summaryrefslogtreecommitdiff
path: root/usr/src/uts/common/io/pciex/pcieb.c
blob: 513b6dae3b3f2e86c49fa4ac9d34c8e18546b099 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
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
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
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
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
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
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
438
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
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
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
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
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
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
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
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
/*
 * CDDL HEADER START
 *
 * The contents of this file are subject to the terms of the
 * Common Development and Distribution License (the "License").
 * You may not use this file except in compliance with the License.
 *
 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
 * or http://www.opensolaris.org/os/licensing.
 * See the License for the specific language governing permissions
 * and limitations under the License.
 *
 * When distributing Covered Code, include this CDDL HEADER in each
 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
 * If applicable, add the following below this CDDL HEADER, with the
 * fields enclosed by brackets "[]" replaced with your own identifying
 * information: Portions Copyright [yyyy] [name of copyright owner]
 *
 * CDDL HEADER END
 */
/*
 * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
 */
/*
 * Copyright 2012 Garrett D'Amore <garrett@damore.org>.  All rights reserved.
 * Copyright 2019 Joyent, Inc.
 * Copyright 2022 Oxide Computer Company
 */

/*
 * Common x86 and SPARC PCI-E to PCI bus bridge nexus driver
 *
 * Background
 * ----------
 *
 * The PCI Express (PCIe) specification defines that all of the PCIe devices in
 * the system are connected together in a series of different fabrics. A way to
 * think of these fabrics is that they are small networks where there are links
 * between different devices and switches that allow fan out or fan in of the
 * fabric. The entry point to that fabric is called a root complex and the
 * fabric terminates at a what is called an endpoint, which is really just PCIe
 * terminology for the common cards that are inserted into the system (HBAs,
 * NICs, USB, NVMe, etc.).
 *
 * The PCIe specification states that every link on the system has a virtual
 * PCI-to-PCI bridge. This allows PCIe devices to still be configured the same
 * way traditional PCI devices are to the operating system and allows them to
 * have a traditional PCI bus, device, and function associated with them, even
 * though there is no actual shared bus. In addition, bridges are also used to
 * connect traditional PCI and PCI-X devices into them.
 *
 * The PCIe specification refers to upstream and downstream ports. Upstream
 * ports are considered closer the root complex and downstream ports are closer
 * to the endpoint. We can divide the devices that the bridge driver attaches to
 * into two groups. Those that are considered upstream ports, these include root
 * complexes and parts of PCIe switches. And downstream ports, which are the
 * other half of PCIe switches and endpoints (which this driver does not attach
 * to, normal hardware-specific or class-specific drivers attach to those).
 *
 * Interrupt Management
 * --------------------
 *
 * Upstream ports of bridges have additional things that we care about.
 * Specifically they're the means through which we find out about:
 *
 *  - Advanced Error Reporting (AERs)
 *  - Hotplug events
 *  - Link Bandwidth Events
 *  - Power Management Events (PME)
 *
 * Each of these features is an optional feature (though ones we hope are
 * implemented). The features above are grouped into two different buckets based
 * on which PCI capability they appear in. AER management is done through a PCI
 * Express extended configuration header (it lives in extended PCI configuration
 * space) called the 'Advanced Error Reporting Extended Capability'. The other
 * events are all managed as part of the 'PCI Express Capability Structure'.
 * This structure is found in traditional PCI configuration space.
 *
 * The way that the interrupts are programmed for these types of events differs
 * a bit from the way one might expect a normal device to operate. For most
 * devices, one allocates a number of interrupts based on a combination of what
 * the device supports, what the OS supports per device, and the number the
 * driver needs. Then the driver programs the device in a device-specific manner
 * to indicate which events should trigger a specific interrupt vector.
 *
 * However, for both the AER and PCI capabilities, the driver has to do
 * something different. The driver first allocates interrupts by programming the
 * MSI or MSI-X table and then asks the device which interrupts have been
 * assigned to these purposes. Because these events are only supported in
 * 'upstream' devices, this does not interfere with the traditional management
 * of MSI and MSI-X interrupts. At this time, the pcieb driver only supports the
 * use of MSI interrupts.
 *
 * Once the interrupts have been allocated, we read back which vectors have been
 * nominated by the device to cover the corresponding capability. The interrupt
 * is allocated on a per-capability basis. Therefore, one interrupt would cover
 * AERs, while another interrupt would cover the rest of the desired functions.
 *
 * To track which interrupts cover which behaviors, each driver state
 * (pcieb_devstate_t) has a member called 'pcieb_isr_tab'. Each index represents
 * an interrupt vector and there are a series of flags that represent the
 * different possible interrupt sources: PCIEB_INTR_SRC_HP (hotplug),
 * PCEIB_INTR_SRC_PME (power management event), PCIEB_INTR_SRC_AER (error
 * reporting), PCIEB_INTR_SRC_LBW (link bandwidth).
 *
 * Because the hotplug, link bandwidth, and power management events all share
 * the same vector, if an interrupt comes in, we must check all of the enabled
 * sources that might generate this interrupt. It is highly likely that more
 * than one will fire at the same time, for example, a hotplug event that fires
 * because a device has been inserted or removed, will likely trigger a link
 * bandwidth event.
 *
 * The pcieb driver itself does not actually have much logic to deal with and
 * clear the interrupts in question. It generally speaking will vector most
 * events back to the more general pcie driver or, in the case of AERs, initiate
 * a scan of the fabric itself (also part of the pcie driver).
 *
 * Link Management
 * ---------------
 *
 * The pcieb driver is used to take care of two different aspects of link
 * management. The first of these, as described briefly above, is to monitor for
 * changes to the negotiated link bandwidth. These events are managed by
 * enabling support for the interrupts in the PCI Express Capability Structure.
 * This is all taken care of by the pcie driver through functions like
 * pcie_link_bw_enable().
 *
 * The second aspect of link management the pcieb driver enables is the ability
 * to retrain the link and optionally limit the speed. This is enabled through a
 * series of private ioctls that are driven through a private userland utility,
 * /usr/lib/pci/pcieb. Eventually, this should be more fleshed out and a more
 * uniform interface based around the devctls that can be leveraged across
 * different classes of devices should be used.
 *
 * Under the hood this basically leverages the ability of the upstream port to
 * retrain a link by writing a bit to the PCIe link control register. See
 * pcieb_ioctl_retrain(). From there, if the driver ever receives a request to
 * change the maximum speed, that is updated in the card; however, it does not
 * immediately retrain the link. A separate ioctl request is required to do so.
 * Once the speed has been changed, regardless of whether or not it has been
 * retrained, that fact will always be noted.
 */

#include <sys/sysmacros.h>
#include <sys/conf.h>
#include <sys/kmem.h>
#include <sys/debug.h>
#include <sys/modctl.h>
#include <sys/autoconf.h>
#include <sys/ddi_impldefs.h>
#include <sys/pci.h>
#include <sys/ddi.h>
#include <sys/sunddi.h>
#include <sys/sunndi.h>
#include <sys/fm/util.h>
#include <sys/pci_cap.h>
#include <sys/pci_impl.h>
#include <sys/pcie_impl.h>
#include <sys/open.h>
#include <sys/stat.h>
#include <sys/file.h>
#include <sys/promif.h>		/* prom_printf */
#include <sys/disp.h>
#include <sys/pcie_pwr.h>
#include <sys/hotplug/pci/pcie_hp.h>
#include "pcieb.h"
#include "pcieb_ioctl.h"
#ifdef PX_PLX
#include <io/pciex/pcieb_plx.h>
#endif /* PX_PLX */

/*LINTLIBRARY*/

/* panic flag */
int pcieb_die = PF_ERR_FATAL_FLAGS;
int pcieb_disable_41210_wkarnd = 0;

/* flag to turn on MSI support */
int pcieb_enable_msi = 1;

#if defined(DEBUG)
uint_t pcieb_dbg_print = 0;

static char *pcieb_debug_sym [] = {	/* same sequence as pcieb_debug_bit */
	/*  0 */ "attach",
	/*  1 */ "pwr",
	/*  2 */ "intr"
};
#endif /* DEBUG */

static int pcieb_bus_map(dev_info_t *, dev_info_t *, ddi_map_req_t *, off_t,
	off_t, caddr_t *);
static int pcieb_ctlops(dev_info_t *, dev_info_t *, ddi_ctl_enum_t, void *,
	void *);
static int pcieb_fm_init(pcieb_devstate_t *pcieb_p);
static void pcieb_fm_fini(pcieb_devstate_t *pcieb_p);
static int pcieb_fm_init_child(dev_info_t *dip, dev_info_t *cdip, int cap,
    ddi_iblock_cookie_t *ibc_p);
static int pcieb_dma_allochdl(dev_info_t *dip, dev_info_t *rdip,
	ddi_dma_attr_t *attr_p, int (*waitfp)(caddr_t), caddr_t arg,
	ddi_dma_handle_t *handlep);
static int pcieb_dma_mctl(dev_info_t *dip, dev_info_t *rdip,
	ddi_dma_handle_t handle, enum ddi_dma_ctlops cmd, off_t *offp,
	size_t *lenp, caddr_t *objp, uint_t cache_flags);
static int pcieb_intr_ops(dev_info_t *dip, dev_info_t *rdip,
	ddi_intr_op_t intr_op, ddi_intr_handle_impl_t *hdlp, void *result);

static struct bus_ops pcieb_bus_ops = {
	BUSO_REV,
	pcieb_bus_map,
	0,
	0,
	0,
	i_ddi_map_fault,
	0,
	pcieb_dma_allochdl,
	ddi_dma_freehdl,
	ddi_dma_bindhdl,
	ddi_dma_unbindhdl,
	ddi_dma_flush,
	ddi_dma_win,
	pcieb_dma_mctl,
	pcieb_ctlops,
	ddi_bus_prop_op,
	ndi_busop_get_eventcookie,	/* (*bus_get_eventcookie)();	*/
	ndi_busop_add_eventcall,	/* (*bus_add_eventcall)();	*/
	ndi_busop_remove_eventcall,	/* (*bus_remove_eventcall)();	*/
	ndi_post_event,			/* (*bus_post_event)();		*/
	NULL,				/* (*bus_intr_ctl)();		*/
	NULL,				/* (*bus_config)();		*/
	NULL,				/* (*bus_unconfig)();		*/
	pcieb_fm_init_child,		/* (*bus_fm_init)();		*/
	NULL,				/* (*bus_fm_fini)();		*/
	i_ndi_busop_access_enter,	/* (*bus_fm_access_enter)();	*/
	i_ndi_busop_access_exit,	/* (*bus_fm_access_exit)();	*/
	pcie_bus_power,			/* (*bus_power)();		*/
	pcieb_intr_ops,			/* (*bus_intr_op)();		*/
	pcie_hp_common_ops		/* (*bus_hp_op)();		*/
};

static int	pcieb_open(dev_t *, int, int, cred_t *);
static int	pcieb_close(dev_t, int, int, cred_t *);
static int	pcieb_ioctl(dev_t, int, intptr_t, int, cred_t *, int *);
static int	pcieb_info(dev_info_t *, ddi_info_cmd_t, void *, void **);
static uint_t	pcieb_intr_handler(caddr_t arg1, caddr_t arg2);

/* PM related functions */
static int	pcieb_pwr_setup(dev_info_t *dip);
static int	pcieb_pwr_init_and_raise(dev_info_t *dip, pcie_pwr_t *pwr_p);
static void	pcieb_pwr_teardown(dev_info_t *dip);
static int	pcieb_pwr_disable(dev_info_t *dip);

/* Hotplug related functions */
static void pcieb_id_props(pcieb_devstate_t *pcieb);

/*
 * soft state pointer
 */
void *pcieb_state;

static struct cb_ops pcieb_cb_ops = {
	pcieb_open,			/* open */
	pcieb_close,			/* close */
	nodev,				/* strategy */
	nodev,				/* print */
	nodev,				/* dump */
	nodev,				/* read */
	nodev,				/* write */
	pcieb_ioctl,			/* ioctl */
	nodev,				/* devmap */
	nodev,				/* mmap */
	nodev,				/* segmap */
	nochpoll,			/* poll */
	pcie_prop_op,			/* cb_prop_op */
	NULL,				/* streamtab */
	D_NEW | D_MP | D_HOTPLUG,	/* Driver compatibility flag */
	CB_REV,				/* rev */
	nodev,				/* int (*cb_aread)() */
	nodev				/* int (*cb_awrite)() */
};

static int	pcieb_probe(dev_info_t *);
static int	pcieb_attach(dev_info_t *devi, ddi_attach_cmd_t cmd);
static int	pcieb_detach(dev_info_t *devi, ddi_detach_cmd_t cmd);

static struct dev_ops pcieb_ops = {
	DEVO_REV,		/* devo_rev */
	0,			/* refcnt  */
	pcieb_info,		/* info */
	nulldev,		/* identify */
	pcieb_probe,		/* probe */
	pcieb_attach,		/* attach */
	pcieb_detach,		/* detach */
	nulldev,		/* reset */
	&pcieb_cb_ops,		/* driver operations */
	&pcieb_bus_ops,		/* bus operations */
	pcie_power,		/* power */
	ddi_quiesce_not_needed,		/* quiesce */
};

/*
 * Module linkage information for the kernel.
 */

static struct modldrv modldrv = {
	&mod_driverops, /* Type of module */
	"PCIe bridge/switch driver",
	&pcieb_ops,	/* driver ops */
};

static struct modlinkage modlinkage = {
	MODREV_1,
	(void *)&modldrv,
	NULL
};

/*
 * forward function declarations:
 */
static void	pcieb_uninitchild(dev_info_t *);
static int	pcieb_initchild(dev_info_t *child);
static void	pcieb_create_ranges_prop(dev_info_t *, ddi_acc_handle_t);
static boolean_t pcieb_is_pcie_device_type(dev_info_t *dip);

/* interrupt related declarations */
static int	pcieb_msi_supported(dev_info_t *);
static int	pcieb_intr_attach(pcieb_devstate_t *pcieb);
static int	pcieb_intr_init(pcieb_devstate_t *pcieb_p, int intr_type);
static void	pcieb_intr_fini(pcieb_devstate_t *pcieb_p);

int
_init(void)
{
	int e;

	if ((e = ddi_soft_state_init(&pcieb_state, sizeof (pcieb_devstate_t),
	    1)) == 0 && (e = mod_install(&modlinkage)) != 0)
		ddi_soft_state_fini(&pcieb_state);
	return (e);
}

int
_fini(void)
{
	int e;

	if ((e = mod_remove(&modlinkage)) == 0) {
		ddi_soft_state_fini(&pcieb_state);
	}
	return (e);
}

int
_info(struct modinfo *modinfop)
{
	return (mod_info(&modlinkage, modinfop));
}

/* ARGSUSED */
static int
pcieb_info(dev_info_t *dip, ddi_info_cmd_t infocmd, void *arg, void **result)
{
	minor_t		minor = getminor((dev_t)arg);
	int		instance = PCI_MINOR_NUM_TO_INSTANCE(minor);
	pcieb_devstate_t *pcieb = ddi_get_soft_state(pcieb_state, instance);
	int		ret = DDI_SUCCESS;

	switch (infocmd) {
	case DDI_INFO_DEVT2INSTANCE:
		*result = (void *)(intptr_t)instance;
		break;
	case DDI_INFO_DEVT2DEVINFO:
		if (pcieb == NULL) {
			ret = DDI_FAILURE;
			break;
		}

		*result = (void *)pcieb->pcieb_dip;
		break;
	default:
		ret = DDI_FAILURE;
		break;
	}

	return (ret);
}


/*ARGSUSED*/
static int
pcieb_probe(dev_info_t *devi)
{
	return (DDI_PROBE_SUCCESS);
}

/*
 * This is a workaround for an undocumented HW erratum with the
 * multi-function, F0 and F2, Intel 41210 PCIe-to-PCI bridge. When
 * Fn (cdip) attaches, this workaround is called to initialize Fn's
 * sibling (sdip) with MPS/MRRS if it isn't already configured.
 * Doing so prevents a malformed TLP panic.
 */
static void
pcieb_41210_mps_wkrnd(dev_info_t *cdip)
{
	dev_info_t *sdip;
	ddi_acc_handle_t cfg_hdl;
	uint16_t cdip_dev_ctrl, cdip_mrrs_mps;
	pcie_bus_t *cdip_bus_p = PCIE_DIP2BUS(cdip);

	/* Get cdip's MPS/MRRS already setup by pcie_initchild_mps() */
	ASSERT(cdip_bus_p);
	cdip_dev_ctrl  = PCIE_CAP_GET(16, cdip_bus_p, PCIE_DEVCTL);
	cdip_mrrs_mps  = cdip_dev_ctrl &
	    (PCIE_DEVCTL_MAX_READ_REQ_MASK | PCIE_DEVCTL_MAX_PAYLOAD_MASK);

	/* Locate sdip and set its MPS/MRRS when applicable */
	for (sdip = ddi_get_child(ddi_get_parent(cdip)); sdip;
	    sdip = ddi_get_next_sibling(sdip)) {
		uint16_t sdip_dev_ctrl, sdip_mrrs_mps, cap_ptr;
		uint32_t bus_dev_ven_id;

		if (sdip == cdip || pci_config_setup(sdip, &cfg_hdl)
		    != DDI_SUCCESS)
			continue;

		/* must be an Intel 41210 bridge */
		bus_dev_ven_id = pci_config_get32(cfg_hdl, PCI_CONF_VENID);
		if (!PCIEB_IS_41210_BRIDGE(bus_dev_ven_id)) {
			pci_config_teardown(&cfg_hdl);
			continue;
		}

		if (PCI_CAP_LOCATE(cfg_hdl, PCI_CAP_ID_PCI_E, &cap_ptr)
		    != DDI_SUCCESS) {
			pci_config_teardown(&cfg_hdl);
			continue;
		}

		/* get sdip's MPS/MRRS to compare to cdip's */
		sdip_dev_ctrl = PCI_CAP_GET16(cfg_hdl, 0, cap_ptr,
		    PCIE_DEVCTL);
		sdip_mrrs_mps = sdip_dev_ctrl &
		    (PCIE_DEVCTL_MAX_READ_REQ_MASK |
		    PCIE_DEVCTL_MAX_PAYLOAD_MASK);

		/* if sdip already attached then its MPS/MRRS is configured */
		if (i_ddi_devi_attached(sdip)) {
			ASSERT(sdip_mrrs_mps == cdip_mrrs_mps);
			pci_config_teardown(&cfg_hdl);
			continue;
		}

		/* otherwise, update sdip's MPS/MRRS if different from cdip's */
		if (sdip_mrrs_mps != cdip_mrrs_mps) {
			sdip_dev_ctrl = (sdip_dev_ctrl &
			    ~(PCIE_DEVCTL_MAX_READ_REQ_MASK |
			    PCIE_DEVCTL_MAX_PAYLOAD_MASK)) | cdip_mrrs_mps;

			(void) PCI_CAP_PUT16(cfg_hdl, 0, cap_ptr, PCIE_DEVCTL,
			    sdip_dev_ctrl);
		}

		/*
		 * note: sdip's bus_mps will be updated by
		 * pcie_initchild_mps()
		 */

		pci_config_teardown(&cfg_hdl);

		break;
	}
}

static int
pcieb_attach(dev_info_t *devi, ddi_attach_cmd_t cmd)
{
	int			instance;
	char			device_type[8];
	pcieb_devstate_t	*pcieb;
	pcie_bus_t		*bus_p = PCIE_DIP2UPBUS(devi);
	ddi_acc_handle_t	config_handle = bus_p->bus_cfg_hdl;

	switch (cmd) {
	case DDI_RESUME:
		(void) pcie_pwr_resume(devi);
		return (DDI_SUCCESS);

	default:
		return (DDI_FAILURE);

	case DDI_ATTACH:
		break;
	}

	if (!(PCIE_IS_BDG(bus_p))) {
		PCIEB_DEBUG(DBG_ATTACH, devi, "This is not a switch or"
		" bridge\n");
		return (DDI_FAILURE);
	}

	/*
	 * If PCIE_LINKCTL_LINK_DISABLE bit in the PCIe Config
	 * Space (PCIe Capability Link Control Register) is set,
	 * then do not bind the driver.
	 */
	if (PCIE_CAP_GET(16, bus_p, PCIE_LINKCTL) & PCIE_LINKCTL_LINK_DISABLE)
		return (DDI_FAILURE);

	/*
	 * Allocate and get soft state structure.
	 */
	instance = ddi_get_instance(devi);
	if (ddi_soft_state_zalloc(pcieb_state, instance) != DDI_SUCCESS)
		return (DDI_FAILURE);
	pcieb = ddi_get_soft_state(pcieb_state, instance);
	pcieb->pcieb_dip = devi;

	if ((pcieb_fm_init(pcieb)) != DDI_SUCCESS) {
		PCIEB_DEBUG(DBG_ATTACH, devi, "Failed in pcieb_fm_init\n");
		goto fail;
	}
	pcieb->pcieb_init_flags |= PCIEB_INIT_FM;

	mutex_init(&pcieb->pcieb_mutex, NULL, MUTEX_DRIVER, NULL);
	mutex_init(&pcieb->pcieb_err_mutex, NULL, MUTEX_DRIVER,
	    (void *)pcieb->pcieb_fm_ibc);
	mutex_init(&pcieb->pcieb_peek_poke_mutex, NULL, MUTEX_DRIVER,
	    (void *)pcieb->pcieb_fm_ibc);

	/* create special properties for device identification */
	pcieb_id_props(pcieb);

	/*
	 * Power management setup. This also makes sure that switch/bridge
	 * is at D0 during attach.
	 */
	if (pwr_common_setup(devi) != DDI_SUCCESS) {
		PCIEB_DEBUG(DBG_PWR, devi, "pwr_common_setup failed\n");
		goto fail;
	}

	if (pcieb_pwr_setup(devi) != DDI_SUCCESS) {
		PCIEB_DEBUG(DBG_PWR, devi, "pxb_pwr_setup failed \n");
		goto fail;
	}

	/*
	 * Make sure the "device_type" property exists.
	 */
	if (pcieb_is_pcie_device_type(devi))
		(void) strcpy(device_type, "pciex");
	else
		(void) strcpy(device_type, "pci");

	(void) ddi_prop_update_string(DDI_DEV_T_NONE, devi,
	    "device_type", device_type);

	/*
	 * Check whether the "ranges" property is present.
	 * Otherwise create the ranges property by reading
	 * the configuration registers
	 */
	if (ddi_prop_exists(DDI_DEV_T_ANY, devi, DDI_PROP_DONTPASS,
	    "ranges") == 0) {
		pcieb_create_ranges_prop(devi, config_handle);
	}

	if (PCIE_IS_PCI_BDG(bus_p))
		pcieb_set_pci_perf_parameters(devi, config_handle);

#ifdef PX_PLX
	pcieb_attach_plx_workarounds(pcieb);
#endif /* PX_PLX */

	if (pcie_init(devi, NULL) != DDI_SUCCESS)
		goto fail;

	/* Intel PCIe-to-PCI 41210 bridge workaround -- if applicable */
	if (pcieb_disable_41210_wkarnd == 0 &&
	    PCIEB_IS_41210_BRIDGE(bus_p->bus_dev_ven_id))
		pcieb_41210_mps_wkrnd(devi);

	/*
	 * Initialize interrupt handlers. Ignore return value.
	 */
	(void) pcieb_intr_attach(pcieb);

	(void) pcie_hpintr_enable(devi);

	(void) pcie_link_bw_enable(devi);

	/* Do any platform specific workarounds needed at this time */
	pcieb_plat_attach_workaround(devi);

	/*
	 * If this is a root port, we need to go through and at this point in
	 * time set up and initialize all fabric-wide settings such as the max
	 * packet size, tagging, etc. Since this will involve scanning the
	 * fabric, all error enabling and sw workarounds should be in place
	 * before doing this. For hotplug-capable bridges, this will happen
	 * again when a hotplug event occurs. See the pcie theory statement in
	 * uts/common/io/pciex/pcie.c for more information.
	 */
	if (PCIE_IS_RP(bus_p))
		pcie_fabric_setup(devi);

	ddi_report_dev(devi);
	return (DDI_SUCCESS);

fail:
	(void) pcieb_detach(devi, DDI_DETACH);
	return (DDI_FAILURE);
}

static int
pcieb_detach(dev_info_t *devi, ddi_detach_cmd_t cmd)
{
	pcieb_devstate_t *pcieb;
	int error = DDI_SUCCESS;

	switch (cmd) {
	case DDI_SUSPEND:
		error = pcie_pwr_suspend(devi);
		return (error);

	case DDI_DETACH:
		break;

	default:
		return (DDI_FAILURE);
	}

	pcieb = ddi_get_soft_state(pcieb_state, ddi_get_instance(devi));

	/* disable hotplug interrupt */
	(void) pcie_hpintr_disable(devi);

	/* remove interrupt handlers */
	pcieb_intr_fini(pcieb);

	/* uninitialize inband PCI-E HPC if present */
	(void) pcie_uninit(devi);

	(void) ddi_prop_remove(DDI_DEV_T_NONE, devi, "device_type");

	(void) ndi_prop_remove(DDI_DEV_T_NONE, pcieb->pcieb_dip,
	    "pcie_ce_mask");

	if (pcieb->pcieb_init_flags & PCIEB_INIT_FM)
		pcieb_fm_fini(pcieb);

	pcieb_pwr_teardown(devi);
	pwr_common_teardown(devi);

	mutex_destroy(&pcieb->pcieb_peek_poke_mutex);
	mutex_destroy(&pcieb->pcieb_err_mutex);
	mutex_destroy(&pcieb->pcieb_mutex);

	/*
	 * And finally free the per-pci soft state.
	 */
	ddi_soft_state_free(pcieb_state, ddi_get_instance(devi));

	return (DDI_SUCCESS);
}

static int
pcieb_bus_map(dev_info_t *dip, dev_info_t *rdip, ddi_map_req_t *mp,
    off_t offset, off_t len, caddr_t *vaddrp)
{
	dev_info_t *pdip;

	if (PCIE_IS_RP(PCIE_DIP2BUS(dip)) && mp->map_handlep != NULL) {
		ddi_acc_impl_t *hdlp =
		    (ddi_acc_impl_t *)(mp->map_handlep)->ah_platform_private;

		pcieb_set_prot_scan(dip, hdlp);
	}
	pdip = (dev_info_t *)DEVI(dip)->devi_parent;
	return ((DEVI(pdip)->devi_ops->devo_bus_ops->bus_map)(pdip, rdip, mp,
	    offset, len, vaddrp));
}

static int
pcieb_ctlops(dev_info_t *dip, dev_info_t *rdip, ddi_ctl_enum_t ctlop,
    void *arg, void *result)
{
	pci_regspec_t *drv_regp;
	int	reglen;
	int	rn;
	int	totreg;
	pcieb_devstate_t *pcieb = ddi_get_soft_state(pcieb_state,
	    ddi_get_instance(dip));
	struct detachspec *ds;
	struct attachspec *as;

	switch (ctlop) {
	case DDI_CTLOPS_REPORTDEV:
		if (rdip == (dev_info_t *)0)
			return (DDI_FAILURE);

		if (ddi_get_parent(rdip) == dip) {
			cmn_err(CE_CONT, "?PCIE-device: %s@%s, %s%d\n",
			    ddi_node_name(rdip), ddi_get_name_addr(rdip),
			    ddi_driver_name(rdip), ddi_get_instance(rdip));
		}

		/* Pass it up for fabric sync */
		(void) ddi_ctlops(dip, rdip, ctlop, arg, result);
		return (DDI_SUCCESS);

	case DDI_CTLOPS_INITCHILD:
		return (pcieb_initchild((dev_info_t *)arg));

	case DDI_CTLOPS_UNINITCHILD:
		pcieb_uninitchild((dev_info_t *)arg);
		return (DDI_SUCCESS);

	case DDI_CTLOPS_SIDDEV:
		return (DDI_SUCCESS);

	case DDI_CTLOPS_REGSIZE:
	case DDI_CTLOPS_NREGS:
		if (rdip == (dev_info_t *)0)
			return (DDI_FAILURE);
		break;

	case DDI_CTLOPS_PEEK:
	case DDI_CTLOPS_POKE:
		return (pcieb_plat_peekpoke(dip, rdip, ctlop, arg, result));
	case DDI_CTLOPS_ATTACH:
		if (!pcie_is_child(dip, rdip))
			return (DDI_SUCCESS);

		as = (struct attachspec *)arg;
		switch (as->when) {
		case DDI_PRE:
			if (as->cmd == DDI_RESUME) {
				pcie_clear_errors(rdip);
				if (pcieb_plat_ctlops(rdip, ctlop, arg) !=
				    DDI_SUCCESS)
					return (DDI_FAILURE);
			}

			if (as->cmd == DDI_ATTACH)
				return (pcie_pm_hold(dip));

			return (DDI_SUCCESS);

		case DDI_POST:
			if (as->cmd == DDI_ATTACH &&
			    as->result != DDI_SUCCESS) {
				/*
				 * Attach failed for the child device. The child
				 * driver may have made PM calls before the
				 * attach failed. pcie_pm_remove_child() should
				 * cleanup PM state and holds (if any)
				 * associated with the child device.
				 */
				return (pcie_pm_remove_child(dip, rdip));
			}

			if (as->result == DDI_SUCCESS) {
				pf_init(rdip, (void *)pcieb->pcieb_fm_ibc,
				    as->cmd);

				(void) pcieb_plat_ctlops(rdip, ctlop, arg);
			}

			/*
			 * For empty hotplug-capable slots, we should explicitly
			 * disable the errors, so that we won't panic upon
			 * unsupported hotplug messages.
			 */
			if ((!ddi_prop_exists(DDI_DEV_T_ANY, rdip,
			    DDI_PROP_DONTPASS, "hotplug-capable")) ||
			    ddi_get_child(rdip)) {
				(void) pcie_postattach_child(rdip);
				return (DDI_SUCCESS);
			}

			pcie_disable_errors(rdip);

			return (DDI_SUCCESS);
		default:
			break;
		}
		return (DDI_SUCCESS);

	case DDI_CTLOPS_DETACH:
		if (!pcie_is_child(dip, rdip))
			return (DDI_SUCCESS);

		ds = (struct detachspec *)arg;
		switch (ds->when) {
		case DDI_PRE:
			pf_fini(rdip, ds->cmd);
			return (DDI_SUCCESS);

		case DDI_POST:
			if (pcieb_plat_ctlops(rdip, ctlop, arg) != DDI_SUCCESS)
				return (DDI_FAILURE);
			if (ds->cmd == DDI_DETACH &&
			    ds->result == DDI_SUCCESS) {
				return (pcie_pm_remove_child(dip, rdip));
			}
			return (DDI_SUCCESS);
		default:
			break;
		}
		return (DDI_SUCCESS);
	default:
		return (ddi_ctlops(dip, rdip, ctlop, arg, result));
	}

	*(int *)result = 0;
	if (ddi_getlongprop(DDI_DEV_T_ANY, rdip,
	    DDI_PROP_DONTPASS | DDI_PROP_CANSLEEP, "reg", (caddr_t)&drv_regp,
	    &reglen) != DDI_SUCCESS)
		return (DDI_FAILURE);

	totreg = reglen / sizeof (pci_regspec_t);
	if (ctlop == DDI_CTLOPS_NREGS)
		*(int *)result = totreg;
	else if (ctlop == DDI_CTLOPS_REGSIZE) {
		rn = *(int *)arg;
		if (rn >= totreg) {
			kmem_free(drv_regp, reglen);
			return (DDI_FAILURE);
		}

		*(off_t *)result = drv_regp[rn].pci_size_low |
		    ((uint64_t)drv_regp[rn].pci_size_hi << 32);
	}

	kmem_free(drv_regp, reglen);
	return (DDI_SUCCESS);
}

/*
 * name_child
 *
 * This function is called from init_child to name a node. It is
 * also passed as a callback for node merging functions.
 *
 * return value: DDI_SUCCESS, DDI_FAILURE
 */
static int
pcieb_name_child(dev_info_t *child, char *name, int namelen)
{
	pci_regspec_t *pci_rp;
	uint_t device, func;
	char **unit_addr;
	uint_t n;

	/*
	 * For .conf nodes, use unit-address property as name
	 */
	if (ndi_dev_is_persistent_node(child) == 0) {
		if (ddi_prop_lookup_string_array(DDI_DEV_T_ANY, child,
		    DDI_PROP_DONTPASS, "unit-address", &unit_addr, &n) !=
		    DDI_PROP_SUCCESS) {
			cmn_err(CE_WARN,
			    "cannot find unit-address in %s.conf",
			    ddi_driver_name(child));
			return (DDI_FAILURE);
		}
		if (n != 1 || *unit_addr == NULL || **unit_addr == 0) {
			cmn_err(CE_WARN, "unit-address property in %s.conf"
			    " not well-formed", ddi_driver_name(child));
			ddi_prop_free(unit_addr);
			return (DDI_FAILURE);
		}
		(void) snprintf(name, namelen, "%s", *unit_addr);
		ddi_prop_free(unit_addr);
		return (DDI_SUCCESS);
	}

	/*
	 * Get the address portion of the node name based on
	 * the function and device number.
	 */
	if (ddi_prop_lookup_int_array(DDI_DEV_T_ANY, child,
	    DDI_PROP_DONTPASS, "reg", (int **)&pci_rp, &n) != DDI_SUCCESS) {
		return (DDI_FAILURE);
	}

	/* copy the device identifications */
	device = PCI_REG_DEV_G(pci_rp[0].pci_phys_hi);
	func = PCI_REG_FUNC_G(pci_rp[0].pci_phys_hi);

	if (pcie_ari_is_enabled(ddi_get_parent(child))
	    == PCIE_ARI_FORW_ENABLED) {
		func = (device << 3) | func;
		device = 0;
	}

	if (func != 0)
		(void) snprintf(name, namelen, "%x,%x", device, func);
	else
		(void) snprintf(name, namelen, "%x", device);

	ddi_prop_free(pci_rp);
	return (DDI_SUCCESS);
}

static int
pcieb_initchild(dev_info_t *child)
{
	char name[MAXNAMELEN];
	int result = DDI_FAILURE;
	pcieb_devstate_t *pcieb =
	    (pcieb_devstate_t *)ddi_get_soft_state(pcieb_state,
	    ddi_get_instance(ddi_get_parent(child)));

	/*
	 * Name the child
	 */
	if (pcieb_name_child(child, name, MAXNAMELEN) != DDI_SUCCESS) {
		result = DDI_FAILURE;
		goto done;
	}
	ddi_set_name_addr(child, name);

	/*
	 * Pseudo nodes indicate a prototype node with per-instance
	 * properties to be merged into the real h/w device node.
	 * The interpretation of the unit-address is DD[,F]
	 * where DD is the device id and F is the function.
	 */
	if (ndi_dev_is_persistent_node(child) == 0) {
		extern int pci_allow_pseudo_children;

		/*
		 * Try to merge the properties from this prototype
		 * node into real h/w nodes.
		 */
		if (ndi_merge_node(child, pcieb_name_child) == DDI_SUCCESS) {
			/*
			 * Merged ok - return failure to remove the node.
			 */
			ddi_set_name_addr(child, NULL);
			result = DDI_FAILURE;
			goto done;
		}

		/* workaround for ddivs to run under PCI-E */
		if (pci_allow_pseudo_children) {
			result = DDI_SUCCESS;
			goto done;
		}

		/*
		 * The child was not merged into a h/w node,
		 * but there's not much we can do with it other
		 * than return failure to cause the node to be removed.
		 */
		cmn_err(CE_WARN, "!%s@%s: %s.conf properties not merged",
		    ddi_driver_name(child), ddi_get_name_addr(child),
		    ddi_driver_name(child));
		ddi_set_name_addr(child, NULL);
		result = DDI_NOT_WELL_FORMED;
		goto done;
	}

	/* platform specific initchild */
	pcieb_plat_initchild(child);

	if (pcie_pm_hold(pcieb->pcieb_dip) != DDI_SUCCESS) {
		PCIEB_DEBUG(DBG_PWR, pcieb->pcieb_dip,
		    "INITCHILD: px_pm_hold failed\n");
		result = DDI_FAILURE;
		goto done;
	}
	/* Any return from here must call pcie_pm_release */

	/*
	 * If configuration registers were previously saved by
	 * child (before it entered D3), then let the child do the
	 * restore to set up the config regs as it'll first need to
	 * power the device out of D3.
	 */
	if (ddi_prop_exists(DDI_DEV_T_ANY, child, DDI_PROP_DONTPASS,
	    "config-regs-saved-by-child") == 1) {
		PCIEB_DEBUG(DBG_PWR, ddi_get_parent(child),
		    "INITCHILD: config regs to be restored by child"
		    " for %s@%s\n", ddi_node_name(child),
		    ddi_get_name_addr(child));

		result = DDI_SUCCESS;
		goto cleanup;
	}

	PCIEB_DEBUG(DBG_PWR, ddi_get_parent(child),
	    "INITCHILD: config regs setup for %s@%s\n",
	    ddi_node_name(child), ddi_get_name_addr(child));

	pcie_init_dom(child);

	if (pcie_initchild(child) != DDI_SUCCESS) {
		result = DDI_FAILURE;
		pcie_fini_dom(child);
		goto cleanup;
	}

#ifdef PX_PLX
	if (pcieb_init_plx_workarounds(pcieb, child) == DDI_FAILURE) {
		result = DDI_FAILURE;
		pcie_fini_dom(child);
		goto cleanup;
	}
#endif /* PX_PLX */

	result = DDI_SUCCESS;
cleanup:
	pcie_pm_release(pcieb->pcieb_dip);
done:
	return (result);
}

static void
pcieb_uninitchild(dev_info_t *dip)
{

	pcie_uninitchild(dip);

	pcieb_plat_uninitchild(dip);

	ddi_set_name_addr(dip, NULL);

	/*
	 * Strip the node to properly convert it back to prototype form
	 */
	ddi_remove_minor_node(dip, NULL);

	ddi_prop_remove_all(dip);
}

static boolean_t
pcieb_is_pcie_device_type(dev_info_t *dip)
{
	pcie_bus_t	*bus_p = PCIE_DIP2BUS(dip);

	if (PCIE_IS_SW(bus_p) || PCIE_IS_RP(bus_p) || PCIE_IS_PCI2PCIE(bus_p))
		return (B_TRUE);

	return (B_FALSE);
}

static int
pcieb_intr_attach(pcieb_devstate_t *pcieb)
{
	int			intr_types;
	dev_info_t		*dip = pcieb->pcieb_dip;

	/* Allow platform specific code to do any initialization first */
	pcieb_plat_intr_attach(pcieb);

	/*
	 * Initialize interrupt handlers.
	 * If both MSI and FIXED are supported, try to attach MSI first.
	 * If MSI fails for any reason, then try FIXED, but only allow one
	 * type to be attached.
	 */
	if (ddi_intr_get_supported_types(dip, &intr_types) != DDI_SUCCESS) {
		PCIEB_DEBUG(DBG_ATTACH, dip, "ddi_intr_get_supported_types"
		    " failed\n");
		goto FAIL;
	}

	if ((intr_types & DDI_INTR_TYPE_MSI) &&
	    (pcieb_msi_supported(dip) == DDI_SUCCESS)) {
		if (pcieb_intr_init(pcieb, DDI_INTR_TYPE_MSI) == DDI_SUCCESS)
			intr_types = DDI_INTR_TYPE_MSI;
		else {
			PCIEB_DEBUG(DBG_ATTACH, dip, "Unable to attach MSI"
			    " handler\n");
		}
	}

	if (intr_types != DDI_INTR_TYPE_MSI) {
		/*
		 * MSIs are not supported or MSI initialization failed. For Root
		 * Ports mark this so error handling might try to fallback to
		 * some other mechanism if available (machinecheck etc.).
		 */
		if (PCIE_IS_RP(PCIE_DIP2UPBUS(dip)))
			pcieb->pcieb_no_aer_msi = B_TRUE;
	}

	if (intr_types & DDI_INTR_TYPE_FIXED) {
		if (pcieb_intr_init(pcieb, DDI_INTR_TYPE_FIXED) !=
		    DDI_SUCCESS) {
			PCIEB_DEBUG(DBG_ATTACH, dip,
			    "Unable to attach INTx handler\n");
			goto FAIL;
		}
	}
	return (DDI_SUCCESS);

FAIL:
	return (DDI_FAILURE);
}

/*
 * This function initializes internally generated interrupts only.
 * It does not affect any interrupts generated by downstream devices
 * or the forwarding of them.
 *
 * Enable Device Specific Interrupts or Hotplug features here.
 * Enabling features may change how many interrupts are requested
 * by the device.  If features are not enabled first, the
 * device might not ask for any interrupts.
 */
static int
pcieb_intr_init(pcieb_devstate_t *pcieb, int intr_type)
{
	dev_info_t	*dip = pcieb->pcieb_dip;
	int		nintrs, request, count, x;
	int		intr_cap = 0;
	int		inum = 0;
	int		ret;
	pcie_bus_t	*bus_p = PCIE_DIP2UPBUS(dip);
	uint16_t	vendorid = bus_p->bus_dev_ven_id & 0xFFFF;
	boolean_t	is_hp = B_FALSE;
	boolean_t	is_pme = B_FALSE;
	boolean_t	is_lbw = B_FALSE;

	PCIEB_DEBUG(DBG_ATTACH, dip, "pcieb_intr_init: Attaching %s handler\n",
	    (intr_type == DDI_INTR_TYPE_MSI) ? "MSI" : "INTx");

	request = 0;
	if (PCIE_IS_HOTPLUG_ENABLED(dip)) {
		is_hp = B_TRUE;
	}

	if ((intr_type == DDI_INTR_TYPE_MSI) && PCIE_IS_RP(bus_p) &&
	    (vendorid == NVIDIA_VENDOR_ID)) {
		is_pme = B_TRUE;
	}

	if (intr_type == DDI_INTR_TYPE_MSI && pcie_link_bw_supported(dip)) {
		is_lbw = B_TRUE;
	}

	/*
	 * The hot-plug, link bandwidth, and power management events all are
	 * based on the PCI Express capability. Therefore, they all share their
	 * own interrupt.
	 */
	if (is_hp || is_pme || is_lbw) {
		request++;
	}

	/*
	 * If this device is a root port, which means it can have MSI interrupts
	 * enabled for AERs, then we need to request one.
	 */
	if (intr_type == DDI_INTR_TYPE_MSI) {
		if (PCIE_IS_RP(bus_p) && PCIE_HAS_AER(bus_p)) {
			request++;
		}
	}

	if (request == 0)
		return (DDI_SUCCESS);

	/*
	 * Get number of supported interrupts.
	 *
	 * Several Bridges/Switches will not have this property set, resulting
	 * in a FAILURE, if the device is not configured in a way that
	 * interrupts are needed. (eg. hotplugging)
	 */
	ret = ddi_intr_get_nintrs(dip, intr_type, &nintrs);
	if ((ret != DDI_SUCCESS) || (nintrs == 0)) {
		PCIEB_DEBUG(DBG_ATTACH, dip, "ddi_intr_get_nintrs ret:%d"
		    " req:%d\n", ret, nintrs);
		return (DDI_FAILURE);
	}

	PCIEB_DEBUG(DBG_ATTACH, dip, "bdf 0x%x: ddi_intr_get_nintrs: nintrs %d",
	    " request %d\n", bus_p->bus_bdf, nintrs, request);

	if (request > nintrs)
		request = nintrs;

	/* Allocate an array of interrupt handlers */
	pcieb->pcieb_htable_size = sizeof (ddi_intr_handle_t) * request;
	pcieb->pcieb_htable = kmem_zalloc(pcieb->pcieb_htable_size,
	    KM_SLEEP);
	pcieb->pcieb_init_flags |= PCIEB_INIT_HTABLE;

	ret = ddi_intr_alloc(dip, pcieb->pcieb_htable, intr_type, inum,
	    request, &count, DDI_INTR_ALLOC_NORMAL);
	if ((ret != DDI_SUCCESS) || (count == 0)) {
		PCIEB_DEBUG(DBG_ATTACH, dip, "ddi_intr_alloc() ret: %d ask: %d"
		    " actual: %d\n", ret, request, count);
		goto FAIL;
	}
	pcieb->pcieb_init_flags |= PCIEB_INIT_ALLOC;

	/* Save the actual number of interrupts allocated */
	pcieb->pcieb_intr_count = count;
	if (count < request) {
		PCIEB_DEBUG(DBG_ATTACH, dip, "bdf 0%x: Requested Intr: %d"
		    " Received: %d\n", bus_p->bus_bdf, request, count);
	}

	/*
	 * NVidia (MCP55 and other) chipsets have a errata that if the number
	 * of requested MSI intrs is not allocated we have to fall back to INTx.
	 */
	if (intr_type == DDI_INTR_TYPE_MSI) {
		if (PCIE_IS_RP(bus_p) && (vendorid == NVIDIA_VENDOR_ID)) {
			if (request != count)
				goto FAIL;
		}
	}

	/* Get interrupt priority */
	ret = ddi_intr_get_pri(pcieb->pcieb_htable[0],
	    &pcieb->pcieb_intr_priority);
	if (ret != DDI_SUCCESS) {
		PCIEB_DEBUG(DBG_ATTACH, dip, "ddi_intr_get_pri() ret: %d\n",
		    ret);
		goto FAIL;
	}

	if (pcieb->pcieb_intr_priority >= LOCK_LEVEL) {
		pcieb->pcieb_intr_priority = LOCK_LEVEL - 1;
		ret = ddi_intr_set_pri(pcieb->pcieb_htable[0],
		    pcieb->pcieb_intr_priority);
		if (ret != DDI_SUCCESS) {
			PCIEB_DEBUG(DBG_ATTACH, dip, "ddi_intr_set_pri() ret:"
			" %d\n", ret);

			goto FAIL;
		}
	}

	mutex_init(&pcieb->pcieb_intr_mutex, NULL, MUTEX_DRIVER, NULL);

	pcieb->pcieb_init_flags |= PCIEB_INIT_MUTEX;

	for (count = 0; count < pcieb->pcieb_intr_count; count++) {
		ret = ddi_intr_add_handler(pcieb->pcieb_htable[count],
		    pcieb_intr_handler, (caddr_t)pcieb,
		    (caddr_t)(uintptr_t)(inum + count));

		if (ret != DDI_SUCCESS) {
			PCIEB_DEBUG(DBG_ATTACH, dip, "Cannot add "
			    "interrupt(%d)\n", ret);
			break;
		}
	}

	/* If unsucessful, remove the added handlers */
	if (ret != DDI_SUCCESS) {
		for (x = 0; x < count; x++) {
			(void) ddi_intr_remove_handler(pcieb->pcieb_htable[x]);
		}
		goto FAIL;
	}

	pcieb->pcieb_init_flags |= PCIEB_INIT_HANDLER;

	(void) ddi_intr_get_cap(pcieb->pcieb_htable[0], &intr_cap);

	/*
	 * Get this intr lock because we are not quite ready to handle
	 * interrupts immediately after enabling it. The MSI multi register
	 * gets programmed in ddi_intr_enable after which we need to get the
	 * MSI offsets for Hotplug/AER.
	 */
	mutex_enter(&pcieb->pcieb_intr_mutex);

	if (intr_cap & DDI_INTR_FLAG_BLOCK) {
		(void) ddi_intr_block_enable(pcieb->pcieb_htable,
		    pcieb->pcieb_intr_count);
		pcieb->pcieb_init_flags |= PCIEB_INIT_BLOCK;
	} else {
		for (count = 0; count < pcieb->pcieb_intr_count; count++) {
			(void) ddi_intr_enable(pcieb->pcieb_htable[count]);
		}
	}
	pcieb->pcieb_init_flags |= PCIEB_INIT_ENABLE;

	/* Save the interrupt type */
	pcieb->pcieb_intr_type = intr_type;

	/* Get the MSI offset for hotplug/PME from the PCIe cap reg */
	if (intr_type == DDI_INTR_TYPE_MSI) {
		uint16_t pcie_msi_off;
		pcie_msi_off = PCI_CAP_GET16(bus_p->bus_cfg_hdl, 0,
		    bus_p->bus_pcie_off, PCIE_PCIECAP) &
		    PCIE_PCIECAP_INT_MSG_NUM;

		if (pcie_msi_off >= count) {
			PCIEB_DEBUG(DBG_ATTACH, dip, "MSI number %u in PCIe "
			    "cap > max allocated %d\n", pcie_msi_off, count);
			mutex_exit(&pcieb->pcieb_intr_mutex);
			goto FAIL;
		}

		if (is_hp) {
			pcieb->pcieb_isr_tab[pcie_msi_off] |= PCIEB_INTR_SRC_HP;
		}

		if (is_pme) {
			pcieb->pcieb_isr_tab[pcie_msi_off] |=
			    PCIEB_INTR_SRC_PME;
		}

		if (is_lbw) {
			pcieb->pcieb_isr_tab[pcie_msi_off] |=
			    PCIEB_INTR_SRC_LBW;
		}
	} else {
		/* INTx handles only Hotplug interrupts */
		if (is_hp)
			pcieb->pcieb_isr_tab[0] |= PCIEB_INTR_SRC_HP;
	}


	/*
	 * Get the MSI offset for errors from the AER Root Error status
	 * register.
	 */
	if ((intr_type == DDI_INTR_TYPE_MSI) && PCIE_IS_RP(bus_p)) {
		if (PCIE_HAS_AER(bus_p)) {
			int aer_msi_off;
			aer_msi_off = (PCI_XCAP_GET32(bus_p->bus_cfg_hdl, 0,
			    bus_p->bus_aer_off, PCIE_AER_RE_STS) >>
			    PCIE_AER_RE_STS_MSG_NUM_SHIFT) &
			    PCIE_AER_RE_STS_MSG_NUM_MASK;

			if (aer_msi_off >= count) {
				PCIEB_DEBUG(DBG_ATTACH, dip, "MSI number %d in"
				    " AER cap > max allocated %d\n",
				    aer_msi_off, count);
				mutex_exit(&pcieb->pcieb_intr_mutex);
				goto FAIL;
			}
			pcieb->pcieb_isr_tab[aer_msi_off] |= PCIEB_INTR_SRC_AER;
		} else {
			/*
			 * This RP does not have AER. Fallback to the
			 * SERR+Machinecheck approach if available.
			 */
			pcieb->pcieb_no_aer_msi = B_TRUE;
		}
	}

	mutex_exit(&pcieb->pcieb_intr_mutex);
	return (DDI_SUCCESS);

FAIL:
	pcieb_intr_fini(pcieb);
	return (DDI_FAILURE);
}

static void
pcieb_intr_fini(pcieb_devstate_t *pcieb)
{
	int x;
	int count = pcieb->pcieb_intr_count;
	int flags = pcieb->pcieb_init_flags;

	if ((flags & PCIEB_INIT_ENABLE) &&
	    (flags & PCIEB_INIT_BLOCK)) {
		(void) ddi_intr_block_disable(pcieb->pcieb_htable, count);
		flags &= ~(PCIEB_INIT_ENABLE |
		    PCIEB_INIT_BLOCK);
	}

	if (flags & PCIEB_INIT_MUTEX)
		mutex_destroy(&pcieb->pcieb_intr_mutex);

	for (x = 0; x < count; x++) {
		if (flags & PCIEB_INIT_ENABLE)
			(void) ddi_intr_disable(pcieb->pcieb_htable[x]);

		if (flags & PCIEB_INIT_HANDLER)
			(void) ddi_intr_remove_handler(pcieb->pcieb_htable[x]);

		if (flags & PCIEB_INIT_ALLOC)
			(void) ddi_intr_free(pcieb->pcieb_htable[x]);
	}

	flags &= ~(PCIEB_INIT_ENABLE | PCIEB_INIT_HANDLER | PCIEB_INIT_ALLOC |
	    PCIEB_INIT_MUTEX);

	if (flags & PCIEB_INIT_HTABLE)
		kmem_free(pcieb->pcieb_htable, pcieb->pcieb_htable_size);

	flags &= ~PCIEB_INIT_HTABLE;

	pcieb->pcieb_init_flags &= flags;
}

/*
 * Checks if this device needs MSIs enabled or not.
 */
/*ARGSUSED*/
static int
pcieb_msi_supported(dev_info_t *dip)
{
	return ((pcieb_enable_msi && pcieb_plat_msi_supported(dip)) ?
	    DDI_SUCCESS: DDI_FAILURE);
}

/*ARGSUSED*/
static int
pcieb_fm_init_child(dev_info_t *dip, dev_info_t *tdip, int cap,
    ddi_iblock_cookie_t *ibc)
{
	pcieb_devstate_t  *pcieb = ddi_get_soft_state(pcieb_state,
	    ddi_get_instance(dip));

	ASSERT(ibc != NULL);
	*ibc = pcieb->pcieb_fm_ibc;

	return (DEVI(dip)->devi_fmhdl->fh_cap | DDI_FM_ACCCHK_CAPABLE |
	    DDI_FM_DMACHK_CAPABLE);
}

static int
pcieb_fm_init(pcieb_devstate_t *pcieb_p)
{
	dev_info_t	*dip = pcieb_p->pcieb_dip;
	int		fm_cap = DDI_FM_EREPORT_CAPABLE;

	/*
	 * Request our capability level and get our parents capability
	 * and ibc.
	 */
	ddi_fm_init(dip, &fm_cap, &pcieb_p->pcieb_fm_ibc);

	return (DDI_SUCCESS);
}

/*
 * Breakdown our FMA resources
 */
static void
pcieb_fm_fini(pcieb_devstate_t *pcieb_p)
{
	/*
	 * Clean up allocated fm structures
	 */
	ddi_fm_fini(pcieb_p->pcieb_dip);
}

static int
pcieb_open(dev_t *devp, int flags, int otyp, cred_t *credp)
{
	int		inst = PCI_MINOR_NUM_TO_INSTANCE(getminor(*devp));
	pcieb_devstate_t	*pcieb = ddi_get_soft_state(pcieb_state, inst);
	int	rv;

	if (pcieb == NULL)
		return (ENXIO);

	mutex_enter(&pcieb->pcieb_mutex);
	rv = pcie_open(pcieb->pcieb_dip, devp, flags, otyp, credp);
	mutex_exit(&pcieb->pcieb_mutex);

	return (rv);
}

static int
pcieb_close(dev_t dev, int flags, int otyp, cred_t *credp)
{
	int		inst = PCI_MINOR_NUM_TO_INSTANCE(getminor(dev));
	pcieb_devstate_t	*pcieb = ddi_get_soft_state(pcieb_state, inst);
	int	rv;

	if (pcieb == NULL)
		return (ENXIO);

	mutex_enter(&pcieb->pcieb_mutex);
	rv = pcie_close(pcieb->pcieb_dip, dev, flags, otyp, credp);
	mutex_exit(&pcieb->pcieb_mutex);

	return (rv);
}

static int
pcieb_ioctl_retrain(pcieb_devstate_t *pcieb, cred_t *credp)
{
	pcie_bus_t	*bus_p = PCIE_DIP2BUS(pcieb->pcieb_dip);

	if (drv_priv(credp) != 0) {
		return (EPERM);
	}

	if (!PCIE_IS_PCIE(bus_p)) {
		return (ENOTSUP);
	}

	if (!PCIE_IS_RP(bus_p) && !PCIE_IS_SWD(bus_p)) {
		return (ENOTSUP);
	}

	return (pcie_link_retrain(pcieb->pcieb_dip));
}

static int
pcieb_ioctl_get_speed(pcieb_devstate_t *pcieb, intptr_t arg, int mode,
    cred_t *credp)
{
	pcie_bus_t			*bus_p = PCIE_DIP2BUS(pcieb->pcieb_dip);
	pcieb_ioctl_target_speed_t	pits;

	if (drv_priv(credp) != 0) {
		return (EPERM);
	}

	if (!PCIE_IS_PCIE(bus_p)) {
		return (ENOTSUP);
	}

	if (!PCIE_IS_RP(bus_p) && !PCIE_IS_SWD(bus_p)) {
		return (ENOTSUP);
	}

	pits.pits_flags = 0;
	pits.pits_speed = PCIEB_LINK_SPEED_UNKNOWN;

	mutex_enter(&bus_p->bus_speed_mutex);
	if ((bus_p->bus_speed_flags & PCIE_LINK_F_ADMIN_TARGET) != 0) {
		pits.pits_flags |= PCIEB_FLAGS_ADMIN_SET;
	}
	switch (bus_p->bus_target_speed) {
	case PCIE_LINK_SPEED_2_5:
		pits.pits_speed = PCIEB_LINK_SPEED_GEN1;
		break;
	case PCIE_LINK_SPEED_5:
		pits.pits_speed = PCIEB_LINK_SPEED_GEN2;
		break;
	case PCIE_LINK_SPEED_8:
		pits.pits_speed = PCIEB_LINK_SPEED_GEN3;
		break;
	case PCIE_LINK_SPEED_16:
		pits.pits_speed = PCIEB_LINK_SPEED_GEN4;
		break;
	case PCIE_LINK_SPEED_32:
		pits.pits_speed = PCIEB_LINK_SPEED_GEN5;
		break;
	case PCIE_LINK_SPEED_64:
		pits.pits_speed = PCIEB_LINK_SPEED_GEN6;
		break;
	default:
		pits.pits_speed = PCIEB_LINK_SPEED_UNKNOWN;
		break;
	}
	mutex_exit(&bus_p->bus_speed_mutex);

	if (ddi_copyout(&pits, (void *)arg, sizeof (pits),
	    mode & FKIOCTL) != 0) {
		return (EFAULT);
	}

	return (0);
}

static int
pcieb_ioctl_set_speed(pcieb_devstate_t *pcieb, intptr_t arg, int mode,
    cred_t *credp)
{
	pcie_bus_t			*bus_p = PCIE_DIP2BUS(pcieb->pcieb_dip);
	pcieb_ioctl_target_speed_t	pits;
	pcie_link_speed_t		speed;

	if (drv_priv(credp) != 0) {
		return (EPERM);
	}

	if (!PCIE_IS_PCIE(bus_p)) {
		return (ENOTSUP);
	}

	if (!PCIE_IS_RP(bus_p) && !PCIE_IS_SWD(bus_p)) {
		return (ENOTSUP);
	}

	if (ddi_copyin((void *)arg, &pits, sizeof (pits),
	    mode & FKIOCTL) != 0) {
		return (EFAULT);
	}

	if (pits.pits_flags != 0) {
		return (EINVAL);
	}

	switch (pits.pits_speed) {
	case PCIEB_LINK_SPEED_GEN1:
		speed = PCIE_LINK_SPEED_2_5;
		break;
	case PCIEB_LINK_SPEED_GEN2:
		speed = PCIE_LINK_SPEED_5;
		break;
	case PCIEB_LINK_SPEED_GEN3:
		speed = PCIE_LINK_SPEED_8;
		break;
	case PCIEB_LINK_SPEED_GEN4:
		speed = PCIE_LINK_SPEED_16;
		break;
	case PCIEB_LINK_SPEED_GEN5:
		speed = PCIE_LINK_SPEED_32;
		break;
	case PCIEB_LINK_SPEED_GEN6:
		speed = PCIE_LINK_SPEED_64;
		break;
	default:
		return (EINVAL);
	}

	return (pcie_link_set_target(pcieb->pcieb_dip, speed));
}

static int
pcieb_ioctl(dev_t dev, int cmd, intptr_t arg, int mode, cred_t *credp,
    int *rvalp)
{
	int		inst = PCI_MINOR_NUM_TO_INSTANCE(getminor(dev));
	pcieb_devstate_t	*pcieb = ddi_get_soft_state(pcieb_state, inst);
	int		rv;

	if (pcieb == NULL)
		return (ENXIO);

	/*
	 * Check if this is one of the commands that the bridge driver natively
	 * understands. There are only a handful of such private ioctls defined
	 * in pcieb_ioctl.h. Otherwise, this ioctl should be handled by the
	 * general pcie driver.
	 */
	switch (cmd) {
	case PCIEB_IOCTL_RETRAIN:
		rv = pcieb_ioctl_retrain(pcieb, credp);
		break;
	case PCIEB_IOCTL_GET_TARGET_SPEED:
		rv = pcieb_ioctl_get_speed(pcieb, arg, mode, credp);
		break;
	case PCIEB_IOCTL_SET_TARGET_SPEED:
		rv = pcieb_ioctl_set_speed(pcieb, arg, mode, credp);
		break;
	default:
		/* To handle devctl and hotplug related ioctls */
		rv = pcie_ioctl(pcieb->pcieb_dip, dev, cmd, arg, mode, credp,
		    rvalp);
		break;
	}

	return (rv);
}

/*
 * Common interrupt handler for hotplug, PME and errors.
 */
static uint_t
pcieb_intr_handler(caddr_t arg1, caddr_t arg2)
{
	pcieb_devstate_t *pcieb_p = (pcieb_devstate_t *)arg1;
	dev_info_t	*dip = pcieb_p->pcieb_dip;
	ddi_fm_error_t	derr;
	int		sts = 0;
	int		ret = DDI_INTR_UNCLAIMED;
	int		isrc;

	if (!(pcieb_p->pcieb_init_flags & PCIEB_INIT_ENABLE))
		goto FAIL;

	mutex_enter(&pcieb_p->pcieb_intr_mutex);
	isrc = pcieb_p->pcieb_isr_tab[(int)(uintptr_t)arg2];
	mutex_exit(&pcieb_p->pcieb_intr_mutex);

	PCIEB_DEBUG(DBG_INTR, dip, "Received intr number %d\n",
	    (int)(uintptr_t)arg2);

	if (isrc == PCIEB_INTR_SRC_UNKNOWN)
		goto FAIL;

	if (isrc & (PCIEB_INTR_SRC_HP | PCIEB_INTR_SRC_LBW))
		ret = pcie_intr(dip);

	if (isrc & PCIEB_INTR_SRC_PME)
		ret = DDI_INTR_CLAIMED;

	/* AER Error */
	if (isrc & PCIEB_INTR_SRC_AER) {
		/*
		 *  If MSI is shared with PME/hotplug then check Root Error
		 *  Status Reg before claiming it. For now it's ok since
		 *  we know we get 2 MSIs.
		 */
		ret = DDI_INTR_CLAIMED;
		bzero(&derr, sizeof (ddi_fm_error_t));
		derr.fme_version = DDI_FME_VERSION;
		mutex_enter(&pcieb_p->pcieb_peek_poke_mutex);
		mutex_enter(&pcieb_p->pcieb_err_mutex);

		pf_eh_enter(PCIE_DIP2BUS(dip));
		PCIE_ROOT_EH_SRC(PCIE_DIP2PFD(dip))->intr_type =
		    PF_INTR_TYPE_AER;

		if ((DEVI(dip)->devi_fmhdl->fh_cap) & DDI_FM_EREPORT_CAPABLE)
			sts = pf_scan_fabric(dip, &derr, NULL);
		pf_eh_exit(PCIE_DIP2BUS(dip));

		mutex_exit(&pcieb_p->pcieb_err_mutex);
		mutex_exit(&pcieb_p->pcieb_peek_poke_mutex);
		if (pcieb_die & sts)
			fm_panic("%s-%d: PCI(-X) Express Fatal Error. (0x%x)",
			    ddi_driver_name(dip), ddi_get_instance(dip), sts);
	}
FAIL:
	return (ret);
}

/*
 * Some PCI-X to PCI-E bridges do not support full 64-bit addressing on the
 * PCI-X side of the bridge.  We build a special version of this driver for
 * those bridges, which uses PCIEB_ADDR_LIMIT_LO and/or PCIEB_ADDR_LIMIT_HI
 * to define the range of values which the chip can handle.  The code below
 * then clamps the DMA address range supplied by the driver, preventing the
 * PCI-E nexus driver from allocating any memory the bridge can't deal
 * with.
 */
static int
pcieb_dma_allochdl(dev_info_t *dip, dev_info_t *rdip,
    ddi_dma_attr_t *attr_p, int (*waitfp)(caddr_t), caddr_t arg,
    ddi_dma_handle_t *handlep)
{
	int		ret;
#ifdef	PCIEB_BCM
	uint64_t	lim;

	/*
	 * If the leaf device's limits are outside than what the Broadcom
	 * bridge can handle, we need to clip the values passed up the chain.
	 */
	lim = attr_p->dma_attr_addr_lo;
	attr_p->dma_attr_addr_lo = MAX(lim, PCIEB_ADDR_LIMIT_LO);

	lim = attr_p->dma_attr_addr_hi;
	attr_p->dma_attr_addr_hi = MIN(lim, PCIEB_ADDR_LIMIT_HI);

#endif	/* PCIEB_BCM */

	/*
	 * This is a software workaround to fix the Broadcom 5714/5715 PCIe-PCI
	 * bridge prefetch bug. Intercept the DMA alloc handle request and set
	 * PX_DMAI_FLAGS_MAP_BUFZONE flag in the handle. If this flag is set,
	 * the px nexus driver will allocate an extra page & make it valid one,
	 * for any DVMA request that comes from any of the Broadcom bridge child
	 * devices.
	 */
	if ((ret = ddi_dma_allochdl(dip, rdip, attr_p, waitfp, arg,
	    handlep)) == DDI_SUCCESS) {
		ddi_dma_impl_t	*mp = (ddi_dma_impl_t *)*handlep;
#ifdef	PCIEB_BCM
		mp->dmai_inuse |= PX_DMAI_FLAGS_MAP_BUFZONE;
#endif	/* PCIEB_BCM */
		/*
		 * For a given rdip, update mp->dmai_bdf with the bdf value
		 * of pcieb's immediate child or secondary bus-id of the
		 * PCIe2PCI bridge.
		 */
		mp->dmai_minxfer = pcie_get_bdf_for_dma_xfer(dip, rdip);
	}

	return (ret);
}

/*
 * FDVMA feature is not supported for any child device of Broadcom 5714/5715
 * PCIe-PCI bridge due to prefetch bug. Return failure immediately, so that
 * these drivers will switch to regular DVMA path.
 */
/*ARGSUSED*/
static int
pcieb_dma_mctl(dev_info_t *dip, dev_info_t *rdip, ddi_dma_handle_t handle,
    enum ddi_dma_ctlops cmd, off_t *offp, size_t *lenp, caddr_t *objp,
    uint_t cache_flags)
{
	int	ret;

#ifdef	PCIEB_BCM
	if (cmd == DDI_DMA_RESERVE)
		return (DDI_FAILURE);
#endif	/* PCIEB_BCM */

	if (((ret = ddi_dma_mctl(dip, rdip, handle, cmd, offp, lenp, objp,
	    cache_flags)) == DDI_SUCCESS) && (cmd == DDI_DMA_RESERVE)) {
		ddi_dma_impl_t	*mp = (ddi_dma_impl_t *)*objp;

		/*
		 * For a given rdip, update mp->dmai_bdf with the bdf value
		 * of pcieb's immediate child or secondary bus-id of the
		 * PCIe2PCI bridge.
		 */
		mp->dmai_minxfer = pcie_get_bdf_for_dma_xfer(dip, rdip);
	}

	return (ret);
}

static int
pcieb_intr_ops(dev_info_t *dip, dev_info_t *rdip, ddi_intr_op_t intr_op,
    ddi_intr_handle_impl_t *hdlp, void *result)
{
	return (pcieb_plat_intr_ops(dip, rdip, intr_op, hdlp, result));

}

/*
 * Power management related initialization specific to pcieb.
 * Called by pcieb_attach()
 */
static int
pcieb_pwr_setup(dev_info_t *dip)
{
	char *comp_array[5];
	int i;
	ddi_acc_handle_t conf_hdl;
	uint16_t pmcap, cap_ptr;
	pcie_pwr_t *pwr_p;

	/* Some platforms/devices may choose to disable PM */
	if (pcieb_plat_pwr_disable(dip)) {
		(void) pcieb_pwr_disable(dip);
		return (DDI_SUCCESS);
	}

	ASSERT(PCIE_PMINFO(dip));
	pwr_p = PCIE_NEXUS_PMINFO(dip);
	ASSERT(pwr_p);

	/* Code taken from pci_pci driver */
	if (pci_config_setup(dip, &pwr_p->pwr_conf_hdl) != DDI_SUCCESS) {
		PCIEB_DEBUG(DBG_PWR, dip, "pcieb_pwr_setup: pci_config_setup "
		    "failed\n");
		return (DDI_FAILURE);
	}
	conf_hdl = pwr_p->pwr_conf_hdl;

	/*
	 * Walk the capabilities searching for a PM entry.
	 */
	if ((PCI_CAP_LOCATE(conf_hdl, PCI_CAP_ID_PM, &cap_ptr)) ==
	    DDI_FAILURE) {
		PCIEB_DEBUG(DBG_PWR, dip, "switch/bridge does not support PM. "
		    " PCI PM data structure not found in config header\n");
		pci_config_teardown(&conf_hdl);
		return (DDI_SUCCESS);
	}
	/*
	 * Save offset to pmcsr for future references.
	 */
	pwr_p->pwr_pmcsr_offset = cap_ptr + PCI_PMCSR;
	pmcap = PCI_CAP_GET16(conf_hdl, 0, cap_ptr, PCI_PMCAP);
	if (pmcap & PCI_PMCAP_D1) {
		PCIEB_DEBUG(DBG_PWR, dip, "D1 state supported\n");
		pwr_p->pwr_pmcaps |= PCIE_SUPPORTS_D1;
	}
	if (pmcap & PCI_PMCAP_D2) {
		PCIEB_DEBUG(DBG_PWR, dip, "D2 state supported\n");
		pwr_p->pwr_pmcaps |= PCIE_SUPPORTS_D2;
	}

	i = 0;
	comp_array[i++] = "NAME=PCIe switch/bridge PM";
	comp_array[i++] = "0=Power Off (D3)";
	if (pwr_p->pwr_pmcaps & PCIE_SUPPORTS_D2)
		comp_array[i++] = "1=D2";
	if (pwr_p->pwr_pmcaps & PCIE_SUPPORTS_D1)
		comp_array[i++] = "2=D1";
	comp_array[i++] = "3=Full Power D0";

	/*
	 * Create pm-components property, if it does not exist already.
	 */
	if (ddi_prop_update_string_array(DDI_DEV_T_NONE, dip,
	    "pm-components", comp_array, i) != DDI_PROP_SUCCESS) {
		PCIEB_DEBUG(DBG_PWR, dip, "could not create pm-components "
		    " prop\n");
		pci_config_teardown(&conf_hdl);
		return (DDI_FAILURE);
	}
	return (pcieb_pwr_init_and_raise(dip, pwr_p));
}

/*
 * undo whatever is done in pcieb_pwr_setup. called by pcieb_detach()
 */
static void
pcieb_pwr_teardown(dev_info_t *dip)
{
	pcie_pwr_t	*pwr_p;

	if (!PCIE_PMINFO(dip) || !(pwr_p = PCIE_NEXUS_PMINFO(dip)))
		return;

	(void) ddi_prop_remove(DDI_DEV_T_NONE, dip, "pm-components");
	if (pwr_p->pwr_conf_hdl)
		pci_config_teardown(&pwr_p->pwr_conf_hdl);
}

/*
 * Initializes the power level and raise the power to D0, if it is
 * not at D0.
 */
static int
pcieb_pwr_init_and_raise(dev_info_t *dip, pcie_pwr_t *pwr_p)
{
	uint16_t pmcsr;
	int ret = DDI_SUCCESS;

	/*
	 * Intialize our power level from PMCSR. The common code initializes
	 * this to UNKNOWN. There is no guarantee that we will be at full
	 * power at attach. If we are not at D0, raise the power.
	 */
	pmcsr = pci_config_get16(pwr_p->pwr_conf_hdl, pwr_p->pwr_pmcsr_offset);
	pmcsr &= PCI_PMCSR_STATE_MASK;
	switch (pmcsr) {
	case PCI_PMCSR_D0:
		pwr_p->pwr_func_lvl = PM_LEVEL_D0;
		break;

	case PCI_PMCSR_D1:
		pwr_p->pwr_func_lvl = PM_LEVEL_D1;
		break;

	case PCI_PMCSR_D2:
		pwr_p->pwr_func_lvl = PM_LEVEL_D2;
		break;

	case PCI_PMCSR_D3HOT:
		pwr_p->pwr_func_lvl = PM_LEVEL_D3;
		break;

	default:
		break;
	}

	/* Raise the power to D0. */
	if (pwr_p->pwr_func_lvl != PM_LEVEL_D0 &&
	    ((ret = pm_raise_power(dip, 0, PM_LEVEL_D0)) != DDI_SUCCESS)) {
		/*
		 * Read PMCSR again. If it is at D0, ignore the return
		 * value from pm_raise_power.
		 */
		pmcsr = pci_config_get16(pwr_p->pwr_conf_hdl,
		    pwr_p->pwr_pmcsr_offset);
		if ((pmcsr & PCI_PMCSR_STATE_MASK) == PCI_PMCSR_D0)
			ret = DDI_SUCCESS;
		else {
			PCIEB_DEBUG(DBG_PWR, dip, "pcieb_pwr_setup: could not "
			    "raise power to D0 \n");
		}
	}
	if (ret == DDI_SUCCESS)
		pwr_p->pwr_func_lvl = PM_LEVEL_D0;
	return (ret);
}

/*
 * Disable PM for x86 and PLX 8532 switch.
 * For PLX Transitioning one port on this switch to low power causes links
 * on other ports on the same station to die. Due to PLX erratum #34, we
 * can't allow the downstream device go to non-D0 state.
 */
static int
pcieb_pwr_disable(dev_info_t *dip)
{
	pcie_pwr_t *pwr_p;

	ASSERT(PCIE_PMINFO(dip));
	pwr_p = PCIE_NEXUS_PMINFO(dip);
	ASSERT(pwr_p);
	PCIEB_DEBUG(DBG_PWR, dip, "pcieb_pwr_disable: disabling PM\n");
	pwr_p->pwr_func_lvl = PM_LEVEL_D0;
	pwr_p->pwr_flags = PCIE_NO_CHILD_PM;
	return (DDI_SUCCESS);
}

#ifdef DEBUG
int pcieb_dbg_intr_print = 0;
void
pcieb_dbg(uint_t bit, dev_info_t *dip, char *fmt, ...)
{
	va_list ap;

	if (!pcieb_dbg_print)
		return;

	if (dip)
		prom_printf("%s(%d): %s", ddi_driver_name(dip),
		    ddi_get_instance(dip), pcieb_debug_sym[bit]);

	va_start(ap, fmt);
	if (servicing_interrupt()) {
		if (pcieb_dbg_intr_print)
			prom_vprintf(fmt, ap);
	} else {
		prom_vprintf(fmt, ap);
	}

	va_end(ap);
}
#endif

static void
pcieb_id_props(pcieb_devstate_t *pcieb)
{
	uint64_t serialid = 0;	/* 40b field of EUI-64 serial no. register */
	uint16_t cap_ptr;
	uint8_t fic = 0;	/* 1 = first in chassis device */
	pcie_bus_t *bus_p = PCIE_DIP2BUS(pcieb->pcieb_dip);
	ddi_acc_handle_t config_handle = bus_p->bus_cfg_hdl;

	/*
	 * Identify first in chassis.  In the special case of a Sun branded
	 * PLX device, it obviously is first in chassis.  Otherwise, in the
	 * general case, look for an Expansion Slot Register and check its
	 * first-in-chassis bit.
	 */
#ifdef	PX_PLX
	uint16_t vendor_id = bus_p->bus_dev_ven_id & 0xFFFF;
	uint16_t device_id = bus_p->bus_dev_ven_id >> 16;
	if ((vendor_id == PXB_VENDOR_SUN) &&
	    ((device_id == PXB_DEVICE_PLX_PCIX) ||
	    (device_id == PXB_DEVICE_PLX_PCIE))) {
		fic = 1;
	}
#endif	/* PX_PLX */
	if ((fic == 0) && ((PCI_CAP_LOCATE(config_handle,
	    PCI_CAP_ID_SLOT_ID, &cap_ptr)) != DDI_FAILURE)) {
		uint8_t esr = PCI_CAP_GET8(config_handle, 0,
		    cap_ptr, PCI_CAP_ID_REGS_OFF);
		if (PCI_CAPSLOT_FIC(esr))
			fic = 1;
	}

	if ((PCI_CAP_LOCATE(config_handle,
	    PCI_CAP_XCFG_SPC(PCIE_EXT_CAP_ID_SER), &cap_ptr)) != DDI_FAILURE) {
		/* Serialid can be 0 thru a full 40b number */
		serialid = PCI_XCAP_GET32(config_handle, 0,
		    cap_ptr, PCIE_SER_SID_UPPER_DW);
		serialid <<= 32;
		serialid |= PCI_XCAP_GET32(config_handle, 0,
		    cap_ptr, PCIE_SER_SID_LOWER_DW);
	}

	if (fic)
		(void) ndi_prop_create_boolean(DDI_DEV_T_NONE, pcieb->pcieb_dip,
		    "first-in-chassis");
	if (serialid)
		(void) ddi_prop_update_int64(DDI_DEV_T_NONE, pcieb->pcieb_dip,
		    "serialid#", serialid);
}

static void
pcieb_create_ranges_prop(dev_info_t *dip,
    ddi_acc_handle_t config_handle)
{
	uint32_t base, limit;
	ppb_ranges_t	ranges[PCIEB_RANGE_LEN];
	uint8_t io_base_lo, io_limit_lo;
	uint16_t io_base_hi, io_limit_hi, mem_base, mem_limit;
	int i = 0, rangelen = sizeof (ppb_ranges_t)/sizeof (int);

	io_base_lo = pci_config_get8(config_handle, PCI_BCNF_IO_BASE_LOW);
	io_limit_lo = pci_config_get8(config_handle, PCI_BCNF_IO_LIMIT_LOW);
	io_base_hi = pci_config_get16(config_handle, PCI_BCNF_IO_BASE_HI);
	io_limit_hi = pci_config_get16(config_handle, PCI_BCNF_IO_LIMIT_HI);
	mem_base = pci_config_get16(config_handle, PCI_BCNF_MEM_BASE);
	mem_limit = pci_config_get16(config_handle, PCI_BCNF_MEM_LIMIT);

	/*
	 * Create ranges for IO space
	 */
	ranges[i].size_low = ranges[i].size_high = 0;
	ranges[i].parent_mid = ranges[i].child_mid = ranges[i].parent_high = 0;
	ranges[i].child_high = ranges[i].parent_high |=
	    (PCI_REG_REL_M | PCI_ADDR_IO);
	base = PCIEB_16bit_IOADDR(io_base_lo);
	limit = PCIEB_16bit_IOADDR(io_limit_lo);

	if ((io_base_lo & 0xf) == PCIEB_32BIT_IO) {
		base = PCIEB_LADDR(base, io_base_hi);
	}
	if ((io_limit_lo & 0xf) == PCIEB_32BIT_IO) {
		limit = PCIEB_LADDR(limit, io_limit_hi);
	}

	if ((io_base_lo & PCIEB_32BIT_IO) && (io_limit_hi > 0)) {
		base = PCIEB_LADDR(base, io_base_hi);
		limit = PCIEB_LADDR(limit, io_limit_hi);
	}

	/*
	 * Create ranges for 32bit memory space
	 */
	base = PCIEB_32bit_MEMADDR(mem_base);
	limit = PCIEB_32bit_MEMADDR(mem_limit);
	ranges[i].size_low = ranges[i].size_high = 0;
	ranges[i].parent_mid = ranges[i].child_mid = ranges[i].parent_high = 0;
	ranges[i].child_high = ranges[i].parent_high |=
	    (PCI_REG_REL_M | PCI_ADDR_MEM32);
	ranges[i].child_low = ranges[i].parent_low = base;
	if (limit >= base) {
		ranges[i].size_low = limit - base + PCIEB_MEMGRAIN;
		i++;
	}

	if (i) {
		(void) ndi_prop_update_int_array(DDI_DEV_T_NONE, dip, "ranges",
		    (int *)ranges, i * rangelen);
	}
}

/*
 * For PCI and PCI-X devices including PCIe2PCI bridge, initialize
 * cache-line-size and latency timer configuration registers.
 */
void
pcieb_set_pci_perf_parameters(dev_info_t *dip, ddi_acc_handle_t cfg_hdl)
{
	uint_t	n;

	/* Initialize cache-line-size configuration register if needed */
	if (ddi_getprop(DDI_DEV_T_ANY, dip, DDI_PROP_DONTPASS,
	    "cache-line-size", 0) == 0) {
		pci_config_put8(cfg_hdl, PCI_CONF_CACHE_LINESZ,
		    PCIEB_CACHE_LINE_SIZE);
		n = pci_config_get8(cfg_hdl, PCI_CONF_CACHE_LINESZ);
		if (n != 0) {
			(void) ndi_prop_update_int(DDI_DEV_T_NONE, dip,
			    "cache-line-size", n);
		}
	}

	/* Initialize latency timer configuration registers if needed */
	if (ddi_getprop(DDI_DEV_T_ANY, dip, DDI_PROP_DONTPASS,
	    "latency-timer", 0) == 0) {
		uchar_t	min_gnt, latency_timer;
		uchar_t header_type;

		/* Determine the configuration header type */
		header_type = pci_config_get8(cfg_hdl, PCI_CONF_HEADER);

		if ((header_type & PCI_HEADER_TYPE_M) == PCI_HEADER_ONE) {
			latency_timer = PCIEB_LATENCY_TIMER;
			pci_config_put8(cfg_hdl, PCI_BCNF_LATENCY_TIMER,
			    latency_timer);
		} else {
			min_gnt = pci_config_get8(cfg_hdl, PCI_CONF_MIN_G);
			latency_timer = min_gnt * 8;
		}

		pci_config_put8(cfg_hdl, PCI_CONF_LATENCY_TIMER,
		    latency_timer);
		n = pci_config_get8(cfg_hdl, PCI_CONF_LATENCY_TIMER);
		if (n != 0) {
			(void) ndi_prop_update_int(DDI_DEV_T_NONE, dip,
			    "latency-timer", n);
		}
	}
}