1 /*-
2 * Copyright (c) 2012 Hudson River Trading LLC
3 * Written by: John H. Baldwin <jhb@FreeBSD.org>
4 * All rights reserved.
5 *
6 * Redistribution and use in source and binary forms, with or without
7 * modification, are permitted provided that the following conditions
8 * are met:
9 * 1. Redistributions of source code must retain the above copyright
10 * notice, this list of conditions and the following disclaimer.
11 * 2. Redistributions in binary form must reproduce the above copyright
12 * notice, this list of conditions and the following disclaimer in the
13 * documentation and/or other materials provided with the distribution.
14 *
15 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
16 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
17 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
18 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
19 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
20 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
21 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
22 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
23 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
24 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
25 * SUCH DAMAGE.
26 */
27
28 #ifndef lint
29 static const char rcsid[] =
30 "$FreeBSD$";
31 #endif /* not lint */
32
33 #include <sys/param.h>
34 #include <sys/pciio.h>
35
36 #include <err.h>
37 #include <stdio.h>
38
39 #include <dev/pci/pcireg.h>
40
41 #include "pciconf.h"
42
43 struct bit_table {
44 uint32_t mask;
45 const char *desc;
46 };
47
48 /* Error indicators in the PCI status register (PCIR_STATUS). */
49 static struct bit_table pci_status[] = {
50 { PCIM_STATUS_MDPERR, "Master Data Parity Error" },
51 { PCIM_STATUS_STABORT, "Sent Target-Abort" },
52 { PCIM_STATUS_RTABORT, "Received Target-Abort" },
53 { PCIM_STATUS_RMABORT, "Received Master-Abort" },
54 { PCIM_STATUS_SERR, "Signalled System Error" },
55 { PCIM_STATUS_PERR, "Detected Parity Error" },
56 { 0, NULL },
57 };
58
59 /* Valid error indicator bits in PCIR_STATUS. */
60 #define PCI_ERRORS (PCIM_STATUS_MDPERR | PCIM_STATUS_STABORT | \
61 PCIM_STATUS_RTABORT | PCIM_STATUS_RMABORT | \
62 PCIM_STATUS_SERR | PCIM_STATUS_PERR)
63
64 /* Error indicators in the PCI-Express device status register. */
65 static struct bit_table pcie_device_status[] = {
66 { PCIEM_STA_CORRECTABLE_ERROR, "Correctable Error Detected" },
67 { PCIEM_STA_NON_FATAL_ERROR, "Non-Fatal Error Detected" },
68 { PCIEM_STA_FATAL_ERROR, "Fatal Error Detected" },
69 { PCIEM_STA_UNSUPPORTED_REQ, "Unsupported Request Detected" },
70 { 0, NULL },
71 };
72
73 /* Valid error indicator bits in the PCI-Express device status register. */
74 #define PCIE_ERRORS (PCIEM_STA_CORRECTABLE_ERROR | \
75 PCIEM_STA_NON_FATAL_ERROR | \
76 PCIEM_STA_FATAL_ERROR | \
77 PCIEM_STA_UNSUPPORTED_REQ)
78
79 /* AER Uncorrected errors. */
80 static struct bit_table aer_uc[] = {
81 { PCIM_AER_UC_TRAINING_ERROR, "Link Training Error" },
82 { PCIM_AER_UC_DL_PROTOCOL_ERROR, "Data Link Protocol Error" },
83 { PCIM_AER_UC_SURPRISE_LINK_DOWN, "Surprise Link Down Error" },
84 { PCIM_AER_UC_POISONED_TLP, "Poisoned TLP" },
85 { PCIM_AER_UC_FC_PROTOCOL_ERROR, "Flow Control Protocol Error" },
86 { PCIM_AER_UC_COMPLETION_TIMEOUT, "Completion Timeout" },
87 { PCIM_AER_UC_COMPLETER_ABORT, "Completer Abort" },
88 { PCIM_AER_UC_UNEXPECTED_COMPLETION, "Unexpected Completion" },
89 { PCIM_AER_UC_RECEIVER_OVERFLOW, "Receiver Overflow Error" },
90 { PCIM_AER_UC_MALFORMED_TLP, "Malformed TLP" },
91 { PCIM_AER_UC_ECRC_ERROR, "ECRC Error" },
92 { PCIM_AER_UC_UNSUPPORTED_REQUEST, "Unsupported Request" },
93 { PCIM_AER_UC_ACS_VIOLATION, "ACS Violation" },
94 { PCIM_AER_UC_INTERNAL_ERROR, "Uncorrectable Internal Error" },
95 { PCIM_AER_UC_MC_BLOCKED_TLP, "MC Blocked TLP" },
96 { PCIM_AER_UC_ATOMIC_EGRESS_BLK, "AtomicOp Egress Blocked" },
97 { PCIM_AER_UC_TLP_PREFIX_BLOCKED, "TLP Prefix Blocked Error" },
98 { 0, NULL },
99 };
100
101 /* AER Corrected errors. */
102 static struct bit_table aer_cor[] = {
103 { PCIM_AER_COR_RECEIVER_ERROR, "Receiver Error" },
104 { PCIM_AER_COR_BAD_TLP, "Bad TLP" },
105 { PCIM_AER_COR_BAD_DLLP, "Bad DLLP" },
106 { PCIM_AER_COR_REPLAY_ROLLOVER, "REPLAY_NUM Rollover" },
107 { PCIM_AER_COR_REPLAY_TIMEOUT, "Replay Timer Timeout" },
108 { PCIM_AER_COR_ADVISORY_NF_ERROR, "Advisory Non-Fatal Error" },
109 { PCIM_AER_COR_INTERNAL_ERROR, "Corrected Internal Error" },
110 { PCIM_AER_COR_HEADER_LOG_OVFLOW, "Header Log Overflow" },
111 { 0, NULL },
112 };
113
114 static void
print_bits(const char * header,struct bit_table * table,uint32_t mask)115 print_bits(const char *header, struct bit_table *table, uint32_t mask)
116 {
117 int first;
118
119 first = 1;
120 for (; table->desc != NULL; table++)
121 if (mask & table->mask) {
122 if (first) {
123 printf("%14s = ", header);
124 first = 0;
125 } else
126 printf(" ");
127 printf("%s\n", table->desc);
128 mask &= ~table->mask;
129 }
130 if (mask != 0) {
131 if (first)
132 printf("%14s = ", header);
133 else
134 printf(" ");
135 printf("Unknown: 0x%08x\n", mask);
136 }
137 }
138
139 void
list_errors(int fd,struct pci_conf * p)140 list_errors(int fd, struct pci_conf *p)
141 {
142 uint32_t mask, severity;
143 uint16_t sta, aer;
144 uint8_t pcie;
145
146 /* First check for standard PCI errors. */
147 sta = read_config(fd, &p->pc_sel, PCIR_STATUS, 2);
148 print_bits("PCI errors", pci_status, sta & PCI_ERRORS);
149
150 /* See if this is a PCI-express device. */
151 pcie = pci_find_cap(fd, p, PCIY_EXPRESS);
152 if (pcie == 0)
153 return;
154
155 /* Check for PCI-e errors. */
156 sta = read_config(fd, &p->pc_sel, pcie + PCIER_DEVICE_STA, 2);
157 print_bits("PCI-e errors", pcie_device_status, sta & PCIE_ERRORS);
158
159 /* See if this device supports AER. */
160 aer = pcie_find_cap(fd, p, PCIZ_AER);
161 if (aer == 0)
162 return;
163
164 /* Check for uncorrected errors. */
165 mask = read_config(fd, &p->pc_sel, aer + PCIR_AER_UC_STATUS, 4);
166 severity = read_config(fd, &p->pc_sel, aer + PCIR_AER_UC_SEVERITY, 4);
167 print_bits("Fatal", aer_uc, mask & severity);
168 print_bits("Non-fatal", aer_uc, mask & ~severity);
169
170 /* Check for corrected errors. */
171 mask = read_config(fd, &p->pc_sel, aer + PCIR_AER_COR_STATUS, 4);
172 print_bits("Corrected", aer_cor, mask);
173 }
174