mce_amd.c 17.8 KB
Newer Older
1
#include <linux/module.h>
2 3
#include <linux/slab.h>

B
Borislav Petkov 已提交
4
#include "mce_amd.h"
D
Doug Thompson 已提交
5

6 7
static struct amd_decoder_ops *fam_ops;

8
static u8 xec_mask	 = 0xf;
9 10
static u8 nb_err_cpumask = 0xf;

11
static bool report_gart_errors;
12
static void (*nb_bus_decoder)(int node_id, struct mce *m, u32 nbcfg);
13 14 15 16 17 18 19

void amd_report_gart_errors(bool v)
{
	report_gart_errors = v;
}
EXPORT_SYMBOL_GPL(amd_report_gart_errors);

20
void amd_register_ecc_decoder(void (*f)(int, struct mce *, u32))
21 22 23 24 25
{
	nb_bus_decoder = f;
}
EXPORT_SYMBOL_GPL(amd_register_ecc_decoder);

26
void amd_unregister_ecc_decoder(void (*f)(int, struct mce *, u32))
27 28 29 30 31 32 33 34 35
{
	if (nb_bus_decoder) {
		WARN_ON(nb_bus_decoder != f);

		nb_bus_decoder = NULL;
	}
}
EXPORT_SYMBOL_GPL(amd_unregister_ecc_decoder);

D
Doug Thompson 已提交
36 37 38 39
/*
 * string representation for the different MCA reported error types, see F3x48
 * or MSR0000_0411.
 */
B
Borislav Petkov 已提交
40 41 42

/* transaction type */
const char *tt_msgs[] = { "INSN", "DATA", "GEN", "RESV" };
43
EXPORT_SYMBOL_GPL(tt_msgs);
D
Doug Thompson 已提交
44

B
Borislav Petkov 已提交
45 46
/* cache level */
const char *ll_msgs[] = { "RESV", "L1", "L2", "L3/GEN" };
47
EXPORT_SYMBOL_GPL(ll_msgs);
D
Doug Thompson 已提交
48

B
Borislav Petkov 已提交
49
/* memory transaction type */
D
Doug Thompson 已提交
50
const char *rrrr_msgs[] = {
B
Borislav Petkov 已提交
51
       "GEN", "RD", "WR", "DRD", "DWR", "IRD", "PRF", "EV", "SNP"
D
Doug Thompson 已提交
52
};
53
EXPORT_SYMBOL_GPL(rrrr_msgs);
D
Doug Thompson 已提交
54

B
Borislav Petkov 已提交
55 56
/* participating processor */
const char *pp_msgs[] = { "SRC", "RES", "OBS", "GEN" };
57
EXPORT_SYMBOL_GPL(pp_msgs);
D
Doug Thompson 已提交
58

B
Borislav Petkov 已提交
59 60
/* request timeout */
const char *to_msgs[] = { "no timeout",	"timed out" };
61
EXPORT_SYMBOL_GPL(to_msgs);
D
Doug Thompson 已提交
62

B
Borislav Petkov 已提交
63 64
/* memory or i/o */
const char *ii_msgs[] = { "MEM", "RESV", "IO", "GEN" };
65
EXPORT_SYMBOL_GPL(ii_msgs);
D
Doug Thompson 已提交
66

67 68 69 70 71 72 73 74 75
static const char *f10h_nb_mce_desc[] = {
	"HT link data error",
	"Protocol error (link, L3, probe filter, etc.)",
	"Parity error in NB-internal arrays",
	"Link Retry due to IO link transmission error",
	"L3 ECC data cache error",
	"ECC error in L3 cache tag",
	"L3 LRU parity bits error",
	"ECC Error in the Probe Filter directory"
D
Doug Thompson 已提交
76
};
77

78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97
static const char * const f15h_ic_mce_desc[] = {
	"UC during a demand linefill from L2",
	"Parity error during data load from IC",
	"Parity error for IC valid bit",
	"Main tag parity error",
	"Parity error in prediction queue",
	"PFB data/address parity error",
	"Parity error in the branch status reg",
	"PFB promotion address error",
	"Tag error during probe/victimization",
	"Parity error for IC probe tag valid bit",
	"PFB non-cacheable bit parity error",
	"PFB valid bit parity error",			/* xec = 0xd */
	"patch RAM",					/* xec = 010 */
	"uop queue",
	"insn buffer",
	"predecode buffer",
	"fetch address FIFO"
};

98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114
static const char * const f15h_cu_mce_desc[] = {
	"Fill ECC error on data fills",			/* xec = 0x4 */
	"Fill parity error on insn fills",
	"Prefetcher request FIFO parity error",
	"PRQ address parity error",
	"PRQ data parity error",
	"WCC Tag ECC error",
	"WCC Data ECC error",
	"WCB Data parity error",
	"VB Data/ECC error",
	"L2 Tag ECC error",				/* xec = 0x10 */
	"Hard L2 Tag ECC error",
	"Multiple hits on L2 tag",
	"XAB parity error",
	"PRB address parity error"
};

115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
static const char * const fr_ex_mce_desc[] = {
	"CPU Watchdog timer expire",
	"Wakeup array dest tag",
	"AG payload array",
	"EX payload array",
	"IDRF array",
	"Retire dispatch queue",
	"Mapper checkpoint array",
	"Physical register file EX0 port",
	"Physical register file EX1 port",
	"Physical register file AG0 port",
	"Physical register file AG1 port",
	"Flag register file",
	"DE correctable error could not be corrected"
};

131
static bool f12h_dc_mce(u16 ec, u8 xec)
132
{
133
	bool ret = false;
134

135 136 137
	if (MEM_ERROR(ec)) {
		u8 ll = ec & 0x3;
		ret = true;
138

139 140 141 142 143 144 145 146 147
		if (ll == LL_L2)
			pr_cont("during L1 linefill from L2.\n");
		else if (ll == LL_L1)
			pr_cont("Data/Tag %s error.\n", RRRR_MSG(ec));
		else
			ret = false;
	}
	return ret;
}
148

149
static bool f10h_dc_mce(u16 ec, u8 xec)
150 151 152 153 154 155 156 157
{
	u8 r4  = (ec >> 4) & 0xf;
	u8 ll  = ec & 0x3;

	if (r4 == R4_GEN && ll == LL_L1) {
		pr_cont("during data scrub.\n");
		return true;
	}
158
	return f12h_dc_mce(ec, xec);
159 160
}

161
static bool k8_dc_mce(u16 ec, u8 xec)
162 163 164 165 166
{
	if (BUS_ERROR(ec)) {
		pr_cont("during system linefill.\n");
		return true;
	}
167

168
	return f10h_dc_mce(ec, xec);
169 170
}

171
static bool f14h_dc_mce(u16 ec, u8 xec)
172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225
{
	u8 r4	 = (ec >> 4) & 0xf;
	u8 ll	 = ec & 0x3;
	u8 tt	 = (ec >> 2) & 0x3;
	u8 ii	 = tt;
	bool ret = true;

	if (MEM_ERROR(ec)) {

		if (tt != TT_DATA || ll != LL_L1)
			return false;

		switch (r4) {
		case R4_DRD:
		case R4_DWR:
			pr_cont("Data/Tag parity error due to %s.\n",
				(r4 == R4_DRD ? "load/hw prf" : "store"));
			break;
		case R4_EVICT:
			pr_cont("Copyback parity error on a tag miss.\n");
			break;
		case R4_SNOOP:
			pr_cont("Tag parity error during snoop.\n");
			break;
		default:
			ret = false;
		}
	} else if (BUS_ERROR(ec)) {

		if ((ii != II_MEM && ii != II_IO) || ll != LL_LG)
			return false;

		pr_cont("System read data error on a ");

		switch (r4) {
		case R4_RD:
			pr_cont("TLB reload.\n");
			break;
		case R4_DWR:
			pr_cont("store.\n");
			break;
		case R4_DRD:
			pr_cont("load.\n");
			break;
		default:
			ret = false;
		}
	} else {
		ret = false;
	}

	return ret;
}

226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273
static bool f15h_dc_mce(u16 ec, u8 xec)
{
	bool ret = true;

	if (MEM_ERROR(ec)) {

		switch (xec) {
		case 0x0:
			pr_cont("Data Array access error.\n");
			break;

		case 0x1:
			pr_cont("UC error during a linefill from L2/NB.\n");
			break;

		case 0x2:
		case 0x11:
			pr_cont("STQ access error.\n");
			break;

		case 0x3:
			pr_cont("SCB access error.\n");
			break;

		case 0x10:
			pr_cont("Tag error.\n");
			break;

		case 0x12:
			pr_cont("LDQ access error.\n");
			break;

		default:
			ret = false;
		}
	} else if (BUS_ERROR(ec)) {

		if (!xec)
			pr_cont("during system linefill.\n");
		else
			pr_cont(" Internal %s condition.\n",
				((xec == 1) ? "livelock" : "deadlock"));
	} else
		ret = false;

	return ret;
}

274 275 276
static void amd_decode_dc_mce(struct mce *m)
{
	u16 ec = m->status & 0xffff;
277
	u8 xec = (m->status >> 16) & xec_mask;
278 279 280 281 282 283 284 285 286

	pr_emerg(HW_ERR "Data Cache Error: ");

	/* TLB error signatures are the same across families */
	if (TLB_ERROR(ec)) {
		u8 tt = (ec >> 2) & 0x3;

		if (tt == TT_DATA) {
			pr_cont("%s TLB %s.\n", LL_MSG(ec),
287 288
				((xec == 2) ? "locked miss"
					    : (xec ? "multimatch" : "parity")));
289 290
			return;
		}
291 292 293 294
	} else if (fam_ops->dc_mce(ec, xec))
		;
	else
		pr_emerg(HW_ERR "Corrupted DC MCE info?\n");
295 296
}

297
static bool k8_ic_mce(u16 ec, u8 xec)
298
{
299 300 301
	u8 ll	 = ec & 0x3;
	u8 r4	 = (ec >> 4) & 0xf;
	bool ret = true;
302

303 304
	if (!MEM_ERROR(ec))
		return false;
305

306 307 308 309 310 311 312
	if (ll == 0x2)
		pr_cont("during a linefill from L2.\n");
	else if (ll == 0x1) {
		switch (r4) {
		case R4_IRD:
			pr_cont("Parity error during data load.\n");
			break;
313

314 315 316 317 318 319 320 321 322 323 324 325
		case R4_EVICT:
			pr_cont("Copyback Parity/Victim error.\n");
			break;

		case R4_SNOOP:
			pr_cont("Tag Snoop error.\n");
			break;

		default:
			ret = false;
			break;
		}
326
	} else
327
		ret = false;
328

329 330 331
	return ret;
}

332
static bool f14h_ic_mce(u16 ec, u8 xec)
333 334 335 336 337
{
	u8 ll    = ec & 0x3;
	u8 tt    = (ec >> 2) & 0x3;
	u8 r4  = (ec >> 4) & 0xf;
	bool ret = true;
338

339 340 341 342 343 344 345 346 347 348 349 350 351 352
	if (MEM_ERROR(ec)) {
		if (tt != 0 || ll != 1)
			ret = false;

		if (r4 == R4_IRD)
			pr_cont("Data/tag array parity error for a tag hit.\n");
		else if (r4 == R4_SNOOP)
			pr_cont("Tag error during snoop/victimization.\n");
		else
			ret = false;
	}
	return ret;
}

353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378
static bool f15h_ic_mce(u16 ec, u8 xec)
{
	bool ret = true;

	if (!MEM_ERROR(ec))
		return false;

	switch (xec) {
	case 0x0 ... 0xa:
		pr_cont("%s.\n", f15h_ic_mce_desc[xec]);
		break;

	case 0xd:
		pr_cont("%s.\n", f15h_ic_mce_desc[xec-2]);
		break;

	case 0x10 ... 0x14:
		pr_cont("Decoder %s parity error.\n", f15h_ic_mce_desc[xec-4]);
		break;

	default:
		ret = false;
	}
	return ret;
}

379 380 381
static void amd_decode_ic_mce(struct mce *m)
{
	u16 ec = m->status & 0xffff;
382
	u8 xec = (m->status >> 16) & xec_mask;
383 384 385 386 387 388 389

	pr_emerg(HW_ERR "Instruction Cache Error: ");

	if (TLB_ERROR(ec))
		pr_cont("%s TLB %s.\n", LL_MSG(ec),
			(xec ? "multimatch" : "parity error"));
	else if (BUS_ERROR(ec)) {
390
		bool k8 = (boot_cpu_data.x86 == 0xf && (m->status & BIT_64(58)));
391 392

		pr_cont("during %s.\n", (k8 ? "system linefill" : "NB data read"));
393
	} else if (fam_ops->ic_mce(ec, xec))
394 395 396
		;
	else
		pr_emerg(HW_ERR "Corrupted IC MCE info?\n");
397 398
}

399
static void amd_decode_bu_mce(struct mce *m)
400
{
401
	u32 ec = m->status & 0xffff;
402
	u32 xec = (m->status >> 16) & xec_mask;
403

B
Borislav Petkov 已提交
404
	pr_emerg(HW_ERR "Bus Unit Error");
405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437

	if (xec == 0x1)
		pr_cont(" in the write data buffers.\n");
	else if (xec == 0x3)
		pr_cont(" in the victim data buffers.\n");
	else if (xec == 0x2 && MEM_ERROR(ec))
		pr_cont(": %s error in the L2 cache tags.\n", RRRR_MSG(ec));
	else if (xec == 0x0) {
		if (TLB_ERROR(ec))
			pr_cont(": %s error in a Page Descriptor Cache or "
				"Guest TLB.\n", TT_MSG(ec));
		else if (BUS_ERROR(ec))
			pr_cont(": %s/ECC error in data read from NB: %s.\n",
				RRRR_MSG(ec), PP_MSG(ec));
		else if (MEM_ERROR(ec)) {
			u8 rrrr = (ec >> 4) & 0xf;

			if (rrrr >= 0x7)
				pr_cont(": %s error during data copyback.\n",
					RRRR_MSG(ec));
			else if (rrrr <= 0x1)
				pr_cont(": %s parity/ECC error during data "
					"access from L2.\n", RRRR_MSG(ec));
			else
				goto wrong_bu_mce;
		} else
			goto wrong_bu_mce;
	} else
		goto wrong_bu_mce;

	return;

wrong_bu_mce:
B
Borislav Petkov 已提交
438
	pr_emerg(HW_ERR "Corrupted BU MCE info?\n");
439 440
}

441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480
static void amd_decode_cu_mce(struct mce *m)
{
	u16 ec = m->status & 0xffff;
	u8 xec = (m->status >> 16) & xec_mask;

	pr_emerg(HW_ERR "Combined Unit Error: ");

	if (TLB_ERROR(ec)) {
		if (xec == 0x0)
			pr_cont("Data parity TLB read error.\n");
		else if (xec == 0x1)
			pr_cont("Poison data provided for TLB fill.\n");
		else
			goto wrong_cu_mce;
	} else if (BUS_ERROR(ec)) {
		if (xec > 2)
			goto wrong_cu_mce;

		pr_cont("Error during attempted NB data read.\n");
	} else if (MEM_ERROR(ec)) {
		switch (xec) {
		case 0x4 ... 0xc:
			pr_cont("%s.\n", f15h_cu_mce_desc[xec - 0x4]);
			break;

		case 0x10 ... 0x14:
			pr_cont("%s.\n", f15h_cu_mce_desc[xec - 0x7]);
			break;

		default:
			goto wrong_cu_mce;
		}
	}

	return;

wrong_cu_mce:
	pr_emerg(HW_ERR "Corrupted CU MCE info?\n");
}

481
static void amd_decode_ls_mce(struct mce *m)
482
{
483
	u16 ec = m->status & 0xffff;
484
	u8 xec = (m->status >> 16) & xec_mask;
485

486
	if (boot_cpu_data.x86 >= 0x14) {
487 488 489 490
		pr_emerg("You shouldn't be seeing an LS MCE on this cpu family,"
			 " please report on LKML.\n");
		return;
	}
491

B
Borislav Petkov 已提交
492
	pr_emerg(HW_ERR "Load Store Error");
493 494

	if (xec == 0x0) {
495
		u8 r4 = (ec >> 4) & 0xf;
496

497
		if (!BUS_ERROR(ec) || (r4 != R4_DRD && r4 != R4_DWR))
498 499 500
			goto wrong_ls_mce;

		pr_cont(" during %s.\n", RRRR_MSG(ec));
501 502 503
	} else
		goto wrong_ls_mce;

504 505 506
	return;

wrong_ls_mce:
B
Borislav Petkov 已提交
507
	pr_emerg(HW_ERR "Corrupted LS MCE info?\n");
508 509
}

510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528
static bool k8_nb_mce(u16 ec, u8 xec)
{
	bool ret = true;

	switch (xec) {
	case 0x1:
		pr_cont("CRC error detected on HT link.\n");
		break;

	case 0x5:
		pr_cont("Invalid GART PTE entry during GART table walk.\n");
		break;

	case 0x6:
		pr_cont("Unsupported atomic RMW received from an IO link.\n");
		break;

	case 0x0:
	case 0x8:
529 530 531
		if (boot_cpu_data.x86 == 0x11)
			return false;

532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574
		pr_cont("DRAM ECC error detected on the NB.\n");
		break;

	case 0xd:
		pr_cont("Parity error on the DRAM addr/ctl signals.\n");
		break;

	default:
		ret = false;
		break;
	}

	return ret;
}

static bool f10h_nb_mce(u16 ec, u8 xec)
{
	bool ret = true;
	u8 offset = 0;

	if (k8_nb_mce(ec, xec))
		return true;

	switch(xec) {
	case 0xa ... 0xc:
		offset = 10;
		break;

	case 0xe:
		offset = 11;
		break;

	case 0xf:
		if (TLB_ERROR(ec))
			pr_cont("GART Table Walk data error.\n");
		else if (BUS_ERROR(ec))
			pr_cont("DMA Exclusion Vector Table Walk error.\n");
		else
			ret = false;

		goto out;
		break;

575 576 577 578 579 580 581 582 583
	case 0x19:
		if (boot_cpu_data.x86 == 0x15)
			pr_cont("Compute Unit Data Error.\n");
		else
			ret = false;

		goto out;
		break;

584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600
	case 0x1c ... 0x1f:
		offset = 24;
		break;

	default:
		ret = false;

		goto out;
		break;
	}

	pr_cont("%s.\n", f10h_nb_mce_desc[xec - offset]);

out:
	return ret;
}

601
static bool nb_noop_mce(u16 ec, u8 xec)
602 603 604 605
{
	return false;
}

606
void amd_decode_nb_mce(int node_id, struct mce *m, u32 nbcfg)
607
{
608 609
	u8 xec   = (m->status >> 16) & 0x1f;
	u16 ec   = m->status & 0xffff;
610
	u32 nbsh = (u32)(m->status >> 32);
611

612
	pr_emerg(HW_ERR "Northbridge Error, node %d: ", node_id);
613 614 615 616 617 618

	/*
	 * F10h, revD can disable ErrCpu[3:0] so check that first and also the
	 * value encoding has changed so interpret those differently
	 */
	if ((boot_cpu_data.x86 == 0x10) &&
619
	    (boot_cpu_data.x86_model > 7)) {
620
		if (nbsh & K8_NBSH_ERR_CPU_VAL)
621
			pr_cont(", core: %u", (u8)(nbsh & nb_err_cpumask));
622
	} else {
623
		u8 assoc_cpus = nbsh & nb_err_cpumask;
624 625 626

		if (assoc_cpus > 0)
			pr_cont(", core: %d", fls(assoc_cpus) - 1);
627
	}
628

629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651
	switch (xec) {
	case 0x2:
		pr_cont("Sync error (sync packets on HT link detected).\n");
		return;

	case 0x3:
		pr_cont("HT Master abort.\n");
		return;

	case 0x4:
		pr_cont("HT Target abort.\n");
		return;

	case 0x7:
		pr_cont("NB Watchdog timeout.\n");
		return;

	case 0x9:
		pr_cont("SVM DMA Exclusion Vector error.\n");
		return;

	default:
		break;
652 653
	}

654 655 656 657 658 659
	if (!fam_ops->nb_mce(ec, xec))
		goto wrong_nb_mce;

	if (boot_cpu_data.x86 == 0xf || boot_cpu_data.x86 == 0x10)
		if ((xec == 0x8 || xec == 0x0) && nb_bus_decoder)
			nb_bus_decoder(node_id, m, nbcfg);
660

661 662 663 664
	return;

wrong_nb_mce:
	pr_emerg(HW_ERR "Corrupted NB MCE info?\n");
665 666 667
}
EXPORT_SYMBOL_GPL(amd_decode_nb_mce);

668
static void amd_decode_fr_mce(struct mce *m)
B
Borislav Petkov 已提交
669
{
670 671 672 673
	struct cpuinfo_x86 *c = &boot_cpu_data;
	u8 xec = (m->status >> 16) & xec_mask;

	if (c->x86 == 0xf || c->x86 == 0x11)
B
Borislav Petkov 已提交
674 675
		goto wrong_fr_mce;

676 677 678 679 680 681 682 683 684 685 686 687 688 689
	if (c->x86 != 0x15 && xec != 0x0)
		goto wrong_fr_mce;

	pr_emerg(HW_ERR "%s Error: ",
		 (c->x86 == 0x15 ? "Execution Unit" : "FIROB"));

	if (xec == 0x0 || xec == 0xc)
		pr_cont("%s.\n", fr_ex_mce_desc[xec]);
	else if (xec < 0xd)
		pr_cont("%s parity error.\n", fr_ex_mce_desc[xec]);
	else
		goto wrong_fr_mce;

	return;
B
Borislav Petkov 已提交
690 691 692

wrong_fr_mce:
	pr_emerg(HW_ERR "Corrupted FR MCE info?\n");
B
Borislav Petkov 已提交
693 694
}

695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734
static void amd_decode_fp_mce(struct mce *m)
{
	u8 xec = (m->status >> 16) & xec_mask;

	pr_emerg(HW_ERR "Floating Point Unit Error: ");

	switch (xec) {
	case 0x1:
		pr_cont("Free List");
		break;

	case 0x2:
		pr_cont("Physical Register File");
		break;

	case 0x3:
		pr_cont("Retire Queue");
		break;

	case 0x4:
		pr_cont("Scheduler table");
		break;

	case 0x5:
		pr_cont("Status Register File");
		break;

	default:
		goto wrong_fp_mce;
		break;
	}

	pr_cont(" parity error.\n");

	return;

wrong_fp_mce:
	pr_emerg(HW_ERR "Corrupted FP MCE info?\n");
}

B
Borislav Petkov 已提交
735
static inline void amd_decode_err_code(u16 ec)
736
{
737
	if (TLB_ERROR(ec)) {
B
Borislav Petkov 已提交
738
		pr_emerg(HW_ERR "Transaction: %s, Cache Level: %s\n",
739 740
			 TT_MSG(ec), LL_MSG(ec));
	} else if (MEM_ERROR(ec)) {
B
Borislav Petkov 已提交
741
		pr_emerg(HW_ERR "Transaction: %s, Type: %s, Cache Level: %s\n",
742 743
			 RRRR_MSG(ec), TT_MSG(ec), LL_MSG(ec));
	} else if (BUS_ERROR(ec)) {
B
Borislav Petkov 已提交
744
		pr_emerg(HW_ERR "Transaction: %s (%s), %s, Cache Level: %s, "
745 746 747 748
			 "Participating Processor: %s\n",
			  RRRR_MSG(ec), II_MSG(ec), TO_MSG(ec), LL_MSG(ec),
			  PP_MSG(ec));
	} else
B
Borislav Petkov 已提交
749
		pr_emerg(HW_ERR "Huh? Unknown MCE error 0x%x\n", ec);
750 751
}

752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767
/*
 * Filter out unwanted MCE signatures here.
 */
static bool amd_filter_mce(struct mce *m)
{
	u8 xec = (m->status >> 16) & 0x1f;

	/*
	 * NB GART TLB error reporting is disabled by default.
	 */
	if (m->bank == 4 && xec == 0x5 && !report_gart_errors)
		return true;

	return false;
}

B
Borislav Petkov 已提交
768
int amd_decode_mce(struct notifier_block *nb, unsigned long val, void *data)
769
{
770
	struct mce *m = (struct mce *)data;
771
	int node, ecc;
772

773 774 775
	if (amd_filter_mce(m))
		return NOTIFY_STOP;

B
Borislav Petkov 已提交
776
	pr_emerg(HW_ERR "MC%d_STATUS: ", m->bank);
777

778
	pr_cont("%sorrected error, other errors lost: %s, "
779 780
		 "CPU context corrupt: %s",
		 ((m->status & MCI_STATUS_UC) ? "Unc"  : "C"),
781
		 ((m->status & MCI_STATUS_OVER) ? "yes"  : "no"),
782
		 ((m->status & MCI_STATUS_PCC) ? "yes" : "no"));
783

784
	/* do the two bits[14:13] together */
785
	ecc = (m->status >> 45) & 0x3;
786 787 788 789 790
	if (ecc)
		pr_cont(", %sECC Error", ((ecc == 2) ? "C" : "U"));

	pr_cont("\n");

791 792
	switch (m->bank) {
	case 0:
793
		amd_decode_dc_mce(m);
794
		break;
795

796
	case 1:
797
		amd_decode_ic_mce(m);
798 799
		break;

800
	case 2:
801 802 803 804
		if (boot_cpu_data.x86 == 0x15)
			amd_decode_cu_mce(m);
		else
			amd_decode_bu_mce(m);
805 806
		break;

807
	case 3:
808
		amd_decode_ls_mce(m);
809 810
		break;

811
	case 4:
812 813
		node = amd_get_nb_id(m->extcpu);
		amd_decode_nb_mce(node, m, 0);
814 815
		break;

B
Borislav Petkov 已提交
816
	case 5:
817
		amd_decode_fr_mce(m);
B
Borislav Petkov 已提交
818 819
		break;

820 821 822 823
	case 6:
		amd_decode_fp_mce(m);
		break;

824 825
	default:
		break;
826
	}
827 828

	amd_decode_err_code(m->status & 0xffff);
829 830

	return NOTIFY_STOP;
831
}
B
Borislav Petkov 已提交
832
EXPORT_SYMBOL_GPL(amd_decode_mce);
833

834 835 836 837
static struct notifier_block amd_mce_dec_nb = {
	.notifier_call	= amd_decode_mce,
};

838 839
static int __init mce_amd_init(void)
{
840 841 842
	if (boot_cpu_data.x86_vendor != X86_VENDOR_AMD)
		return 0;

843
	if ((boot_cpu_data.x86 < 0xf || boot_cpu_data.x86 > 0x12) &&
844
	    (boot_cpu_data.x86 != 0x14 || boot_cpu_data.x86_model > 0xf))
845 846
		return 0;

847 848 849 850 851 852 853
	fam_ops = kzalloc(sizeof(struct amd_decoder_ops), GFP_KERNEL);
	if (!fam_ops)
		return -ENOMEM;

	switch (boot_cpu_data.x86) {
	case 0xf:
		fam_ops->dc_mce = k8_dc_mce;
854
		fam_ops->ic_mce = k8_ic_mce;
855
		fam_ops->nb_mce = k8_nb_mce;
856 857 858 859
		break;

	case 0x10:
		fam_ops->dc_mce = f10h_dc_mce;
860
		fam_ops->ic_mce = k8_ic_mce;
861
		fam_ops->nb_mce = f10h_nb_mce;
862 863
		break;

864 865 866 867 868 869
	case 0x11:
		fam_ops->dc_mce = k8_dc_mce;
		fam_ops->ic_mce = k8_ic_mce;
		fam_ops->nb_mce = f10h_nb_mce;
		break;

870 871
	case 0x12:
		fam_ops->dc_mce = f12h_dc_mce;
872
		fam_ops->ic_mce = k8_ic_mce;
873
		fam_ops->nb_mce = nb_noop_mce;
874 875
		break;

876
	case 0x14:
877
		nb_err_cpumask  = 0x3;
878
		fam_ops->dc_mce = f14h_dc_mce;
879
		fam_ops->ic_mce = f14h_ic_mce;
880
		fam_ops->nb_mce = nb_noop_mce;
881 882
		break;

883 884
	case 0x15:
		xec_mask = 0x1f;
885
		fam_ops->dc_mce = f15h_dc_mce;
886
		fam_ops->ic_mce = f15h_ic_mce;
887
		fam_ops->nb_mce = f10h_nb_mce;
888 889
		break;

890 891 892 893 894 895 896
	default:
		printk(KERN_WARNING "Huh? What family is that: %d?!\n",
				    boot_cpu_data.x86);
		kfree(fam_ops);
		return -EINVAL;
	}

897 898
	pr_info("MCE: In-kernel MCE decoding enabled.\n");

899
	atomic_notifier_chain_register(&x86_mce_decoder_chain, &amd_mce_dec_nb);
900 901 902 903

	return 0;
}
early_initcall(mce_amd_init);
904 905 906 907

#ifdef MODULE
static void __exit mce_amd_exit(void)
{
908
	atomic_notifier_chain_unregister(&x86_mce_decoder_chain, &amd_mce_dec_nb);
909
	kfree(fam_ops);
910 911 912 913 914 915 916
}

MODULE_DESCRIPTION("AMD MCE decoder");
MODULE_ALIAS("edac-mce-amd");
MODULE_LICENSE("GPL");
module_exit(mce_amd_exit);
#endif