drivers/pci/pcie/aer/aerdrv_errprint.c

   1 /*
   2  * drivers/pci/pcie/aer/aerdrv_errprint.c
   3  *
   4  * This file is subject to the terms and conditions of the GNU General Public
   5  * License.  See the file "COPYING" in the main directory of this archive
   6  * for more details.
   7  *
   8  * Format error messages and print them to console.
   9  *
  10  * Copyright (C) 2006 Intel Corp.
  11  *      Tom Long Nguyen (tom.l.nguyen@intel.com)
  12  *      Zhang Yanmin (yanmin.zhang@intel.com)
  13  *
  14  */
  15
  16 #include <linux/module.h>
  17 #include <linux/pci.h>
  18 #include <linux/kernel.h>
  19 #include <linux/errno.h>
  20 #include <linux/pm.h>
  21 #include <linux/suspend.h>
  22 #include <linux/cper.h>
  23
  24 #include "aerdrv.h"
  25
  26 #define CREATE_TRACE_POINTS
  27 #include <trace/events/ras.h>
  28
  29 #define AER_AGENT_RECEIVER              0
  30 #define AER_AGENT_REQUESTER             1
  31 #define AER_AGENT_COMPLETER             2
  32 #define AER_AGENT_TRANSMITTER           3
  33
  34 #define AER_AGENT_REQUESTER_MASK(t)     ((t == AER_CORRECTABLE) ?       \
  35         0 : (PCI_ERR_UNC_COMP_TIME|PCI_ERR_UNC_UNSUP))
  36 #define AER_AGENT_COMPLETER_MASK(t)     ((t == AER_CORRECTABLE) ?       \
  37         0 : PCI_ERR_UNC_COMP_ABORT)
  38 #define AER_AGENT_TRANSMITTER_MASK(t)   ((t == AER_CORRECTABLE) ?       \
  39         (PCI_ERR_COR_REP_ROLL|PCI_ERR_COR_REP_TIMER) : 0)
  40
  41 #define AER_GET_AGENT(t, e)                                             \
  42         ((e & AER_AGENT_COMPLETER_MASK(t)) ? AER_AGENT_COMPLETER :      \
  43         (e & AER_AGENT_REQUESTER_MASK(t)) ? AER_AGENT_REQUESTER :       \
  44         (e & AER_AGENT_TRANSMITTER_MASK(t)) ? AER_AGENT_TRANSMITTER :   \
  45         AER_AGENT_RECEIVER)
  46
  47 #define AER_PHYSICAL_LAYER_ERROR        0
  48 #define AER_DATA_LINK_LAYER_ERROR       1
  49 #define AER_TRANSACTION_LAYER_ERROR     2
  50
  51 #define AER_PHYSICAL_LAYER_ERROR_MASK(t) ((t == AER_CORRECTABLE) ?      \
  52         PCI_ERR_COR_RCVR : 0)
  53 #define AER_DATA_LINK_LAYER_ERROR_MASK(t) ((t == AER_CORRECTABLE) ?     \
  54         (PCI_ERR_COR_BAD_TLP|                                           \
  55         PCI_ERR_COR_BAD_DLLP|                                           \
  56         PCI_ERR_COR_REP_ROLL|                                           \
  57         PCI_ERR_COR_REP_TIMER) : PCI_ERR_UNC_DLP)
  58
  59 #define AER_GET_LAYER_ERROR(t, e)                                       \
  60         ((e & AER_PHYSICAL_LAYER_ERROR_MASK(t)) ? AER_PHYSICAL_LAYER_ERROR : \
  61         (e & AER_DATA_LINK_LAYER_ERROR_MASK(t)) ? AER_DATA_LINK_LAYER_ERROR : \
  62         AER_TRANSACTION_LAYER_ERROR)
  63
  64 /*
  65  * AER error strings
  66  */
  67 static const char *aer_error_severity_string[] = {
  68         "Uncorrected (Non-Fatal)",
  69         "Uncorrected (Fatal)",
  70         "Corrected"
  71 };
  72
  73 static const char *aer_error_layer[] = {
  74         "Physical Layer",
  75         "Data Link Layer",
  76         "Transaction Layer"
  77 };
  78
  79 static const char *aer_correctable_error_string[] = {
  80         "Receiver Error",               /* Bit Position 0       */
  81         NULL,
  82         NULL,
  83         NULL,
  84         NULL,
  85         NULL,
  86         "Bad TLP",                      /* Bit Position 6       */
  87         "Bad DLLP",                     /* Bit Position 7       */
  88         "RELAY_NUM Rollover",           /* Bit Position 8       */
  89         NULL,
  90         NULL,
  91         NULL,
  92         "Replay Timer Timeout",         /* Bit Position 12      */
  93         "Advisory Non-Fatal",           /* Bit Position 13      */
  94 };
  95
  96 static const char *aer_uncorrectable_error_string[] = {
  97         NULL,
  98         NULL,
  99         NULL,
 100         NULL,
 101         "Data Link Protocol",           /* Bit Position 4       */
 102         NULL,
 103         NULL,
 104         NULL,
 105         NULL,
 106         NULL,
 107         NULL,
 108         NULL,
 109         "Poisoned TLP",                 /* Bit Position 12      */
 110         "Flow Control Protocol",        /* Bit Position 13      */
 111         "Completion Timeout",           /* Bit Position 14      */
 112         "Completer Abort",              /* Bit Position 15      */
 113         "Unexpected Completion",        /* Bit Position 16      */
 114         "Receiver Overflow",            /* Bit Position 17      */
 115         "Malformed TLP",                /* Bit Position 18      */
 116         "ECRC",                         /* Bit Position 19      */
 117         "Unsupported Request",          /* Bit Position 20      */
 118 };
 119
 120 static const char *aer_agent_string[] = {
 121         "Receiver ID",
 122         "Requester ID",
 123         "Completer ID",
 124         "Transmitter ID"
 125 };
 126
 127 static void __print_tlp_header(struct pci_dev *dev,
 128                                struct aer_header_log_regs *t)
 129 {
 130         unsigned char *tlp = (unsigned char *)&t;
 131
 132         dev_err(&dev->dev, "  TLP Header:"
 133                 " %02x%02x%02x%02x %02x%02x%02x%02x"
 134                 " %02x%02x%02x%02x %02x%02x%02x%02x\n",
 135                 *(tlp + 3), *(tlp + 2), *(tlp + 1), *tlp,
 136                 *(tlp + 7), *(tlp + 6), *(tlp + 5), *(tlp + 4),
 137                 *(tlp + 11), *(tlp + 10), *(tlp + 9),
 138                 *(tlp + 8), *(tlp + 15), *(tlp + 14),
 139                 *(tlp + 13), *(tlp + 12));
 140 }
 141
 142 static void __aer_print_error(struct pci_dev *dev,
 143                               struct aer_err_info *info)
 144 {
 145         int i, status;
 146         const char *errmsg = NULL;
 147         status = (info->status & ~info->mask);
 148
 149         for (i = 0; i < 32; i++) {
 150                 if (!(status & (1 << i)))
 151                         continue;
 152
 153                 if (info->severity == AER_CORRECTABLE)
 154                         errmsg = i < ARRAY_SIZE(aer_correctable_error_string) ?
 155                                 aer_correctable_error_string[i] : NULL;
 156                 else
 157                         errmsg = i < ARRAY_SIZE(aer_uncorrectable_error_string) ?
 158                                 aer_uncorrectable_error_string[i] : NULL;
 159
 160                 if (errmsg)
 161                         dev_err(&dev->dev, "   [%2d] %-22s%s\n", i, errmsg,
 162                                 info->first_error == i ? " (First)" : "");
 163                 else
 164                         dev_err(&dev->dev, "   [%2d] Unknown Error Bit%s\n",
 165                                 i, info->first_error == i ? " (First)" : "");
 166         }
 167 }
 168
 169 void aer_print_error(struct pci_dev *dev, struct aer_err_info *info)
 170 {
 171         int layer, agent;
 172         int id = ((dev->bus->number << 8) | dev->devfn);
 173
 174         if (!info->status) {
 175                 dev_err(&dev->dev,
 176                         "PCIe Bus Error: severity=%s, type=Unaccessible, "
 177                         "id=%04x(Unregistered Agent ID)\n",
 178                         aer_error_severity_string[info->severity], id);
 179                 goto out;
 180         }
 181
 182         layer = AER_GET_LAYER_ERROR(info->severity, info->status);
 183         agent = AER_GET_AGENT(info->severity, info->status);
 184
 185         dev_err(&dev->dev,
 186                 "PCIe Bus Error: severity=%s, type=%s, id=%04x(%s)\n",
 187                 aer_error_severity_string[info->severity],
 188                 aer_error_layer[layer], id, aer_agent_string[agent]);
 189
 190         dev_err(&dev->dev,
 191                 "  device [%04x:%04x] error status/mask=%08x/%08x\n",
 192                 dev->vendor, dev->device,
 193                 info->status, info->mask);
 194
 195         __aer_print_error(dev, info);
 196
 197         if (info->tlp_header_valid)
 198                 __print_tlp_header(dev, &info->tlp);
 199
 200 out:
 201         if (info->id && info->error_dev_num > 1 && info->id == id)
 202                 dev_err(&dev->dev, "  Error of this Agent(%04x) is reported first\n", id);
 203
 204         trace_aer_event(dev_name(&dev->dev), (info->status & ~info->mask),
 205                         info->severity);
 206 }
 207
 208 void aer_print_port_info(struct pci_dev *dev, struct aer_err_info *info)
 209 {
 210         dev_info(&dev->dev, "AER: %s%s error received: id=%04x\n",
 211                 info->multi_error_valid ? "Multiple " : "",
 212                 aer_error_severity_string[info->severity], info->id);
 213 }
 214
 215 #ifdef CONFIG_ACPI_APEI_PCIEAER
 216 int cper_severity_to_aer(int cper_severity)
 217 {
 218         switch (cper_severity) {
 219         case CPER_SEV_RECOVERABLE:
 220                 return AER_NONFATAL;
 221         case CPER_SEV_FATAL:
 222                 return AER_FATAL;
 223         default:
 224                 return AER_CORRECTABLE;
 225         }
 226 }
 227 EXPORT_SYMBOL_GPL(cper_severity_to_aer);
 228
 229 void cper_print_aer(struct pci_dev *dev, int cper_severity,
 230                     struct aer_capability_regs *aer)
 231 {
 232         int aer_severity, layer, agent, status_strs_size, tlp_header_valid = 0;
 233         u32 status, mask;
 234         const char **status_strs;
 235
 236         aer_severity = cper_severity_to_aer(cper_severity);
 237
 238         if (aer_severity == AER_CORRECTABLE) {
 239                 status = aer->cor_status;
 240                 mask = aer->cor_mask;
 241                 status_strs = aer_correctable_error_string;
 242                 status_strs_size = ARRAY_SIZE(aer_correctable_error_string);
 243         } else {
 244                 status = aer->uncor_status;
 245                 mask = aer->uncor_mask;
 246                 status_strs = aer_uncorrectable_error_string;
 247                 status_strs_size = ARRAY_SIZE(aer_uncorrectable_error_string);
 248                 tlp_header_valid = status & AER_LOG_TLP_MASKS;
 249         }
 250
 251         layer = AER_GET_LAYER_ERROR(aer_severity, status);
 252         agent = AER_GET_AGENT(aer_severity, status);
 253
 254         dev_err(&dev->dev, "aer_status: 0x%08x, aer_mask: 0x%08x\n", status, mask);
 255         cper_print_bits("", status, status_strs, status_strs_size);
 256         dev_err(&dev->dev, "aer_layer=%s, aer_agent=%s\n",
 257                 aer_error_layer[layer], aer_agent_string[agent]);
 258
 259         if (aer_severity != AER_CORRECTABLE)
 260                 dev_err(&dev->dev, "aer_uncor_severity: 0x%08x\n",
 261                         aer->uncor_severity);
 262
 263         if (tlp_header_valid)
 264                 __print_tlp_header(dev, &aer->header_log);
 265
 266         trace_aer_event(dev_name(&dev->dev), (status & ~mask),
 267                         aer_severity);
 268 }
 269 #endif