1 /* 2 * UEFI Common Platform Error Record (CPER) support 3 * 4 * Copyright (C) 2010, Intel Corp. 5 * Author: Huang Ying <ying.huang@intel.com> 6 * 7 * CPER is the format used to describe platform hardware error by 8 * various tables, such as ERST, BERT and HEST etc. 9 * 10 * For more information about CPER, please refer to Appendix N of UEFI 11 * Specification version 2.4. 12 * 13 * This program is free software; you can redistribute it and/or 14 * modify it under the terms of the GNU General Public License version 15 * 2 as published by the Free Software Foundation. 16 * 17 * This program is distributed in the hope that it will be useful, 18 * but WITHOUT ANY WARRANTY; without even the implied warranty of 19 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 20 * GNU General Public License for more details. 21 * 22 * You should have received a copy of the GNU General Public License 23 * along with this program; if not, write to the Free Software 24 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA 25 */ 26 27 #include <linux/kernel.h> 28 #include <linux/module.h> 29 #include <linux/time.h> 30 #include <linux/cper.h> 31 #include <linux/dmi.h> 32 #include <linux/acpi.h> 33 #include <linux/pci.h> 34 #include <linux/aer.h> 35 36 #define INDENT_SP " " 37 38 static char rcd_decode_str[CPER_REC_LEN]; 39 40 /* 41 * CPER record ID need to be unique even after reboot, because record 42 * ID is used as index for ERST storage, while CPER records from 43 * multiple boot may co-exist in ERST. 44 */ 45 u64 cper_next_record_id(void) 46 { 47 static atomic64_t seq; 48 49 if (!atomic64_read(&seq)) 50 atomic64_set(&seq, ((u64)get_seconds()) << 32); 51 52 return atomic64_inc_return(&seq); 53 } 54 EXPORT_SYMBOL_GPL(cper_next_record_id); 55 56 static const char * const severity_strs[] = { 57 "recoverable", 58 "fatal", 59 "corrected", 60 "info", 61 }; 62 63 const char *cper_severity_str(unsigned int severity) 64 { 65 return severity < ARRAY_SIZE(severity_strs) ? 66 severity_strs[severity] : "unknown"; 67 } 68 EXPORT_SYMBOL_GPL(cper_severity_str); 69 70 /* 71 * cper_print_bits - print strings for set bits 72 * @pfx: prefix for each line, including log level and prefix string 73 * @bits: bit mask 74 * @strs: string array, indexed by bit position 75 * @strs_size: size of the string array: @strs 76 * 77 * For each set bit in @bits, print the corresponding string in @strs. 78 * If the output length is longer than 80, multiple line will be 79 * printed, with @pfx is printed at the beginning of each line. 80 */ 81 void cper_print_bits(const char *pfx, unsigned int bits, 82 const char * const strs[], unsigned int strs_size) 83 { 84 int i, len = 0; 85 const char *str; 86 char buf[84]; 87 88 for (i = 0; i < strs_size; i++) { 89 if (!(bits & (1U << i))) 90 continue; 91 str = strs[i]; 92 if (!str) 93 continue; 94 if (len && len + strlen(str) + 2 > 80) { 95 printk("%s\n", buf); 96 len = 0; 97 } 98 if (!len) 99 len = snprintf(buf, sizeof(buf), "%s%s", pfx, str); 100 else 101 len += snprintf(buf+len, sizeof(buf)-len, ", %s", str); 102 } 103 if (len) 104 printk("%s\n", buf); 105 } 106 107 static const char * const proc_type_strs[] = { 108 "IA32/X64", 109 "IA64", 110 }; 111 112 static const char * const proc_isa_strs[] = { 113 "IA32", 114 "IA64", 115 "X64", 116 }; 117 118 static const char * const proc_error_type_strs[] = { 119 "cache error", 120 "TLB error", 121 "bus error", 122 "micro-architectural error", 123 }; 124 125 static const char * const proc_op_strs[] = { 126 "unknown or generic", 127 "data read", 128 "data write", 129 "instruction execution", 130 }; 131 132 static const char * const proc_flag_strs[] = { 133 "restartable", 134 "precise IP", 135 "overflow", 136 "corrected", 137 }; 138 139 static void cper_print_proc_generic(const char *pfx, 140 const struct cper_sec_proc_generic *proc) 141 { 142 if (proc->validation_bits & CPER_PROC_VALID_TYPE) 143 printk("%s""processor_type: %d, %s\n", pfx, proc->proc_type, 144 proc->proc_type < ARRAY_SIZE(proc_type_strs) ? 145 proc_type_strs[proc->proc_type] : "unknown"); 146 if (proc->validation_bits & CPER_PROC_VALID_ISA) 147 printk("%s""processor_isa: %d, %s\n", pfx, proc->proc_isa, 148 proc->proc_isa < ARRAY_SIZE(proc_isa_strs) ? 149 proc_isa_strs[proc->proc_isa] : "unknown"); 150 if (proc->validation_bits & CPER_PROC_VALID_ERROR_TYPE) { 151 printk("%s""error_type: 0x%02x\n", pfx, proc->proc_error_type); 152 cper_print_bits(pfx, proc->proc_error_type, 153 proc_error_type_strs, 154 ARRAY_SIZE(proc_error_type_strs)); 155 } 156 if (proc->validation_bits & CPER_PROC_VALID_OPERATION) 157 printk("%s""operation: %d, %s\n", pfx, proc->operation, 158 proc->operation < ARRAY_SIZE(proc_op_strs) ? 159 proc_op_strs[proc->operation] : "unknown"); 160 if (proc->validation_bits & CPER_PROC_VALID_FLAGS) { 161 printk("%s""flags: 0x%02x\n", pfx, proc->flags); 162 cper_print_bits(pfx, proc->flags, proc_flag_strs, 163 ARRAY_SIZE(proc_flag_strs)); 164 } 165 if (proc->validation_bits & CPER_PROC_VALID_LEVEL) 166 printk("%s""level: %d\n", pfx, proc->level); 167 if (proc->validation_bits & CPER_PROC_VALID_VERSION) 168 printk("%s""version_info: 0x%016llx\n", pfx, proc->cpu_version); 169 if (proc->validation_bits & CPER_PROC_VALID_ID) 170 printk("%s""processor_id: 0x%016llx\n", pfx, proc->proc_id); 171 if (proc->validation_bits & CPER_PROC_VALID_TARGET_ADDRESS) 172 printk("%s""target_address: 0x%016llx\n", 173 pfx, proc->target_addr); 174 if (proc->validation_bits & CPER_PROC_VALID_REQUESTOR_ID) 175 printk("%s""requestor_id: 0x%016llx\n", 176 pfx, proc->requestor_id); 177 if (proc->validation_bits & CPER_PROC_VALID_RESPONDER_ID) 178 printk("%s""responder_id: 0x%016llx\n", 179 pfx, proc->responder_id); 180 if (proc->validation_bits & CPER_PROC_VALID_IP) 181 printk("%s""IP: 0x%016llx\n", pfx, proc->ip); 182 } 183 184 static const char * const mem_err_type_strs[] = { 185 "unknown", 186 "no error", 187 "single-bit ECC", 188 "multi-bit ECC", 189 "single-symbol chipkill ECC", 190 "multi-symbol chipkill ECC", 191 "master abort", 192 "target abort", 193 "parity error", 194 "watchdog timeout", 195 "invalid address", 196 "mirror Broken", 197 "memory sparing", 198 "scrub corrected error", 199 "scrub uncorrected error", 200 "physical memory map-out event", 201 }; 202 203 const char *cper_mem_err_type_str(unsigned int etype) 204 { 205 return etype < ARRAY_SIZE(mem_err_type_strs) ? 206 mem_err_type_strs[etype] : "unknown"; 207 } 208 EXPORT_SYMBOL_GPL(cper_mem_err_type_str); 209 210 static int cper_mem_err_location(struct cper_mem_err_compact *mem, char *msg) 211 { 212 u32 len, n; 213 214 if (!msg) 215 return 0; 216 217 n = 0; 218 len = CPER_REC_LEN - 1; 219 if (mem->validation_bits & CPER_MEM_VALID_NODE) 220 n += scnprintf(msg + n, len - n, "node: %d ", mem->node); 221 if (mem->validation_bits & CPER_MEM_VALID_CARD) 222 n += scnprintf(msg + n, len - n, "card: %d ", mem->card); 223 if (mem->validation_bits & CPER_MEM_VALID_MODULE) 224 n += scnprintf(msg + n, len - n, "module: %d ", mem->module); 225 if (mem->validation_bits & CPER_MEM_VALID_RANK_NUMBER) 226 n += scnprintf(msg + n, len - n, "rank: %d ", mem->rank); 227 if (mem->validation_bits & CPER_MEM_VALID_BANK) 228 n += scnprintf(msg + n, len - n, "bank: %d ", mem->bank); 229 if (mem->validation_bits & CPER_MEM_VALID_DEVICE) 230 n += scnprintf(msg + n, len - n, "device: %d ", mem->device); 231 if (mem->validation_bits & CPER_MEM_VALID_ROW) 232 n += scnprintf(msg + n, len - n, "row: %d ", mem->row); 233 if (mem->validation_bits & CPER_MEM_VALID_COLUMN) 234 n += scnprintf(msg + n, len - n, "column: %d ", mem->column); 235 if (mem->validation_bits & CPER_MEM_VALID_BIT_POSITION) 236 n += scnprintf(msg + n, len - n, "bit_position: %d ", 237 mem->bit_pos); 238 if (mem->validation_bits & CPER_MEM_VALID_REQUESTOR_ID) 239 n += scnprintf(msg + n, len - n, "requestor_id: 0x%016llx ", 240 mem->requestor_id); 241 if (mem->validation_bits & CPER_MEM_VALID_RESPONDER_ID) 242 n += scnprintf(msg + n, len - n, "responder_id: 0x%016llx ", 243 mem->responder_id); 244 if (mem->validation_bits & CPER_MEM_VALID_TARGET_ID) 245 scnprintf(msg + n, len - n, "target_id: 0x%016llx ", 246 mem->target_id); 247 248 msg[n] = '\0'; 249 return n; 250 } 251 252 static int cper_dimm_err_location(struct cper_mem_err_compact *mem, char *msg) 253 { 254 u32 len, n; 255 const char *bank = NULL, *device = NULL; 256 257 if (!msg || !(mem->validation_bits & CPER_MEM_VALID_MODULE_HANDLE)) 258 return 0; 259 260 n = 0; 261 len = CPER_REC_LEN - 1; 262 dmi_memdev_name(mem->mem_dev_handle, &bank, &device); 263 if (bank && device) 264 n = snprintf(msg, len, "DIMM location: %s %s ", bank, device); 265 else 266 n = snprintf(msg, len, 267 "DIMM location: not present. DMI handle: 0x%.4x ", 268 mem->mem_dev_handle); 269 270 msg[n] = '\0'; 271 return n; 272 } 273 274 void cper_mem_err_pack(const struct cper_sec_mem_err *mem, 275 struct cper_mem_err_compact *cmem) 276 { 277 cmem->validation_bits = mem->validation_bits; 278 cmem->node = mem->node; 279 cmem->card = mem->card; 280 cmem->module = mem->module; 281 cmem->bank = mem->bank; 282 cmem->device = mem->device; 283 cmem->row = mem->row; 284 cmem->column = mem->column; 285 cmem->bit_pos = mem->bit_pos; 286 cmem->requestor_id = mem->requestor_id; 287 cmem->responder_id = mem->responder_id; 288 cmem->target_id = mem->target_id; 289 cmem->rank = mem->rank; 290 cmem->mem_array_handle = mem->mem_array_handle; 291 cmem->mem_dev_handle = mem->mem_dev_handle; 292 } 293 294 const char *cper_mem_err_unpack(struct trace_seq *p, 295 struct cper_mem_err_compact *cmem) 296 { 297 const char *ret = trace_seq_buffer_ptr(p); 298 299 if (cper_mem_err_location(cmem, rcd_decode_str)) 300 trace_seq_printf(p, "%s", rcd_decode_str); 301 if (cper_dimm_err_location(cmem, rcd_decode_str)) 302 trace_seq_printf(p, "%s", rcd_decode_str); 303 trace_seq_putc(p, '\0'); 304 305 return ret; 306 } 307 308 static void cper_print_mem(const char *pfx, const struct cper_sec_mem_err *mem, 309 int len) 310 { 311 struct cper_mem_err_compact cmem; 312 313 /* Don't trust UEFI 2.1/2.2 structure with bad validation bits */ 314 if (len == sizeof(struct cper_sec_mem_err_old) && 315 (mem->validation_bits & ~(CPER_MEM_VALID_RANK_NUMBER - 1))) { 316 pr_err(FW_WARN "valid bits set for fields beyond structure\n"); 317 return; 318 } 319 if (mem->validation_bits & CPER_MEM_VALID_ERROR_STATUS) 320 printk("%s""error_status: 0x%016llx\n", pfx, mem->error_status); 321 if (mem->validation_bits & CPER_MEM_VALID_PA) 322 printk("%s""physical_address: 0x%016llx\n", 323 pfx, mem->physical_addr); 324 if (mem->validation_bits & CPER_MEM_VALID_PA_MASK) 325 printk("%s""physical_address_mask: 0x%016llx\n", 326 pfx, mem->physical_addr_mask); 327 cper_mem_err_pack(mem, &cmem); 328 if (cper_mem_err_location(&cmem, rcd_decode_str)) 329 printk("%s%s\n", pfx, rcd_decode_str); 330 if (mem->validation_bits & CPER_MEM_VALID_ERROR_TYPE) { 331 u8 etype = mem->error_type; 332 printk("%s""error_type: %d, %s\n", pfx, etype, 333 cper_mem_err_type_str(etype)); 334 } 335 if (cper_dimm_err_location(&cmem, rcd_decode_str)) 336 printk("%s%s\n", pfx, rcd_decode_str); 337 } 338 339 static const char * const pcie_port_type_strs[] = { 340 "PCIe end point", 341 "legacy PCI end point", 342 "unknown", 343 "unknown", 344 "root port", 345 "upstream switch port", 346 "downstream switch port", 347 "PCIe to PCI/PCI-X bridge", 348 "PCI/PCI-X to PCIe bridge", 349 "root complex integrated endpoint device", 350 "root complex event collector", 351 }; 352 353 static void cper_print_pcie(const char *pfx, const struct cper_sec_pcie *pcie, 354 const struct acpi_hest_generic_data *gdata) 355 { 356 if (pcie->validation_bits & CPER_PCIE_VALID_PORT_TYPE) 357 printk("%s""port_type: %d, %s\n", pfx, pcie->port_type, 358 pcie->port_type < ARRAY_SIZE(pcie_port_type_strs) ? 359 pcie_port_type_strs[pcie->port_type] : "unknown"); 360 if (pcie->validation_bits & CPER_PCIE_VALID_VERSION) 361 printk("%s""version: %d.%d\n", pfx, 362 pcie->version.major, pcie->version.minor); 363 if (pcie->validation_bits & CPER_PCIE_VALID_COMMAND_STATUS) 364 printk("%s""command: 0x%04x, status: 0x%04x\n", pfx, 365 pcie->command, pcie->status); 366 if (pcie->validation_bits & CPER_PCIE_VALID_DEVICE_ID) { 367 const __u8 *p; 368 printk("%s""device_id: %04x:%02x:%02x.%x\n", pfx, 369 pcie->device_id.segment, pcie->device_id.bus, 370 pcie->device_id.device, pcie->device_id.function); 371 printk("%s""slot: %d\n", pfx, 372 pcie->device_id.slot >> CPER_PCIE_SLOT_SHIFT); 373 printk("%s""secondary_bus: 0x%02x\n", pfx, 374 pcie->device_id.secondary_bus); 375 printk("%s""vendor_id: 0x%04x, device_id: 0x%04x\n", pfx, 376 pcie->device_id.vendor_id, pcie->device_id.device_id); 377 p = pcie->device_id.class_code; 378 printk("%s""class_code: %02x%02x%02x\n", pfx, p[0], p[1], p[2]); 379 } 380 if (pcie->validation_bits & CPER_PCIE_VALID_SERIAL_NUMBER) 381 printk("%s""serial number: 0x%04x, 0x%04x\n", pfx, 382 pcie->serial_number.lower, pcie->serial_number.upper); 383 if (pcie->validation_bits & CPER_PCIE_VALID_BRIDGE_CONTROL_STATUS) 384 printk( 385 "%s""bridge: secondary_status: 0x%04x, control: 0x%04x\n", 386 pfx, pcie->bridge.secondary_status, pcie->bridge.control); 387 } 388 389 static void cper_estatus_print_section( 390 const char *pfx, const struct acpi_hest_generic_data *gdata, int sec_no) 391 { 392 uuid_le *sec_type = (uuid_le *)gdata->section_type; 393 __u16 severity; 394 char newpfx[64]; 395 396 severity = gdata->error_severity; 397 printk("%s""Error %d, type: %s\n", pfx, sec_no, 398 cper_severity_str(severity)); 399 if (gdata->validation_bits & CPER_SEC_VALID_FRU_ID) 400 printk("%s""fru_id: %pUl\n", pfx, (uuid_le *)gdata->fru_id); 401 if (gdata->validation_bits & CPER_SEC_VALID_FRU_TEXT) 402 printk("%s""fru_text: %.20s\n", pfx, gdata->fru_text); 403 404 snprintf(newpfx, sizeof(newpfx), "%s%s", pfx, INDENT_SP); 405 if (!uuid_le_cmp(*sec_type, CPER_SEC_PROC_GENERIC)) { 406 struct cper_sec_proc_generic *proc_err = (void *)(gdata + 1); 407 printk("%s""section_type: general processor error\n", newpfx); 408 if (gdata->error_data_length >= sizeof(*proc_err)) 409 cper_print_proc_generic(newpfx, proc_err); 410 else 411 goto err_section_too_small; 412 } else if (!uuid_le_cmp(*sec_type, CPER_SEC_PLATFORM_MEM)) { 413 struct cper_sec_mem_err *mem_err = (void *)(gdata + 1); 414 printk("%s""section_type: memory error\n", newpfx); 415 if (gdata->error_data_length >= 416 sizeof(struct cper_sec_mem_err_old)) 417 cper_print_mem(newpfx, mem_err, 418 gdata->error_data_length); 419 else 420 goto err_section_too_small; 421 } else if (!uuid_le_cmp(*sec_type, CPER_SEC_PCIE)) { 422 struct cper_sec_pcie *pcie = (void *)(gdata + 1); 423 printk("%s""section_type: PCIe error\n", newpfx); 424 if (gdata->error_data_length >= sizeof(*pcie)) 425 cper_print_pcie(newpfx, pcie, gdata); 426 else 427 goto err_section_too_small; 428 } else 429 printk("%s""section type: unknown, %pUl\n", newpfx, sec_type); 430 431 return; 432 433 err_section_too_small: 434 pr_err(FW_WARN "error section length is too small\n"); 435 } 436 437 void cper_estatus_print(const char *pfx, 438 const struct acpi_hest_generic_status *estatus) 439 { 440 struct acpi_hest_generic_data *gdata; 441 unsigned int data_len, gedata_len; 442 int sec_no = 0; 443 char newpfx[64]; 444 __u16 severity; 445 446 severity = estatus->error_severity; 447 if (severity == CPER_SEV_CORRECTED) 448 printk("%s%s\n", pfx, 449 "It has been corrected by h/w " 450 "and requires no further action"); 451 printk("%s""event severity: %s\n", pfx, cper_severity_str(severity)); 452 data_len = estatus->data_length; 453 gdata = (struct acpi_hest_generic_data *)(estatus + 1); 454 snprintf(newpfx, sizeof(newpfx), "%s%s", pfx, INDENT_SP); 455 while (data_len >= sizeof(*gdata)) { 456 gedata_len = gdata->error_data_length; 457 cper_estatus_print_section(newpfx, gdata, sec_no); 458 data_len -= gedata_len + sizeof(*gdata); 459 gdata = (void *)(gdata + 1) + gedata_len; 460 sec_no++; 461 } 462 } 463 EXPORT_SYMBOL_GPL(cper_estatus_print); 464 465 int cper_estatus_check_header(const struct acpi_hest_generic_status *estatus) 466 { 467 if (estatus->data_length && 468 estatus->data_length < sizeof(struct acpi_hest_generic_data)) 469 return -EINVAL; 470 if (estatus->raw_data_length && 471 estatus->raw_data_offset < sizeof(*estatus) + estatus->data_length) 472 return -EINVAL; 473 474 return 0; 475 } 476 EXPORT_SYMBOL_GPL(cper_estatus_check_header); 477 478 int cper_estatus_check(const struct acpi_hest_generic_status *estatus) 479 { 480 struct acpi_hest_generic_data *gdata; 481 unsigned int data_len, gedata_len; 482 int rc; 483 484 rc = cper_estatus_check_header(estatus); 485 if (rc) 486 return rc; 487 data_len = estatus->data_length; 488 gdata = (struct acpi_hest_generic_data *)(estatus + 1); 489 while (data_len >= sizeof(*gdata)) { 490 gedata_len = gdata->error_data_length; 491 if (gedata_len > data_len - sizeof(*gdata)) 492 return -EINVAL; 493 data_len -= gedata_len + sizeof(*gdata); 494 gdata = (void *)(gdata + 1) + gedata_len; 495 } 496 if (data_len) 497 return -EINVAL; 498 499 return 0; 500 } 501 EXPORT_SYMBOL_GPL(cper_estatus_check); 502