forked from ~ljy/RK356X_SDK_RELEASE

blame | history | patch | commit | commitdiff
hc
2024-05-10 37f49e37ab4cb5d0bc4c60eb5c6d4dd57db767bb 
 kernel/drivers/edac/mce_amd.c
....@@ -1,3 +1,4 @@
1
+// SPDX-License-Identifier: GPL-2.0-only
12 #include <linux/module.h>
23 #include <linux/slab.h>
34 
....@@ -5,18 +6,11 @@
56 
67 #include "mce_amd.h"
78 
8
-static struct amd_decoder_ops *fam_ops;
9
+static struct amd_decoder_ops fam_ops;
910 
1011 static u8 xec_mask	 = 0xf;
1112 
12
-static bool report_gart_errors;
1313 static void (*decode_dram_ecc)(int node_id, struct mce *m);
14
-
15
-void amd_report_gart_errors(bool v)
16
-{
17
-	report_gart_errors = v;
18
-}
19
-EXPORT_SYMBOL_GPL(amd_report_gart_errors);
2014 
2115 void amd_register_ecc_decoder(void (*f)(int, struct mce *))
2216 {
....@@ -151,138 +145,259 @@
151145 
152146 /* Scalable MCA error strings */
153147 static const char * const smca_ls_mce_desc[] = {
154
-	"Load queue parity",
155
-	"Store queue parity",
156
-	"Miss address buffer payload parity",
157
-	"L1 TLB parity",
158
-	"Reserved",
159
-	"DC tag error type 6",
160
-	"DC tag error type 1",
148
+	"Load queue parity error",
149
+	"Store queue parity error",
150
+	"Miss address buffer payload parity error",
151
+	"Level 1 TLB parity error",
152
+	"DC Tag error type 5",
153
+	"DC Tag error type 6",
154
+	"DC Tag error type 1",
161155 	"Internal error type 1",
162156 	"Internal error type 2",
163
-	"Sys Read data error thread 0",
164
-	"Sys read data error thread 1",
165
-	"DC tag error type 2",
166
-	"DC data error type 1 (poison consumption)",
167
-	"DC data error type 2",
168
-	"DC data error type 3",
169
-	"DC tag error type 4",
170
-	"L2 TLB parity",
157
+	"System Read Data Error Thread 0",
158
+	"System Read Data Error Thread 1",
159
+	"DC Tag error type 2",
160
+	"DC Data error type 1 and poison consumption",
161
+	"DC Data error type 2",
162
+	"DC Data error type 3",
163
+	"DC Tag error type 4",
164
+	"Level 2 TLB parity error",
171165 	"PDC parity error",
172
-	"DC tag error type 3",
173
-	"DC tag error type 5",
174
-	"L2 fill data error",
166
+	"DC Tag error type 3",
167
+	"DC Tag error type 5",
168
+	"L2 Fill Data error",
169
+};
170
+
171
+static const char * const smca_ls2_mce_desc[] = {
172
+	"An ECC error was detected on a data cache read by a probe or victimization",
173
+	"An ECC error or L2 poison was detected on a data cache read by a load",
174
+	"An ECC error was detected on a data cache read-modify-write by a store",
175
+	"An ECC error or poison bit mismatch was detected on a tag read by a probe or victimization",
176
+	"An ECC error or poison bit mismatch was detected on a tag read by a load",
177
+	"An ECC error or poison bit mismatch was detected on a tag read by a store",
178
+	"An ECC error was detected on an EMEM read by a load",
179
+	"An ECC error was detected on an EMEM read-modify-write by a store",
180
+	"A parity error was detected in an L1 TLB entry by any access",
181
+	"A parity error was detected in an L2 TLB entry by any access",
182
+	"A parity error was detected in a PWC entry by any access",
183
+	"A parity error was detected in an STQ entry by any access",
184
+	"A parity error was detected in an LDQ entry by any access",
185
+	"A parity error was detected in a MAB entry by any access",
186
+	"A parity error was detected in an SCB entry state field by any access",
187
+	"A parity error was detected in an SCB entry address field by any access",
188
+	"A parity error was detected in an SCB entry data field by any access",
189
+	"A parity error was detected in a WCB entry by any access",
190
+	"A poisoned line was detected in an SCB entry by any access",
191
+	"A SystemReadDataError error was reported on read data returned from L2 for a load",
192
+	"A SystemReadDataError error was reported on read data returned from L2 for an SCB store",
193
+	"A SystemReadDataError error was reported on read data returned from L2 for a WCB store",
194
+	"A hardware assertion error was reported",
195
+	"A parity error was detected in an STLF, SCB EMEM entry or SRB store data by any access",
175196 };
176197 
177198 static const char * const smca_if_mce_desc[] = {
178
-	"microtag probe port parity error",
179
-	"IC microtag or full tag multi-hit error",
180
-	"IC full tag parity",
181
-	"IC data array parity",
182
-	"Decoupling queue phys addr parity error",
183
-	"L0 ITLB parity error",
184
-	"L1 ITLB parity error",
185
-	"L2 ITLB parity error",
186
-	"BPQ snoop parity on Thread 0",
187
-	"BPQ snoop parity on Thread 1",
188
-	"L1 BTB multi-match error",
189
-	"L2 BTB multi-match error",
190
-	"L2 Cache Response Poison error",
191
-	"System Read Data error",
199
+	"Op Cache Microtag Probe Port Parity Error",
200
+	"IC Microtag or Full Tag Multi-hit Error",
201
+	"IC Full Tag Parity Error",
202
+	"IC Data Array Parity Error",
203
+	"Decoupling Queue PhysAddr Parity Error",
204
+	"L0 ITLB Parity Error",
205
+	"L1 ITLB Parity Error",
206
+	"L2 ITLB Parity Error",
207
+	"BPQ Thread 0 Snoop Parity Error",
208
+	"BPQ Thread 1 Snoop Parity Error",
209
+	"L1 BTB Multi-Match Error",
210
+	"L2 BTB Multi-Match Error",
211
+	"L2 Cache Response Poison Error",
212
+	"System Read Data Error",
213
+	"Hardware Assertion Error",
214
+	"L1-TLB Multi-Hit",
215
+	"L2-TLB Multi-Hit",
216
+	"BSR Parity Error",
217
+	"CT MCE",
192218 };
193219 
194220 static const char * const smca_l2_mce_desc[] = {
195
-	"L2M tag multi-way-hit error",
196
-	"L2M tag ECC error",
197
-	"L2M data ECC error",
198
-	"HW assert",
221
+	"L2M Tag Multiple-Way-Hit error",
222
+	"L2M Tag or State Array ECC Error",
223
+	"L2M Data Array ECC Error",
224
+	"Hardware Assert Error",
199225 };
200226 
201227 static const char * const smca_de_mce_desc[] = {
202
-	"uop cache tag parity error",
203
-	"uop cache data parity error",
204
-	"Insn buffer parity error",
205
-	"uop queue parity error",
206
-	"Insn dispatch queue parity error",
207
-	"Fetch address FIFO parity",
208
-	"Patch RAM data parity",
209
-	"Patch RAM sequencer parity",
210
-	"uop buffer parity"
228
+	"Micro-op cache tag parity error",
229
+	"Micro-op cache data parity error",
230
+	"Instruction buffer parity error",
231
+	"Micro-op queue parity error",
232
+	"Instruction dispatch queue parity error",
233
+	"Fetch address FIFO parity error",
234
+	"Patch RAM data parity error",
235
+	"Patch RAM sequencer parity error",
236
+	"Micro-op buffer parity error",
237
+	"Hardware Assertion MCA Error",
211238 };
212239 
213240 static const char * const smca_ex_mce_desc[] = {
214
-	"Watchdog timeout error",
215
-	"Phy register file parity",
216
-	"Flag register file parity",
217
-	"Immediate displacement register file parity",
218
-	"Address generator payload parity",
219
-	"EX payload parity",
220
-	"Checkpoint queue parity",
221
-	"Retire dispatch queue parity",
241
+	"Watchdog Timeout error",
242
+	"Physical register file parity error",
243
+	"Flag register file parity error",
244
+	"Immediate displacement register file parity error",
245
+	"Address generator payload parity error",
246
+	"EX payload parity error",
247
+	"Checkpoint queue parity error",
248
+	"Retire dispatch queue parity error",
222249 	"Retire status queue parity error",
223250 	"Scheduling queue parity error",
224251 	"Branch buffer queue parity error",
252
+	"Hardware Assertion error",
253
+	"Spec Map parity error",
254
+	"Retire Map parity error",
225255 };
226256 
227257 static const char * const smca_fp_mce_desc[] = {
228
-	"Physical register file parity",
229
-	"Freelist parity error",
230
-	"Schedule queue parity",
258
+	"Physical register file (PRF) parity error",
259
+	"Freelist (FL) parity error",
260
+	"Schedule queue parity error",
231261 	"NSQ parity error",
232
-	"Retire queue parity",
233
-	"Status register file parity",
262
+	"Retire queue (RQ) parity error",
263
+	"Status register file (SRF) parity error",
234264 	"Hardware assertion",
235265 };
236266 
237267 static const char * const smca_l3_mce_desc[] = {
238
-	"Shadow tag macro ECC error",
239
-	"Shadow tag macro multi-way-hit error",
240
-	"L3M tag ECC error",
241
-	"L3M tag multi-way-hit error",
242
-	"L3M data ECC error",
243
-	"XI parity, L3 fill done channel error",
244
-	"L3 victim queue parity",
245
-	"L3 HW assert",
268
+	"Shadow Tag Macro ECC Error",
269
+	"Shadow Tag Macro Multi-way-hit Error",
270
+	"L3M Tag ECC Error",
271
+	"L3M Tag Multi-way-hit Error",
272
+	"L3M Data ECC Error",
273
+	"SDP Parity Error or SystemReadDataError from XI",
274
+	"L3 Victim Queue Parity Error",
275
+	"L3 Hardware Assertion",
246276 };
247277 
248278 static const char * const smca_cs_mce_desc[] = {
249
-	"Illegal request from transport layer",
250
-	"Address violation",
251
-	"Security violation",
252
-	"Illegal response from transport layer",
253
-	"Unexpected response",
254
-	"Parity error on incoming request or probe response data",
255
-	"Parity error on incoming read response data",
256
-	"Atomic request parity",
257
-	"ECC error on probe filter access",
279
+	"Illegal Request",
280
+	"Address Violation",
281
+	"Security Violation",
282
+	"Illegal Response",
283
+	"Unexpected Response",
284
+	"Request or Probe Parity Error",
285
+	"Read Response Parity Error",
286
+	"Atomic Request Parity Error",
287
+	"Probe Filter ECC Error",
288
+};
289
+
290
+static const char * const smca_cs2_mce_desc[] = {
291
+	"Illegal Request",
292
+	"Address Violation",
293
+	"Security Violation",
294
+	"Illegal Response",
295
+	"Unexpected Response",
296
+	"Request or Probe Parity Error",
297
+	"Read Response Parity Error",
298
+	"Atomic Request Parity Error",
299
+	"SDP read response had no match in the CS queue",
300
+	"Probe Filter Protocol Error",
301
+	"Probe Filter ECC Error",
302
+	"SDP read response had an unexpected RETRY error",
303
+	"Counter overflow error",
304
+	"Counter underflow error",
258305 };
259306 
260307 static const char * const smca_pie_mce_desc[] = {
261
-	"HW assert",
262
-	"Internal PIE register security violation",
263
-	"Error on GMI link",
264
-	"Poison data written to internal PIE register",
308
+	"Hardware Assert",
309
+	"Register security violation",
310
+	"Link Error",
311
+	"Poison data consumption",
312
+	"A deferred error was detected in the DF"
265313 };
266314 
267315 static const char * const smca_umc_mce_desc[] = {
268316 	"DRAM ECC error",
269
-	"Data poison error on DRAM",
317
+	"Data poison error",
270318 	"SDP parity error",
271319 	"Advanced peripheral bus error",
272
-	"Command/address parity error",
320
+	"Address/Command parity error",
273321 	"Write data CRC error",
322
+	"DCQ SRAM ECC error",
323
+	"AES SRAM ECC error",
274324 };
275325 
276326 static const char * const smca_pb_mce_desc[] = {
277
-	"Parameter Block RAM ECC error",
327
+	"An ECC error in the Parameter Block RAM array",
278328 };
279329 
280330 static const char * const smca_psp_mce_desc[] = {
281
-	"PSP RAM ECC or parity error",
331
+	"An ECC or parity error in a PSP RAM instance",
332
+};
333
+
334
+static const char * const smca_psp2_mce_desc[] = {
335
+	"High SRAM ECC or parity error",
336
+	"Low SRAM ECC or parity error",
337
+	"Instruction Cache Bank 0 ECC or parity error",
338
+	"Instruction Cache Bank 1 ECC or parity error",
339
+	"Instruction Tag Ram 0 parity error",
340
+	"Instruction Tag Ram 1 parity error",
341
+	"Data Cache Bank 0 ECC or parity error",
342
+	"Data Cache Bank 1 ECC or parity error",
343
+	"Data Cache Bank 2 ECC or parity error",
344
+	"Data Cache Bank 3 ECC or parity error",
345
+	"Data Tag Bank 0 parity error",
346
+	"Data Tag Bank 1 parity error",
347
+	"Data Tag Bank 2 parity error",
348
+	"Data Tag Bank 3 parity error",
349
+	"Dirty Data Ram parity error",
350
+	"TLB Bank 0 parity error",
351
+	"TLB Bank 1 parity error",
352
+	"System Hub Read Buffer ECC or parity error",
282353 };
283354 
284355 static const char * const smca_smu_mce_desc[] = {
285
-	"SMU RAM ECC or parity error",
356
+	"An ECC or parity error in an SMU RAM instance",
357
+};
358
+
359
+static const char * const smca_smu2_mce_desc[] = {
360
+	"High SRAM ECC or parity error",
361
+	"Low SRAM ECC or parity error",
362
+	"Data Cache Bank A ECC or parity error",
363
+	"Data Cache Bank B ECC or parity error",
364
+	"Data Tag Cache Bank A ECC or parity error",
365
+	"Data Tag Cache Bank B ECC or parity error",
366
+	"Instruction Cache Bank A ECC or parity error",
367
+	"Instruction Cache Bank B ECC or parity error",
368
+	"Instruction Tag Cache Bank A ECC or parity error",
369
+	"Instruction Tag Cache Bank B ECC or parity error",
370
+	"System Hub Read Buffer ECC or parity error",
371
+	"PHY RAM ECC error",
372
+};
373
+
374
+static const char * const smca_mp5_mce_desc[] = {
375
+	"High SRAM ECC or parity error",
376
+	"Low SRAM ECC or parity error",
377
+	"Data Cache Bank A ECC or parity error",
378
+	"Data Cache Bank B ECC or parity error",
379
+	"Data Tag Cache Bank A ECC or parity error",
380
+	"Data Tag Cache Bank B ECC or parity error",
381
+	"Instruction Cache Bank A ECC or parity error",
382
+	"Instruction Cache Bank B ECC or parity error",
383
+	"Instruction Tag Cache Bank A ECC or parity error",
384
+	"Instruction Tag Cache Bank B ECC or parity error",
385
+};
386
+
387
+static const char * const smca_nbio_mce_desc[] = {
388
+	"ECC or Parity error",
389
+	"PCIE error",
390
+	"SDP ErrEvent error",
391
+	"SDP Egress Poison Error",
392
+	"IOHC Internal Poison Error",
393
+};
394
+
395
+static const char * const smca_pcie_mce_desc[] = {
396
+	"CCIX PER Message logging",
397
+	"CCIX Read Response with Status: Non-Data Error",
398
+	"CCIX Write Response with Status: Non-Data Error",
399
+	"CCIX Read Response with Status: Data Error",
400
+	"CCIX Non-okay write response with data error",
286401 };
287402 
288403 struct smca_mce_desc {
....@@ -292,6 +407,7 @@
292407 
293408 static struct smca_mce_desc smca_mce_descs[] = {
294409 	[SMCA_LS]	= { smca_ls_mce_desc,	ARRAY_SIZE(smca_ls_mce_desc)	},
410
+	[SMCA_LS_V2]	= { smca_ls2_mce_desc,	ARRAY_SIZE(smca_ls2_mce_desc)	},
295411 	[SMCA_IF]	= { smca_if_mce_desc,	ARRAY_SIZE(smca_if_mce_desc)	},
296412 	[SMCA_L2_CACHE]	= { smca_l2_mce_desc,	ARRAY_SIZE(smca_l2_mce_desc)	},
297413 	[SMCA_DE]	= { smca_de_mce_desc,	ARRAY_SIZE(smca_de_mce_desc)	},
....@@ -299,11 +415,17 @@
299415 	[SMCA_FP]	= { smca_fp_mce_desc,	ARRAY_SIZE(smca_fp_mce_desc)	},
300416 	[SMCA_L3_CACHE]	= { smca_l3_mce_desc,	ARRAY_SIZE(smca_l3_mce_desc)	},
301417 	[SMCA_CS]	= { smca_cs_mce_desc,	ARRAY_SIZE(smca_cs_mce_desc)	},
418
+	[SMCA_CS_V2]	= { smca_cs2_mce_desc,	ARRAY_SIZE(smca_cs2_mce_desc)	},
302419 	[SMCA_PIE]	= { smca_pie_mce_desc,	ARRAY_SIZE(smca_pie_mce_desc)	},
303420 	[SMCA_UMC]	= { smca_umc_mce_desc,	ARRAY_SIZE(smca_umc_mce_desc)	},
304421 	[SMCA_PB]	= { smca_pb_mce_desc,	ARRAY_SIZE(smca_pb_mce_desc)	},
305422 	[SMCA_PSP]	= { smca_psp_mce_desc,	ARRAY_SIZE(smca_psp_mce_desc)	},
423
+	[SMCA_PSP_V2]	= { smca_psp2_mce_desc,	ARRAY_SIZE(smca_psp2_mce_desc)	},
306424 	[SMCA_SMU]	= { smca_smu_mce_desc,	ARRAY_SIZE(smca_smu_mce_desc)	},
425
+	[SMCA_SMU_V2]	= { smca_smu2_mce_desc,	ARRAY_SIZE(smca_smu2_mce_desc)	},
426
+	[SMCA_MP5]	= { smca_mp5_mce_desc,	ARRAY_SIZE(smca_mp5_mce_desc)	},
427
+	[SMCA_NBIO]	= { smca_nbio_mce_desc,	ARRAY_SIZE(smca_nbio_mce_desc)	},
428
+	[SMCA_PCIE]	= { smca_pcie_mce_desc,	ARRAY_SIZE(smca_pcie_mce_desc)	},
307429 };
308430 
309431 static bool f12h_mc0_mce(u16 ec, u8 xec)
....@@ -463,7 +585,7 @@
463585 					    : (xec ? "multimatch" : "parity")));
464586 			return;
465587 		}
466
-	} else if (fam_ops->mc0_mce(ec, xec))
588
+	} else if (fam_ops.mc0_mce(ec, xec))
467589 		;
468590 	else
469591 		pr_emerg(HW_ERR "Corrupted MC0 MCE info?\n");
....@@ -577,7 +699,7 @@
577699 			pr_cont("Hardware Assert.\n");
578700 		else
579701 			goto wrong_mc1_mce;
580
-	} else if (fam_ops->mc1_mce(ec, xec))
702
+	} else if (fam_ops.mc1_mce(ec, xec))
581703 		;
582704 	else
583705 		goto wrong_mc1_mce;
....@@ -711,7 +833,7 @@
711833 
712834 	pr_emerg(HW_ERR "MC2 Error: ");
713835 
714
-	if (!fam_ops->mc2_mce(ec, xec))
836
+	if (!fam_ops.mc2_mce(ec, xec))
715837 		pr_cont(HW_ERR "Corrupted MC2 MCE info?\n");
716838 }
717839 
....@@ -874,17 +996,14 @@
874996 
875997 	ip_name = smca_get_long_name(bank_type);
876998 
877
-	pr_emerg(HW_ERR "%s Extended Error Code: %d\n", ip_name, xec);
999
+	pr_emerg(HW_ERR "%s Ext. Error Code: %d", ip_name, xec);
8781000 
8791001 	/* Only print the decode of valid error codes */
880
-	if (xec < smca_mce_descs[bank_type].num_descs &&
881
-			(hwid->xec_bitmap & BIT_ULL(xec))) {
882
-		pr_emerg(HW_ERR "%s Error: ", ip_name);
883
-		pr_cont("%s.\n", smca_mce_descs[bank_type].descs[xec]);
884
-	}
1002
+	if (xec < smca_mce_descs[bank_type].num_descs)
1003
+		pr_cont(", %s.\n", smca_mce_descs[bank_type].descs[xec]);
8851004 
8861005 	if (bank_type == SMCA_UMC && xec == 0 && decode_dram_ecc)
887
-		decode_dram_ecc(cpu_to_node(m->extcpu), m);
1006
+		decode_dram_ecc(topology_die_id(m->extcpu), m);
8881007 }
8891008 
8901009 static inline void amd_decode_err_code(u16 ec)
....@@ -911,20 +1030,6 @@
9111030 	pr_cont("\n");
9121031 }
9131032 
914
-/*
915
- * Filter out unwanted MCE signatures here.
916
- */
917
-static bool amd_filter_mce(struct mce *m)
918
-{
919
-	/*
920
-	 * NB GART TLB error reporting is disabled by default.
921
-	 */
922
-	if (m->bank == 4 && XEC(m->status, 0x1f) == 0x5 && !report_gart_errors)
923
-		return true;
924
-
925
-	return false;
926
-}
927
-
9281033 static const char *decode_error_status(struct mce *m)
9291034 {
9301035 	if (m->status & MCI_STATUS_UC) {
....@@ -948,8 +1053,8 @@
9481053 	unsigned int fam = x86_family(m->cpuid);
9491054 	int ecc;
9501055 
951
-	if (amd_filter_mce(m))
952
-		return NOTIFY_STOP;
1056
+	if (m->kflags & MCE_HANDLED_CEC)
1057
+		return NOTIFY_DONE;
9531058 
9541059 	pr_emerg(HW_ERR "%s\n", decode_error_status(m));
9551060 
....@@ -961,8 +1066,24 @@
9611066 		((m->status & MCI_STATUS_UC)	? "UE"	  :
9621067 		 (m->status & MCI_STATUS_DEFERRED) ? "-"  : "CE"),
9631068 		((m->status & MCI_STATUS_MISCV)	? "MiscV" : "-"),
964
-		((m->status & MCI_STATUS_PCC)	? "PCC"	  : "-"),
965
-		((m->status & MCI_STATUS_ADDRV)	? "AddrV" : "-"));
1069
+		((m->status & MCI_STATUS_ADDRV)	? "AddrV" : "-"),
1070
+		((m->status & MCI_STATUS_PCC)	? "PCC"	  : "-"));
1071
+
1072
+	if (boot_cpu_has(X86_FEATURE_SMCA)) {
1073
+		u32 low, high;
1074
+		u32 addr = MSR_AMD64_SMCA_MCx_CONFIG(m->bank);
1075
+
1076
+		if (!rdmsr_safe(addr, &low, &high) &&
1077
+		    (low & MCI_CONFIG_MCAX))
1078
+			pr_cont("|%s", ((m->status & MCI_STATUS_TCC) ? "TCC" : "-"));
1079
+
1080
+		pr_cont("|%s", ((m->status & MCI_STATUS_SYNDV) ? "SyndV" : "-"));
1081
+	}
1082
+
1083
+	/* do the two bits[14:13] together */
1084
+	ecc = (m->status >> 45) & 0x3;
1085
+	if (ecc)
1086
+		pr_cont("|%sECC", ((ecc == 2) ? "C" : "U"));
9661087 
9671088 	if (fam >= 0x15) {
9681089 		pr_cont("|%s", (m->status & MCI_STATUS_DEFERRED ? "Deferred" : "-"));
....@@ -972,26 +1093,16 @@
9721093 			pr_cont("|%s", (m->status & MCI_STATUS_POISON ? "Poison" : "-"));
9731094 	}
9741095 
975
-	if (boot_cpu_has(X86_FEATURE_SMCA)) {
976
-		u32 low, high;
977
-		u32 addr = MSR_AMD64_SMCA_MCx_CONFIG(m->bank);
978
-
979
-		pr_cont("|%s", ((m->status & MCI_STATUS_SYNDV) ? "SyndV" : "-"));
980
-
981
-		if (!rdmsr_safe(addr, &low, &high) &&
982
-		    (low & MCI_CONFIG_MCAX))
983
-			pr_cont("|%s", ((m->status & MCI_STATUS_TCC) ? "TCC" : "-"));
984
-	}
985
-
986
-	/* do the two bits[14:13] together */
987
-	ecc = (m->status >> 45) & 0x3;
988
-	if (ecc)
989
-		pr_cont("|%sECC", ((ecc == 2) ? "C" : "U"));
1096
+	if (fam >= 0x17)
1097
+		pr_cont("|%s", (m->status & MCI_STATUS_SCRUB ? "Scrub" : "-"));
9901098 
9911099 	pr_cont("]: 0x%016llx\n", m->status);
9921100 
9931101 	if (m->status & MCI_STATUS_ADDRV)
9941102 		pr_emerg(HW_ERR "Error Addr: 0x%016llx\n", m->addr);
1103
+
1104
+	if (m->ppin)
1105
+		pr_emerg(HW_ERR "PPIN: 0x%016llx\n", m->ppin);
9951106 
9961107 	if (boot_cpu_has(X86_FEATURE_SMCA)) {
9971108 		pr_emerg(HW_ERR "IPID: 0x%016llx", m->ipid);
....@@ -1008,7 +1119,8 @@
10081119 	if (m->tsc)
10091120 		pr_emerg(HW_ERR "TSC: %llu\n", m->tsc);
10101121 
1011
-	if (!fam_ops)
1122
+	/* Doesn't matter which member to test. */
1123
+	if (!fam_ops.mc0_mce)
10121124 		goto err_code;
10131125 
10141126 	switch (m->bank) {
....@@ -1047,7 +1159,8 @@
10471159  err_code:
10481160 	amd_decode_err_code(m->status & 0xffff);
10491161 
1050
-	return NOTIFY_STOP;
1162
+	m->kflags |= MCE_HANDLED_EDAC;
1163
+	return NOTIFY_OK;
10511164 }
10521165 
10531166 static struct notifier_block amd_mce_dec_nb = {
....@@ -1059,82 +1172,80 @@
10591172 {
10601173 	struct cpuinfo_x86 *c = &boot_cpu_data;
10611174 
1062
-	if (c->x86_vendor != X86_VENDOR_AMD)
1175
+	if (c->x86_vendor != X86_VENDOR_AMD &&
1176
+	    c->x86_vendor != X86_VENDOR_HYGON)
10631177 		return -ENODEV;
10641178 
1065
-	fam_ops = kzalloc(sizeof(struct amd_decoder_ops), GFP_KERNEL);
1066
-	if (!fam_ops)
1067
-		return -ENOMEM;
1179
+	if (cpu_feature_enabled(X86_FEATURE_HYPERVISOR))
1180
+		return -ENODEV;
1181
+
1182
+	if (boot_cpu_has(X86_FEATURE_SMCA)) {
1183
+		xec_mask = 0x3f;
1184
+		goto out;
1185
+	}
10681186 
10691187 	switch (c->x86) {
10701188 	case 0xf:
1071
-		fam_ops->mc0_mce = k8_mc0_mce;
1072
-		fam_ops->mc1_mce = k8_mc1_mce;
1073
-		fam_ops->mc2_mce = k8_mc2_mce;
1189
+		fam_ops.mc0_mce = k8_mc0_mce;
1190
+		fam_ops.mc1_mce = k8_mc1_mce;
1191
+		fam_ops.mc2_mce = k8_mc2_mce;
10741192 		break;
10751193 
10761194 	case 0x10:
1077
-		fam_ops->mc0_mce = f10h_mc0_mce;
1078
-		fam_ops->mc1_mce = k8_mc1_mce;
1079
-		fam_ops->mc2_mce = k8_mc2_mce;
1195
+		fam_ops.mc0_mce = f10h_mc0_mce;
1196
+		fam_ops.mc1_mce = k8_mc1_mce;
1197
+		fam_ops.mc2_mce = k8_mc2_mce;
10801198 		break;
10811199 
10821200 	case 0x11:
1083
-		fam_ops->mc0_mce = k8_mc0_mce;
1084
-		fam_ops->mc1_mce = k8_mc1_mce;
1085
-		fam_ops->mc2_mce = k8_mc2_mce;
1201
+		fam_ops.mc0_mce = k8_mc0_mce;
1202
+		fam_ops.mc1_mce = k8_mc1_mce;
1203
+		fam_ops.mc2_mce = k8_mc2_mce;
10861204 		break;
10871205 
10881206 	case 0x12:
1089
-		fam_ops->mc0_mce = f12h_mc0_mce;
1090
-		fam_ops->mc1_mce = k8_mc1_mce;
1091
-		fam_ops->mc2_mce = k8_mc2_mce;
1207
+		fam_ops.mc0_mce = f12h_mc0_mce;
1208
+		fam_ops.mc1_mce = k8_mc1_mce;
1209
+		fam_ops.mc2_mce = k8_mc2_mce;
10921210 		break;
10931211 
10941212 	case 0x14:
1095
-		fam_ops->mc0_mce = cat_mc0_mce;
1096
-		fam_ops->mc1_mce = cat_mc1_mce;
1097
-		fam_ops->mc2_mce = k8_mc2_mce;
1213
+		fam_ops.mc0_mce = cat_mc0_mce;
1214
+		fam_ops.mc1_mce = cat_mc1_mce;
1215
+		fam_ops.mc2_mce = k8_mc2_mce;
10981216 		break;
10991217 
11001218 	case 0x15:
11011219 		xec_mask = c->x86_model == 0x60 ? 0x3f : 0x1f;
11021220 
1103
-		fam_ops->mc0_mce = f15h_mc0_mce;
1104
-		fam_ops->mc1_mce = f15h_mc1_mce;
1105
-		fam_ops->mc2_mce = f15h_mc2_mce;
1221
+		fam_ops.mc0_mce = f15h_mc0_mce;
1222
+		fam_ops.mc1_mce = f15h_mc1_mce;
1223
+		fam_ops.mc2_mce = f15h_mc2_mce;
11061224 		break;
11071225 
11081226 	case 0x16:
11091227 		xec_mask = 0x1f;
1110
-		fam_ops->mc0_mce = cat_mc0_mce;
1111
-		fam_ops->mc1_mce = cat_mc1_mce;
1112
-		fam_ops->mc2_mce = f16h_mc2_mce;
1228
+		fam_ops.mc0_mce = cat_mc0_mce;
1229
+		fam_ops.mc1_mce = cat_mc1_mce;
1230
+		fam_ops.mc2_mce = f16h_mc2_mce;
11131231 		break;
11141232 
11151233 	case 0x17:
1116
-		xec_mask = 0x3f;
1117
-		if (!boot_cpu_has(X86_FEATURE_SMCA)) {
1118
-			printk(KERN_WARNING "Decoding supported only on Scalable MCA processors.\n");
1119
-			goto err_out;
1120
-		}
1121
-		break;
1234
+	case 0x18:
1235
+		pr_warn_once("Decoding supported only on Scalable MCA processors.\n");
1236
+		return -EINVAL;
11221237 
11231238 	default:
11241239 		printk(KERN_WARNING "Huh? What family is it: 0x%x?!\n", c->x86);
1125
-		goto err_out;
1240
+		return -EINVAL;
11261241 	}
11271242 
1243
+out:
11281244 	pr_info("MCE: In-kernel MCE decoding enabled.\n");
11291245 
11301246 	mce_register_decode_chain(&amd_mce_dec_nb);
11311247 
11321248 	return 0;
1133
-
1134
-err_out:
1135
-	kfree(fam_ops);
1136
-	fam_ops = NULL;
1137
-	return -EINVAL;
11381249 }
11391250 early_initcall(mce_amd_init);
11401251 
....@@ -1142,7 +1253,6 @@
11421253 static void __exit mce_amd_exit(void)
11431254 {
11441255 	mce_unregister_decode_chain(&amd_mce_dec_nb);
1145
-	kfree(fam_ops);
11461256 }
11471257 
11481258 MODULE_DESCRIPTION("AMD MCE decoder");