probe.c 57.8 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8
/*
 * probe.c - PCI detection and setup code
 */

#include <linux/kernel.h>
#include <linux/delay.h>
#include <linux/init.h>
#include <linux/pci.h>
9
#include <linux/of_pci.h>
10
#include <linux/pci_hotplug.h>
L
Linus Torvalds 已提交
11 12 13
#include <linux/slab.h>
#include <linux/module.h>
#include <linux/cpumask.h>
S
Shaohua Li 已提交
14
#include <linux/pci-aspm.h>
15
#include <asm-generic/pci-bridge.h>
16
#include "pci.h"
L
Linus Torvalds 已提交
17 18 19 20

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

21
static struct resource busn_resource = {
Y
Yinghai Lu 已提交
22 23 24 25 26 27
	.name	= "PCI busn",
	.start	= 0,
	.end	= 255,
	.flags	= IORESOURCE_BUS,
};

L
Linus Torvalds 已提交
28 29 30 31
/* Ugh.  Need to stop exporting this to modules. */
LIST_HEAD(pci_root_buses);
EXPORT_SYMBOL(pci_root_buses);

32 33 34 35 36 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
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;
}

62 63 64 65
static int find_anything(struct device *dev, void *data)
{
	return 1;
}
L
Linus Torvalds 已提交
66

Z
Zhang, Yanmin 已提交
67 68 69
/*
 * Some device drivers need know if pci is initiated.
 * Basically, we think pci is not initiated when there
70
 * is no device to be found on the pci_bus_type.
Z
Zhang, Yanmin 已提交
71 72 73
 */
int no_pci_devices(void)
{
74 75
	struct device *dev;
	int no_devices;
Z
Zhang, Yanmin 已提交
76

77 78 79 80 81
	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
82 83
EXPORT_SYMBOL(no_pci_devices);

L
Linus Torvalds 已提交
84 85 86
/*
 * PCI Bus Class
 */
87
static void release_pcibus_dev(struct device *dev)
L
Linus Torvalds 已提交
88
{
89
	struct pci_bus *pci_bus = to_pci_bus(dev);
L
Linus Torvalds 已提交
90

91
	put_device(pci_bus->bridge);
92
	pci_bus_remove_resources(pci_bus);
93
	pci_release_bus_of_node(pci_bus);
L
Linus Torvalds 已提交
94 95 96 97 98
	kfree(pci_bus);
}

static struct class pcibus_class = {
	.name		= "pci_bus",
99
	.dev_release	= &release_pcibus_dev,
100
	.dev_groups	= pcibus_groups,
L
Linus Torvalds 已提交
101 102 103 104 105 106 107 108
};

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

109
static u64 pci_size(u64 base, u64 maxbase, u64 mask)
L
Linus Torvalds 已提交
110
{
111
	u64 size = mask & maxbase;	/* Find the significant bits */
L
Linus Torvalds 已提交
112 113 114 115 116 117 118 119 120 121 122 123 124 125 126
	if (!size)
		return 0;

	/* Get the lowest of them to find the decode size, and
	   from that the extent.  */
	size = (size & ~(size-1)) - 1;

	/* base == maxbase can be valid only if the BAR has
	   already been programmed with all 1s.  */
	if (base == maxbase && ((base | size) & mask) != mask)
		return 0;

	return size;
}

127
static inline unsigned long decode_bar(struct pci_dev *dev, u32 bar)
128
{
129
	u32 mem_type;
130
	unsigned long flags;
131

132
	if ((bar & PCI_BASE_ADDRESS_SPACE) == PCI_BASE_ADDRESS_SPACE_IO) {
133 134 135
		flags = bar & ~PCI_BASE_ADDRESS_IO_MASK;
		flags |= IORESOURCE_IO;
		return flags;
136
	}
137

138 139 140 141
	flags = bar & ~PCI_BASE_ADDRESS_MEM_MASK;
	flags |= IORESOURCE_MEM;
	if (flags & PCI_BASE_ADDRESS_MEM_PREFETCH)
		flags |= IORESOURCE_PREFETCH;
142

143 144 145 146 147
	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:
148
		/* 1M mem BAR treated as 32-bit BAR */
149 150
		break;
	case PCI_BASE_ADDRESS_MEM_TYPE_64:
151 152
		flags |= IORESOURCE_MEM_64;
		break;
153
	default:
154
		/* mem unknown type treated as 32-bit BAR */
155 156
		break;
	}
157
	return flags;
158 159
}

160 161
#define PCI_COMMAND_DECODE_ENABLE	(PCI_COMMAND_MEMORY | PCI_COMMAND_IO)

Y
Yu Zhao 已提交
162 163 164 165 166 167 168 169
/**
 * pci_read_base - read a PCI BAR
 * @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.
170
 */
Y
Yu Zhao 已提交
171
int __pci_read_base(struct pci_dev *dev, enum pci_bar_type type,
R
Ryan Desfosses 已提交
172
		    struct resource *res, unsigned int pos)
173
{
174
	u32 l, sz, mask;
175
	u64 l64, sz64, mask64;
J
Jacob Pan 已提交
176
	u16 orig_cmd;
177
	struct pci_bus_region region, inverted_region;
178

179
	mask = type ? PCI_ROM_ADDRESS_MASK : ~0;
180

181
	/* No printks while decoding is disabled! */
J
Jacob Pan 已提交
182 183
	if (!dev->mmio_always_on) {
		pci_read_config_word(dev, PCI_COMMAND, &orig_cmd);
184 185 186 187
		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
188 189
	}

190 191 192
	res->name = pci_name(dev);

	pci_read_config_dword(dev, pos, &l);
193
	pci_write_config_dword(dev, pos, l | mask);
194 195 196 197 198
	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.
199 200 201
	 * 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.
202
	 */
203 204
	if (sz == 0xffffffff)
		sz = 0;
205 206 207 208 209 210 211 212 213

	/*
	 * 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) {
214 215 216
		res->flags = decode_bar(dev, l);
		res->flags |= IORESOURCE_SIZEALIGN;
		if (res->flags & IORESOURCE_IO) {
217 218 219
			l64 = l & PCI_BASE_ADDRESS_IO_MASK;
			sz64 = sz & PCI_BASE_ADDRESS_IO_MASK;
			mask64 = PCI_BASE_ADDRESS_IO_MASK & (u32)IO_SPACE_LIMIT;
220
		} else {
221 222 223
			l64 = l & PCI_BASE_ADDRESS_MEM_MASK;
			sz64 = sz & PCI_BASE_ADDRESS_MEM_MASK;
			mask64 = (u32)PCI_BASE_ADDRESS_MEM_MASK;
224 225 226
		}
	} else {
		res->flags |= (l & IORESOURCE_ROM_ENABLE);
227 228 229
		l64 = l & PCI_ROM_ADDRESS_MASK;
		sz64 = sz & PCI_ROM_ADDRESS_MASK;
		mask64 = (u32)PCI_ROM_ADDRESS_MASK;
230 231
	}

232
	if (res->flags & IORESOURCE_MEM_64) {
233 234 235 236 237 238 239
		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);
240 241
		mask64 |= ((u64)~0 << 32);
	}
242

243 244
	if (!dev->mmio_always_on && (orig_cmd & PCI_COMMAND_DECODE_ENABLE))
		pci_write_config_word(dev, PCI_COMMAND, orig_cmd);
245

246 247
	if (!sz64)
		goto fail;
248

249
	sz64 = pci_size(l64, sz64, mask64);
250 251 252
	if (!sz64) {
		dev_info(&dev->dev, FW_BUG "reg 0x%x: invalid BAR (can't size)\n",
			 pos);
253
		goto fail;
254
	}
255 256

	if (res->flags & IORESOURCE_MEM_64) {
Y
Yinghai Lu 已提交
257 258
		if ((sizeof(pci_bus_addr_t) < 8 || sizeof(resource_size_t) < 8)
		    && sz64 > 0x100000000ULL) {
259 260 261
			res->flags |= IORESOURCE_UNSET | IORESOURCE_DISABLED;
			res->start = 0;
			res->end = 0;
262 263
			dev_err(&dev->dev, "reg 0x%x: can't handle BAR larger than 4GB (size %#010llx)\n",
				pos, (unsigned long long)sz64);
264
			goto out;
265 266
		}

Y
Yinghai Lu 已提交
267
		if ((sizeof(pci_bus_addr_t) < 8) && l) {
268
			/* Above 32-bit boundary; try to reallocate */
269
			res->flags |= IORESOURCE_UNSET;
270 271
			res->start = 0;
			res->end = sz64;
272 273
			dev_info(&dev->dev, "reg 0x%x: can't handle BAR above 4GB (bus address %#010llx)\n",
				 pos, (unsigned long long)l64);
274
			goto out;
275 276 277
		}
	}

278 279 280
	region.start = l64;
	region.end = l64 + sz64;

281 282
	pcibios_bus_to_resource(dev->bus, res, &region);
	pcibios_resource_to_bus(dev->bus, &inverted_region, res);
283 284 285 286 287 288 289 290 291 292 293 294 295 296 297

	/*
	 * 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;
298
		res->end = region.end - region.start;
299 300
		dev_info(&dev->dev, "reg 0x%x: initial BAR value %#010llx invalid\n",
			 pos, (unsigned long long)region.start);
301
	}
302

303 304 305 306 307 308
	goto out;


fail:
	res->flags = 0;
out:
309
	if (res->flags)
310
		dev_printk(KERN_DEBUG, &dev->dev, "reg 0x%x: %pR\n", pos, res);
311

312
	return (res->flags & IORESOURCE_MEM_64) ? 1 : 0;
313 314
}

L
Linus Torvalds 已提交
315 316
static void pci_read_bases(struct pci_dev *dev, unsigned int howmany, int rom)
{
317
	unsigned int pos, reg;
318

319 320
	for (pos = 0; pos < howmany; pos++) {
		struct resource *res = &dev->resource[pos];
L
Linus Torvalds 已提交
321
		reg = PCI_BASE_ADDRESS_0 + (pos << 2);
322
		pos += __pci_read_base(dev, pci_bar_unknown, res, reg);
L
Linus Torvalds 已提交
323
	}
324

L
Linus Torvalds 已提交
325
	if (rom) {
326
		struct resource *res = &dev->resource[PCI_ROM_RESOURCE];
L
Linus Torvalds 已提交
327
		dev->rom_base_reg = rom;
328 329 330 331
		res->flags = IORESOURCE_MEM | IORESOURCE_PREFETCH |
				IORESOURCE_READONLY | IORESOURCE_CACHEABLE |
				IORESOURCE_SIZEALIGN;
		__pci_read_base(dev, pci_bar_mem32, res, rom);
L
Linus Torvalds 已提交
332 333 334
	}
}

B
Bill Pemberton 已提交
335
static void pci_read_bridge_io(struct pci_bus *child)
L
Linus Torvalds 已提交
336 337 338
{
	struct pci_dev *dev = child->self;
	u8 io_base_lo, io_limit_lo;
339
	unsigned long io_mask, io_granularity, base, limit;
340
	struct pci_bus_region region;
341 342 343 344 345 346 347 348 349
	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
350 351 352 353

	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);
354 355
	base = (io_base_lo & io_mask) << 8;
	limit = (io_limit_lo & io_mask) << 8;
L
Linus Torvalds 已提交
356 357 358

	if ((io_base_lo & PCI_IO_RANGE_TYPE_MASK) == PCI_IO_RANGE_TYPE_32) {
		u16 io_base_hi, io_limit_hi;
359

L
Linus Torvalds 已提交
360 361
		pci_read_config_word(dev, PCI_IO_BASE_UPPER16, &io_base_hi);
		pci_read_config_word(dev, PCI_IO_LIMIT_UPPER16, &io_limit_hi);
362 363
		base |= ((unsigned long) io_base_hi << 16);
		limit |= ((unsigned long) io_limit_hi << 16);
L
Linus Torvalds 已提交
364 365
	}

366
	if (base <= limit) {
L
Linus Torvalds 已提交
367
		res->flags = (io_base_lo & PCI_IO_RANGE_TYPE_MASK) | IORESOURCE_IO;
368
		region.start = base;
369
		region.end = limit + io_granularity - 1;
370
		pcibios_bus_to_resource(dev->bus, res, &region);
371
		dev_printk(KERN_DEBUG, &dev->dev, "  bridge window %pR\n", res);
L
Linus Torvalds 已提交
372
	}
373 374
}

B
Bill Pemberton 已提交
375
static void pci_read_bridge_mmio(struct pci_bus *child)
376 377 378 379
{
	struct pci_dev *dev = child->self;
	u16 mem_base_lo, mem_limit_lo;
	unsigned long base, limit;
380
	struct pci_bus_region region;
381
	struct resource *res;
L
Linus Torvalds 已提交
382 383 384 385

	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);
386 387
	base = ((unsigned long) mem_base_lo & PCI_MEMORY_RANGE_MASK) << 16;
	limit = ((unsigned long) mem_limit_lo & PCI_MEMORY_RANGE_MASK) << 16;
388
	if (base <= limit) {
L
Linus Torvalds 已提交
389
		res->flags = (mem_base_lo & PCI_MEMORY_RANGE_TYPE_MASK) | IORESOURCE_MEM;
390 391
		region.start = base;
		region.end = limit + 0xfffff;
392
		pcibios_bus_to_resource(dev->bus, res, &region);
393
		dev_printk(KERN_DEBUG, &dev->dev, "  bridge window %pR\n", res);
L
Linus Torvalds 已提交
394
	}
395 396
}

B
Bill Pemberton 已提交
397
static void pci_read_bridge_mmio_pref(struct pci_bus *child)
398 399 400
{
	struct pci_dev *dev = child->self;
	u16 mem_base_lo, mem_limit_lo;
401
	u64 base64, limit64;
Y
Yinghai Lu 已提交
402
	pci_bus_addr_t base, limit;
403
	struct pci_bus_region region;
404
	struct resource *res;
L
Linus Torvalds 已提交
405 406 407 408

	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);
409 410
	base64 = (mem_base_lo & PCI_PREF_RANGE_MASK) << 16;
	limit64 = (mem_limit_lo & PCI_PREF_RANGE_MASK) << 16;
L
Linus Torvalds 已提交
411 412 413

	if ((mem_base_lo & PCI_PREF_RANGE_TYPE_MASK) == PCI_PREF_RANGE_TYPE_64) {
		u32 mem_base_hi, mem_limit_hi;
414

L
Linus Torvalds 已提交
415 416 417 418 419 420 421 422 423
		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) {
424 425
			base64 |= (u64) mem_base_hi << 32;
			limit64 |= (u64) mem_limit_hi << 32;
L
Linus Torvalds 已提交
426 427
		}
	}
428

Y
Yinghai Lu 已提交
429 430
	base = (pci_bus_addr_t) base64;
	limit = (pci_bus_addr_t) limit64;
431 432 433 434 435 436 437

	if (base != base64) {
		dev_err(&dev->dev, "can't handle bridge window above 4GB (bus address %#010llx)\n",
			(unsigned long long) base64);
		return;
	}

438
	if (base <= limit) {
439 440 441 442
		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;
443 444
		region.start = base;
		region.end = limit + 0xfffff;
445
		pcibios_bus_to_resource(dev->bus, res, &region);
446
		dev_printk(KERN_DEBUG, &dev->dev, "  bridge window %pR\n", res);
L
Linus Torvalds 已提交
447 448 449
	}
}

B
Bill Pemberton 已提交
450
void pci_read_bridge_bases(struct pci_bus *child)
451 452
{
	struct pci_dev *dev = child->self;
453
	struct resource *res;
454 455 456 457 458
	int i;

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

459 460
	dev_info(&dev->dev, "PCI bridge to %pR%s\n",
		 &child->busn_res,
461 462
		 dev->transparent ? " (subtractive decode)" : "");

463 464 465 466
	pci_bus_remove_resources(child);
	for (i = 0; i < PCI_BRIDGE_RESOURCE_NUM; i++)
		child->resource[i] = &dev->resource[PCI_BRIDGE_RESOURCES+i];

467 468 469
	pci_read_bridge_io(child);
	pci_read_bridge_mmio(child);
	pci_read_bridge_mmio_pref(child);
470 471

	if (dev->transparent) {
472
		pci_bus_for_each_resource(child->parent, res, i) {
473
			if (res && res->flags) {
474 475
				pci_bus_add_resource(child, res,
						     PCI_SUBTRACTIVE_DECODE);
476 477
				dev_printk(KERN_DEBUG, &dev->dev,
					   "  bridge window %pR (subtractive decode)\n",
478 479
					   res);
			}
480 481
		}
	}
482 483
}

484
static struct pci_bus *pci_alloc_bus(struct pci_bus *parent)
L
Linus Torvalds 已提交
485 486 487
{
	struct pci_bus *b;

488
	b = kzalloc(sizeof(*b), GFP_KERNEL);
489 490 491 492 493 494 495 496 497 498
	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;
499 500 501 502
#ifdef CONFIG_PCI_DOMAINS_GENERIC
	if (parent)
		b->domain_nr = parent->domain_nr;
#endif
L
Linus Torvalds 已提交
503 504 505
	return b;
}

506 507 508 509 510 511 512 513 514 515 516 517
static void pci_release_host_bridge_dev(struct device *dev)
{
	struct pci_host_bridge *bridge = to_pci_host_bridge(dev);

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

	pci_free_resource_list(&bridge->windows);

	kfree(bridge);
}

518 519 520 521 522
static struct pci_host_bridge *pci_alloc_host_bridge(struct pci_bus *b)
{
	struct pci_host_bridge *bridge;

	bridge = kzalloc(sizeof(*bridge), GFP_KERNEL);
523 524
	if (!bridge)
		return NULL;
525

526 527
	INIT_LIST_HEAD(&bridge->windows);
	bridge->bus = b;
528 529 530
	return bridge;
}

531
static const unsigned char pcix_bus_speed[] = {
532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549
	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 */
};

550
const unsigned char pcie_link_speed[] = {
551 552 553
	PCI_SPEED_UNKNOWN,		/* 0 */
	PCIE_SPEED_2_5GT,		/* 1 */
	PCIE_SPEED_5_0GT,		/* 2 */
554
	PCIE_SPEED_8_0GT,		/* 3 */
555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570
	PCI_SPEED_UNKNOWN,		/* 4 */
	PCI_SPEED_UNKNOWN,		/* 5 */
	PCI_SPEED_UNKNOWN,		/* 6 */
	PCI_SPEED_UNKNOWN,		/* 7 */
	PCI_SPEED_UNKNOWN,		/* 8 */
	PCI_SPEED_UNKNOWN,		/* 9 */
	PCI_SPEED_UNKNOWN,		/* A */
	PCI_SPEED_UNKNOWN,		/* B */
	PCI_SPEED_UNKNOWN,		/* C */
	PCI_SPEED_UNKNOWN,		/* D */
	PCI_SPEED_UNKNOWN,		/* E */
	PCI_SPEED_UNKNOWN		/* F */
};

void pcie_update_link_speed(struct pci_bus *bus, u16 linksta)
{
571
	bus->cur_bus_speed = pcie_link_speed[linksta & PCI_EXP_LNKSTA_CLS];
572 573 574
}
EXPORT_SYMBOL_GPL(pcie_update_link_speed);

575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594
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;
595

596 597 598 599 600 601 602 603 604 605
	if (agp3) {
		index += 2;
		if (index == 5)
			index = 0;
	}

 out:
	return agp_speeds[index];
}

606 607 608 609 610
static void pci_set_bus_speed(struct pci_bus *bus)
{
	struct pci_dev *bridge = bus->self;
	int pos;

611 612 613 614 615 616 617 618 619 620 621 622 623
	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);
	}

624 625 626 627 628
	pos = pci_find_capability(bridge, PCI_CAP_ID_PCIX);
	if (pos) {
		u16 status;
		enum pci_bus_speed max;

629 630 631 632
		pci_read_config_word(bridge, pos + PCI_X_BRIDGE_SSTATUS,
				     &status);

		if (status & PCI_X_SSTATUS_533MHZ) {
633
			max = PCI_SPEED_133MHz_PCIX_533;
634
		} else if (status & PCI_X_SSTATUS_266MHZ) {
635
			max = PCI_SPEED_133MHz_PCIX_266;
636
		} else if (status & PCI_X_SSTATUS_133MHZ) {
R
Ryan Desfosses 已提交
637
			if ((status & PCI_X_SSTATUS_VERS) == PCI_X_SSTATUS_V2)
638
				max = PCI_SPEED_133MHz_PCIX_ECC;
R
Ryan Desfosses 已提交
639
			else
640 641 642 643 644 645
				max = PCI_SPEED_133MHz_PCIX;
		} else {
			max = PCI_SPEED_66MHz_PCIX;
		}

		bus->max_bus_speed = max;
646 647
		bus->cur_bus_speed = pcix_bus_speed[
			(status & PCI_X_SSTATUS_FREQ) >> 6];
648 649 650 651

		return;
	}

652
	if (pci_is_pcie(bridge)) {
653 654 655
		u32 linkcap;
		u16 linksta;

656
		pcie_capability_read_dword(bridge, PCI_EXP_LNKCAP, &linkcap);
657
		bus->max_bus_speed = pcie_link_speed[linkcap & PCI_EXP_LNKCAP_SLS];
658

659
		pcie_capability_read_word(bridge, PCI_EXP_LNKSTA, &linksta);
660 661 662 663
		pcie_update_link_speed(bus, linksta);
	}
}

664 665
static struct pci_bus *pci_alloc_child_bus(struct pci_bus *parent,
					   struct pci_dev *bridge, int busnr)
L
Linus Torvalds 已提交
666 667 668
{
	struct pci_bus *child;
	int i;
669
	int ret;
L
Linus Torvalds 已提交
670 671 672 673

	/*
	 * Allocate a new bus, and inherit stuff from the parent..
	 */
674
	child = pci_alloc_bus(parent);
L
Linus Torvalds 已提交
675 676 677 678 679
	if (!child)
		return NULL;

	child->parent = parent;
	child->ops = parent->ops;
680
	child->msi = parent->msi;
L
Linus Torvalds 已提交
681
	child->sysdata = parent->sysdata;
682
	child->bus_flags = parent->bus_flags;
L
Linus Torvalds 已提交
683

684
	/* initialize some portions of the bus device, but don't register it
685
	 * now as the parent is not properly set up yet.
686 687
	 */
	child->dev.class = &pcibus_class;
688
	dev_set_name(&child->dev, "%04x:%02x", pci_domain_nr(child), busnr);
L
Linus Torvalds 已提交
689 690 691 692 693

	/*
	 * Set up the primary, secondary and subordinate
	 * bus numbers.
	 */
694 695 696
	child->number = child->busn_res.start = busnr;
	child->primary = parent->busn_res.start;
	child->busn_res.end = 0xff;
L
Linus Torvalds 已提交
697

698 699 700 701
	if (!bridge) {
		child->dev.parent = parent->bridge;
		goto add_dev;
	}
702 703 704

	child->self = bridge;
	child->bridge = get_device(&bridge->dev);
705
	child->dev.parent = child->bridge;
706
	pci_set_bus_of_node(child);
707 708
	pci_set_bus_speed(child);

L
Linus Torvalds 已提交
709
	/* Set up default resource pointers and names.. */
710
	for (i = 0; i < PCI_BRIDGE_RESOURCE_NUM; i++) {
L
Linus Torvalds 已提交
711 712 713 714 715
		child->resource[i] = &bridge->resource[PCI_BRIDGE_RESOURCES+i];
		child->resource[i]->name = child->name;
	}
	bridge->subordinate = child;

716 717 718 719
add_dev:
	ret = device_register(&child->dev);
	WARN_ON(ret < 0);

720 721
	pcibios_add_bus(child);

722 723 724
	/* Create legacy_io and legacy_mem files for this bus */
	pci_create_legacy_files(child);

L
Linus Torvalds 已提交
725 726 727
	return child;
}

R
Ryan Desfosses 已提交
728 729
struct pci_bus *pci_add_new_bus(struct pci_bus *parent, struct pci_dev *dev,
				int busnr)
L
Linus Torvalds 已提交
730 731 732 733
{
	struct pci_bus *child;

	child = pci_alloc_child_bus(parent, dev, busnr);
734
	if (child) {
735
		down_write(&pci_bus_sem);
L
Linus Torvalds 已提交
736
		list_add_tail(&child->node, &parent->children);
737
		up_write(&pci_bus_sem);
738
	}
L
Linus Torvalds 已提交
739 740
	return child;
}
741
EXPORT_SYMBOL(pci_add_new_bus);
L
Linus Torvalds 已提交
742

743 744 745 746 747 748 749 750 751 752 753
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);
}

L
Linus Torvalds 已提交
754 755 756 757 758 759 760 761 762 763
/*
 * 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.
 */
B
Bill Pemberton 已提交
764
int pci_scan_bridge(struct pci_bus *bus, struct pci_dev *dev, int max, int pass)
L
Linus Torvalds 已提交
765 766 767
{
	struct pci_bus *child;
	int is_cardbus = (dev->hdr_type == PCI_HEADER_TYPE_CARDBUS);
768
	u32 buses, i, j = 0;
L
Linus Torvalds 已提交
769
	u16 bctl;
770
	u8 primary, secondary, subordinate;
771
	int broken = 0;
L
Linus Torvalds 已提交
772 773

	pci_read_config_dword(dev, PCI_PRIMARY_BUS, &buses);
774 775 776
	primary = buses & 0xFF;
	secondary = (buses >> 8) & 0xFF;
	subordinate = (buses >> 16) & 0xFF;
L
Linus Torvalds 已提交
777

778 779
	dev_dbg(&dev->dev, "scanning [bus %02x-%02x] behind bridge, pass %d\n",
		secondary, subordinate, pass);
L
Linus Torvalds 已提交
780

781 782 783 784 785
	if (!primary && (primary != bus->number) && secondary && subordinate) {
		dev_warn(&dev->dev, "Primary bus is hard wired to 0\n");
		primary = bus->number;
	}

786 787
	/* Check if setup is sensible at all */
	if (!pass &&
788
	    (primary != bus->number || secondary <= bus->number ||
789
	     secondary > subordinate)) {
790 791
		dev_info(&dev->dev, "bridge configuration invalid ([bus %02x-%02x]), reconfiguring\n",
			 secondary, subordinate);
792 793 794
		broken = 1;
	}

L
Linus Torvalds 已提交
795
	/* Disable MasterAbortMode during probing to avoid reporting
796
	   of bus errors (in some architectures) */
L
Linus Torvalds 已提交
797 798 799 800
	pci_read_config_word(dev, PCI_BRIDGE_CONTROL, &bctl);
	pci_write_config_word(dev, PCI_BRIDGE_CONTROL,
			      bctl & ~PCI_BRIDGE_CTL_MASTER_ABORT);

801 802
	pci_enable_crs(dev);

803 804 805
	if ((secondary || subordinate) && !pcibios_assign_all_busses() &&
	    !is_cardbus && !broken) {
		unsigned int cmax;
L
Linus Torvalds 已提交
806 807 808 809 810
		/*
		 * Bus already configured by firmware, process it in the first
		 * pass and just note the configuration.
		 */
		if (pass)
811
			goto out;
L
Linus Torvalds 已提交
812 813

		/*
814 815 816 817
		 * 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
818
		 */
819
		child = pci_find_bus(pci_domain_nr(bus), secondary);
A
Alex Chiang 已提交
820
		if (!child) {
821
			child = pci_add_new_bus(bus, dev, secondary);
A
Alex Chiang 已提交
822 823
			if (!child)
				goto out;
824
			child->primary = primary;
Y
Yinghai Lu 已提交
825
			pci_bus_insert_busn_res(child, secondary, subordinate);
A
Alex Chiang 已提交
826
			child->bridge_ctl = bctl;
L
Linus Torvalds 已提交
827 828 829
		}

		cmax = pci_scan_child_bus(child);
830 831 832 833 834 835
		if (cmax > subordinate)
			dev_warn(&dev->dev, "bridge has subordinate %02x but max busn %02x\n",
				 subordinate, cmax);
		/* subordinate should equal child->busn_res.end */
		if (subordinate > max)
			max = subordinate;
L
Linus Torvalds 已提交
836 837 838 839 840
	} else {
		/*
		 * We need to assign a number to this bus which we always
		 * do in the second pass.
		 */
841
		if (!pass) {
842
			if (pcibios_assign_all_busses() || broken || is_cardbus)
843 844 845 846 847 848 849 850
				/* 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. */
				pci_write_config_dword(dev, PCI_PRIMARY_BUS,
						       buses & ~0xffffff);
851
			goto out;
852
		}
L
Linus Torvalds 已提交
853 854 855 856

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

857 858 859
		/* 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. */
860 861
		child = pci_find_bus(pci_domain_nr(bus), max+1);
		if (!child) {
862
			child = pci_add_new_bus(bus, dev, max+1);
863 864
			if (!child)
				goto out;
865
			pci_bus_insert_busn_res(child, max+1, 0xff);
866
		}
867
		max++;
L
Linus Torvalds 已提交
868 869
		buses = (buses & 0xff000000)
		      | ((unsigned int)(child->primary)     <<  0)
870 871
		      | ((unsigned int)(child->busn_res.start)   <<  8)
		      | ((unsigned int)(child->busn_res.end) << 16);
L
Linus Torvalds 已提交
872 873 874 875 876 877 878 879 880

		/*
		 * yenta.c forces a secondary latency timer of 176.
		 * Copy that behaviour here.
		 */
		if (is_cardbus) {
			buses &= ~0xff000000;
			buses |= CARDBUS_LATENCY_TIMER << 24;
		}
881

L
Linus Torvalds 已提交
882 883 884 885 886 887
		/*
		 * We need to blast all three values with a single write.
		 */
		pci_write_config_dword(dev, PCI_PRIMARY_BUS, buses);

		if (!is_cardbus) {
888
			child->bridge_ctl = bctl;
L
Linus Torvalds 已提交
889 890 891 892 893 894 895
			max = pci_scan_child_bus(child);
		} else {
			/*
			 * For CardBus bridges, we leave 4 bus numbers
			 * as cards with a PCI-to-PCI bridge can be
			 * inserted later.
			 */
R
Ryan Desfosses 已提交
896
			for (i = 0; i < CARDBUS_RESERVE_BUSNR; i++) {
897
				struct pci_bus *parent = bus;
898 899 900
				if (pci_find_bus(pci_domain_nr(bus),
							max+i+1))
					break;
901 902
				while (parent->parent) {
					if ((!pcibios_assign_all_busses()) &&
903 904
					    (parent->busn_res.end > max) &&
					    (parent->busn_res.end <= max+i)) {
905 906 907 908 909 910 911 912 913 914 915 916 917 918
						j = 1;
					}
					parent = parent->parent;
				}
				if (j) {
					/*
					 * Often, there are two cardbus bridges
					 * -- try to leave one valid bus number
					 * for each one.
					 */
					i /= 2;
					break;
				}
			}
919
			max += i;
L
Linus Torvalds 已提交
920 921 922 923
		}
		/*
		 * Set the subordinate bus number to its real value.
		 */
Y
Yinghai Lu 已提交
924
		pci_bus_update_busn_res_end(child, max);
L
Linus Torvalds 已提交
925 926 927
		pci_write_config_byte(dev, PCI_SUBORDINATE_BUS, max);
	}

928 929 930
	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
931

932
	/* Has only triggered on CardBus, fixup is in yenta_socket */
933
	while (bus->parent) {
934 935
		if ((child->busn_res.end > bus->busn_res.end) ||
		    (child->number > bus->busn_res.end) ||
936
		    (child->number < bus->number) ||
937
		    (child->busn_res.end < bus->number)) {
938
			dev_info(&child->dev, "%pR %s hidden behind%s bridge %s %pR\n",
939 940 941
				&child->busn_res,
				(bus->number > child->busn_res.end &&
				 bus->busn_res.end < child->number) ?
942 943
					"wholly" : "partially",
				bus->self->transparent ? " transparent" : "",
944
				dev_name(&bus->dev),
945
				&bus->busn_res);
946 947 948 949
		}
		bus = bus->parent;
	}

950 951 952
out:
	pci_write_config_word(dev, PCI_BRIDGE_CONTROL, bctl);

L
Linus Torvalds 已提交
953 954
	return max;
}
955
EXPORT_SYMBOL(pci_scan_bridge);
L
Linus Torvalds 已提交
956 957 958 959 960 961 962 963 964 965

/*
 * 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;

	pci_read_config_byte(dev, PCI_INTERRUPT_PIN, &irq);
966
	dev->pin = irq;
L
Linus Torvalds 已提交
967 968 969 970 971
	if (irq)
		pci_read_config_byte(dev, PCI_INTERRUPT_LINE, &irq);
	dev->irq = irq;
}

972
void set_pcie_port_type(struct pci_dev *pdev)
Y
Yu Zhao 已提交
973 974 975
{
	int pos;
	u16 reg16;
976 977
	int type;
	struct pci_dev *parent;
Y
Yu Zhao 已提交
978 979 980 981

	pos = pci_find_capability(pdev, PCI_CAP_ID_EXP);
	if (!pos)
		return;
982
	pdev->pcie_cap = pos;
Y
Yu Zhao 已提交
983
	pci_read_config_word(pdev, pos + PCI_EXP_FLAGS, &reg16);
984
	pdev->pcie_flags_reg = reg16;
985 986
	pci_read_config_word(pdev, pos + PCI_EXP_DEVCAP, &reg16);
	pdev->pcie_mpss = reg16 & PCI_EXP_DEVCAP_PAYLOAD;
987 988 989 990 991 992 993 994 995 996 997 998 999

	/*
	 * A Root Port 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.
	 */
	type = pci_pcie_type(pdev);
	if (type == PCI_EXP_TYPE_ROOT_PORT)
		pdev->has_secondary_link = 1;
	else if (type == PCI_EXP_TYPE_UPSTREAM ||
		 type == PCI_EXP_TYPE_DOWNSTREAM) {
		parent = pci_upstream_bridge(pdev);
1000 1001 1002 1003 1004 1005

		/*
		 * 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)
1006 1007
			pdev->has_secondary_link = 1;
	}
Y
Yu Zhao 已提交
1008 1009
}

1010
void set_pcie_hotplug_bridge(struct pci_dev *pdev)
1011 1012 1013
{
	u32 reg32;

1014
	pcie_capability_read_dword(pdev, PCI_EXP_SLTCAP, &reg32);
1015 1016 1017 1018
	if (reg32 & PCI_EXP_SLTCAP_HPC)
		pdev->is_hotplug_bridge = 1;
}

1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055
/**
 * pci_ext_cfg_is_aliased - is ext config space just an alias of std config?
 * @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
}

1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073
/**
 * pci_cfg_space_size - get the configuration space size of the PCI device.
 * @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)
		goto fail;
1074
	if (status == 0xffffffff || pci_ext_cfg_is_aliased(dev))
1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108
		goto fail;

	return PCI_CFG_SPACE_EXP_SIZE;

 fail:
	return PCI_CFG_SPACE_SIZE;
}

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

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

	if (!pci_is_pcie(dev)) {
		pos = pci_find_capability(dev, PCI_CAP_ID_PCIX);
		if (!pos)
			goto fail;

		pci_read_config_dword(dev, pos + PCI_X_STATUS, &status);
		if (!(status & (PCI_X_STATUS_266MHZ | PCI_X_STATUS_533MHZ)))
			goto fail;
	}

	return pci_cfg_space_size_ext(dev);

 fail:
	return PCI_CFG_SPACE_SIZE;
}

1109
#define LEGACY_IO_RESOURCE	(IORESOURCE_IO | IORESOURCE_PCI_FIXED)
1110

1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126
static void pci_msi_setup_pci_dev(struct pci_dev *dev)
{
	/*
	 * 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);
}

L
Linus Torvalds 已提交
1127 1128 1129 1130
/**
 * pci_setup_device - fill in class and map information of a device
 * @dev: the device structure to fill
 *
1131
 * Initialize the device structure with information about the device's
L
Linus Torvalds 已提交
1132 1133
 * vendor,class,memory and IO-space addresses,IRQ lines etc.
 * Called at initialisation of the PCI subsystem and by CardBus services.
Y
Yu Zhao 已提交
1134 1135
 * Returns 0 on success and negative if unknown type of device (not normal,
 * bridge or CardBus).
L
Linus Torvalds 已提交
1136
 */
Y
Yu Zhao 已提交
1137
int pci_setup_device(struct pci_dev *dev)
L
Linus Torvalds 已提交
1138 1139
{
	u32 class;
Y
Yu Zhao 已提交
1140 1141
	u8 hdr_type;
	struct pci_slot *slot;
1142
	int pos = 0;
1143 1144
	struct pci_bus_region region;
	struct resource *res;
Y
Yu Zhao 已提交
1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163

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

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

	list_for_each_entry(slot, &dev->bus->slots, list)
		if (PCI_SLOT(dev->devfn) == slot->number)
			dev->slot = slot;

	/* Assume 32-bit PCI; let 64-bit PCI cards (which are far rarer)
	   set this higher, assuming the system even supports it.  */
	dev->dma_mask = 0xffffffff;
L
Linus Torvalds 已提交
1164

1165 1166 1167
	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
1168 1169

	pci_read_config_dword(dev, PCI_CLASS_REVISION, &class);
A
Auke Kok 已提交
1170
	dev->revision = class & 0xff;
Y
Yinghai Lu 已提交
1171
	dev->class = class >> 8;		    /* upper 3 bytes */
L
Linus Torvalds 已提交
1172

Y
Yinghai Lu 已提交
1173 1174
	dev_printk(KERN_DEBUG, &dev->dev, "[%04x:%04x] type %02x class %#08x\n",
		   dev->vendor, dev->device, dev->hdr_type, dev->class);
L
Linus Torvalds 已提交
1175

1176 1177 1178
	/* need to have dev->class ready */
	dev->cfg_size = pci_cfg_space_size(dev);

L
Linus Torvalds 已提交
1179
	/* "Unknown power state" */
1180
	dev->current_state = PCI_UNKNOWN;
L
Linus Torvalds 已提交
1181

1182 1183
	pci_msi_setup_pci_dev(dev);

L
Linus Torvalds 已提交
1184 1185
	/* Early fixups, before probing the BARs */
	pci_fixup_device(pci_fixup_early, dev);
1186 1187
	/* device class may be changed after fixup */
	class = dev->class >> 8;
L
Linus Torvalds 已提交
1188 1189 1190 1191 1192 1193 1194 1195 1196

	switch (dev->hdr_type) {		    /* header type */
	case PCI_HEADER_TYPE_NORMAL:		    /* standard header */
		if (class == PCI_CLASS_BRIDGE_PCI)
			goto bad;
		pci_read_irq(dev);
		pci_read_bases(dev, 6, PCI_ROM_ADDRESS);
		pci_read_config_word(dev, PCI_SUBSYSTEM_VENDOR_ID, &dev->subsystem_vendor);
		pci_read_config_word(dev, PCI_SUBSYSTEM_ID, &dev->subsystem_device);
1197 1198

		/*
1199 1200 1201 1202
		 * 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!
1203 1204 1205 1206 1207
		 */
		if (class == PCI_CLASS_STORAGE_IDE) {
			u8 progif;
			pci_read_config_byte(dev, PCI_CLASS_PROG, &progif);
			if ((progif & 1) == 0) {
1208 1209 1210 1211
				region.start = 0x1F0;
				region.end = 0x1F7;
				res = &dev->resource[0];
				res->flags = LEGACY_IO_RESOURCE;
1212
				pcibios_bus_to_resource(dev->bus, res, &region);
1213 1214
				dev_info(&dev->dev, "legacy IDE quirk: reg 0x10: %pR\n",
					 res);
1215 1216 1217 1218
				region.start = 0x3F6;
				region.end = 0x3F6;
				res = &dev->resource[1];
				res->flags = LEGACY_IO_RESOURCE;
1219
				pcibios_bus_to_resource(dev->bus, res, &region);
1220 1221
				dev_info(&dev->dev, "legacy IDE quirk: reg 0x14: %pR\n",
					 res);
1222 1223
			}
			if ((progif & 4) == 0) {
1224 1225 1226 1227
				region.start = 0x170;
				region.end = 0x177;
				res = &dev->resource[2];
				res->flags = LEGACY_IO_RESOURCE;
1228
				pcibios_bus_to_resource(dev->bus, res, &region);
1229 1230
				dev_info(&dev->dev, "legacy IDE quirk: reg 0x18: %pR\n",
					 res);
1231 1232 1233 1234
				region.start = 0x376;
				region.end = 0x376;
				res = &dev->resource[3];
				res->flags = LEGACY_IO_RESOURCE;
1235
				pcibios_bus_to_resource(dev->bus, res, &region);
1236 1237
				dev_info(&dev->dev, "legacy IDE quirk: reg 0x1c: %pR\n",
					 res);
1238 1239
			}
		}
L
Linus Torvalds 已提交
1240 1241 1242 1243 1244 1245 1246
		break;

	case PCI_HEADER_TYPE_BRIDGE:		    /* bridge header */
		if (class != PCI_CLASS_BRIDGE_PCI)
			goto bad;
		/* The PCI-to-PCI bridge spec requires that subtractive
		   decoding (i.e. transparent) bridge must have programming
1247
		   interface code of 0x01. */
1248
		pci_read_irq(dev);
L
Linus Torvalds 已提交
1249 1250
		dev->transparent = ((dev->class & 0xff) == 1);
		pci_read_bases(dev, 2, PCI_ROM_ADDRESS1);
1251
		set_pcie_hotplug_bridge(dev);
1252 1253 1254 1255 1256
		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
1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268
		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 */
1269 1270
		dev_err(&dev->dev, "unknown header type %02x, ignoring device\n",
			dev->hdr_type);
Y
Yu Zhao 已提交
1271
		return -EIO;
L
Linus Torvalds 已提交
1272 1273

	bad:
1274 1275
		dev_err(&dev->dev, "ignoring class %#08x (doesn't match header type %02x)\n",
			dev->class, dev->hdr_type);
L
Linus Torvalds 已提交
1276 1277 1278 1279 1280 1281 1282
		dev->class = PCI_CLASS_NOT_DEFINED;
	}

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

1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294
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;

1295
	if (!hpp)
1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346
		hpp = &pci_default_type0;

	if (hpp->revision > 1) {
		dev_warn(&dev->dev,
			 "PCI settings rev %d not supported; using defaults\n",
			 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)
{
	if (hpp)
		dev_warn(&dev->dev, "PCI-X settings not supported\n");
}

static void program_hpp_type2(struct pci_dev *dev, struct hpp_type2 *hpp)
{
	int pos;
	u32 reg32;

	if (!hpp)
		return;

	if (hpp->revision > 1) {
		dev_warn(&dev->dev, "PCIe settings rev %d not supported\n",
			 hpp->revision);
		return;
	}

1347 1348 1349 1350 1351 1352 1353 1354 1355 1356
	/*
	 * 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);

1357 1358 1359 1360 1361
	/* 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 */
1362
	if (pcie_cap_has_lnkctl(dev))
1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398
		pcie_capability_clear_and_set_word(dev, PCI_EXP_LNKCTL,
			~hpp->pci_exp_lnkctl_and, hpp->pci_exp_lnkctl_or);

	/* 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;
	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
	 */
}

1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413
static void pci_configure_device(struct pci_dev *dev)
{
	struct hotplug_params hpp;
	int ret;

	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);
}

1414 1415 1416
static void pci_release_capabilities(struct pci_dev *dev)
{
	pci_vpd_release(dev);
1417
	pci_iov_release(dev);
1418
	pci_free_cap_save_buffers(dev);
1419 1420
}

L
Linus Torvalds 已提交
1421 1422 1423 1424 1425 1426 1427 1428 1429
/**
 * pci_release_dev - free a pci device structure when all users of it are finished.
 * @dev: device that's been disconnected
 *
 * Will be called only by the device core when all users of this pci device are
 * done.
 */
static void pci_release_dev(struct device *dev)
{
1430
	struct pci_dev *pci_dev;
L
Linus Torvalds 已提交
1431

1432
	pci_dev = to_pci_dev(dev);
1433
	pci_release_capabilities(pci_dev);
1434
	pci_release_of_node(pci_dev);
1435
	pcibios_release_device(pci_dev);
1436
	pci_bus_put(pci_dev->bus);
1437
	kfree(pci_dev->driver_override);
L
Linus Torvalds 已提交
1438 1439 1440
	kfree(pci_dev);
}

1441
struct pci_dev *pci_alloc_dev(struct pci_bus *bus)
1442 1443 1444 1445 1446 1447 1448 1449
{
	struct pci_dev *dev;

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

	INIT_LIST_HEAD(&dev->bus_list);
1450
	dev->dev.type = &pci_dev_type;
1451
	dev->bus = pci_bus_get(bus);
1452 1453 1454

	return dev;
}
1455 1456
EXPORT_SYMBOL(pci_alloc_dev);

1457
bool pci_bus_read_dev_vendor_id(struct pci_bus *bus, int devfn, u32 *l,
R
Ryan Desfosses 已提交
1458
				int crs_timeout)
L
Linus Torvalds 已提交
1459 1460 1461
{
	int delay = 1;

1462 1463
	if (pci_bus_read_config_dword(bus, devfn, PCI_VENDOR_ID, l))
		return false;
L
Linus Torvalds 已提交
1464 1465

	/* some broken boards return 0 or ~0 if a slot is empty: */
1466 1467 1468
	if (*l == 0xffffffff || *l == 0x00000000 ||
	    *l == 0x0000ffff || *l == 0xffff0000)
		return false;
L
Linus Torvalds 已提交
1469

1470 1471 1472 1473 1474 1475 1476
	/*
	 * Configuration Request Retry Status.  Some root ports return the
	 * actual device ID instead of the synthetic ID (0xFFFF) required
	 * by the PCIe spec.  Ignore the device ID and only check for
	 * (vendor id == 1).
	 */
	while ((*l & 0xffff) == 0x0001) {
1477 1478 1479
		if (!crs_timeout)
			return false;

L
Linus Torvalds 已提交
1480 1481
		msleep(delay);
		delay *= 2;
1482 1483
		if (pci_bus_read_config_dword(bus, devfn, PCI_VENDOR_ID, l))
			return false;
L
Linus Torvalds 已提交
1484
		/* Card hasn't responded in 60 seconds?  Must be stuck. */
1485
		if (delay > crs_timeout) {
1486 1487 1488
			printk(KERN_WARNING "pci %04x:%02x:%02x.%d: not responding\n",
			       pci_domain_nr(bus), bus->number, PCI_SLOT(devfn),
			       PCI_FUNC(devfn));
1489
			return false;
L
Linus Torvalds 已提交
1490 1491 1492
		}
	}

1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508
	return true;
}
EXPORT_SYMBOL(pci_bus_read_dev_vendor_id);

/*
 * Read the config data for a PCI device, sanity-check it
 * and fill in the dev structure...
 */
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;

1509
	dev = pci_alloc_dev(bus);
L
Linus Torvalds 已提交
1510 1511 1512 1513 1514 1515
	if (!dev)
		return NULL;

	dev->devfn = devfn;
	dev->vendor = l & 0xffff;
	dev->device = (l >> 16) & 0xffff;
1516

1517 1518
	pci_set_of_node(dev);

Y
Yu Zhao 已提交
1519
	if (pci_setup_device(dev)) {
1520
		pci_bus_put(dev->bus);
L
Linus Torvalds 已提交
1521 1522 1523 1524 1525 1526 1527
		kfree(dev);
		return NULL;
	}

	return dev;
}

1528 1529 1530 1531 1532
static void pci_init_capabilities(struct pci_dev *dev)
{
	/* MSI/MSI-X list */
	pci_msi_init_pci_dev(dev);

1533 1534 1535
	/* Buffers for saving PCIe and PCI-X capabilities */
	pci_allocate_cap_save_buffers(dev);

1536 1537 1538 1539 1540
	/* Power Management */
	pci_pm_init(dev);

	/* Vital Product Data */
	pci_vpd_pci22_init(dev);
Y
Yu Zhao 已提交
1541 1542

	/* Alternative Routing-ID Forwarding */
1543
	pci_configure_ari(dev);
1544 1545 1546

	/* Single Root I/O Virtualization */
	pci_iov_init(dev);
1547 1548

	/* Enable ACS P2P upstream forwarding */
C
Chris Wright 已提交
1549
	pci_enable_acs(dev);
1550 1551
}

1552
void pci_device_add(struct pci_dev *dev, struct pci_bus *bus)
L
Linus Torvalds 已提交
1553
{
1554 1555
	int ret;

1556 1557
	pci_configure_device(dev);

1558 1559
	device_initialize(&dev->dev);
	dev->dev.release = pci_release_dev;
L
Linus Torvalds 已提交
1560

1561
	set_dev_node(&dev->dev, pcibus_to_node(bus));
1562
	dev->dev.dma_mask = &dev->dma_mask;
1563
	dev->dev.dma_parms = &dev->dma_parms;
1564
	dev->dev.coherent_dma_mask = 0xffffffffull;
1565
	of_pci_dma_configure(dev);
L
Linus Torvalds 已提交
1566

1567
	pci_set_dma_max_seg_size(dev, 65536);
1568
	pci_set_dma_seg_boundary(dev, 0xffffffff);
1569

L
Linus Torvalds 已提交
1570 1571 1572
	/* Fix up broken headers */
	pci_fixup_device(pci_fixup_header, dev);

1573 1574 1575
	/* moved out from quirk header fixup code */
	pci_reassigndev_resource_alignment(dev);

1576 1577 1578
	/* Clear the state_saved flag. */
	dev->state_saved = false;

1579 1580
	/* Initialize various capabilities */
	pci_init_capabilities(dev);
1581

L
Linus Torvalds 已提交
1582 1583 1584 1585
	/*
	 * Add the device to our list of discovered devices
	 * and the bus list for fixup functions, etc.
	 */
1586
	down_write(&pci_bus_sem);
L
Linus Torvalds 已提交
1587
	list_add_tail(&dev->bus_list, &bus->devices);
1588
	up_write(&pci_bus_sem);
1589 1590 1591 1592 1593 1594 1595 1596

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

	/* Notifier could use PCI capabilities */
	dev->match_driver = false;
	ret = device_add(&dev->dev);
	WARN_ON(ret < 0);
1597 1598
}

1599
struct pci_dev *pci_scan_single_device(struct pci_bus *bus, int devfn)
1600 1601 1602
{
	struct pci_dev *dev;

T
Trent Piepho 已提交
1603 1604 1605 1606 1607 1608
	dev = pci_get_slot(bus, devfn);
	if (dev) {
		pci_dev_put(dev);
		return dev;
	}

1609 1610 1611 1612 1613
	dev = pci_scan_device(bus, devfn);
	if (!dev)
		return NULL;

	pci_device_add(dev, bus);
L
Linus Torvalds 已提交
1614 1615 1616

	return dev;
}
1617
EXPORT_SYMBOL(pci_scan_single_device);
L
Linus Torvalds 已提交
1618

1619
static unsigned next_fn(struct pci_bus *bus, struct pci_dev *dev, unsigned fn)
M
Matthew Wilcox 已提交
1620
{
1621 1622 1623
	int pos;
	u16 cap = 0;
	unsigned next_fn;
1624

1625 1626 1627 1628 1629 1630
	if (pci_ari_enabled(bus)) {
		if (!dev)
			return 0;
		pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ARI);
		if (!pos)
			return 0;
1631

1632 1633 1634 1635
		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
1636

1637 1638 1639 1640 1641 1642
		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
1643 1644 1645 1646 1647 1648 1649

	return 0;
}

static int only_one_child(struct pci_bus *bus)
{
	struct pci_dev *parent = bus->self;
1650

M
Matthew Wilcox 已提交
1651 1652
	if (!parent || !pci_is_pcie(parent))
		return 0;
1653
	if (pci_pcie_type(parent) == PCI_EXP_TYPE_ROOT_PORT)
1654
		return 1;
1655
	if (parent->has_secondary_link &&
1656
	    !pci_has_flag(PCI_SCAN_ALL_PCIE_DEVS))
M
Matthew Wilcox 已提交
1657 1658 1659 1660
		return 1;
	return 0;
}

L
Linus Torvalds 已提交
1661 1662 1663 1664 1665 1666 1667
/**
 * pci_scan_slot - scan a PCI slot on a bus for devices.
 * @bus: PCI bus to scan
 * @devfn: slot number to scan (must have zero function.)
 *
 * Scan a PCI slot on the specified PCI bus for devices, adding
 * discovered devices to the @bus->devices list.  New devices
1668
 * will not have is_added set.
1669 1670
 *
 * Returns the number of new devices found.
L
Linus Torvalds 已提交
1671
 */
1672
int pci_scan_slot(struct pci_bus *bus, int devfn)
L
Linus Torvalds 已提交
1673
{
M
Matthew Wilcox 已提交
1674
	unsigned fn, nr = 0;
1675
	struct pci_dev *dev;
M
Matthew Wilcox 已提交
1676 1677 1678

	if (only_one_child(bus) && (devfn > 0))
		return 0; /* Already scanned the entire slot */
L
Linus Torvalds 已提交
1679

1680
	dev = pci_scan_single_device(bus, devfn);
1681 1682 1683
	if (!dev)
		return 0;
	if (!dev->is_added)
1684 1685
		nr++;

1686
	for (fn = next_fn(bus, dev, 0); fn > 0; fn = next_fn(bus, dev, fn)) {
M
Matthew Wilcox 已提交
1687 1688 1689 1690 1691
		dev = pci_scan_single_device(bus, devfn + fn);
		if (dev) {
			if (!dev->is_added)
				nr++;
			dev->multifunction = 1;
L
Linus Torvalds 已提交
1692 1693
		}
	}
S
Shaohua Li 已提交
1694

1695 1696
	/* only one slot has pcie device */
	if (bus->self && nr)
S
Shaohua Li 已提交
1697 1698
		pcie_aspm_init_link_state(bus->self);

L
Linus Torvalds 已提交
1699 1700
	return nr;
}
1701
EXPORT_SYMBOL(pci_scan_slot);
L
Linus Torvalds 已提交
1702

1703 1704 1705 1706 1707 1708 1709
static int pcie_find_smpss(struct pci_dev *dev, void *data)
{
	u8 *smpss = data;

	if (!pci_is_pcie(dev))
		return 0;

1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723
	/*
	 * 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.
1724
	 */
1725 1726
	if (dev->is_hotplug_bridge &&
	    pci_pcie_type(dev) != PCI_EXP_TYPE_ROOT_PORT)
1727 1728 1729 1730 1731 1732 1733 1734 1735 1736
		*smpss = 0;

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

	return 0;
}

static void pcie_write_mps(struct pci_dev *dev, int mps)
{
1737
	int rc;
1738 1739

	if (pcie_bus_config == PCIE_BUS_PERFORMANCE) {
1740
		mps = 128 << dev->pcie_mpss;
1741

1742 1743
		if (pci_pcie_type(dev) != PCI_EXP_TYPE_ROOT_PORT &&
		    dev->bus->self)
1744
			/* For "Performance", the assumption is made that
1745 1746 1747 1748 1749
			 * 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.
1750 1751 1752 1753 1754
			 *
			 * 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).
1755
			 */
1756
			mps = min(mps, pcie_get_mps(dev->bus->self));
1757 1758 1759 1760 1761 1762 1763
	}

	rc = pcie_set_mps(dev, mps);
	if (rc)
		dev_err(&dev->dev, "Failed attempting to set the MPS\n");
}

1764
static void pcie_write_mrrs(struct pci_dev *dev)
1765
{
1766
	int rc, mrrs;
1767

1768 1769 1770 1771 1772 1773 1774 1775
	/* In the "safe" case, do not configure the MRRS.  There appear to be
	 * issues with setting MRRS to 0 on a number of devices.
	 */
	if (pcie_bus_config != PCIE_BUS_PERFORMANCE)
		return;

	/* For Max performance, the MRRS must be set to the largest supported
	 * value.  However, it cannot be configured larger than the MPS the
1776 1777
	 * device or the bus can support.  This should already be properly
	 * configured by a prior call to pcie_write_mps.
1778
	 */
1779
	mrrs = pcie_get_mps(dev);
1780 1781

	/* MRRS is a R/W register.  Invalid values can be written, but a
1782
	 * subsequent read will verify if the value is acceptable or not.
1783 1784
	 * If the MRRS value provided is not acceptable (e.g., too large),
	 * shrink the value until it is acceptable to the HW.
1785
	 */
1786 1787
	while (mrrs != pcie_get_readrq(dev) && mrrs >= 128) {
		rc = pcie_set_readrq(dev, mrrs);
1788 1789
		if (!rc)
			break;
1790

1791
		dev_warn(&dev->dev, "Failed attempting to set the MRRS\n");
1792 1793
		mrrs /= 2;
	}
1794 1795

	if (mrrs < 128)
1796
		dev_err(&dev->dev, "MRRS was unable to be configured with a safe value.  If problems are experienced, try running with pci=pcie_bus_safe\n");
1797 1798
}

1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814
static void pcie_bus_detect_mps(struct pci_dev *dev)
{
	struct pci_dev *bridge = dev->bus->self;
	int mps, p_mps;

	if (!bridge)
		return;

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

	if (mps != p_mps)
		dev_warn(&dev->dev, "Max Payload Size %d, but upstream %s set to %d; if necessary, use \"pci=pcie_bus_safe\" and report a bug\n",
			 mps, pci_name(bridge), p_mps);
}

1815 1816
static int pcie_bus_configure_set(struct pci_dev *dev, void *data)
{
J
Jon Mason 已提交
1817
	int mps, orig_mps;
1818 1819 1820 1821

	if (!pci_is_pcie(dev))
		return 0;

1822 1823 1824 1825 1826
	if (pcie_bus_config == PCIE_BUS_TUNE_OFF) {
		pcie_bus_detect_mps(dev);
		return 0;
	}

J
Jon Mason 已提交
1827 1828
	mps = 128 << *(u8 *)data;
	orig_mps = pcie_get_mps(dev);
1829 1830

	pcie_write_mps(dev, mps);
1831
	pcie_write_mrrs(dev);
1832

1833 1834
	dev_info(&dev->dev, "Max Payload Size set to %4d/%4d (was %4d), Max Read Rq %4d\n",
		 pcie_get_mps(dev), 128 << dev->pcie_mpss,
J
Jon Mason 已提交
1835
		 orig_mps, pcie_get_readrq(dev));
1836 1837 1838 1839

	return 0;
}

J
Jon Mason 已提交
1840
/* pcie_bus_configure_settings requires that pci_walk_bus work in a top-down,
1841 1842 1843
 * parents then children fashion.  If this changes, then this code will not
 * work as designed.
 */
1844
void pcie_bus_configure_settings(struct pci_bus *bus)
1845
{
1846
	u8 smpss = 0;
1847

1848
	if (!bus->self)
1849 1850 1851
		return;

	if (!pci_is_pcie(bus->self))
1852 1853 1854
		return;

	/* FIXME - Peer to peer DMA is possible, though the endpoint would need
1855
	 * to be aware of the MPS of the destination.  To work around this,
1856 1857 1858 1859 1860
	 * simply force the MPS of the entire system to the smallest possible.
	 */
	if (pcie_bus_config == PCIE_BUS_PEER2PEER)
		smpss = 0;

1861
	if (pcie_bus_config == PCIE_BUS_SAFE) {
1862
		smpss = bus->self->pcie_mpss;
1863

1864 1865 1866 1867 1868 1869 1870
		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);
}
1871
EXPORT_SYMBOL_GPL(pcie_bus_configure_settings);
1872

B
Bill Pemberton 已提交
1873
unsigned int pci_scan_child_bus(struct pci_bus *bus)
L
Linus Torvalds 已提交
1874
{
1875
	unsigned int devfn, pass, max = bus->busn_res.start;
L
Linus Torvalds 已提交
1876 1877
	struct pci_dev *dev;

B
Bjorn Helgaas 已提交
1878
	dev_dbg(&bus->dev, "scanning bus\n");
L
Linus Torvalds 已提交
1879 1880 1881 1882 1883

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

1884 1885 1886
	/* Reserve buses for SR-IOV capability. */
	max += pci_iov_bus_range(bus);

L
Linus Torvalds 已提交
1887 1888 1889 1890
	/*
	 * After performing arch-dependent fixup of the bus, look behind
	 * all PCI-to-PCI bridges on this bus.
	 */
A
Alex Chiang 已提交
1891
	if (!bus->is_added) {
B
Bjorn Helgaas 已提交
1892
		dev_dbg(&bus->dev, "fixups for bus\n");
A
Alex Chiang 已提交
1893
		pcibios_fixup_bus(bus);
1894
		bus->is_added = 1;
A
Alex Chiang 已提交
1895 1896
	}

R
Ryan Desfosses 已提交
1897
	for (pass = 0; pass < 2; pass++)
L
Linus Torvalds 已提交
1898
		list_for_each_entry(dev, &bus->devices, bus_list) {
1899
			if (pci_is_bridge(dev))
L
Linus Torvalds 已提交
1900 1901 1902 1903 1904 1905 1906 1907 1908 1909
				max = pci_scan_bridge(bus, dev, max, pass);
		}

	/*
	 * 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
1910
	dev_dbg(&bus->dev, "bus scan returning with max=%02x\n", max);
L
Linus Torvalds 已提交
1911 1912
	return max;
}
1913
EXPORT_SYMBOL_GPL(pci_scan_child_bus);
L
Linus Torvalds 已提交
1914

1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926
/**
 * pcibios_root_bridge_prepare - Platform-specific host bridge setup.
 * @bridge: Host bridge to set up.
 *
 * 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;
}

1927 1928 1929 1930 1931 1932 1933 1934
void __weak pcibios_add_bus(struct pci_bus *bus)
{
}

void __weak pcibios_remove_bus(struct pci_bus *bus)
{
}

1935 1936
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
1937
{
1938
	int error;
1939
	struct pci_host_bridge *bridge;
B
Bjorn Helgaas 已提交
1940
	struct pci_bus *b, *b2;
1941
	struct resource_entry *window, *n;
1942
	struct resource *res;
1943 1944 1945
	resource_size_t offset;
	char bus_addr[64];
	char *fmt;
L
Linus Torvalds 已提交
1946

1947
	b = pci_alloc_bus(NULL);
L
Linus Torvalds 已提交
1948
	if (!b)
1949
		return NULL;
L
Linus Torvalds 已提交
1950 1951 1952

	b->sysdata = sysdata;
	b->ops = ops;
1953
	b->number = b->busn_res.start = bus;
1954
	pci_bus_assign_domain_nr(b, parent);
B
Bjorn Helgaas 已提交
1955 1956
	b2 = pci_find_bus(pci_domain_nr(b), bus);
	if (b2) {
L
Linus Torvalds 已提交
1957
		/* If we already got to this bus through a different bridge, ignore it */
B
Bjorn Helgaas 已提交
1958
		dev_dbg(&b2->dev, "bus already known\n");
L
Linus Torvalds 已提交
1959 1960
		goto err_out;
	}
1961

1962 1963 1964 1965 1966
	bridge = pci_alloc_host_bridge(b);
	if (!bridge)
		goto err_out;

	bridge->dev.parent = parent;
1967
	bridge->dev.release = pci_release_host_bridge_dev;
1968
	dev_set_name(&bridge->dev, "pci%04x:%02x", pci_domain_nr(b), bus);
1969
	error = pcibios_root_bridge_prepare(bridge);
1970 1971 1972 1973
	if (error) {
		kfree(bridge);
		goto err_out;
	}
1974

1975
	error = device_register(&bridge->dev);
1976 1977 1978 1979
	if (error) {
		put_device(&bridge->dev);
		goto err_out;
	}
1980
	b->bridge = get_device(&bridge->dev);
1981
	device_enable_async_suspend(b->bridge);
1982
	pci_set_bus_of_node(b);
L
Linus Torvalds 已提交
1983

1984 1985 1986
	if (!parent)
		set_dev_node(b->bridge, pcibus_to_node(b));

1987 1988
	b->dev.class = &pcibus_class;
	b->dev.parent = b->bridge;
1989
	dev_set_name(&b->dev, "%04x:%02x", pci_domain_nr(b), bus);
1990
	error = device_register(&b->dev);
L
Linus Torvalds 已提交
1991 1992 1993
	if (error)
		goto class_dev_reg_err;

1994 1995
	pcibios_add_bus(b);

L
Linus Torvalds 已提交
1996 1997 1998
	/* Create legacy_io and legacy_mem files for this bus */
	pci_create_legacy_files(b);

1999 2000 2001 2002 2003
	if (parent)
		dev_info(parent, "PCI host bridge to bus %s\n", dev_name(&b->dev));
	else
		printk(KERN_INFO "PCI host bridge to bus %s\n", dev_name(&b->dev));

2004
	/* Add initial resources to the bus */
2005 2006
	resource_list_for_each_entry_safe(window, n, resources) {
		list_move_tail(&window->node, &bridge->windows);
2007 2008
		res = window->res;
		offset = window->offset;
2009 2010 2011 2012
		if (res->flags & IORESOURCE_BUS)
			pci_bus_insert_busn_res(b, bus, res->end);
		else
			pci_bus_add_resource(b, res, 0);
2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023
		if (offset) {
			if (resource_type(res) == IORESOURCE_IO)
				fmt = " (bus address [%#06llx-%#06llx])";
			else
				fmt = " (bus address [%#010llx-%#010llx])";
			snprintf(bus_addr, sizeof(bus_addr), fmt,
				 (unsigned long long) (res->start - offset),
				 (unsigned long long) (res->end - offset));
		} else
			bus_addr[0] = '\0';
		dev_info(&b->dev, "root bus resource %pR%s\n", res, bus_addr);
2024 2025
	}

2026 2027 2028 2029
	down_write(&pci_bus_sem);
	list_add_tail(&b->node, &pci_root_buses);
	up_write(&pci_bus_sem);

L
Linus Torvalds 已提交
2030 2031 2032
	return b;

class_dev_reg_err:
2033 2034
	put_device(&bridge->dev);
	device_unregister(&bridge->dev);
L
Linus Torvalds 已提交
2035 2036 2037 2038
err_out:
	kfree(b);
	return NULL;
}
2039
EXPORT_SYMBOL_GPL(pci_create_root_bus);
2040

2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056
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;
	}

2057
	conflict = request_resource_conflict(parent_res, res);
2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103

	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");
}

B
Bill Pemberton 已提交
2104
struct pci_bus *pci_scan_root_bus(struct device *parent, int bus,
2105 2106
		struct pci_ops *ops, void *sysdata, struct list_head *resources)
{
2107
	struct resource_entry *window;
2108
	bool found = false;
2109
	struct pci_bus *b;
2110 2111
	int max;

2112
	resource_list_for_each_entry(window, resources)
2113 2114 2115 2116
		if (window->res->flags & IORESOURCE_BUS) {
			found = true;
			break;
		}
2117 2118 2119 2120 2121

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

2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133
	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);

2134 2135 2136 2137
	return b;
}
EXPORT_SYMBOL(pci_scan_root_bus);

B
Bill Pemberton 已提交
2138
struct pci_bus *pci_scan_bus(int bus, struct pci_ops *ops,
2139 2140 2141 2142 2143 2144 2145
					void *sysdata)
{
	LIST_HEAD(resources);
	struct pci_bus *b;

	pci_add_resource(&resources, &ioport_resource);
	pci_add_resource(&resources, &iomem_resource);
2146
	pci_add_resource(&resources, &busn_resource);
2147 2148
	b = pci_create_root_bus(NULL, bus, ops, sysdata, &resources);
	if (b) {
2149
		pci_scan_child_bus(b);
2150 2151 2152 2153 2154 2155 2156
	} else {
		pci_free_resource_list(&resources);
	}
	return b;
}
EXPORT_SYMBOL(pci_scan_bus);

2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167
/**
 * pci_rescan_bus_bridge_resize - scan a PCI bus for devices.
 * @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.
 */
2168
unsigned int pci_rescan_bus_bridge_resize(struct pci_dev *bridge)
2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181
{
	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;
}

2182 2183 2184 2185 2186 2187 2188 2189 2190
/**
 * pci_rescan_bus - scan a PCI bus for devices.
 * @bus: PCI bus to scan
 *
 * Scan a PCI bus and child buses for new devices, adds them,
 * and enables them.
 *
 * Returns the max number of subordinate bus discovered.
 */
2191
unsigned int pci_rescan_bus(struct pci_bus *bus)
2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202
{
	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);

2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220
/*
 * 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
2221 2222
static int __init pci_sort_bf_cmp(const struct device *d_a,
				  const struct device *d_b)
2223
{
2224 2225 2226
	const struct pci_dev *a = to_pci_dev(d_a);
	const struct pci_dev *b = to_pci_dev(d_b);

2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238
	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;
}

2239
void __init pci_sort_breadthfirst(void)
2240
{
2241
	bus_sort_breadthfirst(&pci_bus_type, &pci_sort_bf_cmp);
2242
}