aboutsummaryrefslogtreecommitdiffstats
path: root/drivers/infiniband/hw/ipath/ipath_mad.c
blob: b34b91d3723a58f4a2e58f0d6233ed3e264f9338 (plain) (blame)
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
/*
 * Copyright (c) 2006, 2007 QLogic Corporation. All rights reserved.
 * Copyright (c) 2005, 2006 PathScale, Inc. All rights reserved.
 *
 * This software is available to you under a choice of one of two
 * licenses.  You may choose to be licensed under the terms of the GNU
 * General Public License (GPL) Version 2, available from the file
 * COPYING in the main directory of this source tree, or the
 * OpenIB.org BSD license below:
 *
 *     Redistribution and use in source and binary forms, with or
 *     without modification, are permitted provided that the following
 *     conditions are met:
 *
 *      - Redistributions of source code must retain the above
 *        copyright notice, this list of conditions and the following
 *        disclaimer.
 *
 *      - Redistributions in binary form must reproduce the above
 *        copyright notice, this list of conditions and the following
 *        disclaimer in the documentation and/or other materials
 *        provided with the distribution.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 * SOFTWARE.
 */

#include <rdma/ib_smi.h>

#include "ipath_kernel.h"
#include "ipath_verbs.h"
#include "ipath_common.h"

#define IB_SMP_UNSUP_VERSION	__constant_htons(0x0004)
#define IB_SMP_UNSUP_METHOD	__constant_htons(0x0008)
#define IB_SMP_UNSUP_METH_ATTR	__constant_htons(0x000C)
#define IB_SMP_INVALID_FIELD	__constant_htons(0x001C)

static int reply(struct ib_smp *smp)
{
	/*
	 * The verbs framework will handle the directed/LID route
	 * packet changes.
	 */
	smp->method = IB_MGMT_METHOD_GET_RESP;
	if (smp->mgmt_class == IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE)
		smp->status |= IB_SMP_DIRECTION;
	return IB_MAD_RESULT_SUCCESS | IB_MAD_RESULT_REPLY;
}

static int recv_subn_get_nodedescription(struct ib_smp *smp,
					 struct ib_device *ibdev)
{
	if (smp->attr_mod)
		smp->status |= IB_SMP_INVALID_FIELD;

	strncpy(smp->data, ibdev->node_desc, sizeof(smp->data));

	return reply(smp);
}

struct nodeinfo {
	u8 base_version;
	u8 class_version;
	u8 node_type;
	u8 num_ports;
	__be64 sys_guid;
	__be64 node_guid;
	__be64 port_guid;
	__be16 partition_cap;
	__be16 device_id;
	__be32 revision;
	u8 local_port_num;
	u8 vendor_id[3];
} __attribute__ ((packed));

static int recv_subn_get_nodeinfo(struct ib_smp *smp,
				  struct ib_device *ibdev, u8 port)
{
	struct nodeinfo *nip = (struct nodeinfo *)&smp->data;
	struct ipath_devdata *dd = to_idev(ibdev)->dd;
	u32 vendor, majrev, minrev;

	/* GUID 0 is illegal */
	if (smp->attr_mod || (dd->ipath_guid == 0))
		smp->status |= IB_SMP_INVALID_FIELD;

	nip->base_version = 1;
	nip->class_version = 1;
	nip->node_type = 1;	/* channel adapter */
	/*
	 * XXX The num_ports value will need a layer function to get
	 * the value if we ever have more than one IB port on a chip.
	 * We will also need to get the GUID for the port.
	 */
	nip->num_ports = ibdev->phys_port_cnt;
	/* This is already in network order */
	nip->sys_guid = to_idev(ibdev)->sys_image_guid;
	nip->node_guid = dd->ipath_guid;
	nip->port_guid = dd->ipath_guid;
	nip->partition_cap = cpu_to_be16(ipath_get_npkeys(dd));
	nip->device_id = cpu_to_be16(dd->ipath_deviceid);
	majrev = dd->ipath_majrev;
	minrev = dd->ipath_minrev;
	nip->revision = cpu_to_be32((majrev << 16) | minrev);
	nip->local_port_num = port;
	vendor = dd->ipath_vendorid;
	nip->vendor_id[0] = 0;
	nip->vendor_id[1] = vendor >> 8;
	nip->vendor_id[2] = vendor;

	return reply(smp);
}

static int recv_subn_get_guidinfo(struct ib_smp *smp,
				  struct ib_device *ibdev)
{
	u32 startgx = 8 * be32_to_cpu(smp->attr_mod);
	__be64 *p = (__be64 *) smp->data;

	/* 32 blocks of 8 64-bit GUIDs per block */

	memset(smp->data, 0, sizeof(smp->data));

	/*
	 * We only support one GUID for now.  If this changes, the
	 * portinfo.guid_cap field needs to be updated too.
	 */
	if (startgx == 0) {
		__be64 g = to_idev(ibdev)->dd->ipath_guid;
		if (g == 0)
			/* GUID 0 is illegal */
			smp->status |= IB_SMP_INVALID_FIELD;
		else
			/* The first is a copy of the read-only HW GUID. */
			*p = g;
	} else
		smp->status |= IB_SMP_INVALID_FIELD;

	return reply(smp);
}


static int get_overrunthreshold(struct ipath_devdata *dd)
{
	return (dd->ipath_ibcctrl >>
		INFINIPATH_IBCC_OVERRUNTHRESHOLD_SHIFT) &
		INFINIPATH_IBCC_OVERRUNTHRESHOLD_MASK;
}

/**
 * set_overrunthreshold - set the overrun threshold
 * @dd: the infinipath device
 * @n: the new threshold
 *
 * Note that this will only take effect when the link state changes.
 */
static int set_overrunthreshold(struct ipath_devdata *dd, unsigned n)
{
	unsigned v;

	v = (dd->ipath_ibcctrl >> INFINIPATH_IBCC_OVERRUNTHRESHOLD_SHIFT) &
		INFINIPATH_IBCC_OVERRUNTHRESHOLD_MASK;
	if (v != n) {
		dd->ipath_ibcctrl &=
			~(INFINIPATH_IBCC_OVERRUNTHRESHOLD_MASK <<
			  INFINIPATH_IBCC_OVERRUNTHRESHOLD_SHIFT);
		dd->ipath_ibcctrl |=
			(u64) n << INFINIPATH_IBCC_OVERRUNTHRESHOLD_SHIFT;
		ipath_write_kreg(dd, dd->ipath_kregs->kr_ibcctrl,
				 dd->ipath_ibcctrl);
	}
	return 0;
}

static int get_phyerrthreshold(struct ipath_devdata *dd)
{
	return (dd->ipath_ibcctrl >>
		INFINIPATH_IBCC_PHYERRTHRESHOLD_SHIFT) &
		INFINIPATH_IBCC_PHYERRTHRESHOLD_MASK;
}

/**
 * set_phyerrthreshold - set the physical error threshold
 * @dd: the infinipath device
 * @n: the new threshold
 *
 * Note that this will only take effect when the link state changes.
 */
static int set_phyerrthreshold(struct ipath_devdata *dd, unsigned n)
{
	unsigned v;

	v = (dd->ipath_ibcctrl >> INFINIPATH_IBCC_PHYERRTHRESHOLD_SHIFT) &
		INFINIPATH_IBCC_PHYERRTHRESHOLD_MASK;
	if (v != n) {
		dd->ipath_ibcctrl &=
			~(INFINIPATH_IBCC_PHYERRTHRESHOLD_MASK <<
			  INFINIPATH_IBCC_PHYERRTHRESHOLD_SHIFT);
		dd->ipath_ibcctrl |=
			(u64) n << INFINIPATH_IBCC_PHYERRTHRESHOLD_SHIFT;
		ipath_write_kreg(dd, dd->ipath_kregs->kr_ibcctrl,
				 dd->ipath_ibcctrl);
	}
	return 0;
}

/**
 * get_linkdowndefaultstate - get the default linkdown state
 * @dd: the infinipath device
 *
 * Returns zero if the default is POLL, 1 if the default is SLEEP.
 */
static int get_linkdowndefaultstate(struct ipath_devdata *dd)
{
	return !!(dd->ipath_ibcctrl & INFINIPATH_IBCC_LINKDOWNDEFAULTSTATE);
}

static int recv_subn_get_portinfo(struct ib_smp *smp,
				  struct ib_device *ibdev, u8 port)
{
	struct ipath_ibdev *dev;
	struct ib_port_info *pip = (struct ib_port_info *)smp->data;
	u16 lid;
	u8 ibcstat;
	u8 mtu;
	int ret;

	if (be32_to_cpu(smp->attr_mod) > ibdev->phys_port_cnt) {
		smp->status |= IB_SMP_INVALID_FIELD;
		ret = reply(smp);
		goto bail;
	}

	dev = to_idev(ibdev);

	/* Clear all fields.  Only set the non-zero fields. */
	memset(smp->data, 0, sizeof(smp->data));

	/* Only return the mkey if the protection field allows it. */
	if (smp->method == IB_MGMT_METHOD_SET || dev->mkey == smp->mkey ||
	    dev->mkeyprot == 0)
		pip->mkey = dev->mkey;
	pip->gid_prefix = dev->gid_prefix;
	lid = dev->dd->ipath_lid;
	pip->lid = lid ? cpu_to_be16(lid) : IB_LID_PERMISSIVE;
	pip->sm_lid = cpu_to_be16(dev->sm_lid);
	pip->cap_mask = cpu_to_be32(dev->port_cap_flags);
	/* pip->diag_code; */
	pip->mkey_lease_period = cpu_to_be16(dev->mkey_lease_period);
	pip->local_port_num = port;
	pip->link_width_enabled = dev->link_width_enabled;
	pip->link_width_supported = 3;	/* 1x or 4x */
	pip->link_width_active = 2;	/* 4x */
	pip->linkspeed_portstate = 0x10;	/* 2.5Gbps */
	ibcstat = dev->dd->ipath_lastibcstat;
	pip->linkspeed_portstate |= ((ibcstat >> 4) & 0x3) + 1;
	pip->portphysstate_linkdown =
		(ipath_cvt_physportstate[ibcstat & 0xf] << 4) |
		(get_linkdowndefaultstate(dev->dd) ? 1 : 2);
	pip->mkeyprot_resv_lmc = (dev->mkeyprot << 6) | dev->dd->ipath_lmc;
	pip->linkspeedactive_enabled = 0x11;	/* 2.5Gbps, 2.5Gbps */
	switch (dev->dd->ipath_ibmtu) {
	case 4096:
		mtu = IB_MTU_4096;
		break;
	case 2048:
		mtu = IB_MTU_2048;
		break;
	case 1024:
		mtu = IB_MTU_1024;
		break;
	case 512:
		mtu = IB_MTU_512;
		break;
	case 256:
		mtu = IB_MTU_256;
		break;
	default:		/* oops, something is wrong */
		mtu = IB_MTU_2048;
		break;
	}
	pip->neighbormtu_mastersmsl = (mtu << 4) | dev->sm_sl;
	pip->vlcap_inittype = 0x10;	/* VLCap = VL0, InitType = 0 */
	pip->vl_high_limit = dev->vl_high_limit;
	/* pip->vl_arb_high_cap; // only one VL */
	/* pip->vl_arb_low_cap; // only one VL */
	/* InitTypeReply = 0 */
	/*
	 * Note: the chips support a maximum MTU of 4096, but the driver
	 * hasn't implemented this feature yet, so set the maximum value
	 * to 2048.
	 */
	pip->inittypereply_mtucap = IB_MTU_2048;
	// HCAs ignore VLStallCount and HOQLife
	/* pip->vlstallcnt_hoqlife; */
	pip->operationalvl_pei_peo_fpi_fpo = 0x10;	/* OVLs = 1 */
	pip->mkey_violations = cpu_to_be16(dev->mkey_violations);
	/* P_KeyViolations are counted by hardware. */
	pip->pkey_violations =
		cpu_to_be16((ipath_get_cr_errpkey(dev->dd) -
			     dev->z_pkey_violations) & 0xFFFF);
	pip->qkey_violations = cpu_to_be16(dev->qkey_violations);
	/* Only the hardware GUID is supported for now */
	pip->guid_cap = 1;
	pip->clientrereg_resv_subnetto = dev->subnet_timeout;
	/* 32.768 usec. response time (guessing) */
	pip->resv_resptimevalue = 3;
	pip->localphyerrors_overrunerrors =
		(get_phyerrthreshold(dev->dd) << 4) |
		get_overrunthreshold(dev->dd);
	/* pip->max_credit_hint; */
	/* pip->link_roundtrip_latency[3]; */

	ret = reply(smp);

bail:
	return ret;
}

/**
 * get_pkeys - return the PKEY table for port 0
 * @dd: the infinipath device
 * @pkeys: the pkey table is placed here
 */
static int get_pkeys(struct ipath_devdata *dd, u16 * pkeys)
{
	struct ipath_portdata *pd = dd->ipath_pd[0];

	memcpy(pkeys, pd->port_pkeys, sizeof(pd->port_pkeys));

	return 0;
}

static int recv_subn_get_pkeytable(struct ib_smp *smp,
				   struct ib_device *ibdev)
{
	u32 startpx = 32 * (be32_to_cpu(smp->attr_mod) & 0xffff);
	u16 *p = (u16 *) smp->data;
	__be16 *q = (__be16 *) smp->data;

	/* 64 blocks of 32 16-bit P_Key entries */

	memset(smp->data, 0, sizeof(smp->data));
	if (startpx == 0) {
		struct ipath_ibdev *dev = to_idev(ibdev);
		unsigned i, n = ipath_get_npkeys(dev->dd);

		get_pkeys(dev->dd, p);

		for (i = 0; i < n; i++)
			q[i] = cpu_to_be16(p[i]);
	} else
		smp->status |= IB_SMP_INVALID_FIELD;

	return reply(smp);
}

static int recv_subn_set_guidinfo(struct ib_smp *smp,
				  struct ib_device *ibdev)
{
	/* The only GUID we support is the first read-only entry. */
	return recv_subn_get_guidinfo(smp, ibdev);
}

/**
 * set_linkdowndefaultstate - set the default linkdown state
 * @dd: the infinipath device
 * @sleep: the new state
 *
 * Note that this will only take effect when the link state changes.
 */
static int set_linkdowndefaultstate(struct ipath_devdata *dd, int sleep)
{
	if (sleep)
		dd->ipath_ibcctrl |= INFINIPATH_IBCC_LINKDOWNDEFAULTSTATE;
	else
		dd->ipath_ibcctrl &= ~INFINIPATH_IBCC_LINKDOWNDEFAULTSTATE;
	ipath_write_kreg(dd, dd->ipath_kregs->kr_ibcctrl,
			 dd->ipath_ibcctrl);
	return 0;
}

/**
 * recv_subn_set_portinfo - set port information
 * @smp: the incoming SM packet
 * @ibdev: the infiniband device
 * @port: the port on the device
 *
 * Set Portinfo (see ch. 14.2.5.6).
 */
static int recv_subn_set_portinfo(struct ib_smp *smp,
				  struct ib_device *ibdev, u8 port)
{
	struct ib_port_info *pip = (struct ib_port_info *)smp->data;
	struct ib_event event;
	struct ipath_ibdev *dev;
	struct ipath_devdata *dd;
	char clientrereg = 0;
	u16 lid, smlid;
	u8 lwe;
	u8 lse;
	u8 state;
	u16 lstate;
	u32 mtu;
	int ret, ore;

	if (be32_to_cpu(smp->attr_mod) > ibdev->phys_port_cnt)
		goto err;

	dev = to_idev(ibdev);
	dd = dev->dd;
	event.device = ibdev;
	event.element.port_num = port;

	dev->mkey = pip->mkey;
	dev->gid_prefix = pip->gid_prefix;
	dev->mkey_lease_period = be16_to_cpu(pip->mkey_lease_period);

	lid = be16_to_cpu(pip->lid);
	if (dd->ipath_lid != lid ||
	    dd->ipath_lmc != (pip->mkeyprot_resv_lmc & 7)) {
		/* Must be a valid unicast LID address. */
		if (lid == 0 || lid >= IPATH_MULTICAST_LID_BASE)
			goto err;
		ipath_set_lid(dd, lid, pip->mkeyprot_resv_lmc & 7);
		event.event = IB_EVENT_LID_CHANGE;
		ib_dispatch_event(&event);
	}

	smlid = be16_to_cpu(pip->sm_lid);
	if (smlid != dev->sm_lid) {
		/* Must be a valid unicast LID address. */
		if (smlid == 0 || smlid >= IPATH_MULTICAST_LID_BASE)
			goto err;
		dev->sm_lid = smlid;
		event.event = IB_EVENT_SM_CHANGE;
		ib_dispatch_event(&event);
	}

	/* Only 4x supported but allow 1x or 4x to be set (see 14.2.6.6). */
	lwe = pip->link_width_enabled;
	if ((lwe >= 4 && lwe <= 8) || (lwe >= 0xC && lwe <= 0xFE))
		goto err;
	if (lwe == 0xFF)
		dev->link_width_enabled = 3;	/* 1x or 4x */
	else if (lwe)
		dev->link_width_enabled = lwe;

	/* Only 2.5 Gbs supported. */
	lse = pip->linkspeedactive_enabled & 0xF;
	if (lse >= 2 && lse <= 0xE)
		goto err;

	/* Set link down default state. */
	switch (pip->portphysstate_linkdown & 0xF) {
	case 0: /* NOP */
		break;
	case 1: /* SLEEP */
		if (set_linkdowndefaultstate(dd, 1))
			goto err;
		break;
	case 2: /* POLL */
		if (set_linkdowndefaultstate(dd, 0))
			goto err;
		break;
	default:
		goto err;
	}

	dev->mkeyprot = pip->mkeyprot_resv_lmc >> 6;
	dev->vl_high_limit = pip->vl_high_limit;

	switch ((pip->neighbormtu_mastersmsl >> 4) & 0xF) {
	case IB_MTU_256:
		mtu = 256;
		break;
	case IB_MTU_512:
		mtu = 512;
		break;
	case IB_MTU_1024:
		mtu = 1024;
		break;
	case IB_MTU_2048:
		mtu = 2048;
		break;
	case IB_MTU_4096:
		mtu = 4096;
		break;
	default:
		/* XXX We have already partially updated our state! */
		goto err;
	}
	ipath_set_mtu(dd, mtu);

	dev->sm_sl = pip->neighbormtu_mastersmsl & 0xF;

	/* We only support VL0 */
	if (((pip->operationalvl_pei_peo_fpi_fpo >> 4) & 0xF) > 1)
		goto err;

	if (pip->mkey_violations == 0)
		dev->mkey_violations = 0;

	/*
	 * Hardware counter can't be reset so snapshot and subtract
	 * later.
	 */
	if (pip->pkey_violations == 0)
		dev->z_pkey_violations = ipath_get_cr_errpkey(dd);

	if (pip->qkey_violations == 0)
		dev->qkey_violations = 0;

	ore = pip->localphyerrors_overrunerrors;
	if (set_phyerrthreshold(dd, (ore >> 4) & 0xF))
		goto err;

	if (set_overrunthreshold(dd, (ore & 0xF)))
		goto err;

	dev->subnet_timeout = pip->clientrereg_resv_subnetto & 0x1F;

	if (pip->clientrereg_resv_subnetto & 0x80) {
		clientrereg = 1;
		event.event = IB_EVENT_CLIENT_REREGISTER;
		ib_dispatch_event(&event);
	}

	/*
	 * Do the port state change now that the other link parameters
	 * have been set.
	 * Changing the port physical state only makes sense if the link
	 * is down or is being set to down.
	 */
	state = pip->linkspeed_portstate & 0xF;
	lstate = (pip->portphysstate_linkdown >> 4) & 0xF;
	if (lstate && !(state == IB_PORT_DOWN || state == IB_PORT_NOP))
		goto err;

	/*
	 * Only state changes of DOWN, ARM, and ACTIVE are valid
	 * and must be in the correct state to take effect (see 7.2.6).
	 */
	switch (state) {
	case IB_PORT_NOP:
		if (lstate == 0)
			break;
		/* FALLTHROUGH */
	case IB_PORT_DOWN:
		if (lstate == 0)
			lstate = IPATH_IB_LINKDOWN_ONLY;
		else if (lstate == 1)
			lstate = IPATH_IB_LINKDOWN_SLEEP;
		else if (lstate == 2)
			lstate = IPATH_IB_LINKDOWN;
		else if (lstate == 3)
			lstate = IPATH_IB_LINKDOWN_DISABLE;
		else
			goto err;
		ipath_set_linkstate(dd, lstate);
		ipath_wait_linkstate(dd, IPATH_LINKINIT | IPATH_LINKARMED |
				IPATH_LINKACTIVE, 1000);
		break;
	case IB_PORT_ARMED:
		ipath_set_linkstate(dd, IPATH_IB_LINKARM);
		break;
	case IB_PORT_ACTIVE:
		ipath_set_linkstate(dd, IPATH_IB_LINKACTIVE);
		break;
	default:
		/* XXX We have already partially updated our state! */
		goto err;
	}

	ret = recv_subn_get_portinfo(smp, ibdev, port);

	if (clientrereg)
		pip->clientrereg_resv_subnetto |= 0x80;

	goto done;

err:
	smp->status |= IB_SMP_INVALID_FIELD;
	ret = recv_subn_get_portinfo(smp, ibdev, port);

done:
	return ret;
}

/**
 * rm_pkey - decrecment the reference count for the given PKEY
 * @dd: the infinipath device
 * @key: the PKEY index
 *
 * Return true if this was the last reference and the hardware table entry
 * needs to be changed.
 */
static int rm_pkey(struct ipath_devdata *dd, u16 key)
{
	int i;
	int ret;

	for (i = 0; i < ARRAY_SIZE(dd->ipath_pkeys); i++) {
		if (dd->ipath_pkeys[i] != key)
			continue;
		if (atomic_dec_and_test(&dd->ipath_pkeyrefs[i])) {
			dd->ipath_pkeys[i] = 0;
			ret = 1;
			goto bail;
		}
		break;
	}

	ret = 0;

bail:
	return ret;
}

/**
 * add_pkey - add the given PKEY to the hardware table
 * @dd: the infinipath device
 * @key: the PKEY
 *
 * Return an error code if unable to add the entry, zero if no change,
 * or 1 if the hardware PKEY register needs to be updated.
 */
static int add_pkey(struct ipath_devdata *dd, u16 key)
{
	int i;
	u16 lkey = key & 0x7FFF;
	int any = 0;
	int ret;

	if (lkey == 0x7FFF) {
		ret = 0;
		goto bail;
	}

	/* Look for an empty slot or a matching PKEY. */
	for (i = 0; i < ARRAY_SIZE(dd->ipath_pkeys); i++) {
		if (!dd->ipath_pkeys[i]) {
			any++;
			continue;
		}
		/* If it matches exactly, try to increment the ref count */
		if (dd->ipath_pkeys[i] == key) {
			if (atomic_inc_return(&dd->ipath_pkeyrefs[i]) > 1) {
				ret = 0;
				goto bail;
			}
			/* Lost the race. Look for an empty slot below. */
			atomic_dec(&dd->ipath_pkeyrefs[i]);
			any++;
		}
		/*
		 * It makes no sense to have both the limited and unlimited
		 * PKEY set at the same time since the unlimited one will
		 * disable the limited one.
		 */
		if ((dd->ipath_pkeys[i] & 0x7FFF) == lkey) {
			ret = -EEXIST;
			goto bail;
		}
	}
	if (!any) {
		ret = -EBUSY;
		goto bail;
	}
	for (i = 0; i < ARRAY_SIZE(dd->ipath_pkeys); i++) {
		if (!dd->ipath_pkeys[i] &&
		    atomic_inc_return(&dd->ipath_pkeyrefs[i]) == 1) {
			/* for ipathstats, etc. */
			ipath_stats.sps_pkeys[i] = lkey;
			dd->ipath_pkeys[i] = key;
			ret = 1;
			goto bail;
		}
	}
	ret = -EBUSY;

bail:
	return ret;
}

/**
 * set_pkeys - set the PKEY table for port 0
 * @dd: the infinipath device
 * @pkeys: the PKEY table
 */
static int set_pkeys(struct ipath_devdata *dd, u16 *pkeys)
{
	struct ipath_portdata *pd;
	int i;
	int changed = 0;

	pd = dd->ipath_pd[0];

	for (i = 0; i < ARRAY_SIZE(pd->port_pkeys); i++) {
		u16 key = pkeys[i];
		u16 okey = pd->port_pkeys[i];

		if (key == okey)
			continue;
		/*
		 * The value of this PKEY table entry is changing.
		 * Remove the old entry in the hardware's array of PKEYs.
		 */
		if (okey & 0x7FFF)
			changed |= rm_pkey(dd, okey);
		if (key & 0x7FFF) {
			int ret = add_pkey(dd, key);

			if (ret < 0)
				key = 0;
			else
				changed |= ret;
		}
		pd->port_pkeys[i] = key;
	}
	if (changed) {
		u64 pkey;

		pkey = (u64) dd->ipath_pkeys[0] |
			((u64) dd->ipath_pkeys[1] << 16) |
			((u64) dd->ipath_pkeys[2] << 32) |
			((u64) dd->ipath_pkeys[3] << 48);
		ipath_cdbg(VERBOSE, "p0 new pkey reg %llx\n",
			   (unsigned long long) pkey);
		ipath_write_kreg(dd, dd->ipath_kregs->kr_partitionkey,
				 pkey);
	}
	return 0;
}

static int recv_subn_set_pkeytable(struct ib_smp *smp,
				   struct ib_device *ibdev)
{
	u32 startpx = 32 * (be32_to_cpu(smp->attr_mod) & 0xffff);
	__be16 *p = (__be16 *) smp->data;
	u16 *q = (u16 *) smp->data;
	struct ipath_ibdev *dev = to_idev(ibdev);
	unsigned i, n = ipath_get_npkeys(dev->dd);

	for (i = 0; i < n; i++)
		q[i] = be16_to_cpu(p[i]);

	if (startpx != 0 || set_pkeys(dev->dd, q) != 0)
		smp->status |= IB_SMP_INVALID_FIELD;

	return recv_subn_get_pkeytable(smp, ibdev);
}

#define IB_PMA_CLASS_PORT_INFO		__constant_htons(0x0001)
#define IB_PMA_PORT_SAMPLES_CONTROL	__constant_htons(0x0010)
#define IB_PMA_PORT_SAMPLES_RESULT	__constant_htons(0x0011)
#define IB_PMA_PORT_COUNTERS		__constant_htons(0x0012)
#define IB_PMA_PORT_COUNTERS_EXT	__constant_htons(0x001D)
#define IB_PMA_PORT_SAMPLES_RESULT_EXT	__constant_htons(0x001E)

struct ib_perf {
	u8 base_version;
	u8 mgmt_class;
	u8 class_version;
	u8 method;
	__be16 status;
	__be16 unused;
	__be64 tid;
	__be16 attr_id;
	__be16 resv;
	__be32 attr_mod;
	u8 reserved[40];
	u8 data[192];
} __attribute__ ((packed));

struct ib_pma_classportinfo {
	u8 base_version;
	u8 class_version;
	__be16 cap_mask;
	u8 reserved[3];
	u8 resp_time_value;	/* only lower 5 bits */
	union ib_gid redirect_gid;
	__be32 redirect_tc_sl_fl;	/* 8, 4, 20 bits respectively */
	__be16 redirect_lid;
	__be16 redirect_pkey;
	__be32 redirect_qp;	/* only lower 24 bits */
	__be32 redirect_qkey;
	union ib_gid trap_gid;
	__be32 trap_tc_sl_fl;	/* 8, 4, 20 bits respectively */
	__be16 trap_lid;
	__be16 trap_pkey;
	__be32 trap_hl_qp;	/* 8, 24 bits respectively */
	__be32 trap_qkey;
} __attribute__ ((packed));

struct ib_pma_portsamplescontrol {
	u8 opcode;
	u8 port_select;
	u8 tick;
	u8 counter_width;	/* only lower 3 bits */
	__be32 counter_mask0_9;	/* 2, 10 * 3, bits */
	__be16 counter_mask10_14;	/* 1, 5 * 3, bits */
	u8 sample_mechanisms;
	u8 sample_status;	/* only lower 2 bits */
	__be64 option_mask;
	__be64 vendor_mask;
	__be32 sample_start;
	__be32 sample_interval;
	__be16 tag;
	__be16 counter_select[15];
} __attribute__ ((packed));

struct ib_pma_portsamplesresult {
	__be16 tag;
	__be16 sample_status;	/* only lower 2 bits */
	__be32 counter[15];
} __attribute__ ((packed));

struct ib_pma_portsamplesresult_ext {
	__be16 tag;
	__be16 sample_status;	/* only lower 2 bits */
	__be32 extended_width;	/* only upper 2 bits */
	__be64 counter[15];
} __attribute__ ((packed));

struct ib_pma_portcounters {
	u8 reserved;
	u8 port_select;
	__be16 counter_select;
	__be16 symbol_error_counter;
	u8 link_error_recovery_counter;
	u8 link_downed_counter;
	__be16 port_rcv_errors;
	__be16 port_rcv_remphys_errors;
	__be16 port_rcv_switch_relay_errors;
	__be16 port_xmit_discards;
	u8 port_xmit_constraint_errors;
	u8 port_rcv_constraint_errors;
	u8 reserved1;
	u8 lli_ebor_errors;	/* 4, 4, bits */
	__be16 reserved2;
	__be16 vl15_dropped;
	__be32 port_xmit_data;
	__be32 port_rcv_data;
	__be32 port_xmit_packets;
	__be32 port_rcv_packets;
} __attribute__ ((packed));

#define IB_PMA_SEL_SYMBOL_ERROR			__constant_htons(0x0001)
#define IB_PMA_SEL_LINK_ERROR_RECOVERY		__constant_htons(0x0002)
#define IB_PMA_SEL_LINK_DOWNED			__constant_htons(0x0004)
#define IB_PMA_SEL_PORT_RCV_ERRORS		__constant_htons(0x0008)
#define IB_PMA_SEL_PORT_RCV_REMPHYS_ERRORS	__constant_htons(0x0010)
#define IB_PMA_SEL_PORT_XMIT_DISCARDS		__constant_htons(0x0040)
#define IB_PMA_SEL_LOCAL_LINK_INTEGRITY_ERRORS	__constant_htons(0x0200)
#define IB_PMA_SEL_EXCESSIVE_BUFFER_OVERRUNS	__constant_htons(0x0400)
#define IB_PMA_SEL_PORT_VL15_DROPPED		__constant_htons(0x0800)
#define IB_PMA_SEL_PORT_XMIT_DATA		__constant_htons(0x1000)
#define IB_PMA_SEL_PORT_RCV_DATA		__constant_htons(0x2000)
#define IB_PMA_SEL_PORT_XMIT_PACKETS		__constant_htons(0x4000)
#define IB_PMA_SEL_PORT_RCV_PACKETS		__constant_htons(0x8000)

struct ib_pma_portcounters_ext {
	u8 reserved;
	u8 port_select;
	__be16 counter_select;
	__be32 reserved1;
	__be64 port_xmit_data;
	__be64 port_rcv_data;
	__be64 port_xmit_packets;
	__be64 port_rcv_packets;
	__be64 port_unicast_xmit_packets;
	__be64 port_unicast_rcv_packets;
	__be64 port_multicast_xmit_packets;
	__be64 port_multicast_rcv_packets;
} __attribute__ ((packed));

#define IB_PMA_SELX_PORT_XMIT_DATA		__constant_htons(0x0001)
#define IB_PMA_SELX_PORT_RCV_DATA		__constant_htons(0x0002)
#define IB_PMA_SELX_PORT_XMIT_PACKETS		__constant_htons(0x0004)
#define IB_PMA_SELX_PORT_RCV_PACKETS		__constant_htons(0x0008)
#define IB_PMA_SELX_PORT_UNI_XMIT_PACKETS	__constant_htons(0x0010)
#define IB_PMA_SELX_PORT_UNI_RCV_PACKETS	__constant_htons(0x0020)
#define IB_PMA_SELX_PORT_MULTI_XMIT_PACKETS	__constant_htons(0x0040)
#define IB_PMA_SELX_PORT_MULTI_RCV_PACKETS	__constant_htons(0x0080)

static int recv_pma_get_classportinfo(struct ib_perf *pmp)
{
	struct ib_pma_classportinfo *p =
		(struct ib_pma_classportinfo *)pmp->data;

	memset(pmp->data, 0, sizeof(pmp->data));

	if (pmp->attr_mod != 0)
		pmp->status |= IB_SMP_INVALID_FIELD;

	/* Indicate AllPortSelect is valid (only one port anyway) */
	p->cap_mask = __constant_cpu_to_be16(1 << 8);
	p->base_version = 1;
	p->class_version = 1;
	/*
	 * Expected response time is 4.096 usec. * 2^18 == 1.073741824
	 * sec.
	 */
	p->resp_time_value = 18;

	return reply((struct ib_smp *) pmp);
}

/*
 * The PortSamplesControl.CounterMasks field is an array of 3 bit fields
 * which specify the N'th counter's capabilities. See ch. 16.1.3.2.
 * We support 5 counters which only count the mandatory quantities.
 */
#define COUNTER_MASK(q, n) (q << ((9 - n) * 3))
#define COUNTER_MASK0_9 \
	__constant_cpu_to_be32(COUNTER_MASK(1, 0) | \
			       COUNTER_MASK(1, 1) | \
			       COUNTER_MASK(1, 2) | \
			       COUNTER_MASK(1, 3) | \
			       COUNTER_MASK(1, 4))

static int recv_pma_get_portsamplescontrol(struct ib_perf *pmp,
					   struct ib_device *ibdev, u8 port)
{
	struct ib_pma_portsamplescontrol *p =
		(struct ib_pma_portsamplescontrol *)pmp->data;
	struct ipath_ibdev *dev = to_idev(ibdev);
	struct ipath_cregs const *crp = dev->dd->ipath_cregs;
	unsigned long flags;
	u8 port_select = p->port_select;

	memset(pmp->data, 0, sizeof(pmp->data));

	p->port_select = port_select;
	if (pmp->attr_mod != 0 ||
	    (port_select != port && port_select != 0xFF))
		pmp->status |= IB_SMP_INVALID_FIELD;
	/*
	 * Ticks are 10x the link transfer period which for 2.5Gbs is 4
	 * nsec.  0 == 4 nsec., 1 == 8 nsec., ..., 255 == 1020 nsec.  Sample
	 * intervals are counted in ticks.  Since we use Linux timers, that
	 * count in jiffies, we can't sample for less than 1000 ticks if HZ
	 * == 1000 (4000 ticks if HZ is 250).
	 */
	/* XXX This is WRONG. */
	p->tick = 250;		/* 1 usec. */
	p->counter_width = 4;	/* 32 bit counters */
	p->counter_mask0_9 = COUNTER_MASK0_9;
	spin_lock_irqsave(&dev->pending_lock, flags);
	if (crp->cr_psstat)
		p->sample_status = ipath_read_creg32(dev->dd, crp->cr_psstat);
	else
		p->sample_status = dev->pma_sample_status;
	p->sample_start = cpu_to_be32(dev->pma_sample_start);
	p->sample_interval = cpu_to_be32(dev->pma_sample_interval);
	p->tag = cpu_to_be16(dev->pma_tag);
	p->counter_select[0] = dev->pma_counter_select[0];
	p->counter_select[1] = dev->pma_counter_select[1];
	p->counter_select[2] = dev->pma_counter_select[2];
	p->counter_select[3] = dev->pma_counter_select[3];
	p->counter_select[4] = dev->pma_counter_select[4];
	spin_unlock_irqrestore(&dev->pending_lock, flags);

	return reply((struct ib_smp *) pmp);
}

static int recv_pma_set_portsamplescontrol(struct ib_perf *pmp,
					   struct ib_device *ibdev, u8 port)
{
	struct ib_pma_portsamplescontrol *p =
		(struct ib_pma_portsamplescontrol *)pmp->data;
	struct ipath_ibdev *dev = to_idev(ibdev);
	struct ipath_cregs const *crp = dev->dd->ipath_cregs;
	unsigned long flags;
	u8 status;
	int ret;

	if (pmp->attr_mod != 0 ||
	    (p->port_select != port && p->port_select != 0xFF)) {
		pmp->status |= IB_SMP_INVALID_FIELD;
		ret = reply((struct ib_smp *) pmp);
		goto bail;
	}

	spin_lock_irqsave(&dev->pending_lock, flags);
	if (crp->cr_psstat)
		status = ipath_read_creg32(dev->dd, crp->cr_psstat);
	else
		status = dev->pma_sample_status;
	if (status == IB_PMA_SAMPLE_STATUS_DONE) {
		dev->pma_sample_start = be32_to_cpu(p->sample_start);
		dev->pma_sample_interval = be32_to_cpu(p->sample_interval);
		dev->pma_tag = be16_to_cpu(p->tag);
		dev->pma_counter_select[0] = p->counter_select[0];
		dev->pma_counter_select[1] = p->counter_select[1];
		dev->pma_counter_select[2] = p->counter_select[2];
		dev->pma_counter_select[3] = p->counter_select[3];
		dev->pma_counter_select[4] = p->counter_select[4];
		if (crp->cr_psstat) {
			ipath_write_creg(dev->dd, crp->cr_psinterval,
					 dev->pma_sample_interval);
			ipath_write_creg(dev->dd, crp->cr_psstart,
					 dev->pma_sample_start);
		} else
			dev->pma_sample_status = IB_PMA_SAMPLE_STATUS_STARTED;
	}
	spin_unlock_irqrestore(&dev->pending_lock, flags);

	ret = recv_pma_get_portsamplescontrol(pmp, ibdev, port);

bail:
	return ret;
}

static u64 get_counter(struct ipath_ibdev *dev,
		       struct ipath_cregs const *crp,
		       __be16 sel)
{
	u64 ret;

	switch (sel) {
	case IB_PMA_PORT_XMIT_DATA:
		ret = (crp->cr_psxmitdatacount) ?
			ipath_read_creg32(dev->dd, crp->cr_psxmitdatacount) :
			dev->ipath_sword;
		break;
	case IB_PMA_PORT_RCV_DATA:
		ret = (crp->cr_psrcvdatacount) ?
			ipath_read_creg32(dev->dd, crp->cr_psrcvdatacount) :
			dev->ipath_rword;
		break;
	case IB_PMA_PORT_XMIT_PKTS:
		ret = (crp->cr_psxmitpktscount) ?
			ipath_read_creg32(dev->dd, crp->cr_psxmitpktscount) :
			dev->ipath_spkts;
		break;
	case IB_PMA_PORT_RCV_PKTS:
		ret = (crp->cr_psrcvpktscount) ?
			ipath_read_creg32(dev->dd, crp->cr_psrcvpktscount) :
			dev->ipath_rpkts;
		break;
	case IB_PMA_PORT_XMIT_WAIT:
		ret = (crp->cr_psxmitwaitcount) ?
			ipath_read_creg32(dev->dd, crp->cr_psxmitwaitcount) :
			dev->ipath_xmit_wait;
		break;
	default:
		ret = 0;
	}

	return ret;
}

static int recv_pma_get_portsamplesresult(struct ib_perf *pmp,
					  struct ib_device *ibdev)
{
	struct ib_pma_portsamplesresult *p =
		(struct ib_pma_portsamplesresult *)pmp->data;
	struct ipath_ibdev *dev = to_idev(ibdev);
	struct ipath_cregs const *crp = dev->dd->ipath_cregs;
	u8 status;
	int i;

	memset(pmp->data, 0, sizeof(pmp->data));
	p->tag = cpu_to_be16(dev->pma_tag);
	if (crp->cr_psstat)
		status = ipath_read_creg32(dev->dd, crp->cr_psstat);
	else
		status = dev->pma_sample_status;
	p->sample_status = cpu_to_be16(status);
	for (i = 0; i < ARRAY_SIZE(dev->pma_counter_select); i++)
		p->counter[i] = (status != IB_PMA_SAMPLE_STATUS_DONE) ? 0 :
		    cpu_to_be32(
			get_counter(dev, crp, dev->pma_counter_select[i]));

	return reply((struct ib_smp *) pmp);
}

static int recv_pma_get_portsamplesresult_ext(struct ib_perf *pmp,
					      struct ib_device *ibdev)
{
	struct ib_pma_portsamplesresult_ext *p =
		(struct ib_pma_portsamplesresult_ext *)pmp->data;
	struct ipath_ibdev *dev = to_idev(ibdev);
	struct ipath_cregs const *crp = dev->dd->ipath_cregs;
	u8 status;
	int i;

	memset(pmp->data, 0, sizeof(pmp->data));
	p->tag = cpu_to_be16(dev->pma_tag);
	if (crp->cr_psstat)
		status = ipath_read_creg32(dev->dd, crp->cr_psstat);
	else
		status = dev->pma_sample_status;
	p->sample_status = cpu_to_be16(status);
	/* 64 bits */
	p->extended_width = __constant_cpu_to_be32(0x80000000);
	for (i = 0; i < ARRAY_SIZE(dev->pma_counter_select); i++)
		p->counter[i] = (status != IB_PMA_SAMPLE_STATUS_DONE) ? 0 :
		    cpu_to_be64(
			get_counter(dev, crp, dev->pma_counter_select[i]));

	return reply((struct ib_smp *) pmp);
}

static int recv_pma_get_portcounters(struct ib_perf *pmp,
				     struct ib_device *ibdev, u8 port)
{
	struct ib_pma_portcounters *p = (struct ib_pma_portcounters *)
		pmp->data;
	struct ipath_ibdev *dev = to_idev(ibdev);
	struct ipath_verbs_counters cntrs;
	u8 port_select = p->port_select;

	ipath_get_counters(dev->dd, &cntrs);

	/* Adjust counters for any resets done. */
	cntrs.symbol_error_counter -= dev->z_symbol_error_counter;
	cntrs.link_error_recovery_counter -=
		dev->z_link_error_recovery_counter;
	cntrs.link_downed_counter -= dev->z_link_downed_counter;
	cntrs.port_rcv_errors += dev->rcv_errors;
	cntrs.port_rcv_errors -= dev->z_port_rcv_errors;
	cntrs.port_rcv_remphys_errors -= dev->z_port_rcv_remphys_errors;
	cntrs.port_xmit_discards -= dev->z_port_xmit_discards;
	cntrs.port_xmit_data -= dev->z_port_xmit_data;
	cntrs.port_rcv_data -= dev->z_port_rcv_data;
	cntrs.port_xmit_packets -= dev->z_port_xmit_packets;
	cntrs.port_rcv_packets -= dev->z_port_rcv_packets;
	cntrs.local_link_integrity_errors -=
		dev->z_local_link_integrity_errors;
	cntrs.excessive_buffer_overrun_errors -=
		dev->z_excessive_buffer_overrun_errors;
	cntrs.vl15_dropped -= dev->z_vl15_dropped;
	cntrs.vl15_dropped += dev->n_vl15_dropped;

	memset(pmp->data, 0, sizeof(pmp->data));

	p->port_select = port_select;
	if (pmp->attr_mod != 0 ||
	    (port_select != port && port_select != 0xFF))
		pmp->status |= IB_SMP_INVALID_FIELD;

	if (cntrs.symbol_error_counter > 0xFFFFUL)
		p->symbol_error_counter = __constant_cpu_to_be16(0xFFFF);
	else
		p->symbol_error_counter =
			cpu_to_be16((u16)cntrs.symbol_error_counter);
	if (cntrs.link_error_recovery_counter > 0xFFUL)
		p->link_error_recovery_counter = 0xFF;
	else
		p->link_error_recovery_counter =
			(u8)cntrs.link_error_recovery_counter;
	if (cntrs.link_downed_counter > 0xFFUL)
		p->link_downed_counter = 0xFF;
	else
		p->link_downed_counter = (u8)cntrs.link_downed_counter;
	if (cntrs.port_rcv_errors > 0xFFFFUL)
		p->port_rcv_errors = __constant_cpu_to_be16(0xFFFF);
	else
		p->port_rcv_errors =
			cpu_to_be16((u16) cntrs.port_rcv_errors);
	if (cntrs.port_rcv_remphys_errors > 0xFFFFUL)
		p->port_rcv_remphys_errors = __constant_cpu_to_be16(0xFFFF);
	else
		p->port_rcv_remphys_errors =
			cpu_to_be16((u16)cntrs.port_rcv_remphys_errors);
	if (cntrs.port_xmit_discards > 0xFFFFUL)
		p->port_xmit_discards = __constant_cpu_to_be16(0xFFFF);
	else
		p->port_xmit_discards =
			cpu_to_be16((u16)cntrs.port_xmit_discards);
	if (cntrs.local_link_integrity_errors > 0xFUL)
		cntrs.local_link_integrity_errors = 0xFUL;
	if (cntrs.excessive_buffer_overrun_errors > 0xFUL)
		cntrs.excessive_buffer_overrun_errors = 0xFUL;
	p->lli_ebor_errors = (cntrs.local_link_integrity_errors << 4) |
		cntrs.excessive_buffer_overrun_errors;
	if (cntrs.vl15_dropped > 0xFFFFUL)
		p->vl15_dropped = __constant_cpu_to_be16(0xFFFF);
	else
		p->vl15_dropped = cpu_to_be16((u16)cntrs.vl15_dropped);
	if (cntrs.port_xmit_data > 0xFFFFFFFFUL)
		p->port_xmit_data = __constant_cpu_to_be32(0xFFFFFFFF);
	else
		p->port_xmit_data = cpu_to_be32((u32)cntrs.port_xmit_data);
	if (cntrs.port_rcv_data > 0xFFFFFFFFUL)
		p->port_rcv_data = __constant_cpu_to_be32(0xFFFFFFFF);
	else
		p->port_rcv_data = cpu_to_be32((u32)cntrs.port_rcv_data);
	if (cntrs.port_xmit_packets > 0xFFFFFFFFUL)
		p->port_xmit_packets = __constant_cpu_to_be32(0xFFFFFFFF);
	else
		p->port_xmit_packets =
			cpu_to_be32((u32)cntrs.port_xmit_packets);
	if (cntrs.port_rcv_packets > 0xFFFFFFFFUL)
		p->port_rcv_packets = __constant_cpu_to_be32(0xFFFFFFFF);
	else
		p->port_rcv_packets =
			cpu_to_be32((u32) cntrs.port_rcv_packets);

	return reply((struct ib_smp *) pmp);
}

static int recv_pma_get_portcounters_ext(struct ib_perf *pmp,
					 struct ib_device *ibdev, u8 port)
{
	struct ib_pma_portcounters_ext *p =
		(struct ib_pma_portcounters_ext *)pmp->data;
	struct ipath_ibdev *dev = to_idev(ibdev);
	u64 swords, rwords, spkts, rpkts, xwait;
	u8 port_select = p->port_select;

	ipath_snapshot_counters(dev->dd, &swords, &rwords, &spkts,
				&rpkts, &xwait);

	/* Adjust counters for any resets done. */
	swords -= dev->z_port_xmit_data;
	rwords -= dev->z_port_rcv_data;
	spkts -= dev->z_port_xmit_packets;
	rpkts -= dev->z_port_rcv_packets;

	memset(pmp->data, 0, sizeof(pmp->data));

	p->port_select = port_select;
	if (pmp->attr_mod != 0 ||
	    (port_select != port && port_select != 0xFF))
		pmp->status |= IB_SMP_INVALID_FIELD;

	p->port_xmit_data = cpu_to_be64(swords);
	p->port_rcv_data = cpu_to_be64(rwords);
	p->port_xmit_packets = cpu_to_be64(spkts);
	p->port_rcv_packets = cpu_to_be64(rpkts);
	p->port_unicast_xmit_packets = cpu_to_be64(dev->n_unicast_xmit);
	p->port_unicast_rcv_packets = cpu_to_be64(dev->n_unicast_rcv);
	p->port_multicast_xmit_packets = cpu_to_be64(dev->n_multicast_xmit);
	p->port_multicast_rcv_packets = cpu_to_be64(dev->n_multicast_rcv);

	return reply((struct ib_smp *) pmp);
}

static int recv_pma_set_portcounters(struct ib_perf *pmp,
				     struct ib_device *ibdev, u8 port)
{
	struct ib_pma_portcounters *p = (struct ib_pma_portcounters *)
		pmp->data;
	struct ipath_ibdev *dev = to_idev(ibdev);
	struct ipath_verbs_counters cntrs;

	/*
	 * Since the HW doesn't support clearing counters, we save the
	 * current count and subtract it from future responses.
	 */
	ipath_get_counters(dev->dd, &cntrs);

	if (p->counter_select & IB_PMA_SEL_SYMBOL_ERROR)
		dev->z_symbol_error_counter = cntrs.symbol_error_counter;

	if (p->counter_select & IB_PMA_SEL_LINK_ERROR_RECOVERY)
		dev->z_link_error_recovery_counter =
			cntrs.link_error_recovery_counter;

	if (p->counter_select & IB_PMA_SEL_LINK_DOWNED)
		dev->z_link_downed_counter = cntrs.link_downed_counter;

	if (p->counter_select & IB_PMA_SEL_PORT_RCV_ERRORS)
		dev->z_port_rcv_errors =
			cntrs.port_rcv_errors + dev->rcv_errors;

	if (p->counter_select & IB_PMA_SEL_PORT_RCV_REMPHYS_ERRORS)
		dev->z_port_rcv_remphys_errors =
			cntrs.port_rcv_remphys_errors;

	if (p->counter_select & IB_PMA_SEL_PORT_XMIT_DISCARDS)
		dev->z_port_xmit_discards = cntrs.port_xmit_discards;

	if (p->counter_select & IB_PMA_SEL_LOCAL_LINK_INTEGRITY_ERRORS)
		dev->z_local_link_integrity_errors =
			cntrs.local_link_integrity_errors;

	if (p->counter_select & IB_PMA_SEL_EXCESSIVE_BUFFER_OVERRUNS)
		dev->z_excessive_buffer_overrun_errors =
			cntrs.excessive_buffer_overrun_errors;

	if (p->counter_select & IB_PMA_SEL_PORT_VL15_DROPPED) {
		dev->n_vl15_dropped = 0;
		dev->z_vl15_dropped = cntrs.vl15_dropped;
	}

	if (p->counter_select & IB_PMA_SEL_PORT_XMIT_DATA)
		dev->z_port_xmit_data = cntrs.port_xmit_data;

	if (p->counter_select & IB_PMA_SEL_PORT_RCV_DATA)
		dev->z_port_rcv_data = cntrs.port_rcv_data;

	if (p->counter_select & IB_PMA_SEL_PORT_XMIT_PACKETS)
		dev->z_port_xmit_packets = cntrs.port_xmit_packets;

	if (p->counter_select & IB_PMA_SEL_PORT_RCV_PACKETS)
		dev->z_port_rcv_packets = cntrs.port_rcv_packets;

	return recv_pma_get_portcounters(pmp, ibdev, port);
}

static int recv_pma_set_portcounters_ext(struct ib_perf *pmp,
					 struct ib_device *ibdev, u8 port)
{
	struct ib_pma_portcounters *p = (struct ib_pma_portcounters *)
		pmp->data;
	struct ipath_ibdev *dev = to_idev(ibdev);
	u64 swords, rwords, spkts, rpkts, xwait;

	ipath_snapshot_counters(dev->dd, &swords, &rwords, &spkts,
				&rpkts, &xwait);

	if (p->counter_select & IB_PMA_SELX_PORT_XMIT_DATA)
		dev->z_port_xmit_data = swords;

	if (p->counter_select & IB_PMA_SELX_PORT_RCV_DATA)
		dev->z_port_rcv_data = rwords;

	if (p->counter_select & IB_PMA_SELX_PORT_XMIT_PACKETS)
		dev->z_port_xmit_packets = spkts;

	if (p->counter_select & IB_PMA_SELX_PORT_RCV_PACKETS)
		dev->z_port_rcv_packets = rpkts;

	if (p->counter_select & IB_PMA_SELX_PORT_UNI_XMIT_PACKETS)
		dev->n_unicast_xmit = 0;

	if (p->counter_select & IB_PMA_SELX_PORT_UNI_RCV_PACKETS)
		dev->n_unicast_rcv = 0;

	if (p->counter_select & IB_PMA_SELX_PORT_MULTI_XMIT_PACKETS)
		dev->n_multicast_xmit = 0;

	if (p->counter_select & IB_PMA_SELX_PORT_MULTI_RCV_PACKETS)
		dev->n_multicast_rcv = 0;

	return recv_pma_get_portcounters_ext(pmp, ibdev, port);
}

static int process_subn(struct ib_device *ibdev, int mad_flags,
			u8 port_num, struct ib_mad *in_mad,
			struct ib_mad *out_mad)
{
	struct ib_smp *smp = (struct ib_smp *)out_mad;
	struct ipath_ibdev *dev = to_idev(ibdev);
	int ret;

	*out_mad = *in_mad;
	if (smp->class_version != 1) {
		smp->status |= IB_SMP_UNSUP_VERSION;
		ret = reply(smp);
		goto bail;
	}

	/* Is the mkey in the process of expiring? */
	if (dev->mkey_lease_timeout && jiffies >= dev->mkey_lease_timeout) {
		/* Clear timeout and mkey protection field. */
		dev->mkey_lease_timeout = 0;
		dev->mkeyprot = 0;
	}

	/*
	 * M_Key checking depends on
	 * Portinfo:M_Key_protect_bits
	 */
	if ((mad_flags & IB_MAD_IGNORE_MKEY) == 0 && dev->mkey != 0 &&
	    dev->mkey != smp->mkey &&
	    (smp->method == IB_MGMT_METHOD_SET ||
	     (smp->method == IB_MGMT_METHOD_GET &&
	      dev->mkeyprot >= 2))) {
		if (dev->mkey_violations != 0xFFFF)
			++dev->mkey_violations;
		if (dev->mkey_lease_timeout ||
		    dev->mkey_lease_period == 0) {
			ret = IB_MAD_RESULT_SUCCESS |
				IB_MAD_RESULT_CONSUMED;
			goto bail;
		}
		dev->mkey_lease_timeout = jiffies +
			dev->mkey_lease_period * HZ;
		/* Future: Generate a trap notice. */
		ret = IB_MAD_RESULT_SUCCESS | IB_MAD_RESULT_CONSUMED;
		goto bail;
	} else if (dev->mkey_lease_timeout)
		dev->mkey_lease_timeout = 0;

	switch (smp->method) {
	case IB_MGMT_METHOD_GET:
		switch (smp->attr_id) {
		case IB_SMP_ATTR_NODE_DESC:
			ret = recv_subn_get_nodedescription(smp, ibdev);
			goto bail;
		case IB_SMP_ATTR_NODE_INFO:
			ret = recv_subn_get_nodeinfo(smp, ibdev, port_num);
			goto bail;
		case IB_SMP_ATTR_GUID_INFO:
			ret = recv_subn_get_guidinfo(smp, ibdev);
			goto bail;
		case IB_SMP_ATTR_PORT_INFO:
			ret = recv_subn_get_portinfo(smp, ibdev, port_num);
			goto bail;
		case IB_SMP_ATTR_PKEY_TABLE:
			ret = recv_subn_get_pkeytable(smp, ibdev);
			goto bail;
		case IB_SMP_ATTR_SM_INFO:
			if (dev->port_cap_flags & IB_PORT_SM_DISABLED) {
				ret = IB_MAD_RESULT_SUCCESS |
					IB_MAD_RESULT_CONSUMED;
				goto bail;
			}
			if (dev->port_cap_flags & IB_PORT_SM) {
				ret = IB_MAD_RESULT_SUCCESS;
				goto bail;
			}
			/* FALLTHROUGH */
		default:
			smp->status |= IB_SMP_UNSUP_METH_ATTR;
			ret = reply(smp);
			goto bail;
		}

	case IB_MGMT_METHOD_SET:
		switch (smp->attr_id) {
		case IB_SMP_ATTR_GUID_INFO:
			ret = recv_subn_set_guidinfo(smp, ibdev);
			goto bail;
		case IB_SMP_ATTR_PORT_INFO:
			ret = recv_subn_set_portinfo(smp, ibdev, port_num);
			goto bail;
		case IB_SMP_ATTR_PKEY_TABLE:
			ret = recv_subn_set_pkeytable(smp, ibdev);
			goto bail;
		case IB_SMP_ATTR_SM_INFO:
			if (dev->port_cap_flags & IB_PORT_SM_DISABLED) {
				ret = IB_MAD_RESULT_SUCCESS |
					IB_MAD_RESULT_CONSUMED;
				goto bail;
			}
			if (dev->port_cap_flags & IB_PORT_SM) {
				ret = IB_MAD_RESULT_SUCCESS;
				goto bail;
			}
			/* FALLTHROUGH */
		default:
			smp->status |= IB_SMP_UNSUP_METH_ATTR;
			ret = reply(smp);
			goto bail;
		}

	case IB_MGMT_METHOD_GET_RESP:
		/*
		 * The ib_mad module will call us to process responses
		 * before checking for other consumers.
		 * Just tell the caller to process it normally.
		 */
		ret = IB_MAD_RESULT_SUCCESS;
		goto bail;
	default:
		smp->status |= IB_SMP_UNSUP_METHOD;
		ret = reply(smp);
	}

bail:
	return ret;
}

static int process_perf(struct ib_device *ibdev, u8 port_num,
			struct ib_mad *in_mad,
			struct ib_mad *out_mad)
{
	struct ib_perf *pmp = (struct ib_perf *)out_mad;
	int ret;

	*out_mad = *in_mad;
	if (pmp->class_version != 1) {
		pmp->status |= IB_SMP_UNSUP_VERSION;
		ret = reply((struct ib_smp *) pmp);
		goto bail;
	}

	switch (pmp->method) {
	case IB_MGMT_METHOD_GET:
		switch (pmp->attr_id) {
		case IB_PMA_CLASS_PORT_INFO:
			ret = recv_pma_get_classportinfo(pmp);
			goto bail;
		case IB_PMA_PORT_SAMPLES_CONTROL:
			ret = recv_pma_get_portsamplescontrol(pmp, ibdev,
							      port_num);
			goto bail;
		case IB_PMA_PORT_SAMPLES_RESULT:
			ret = recv_pma_get_portsamplesresult(pmp, ibdev);
			goto bail;
		case IB_PMA_PORT_SAMPLES_RESULT_EXT:
			ret = recv_pma_get_portsamplesresult_ext(pmp,
								 ibdev);
			goto bail;
		case IB_PMA_PORT_COUNTERS:
			ret = recv_pma_get_portcounters(pmp, ibdev,
							port_num);
			goto bail;
		case IB_PMA_PORT_COUNTERS_EXT:
			ret = recv_pma_get_portcounters_ext(pmp, ibdev,
							    port_num);
			goto bail;
		default:
			pmp->status |= IB_SMP_UNSUP_METH_ATTR;
			ret = reply((struct ib_smp *) pmp);
			goto bail;
		}

	case IB_MGMT_METHOD_SET:
		switch (pmp->attr_id) {
		case IB_PMA_PORT_SAMPLES_CONTROL:
			ret = recv_pma_set_portsamplescontrol(pmp, ibdev,
							      port_num);
			goto bail;
		case IB_PMA_PORT_COUNTERS:
			ret = recv_pma_set_portcounters(pmp, ibdev,
							port_num);
			goto bail;
		case IB_PMA_PORT_COUNTERS_EXT:
			ret = recv_pma_set_portcounters_ext(pmp, ibdev,
							    port_num);
			goto bail;
		default:
			pmp->status |= IB_SMP_UNSUP_METH_ATTR;
			ret = reply((struct ib_smp *) pmp);
			goto bail;
		}

	case IB_MGMT_METHOD_GET_RESP:
		/*
		 * The ib_mad module will call us to process responses
		 * before checking for other consumers.
		 * Just tell the caller to process it normally.
		 */
		ret = IB_MAD_RESULT_SUCCESS;
		goto bail;
	default:
		pmp->status |= IB_SMP_UNSUP_METHOD;
		ret = reply((struct ib_smp *) pmp);
	}

bail:
	return ret;
}

/**
 * ipath_process_mad - process an incoming MAD packet
 * @ibdev: the infiniband device this packet came in on
 * @mad_flags: MAD flags
 * @port_num: the port number this packet came in on
 * @in_wc: the work completion entry for this packet
 * @in_grh: the global route header for this packet
 * @in_mad: the incoming MAD
 * @out_mad: any outgoing MAD reply
 *
 * Returns IB_MAD_RESULT_SUCCESS if this is a MAD that we are not
 * interested in processing.
 *
 * Note that the verbs framework has already done the MAD sanity checks,
 * and hop count/pointer updating for IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE
 * MADs.
 *
 * This is called by the ib_mad module.
 */
int ipath_process_mad(struct ib_device *ibdev, int mad_flags, u8 port_num,
		      struct ib_wc *in_wc, struct ib_grh *in_grh,
		      struct ib_mad *in_mad, struct ib_mad *out_mad)
{
	int ret;

	switch (in_mad->mad_hdr.mgmt_class) {
	case IB_MGMT_CLASS_SUBN_DIRECTED_ROUTE:
	case IB_MGMT_CLASS_SUBN_LID_ROUTED:
		ret = process_subn(ibdev, mad_flags, port_num,
				   in_mad, out_mad);
		goto bail;
	case IB_MGMT_CLASS_PERF_MGMT:
		ret = process_perf(ibdev, port_num, in_mad, out_mad);
		goto bail;
	default:
		ret = IB_MAD_RESULT_SUCCESS;
	}

bail:
	return ret;
}