_pme = 1;
559
	bridge->native_ltr = 1;
560

561 562
	return bridge;
}
563
EXPORT_SYMBOL(pci_alloc_host_bridge);
564

565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580
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);

581 582 583 584 585 586 587 588
void pci_free_host_bridge(struct pci_host_bridge *bridge)
{
	pci_free_resource_list(&bridge->windows);

	kfree(bridge);
}
EXPORT_SYMBOL(pci_free_host_bridge);

589
static const unsigned char pcix_bus_speed[] = {
590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607
	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 */
};

608
const unsigned char pcie_link_speed[] = {
609 610 611
	PCI_SPEED_UNKNOWN,		/* 0 */
	PCIE_SPEED_2_5GT,		/* 1 */
	PCIE_SPEED_5_0GT,		/* 2 */
612
	PCIE_SPEED_8_0GT,		/* 3 */
613
	PCIE_SPEED_16_0GT,		/* 4 */
614 615 616 617 618 619 620 621 622 623 624 625 626 627 628
	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)
{
629
	bus->cur_bus_speed = pcie_link_speed[linksta & PCI_EXP_LNKSTA_CLS];
630 631 632
}
EXPORT_SYMBOL_GPL(pcie_update_link_speed);

633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652
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;
653

654 655 656 657 658 659 660 661 662 663
	if (agp3) {
		index += 2;
		if (index == 5)
			index = 0;
	}

 out:
	return agp_speeds[index];
}

664 665 666 667 668
static void pci_set_bus_speed(struct pci_bus *bus)
{
	struct pci_dev *bridge = bus->self;
	int pos;

669 670 671 672 673 674 675 676 677 678 679 680 681
	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);
	}

682 683 684 685 686
	pos = pci_find_capability(bridge, PCI_CAP_ID_PCIX);
	if (pos) {
		u16 status;
		enum pci_bus_speed max;

687 688 689 690
		pci_read_config_word(bridge, pos + PCI_X_BRIDGE_SSTATUS,
				     &status);

		if (status & PCI_X_SSTATUS_533MHZ) {
691
			max = PCI_SPEED_133MHz_PCIX_533;
692
		} else if (status & PCI_X_SSTATUS_266MHZ) {
693
			max = PCI_SPEED_133MHz_PCIX_266;
694
		} else if (status & PCI_X_SSTATUS_133MHZ) {
R
Ryan Desfosses 已提交
695
			if ((status & PCI_X_SSTATUS_VERS) == PCI_X_SSTATUS_V2)
696
				max = PCI_SPEED_133MHz_PCIX_ECC;
R
Ryan Desfosses 已提交
697
			else
698 699 700 701 702 703
				max = PCI_SPEED_133MHz_PCIX;
		} else {
			max = PCI_SPEED_66MHz_PCIX;
		}

		bus->max_bus_speed = max;
704 705
		bus->cur_bus_speed = pcix_bus_speed[
			(status & PCI_X_SSTATUS_FREQ) >> 6];
706 707 708 709

		return;
	}

710
	if (pci_is_pcie(bridge)) {
711 712 713
		u32 linkcap;
		u16 linksta;

714
		pcie_capability_read_dword(bridge, PCI_EXP_LNKCAP, &linkcap);
715
		bus->max_bus_speed = pcie_link_speed[linkcap & PCI_EXP_LNKCAP_SLS];
716
		bridge->link_active_reporting = !!(linkcap & PCI_EXP_LNKCAP_DLLLARC);
717

718
		pcie_capability_read_word(bridge, PCI_EXP_LNKSTA, &linksta);
719 720 721 722
		pcie_update_link_speed(bus, linksta);
	}
}

723 724
static struct irq_domain *pci_host_bridge_msi_domain(struct pci_bus *bus)
{
725 726
	struct irq_domain *d;

727 728 729 730
	/*
	 * Any firmware interface that can resolve the msi_domain
	 * should be called from here.
	 */
731
	d = pci_host_bridge_of_msi_domain(bus);
732 733
	if (!d)
		d = pci_host_bridge_acpi_msi_domain(bus);
734

735 736 737 738 739 740 741 742 743 744 745 746 747 748
#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

749
	return d;
750 751 752 753 754
}

static void pci_set_bus_msi_domain(struct pci_bus *bus)
{
	struct irq_domain *d;
755
	struct pci_bus *b;
756 757

	/*
758 759 760
	 * 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.
761
	 */
762 763 764 765 766 767 768
	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);
769 770 771 772

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

773
static int pci_register_host_bridge(struct pci_host_bridge *bridge)
774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790
{
	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
791
	/* Temporarily move resources off the list */
792 793 794 795 796 797 798 799 800 801 802
	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
803
		/* Ignore it if we already got here via a different bridge */
804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888
		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;
}

889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927
static bool pci_bridge_child_ext_cfg_accessible(struct pci_dev *bridge)
{
	int pos;
	u32 status;

	/*
	 * If extended config space isn't accessible on a bridge's primary
	 * bus, we certainly can't access it on the secondary bus.
	 */
	if (bridge->bus->bus_flags & PCI_BUS_FLAGS_NO_EXTCFG)
		return false;

	/*
	 * PCIe Root Ports and switch ports are PCIe on both sides, so if
	 * extended config space is accessible on the primary, it's also
	 * accessible on the secondary.
	 */
	if (pci_is_pcie(bridge) &&
	    (pci_pcie_type(bridge) == PCI_EXP_TYPE_ROOT_PORT ||
	     pci_pcie_type(bridge) == PCI_EXP_TYPE_UPSTREAM ||
	     pci_pcie_type(bridge) == PCI_EXP_TYPE_DOWNSTREAM))
		return true;

	/*
	 * For the other bridge types:
	 *   - PCI-to-PCI bridges
	 *   - PCIe-to-PCI/PCI-X forward bridges
	 *   - PCI/PCI-X-to-PCIe reverse bridges
	 * extended config space on the secondary side is only accessible
	 * if the bridge supports PCI-X Mode 2.
	 */
	pos = pci_find_capability(bridge, PCI_CAP_ID_PCIX);
	if (!pos)
		return false;

	pci_read_config_dword(bridge, pos + PCI_X_STATUS, &status);
	return status & (PCI_X_STATUS_266MHZ | PCI_X_STATUS_533MHZ);
}

928 929
static struct pci_bus *pci_alloc_child_bus(struct pci_bus *parent,
					   struct pci_dev *bridge, int busnr)
L
Linus Torvalds 已提交
930 931 932
{
	struct pci_bus *child;
	int i;
933
	int ret;
L
Linus Torvalds 已提交
934

B
Bjorn Helgaas 已提交
935
	/* Allocate a new bus and inherit stuff from the parent */
936
	child = pci_alloc_bus(parent);
L
Linus Torvalds 已提交
937 938 939 940 941
	if (!child)
		return NULL;

	child->parent = parent;
	child->ops = parent->ops;
942
	child->msi = parent->msi;
L
Linus Torvalds 已提交
943
	child->sysdata = parent->sysdata;
944
	child->bus_flags = parent->bus_flags;
L
Linus Torvalds 已提交
945

B
Bjorn Helgaas 已提交
946 947 948
	/*
	 * Initialize some portions of the bus device, but don't register
	 * it now as the parent is not properly set up yet.
949 950
	 */
	child->dev.class = &pcibus_class;
951
	dev_set_name(&child->dev, "%04x:%02x", pci_domain_nr(child), busnr);
L
Linus Torvalds 已提交
952

B
Bjorn Helgaas 已提交
953
	/* Set up the primary, secondary and subordinate bus numbers */
954 955 956
	child->number = child->busn_res.start = busnr;
	child->primary = parent->busn_res.start;
	child->busn_res.end = 0xff;
L
Linus Torvalds 已提交
957

958 959 960 961
	if (!bridge) {
		child->dev.parent = parent->bridge;
		goto add_dev;
	}
962 963 964

	child->self = bridge;
	child->bridge = get_device(&bridge->dev);
965
	child->dev.parent = child->bridge;
966
	pci_set_bus_of_node(child);
967 968
	pci_set_bus_speed(child);

969 970 971 972 973 974 975 976 977 978
	/*
	 * Check whether extended config space is accessible on the child
	 * bus.  Note that we currently assume it is always accessible on
	 * the root bus.
	 */
	if (!pci_bridge_child_ext_cfg_accessible(bridge)) {
		child->bus_flags |= PCI_BUS_FLAGS_NO_EXTCFG;
		pci_info(child, "extended config space not accessible\n");
	}

B
Bjorn Helgaas 已提交
979
	/* Set up default resource pointers and names */
980
	for (i = 0; i < PCI_BRIDGE_RESOURCE_NUM; i++) {
L
Linus Torvalds 已提交
981 982 983 984 985
		child->resource[i] = &bridge->resource[PCI_BRIDGE_RESOURCES+i];
		child->resource[i]->name = child->name;
	}
	bridge->subordinate = child;

986
add_dev:
987
	pci_set_bus_msi_domain(child);
988 989 990
	ret = device_register(&child->dev);
	WARN_ON(ret < 0);

991 992
	pcibios_add_bus(child);

993 994 995 996 997 998
	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);
	}

999 1000 1001
	/* Create legacy_io and legacy_mem files for this bus */
	pci_create_legacy_files(child);

L
Linus Torvalds 已提交
1002 1003 1004
	return child;
}

R
Ryan Desfosses 已提交
1005 1006
struct pci_bus *pci_add_new_bus(struct pci_bus *parent, struct pci_dev *dev,
				int busnr)
L
Linus Torvalds 已提交
1007 1008 1009 1010
{
	struct pci_bus *child;

	child = pci_alloc_child_bus(parent, dev, busnr);
1011
	if (child) {
1012
		down_write(&pci_bus_sem);
L
Linus Torvalds 已提交
1013
		list_add_tail(&child->node, &parent->children);
1014
		up_write(&pci_bus_sem);
1015
	}
L
Linus Torvalds 已提交
1016 1017
	return child;
}
1018
EXPORT_SYMBOL(pci_add_new_bus);
L
Linus Torvalds 已提交
1019

1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030
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);
}

1031 1032 1033
static unsigned int pci_scan_child_bus_extend(struct pci_bus *bus,
					      unsigned int available_buses);

L
Linus Torvalds 已提交
1034
/*
1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045
 * 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
1046 1047 1048 1049 1050 1051 1052 1053
 * 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.
1054 1055
 *
 * Return: New subordinate number covering all buses behind this bridge.
L
Linus Torvalds 已提交
1056
 */
1057 1058 1059
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
1060 1061 1062
{
	struct pci_bus *child;
	int is_cardbus = (dev->hdr_type == PCI_HEADER_TYPE_CARDBUS);
1063
	u32 buses, i, j = 0;
L
Linus Torvalds 已提交
1064
	u16 bctl;
1065
	u8 primary, secondary, subordinate;
1066
	int broken = 0;
L
Linus Torvalds 已提交
1067

1068 1069 1070 1071 1072 1073
	/*
	 * 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
1074
	pci_read_config_dword(dev, PCI_PRIMARY_BUS, &buses);
1075 1076 1077
	primary = buses & 0xFF;
	secondary = (buses >> 8) & 0xFF;
	subordinate = (buses >> 16) & 0xFF;
L
Linus Torvalds 已提交
1078

1079
	pci_dbg(dev, "scanning [bus %02x-%02x] behind bridge, pass %d\n",
1080
		secondary, subordinate, pass);
L
Linus Torvalds 已提交
1081

1082
	if (!primary && (primary != bus->number) && secondary && subordinate) {
1083
		pci_warn(dev, "Primary bus is hard wired to 0\n");
1084 1085 1086
		primary = bus->number;
	}

1087 1088
	/* Check if setup is sensible at all */
	if (!pass &&
1089
	    (primary != bus->number || secondary <= bus->number ||
1090
	     secondary > subordinate)) {
1091
		pci_info(dev, "bridge configuration invalid ([bus %02x-%02x]), reconfiguring\n",
1092
			 secondary, subordinate);
1093 1094 1095
		broken = 1;
	}

B
Bjorn Helgaas 已提交
1096 1097 1098 1099
	/*
	 * Disable Master-Abort Mode during probing to avoid reporting of
	 * bus errors in some architectures.
	 */
L
Linus Torvalds 已提交
1100 1101 1102 1103
	pci_read_config_word(dev, PCI_BRIDGE_CONTROL, &bctl);
	pci_write_config_word(dev, PCI_BRIDGE_CONTROL,
			      bctl & ~PCI_BRIDGE_CTL_MASTER_ABORT);

1104 1105
	pci_enable_crs(dev);

1106 1107 1108
	if ((secondary || subordinate) && !pcibios_assign_all_busses() &&
	    !is_cardbus && !broken) {
		unsigned int cmax;
B
Bjorn Helgaas 已提交
1109

L
Linus Torvalds 已提交
1110
		/*
B
Bjorn Helgaas 已提交
1111 1112
		 * Bus already configured by firmware, process it in the
		 * first pass and just note the configuration.
L
Linus Torvalds 已提交
1113 1114
		 */
		if (pass)
1115
			goto out;
L
Linus Torvalds 已提交
1116 1117

		/*
B
Bjorn Helgaas 已提交
1118 1119 1120 1121
		 * 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
1122
		 */
1123
		child = pci_find_bus(pci_domain_nr(bus), secondary);
A
Alex Chiang 已提交
1124
		if (!child) {
1125
			child = pci_add_new_bus(bus, dev, secondary);
A
Alex Chiang 已提交
1126 1127
			if (!child)
				goto out;
1128
			child->primary = primary;
Y
Yinghai Lu 已提交
1129
			pci_bus_insert_busn_res(child, secondary, subordinate);
A
Alex Chiang 已提交
1130
			child->bridge_ctl = bctl;
L
Linus Torvalds 已提交
1131 1132 1133
		}

		cmax = pci_scan_child_bus(child);
1134
		if (cmax > subordinate)
1135
			pci_warn(dev, "bridge has subordinate %02x but max busn %02x\n",
1136
				 subordinate, cmax);
B
Bjorn Helgaas 已提交
1137 1138

		/* Subordinate should equal child->busn_res.end */
1139 1140
		if (subordinate > max)
			max = subordinate;
L
Linus Torvalds 已提交
1141
	} else {
B
Bjorn Helgaas 已提交
1142

L
Linus Torvalds 已提交
1143 1144 1145 1146
		/*
		 * We need to assign a number to this bus which we always
		 * do in the second pass.
		 */
1147
		if (!pass) {
1148
			if (pcibios_assign_all_busses() || broken || is_cardbus)
B
Bjorn Helgaas 已提交
1149 1150 1151 1152 1153 1154 1155 1156 1157

				/*
				 * 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.
				 */
1158 1159
				pci_write_config_dword(dev, PCI_PRIMARY_BUS,
						       buses & ~0xffffff);
1160
			goto out;
1161
		}
L
Linus Torvalds 已提交
1162 1163 1164 1165

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

B
Bjorn Helgaas 已提交
1166 1167 1168 1169 1170
		/*
		 * 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.
		 */
1171 1172
		child = pci_find_bus(pci_domain_nr(bus), max+1);
		if (!child) {
1173
			child = pci_add_new_bus(bus, dev, max+1);
1174 1175
			if (!child)
				goto out;
1176 1177
			pci_bus_insert_busn_res(child, max+1,
						bus->busn_res.end);
1178
		}
1179
		max++;
1180 1181 1182
		if (available_buses)
			available_buses--;

L
Linus Torvalds 已提交
1183 1184
		buses = (buses & 0xff000000)
		      | ((unsigned int)(child->primary)     <<  0)
1185 1186
		      | ((unsigned int)(child->busn_res.start)   <<  8)
		      | ((unsigned int)(child->busn_res.end) << 16);
L
Linus Torvalds 已提交
1187 1188 1189 1190 1191 1192 1193 1194 1195

		/*
		 * yenta.c forces a secondary latency timer of 176.
		 * Copy that behaviour here.
		 */
		if (is_cardbus) {
			buses &= ~0xff000000;
			buses |= CARDBUS_LATENCY_TIMER << 24;
		}
1196

B
Bjorn Helgaas 已提交
1197
		/* We need to blast all three values with a single write */
L
Linus Torvalds 已提交
1198 1199 1200
		pci_write_config_dword(dev, PCI_PRIMARY_BUS, buses);

		if (!is_cardbus) {
1201
			child->bridge_ctl = bctl;
1202
			max = pci_scan_child_bus_extend(child, available_buses);
L
Linus Torvalds 已提交
1203
		} else {
B
Bjorn Helgaas 已提交
1204

L
Linus Torvalds 已提交
1205
			/*
B
Bjorn Helgaas 已提交
1206 1207 1208
			 * 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
1209
			 */
R
Ryan Desfosses 已提交
1210
			for (i = 0; i < CARDBUS_RESERVE_BUSNR; i++) {
1211
				struct pci_bus *parent = bus;
1212 1213 1214
				if (pci_find_bus(pci_domain_nr(bus),
							max+i+1))
					break;
1215 1216
				while (parent->parent) {
					if ((!pcibios_assign_all_busses()) &&
1217 1218
					    (parent->busn_res.end > max) &&
					    (parent->busn_res.end <= max+i)) {
1219 1220 1221 1222 1223
						j = 1;
					}
					parent = parent->parent;
				}
				if (j) {
B
Bjorn Helgaas 已提交
1224

1225
					/*
B
Bjorn Helgaas 已提交
1226 1227 1228
					 * Often, there are two CardBus
					 * bridges -- try to leave one
					 * valid bus number for each one.
1229 1230 1231 1232 1233
					 */
					i /= 2;
					break;
				}
			}
1234
			max += i;
L
Linus Torvalds 已提交
1235
		}
B
Bjorn Helgaas 已提交
1236 1237

		/* Set subordinate bus number to its real value */
Y
Yinghai Lu 已提交
1238
		pci_bus_update_busn_res_end(child, max);
L
Linus Torvalds 已提交
1239 1240 1241
		pci_write_config_byte(dev, PCI_SUBORDINATE_BUS, max);
	}

1242 1243 1244
	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
1245

1246
	/* Check that all devices are accessible */
1247
	while (bus->parent) {
1248 1249
		if ((child->busn_res.end > bus->busn_res.end) ||
		    (child->number > bus->busn_res.end) ||
1250
		    (child->number < bus->number) ||
1251
		    (child->busn_res.end < bus->number)) {
1252 1253 1254
			dev_info(&dev->dev, "devices behind bridge are unusable because %pR cannot be assigned for them\n",
				 &child->busn_res);
			break;
1255 1256 1257 1258
		}
		bus = bus->parent;
	}

1259 1260 1261
out:
	pci_write_config_word(dev, PCI_BRIDGE_CONTROL, bctl);

1262 1263
	pm_runtime_put(&dev->dev);

L
Linus Torvalds 已提交
1264 1265
	return max;
}
1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282

/*
 * 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.
1283 1284
 *
 * Return: New subordinate number covering all buses behind this bridge.
1285 1286 1287 1288 1289
 */
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);
}
1290
EXPORT_SYMBOL(pci_scan_bridge);
L
Linus Torvalds 已提交
1291 1292 1293 1294 1295 1296 1297 1298 1299

/*
 * 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;

1300 1301 1302 1303 1304 1305 1306
	/* 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
1307
	pci_read_config_byte(dev, PCI_INTERRUPT_PIN, &irq);
1308
	dev->pin = irq;
L
Linus Torvalds 已提交
1309 1310 1311 1312 1313
	if (irq)
		pci_read_config_byte(dev, PCI_INTERRUPT_LINE, &irq);
	dev->irq = irq;
}

1314
void set_pcie_port_type(struct pci_dev *pdev)
Y
Yu Zhao 已提交
1315 1316 1317
{
	int pos;
	u16 reg16;
1318 1319
	int type;
	struct pci_dev *parent;
Y
Yu Zhao 已提交
1320 1321 1322 1323

	pos = pci_find_capability(pdev, PCI_CAP_ID_EXP);
	if (!pos)
		return;
1324

1325
	pdev->pcie_cap = pos;
Y
Yu Zhao 已提交
1326
	pci_read_config_word(pdev, pos + PCI_EXP_FLAGS, &reg16);
1327
	pdev->pcie_flags_reg = reg16;
1328 1329
	pci_read_config_word(pdev, pos + PCI_EXP_DEVCAP, &reg16);
	pdev->pcie_mpss = reg16 & PCI_EXP_DEVCAP_PAYLOAD;
1330 1331

	/*
1332 1333 1334 1335
	 * 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.
1336 1337
	 */
	type = pci_pcie_type(pdev);
1338 1339
	if (type == PCI_EXP_TYPE_ROOT_PORT ||
	    type == PCI_EXP_TYPE_PCIE_BRIDGE)
1340 1341 1342 1343
		pdev->has_secondary_link = 1;
	else if (type == PCI_EXP_TYPE_UPSTREAM ||
		 type == PCI_EXP_TYPE_DOWNSTREAM) {
		parent = pci_upstream_bridge(pdev);
1344 1345 1346 1347 1348 1349

		/*
		 * 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)
1350 1351
			pdev->has_secondary_link = 1;
	}
Y
Yu Zhao 已提交
1352 1353
}

1354
void set_pcie_hotplug_bridge(struct pci_dev *pdev)
1355 1356 1357
{
	u32 reg32;

1358
	pcie_capability_read_dword(pdev, PCI_EXP_SLTCAP, &reg32);
1359 1360 1361 1362
	if (reg32 & PCI_EXP_SLTCAP_HPC)
		pdev->is_hotplug_bridge = 1;
}

1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380
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;
		}
	}
}

1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393
static void set_pcie_untrusted(struct pci_dev *dev)
{
	struct pci_dev *parent;

	/*
	 * If the upstream bridge is untrusted we treat this device
	 * untrusted as well.
	 */
	parent = pci_upstream_bridge(dev);
	if (parent && parent->untrusted)
		dev->untrusted = true;
}

1394
/**
B
Bjorn Helgaas 已提交
1395
 * pci_ext_cfg_is_aliased - Is ext config space just an alias of std config?
1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430
 * @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
}

1431
/**
B
Bjorn Helgaas 已提交
1432
 * pci_cfg_space_size - Get the configuration space size of the PCI device
1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447
 * @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)
1448
		return PCI_CFG_SPACE_SIZE;
1449
	if (status == 0xffffffff || pci_ext_cfg_is_aliased(dev))
1450
		return PCI_CFG_SPACE_SIZE;
1451 1452 1453 1454

	return PCI_CFG_SPACE_EXP_SIZE;
}

1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465
#ifdef CONFIG_PCI_IOV
static bool is_vf0(struct pci_dev *dev)
{
	if (pci_iov_virtfn_devfn(dev->physfn, 0) == dev->devfn &&
	    pci_iov_virtfn_bus(dev->physfn, 0) == dev->bus->number)
		return true;

	return false;
}
#endif

1466 1467 1468 1469 1470 1471
int pci_cfg_space_size(struct pci_dev *dev)
{
	int pos;
	u32 status;
	u16 class;

1472 1473 1474 1475 1476 1477
#ifdef CONFIG_PCI_IOV
	/* Read cached value for all VFs except for VF0 */
	if (dev->is_virtfn && !is_vf0(dev))
		return dev->physfn->sriov->cfg_size;
#endif

1478 1479 1480
	if (dev->bus->bus_flags & PCI_BUS_FLAGS_NO_EXTCFG)
		return PCI_CFG_SPACE_SIZE;

1481 1482 1483 1484
	class = dev->class >> 8;
	if (class == PCI_CLASS_BRIDGE_HOST)
		return pci_cfg_space_size_ext(dev);

1485 1486
	if (pci_is_pcie(dev))
		return pci_cfg_space_size_ext(dev);
1487

1488 1489 1490
	pos = pci_find_capability(dev, PCI_CAP_ID_PCIX);
	if (!pos)
		return PCI_CFG_SPACE_SIZE;
1491

1492 1493 1494
	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);
1495 1496 1497 1498

	return PCI_CFG_SPACE_SIZE;
}

1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535
static u32 pci_class(struct pci_dev *dev)
{
	u32 class;

#ifdef CONFIG_PCI_IOV
	if (dev->is_virtfn)
		return dev->physfn->sriov->class;
#endif
	pci_read_config_dword(dev, PCI_CLASS_REVISION, &class);
	return class;
}

static void pci_subsystem_ids(struct pci_dev *dev, u16 *vendor, u16 *device)
{
#ifdef CONFIG_PCI_IOV
	if (dev->is_virtfn) {
		*vendor = dev->physfn->sriov->subsystem_vendor;
		*device = dev->physfn->sriov->subsystem_device;
		return;
	}
#endif
	pci_read_config_word(dev, PCI_SUBSYSTEM_VENDOR_ID, vendor);
	pci_read_config_word(dev, PCI_SUBSYSTEM_ID, device);
}

static u8 pci_hdr_type(struct pci_dev *dev)
{
	u8 hdr_type;

#ifdef CONFIG_PCI_IOV
	if (dev->is_virtfn)
		return dev->physfn->sriov->hdr_type;
#endif
	pci_read_config_byte(dev, PCI_HEADER_TYPE, &hdr_type);
	return hdr_type;
}

1536
#define LEGACY_IO_RESOURCE	(IORESOURCE_IO | IORESOURCE_PCI_FIXED)
1537

1538
static void pci_msi_setup_pci_dev(struct pci_dev *dev)
1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553
{
	/*
	 * 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);
}

1554
/**
B
Bjorn Helgaas 已提交
1555
 * pci_intx_mask_broken - Test PCI_COMMAND_INTX_DISABLE writability
1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581
 * @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;
}

1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595
static void early_dump_pci_device(struct pci_dev *pdev)
{
	u32 value[256 / 4];
	int i;

	pci_info(pdev, "config space:\n");

	for (i = 0; i < 256; i += 4)
		pci_read_config_dword(pdev, i, &value[i / 4]);

	print_hex_dump(KERN_INFO, "", DUMP_PREFIX_OFFSET, 16, 1,
		       value, 256, false);
}

L
Linus Torvalds 已提交
1596
/**
B
Bjorn Helgaas 已提交
1597
 * pci_setup_device - Fill in class and map information of a device
L
Linus Torvalds 已提交
1598 1599
 * @dev: the device structure to fill
 *
1600
 * Initialize the device structure with information about the device's
B
Bjorn Helgaas 已提交
1601
 * vendor,class,memory and IO-space addresses, IRQ lines etc.
L
Linus Torvalds 已提交
1602
 * Called at initialisation of the PCI subsystem and by CardBus services.
Y
Yu Zhao 已提交
1603 1604
 * Returns 0 on success and negative if unknown type of device (not normal,
 * bridge or CardBus).
L
Linus Torvalds 已提交
1605
 */
Y
Yu Zhao 已提交
1606
int pci_setup_device(struct pci_dev *dev)
L
Linus Torvalds 已提交
1607 1608
{
	u32 class;
1609
	u16 cmd;
Y
Yu Zhao 已提交
1610
	u8 hdr_type;
1611
	int pos = 0;
1612 1613
	struct pci_bus_region region;
	struct resource *res;
Y
Yu Zhao 已提交
1614

1615
	hdr_type = pci_hdr_type(dev);
Y
Yu Zhao 已提交
1616 1617 1618 1619 1620 1621 1622 1623 1624

	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);

1625
	pci_dev_assign_slot(dev);
B
Bjorn Helgaas 已提交
1626 1627 1628 1629 1630

	/*
	 * 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
1631
	dev->dma_mask = 0xffffffff;
L
Linus Torvalds 已提交
1632

1633 1634 1635
	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
1636

1637 1638
	class = pci_class(dev);

A
Auke Kok 已提交
1639
	dev->revision = class & 0xff;
Y
Yinghai Lu 已提交
1640
	dev->class = class >> 8;		    /* upper 3 bytes */
L
Linus Torvalds 已提交
1641

1642
	pci_printk(KERN_DEBUG, dev, "[%04x:%04x] type %02x class %#08x\n",
Y
Yinghai Lu 已提交
1643
		   dev->vendor, dev->device, dev->hdr_type, dev->class);
L
Linus Torvalds 已提交
1644

1645 1646 1647
	if (pci_early_dump)
		early_dump_pci_device(dev);

B
Bjorn Helgaas 已提交
1648
	/* Need to have dev->class ready */
1649 1650
	dev->cfg_size = pci_cfg_space_size(dev);

B
Bjorn Helgaas 已提交
1651
	/* Need to have dev->cfg_size ready */
1652 1653
	set_pcie_thunderbolt(dev);

1654 1655
	set_pcie_untrusted(dev);

L
Linus Torvalds 已提交
1656
	/* "Unknown power state" */
1657
	dev->current_state = PCI_UNKNOWN;
L
Linus Torvalds 已提交
1658 1659 1660

	/* Early fixups, before probing the BARs */
	pci_fixup_device(pci_fixup_early, dev);
B
Bjorn Helgaas 已提交
1661 1662

	/* Device class may be changed after fixup */
1663
	class = dev->class >> 8;
L
Linus Torvalds 已提交
1664

1665 1666 1667
	if (dev->non_compliant_bars) {
		pci_read_config_word(dev, PCI_COMMAND, &cmd);
		if (cmd & (PCI_COMMAND_IO | PCI_COMMAND_MEMORY)) {
1668
			pci_info(dev, "device has non-compliant BARs; disabling IO/MEM decoding\n");
1669 1670 1671 1672 1673 1674
			cmd &= ~PCI_COMMAND_IO;
			cmd &= ~PCI_COMMAND_MEMORY;
			pci_write_config_word(dev, PCI_COMMAND, cmd);
		}
	}

1675 1676
	dev->broken_intx_masking = pci_intx_mask_broken(dev);

L
Linus Torvalds 已提交
1677 1678 1679 1680 1681 1682
	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);
1683 1684

		pci_subsystem_ids(dev, &dev->subsystem_vendor, &dev->subsystem_device);
1685 1686

		/*
1687 1688 1689 1690
		 * 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!
1691 1692 1693 1694 1695
		 */
		if (class == PCI_CLASS_STORAGE_IDE) {
			u8 progif;
			pci_read_config_byte(dev, PCI_CLASS_PROG, &progif);
			if ((progif & 1) == 0) {
1696 1697 1698 1699
				region.start = 0x1F0;
				region.end = 0x1F7;
				res = &dev->resource[0];
				res->flags = LEGACY_IO_RESOURCE;
1700
				pcibios_bus_to_resource(dev->bus, res, &region);
1701
				pci_info(dev, "legacy IDE quirk: reg 0x10: %pR\n",
1702
					 res);
1703 1704 1705 1706
				region.start = 0x3F6;
				region.end = 0x3F6;
				res = &dev->resource[1];
				res->flags = LEGACY_IO_RESOURCE;
1707
				pcibios_bus_to_resource(dev->bus, res, &region);
1708
				pci_info(dev, "legacy IDE quirk: reg 0x14: %pR\n",
1709
					 res);
1710 1711
			}
			if ((progif & 4) == 0) {
1712 1713 1714 1715
				region.start = 0x170;
				region.end = 0x177;
				res = &dev->resource[2];
				res->flags = LEGACY_IO_RESOURCE;
1716
				pcibios_bus_to_resource(dev->bus, res, &region);
1717
				pci_info(dev, "legacy IDE quirk: reg 0x18: %pR\n",
1718
					 res);
1719 1720 1721 1722
				region.start = 0x376;
				region.end = 0x376;
				res = &dev->resource[3];
				res->flags = LEGACY_IO_RESOURCE;
1723
				pcibios_bus_to_resource(dev->bus, res, &region);
1724
				pci_info(dev, "legacy IDE quirk: reg 0x1c: %pR\n",
1725
					 res);
1726 1727
			}
		}
L
Linus Torvalds 已提交
1728 1729 1730 1731 1732
		break;

	case PCI_HEADER_TYPE_BRIDGE:		    /* bridge header */
		if (class != PCI_CLASS_BRIDGE_PCI)
			goto bad;
B
Bjorn Helgaas 已提交
1733 1734 1735 1736 1737 1738

		/*
		 * The PCI-to-PCI bridge spec requires that subtractive
		 * decoding (i.e. transparent) bridge must have programming
		 * interface code of 0x01.
		 */
1739
		pci_read_irq(dev);
L
Linus Torvalds 已提交
1740 1741
		dev->transparent = ((dev->class & 0xff) == 1);
		pci_read_bases(dev, 2, PCI_ROM_ADDRESS1);
1742
		set_pcie_hotplug_bridge(dev);
1743 1744 1745 1746 1747
		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
1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759
		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 */
1760
		pci_err(dev, "unknown header type %02x, ignoring device\n",
1761
			dev->hdr_type);
Y
Yu Zhao 已提交
1762
		return -EIO;
L
Linus Torvalds 已提交
1763 1764

	bad:
1765
		pci_err(dev, "ignoring class %#08x (doesn't match header type %02x)\n",
1766
			dev->class, dev->hdr_type);
1767
		dev->class = PCI_CLASS_NOT_DEFINED << 8;
L
Linus Torvalds 已提交
1768 1769 1770 1771 1772 1773
	}

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

1774 1775 1776
static void pci_configure_mps(struct pci_dev *dev)
{
	struct pci_dev *bridge = pci_upstream_bridge(dev);
1777
	int mps, mpss, p_mps, rc;
1778 1779 1780 1781

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

1782 1783 1784 1785
	/* MPS and MRRS fields are of type 'RsvdP' for VFs, short-circuit out */
	if (dev->is_virtfn)
		return;

1786 1787 1788 1789 1790 1791 1792
	mps = pcie_get_mps(dev);
	p_mps = pcie_get_mps(bridge);

	if (mps == p_mps)
		return;

	if (pcie_bus_config == PCIE_BUS_TUNE_OFF) {
1793
		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",
1794 1795 1796
			 mps, pci_name(bridge), p_mps);
		return;
	}
1797 1798 1799 1800 1801 1802 1803 1804

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

1805 1806 1807 1808 1809 1810 1811 1812
	mpss = 128 << dev->pcie_mpss;
	if (mpss < p_mps && pci_pcie_type(bridge) == PCI_EXP_TYPE_ROOT_PORT) {
		pcie_set_mps(bridge, mpss);
		pci_info(dev, "Upstream bridge's Max Payload Size set to %d (was %d, max %d)\n",
			 mpss, p_mps, 128 << bridge->pcie_mpss);
		p_mps = pcie_get_mps(bridge);
	}

1813 1814
	rc = pcie_set_mps(dev, p_mps);
	if (rc) {
1815
		pci_warn(dev, "can't set Max Payload Size to %d; if necessary, use \"pci=pcie_bus_safe\" and report a bug\n",
1816 1817 1818 1819
			 p_mps);
		return;
	}

1820
	pci_info(dev, "Max Payload Size set to %d (was %d, max %d)\n",
1821
		 p_mps, mps, mpss);
1822 1823
}

1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835
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;

1836
	if (!hpp)
1837 1838 1839
		hpp = &pci_default_type0;

	if (hpp->revision > 1) {
1840
		pci_warn(dev, "PCI settings rev %d not supported; using defaults\n",
1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868
			 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)
{
1869 1870 1871 1872 1873 1874 1875 1876 1877
	int pos;

	if (!hpp)
		return;

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

1878
	pci_warn(dev, "PCI-X settings not supported\n");
1879 1880
}

1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895
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;
}

1896 1897 1898 1899 1900 1901 1902 1903
static void program_hpp_type2(struct pci_dev *dev, struct hpp_type2 *hpp)
{
	int pos;
	u32 reg32;

	if (!hpp)
		return;

1904 1905 1906
	if (!pci_is_pcie(dev))
		return;

1907
	if (hpp->revision > 1) {
1908
		pci_warn(dev, "PCIe settings rev %d not supported\n",
1909 1910 1911 1912
			 hpp->revision);
		return;
	}

1913 1914 1915 1916 1917 1918 1919 1920 1921 1922
	/*
	 * 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);

1923 1924 1925 1926 1927
	/* 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 */
1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938
	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;

1939 1940
		pcie_capability_clear_and_set_word(dev, PCI_EXP_LNKCTL,
			~hpp->pci_exp_lnkctl_and, hpp->pci_exp_lnkctl_or);
1941
	}
1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965

	/* 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
1966

1967 1968 1969 1970 1971
	/* 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;
1972 1973 1974 1975 1976 1977 1978 1979 1980 1981
	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
	 */
}

1982
int pci_configure_extended_tags(struct pci_dev *dev, void *ign)
1983
{
1984 1985 1986
	struct pci_host_bridge *host;
	u32 cap;
	u16 ctl;
1987 1988 1989
	int ret;

	if (!pci_is_pcie(dev))
1990
		return 0;
1991

1992
	ret = pcie_capability_read_dword(dev, PCI_EXP_DEVCAP, &cap);
1993
	if (ret)
1994 1995 1996 1997
		return 0;

	if (!(cap & PCI_EXP_DEVCAP_EXT_TAG))
		return 0;
1998

1999 2000 2001 2002 2003 2004 2005
	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;
2006

2007 2008 2009 2010 2011 2012
	/*
	 * 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) {
2013
			pci_info(dev, "disabling Extended Tags\n");
2014 2015 2016 2017 2018 2019 2020
			pcie_capability_clear_word(dev, PCI_EXP_DEVCTL,
						   PCI_EXP_DEVCTL_EXT_TAG);
		}
		return 0;
	}

	if (!(ctl & PCI_EXP_DEVCTL_EXT_TAG)) {
2021
		pci_info(dev, "enabling Extended Tags\n");
2022 2023
		pcie_capability_set_word(dev, PCI_EXP_DEVCTL,
					 PCI_EXP_DEVCTL_EXT_TAG);
2024 2025
	}
	return 0;
2026 2027
}

2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065
/**
 * 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);
2066
		pci_info(dev, "Relaxed Ordering disabled because the Root Port didn't support it\n");
2067 2068 2069
	}
}

2070 2071 2072
static void pci_configure_ltr(struct pci_dev *dev)
{
#ifdef CONFIG_PCIEASPM
2073
	struct pci_host_bridge *host = pci_find_host_bridge(dev->bus);
2074 2075 2076
	u32 cap;
	struct pci_dev *bridge;

2077 2078 2079
	if (!host->native_ltr)
		return;

2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105
	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
}

2106 2107 2108 2109
static void pci_configure_eetlp_prefix(struct pci_dev *dev)
{
#ifdef CONFIG_PCI_PASID
	struct pci_dev *bridge;
2110
	int pcie_type;
2111 2112 2113 2114 2115 2116 2117 2118 2119
	u32 cap;

	if (!pci_is_pcie(dev))
		return;

	pcie_capability_read_dword(dev, PCI_EXP_DEVCAP2, &cap);
	if (!(cap & PCI_EXP_DEVCAP2_EE_PREFIX))
		return;

2120 2121 2122
	pcie_type = pci_pcie_type(dev);
	if (pcie_type == PCI_EXP_TYPE_ROOT_PORT ||
	    pcie_type == PCI_EXP_TYPE_RC_END)
2123 2124 2125 2126 2127 2128 2129 2130 2131
		dev->eetlp_prefix_path = 1;
	else {
		bridge = pci_upstream_bridge(dev);
		if (bridge && bridge->eetlp_prefix_path)
			dev->eetlp_prefix_path = 1;
	}
#endif
}

2132 2133 2134 2135 2136
static void pci_configure_device(struct pci_dev *dev)
{
	struct hotplug_params hpp;
	int ret;

2137
	pci_configure_mps(dev);
2138
	pci_configure_extended_tags(dev, NULL);
2139
	pci_configure_relaxed_ordering(dev);
2140
	pci_configure_ltr(dev);
2141
	pci_configure_eetlp_prefix(dev);
2142

2143 2144 2145 2146 2147 2148 2149 2150 2151 2152
	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);
}

2153 2154
static void pci_release_capabilities(struct pci_dev *dev)
{
2155
	pci_aer_exit(dev);
2156
	pci_vpd_release(dev);
2157
	pci_iov_release(dev);
2158
	pci_free_cap_save_buffers(dev);
2159 2160
}

L
Linus Torvalds 已提交
2161
/**
B
Bjorn Helgaas 已提交
2162 2163
 * pci_release_dev - Free a PCI device structure when all users of it are
 *		     finished
L
Linus Torvalds 已提交
2164 2165
 * @dev: device that's been disconnected
 *
B
Bjorn Helgaas 已提交
2166
 * Will be called only by the device core when all users of this PCI device are
L
Linus Torvalds 已提交
2167 2168 2169 2170
 * done.
 */
static void pci_release_dev(struct device *dev)
{
2171
	struct pci_dev *pci_dev;
L
Linus Torvalds 已提交
2172

2173
	pci_dev = to_pci_dev(dev);
2174
	pci_release_capabilities(pci_dev);
2175
	pci_release_of_node(pci_dev);
2176
	pcibios_release_device(pci_dev);
2177
	pci_bus_put(pci_dev->bus);
2178
	kfree(pci_dev->driver_override);
2179
	bitmap_free(pci_dev->dma_alias_mask);
L
Linus Torvalds 已提交
2180 2181 2182
	kfree(pci_dev);
}

2183
struct pci_dev *pci_alloc_dev(struct pci_bus *bus)
2184 2185 2186 2187 2188 2189 2190 2191
{
	struct pci_dev *dev;

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

	INIT_LIST_HEAD(&dev->bus_list);
2192
	dev->dev.type = &pci_dev_type;
2193
	dev->bus = pci_bus_get(bus);
2194 2195 2196

	return dev;
}
2197 2198
EXPORT_SYMBOL(pci_alloc_dev);

2199 2200 2201 2202 2203
static bool pci_bus_crs_vendor_id(u32 l)
{
	return (l & 0xffff) == 0x0001;
}

S
Sinan Kaya 已提交
2204 2205
static bool pci_bus_wait_crs(struct pci_bus *bus, int devfn, u32 *l,
			     int timeout)
L
Linus Torvalds 已提交
2206 2207 2208
{
	int delay = 1;

S
Sinan Kaya 已提交
2209 2210
	if (!pci_bus_crs_vendor_id(*l))
		return true;	/* not a CRS completion */
L
Linus Torvalds 已提交
2211

S
Sinan Kaya 已提交
2212 2213
	if (!timeout)
		return false;	/* CRS, but caller doesn't want to wait */
L
Linus Torvalds 已提交
2214

2215
	/*
S
Sinan Kaya 已提交
2216 2217 2218
	 * 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.
2219
	 */
2220
	while (pci_bus_crs_vendor_id(*l)) {
S
Sinan Kaya 已提交
2221
		if (delay > timeout) {
2222 2223 2224 2225
			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);

2226
			return false;
L
Linus Torvalds 已提交
2227
		}
2228 2229 2230 2231
		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);
2232

L
Linus Torvalds 已提交
2233 2234
		msleep(delay);
		delay *= 2;
2235

2236 2237
		if (pci_bus_read_config_dword(bus, devfn, PCI_VENDOR_ID, l))
			return false;
L
Linus Torvalds 已提交
2238 2239
	}

2240 2241 2242 2243 2244
	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);

2245 2246
	return true;
}
S
Sinan Kaya 已提交
2247

2248 2249
bool pci_bus_generic_read_dev_vendor_id(struct pci_bus *bus, int devfn, u32 *l,
					int timeout)
S
Sinan Kaya 已提交
2250 2251 2252 2253
{
	if (pci_bus_read_config_dword(bus, devfn, PCI_VENDOR_ID, l))
		return false;

B
Bjorn Helgaas 已提交
2254
	/* Some broken boards return 0 or ~0 if a slot is empty: */
S
Sinan Kaya 已提交
2255 2256 2257 2258 2259 2260 2261
	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);

2262 2263
	return true;
}
2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281

bool pci_bus_read_dev_vendor_id(struct pci_bus *bus, int devfn, u32 *l,
				int timeout)
{
#ifdef CONFIG_PCI_QUIRKS
	struct pci_dev *bridge = bus->self;

	/*
	 * Certain IDT switches have an issue where they improperly trigger
	 * ACS Source Validation errors on completions for config reads.
	 */
	if (bridge && bridge->vendor == PCI_VENDOR_ID_IDT &&
	    bridge->device == 0x80b5)
		return pci_idt_bus_quirk(bus, devfn, l, timeout);
#endif

	return pci_bus_generic_read_dev_vendor_id(bus, devfn, l, timeout);
}
2282 2283 2284
EXPORT_SYMBOL(pci_bus_read_dev_vendor_id);

/*
B
Bjorn Helgaas 已提交
2285 2286
 * Read the config data for a PCI device, sanity-check it,
 * and fill in the dev structure.
2287 2288 2289 2290 2291 2292 2293 2294 2295
 */
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;

2296
	dev = pci_alloc_dev(bus);
L
Linus Torvalds 已提交
2297 2298 2299 2300 2301 2302
	if (!dev)
		return NULL;

	dev->devfn = devfn;
	dev->vendor = l & 0xffff;
	dev->device = (l >> 16) & 0xffff;
2303

2304 2305
	pci_set_of_node(dev);

Y
Yu Zhao 已提交
2306
	if (pci_setup_device(dev)) {
2307
		pci_bus_put(dev->bus);
L
Linus Torvalds 已提交
2308 2309 2310 2311 2312 2313 2314
		kfree(dev);
		return NULL;
	}

	return dev;
}

2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333
static void pcie_report_downtraining(struct pci_dev *dev)
{
	if (!pci_is_pcie(dev))
		return;

	/* Look from the device up to avoid downstream ports with no devices */
	if ((pci_pcie_type(dev) != PCI_EXP_TYPE_ENDPOINT) &&
	    (pci_pcie_type(dev) != PCI_EXP_TYPE_LEG_END) &&
	    (pci_pcie_type(dev) != PCI_EXP_TYPE_UPSTREAM))
		return;

	/* Multi-function PCIe devices share the same link/status */
	if (PCI_FUNC(dev->devfn) != 0 || dev->is_virtfn)
		return;

	/* Print link status only if the device is constrained by the fabric */
	__pcie_print_link_status(dev, false);
}

2334 2335
static void pci_init_capabilities(struct pci_dev *dev)
{
2336 2337 2338
	/* Enhanced Allocation */
	pci_ea_init(dev);

2339 2340
	/* Setup MSI caps & disable MSI/MSI-X interrupts */
	pci_msi_setup_pci_dev(dev);
2341

2342 2343 2344
	/* Buffers for saving PCIe and PCI-X capabilities */
	pci_allocate_cap_save_buffers(dev);

2345 2346 2347 2348
	/* Power Management */
	pci_pm_init(dev);

	/* Vital Product Data */
2349
	pci_vpd_init(dev);
Y
Yu Zhao 已提交
2350 2351

	/* Alternative Routing-ID Forwarding */
2352
	pci_configure_ari(dev);
2353 2354 2355

	/* Single Root I/O Virtualization */
	pci_iov_init(dev);
2356

2357 2358 2359
	/* Address Translation Services */
	pci_ats_init(dev);

2360
	/* Enable ACS P2P upstream forwarding */
C
Chris Wright 已提交
2361
	pci_enable_acs(dev);
2362

2363 2364
	/* Precision Time Measurement */
	pci_ptm_init(dev);
B
Bjorn Helgaas 已提交
2365

K
Keith Busch 已提交
2366 2367
	/* Advanced Error Reporting */
	pci_aer_init(dev);
2368

2369 2370
	pcie_report_downtraining(dev);

2371 2372
	if (pci_probe_reset_function(dev) == 0)
		dev->reset_fn = 1;
2373 2374
}

M
Marc Zyngier 已提交
2375
/*
B
Bjorn Helgaas 已提交
2376
 * This is the equivalent of pci_host_bridge_msi_domain() that acts on
M
Marc Zyngier 已提交
2377 2378 2379 2380 2381 2382 2383 2384
 * 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
2385
	 * If a domain has been set through the pcibios_add_device()
M
Marc Zyngier 已提交
2386 2387 2388 2389 2390 2391
	 * callback, then this is the one (platform code knows best).
	 */
	d = dev_get_msi_domain(&dev->dev);
	if (d)
		return d;

2392 2393 2394 2395 2396 2397 2398 2399
	/*
	 * 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
2400 2401 2402
	return NULL;
}

2403 2404
static void pci_set_msi_domain(struct pci_dev *dev)
{
M
Marc Zyngier 已提交
2405 2406
	struct irq_domain *d;

2407
	/*
M
Marc Zyngier 已提交
2408 2409 2410
	 * If the platform or firmware interfaces cannot supply a
	 * device-specific MSI domain, then inherit the default domain
	 * from the host bridge itself.
2411
	 */
M
Marc Zyngier 已提交
2412 2413 2414 2415 2416
	d = pci_dev_msi_domain(dev);
	if (!d)
		d = dev_get_msi_domain(&dev->bus->dev);

	dev_set_msi_domain(&dev->dev, d);
2417 2418
}

2419
void pci_device_add(struct pci_dev *dev, struct pci_bus *bus)
L
Linus Torvalds 已提交
2420
{
2421 2422
	int ret;

2423 2424
	pci_configure_device(dev);

2425 2426
	device_initialize(&dev->dev);
	dev->dev.release = pci_release_dev;
L
Linus Torvalds 已提交
2427

2428
	set_dev_node(&dev->dev, pcibus_to_node(bus));
2429
	dev->dev.dma_mask = &dev->dma_mask;
2430
	dev->dev.dma_parms = &dev->dma_parms;
2431
	dev->dev.coherent_dma_mask = 0xffffffffull;
L
Linus Torvalds 已提交
2432

2433
	dma_set_max_seg_size(&dev->dev, 65536);
2434
	dma_set_seg_boundary(&dev->dev, 0xffffffff);
2435

L
Linus Torvalds 已提交
2436 2437 2438
	/* Fix up broken headers */
	pci_fixup_device(pci_fixup_header, dev);

B
Bjorn Helgaas 已提交
2439
	/* Moved out from quirk header fixup code */
2440 2441
	pci_reassigndev_resource_alignment(dev);

B
Bjorn Helgaas 已提交
2442
	/* Clear the state_saved flag */
2443 2444
	dev->state_saved = false;

2445 2446
	/* Initialize various capabilities */
	pci_init_capabilities(dev);
2447

L
Linus Torvalds 已提交
2448 2449 2450 2451
	/*
	 * Add the device to our list of discovered devices
	 * and the bus list for fixup functions, etc.
	 */
2452
	down_write(&pci_bus_sem);
L
Linus Torvalds 已提交
2453
	list_add_tail(&dev->bus_list, &bus->devices);
2454
	up_write(&pci_bus_sem);
2455 2456 2457 2458

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

B
Bjorn Helgaas 已提交
2459
	/* Set up MSI IRQ domain */
2460 2461
	pci_set_msi_domain(dev);

2462 2463 2464 2465
	/* Notifier could use PCI capabilities */
	dev->match_driver = false;
	ret = device_add(&dev->dev);
	WARN_ON(ret < 0);
2466 2467
}

2468
struct pci_dev *pci_scan_single_device(struct pci_bus *bus, int devfn)
2469 2470 2471
{
	struct pci_dev *dev;

T
Trent Piepho 已提交
2472 2473 2474 2475 2476 2477
	dev = pci_get_slot(bus, devfn);
	if (dev) {
		pci_dev_put(dev);
		return dev;
	}

2478 2479 2480 2481 2482
	dev = pci_scan_device(bus, devfn);
	if (!dev)
		return NULL;

	pci_device_add(dev, bus);
L
Linus Torvalds 已提交
2483 2484 2485

	return dev;
}
2486
EXPORT_SYMBOL(pci_scan_single_device);
L
Linus Torvalds 已提交
2487

2488
static unsigned next_fn(struct pci_bus *bus, struct pci_dev *dev, unsigned fn)
M
Matthew Wilcox 已提交
2489
{
2490 2491 2492
	int pos;
	u16 cap = 0;
	unsigned next_fn;
2493

2494 2495 2496 2497 2498 2499
	if (pci_ari_enabled(bus)) {
		if (!dev)
			return 0;
		pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ARI);
		if (!pos)
			return 0;
2500

2501 2502 2503 2504
		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
2505

2506 2507 2508 2509 2510 2511
		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
2512 2513 2514 2515 2516 2517

	return 0;
}

static int only_one_child(struct pci_bus *bus)
{
2518
	struct pci_dev *bridge = bus->self;
2519

2520 2521 2522 2523 2524
	/*
	 * 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
2525
		return 0;
2526 2527

	/*
2528 2529 2530 2531 2532 2533 2534
	 * 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.
2535
	 */
2536
	if (bridge && pci_is_pcie(bridge) && bridge->has_secondary_link)
M
Matthew Wilcox 已提交
2537
		return 1;
2538

M
Matthew Wilcox 已提交
2539 2540 2541
	return 0;
}

L
Linus Torvalds 已提交
2542
/**
B
Bjorn Helgaas 已提交
2543
 * pci_scan_slot - Scan a PCI slot on a bus for devices
L
Linus Torvalds 已提交
2544
 * @bus: PCI bus to scan
B
Bjorn Helgaas 已提交
2545
 * @devfn: slot number to scan (must have zero function)
L
Linus Torvalds 已提交
2546 2547 2548
 *
 * Scan a PCI slot on the specified PCI bus for devices, adding
 * discovered devices to the @bus->devices list.  New devices
2549
 * will not have is_added set.
2550 2551
 *
 * Returns the number of new devices found.
L
Linus Torvalds 已提交
2552
 */
2553
int pci_scan_slot(struct pci_bus *bus, int devfn)
L
Linus Torvalds 已提交
2554
{
M
Matthew Wilcox 已提交
2555
	unsigned fn, nr = 0;
2556
	struct pci_dev *dev;
M
Matthew Wilcox 已提交
2557 2558 2559

	if (only_one_child(bus) && (devfn > 0))
		return 0; /* Already scanned the entire slot */
L
Linus Torvalds 已提交
2560

2561
	dev = pci_scan_single_device(bus, devfn);
2562 2563
	if (!dev)
		return 0;
2564
	if (!pci_dev_is_added(dev))
2565 2566
		nr++;

2567
	for (fn = next_fn(bus, dev, 0); fn > 0; fn = next_fn(bus, dev, fn)) {
M
Matthew Wilcox 已提交
2568 2569
		dev = pci_scan_single_device(bus, devfn + fn);
		if (dev) {
2570
			if (!pci_dev_is_added(dev))
M
Matthew Wilcox 已提交
2571 2572
				nr++;
			dev->multifunction = 1;
L
Linus Torvalds 已提交
2573 2574
		}
	}
S
Shaohua Li 已提交
2575

B
Bjorn Helgaas 已提交
2576
	/* Only one slot has PCIe device */
2577
	if (bus->self && nr)
S
Shaohua Li 已提交
2578 2579
		pcie_aspm_init_link_state(bus->self);

L
Linus Torvalds 已提交
2580 2581
	return nr;
}
2582
EXPORT_SYMBOL(pci_scan_slot);
L
Linus Torvalds 已提交
2583

2584 2585 2586 2587 2588 2589 2590
static int pcie_find_smpss(struct pci_dev *dev, void *data)
{
	u8 *smpss = data;

	if (!pci_is_pcie(dev))
		return 0;

2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604
	/*
	 * 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.
2605
	 */
2606 2607
	if (dev->is_hotplug_bridge &&
	    pci_pcie_type(dev) != PCI_EXP_TYPE_ROOT_PORT)
2608 2609 2610 2611 2612 2613 2614 2615 2616 2617
		*smpss = 0;

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

	return 0;
}

static void pcie_write_mps(struct pci_dev *dev, int mps)
{
2618
	int rc;
2619 2620

	if (pcie_bus_config == PCIE_BUS_PERFORMANCE) {
2621
		mps = 128 << dev->pcie_mpss;
2622

2623 2624
		if (pci_pcie_type(dev) != PCI_EXP_TYPE_ROOT_PORT &&
		    dev->bus->self)
B
Bjorn Helgaas 已提交
2625 2626 2627

			/*
			 * For "Performance", the assumption is made that
2628 2629 2630 2631 2632
			 * 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.
2633 2634 2635 2636 2637
			 *
			 * 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).
2638
			 */
2639
			mps = min(mps, pcie_get_mps(dev->bus->self));
2640 2641 2642 2643
	}

	rc = pcie_set_mps(dev, mps);
	if (rc)
2644
		pci_err(dev, "Failed attempting to set the MPS\n");
2645 2646
}

2647
static void pcie_write_mrrs(struct pci_dev *dev)
2648
{
2649
	int rc, mrrs;
2650

B
Bjorn Helgaas 已提交
2651 2652
	/*
	 * In the "safe" case, do not configure the MRRS.  There appear to be
2653 2654 2655 2656 2657
	 * 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
2658 2659
	/*
	 * For max performance, the MRRS must be set to the largest supported
2660
	 * value.  However, it cannot be configured larger than the MPS the
2661
	 * device or the bus can support.  This should already be properly
B
Bjorn Helgaas 已提交
2662
	 * configured by a prior call to pcie_write_mps().
2663
	 */
2664
	mrrs = pcie_get_mps(dev);
2665

B
Bjorn Helgaas 已提交
2666 2667
	/*
	 * MRRS is a R/W register.  Invalid values can be written, but a
2668
	 * subsequent read will verify if the value is acceptable or not.
2669 2670
	 * If the MRRS value provided is not acceptable (e.g., too large),
	 * shrink the value until it is acceptable to the HW.
2671
	 */
2672 2673
	while (mrrs != pcie_get_readrq(dev) && mrrs >= 128) {
		rc = pcie_set_readrq(dev, mrrs);
2674 2675
		if (!rc)
			break;
2676

2677
		pci_warn(dev, "Failed attempting to set the MRRS\n");
2678 2679
		mrrs /= 2;
	}
2680 2681

	if (mrrs < 128)
2682
		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");
2683 2684 2685 2686
}

static int pcie_bus_configure_set(struct pci_dev *dev, void *data)
{
J
Jon Mason 已提交
2687
	int mps, orig_mps;
2688 2689 2690 2691

	if (!pci_is_pcie(dev))
		return 0;

2692 2693
	if (pcie_bus_config == PCIE_BUS_TUNE_OFF ||
	    pcie_bus_config == PCIE_BUS_DEFAULT)
2694 2695
		return 0;

J
Jon Mason 已提交
2696 2697
	mps = 128 << *(u8 *)data;
	orig_mps = pcie_get_mps(dev);
2698 2699

	pcie_write_mps(dev, mps);
2700
	pcie_write_mrrs(dev);
2701

2702
	pci_info(dev, "Max Payload Size set to %4d/%4d (was %4d), Max Read Rq %4d\n",
2703
		 pcie_get_mps(dev), 128 << dev->pcie_mpss,
J
Jon Mason 已提交
2704
		 orig_mps, pcie_get_readrq(dev));
2705 2706 2707 2708

	return 0;
}

B
Bjorn Helgaas 已提交
2709 2710
/*
 * pcie_bus_configure_settings() requires that pci_walk_bus work in a top-down,
2711 2712 2713
 * parents then children fashion.  If this changes, then this code will not
 * work as designed.
 */
2714
void pcie_bus_configure_settings(struct pci_bus *bus)
2715
{
2716
	u8 smpss = 0;
2717

2718
	if (!bus->self)
2719 2720 2721
		return;

	if (!pci_is_pcie(bus->self))
2722 2723
		return;

B
Bjorn Helgaas 已提交
2724 2725
	/*
	 * FIXME - Peer to peer DMA is possible, though the endpoint would need
2726
	 * to be aware of the MPS of the destination.  To work around this,
2727 2728 2729 2730 2731
	 * simply force the MPS of the entire system to the smallest possible.
	 */
	if (pcie_bus_config == PCIE_BUS_PEER2PEER)
		smpss = 0;

2732
	if (pcie_bus_config == PCIE_BUS_SAFE) {
2733
		smpss = bus->self->pcie_mpss;
2734

2735 2736 2737 2738 2739 2740 2741
		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);
}
2742
EXPORT_SYMBOL_GPL(pcie_bus_configure_settings);
2743

2744 2745 2746 2747 2748 2749 2750 2751 2752
/*
 * 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 */
}

2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766
/**
 * 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
2767
{
2768 2769
	unsigned int used_buses, normal_bridges = 0, hotplug_bridges = 0;
	unsigned int start = bus->busn_res.start;
2770
	unsigned int devfn, fn, cmax, max = start;
L
Linus Torvalds 已提交
2771
	struct pci_dev *dev;
2772
	int nr_devs;
L
Linus Torvalds 已提交
2773

B
Bjorn Helgaas 已提交
2774
	dev_dbg(&bus->dev, "scanning bus\n");
L
Linus Torvalds 已提交
2775 2776

	/* Go find them, Rover! */
2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792
	for (devfn = 0; devfn < 256; devfn += 8) {
		nr_devs = pci_scan_slot(bus, devfn);

		/*
		 * The Jailhouse hypervisor may pass individual functions of a
		 * multi-function device to a guest without passing function 0.
		 * Look for them as well.
		 */
		if (jailhouse_paravirt() && nr_devs == 0) {
			for (fn = 1; fn < 8; fn++) {
				dev = pci_scan_single_device(bus, devfn + fn);
				if (dev)
					dev->multifunction = 1;
			}
		}
	}
L
Linus Torvalds 已提交
2793

B
Bjorn Helgaas 已提交
2794
	/* Reserve buses for SR-IOV capability */
2795 2796
	used_buses = pci_iov_bus_range(bus);
	max += used_buses;
2797

L
Linus Torvalds 已提交
2798 2799 2800 2801
	/*
	 * After performing arch-dependent fixup of the bus, look behind
	 * all PCI-to-PCI bridges on this bus.
	 */
A
Alex Chiang 已提交
2802
	if (!bus->is_added) {
B
Bjorn Helgaas 已提交
2803
		dev_dbg(&bus->dev, "fixups for bus\n");
A
Alex Chiang 已提交
2804
		pcibios_fixup_bus(bus);
2805
		bus->is_added = 1;
A
Alex Chiang 已提交
2806 2807
	}

2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819
	/*
	 * 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++;
	}

2820 2821 2822 2823 2824
	/*
	 * Scan bridges that are already configured. We don't touch them
	 * unless they are misconfigured (which will be done in the second
	 * scan below).
	 */
2825 2826 2827
	for_each_pci_bridge(dev, bus) {
		cmax = max;
		max = pci_scan_bridge_extend(bus, dev, max, 0, 0);
2828 2829 2830 2831 2832 2833 2834 2835

		/*
		 * Reserve one bus for each bridge now to avoid extending
		 * hotplug bridges too much during the second scan below.
		 */
		used_buses++;
		if (cmax - max > 1)
			used_buses += cmax - max - 1;
2836
	}
2837 2838

	/* Scan bridges that need to be reconfigured */
2839 2840 2841 2842
	for_each_pci_bridge(dev, bus) {
		unsigned int buses = 0;

		if (!hotplug_bridges && normal_bridges == 1) {
B
Bjorn Helgaas 已提交
2843

2844 2845 2846 2847 2848 2849 2850 2851
			/*
			 * 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
2852

2853 2854 2855 2856 2857
			/*
			 * Distribute the extra buses between hotplug
			 * bridges if any.
			 */
			buses = available_buses / hotplug_bridges;
2858
			buses = min(buses, available_buses - used_buses + 1);
2859 2860 2861 2862
		}

		cmax = max;
		max = pci_scan_bridge_extend(bus, dev, cmax, buses, 1);
2863 2864 2865
		/* One bus is already accounted so don't add it again */
		if (max - cmax > 1)
			used_buses += max - cmax - 1;
2866
	}
L
Linus Torvalds 已提交
2867

2868 2869
	/*
	 * Make sure a hotplug bridge has at least the minimum requested
2870 2871
	 * number of buses but allow it to grow up to the maximum available
	 * bus number of there is room.
2872
	 */
2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885
	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);
		}
2886 2887
	}

L
Linus Torvalds 已提交
2888 2889 2890 2891 2892 2893 2894
	/*
	 * 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
2895
	dev_dbg(&bus->dev, "bus scan returning with max=%02x\n", max);
L
Linus Torvalds 已提交
2896 2897
	return max;
}
2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909

/**
 * 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);
}
2910
EXPORT_SYMBOL_GPL(pci_scan_child_bus);
L
Linus Torvalds 已提交
2911

2912
/**
B
Bjorn Helgaas 已提交
2913 2914
 * pcibios_root_bridge_prepare - Platform-specific host bridge setup
 * @bridge: Host bridge to set up
2915 2916 2917 2918 2919 2920 2921 2922 2923
 *
 * 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;
}

2924 2925 2926 2927 2928 2929 2930 2931
void __weak pcibios_add_bus(struct pci_bus *bus)
{
}

void __weak pcibios_remove_bus(struct pci_bus *bus)
{
}

2932 2933
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
2934
{
2935
	int error;
2936
	struct pci_host_bridge *bridge;
L
Linus Torvalds 已提交
2937

2938
	bridge = pci_alloc_host_bridge(0);
2939
	if (!bridge)
2940
		return NULL;
2941 2942

	bridge->dev.parent = parent;
2943

2944 2945 2946 2947
	list_splice_init(resources, &bridge->windows);
	bridge->sysdata = sysdata;
	bridge->busnr = bus;
	bridge->ops = ops;
2948

2949 2950 2951
	error = pci_register_host_bridge(bridge);
	if (error < 0)
		goto err_out;
2952

2953
	return bridge->bus;
L
Linus Torvalds 已提交
2954 2955

err_out:
2956
	kfree(bridge);
L
Linus Torvalds 已提交
2957 2958
	return NULL;
}
2959
EXPORT_SYMBOL_GPL(pci_create_root_bus);
2960

2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993
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);

2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009
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;
	}

3010
	conflict = request_resource_conflict(parent_res, res);
3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056

	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");
}

3057
int pci_scan_root_bus_bridge(struct pci_host_bridge *bridge)
3058
{
3059
	struct resource_entry *window;
3060
	bool found = false;
3061
	struct pci_bus *b;
3062
	int max, bus, ret;
3063

3064 3065 3066 3067
	if (!bridge)
		return -EINVAL;

	resource_list_for_each_entry(window, &bridge->windows)
3068 3069 3070 3071
		if (window->res->flags & IORESOURCE_BUS) {
			found = true;
			break;
		}
3072

3073 3074 3075 3076 3077 3078
	ret = pci_register_host_bridge(bridge);
	if (ret < 0)
		return ret;

	b = bridge->bus;
	bus = bridge->busnr;
3079

3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091
	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);

3092
	return 0;
3093
}
3094
EXPORT_SYMBOL(pci_scan_root_bus_bridge);
3095 3096 3097 3098

struct pci_bus *pci_scan_root_bus(struct device *parent, int bus,
		struct pci_ops *ops, void *sysdata, struct list_head *resources)
{
3099
	struct resource_entry *window;
3100
	bool found = false;
3101
	struct pci_bus *b;
3102 3103
	int max;

3104
	resource_list_for_each_entry(window, resources)
3105 3106 3107 3108
		if (window->res->flags & IORESOURCE_BUS) {
			found = true;
			break;
		}
3109

3110
	b = pci_create_root_bus(parent, bus, ops, sysdata, resources);
3111 3112 3113
	if (!b)
		return NULL;

3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125
	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);

3126
	return b;
3127
}
3128 3129
EXPORT_SYMBOL(pci_scan_root_bus);

B
Bill Pemberton 已提交
3130
struct pci_bus *pci_scan_bus(int bus, struct pci_ops *ops,
3131 3132 3133 3134 3135 3136 3137
					void *sysdata)
{
	LIST_HEAD(resources);
	struct pci_bus *b;

	pci_add_resource(&resources, &ioport_resource);
	pci_add_resource(&resources, &iomem_resource);
3138
	pci_add_resource(&resources, &busn_resource);
3139 3140
	b = pci_create_root_bus(NULL, bus, ops, sysdata, &resources);
	if (b) {
3141
		pci_scan_child_bus(b);
3142 3143 3144 3145 3146 3147 3148
	} else {
		pci_free_resource_list(&resources);
	}
	return b;
}
EXPORT_SYMBOL(pci_scan_bus);

3149
/**
B
Bjorn Helgaas 已提交
3150
 * pci_rescan_bus_bridge_resize - Scan a PCI bus for devices
3151 3152 3153 3154 3155 3156 3157 3158 3159
 * @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.
 */
3160
unsigned int pci_rescan_bus_bridge_resize(struct pci_dev *bridge)
3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173
{
	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;
}

3174
/**
B
Bjorn Helgaas 已提交
3175
 * pci_rescan_bus - Scan a PCI bus for devices
3176 3177
 * @bus: PCI bus to scan
 *
B
Bjorn Helgaas 已提交
3178 3179
 * Scan a PCI bus and child buses for new devices, add them,
 * and enable them.
3180 3181 3182
 *
 * Returns the max number of subordinate bus discovered.
 */
3183
unsigned int pci_rescan_bus(struct pci_bus *bus)
3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194
{
	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);

3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212
/*
 * 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
3213 3214
static int __init pci_sort_bf_cmp(const struct device *d_a,
				  const struct device *d_b)
3215
{
3216 3217 3218
	const struct pci_dev *a = to_pci_dev(d_a);
	const struct pci_dev *b = to_pci_dev(d_b);

3219 3220 3221 3222 3223 3224 3225 3226 3227 3228 3229 3230
	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;
}

3231
void __init pci_sort_breadthfirst(void)
3232
{
3233
	bus_sort_breadthfirst(&pci_bus_type, &pci_sort_bf_cmp);
3234
}
3235 3236 3237 3238

int pci_hp_add_bridge(struct pci_dev *dev)
{
	struct pci_bus *parent = dev->bus;
3239
	int busnr, start = parent->busn_res.start;
3240
	unsigned int available_buses = 0;
3241 3242 3243 3244 3245 3246 3247
	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) {
3248
		pci_err(dev, "No bus number available for hot-added bridge\n");
3249 3250
		return -1;
	}
3251 3252 3253 3254

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

3255 3256 3257 3258 3259 3260
	/*
	 * Distribute the available bus numbers between hotplug-capable
	 * bridges to make extending the chain later possible.
	 */
	available_buses = end - busnr;

3261
	/* Scan bridges that need to be reconfigured */
3262
	pci_scan_bridge_extend(parent, dev, busnr, available_buses, 1);
3263

3264 3265 3266 3267 3268 3269
	if (!dev->subordinate)
		return -1;

	return 0;
}
EXPORT_SYMBOL_GPL(pci_hp_add_bridge);