probe.c 57.3 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/pci_hotplug.h>
L
Linus Torvalds 已提交
10 11 12
#include <linux/slab.h>
#include <linux/module.h>
#include <linux/cpumask.h>
S
Shaohua Li 已提交
13
#include <linux/pci-aspm.h>
14
#include <asm-generic/pci-bridge.h>
15
#include "pci.h"
L
Linus Torvalds 已提交
16 17 18 19

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

20
static struct resource busn_resource = {
Y
Yinghai Lu 已提交
21 22 23 24 25 26
	.name	= "PCI busn",
	.start	= 0,
	.end	= 255,
	.flags	= IORESOURCE_BUS,
};

L
Linus Torvalds 已提交
27 28 29 30
/* Ugh.  Need to stop exporting this to modules. */
LIST_HEAD(pci_root_buses);
EXPORT_SYMBOL(pci_root_buses);

31 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
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;
}

61 62 63 64
static int find_anything(struct device *dev, void *data)
{
	return 1;
}
L
Linus Torvalds 已提交
65

Z
Zhang, Yanmin 已提交
66 67 68
/*
 * Some device drivers need know if pci is initiated.
 * Basically, we think pci is not initiated when there
69
 * is no device to be found on the pci_bus_type.
Z
Zhang, Yanmin 已提交
70 71 72
 */
int no_pci_devices(void)
{
73 74
	struct device *dev;
	int no_devices;
Z
Zhang, Yanmin 已提交
75

76 77 78 79 80
	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
81 82
EXPORT_SYMBOL(no_pci_devices);

L
Linus Torvalds 已提交
83 84 85
/*
 * PCI Bus Class
 */
86
static void release_pcibus_dev(struct device *dev)
L
Linus Torvalds 已提交
87
{
88
	struct pci_bus *pci_bus = to_pci_bus(dev);
L
Linus Torvalds 已提交
89 90 91

	if (pci_bus->bridge)
		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
	bool bar_too_big = false, bar_too_high = false, bar_invalid = false;
179

180
	mask = type ? PCI_ROM_ADDRESS_MASK : ~0;
181

182
	/* No printks while decoding is disabled! */
J
Jacob Pan 已提交
183 184
	if (!dev->mmio_always_on) {
		pci_read_config_word(dev, PCI_COMMAND, &orig_cmd);
185 186 187 188
		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
189 190
	}

191 192 193
	res->name = pci_name(dev);

	pci_read_config_dword(dev, pos, &l);
194
	pci_write_config_dword(dev, pos, l | mask);
195 196 197 198 199
	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.
200 201 202
	 * 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.
203
	 */
204
	if (!sz || sz == 0xffffffff)
205 206 207 208 209 210 211 212 213 214
		goto fail;

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

230
	if (res->flags & IORESOURCE_MEM_64) {
231 232 233
		l64 = l;
		sz64 = sz;
		mask64 = mask | (u64)~0 << 32;
234 235 236 237 238 239 240 241 242 243 244 245 246 247

		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);

		sz64 = pci_size(l64, sz64, mask64);

		if (!sz64)
			goto fail;

248 249 250 251 252
		if ((sizeof(dma_addr_t) < 8 || sizeof(resource_size_t) < 8) &&
		    sz64 > 0x100000000ULL) {
			res->flags |= IORESOURCE_UNSET | IORESOURCE_DISABLED;
			res->start = 0;
			res->end = 0;
253
			bar_too_big = true;
254
			goto out;
255 256
		}

257
		if ((sizeof(dma_addr_t) < 8) && l) {
258
			/* Above 32-bit boundary; try to reallocate */
259
			res->flags |= IORESOURCE_UNSET;
260 261
			res->start = 0;
			res->end = sz64;
262
			bar_too_high = true;
263
			goto out;
264
		} else {
265 266
			region.start = l64;
			region.end = l64 + sz64;
267 268
		}
	} else {
269
		sz = pci_size(l, sz, mask);
270

271
		if (!sz)
272 273
			goto fail;

274 275
		region.start = l;
		region.end = l + sz;
276 277
	}

278 279
	pcibios_bus_to_resource(dev->bus, res, &region);
	pcibios_resource_to_bus(dev->bus, &inverted_region, res);
280 281 282 283 284 285 286 287 288 289 290 291 292 293 294

	/*
	 * 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;
295 296
		res->end = region.end - region.start;
		bar_invalid = true;
297
	}
298

299 300 301 302 303 304
	goto out;


fail:
	res->flags = 0;
out:
305 306
	if (!dev->mmio_always_on &&
	    (orig_cmd & PCI_COMMAND_DECODE_ENABLE))
307 308
		pci_write_config_word(dev, PCI_COMMAND, orig_cmd);

309
	if (bar_too_big)
310 311
		dev_err(&dev->dev, "reg 0x%x: can't handle BAR larger than 4GB (size %#010llx)\n",
			pos, (unsigned long long) sz64);
312 313 314
	if (bar_too_high)
		dev_info(&dev->dev, "reg 0x%x: can't handle BAR above 4G (bus address %#010llx)\n",
			 pos, (unsigned long long) l64);
315 316 317
	if (bar_invalid)
		dev_info(&dev->dev, "reg 0x%x: initial BAR value %#010llx invalid\n",
			 pos, (unsigned long long) region.start);
318
	if (res->flags)
319
		dev_printk(KERN_DEBUG, &dev->dev, "reg 0x%x: %pR\n", pos, res);
320

321
	return (res->flags & IORESOURCE_MEM_64) ? 1 : 0;
322 323
}

L
Linus Torvalds 已提交
324 325
static void pci_read_bases(struct pci_dev *dev, unsigned int howmany, int rom)
{
326
	unsigned int pos, reg;
327

328 329
	for (pos = 0; pos < howmany; pos++) {
		struct resource *res = &dev->resource[pos];
L
Linus Torvalds 已提交
330
		reg = PCI_BASE_ADDRESS_0 + (pos << 2);
331
		pos += __pci_read_base(dev, pci_bar_unknown, res, reg);
L
Linus Torvalds 已提交
332
	}
333

L
Linus Torvalds 已提交
334
	if (rom) {
335
		struct resource *res = &dev->resource[PCI_ROM_RESOURCE];
L
Linus Torvalds 已提交
336
		dev->rom_base_reg = rom;
337 338 339 340
		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
341 342 343
	}
}

B
Bill Pemberton 已提交
344
static void pci_read_bridge_io(struct pci_bus *child)
L
Linus Torvalds 已提交
345 346 347
{
	struct pci_dev *dev = child->self;
	u8 io_base_lo, io_limit_lo;
348
	unsigned long io_mask, io_granularity, base, limit;
349
	struct pci_bus_region region;
350 351 352 353 354 355 356 357 358
	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
359 360 361 362

	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);
363 364
	base = (io_base_lo & io_mask) << 8;
	limit = (io_limit_lo & io_mask) << 8;
L
Linus Torvalds 已提交
365 366 367

	if ((io_base_lo & PCI_IO_RANGE_TYPE_MASK) == PCI_IO_RANGE_TYPE_32) {
		u16 io_base_hi, io_limit_hi;
368

L
Linus Torvalds 已提交
369 370
		pci_read_config_word(dev, PCI_IO_BASE_UPPER16, &io_base_hi);
		pci_read_config_word(dev, PCI_IO_LIMIT_UPPER16, &io_limit_hi);
371 372
		base |= ((unsigned long) io_base_hi << 16);
		limit |= ((unsigned long) io_limit_hi << 16);
L
Linus Torvalds 已提交
373 374
	}

375
	if (base <= limit) {
L
Linus Torvalds 已提交
376
		res->flags = (io_base_lo & PCI_IO_RANGE_TYPE_MASK) | IORESOURCE_IO;
377
		region.start = base;
378
		region.end = limit + io_granularity - 1;
379
		pcibios_bus_to_resource(dev->bus, res, &region);
380
		dev_printk(KERN_DEBUG, &dev->dev, "  bridge window %pR\n", res);
L
Linus Torvalds 已提交
381
	}
382 383
}

B
Bill Pemberton 已提交
384
static void pci_read_bridge_mmio(struct pci_bus *child)
385 386 387 388
{
	struct pci_dev *dev = child->self;
	u16 mem_base_lo, mem_limit_lo;
	unsigned long base, limit;
389
	struct pci_bus_region region;
390
	struct resource *res;
L
Linus Torvalds 已提交
391 392 393 394

	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);
395 396
	base = ((unsigned long) mem_base_lo & PCI_MEMORY_RANGE_MASK) << 16;
	limit = ((unsigned long) mem_limit_lo & PCI_MEMORY_RANGE_MASK) << 16;
397
	if (base <= limit) {
L
Linus Torvalds 已提交
398
		res->flags = (mem_base_lo & PCI_MEMORY_RANGE_TYPE_MASK) | IORESOURCE_MEM;
399 400
		region.start = base;
		region.end = limit + 0xfffff;
401
		pcibios_bus_to_resource(dev->bus, res, &region);
402
		dev_printk(KERN_DEBUG, &dev->dev, "  bridge window %pR\n", res);
L
Linus Torvalds 已提交
403
	}
404 405
}

B
Bill Pemberton 已提交
406
static void pci_read_bridge_mmio_pref(struct pci_bus *child)
407 408 409 410
{
	struct pci_dev *dev = child->self;
	u16 mem_base_lo, mem_limit_lo;
	unsigned long base, limit;
411
	struct pci_bus_region region;
412
	struct resource *res;
L
Linus Torvalds 已提交
413 414 415 416

	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);
417 418
	base = ((unsigned long) mem_base_lo & PCI_PREF_RANGE_MASK) << 16;
	limit = ((unsigned long) mem_limit_lo & PCI_PREF_RANGE_MASK) << 16;
L
Linus Torvalds 已提交
419 420 421

	if ((mem_base_lo & PCI_PREF_RANGE_TYPE_MASK) == PCI_PREF_RANGE_TYPE_64) {
		u32 mem_base_hi, mem_limit_hi;
422

L
Linus Torvalds 已提交
423 424 425 426 427 428 429 430 431 432
		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) {
#if BITS_PER_LONG == 64
433 434
			base |= ((unsigned long) mem_base_hi) << 32;
			limit |= ((unsigned long) mem_limit_hi) << 32;
L
Linus Torvalds 已提交
435 436
#else
			if (mem_base_hi || mem_limit_hi) {
437
				dev_err(&dev->dev, "can't handle 64-bit address space for bridge\n");
L
Linus Torvalds 已提交
438 439 440 441 442
				return;
			}
#endif
		}
	}
443
	if (base <= limit) {
444 445 446 447
		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;
448 449
		region.start = base;
		region.end = limit + 0xfffff;
450
		pcibios_bus_to_resource(dev->bus, res, &region);
451
		dev_printk(KERN_DEBUG, &dev->dev, "  bridge window %pR\n", res);
L
Linus Torvalds 已提交
452 453 454
	}
}

B
Bill Pemberton 已提交
455
void pci_read_bridge_bases(struct pci_bus *child)
456 457
{
	struct pci_dev *dev = child->self;
458
	struct resource *res;
459 460 461 462 463
	int i;

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

464 465
	dev_info(&dev->dev, "PCI bridge to %pR%s\n",
		 &child->busn_res,
466 467
		 dev->transparent ? " (subtractive decode)" : "");

468 469 470 471
	pci_bus_remove_resources(child);
	for (i = 0; i < PCI_BRIDGE_RESOURCE_NUM; i++)
		child->resource[i] = &dev->resource[PCI_BRIDGE_RESOURCES+i];

472 473 474
	pci_read_bridge_io(child);
	pci_read_bridge_mmio(child);
	pci_read_bridge_mmio_pref(child);
475 476

	if (dev->transparent) {
477
		pci_bus_for_each_resource(child->parent, res, i) {
478
			if (res && res->flags) {
479 480
				pci_bus_add_resource(child, res,
						     PCI_SUBTRACTIVE_DECODE);
481 482
				dev_printk(KERN_DEBUG, &dev->dev,
					   "  bridge window %pR (subtractive decode)\n",
483 484
					   res);
			}
485 486
		}
	}
487 488
}

489
static struct pci_bus *pci_alloc_bus(struct pci_bus *parent)
L
Linus Torvalds 已提交
490 491 492
{
	struct pci_bus *b;

493
	b = kzalloc(sizeof(*b), GFP_KERNEL);
494 495 496 497 498 499 500 501 502 503
	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;
504 505 506 507
#ifdef CONFIG_PCI_DOMAINS_GENERIC
	if (parent)
		b->domain_nr = parent->domain_nr;
#endif
L
Linus Torvalds 已提交
508 509 510
	return b;
}

511 512 513 514 515 516 517 518 519 520 521 522
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);
}

523 524 525 526 527
static struct pci_host_bridge *pci_alloc_host_bridge(struct pci_bus *b)
{
	struct pci_host_bridge *bridge;

	bridge = kzalloc(sizeof(*bridge), GFP_KERNEL);
528 529
	if (!bridge)
		return NULL;
530

531 532
	INIT_LIST_HEAD(&bridge->windows);
	bridge->bus = b;
533 534 535
	return bridge;
}

536
static const unsigned char pcix_bus_speed[] = {
537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554
	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 */
};

555
const unsigned char pcie_link_speed[] = {
556 557 558
	PCI_SPEED_UNKNOWN,		/* 0 */
	PCIE_SPEED_2_5GT,		/* 1 */
	PCIE_SPEED_5_0GT,		/* 2 */
559
	PCIE_SPEED_8_0GT,		/* 3 */
560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575
	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)
{
576
	bus->cur_bus_speed = pcie_link_speed[linksta & PCI_EXP_LNKSTA_CLS];
577 578 579
}
EXPORT_SYMBOL_GPL(pcie_update_link_speed);

580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599
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;
600

601 602 603 604 605 606 607 608 609 610
	if (agp3) {
		index += 2;
		if (index == 5)
			index = 0;
	}

 out:
	return agp_speeds[index];
}

611 612 613 614 615
static void pci_set_bus_speed(struct pci_bus *bus)
{
	struct pci_dev *bridge = bus->self;
	int pos;

616 617 618 619 620 621 622 623 624 625 626 627 628
	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);
	}

629 630 631 632 633
	pos = pci_find_capability(bridge, PCI_CAP_ID_PCIX);
	if (pos) {
		u16 status;
		enum pci_bus_speed max;

634 635 636 637
		pci_read_config_word(bridge, pos + PCI_X_BRIDGE_SSTATUS,
				     &status);

		if (status & PCI_X_SSTATUS_533MHZ) {
638
			max = PCI_SPEED_133MHz_PCIX_533;
639
		} else if (status & PCI_X_SSTATUS_266MHZ) {
640
			max = PCI_SPEED_133MHz_PCIX_266;
641
		} else if (status & PCI_X_SSTATUS_133MHZ) {
R
Ryan Desfosses 已提交
642
			if ((status & PCI_X_SSTATUS_VERS) == PCI_X_SSTATUS_V2)
643
				max = PCI_SPEED_133MHz_PCIX_ECC;
R
Ryan Desfosses 已提交
644
			else
645 646 647 648 649 650
				max = PCI_SPEED_133MHz_PCIX;
		} else {
			max = PCI_SPEED_66MHz_PCIX;
		}

		bus->max_bus_speed = max;
651 652
		bus->cur_bus_speed = pcix_bus_speed[
			(status & PCI_X_SSTATUS_FREQ) >> 6];
653 654 655 656

		return;
	}

657
	if (pci_is_pcie(bridge)) {
658 659 660
		u32 linkcap;
		u16 linksta;

661
		pcie_capability_read_dword(bridge, PCI_EXP_LNKCAP, &linkcap);
662
		bus->max_bus_speed = pcie_link_speed[linkcap & PCI_EXP_LNKCAP_SLS];
663

664
		pcie_capability_read_word(bridge, PCI_EXP_LNKSTA, &linksta);
665 666 667 668
		pcie_update_link_speed(bus, linksta);
	}
}

669 670
static struct pci_bus *pci_alloc_child_bus(struct pci_bus *parent,
					   struct pci_dev *bridge, int busnr)
L
Linus Torvalds 已提交
671 672 673
{
	struct pci_bus *child;
	int i;
674
	int ret;
L
Linus Torvalds 已提交
675 676 677 678

	/*
	 * Allocate a new bus, and inherit stuff from the parent..
	 */
679
	child = pci_alloc_bus(parent);
L
Linus Torvalds 已提交
680 681 682 683 684
	if (!child)
		return NULL;

	child->parent = parent;
	child->ops = parent->ops;
685
	child->msi = parent->msi;
L
Linus Torvalds 已提交
686
	child->sysdata = parent->sysdata;
687
	child->bus_flags = parent->bus_flags;
L
Linus Torvalds 已提交
688

689
	/* initialize some portions of the bus device, but don't register it
690
	 * now as the parent is not properly set up yet.
691 692
	 */
	child->dev.class = &pcibus_class;
693
	dev_set_name(&child->dev, "%04x:%02x", pci_domain_nr(child), busnr);
L
Linus Torvalds 已提交
694 695 696 697 698

	/*
	 * Set up the primary, secondary and subordinate
	 * bus numbers.
	 */
699 700 701
	child->number = child->busn_res.start = busnr;
	child->primary = parent->busn_res.start;
	child->busn_res.end = 0xff;
L
Linus Torvalds 已提交
702

703 704 705 706
	if (!bridge) {
		child->dev.parent = parent->bridge;
		goto add_dev;
	}
707 708 709

	child->self = bridge;
	child->bridge = get_device(&bridge->dev);
710
	child->dev.parent = child->bridge;
711
	pci_set_bus_of_node(child);
712 713
	pci_set_bus_speed(child);

L
Linus Torvalds 已提交
714
	/* Set up default resource pointers and names.. */
715
	for (i = 0; i < PCI_BRIDGE_RESOURCE_NUM; i++) {
L
Linus Torvalds 已提交
716 717 718 719 720
		child->resource[i] = &bridge->resource[PCI_BRIDGE_RESOURCES+i];
		child->resource[i]->name = child->name;
	}
	bridge->subordinate = child;

721 722 723 724
add_dev:
	ret = device_register(&child->dev);
	WARN_ON(ret < 0);

725 726
	pcibios_add_bus(child);

727 728 729
	/* Create legacy_io and legacy_mem files for this bus */
	pci_create_legacy_files(child);

L
Linus Torvalds 已提交
730 731 732
	return child;
}

R
Ryan Desfosses 已提交
733 734
struct pci_bus *pci_add_new_bus(struct pci_bus *parent, struct pci_dev *dev,
				int busnr)
L
Linus Torvalds 已提交
735 736 737 738
{
	struct pci_bus *child;

	child = pci_alloc_child_bus(parent, dev, busnr);
739
	if (child) {
740
		down_write(&pci_bus_sem);
L
Linus Torvalds 已提交
741
		list_add_tail(&child->node, &parent->children);
742
		up_write(&pci_bus_sem);
743
	}
L
Linus Torvalds 已提交
744 745
	return child;
}
746
EXPORT_SYMBOL(pci_add_new_bus);
L
Linus Torvalds 已提交
747

748 749 750 751 752 753 754 755 756 757 758
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
759 760 761 762 763 764 765 766 767 768
/*
 * 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
769
int pci_scan_bridge(struct pci_bus *bus, struct pci_dev *dev, int max, int pass)
L
Linus Torvalds 已提交
770 771 772
{
	struct pci_bus *child;
	int is_cardbus = (dev->hdr_type == PCI_HEADER_TYPE_CARDBUS);
773
	u32 buses, i, j = 0;
L
Linus Torvalds 已提交
774
	u16 bctl;
775
	u8 primary, secondary, subordinate;
776
	int broken = 0;
L
Linus Torvalds 已提交
777 778

	pci_read_config_dword(dev, PCI_PRIMARY_BUS, &buses);
779 780 781
	primary = buses & 0xFF;
	secondary = (buses >> 8) & 0xFF;
	subordinate = (buses >> 16) & 0xFF;
L
Linus Torvalds 已提交
782

783 784
	dev_dbg(&dev->dev, "scanning [bus %02x-%02x] behind bridge, pass %d\n",
		secondary, subordinate, pass);
L
Linus Torvalds 已提交
785

786 787 788 789 790
	if (!primary && (primary != bus->number) && secondary && subordinate) {
		dev_warn(&dev->dev, "Primary bus is hard wired to 0\n");
		primary = bus->number;
	}

791 792
	/* Check if setup is sensible at all */
	if (!pass &&
793
	    (primary != bus->number || secondary <= bus->number ||
794
	     secondary > subordinate)) {
795 796
		dev_info(&dev->dev, "bridge configuration invalid ([bus %02x-%02x]), reconfiguring\n",
			 secondary, subordinate);
797 798 799
		broken = 1;
	}

L
Linus Torvalds 已提交
800
	/* Disable MasterAbortMode during probing to avoid reporting
801
	   of bus errors (in some architectures) */
L
Linus Torvalds 已提交
802 803 804 805
	pci_read_config_word(dev, PCI_BRIDGE_CONTROL, &bctl);
	pci_write_config_word(dev, PCI_BRIDGE_CONTROL,
			      bctl & ~PCI_BRIDGE_CTL_MASTER_ABORT);

806 807
	pci_enable_crs(dev);

808 809 810
	if ((secondary || subordinate) && !pcibios_assign_all_busses() &&
	    !is_cardbus && !broken) {
		unsigned int cmax;
L
Linus Torvalds 已提交
811 812 813 814 815
		/*
		 * Bus already configured by firmware, process it in the first
		 * pass and just note the configuration.
		 */
		if (pass)
816
			goto out;
L
Linus Torvalds 已提交
817 818

		/*
819 820 821 822
		 * 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
823
		 */
824
		child = pci_find_bus(pci_domain_nr(bus), secondary);
A
Alex Chiang 已提交
825
		if (!child) {
826
			child = pci_add_new_bus(bus, dev, secondary);
A
Alex Chiang 已提交
827 828
			if (!child)
				goto out;
829
			child->primary = primary;
Y
Yinghai Lu 已提交
830
			pci_bus_insert_busn_res(child, secondary, subordinate);
A
Alex Chiang 已提交
831
			child->bridge_ctl = bctl;
L
Linus Torvalds 已提交
832 833 834
		}

		cmax = pci_scan_child_bus(child);
835 836 837 838 839 840
		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
841 842 843 844 845
	} else {
		/*
		 * We need to assign a number to this bus which we always
		 * do in the second pass.
		 */
846
		if (!pass) {
847
			if (pcibios_assign_all_busses() || broken || is_cardbus)
848 849 850 851 852 853 854 855
				/* 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);
856
			goto out;
857
		}
L
Linus Torvalds 已提交
858 859 860 861

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

862 863 864
		/* 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. */
865 866
		child = pci_find_bus(pci_domain_nr(bus), max+1);
		if (!child) {
867
			child = pci_add_new_bus(bus, dev, max+1);
868 869
			if (!child)
				goto out;
870
			pci_bus_insert_busn_res(child, max+1, 0xff);
871
		}
872
		max++;
L
Linus Torvalds 已提交
873 874
		buses = (buses & 0xff000000)
		      | ((unsigned int)(child->primary)     <<  0)
875 876
		      | ((unsigned int)(child->busn_res.start)   <<  8)
		      | ((unsigned int)(child->busn_res.end) << 16);
L
Linus Torvalds 已提交
877 878 879 880 881 882 883 884 885

		/*
		 * yenta.c forces a secondary latency timer of 176.
		 * Copy that behaviour here.
		 */
		if (is_cardbus) {
			buses &= ~0xff000000;
			buses |= CARDBUS_LATENCY_TIMER << 24;
		}
886

L
Linus Torvalds 已提交
887 888 889 890 891 892
		/*
		 * We need to blast all three values with a single write.
		 */
		pci_write_config_dword(dev, PCI_PRIMARY_BUS, buses);

		if (!is_cardbus) {
893
			child->bridge_ctl = bctl;
L
Linus Torvalds 已提交
894 895 896 897 898 899 900
			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
901
			for (i = 0; i < CARDBUS_RESERVE_BUSNR; i++) {
902
				struct pci_bus *parent = bus;
903 904 905
				if (pci_find_bus(pci_domain_nr(bus),
							max+i+1))
					break;
906 907
				while (parent->parent) {
					if ((!pcibios_assign_all_busses()) &&
908 909
					    (parent->busn_res.end > max) &&
					    (parent->busn_res.end <= max+i)) {
910 911 912 913 914 915 916 917 918 919 920 921 922 923
						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;
				}
			}
924
			max += i;
L
Linus Torvalds 已提交
925 926 927 928
		}
		/*
		 * Set the subordinate bus number to its real value.
		 */
Y
Yinghai Lu 已提交
929
		pci_bus_update_busn_res_end(child, max);
L
Linus Torvalds 已提交
930 931 932
		pci_write_config_byte(dev, PCI_SUBORDINATE_BUS, max);
	}

933 934 935
	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
936

937
	/* Has only triggered on CardBus, fixup is in yenta_socket */
938
	while (bus->parent) {
939 940
		if ((child->busn_res.end > bus->busn_res.end) ||
		    (child->number > bus->busn_res.end) ||
941
		    (child->number < bus->number) ||
942
		    (child->busn_res.end < bus->number)) {
943
			dev_info(&child->dev, "%pR %s hidden behind%s bridge %s %pR\n",
944 945 946
				&child->busn_res,
				(bus->number > child->busn_res.end &&
				 bus->busn_res.end < child->number) ?
947 948
					"wholly" : "partially",
				bus->self->transparent ? " transparent" : "",
949
				dev_name(&bus->dev),
950
				&bus->busn_res);
951 952 953 954
		}
		bus = bus->parent;
	}

955 956 957
out:
	pci_write_config_word(dev, PCI_BRIDGE_CONTROL, bctl);

L
Linus Torvalds 已提交
958 959
	return max;
}
960
EXPORT_SYMBOL(pci_scan_bridge);
L
Linus Torvalds 已提交
961 962 963 964 965 966 967 968 969 970

/*
 * 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);
971
	dev->pin = irq;
L
Linus Torvalds 已提交
972 973 974 975 976
	if (irq)
		pci_read_config_byte(dev, PCI_INTERRUPT_LINE, &irq);
	dev->irq = irq;
}

977
void set_pcie_port_type(struct pci_dev *pdev)
Y
Yu Zhao 已提交
978 979 980 981 982 983 984
{
	int pos;
	u16 reg16;

	pos = pci_find_capability(pdev, PCI_CAP_ID_EXP);
	if (!pos)
		return;
985
	pdev->pcie_cap = pos;
Y
Yu Zhao 已提交
986
	pci_read_config_word(pdev, pos + PCI_EXP_FLAGS, &reg16);
987
	pdev->pcie_flags_reg = reg16;
988 989
	pci_read_config_word(pdev, pos + PCI_EXP_DEVCAP, &reg16);
	pdev->pcie_mpss = reg16 & PCI_EXP_DEVCAP_PAYLOAD;
Y
Yu Zhao 已提交
990 991
}

992
void set_pcie_hotplug_bridge(struct pci_dev *pdev)
993 994 995
{
	u32 reg32;

996
	pcie_capability_read_dword(pdev, PCI_EXP_SLTCAP, &reg32);
997 998 999 1000
	if (reg32 & PCI_EXP_SLTCAP_HPC)
		pdev->is_hotplug_bridge = 1;
}

1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037
/**
 * 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
}

1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055
/**
 * 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;
1056
	if (status == 0xffffffff || pci_ext_cfg_is_aliased(dev))
1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090
		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;
}

1091
#define LEGACY_IO_RESOURCE	(IORESOURCE_IO | IORESOURCE_PCI_FIXED)
1092

L
Linus Torvalds 已提交
1093 1094 1095 1096
/**
 * pci_setup_device - fill in class and map information of a device
 * @dev: the device structure to fill
 *
1097
 * Initialize the device structure with information about the device's
L
Linus Torvalds 已提交
1098 1099
 * 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
1100 1101
 * Returns 0 on success and negative if unknown type of device (not normal,
 * bridge or CardBus).
L
Linus Torvalds 已提交
1102
 */
Y
Yu Zhao 已提交
1103
int pci_setup_device(struct pci_dev *dev)
L
Linus Torvalds 已提交
1104 1105
{
	u32 class;
Y
Yu Zhao 已提交
1106 1107
	u8 hdr_type;
	struct pci_slot *slot;
1108
	int pos = 0;
1109 1110
	struct pci_bus_region region;
	struct resource *res;
Y
Yu Zhao 已提交
1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129

	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
1130

1131 1132 1133
	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
1134 1135

	pci_read_config_dword(dev, PCI_CLASS_REVISION, &class);
A
Auke Kok 已提交
1136
	dev->revision = class & 0xff;
Y
Yinghai Lu 已提交
1137
	dev->class = class >> 8;		    /* upper 3 bytes */
L
Linus Torvalds 已提交
1138

Y
Yinghai Lu 已提交
1139 1140
	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
1141

1142 1143 1144
	/* need to have dev->class ready */
	dev->cfg_size = pci_cfg_space_size(dev);

L
Linus Torvalds 已提交
1145
	/* "Unknown power state" */
1146
	dev->current_state = PCI_UNKNOWN;
L
Linus Torvalds 已提交
1147 1148 1149

	/* Early fixups, before probing the BARs */
	pci_fixup_device(pci_fixup_early, dev);
1150 1151
	/* device class may be changed after fixup */
	class = dev->class >> 8;
L
Linus Torvalds 已提交
1152 1153 1154 1155 1156 1157 1158 1159 1160

	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);
1161 1162

		/*
1163 1164 1165 1166
		 * 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!
1167 1168 1169 1170 1171
		 */
		if (class == PCI_CLASS_STORAGE_IDE) {
			u8 progif;
			pci_read_config_byte(dev, PCI_CLASS_PROG, &progif);
			if ((progif & 1) == 0) {
1172 1173 1174 1175
				region.start = 0x1F0;
				region.end = 0x1F7;
				res = &dev->resource[0];
				res->flags = LEGACY_IO_RESOURCE;
1176
				pcibios_bus_to_resource(dev->bus, res, &region);
1177 1178
				dev_info(&dev->dev, "legacy IDE quirk: reg 0x10: %pR\n",
					 res);
1179 1180 1181 1182
				region.start = 0x3F6;
				region.end = 0x3F6;
				res = &dev->resource[1];
				res->flags = LEGACY_IO_RESOURCE;
1183
				pcibios_bus_to_resource(dev->bus, res, &region);
1184 1185
				dev_info(&dev->dev, "legacy IDE quirk: reg 0x14: %pR\n",
					 res);
1186 1187
			}
			if ((progif & 4) == 0) {
1188 1189 1190 1191
				region.start = 0x170;
				region.end = 0x177;
				res = &dev->resource[2];
				res->flags = LEGACY_IO_RESOURCE;
1192
				pcibios_bus_to_resource(dev->bus, res, &region);
1193 1194
				dev_info(&dev->dev, "legacy IDE quirk: reg 0x18: %pR\n",
					 res);
1195 1196 1197 1198
				region.start = 0x376;
				region.end = 0x376;
				res = &dev->resource[3];
				res->flags = LEGACY_IO_RESOURCE;
1199
				pcibios_bus_to_resource(dev->bus, res, &region);
1200 1201
				dev_info(&dev->dev, "legacy IDE quirk: reg 0x1c: %pR\n",
					 res);
1202 1203
			}
		}
L
Linus Torvalds 已提交
1204 1205 1206 1207 1208 1209 1210
		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
1211
		   interface code of 0x01. */
1212
		pci_read_irq(dev);
L
Linus Torvalds 已提交
1213 1214
		dev->transparent = ((dev->class & 0xff) == 1);
		pci_read_bases(dev, 2, PCI_ROM_ADDRESS1);
1215
		set_pcie_hotplug_bridge(dev);
1216 1217 1218 1219 1220
		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
1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232
		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 */
1233 1234
		dev_err(&dev->dev, "unknown header type %02x, ignoring device\n",
			dev->hdr_type);
Y
Yu Zhao 已提交
1235
		return -EIO;
L
Linus Torvalds 已提交
1236 1237

	bad:
1238 1239
		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
1240 1241 1242 1243 1244 1245 1246
		dev->class = PCI_CLASS_NOT_DEFINED;
	}

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

1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258
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;

1259
	if (!hpp)
1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310
		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;
	}

1311 1312 1313 1314 1315 1316 1317 1318 1319 1320
	/*
	 * 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);

1321 1322 1323 1324 1325
	/* 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 */
1326
	if (pcie_cap_has_lnkctl(dev))
1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362
		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
	 */
}

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

1378 1379 1380
static void pci_release_capabilities(struct pci_dev *dev)
{
	pci_vpd_release(dev);
1381
	pci_iov_release(dev);
1382
	pci_free_cap_save_buffers(dev);
1383 1384
}

L
Linus Torvalds 已提交
1385 1386 1387 1388 1389 1390 1391 1392 1393
/**
 * 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)
{
1394
	struct pci_dev *pci_dev;
L
Linus Torvalds 已提交
1395

1396
	pci_dev = to_pci_dev(dev);
1397
	pci_release_capabilities(pci_dev);
1398
	pci_release_of_node(pci_dev);
1399
	pcibios_release_device(pci_dev);
1400
	pci_bus_put(pci_dev->bus);
1401
	kfree(pci_dev->driver_override);
L
Linus Torvalds 已提交
1402 1403 1404
	kfree(pci_dev);
}

1405
struct pci_dev *pci_alloc_dev(struct pci_bus *bus)
1406 1407 1408 1409 1410 1411 1412 1413
{
	struct pci_dev *dev;

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

	INIT_LIST_HEAD(&dev->bus_list);
1414
	dev->dev.type = &pci_dev_type;
1415
	dev->bus = pci_bus_get(bus);
1416 1417 1418

	return dev;
}
1419 1420
EXPORT_SYMBOL(pci_alloc_dev);

1421
bool pci_bus_read_dev_vendor_id(struct pci_bus *bus, int devfn, u32 *l,
R
Ryan Desfosses 已提交
1422
				int crs_timeout)
L
Linus Torvalds 已提交
1423 1424 1425
{
	int delay = 1;

1426 1427
	if (pci_bus_read_config_dword(bus, devfn, PCI_VENDOR_ID, l))
		return false;
L
Linus Torvalds 已提交
1428 1429

	/* some broken boards return 0 or ~0 if a slot is empty: */
1430 1431 1432
	if (*l == 0xffffffff || *l == 0x00000000 ||
	    *l == 0x0000ffff || *l == 0xffff0000)
		return false;
L
Linus Torvalds 已提交
1433

1434 1435 1436 1437 1438 1439 1440
	/*
	 * 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) {
1441 1442 1443
		if (!crs_timeout)
			return false;

L
Linus Torvalds 已提交
1444 1445
		msleep(delay);
		delay *= 2;
1446 1447
		if (pci_bus_read_config_dword(bus, devfn, PCI_VENDOR_ID, l))
			return false;
L
Linus Torvalds 已提交
1448
		/* Card hasn't responded in 60 seconds?  Must be stuck. */
1449
		if (delay > crs_timeout) {
1450 1451 1452
			printk(KERN_WARNING "pci %04x:%02x:%02x.%d: not responding\n",
			       pci_domain_nr(bus), bus->number, PCI_SLOT(devfn),
			       PCI_FUNC(devfn));
1453
			return false;
L
Linus Torvalds 已提交
1454 1455 1456
		}
	}

1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472
	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;

1473
	dev = pci_alloc_dev(bus);
L
Linus Torvalds 已提交
1474 1475 1476 1477 1478 1479
	if (!dev)
		return NULL;

	dev->devfn = devfn;
	dev->vendor = l & 0xffff;
	dev->device = (l >> 16) & 0xffff;
1480

1481 1482
	pci_set_of_node(dev);

Y
Yu Zhao 已提交
1483
	if (pci_setup_device(dev)) {
1484
		pci_bus_put(dev->bus);
L
Linus Torvalds 已提交
1485 1486 1487 1488 1489 1490 1491
		kfree(dev);
		return NULL;
	}

	return dev;
}

1492 1493 1494 1495 1496
static void pci_init_capabilities(struct pci_dev *dev)
{
	/* MSI/MSI-X list */
	pci_msi_init_pci_dev(dev);

1497 1498 1499
	/* Buffers for saving PCIe and PCI-X capabilities */
	pci_allocate_cap_save_buffers(dev);

1500 1501 1502 1503 1504
	/* Power Management */
	pci_pm_init(dev);

	/* Vital Product Data */
	pci_vpd_pci22_init(dev);
Y
Yu Zhao 已提交
1505 1506

	/* Alternative Routing-ID Forwarding */
1507
	pci_configure_ari(dev);
1508 1509 1510

	/* Single Root I/O Virtualization */
	pci_iov_init(dev);
1511 1512

	/* Enable ACS P2P upstream forwarding */
C
Chris Wright 已提交
1513
	pci_enable_acs(dev);
1514 1515
}

1516
void pci_device_add(struct pci_dev *dev, struct pci_bus *bus)
L
Linus Torvalds 已提交
1517
{
1518 1519
	int ret;

1520 1521
	pci_configure_device(dev);

1522 1523
	device_initialize(&dev->dev);
	dev->dev.release = pci_release_dev;
L
Linus Torvalds 已提交
1524

1525
	set_dev_node(&dev->dev, pcibus_to_node(bus));
1526
	dev->dev.dma_mask = &dev->dma_mask;
1527
	dev->dev.dma_parms = &dev->dma_parms;
1528
	dev->dev.coherent_dma_mask = 0xffffffffull;
L
Linus Torvalds 已提交
1529

1530
	pci_set_dma_max_seg_size(dev, 65536);
1531
	pci_set_dma_seg_boundary(dev, 0xffffffff);
1532

L
Linus Torvalds 已提交
1533 1534 1535
	/* Fix up broken headers */
	pci_fixup_device(pci_fixup_header, dev);

1536 1537 1538
	/* moved out from quirk header fixup code */
	pci_reassigndev_resource_alignment(dev);

1539 1540 1541
	/* Clear the state_saved flag. */
	dev->state_saved = false;

1542 1543
	/* Initialize various capabilities */
	pci_init_capabilities(dev);
1544

L
Linus Torvalds 已提交
1545 1546 1547 1548
	/*
	 * Add the device to our list of discovered devices
	 * and the bus list for fixup functions, etc.
	 */
1549
	down_write(&pci_bus_sem);
L
Linus Torvalds 已提交
1550
	list_add_tail(&dev->bus_list, &bus->devices);
1551
	up_write(&pci_bus_sem);
1552 1553 1554 1555 1556 1557 1558 1559

	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);
1560 1561
}

1562
struct pci_dev *pci_scan_single_device(struct pci_bus *bus, int devfn)
1563 1564 1565
{
	struct pci_dev *dev;

T
Trent Piepho 已提交
1566 1567 1568 1569 1570 1571
	dev = pci_get_slot(bus, devfn);
	if (dev) {
		pci_dev_put(dev);
		return dev;
	}

1572 1573 1574 1575 1576
	dev = pci_scan_device(bus, devfn);
	if (!dev)
		return NULL;

	pci_device_add(dev, bus);
L
Linus Torvalds 已提交
1577 1578 1579

	return dev;
}
1580
EXPORT_SYMBOL(pci_scan_single_device);
L
Linus Torvalds 已提交
1581

1582
static unsigned next_fn(struct pci_bus *bus, struct pci_dev *dev, unsigned fn)
M
Matthew Wilcox 已提交
1583
{
1584 1585 1586
	int pos;
	u16 cap = 0;
	unsigned next_fn;
1587

1588 1589 1590 1591 1592 1593
	if (pci_ari_enabled(bus)) {
		if (!dev)
			return 0;
		pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ARI);
		if (!pos)
			return 0;
1594

1595 1596 1597 1598
		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
1599

1600 1601 1602 1603 1604 1605
		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
1606 1607 1608 1609 1610 1611 1612

	return 0;
}

static int only_one_child(struct pci_bus *bus)
{
	struct pci_dev *parent = bus->self;
1613

M
Matthew Wilcox 已提交
1614 1615
	if (!parent || !pci_is_pcie(parent))
		return 0;
1616
	if (pci_pcie_type(parent) == PCI_EXP_TYPE_ROOT_PORT)
1617
		return 1;
1618
	if (pci_pcie_type(parent) == PCI_EXP_TYPE_DOWNSTREAM &&
1619
	    !pci_has_flag(PCI_SCAN_ALL_PCIE_DEVS))
M
Matthew Wilcox 已提交
1620 1621 1622 1623
		return 1;
	return 0;
}

L
Linus Torvalds 已提交
1624 1625 1626 1627 1628 1629 1630
/**
 * 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
1631
 * will not have is_added set.
1632 1633
 *
 * Returns the number of new devices found.
L
Linus Torvalds 已提交
1634
 */
1635
int pci_scan_slot(struct pci_bus *bus, int devfn)
L
Linus Torvalds 已提交
1636
{
M
Matthew Wilcox 已提交
1637
	unsigned fn, nr = 0;
1638
	struct pci_dev *dev;
M
Matthew Wilcox 已提交
1639 1640 1641

	if (only_one_child(bus) && (devfn > 0))
		return 0; /* Already scanned the entire slot */
L
Linus Torvalds 已提交
1642

1643
	dev = pci_scan_single_device(bus, devfn);
1644 1645 1646
	if (!dev)
		return 0;
	if (!dev->is_added)
1647 1648
		nr++;

1649
	for (fn = next_fn(bus, dev, 0); fn > 0; fn = next_fn(bus, dev, fn)) {
M
Matthew Wilcox 已提交
1650 1651 1652 1653 1654
		dev = pci_scan_single_device(bus, devfn + fn);
		if (dev) {
			if (!dev->is_added)
				nr++;
			dev->multifunction = 1;
L
Linus Torvalds 已提交
1655 1656
		}
	}
S
Shaohua Li 已提交
1657

1658 1659
	/* only one slot has pcie device */
	if (bus->self && nr)
S
Shaohua Li 已提交
1660 1661
		pcie_aspm_init_link_state(bus->self);

L
Linus Torvalds 已提交
1662 1663
	return nr;
}
1664
EXPORT_SYMBOL(pci_scan_slot);
L
Linus Torvalds 已提交
1665

1666 1667 1668 1669 1670 1671 1672
static int pcie_find_smpss(struct pci_dev *dev, void *data)
{
	u8 *smpss = data;

	if (!pci_is_pcie(dev))
		return 0;

1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686
	/*
	 * 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.
1687
	 */
1688 1689
	if (dev->is_hotplug_bridge &&
	    pci_pcie_type(dev) != PCI_EXP_TYPE_ROOT_PORT)
1690 1691 1692 1693 1694 1695 1696 1697 1698 1699
		*smpss = 0;

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

	return 0;
}

static void pcie_write_mps(struct pci_dev *dev, int mps)
{
1700
	int rc;
1701 1702

	if (pcie_bus_config == PCIE_BUS_PERFORMANCE) {
1703
		mps = 128 << dev->pcie_mpss;
1704

1705 1706
		if (pci_pcie_type(dev) != PCI_EXP_TYPE_ROOT_PORT &&
		    dev->bus->self)
1707
			/* For "Performance", the assumption is made that
1708 1709 1710 1711 1712
			 * 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.
1713 1714 1715 1716 1717
			 *
			 * 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).
1718
			 */
1719
			mps = min(mps, pcie_get_mps(dev->bus->self));
1720 1721 1722 1723 1724 1725 1726
	}

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

1727
static void pcie_write_mrrs(struct pci_dev *dev)
1728
{
1729
	int rc, mrrs;
1730

1731 1732 1733 1734 1735 1736 1737 1738
	/* 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
1739 1740
	 * device or the bus can support.  This should already be properly
	 * configured by a prior call to pcie_write_mps.
1741
	 */
1742
	mrrs = pcie_get_mps(dev);
1743 1744

	/* MRRS is a R/W register.  Invalid values can be written, but a
1745
	 * subsequent read will verify if the value is acceptable or not.
1746 1747
	 * If the MRRS value provided is not acceptable (e.g., too large),
	 * shrink the value until it is acceptable to the HW.
1748
	 */
1749 1750
	while (mrrs != pcie_get_readrq(dev) && mrrs >= 128) {
		rc = pcie_set_readrq(dev, mrrs);
1751 1752
		if (!rc)
			break;
1753

1754
		dev_warn(&dev->dev, "Failed attempting to set the MRRS\n");
1755 1756
		mrrs /= 2;
	}
1757 1758

	if (mrrs < 128)
1759
		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");
1760 1761
}

1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777
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);
}

1778 1779
static int pcie_bus_configure_set(struct pci_dev *dev, void *data)
{
J
Jon Mason 已提交
1780
	int mps, orig_mps;
1781 1782 1783 1784

	if (!pci_is_pcie(dev))
		return 0;

1785 1786 1787 1788 1789
	if (pcie_bus_config == PCIE_BUS_TUNE_OFF) {
		pcie_bus_detect_mps(dev);
		return 0;
	}

J
Jon Mason 已提交
1790 1791
	mps = 128 << *(u8 *)data;
	orig_mps = pcie_get_mps(dev);
1792 1793

	pcie_write_mps(dev, mps);
1794
	pcie_write_mrrs(dev);
1795

1796 1797
	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
1798
		 orig_mps, pcie_get_readrq(dev));
1799 1800 1801 1802

	return 0;
}

J
Jon Mason 已提交
1803
/* pcie_bus_configure_settings requires that pci_walk_bus work in a top-down,
1804 1805 1806
 * parents then children fashion.  If this changes, then this code will not
 * work as designed.
 */
1807
void pcie_bus_configure_settings(struct pci_bus *bus)
1808
{
1809
	u8 smpss = 0;
1810

1811
	if (!bus->self)
1812 1813 1814
		return;

	if (!pci_is_pcie(bus->self))
1815 1816 1817
		return;

	/* FIXME - Peer to peer DMA is possible, though the endpoint would need
1818
	 * to be aware of the MPS of the destination.  To work around this,
1819 1820 1821 1822 1823
	 * simply force the MPS of the entire system to the smallest possible.
	 */
	if (pcie_bus_config == PCIE_BUS_PEER2PEER)
		smpss = 0;

1824
	if (pcie_bus_config == PCIE_BUS_SAFE) {
1825
		smpss = bus->self->pcie_mpss;
1826

1827 1828 1829 1830 1831 1832 1833
		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);
}
1834
EXPORT_SYMBOL_GPL(pcie_bus_configure_settings);
1835

B
Bill Pemberton 已提交
1836
unsigned int pci_scan_child_bus(struct pci_bus *bus)
L
Linus Torvalds 已提交
1837
{
1838
	unsigned int devfn, pass, max = bus->busn_res.start;
L
Linus Torvalds 已提交
1839 1840
	struct pci_dev *dev;

B
Bjorn Helgaas 已提交
1841
	dev_dbg(&bus->dev, "scanning bus\n");
L
Linus Torvalds 已提交
1842 1843 1844 1845 1846

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

1847 1848 1849
	/* Reserve buses for SR-IOV capability. */
	max += pci_iov_bus_range(bus);

L
Linus Torvalds 已提交
1850 1851 1852 1853
	/*
	 * After performing arch-dependent fixup of the bus, look behind
	 * all PCI-to-PCI bridges on this bus.
	 */
A
Alex Chiang 已提交
1854
	if (!bus->is_added) {
B
Bjorn Helgaas 已提交
1855
		dev_dbg(&bus->dev, "fixups for bus\n");
A
Alex Chiang 已提交
1856
		pcibios_fixup_bus(bus);
1857
		bus->is_added = 1;
A
Alex Chiang 已提交
1858 1859
	}

R
Ryan Desfosses 已提交
1860
	for (pass = 0; pass < 2; pass++)
L
Linus Torvalds 已提交
1861
		list_for_each_entry(dev, &bus->devices, bus_list) {
1862
			if (pci_is_bridge(dev))
L
Linus Torvalds 已提交
1863 1864 1865 1866 1867 1868 1869 1870 1871 1872
				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
1873
	dev_dbg(&bus->dev, "bus scan returning with max=%02x\n", max);
L
Linus Torvalds 已提交
1874 1875
	return max;
}
1876
EXPORT_SYMBOL_GPL(pci_scan_child_bus);
L
Linus Torvalds 已提交
1877

1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889
/**
 * 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;
}

1890 1891 1892 1893 1894 1895 1896 1897
void __weak pcibios_add_bus(struct pci_bus *bus)
{
}

void __weak pcibios_remove_bus(struct pci_bus *bus)
{
}

1898 1899
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
1900
{
1901
	int error;
1902
	struct pci_host_bridge *bridge;
B
Bjorn Helgaas 已提交
1903
	struct pci_bus *b, *b2;
1904
	struct pci_host_bridge_window *window, *n;
1905
	struct resource *res;
1906 1907 1908
	resource_size_t offset;
	char bus_addr[64];
	char *fmt;
L
Linus Torvalds 已提交
1909

1910
	b = pci_alloc_bus(NULL);
L
Linus Torvalds 已提交
1911
	if (!b)
1912
		return NULL;
L
Linus Torvalds 已提交
1913 1914 1915

	b->sysdata = sysdata;
	b->ops = ops;
1916
	b->number = b->busn_res.start = bus;
1917
	pci_bus_assign_domain_nr(b, parent);
B
Bjorn Helgaas 已提交
1918 1919
	b2 = pci_find_bus(pci_domain_nr(b), bus);
	if (b2) {
L
Linus Torvalds 已提交
1920
		/* If we already got to this bus through a different bridge, ignore it */
B
Bjorn Helgaas 已提交
1921
		dev_dbg(&b2->dev, "bus already known\n");
L
Linus Torvalds 已提交
1922 1923
		goto err_out;
	}
1924

1925 1926 1927 1928 1929
	bridge = pci_alloc_host_bridge(b);
	if (!bridge)
		goto err_out;

	bridge->dev.parent = parent;
1930
	bridge->dev.release = pci_release_host_bridge_dev;
1931
	dev_set_name(&bridge->dev, "pci%04x:%02x", pci_domain_nr(b), bus);
1932
	error = pcibios_root_bridge_prepare(bridge);
1933 1934 1935 1936
	if (error) {
		kfree(bridge);
		goto err_out;
	}
1937

1938
	error = device_register(&bridge->dev);
1939 1940 1941 1942
	if (error) {
		put_device(&bridge->dev);
		goto err_out;
	}
1943
	b->bridge = get_device(&bridge->dev);
1944
	device_enable_async_suspend(b->bridge);
1945
	pci_set_bus_of_node(b);
L
Linus Torvalds 已提交
1946

1947 1948 1949
	if (!parent)
		set_dev_node(b->bridge, pcibus_to_node(b));

1950 1951
	b->dev.class = &pcibus_class;
	b->dev.parent = b->bridge;
1952
	dev_set_name(&b->dev, "%04x:%02x", pci_domain_nr(b), bus);
1953
	error = device_register(&b->dev);
L
Linus Torvalds 已提交
1954 1955 1956
	if (error)
		goto class_dev_reg_err;

1957 1958
	pcibios_add_bus(b);

L
Linus Torvalds 已提交
1959 1960 1961
	/* Create legacy_io and legacy_mem files for this bus */
	pci_create_legacy_files(b);

1962 1963 1964 1965 1966
	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));

1967 1968 1969 1970 1971
	/* Add initial resources to the bus */
	list_for_each_entry_safe(window, n, resources, list) {
		list_move_tail(&window->list, &bridge->windows);
		res = window->res;
		offset = window->offset;
1972 1973 1974 1975
		if (res->flags & IORESOURCE_BUS)
			pci_bus_insert_busn_res(b, bus, res->end);
		else
			pci_bus_add_resource(b, res, 0);
1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986
		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);
1987 1988
	}

1989 1990 1991 1992
	down_write(&pci_bus_sem);
	list_add_tail(&b->node, &pci_root_buses);
	up_write(&pci_bus_sem);

L
Linus Torvalds 已提交
1993 1994 1995
	return b;

class_dev_reg_err:
1996 1997
	put_device(&bridge->dev);
	device_unregister(&bridge->dev);
L
Linus Torvalds 已提交
1998 1999 2000 2001
err_out:
	kfree(b);
	return NULL;
}
2002

2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
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;
	}

2019
	conflict = request_resource_conflict(parent_res, res);
2020 2021 2022 2023 2024 2025 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

	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
2066
struct pci_bus *pci_scan_root_bus(struct device *parent, int bus,
2067 2068
		struct pci_ops *ops, void *sysdata, struct list_head *resources)
{
2069 2070
	struct pci_host_bridge_window *window;
	bool found = false;
2071
	struct pci_bus *b;
2072 2073 2074 2075 2076 2077 2078
	int max;

	list_for_each_entry(window, resources, list)
		if (window->res->flags & IORESOURCE_BUS) {
			found = true;
			break;
		}
2079 2080 2081 2082 2083

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

2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095
	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);

2096 2097 2098 2099 2100
	pci_bus_add_devices(b);
	return b;
}
EXPORT_SYMBOL(pci_scan_root_bus);

2101
/* Deprecated; use pci_scan_root_bus() instead */
B
Bill Pemberton 已提交
2102
struct pci_bus *pci_scan_bus_parented(struct device *parent,
2103 2104
		int bus, struct pci_ops *ops, void *sysdata)
{
2105
	LIST_HEAD(resources);
2106 2107
	struct pci_bus *b;

2108 2109
	pci_add_resource(&resources, &ioport_resource);
	pci_add_resource(&resources, &iomem_resource);
2110
	pci_add_resource(&resources, &busn_resource);
2111
	b = pci_create_root_bus(parent, bus, ops, sysdata, &resources);
2112
	if (b)
2113
		pci_scan_child_bus(b);
2114 2115
	else
		pci_free_resource_list(&resources);
2116 2117
	return b;
}
L
Linus Torvalds 已提交
2118 2119
EXPORT_SYMBOL(pci_scan_bus_parented);

B
Bill Pemberton 已提交
2120
struct pci_bus *pci_scan_bus(int bus, struct pci_ops *ops,
2121 2122 2123 2124 2125 2126 2127
					void *sysdata)
{
	LIST_HEAD(resources);
	struct pci_bus *b;

	pci_add_resource(&resources, &ioport_resource);
	pci_add_resource(&resources, &iomem_resource);
2128
	pci_add_resource(&resources, &busn_resource);
2129 2130
	b = pci_create_root_bus(NULL, bus, ops, sysdata, &resources);
	if (b) {
2131
		pci_scan_child_bus(b);
2132 2133 2134 2135 2136 2137 2138 2139
		pci_bus_add_devices(b);
	} else {
		pci_free_resource_list(&resources);
	}
	return b;
}
EXPORT_SYMBOL(pci_scan_bus);

2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150
/**
 * 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.
 */
2151
unsigned int pci_rescan_bus_bridge_resize(struct pci_dev *bridge)
2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164
{
	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;
}

2165 2166 2167 2168 2169 2170 2171 2172 2173
/**
 * 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.
 */
2174
unsigned int pci_rescan_bus(struct pci_bus *bus)
2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185
{
	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);

2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203
/*
 * 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
2204 2205
static int __init pci_sort_bf_cmp(const struct device *d_a,
				  const struct device *d_b)
2206
{
2207 2208 2209
	const struct pci_dev *a = to_pci_dev(d_a);
	const struct pci_dev *b = to_pci_dev(d_b);

2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221
	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;
}

2222
void __init pci_sort_breadthfirst(void)
2223
{
2224
	bus_sort_breadthfirst(&pci_bus_type, &pci_sort_bf_cmp);
2225
}