aboutsummaryrefslogtreecommitdiffstats
path: root/drivers/infiniband/hw/vmw_pvrdma/pvrdma_main.c
blob: 47e653d2495c41611f3448489ee97d7025fb59d5 (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
/*
 * Copyright (c) 2012-2016 VMware, Inc.  All rights reserved.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of EITHER the GNU General Public License
 * version 2 as published by the Free Software Foundation or the BSD
 * 2-Clause License. This program is distributed in the hope that it
 * will be useful, but WITHOUT ANY WARRANTY; WITHOUT EVEN THE IMPLIED
 * WARRANTY OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
 * See the GNU General Public License version 2 for more details at
 * http://www.gnu.org/licenses/old-licenses/gpl-2.0.en.html.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program available in the file COPYING in the main
 * directory of this source tree.
 *
 * The BSD 2-Clause License
 *
 *     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.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
 * COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT,
 * INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
 * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
 * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
 * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
 * OF THE POSSIBILITY OF SUCH DAMAGE.
 */

#include <linux/errno.h>
#include <linux/inetdevice.h>
#include <linux/init.h>
#include <linux/module.h>
#include <linux/slab.h>
#include <rdma/ib_addr.h>
#include <rdma/ib_smi.h>
#include <rdma/ib_user_verbs.h>
#include <net/addrconf.h>

#include "pvrdma.h"

#define DRV_NAME	"vmw_pvrdma"
#define DRV_VERSION	"1.0.1.0-k"

static DEFINE_MUTEX(pvrdma_device_list_lock);
static LIST_HEAD(pvrdma_device_list);
static struct workqueue_struct *event_wq;

static int pvrdma_add_gid(const struct ib_gid_attr *attr, void **context);
static int pvrdma_del_gid(const struct ib_gid_attr *attr, void **context);

static ssize_t hca_type_show(struct device *device,
			     struct device_attribute *attr, char *buf)
{
	return sprintf(buf, "VMW_PVRDMA-%s\n", DRV_VERSION);
}
static DEVICE_ATTR_RO(hca_type);

static ssize_t hw_rev_show(struct device *device,
			   struct device_attribute *attr, char *buf)
{
	return sprintf(buf, "%d\n", PVRDMA_REV_ID);
}
static DEVICE_ATTR_RO(hw_rev);

static ssize_t board_id_show(struct device *device,
			     struct device_attribute *attr, char *buf)
{
	return sprintf(buf, "%d\n", PVRDMA_BOARD_ID);
}
static DEVICE_ATTR_RO(board_id);

static struct attribute *pvrdma_class_attributes[] = {
	&dev_attr_hw_rev.attr,
	&dev_attr_hca_type.attr,
	&dev_attr_board_id.attr,
	NULL,
};

static const struct attribute_group pvrdma_attr_group = {
	.attrs = pvrdma_class_attributes,
};

static void pvrdma_get_fw_ver_str(struct ib_device *device, char *str)
{
	struct pvrdma_dev *dev =
		container_of(device, struct pvrdma_dev, ib_dev);
	snprintf(str, IB_FW_VERSION_NAME_MAX, "%d.%d.%d\n",
		 (int) (dev->dsr->caps.fw_ver >> 32),
		 (int) (dev->dsr->caps.fw_ver >> 16) & 0xffff,
		 (int) dev->dsr->caps.fw_ver & 0xffff);
}

static int pvrdma_init_device(struct pvrdma_dev *dev)
{
	/*  Initialize some device related stuff */
	spin_lock_init(&dev->cmd_lock);
	sema_init(&dev->cmd_sema, 1);
	atomic_set(&dev->num_qps, 0);
	atomic_set(&dev->num_srqs, 0);
	atomic_set(&dev->num_cqs, 0);
	atomic_set(&dev->num_pds, 0);
	atomic_set(&dev->num_ahs, 0);

	return 0;
}

static int pvrdma_port_immutable(struct ib_device *ibdev, u8 port_num,
				 struct ib_port_immutable *immutable)
{
	struct pvrdma_dev *dev = to_vdev(ibdev);
	struct ib_port_attr attr;
	int err;

	if (dev->dsr->caps.gid_types == PVRDMA_GID_TYPE_FLAG_ROCE_V1)
		immutable->core_cap_flags |= RDMA_CORE_PORT_IBA_ROCE;
	else if (dev->dsr->caps.gid_types == PVRDMA_GID_TYPE_FLAG_ROCE_V2)
		immutable->core_cap_flags |= RDMA_CORE_PORT_IBA_ROCE_UDP_ENCAP;

	err = ib_query_port(ibdev, port_num, &attr);
	if (err)
		return err;

	immutable->pkey_tbl_len = attr.pkey_tbl_len;
	immutable->gid_tbl_len = attr.gid_tbl_len;
	immutable->max_mad_size = IB_MGMT_MAD_SIZE;
	return 0;
}

static struct net_device *pvrdma_get_netdev(struct ib_device *ibdev,
					    u8 port_num)
{
	struct net_device *netdev;
	struct pvrdma_dev *dev = to_vdev(ibdev);

	if (port_num != 1)
		return NULL;

	rcu_read_lock();
	netdev = dev->netdev;
	if (netdev)
		dev_hold(netdev);
	rcu_read_unlock();

	return netdev;
}

static const struct ib_device_ops pvrdma_dev_ops = {
	.add_gid = pvrdma_add_gid,
	.alloc_mr = pvrdma_alloc_mr,
	.alloc_pd = pvrdma_alloc_pd,
	.alloc_ucontext = pvrdma_alloc_ucontext,
	.create_ah = pvrdma_create_ah,
	.create_cq = pvrdma_create_cq,
	.create_qp = pvrdma_create_qp,
	.dealloc_pd = pvrdma_dealloc_pd,
	.dealloc_ucontext = pvrdma_dealloc_ucontext,
	.del_gid = pvrdma_del_gid,
	.dereg_mr = pvrdma_dereg_mr,
	.destroy_ah = pvrdma_destroy_ah,
	.destroy_cq = pvrdma_destroy_cq,
	.destroy_qp = pvrdma_destroy_qp,
	.get_dev_fw_str = pvrdma_get_fw_ver_str,
	.get_dma_mr = pvrdma_get_dma_mr,
	.get_link_layer = pvrdma_port_link_layer,
	.get_netdev = pvrdma_get_netdev,
	.get_port_immutable = pvrdma_port_immutable,
	.map_mr_sg = pvrdma_map_mr_sg,
	.mmap = pvrdma_mmap,
	.modify_port = pvrdma_modify_port,
	.modify_qp = pvrdma_modify_qp,
	.poll_cq = pvrdma_poll_cq,
	.post_recv = pvrdma_post_recv,
	.post_send = pvrdma_post_send,
	.query_device = pvrdma_query_device,
	.query_gid = pvrdma_query_gid,
	.query_pkey = pvrdma_query_pkey,
	.query_port = pvrdma_query_port,
	.query_qp = pvrdma_query_qp,
	.reg_user_mr = pvrdma_reg_user_mr,
	.req_notify_cq = pvrdma_req_notify_cq,
	INIT_RDMA_OBJ_SIZE(ib_pd, pvrdma_pd, ibpd),
};

static const struct ib_device_ops pvrdma_dev_srq_ops = {
	.create_srq = pvrdma_create_srq,
	.destroy_srq = pvrdma_destroy_srq,
	.modify_srq = pvrdma_modify_srq,
	.query_srq = pvrdma_query_srq,
};

static int pvrdma_register_device(struct pvrdma_dev *dev)
{
	int ret = -1;

	dev->ib_dev.node_guid = dev->dsr->caps.node_guid;
	dev->sys_image_guid = dev->dsr->caps.sys_image_guid;
	dev->flags = 0;
	dev->ib_dev.owner = THIS_MODULE;
	dev->ib_dev.num_comp_vectors = 1;
	dev->ib_dev.dev.parent = &dev->pdev->dev;
	dev->ib_dev.uverbs_abi_ver = PVRDMA_UVERBS_ABI_VERSION;
	dev->ib_dev.uverbs_cmd_mask =
		(1ull << IB_USER_VERBS_CMD_GET_CONTEXT)		|
		(1ull << IB_USER_VERBS_CMD_QUERY_DEVICE)	|
		(1ull << IB_USER_VERBS_CMD_QUERY_PORT)		|
		(1ull << IB_USER_VERBS_CMD_ALLOC_PD)		|
		(1ull << IB_USER_VERBS_CMD_DEALLOC_PD)		|
		(1ull << IB_USER_VERBS_CMD_REG_MR)		|
		(1ull << IB_USER_VERBS_CMD_DEREG_MR)		|
		(1ull << IB_USER_VERBS_CMD_CREATE_COMP_CHANNEL)	|
		(1ull << IB_USER_VERBS_CMD_CREATE_CQ)		|
		(1ull << IB_USER_VERBS_CMD_POLL_CQ)		|
		(1ull << IB_USER_VERBS_CMD_REQ_NOTIFY_CQ)	|
		(1ull << IB_USER_VERBS_CMD_DESTROY_CQ)		|
		(1ull << IB_USER_VERBS_CMD_CREATE_QP)		|
		(1ull << IB_USER_VERBS_CMD_MODIFY_QP)		|
		(1ull << IB_USER_VERBS_CMD_QUERY_QP)		|
		(1ull << IB_USER_VERBS_CMD_DESTROY_QP)		|
		(1ull << IB_USER_VERBS_CMD_POST_SEND)		|
		(1ull << IB_USER_VERBS_CMD_POST_RECV)		|
		(1ull << IB_USER_VERBS_CMD_CREATE_AH)		|
		(1ull << IB_USER_VERBS_CMD_DESTROY_AH);

	dev->ib_dev.node_type = RDMA_NODE_IB_CA;
	dev->ib_dev.phys_port_cnt = dev->dsr->caps.phys_port_cnt;

	ib_set_device_ops(&dev->ib_dev, &pvrdma_dev_ops);

	mutex_init(&dev->port_mutex);
	spin_lock_init(&dev->desc_lock);

	dev->cq_tbl = kcalloc(dev->dsr->caps.max_cq, sizeof(struct pvrdma_cq *),
			      GFP_KERNEL);
	if (!dev->cq_tbl)
		return ret;
	spin_lock_init(&dev->cq_tbl_lock);

	dev->qp_tbl = kcalloc(dev->dsr->caps.max_qp, sizeof(struct pvrdma_qp *),
			      GFP_KERNEL);
	if (!dev->qp_tbl)
		goto err_cq_free;
	spin_lock_init(&dev->qp_tbl_lock);

	/* Check if SRQ is supported by backend */
	if (dev->dsr->caps.max_srq) {
		dev->ib_dev.uverbs_cmd_mask |=
			(1ull << IB_USER_VERBS_CMD_CREATE_SRQ)	|
			(1ull << IB_USER_VERBS_CMD_MODIFY_SRQ)	|
			(1ull << IB_USER_VERBS_CMD_QUERY_SRQ)	|
			(1ull << IB_USER_VERBS_CMD_DESTROY_SRQ)	|
			(1ull << IB_USER_VERBS_CMD_POST_SRQ_RECV);

		ib_set_device_ops(&dev->ib_dev, &pvrdma_dev_srq_ops);

		dev->srq_tbl = kcalloc(dev->dsr->caps.max_srq,
				       sizeof(struct pvrdma_srq *),
				       GFP_KERNEL);
		if (!dev->srq_tbl)
			goto err_qp_free;
	}
	dev->ib_dev.driver_id = RDMA_DRIVER_VMW_PVRDMA;
	spin_lock_init(&dev->srq_tbl_lock);
	rdma_set_device_sysfs_group(&dev->ib_dev, &pvrdma_attr_group);

	ret = ib_register_device(&dev->ib_dev, "vmw_pvrdma%d");
	if (ret)
		goto err_srq_free;

	dev->ib_active = true;

	return 0;

err_srq_free:
	kfree(dev->srq_tbl);
err_qp_free:
	kfree(dev->qp_tbl);
err_cq_free:
	kfree(dev->cq_tbl);

	return ret;
}

static irqreturn_t pvrdma_intr0_handler(int irq, void *dev_id)
{
	u32 icr = PVRDMA_INTR_CAUSE_RESPONSE;
	struct pvrdma_dev *dev = dev_id;

	dev_dbg(&dev->pdev->dev, "interrupt 0 (response) handler\n");

	if (!dev->pdev->msix_enabled) {
		/* Legacy intr */
		icr = pvrdma_read_reg(dev, PVRDMA_REG_ICR);
		if (icr == 0)
			return IRQ_NONE;
	}

	if (icr == PVRDMA_INTR_CAUSE_RESPONSE)
		complete(&dev->cmd_done);

	return IRQ_HANDLED;
}

static void pvrdma_qp_event(struct pvrdma_dev *dev, u32 qpn, int type)
{
	struct pvrdma_qp *qp;
	unsigned long flags;

	spin_lock_irqsave(&dev->qp_tbl_lock, flags);
	qp = dev->qp_tbl[qpn % dev->dsr->caps.max_qp];
	if (qp)
		refcount_inc(&qp->refcnt);
	spin_unlock_irqrestore(&dev->qp_tbl_lock, flags);

	if (qp && qp->ibqp.event_handler) {
		struct ib_qp *ibqp = &qp->ibqp;
		struct ib_event e;

		e.device = ibqp->device;
		e.element.qp = ibqp;
		e.event = type; /* 1:1 mapping for now. */
		ibqp->event_handler(&e, ibqp->qp_context);
	}
	if (qp) {
		if (refcount_dec_and_test(&qp->refcnt))
			complete(&qp->free);
	}
}

static void pvrdma_cq_event(struct pvrdma_dev *dev, u32 cqn, int type)
{
	struct pvrdma_cq *cq;
	unsigned long flags;

	spin_lock_irqsave(&dev->cq_tbl_lock, flags);
	cq = dev->cq_tbl[cqn % dev->dsr->caps.max_cq];
	if (cq)
		refcount_inc(&cq->refcnt);
	spin_unlock_irqrestore(&dev->cq_tbl_lock, flags);

	if (cq && cq->ibcq.event_handler) {
		struct ib_cq *ibcq = &cq->ibcq;
		struct ib_event e;

		e.device = ibcq->device;
		e.element.cq = ibcq;
		e.event = type; /* 1:1 mapping for now. */
		ibcq->event_handler(&e, ibcq->cq_context);
	}
	if (cq) {
		if (refcount_dec_and_test(&cq->refcnt))
			complete(&cq->free);
	}
}

static void pvrdma_srq_event(struct pvrdma_dev *dev, u32 srqn, int type)
{
	struct pvrdma_srq *srq;
	unsigned long flags;

	spin_lock_irqsave(&dev->srq_tbl_lock, flags);
	if (dev->srq_tbl)
		srq = dev->srq_tbl[srqn % dev->dsr->caps.max_srq];
	else
		srq = NULL;
	if (srq)
		refcount_inc(&srq->refcnt);
	spin_unlock_irqrestore(&dev->srq_tbl_lock, flags);

	if (srq && srq->ibsrq.event_handler) {
		struct ib_srq *ibsrq = &srq->ibsrq;
		struct ib_event e;

		e.device = ibsrq->device;
		e.element.srq = ibsrq;
		e.event = type; /* 1:1 mapping for now. */
		ibsrq->event_handler(&e, ibsrq->srq_context);
	}
	if (srq) {
		if (refcount_dec_and_test(&srq->refcnt))
			complete(&srq->free);
	}
}

static void pvrdma_dispatch_event(struct pvrdma_dev *dev, int port,
				  enum ib_event_type event)
{
	struct ib_event ib_event;

	memset(&ib_event, 0, sizeof(ib_event));
	ib_event.device = &dev->ib_dev;
	ib_event.element.port_num = port;
	ib_event.event = event;
	ib_dispatch_event(&ib_event);
}

static void pvrdma_dev_event(struct pvrdma_dev *dev, u8 port, int type)
{
	if (port < 1 || port > dev->dsr->caps.phys_port_cnt) {
		dev_warn(&dev->pdev->dev, "event on port %d\n", port);
		return;
	}

	pvrdma_dispatch_event(dev, port, type);
}

static inline struct pvrdma_eqe *get_eqe(struct pvrdma_dev *dev, unsigned int i)
{
	return (struct pvrdma_eqe *)pvrdma_page_dir_get_ptr(
					&dev->async_pdir,
					PAGE_SIZE +
					sizeof(struct pvrdma_eqe) * i);
}

static irqreturn_t pvrdma_intr1_handler(int irq, void *dev_id)
{
	struct pvrdma_dev *dev = dev_id;
	struct pvrdma_ring *ring = &dev->async_ring_state->rx;
	int ring_slots = (dev->dsr->async_ring_pages.num_pages - 1) *
			 PAGE_SIZE / sizeof(struct pvrdma_eqe);
	unsigned int head;

	dev_dbg(&dev->pdev->dev, "interrupt 1 (async event) handler\n");

	/*
	 * Don't process events until the IB device is registered. Otherwise
	 * we'll try to ib_dispatch_event() on an invalid device.
	 */
	if (!dev->ib_active)
		return IRQ_HANDLED;

	while (pvrdma_idx_ring_has_data(ring, ring_slots, &head) > 0) {
		struct pvrdma_eqe *eqe;

		eqe = get_eqe(dev, head);

		switch (eqe->type) {
		case PVRDMA_EVENT_QP_FATAL:
		case PVRDMA_EVENT_QP_REQ_ERR:
		case PVRDMA_EVENT_QP_ACCESS_ERR:
		case PVRDMA_EVENT_COMM_EST:
		case PVRDMA_EVENT_SQ_DRAINED:
		case PVRDMA_EVENT_PATH_MIG:
		case PVRDMA_EVENT_PATH_MIG_ERR:
		case PVRDMA_EVENT_QP_LAST_WQE_REACHED:
			pvrdma_qp_event(dev, eqe->info, eqe->type);
			break;

		case PVRDMA_EVENT_CQ_ERR:
			pvrdma_cq_event(dev, eqe->info, eqe->type);
			break;

		case PVRDMA_EVENT_SRQ_ERR:
		case PVRDMA_EVENT_SRQ_LIMIT_REACHED:
			pvrdma_srq_event(dev, eqe->info, eqe->type);
			break;

		case PVRDMA_EVENT_PORT_ACTIVE:
		case PVRDMA_EVENT_PORT_ERR:
		case PVRDMA_EVENT_LID_CHANGE:
		case PVRDMA_EVENT_PKEY_CHANGE:
		case PVRDMA_EVENT_SM_CHANGE:
		case PVRDMA_EVENT_CLIENT_REREGISTER:
		case PVRDMA_EVENT_GID_CHANGE:
			pvrdma_dev_event(dev, eqe->info, eqe->type);
			break;

		case PVRDMA_EVENT_DEVICE_FATAL:
			pvrdma_dev_event(dev, 1, eqe->type);
			break;

		default:
			break;
		}

		pvrdma_idx_ring_inc(&ring->cons_head, ring_slots);
	}

	return IRQ_HANDLED;
}

static inline struct pvrdma_cqne *get_cqne(struct pvrdma_dev *dev,
					   unsigned int i)
{
	return (struct pvrdma_cqne *)pvrdma_page_dir_get_ptr(
					&dev->cq_pdir,
					PAGE_SIZE +
					sizeof(struct pvrdma_cqne) * i);
}

static irqreturn_t pvrdma_intrx_handler(int irq, void *dev_id)
{
	struct pvrdma_dev *dev = dev_id;
	struct pvrdma_ring *ring = &dev->cq_ring_state->rx;
	int ring_slots = (dev->dsr->cq_ring_pages.num_pages - 1) * PAGE_SIZE /
			 sizeof(struct pvrdma_cqne);
	unsigned int head;
	unsigned long flags;

	dev_dbg(&dev->pdev->dev, "interrupt x (completion) handler\n");

	while (pvrdma_idx_ring_has_data(ring, ring_slots, &head) > 0) {
		struct pvrdma_cqne *cqne;
		struct pvrdma_cq *cq;

		cqne = get_cqne(dev, head);
		spin_lock_irqsave(&dev->cq_tbl_lock, flags);
		cq = dev->cq_tbl[cqne->info % dev->dsr->caps.max_cq];
		if (cq)
			refcount_inc(&cq->refcnt);
		spin_unlock_irqrestore(&dev->cq_tbl_lock, flags);

		if (cq && cq->ibcq.comp_handler)
			cq->ibcq.comp_handler(&cq->ibcq, cq->ibcq.cq_context);
		if (cq) {
			if (refcount_dec_and_test(&cq->refcnt))
				complete(&cq->free);
		}
		pvrdma_idx_ring_inc(&ring->cons_head, ring_slots);
	}

	return IRQ_HANDLED;
}

static void pvrdma_free_irq(struct pvrdma_dev *dev)
{
	int i;

	dev_dbg(&dev->pdev->dev, "freeing interrupts\n");
	for (i = 0; i < dev->nr_vectors; i++)
		free_irq(pci_irq_vector(dev->pdev, i), dev);
}

static void pvrdma_enable_intrs(struct pvrdma_dev *dev)
{
	dev_dbg(&dev->pdev->dev, "enable interrupts\n");
	pvrdma_write_reg(dev, PVRDMA_REG_IMR, 0);
}

static void pvrdma_disable_intrs(struct pvrdma_dev *dev)
{
	dev_dbg(&dev->pdev->dev, "disable interrupts\n");
	pvrdma_write_reg(dev, PVRDMA_REG_IMR, ~0);
}

static int pvrdma_alloc_intrs(struct pvrdma_dev *dev)
{
	struct pci_dev *pdev = dev->pdev;
	int ret = 0, i;

	ret = pci_alloc_irq_vectors(pdev, 1, PVRDMA_MAX_INTERRUPTS,
			PCI_IRQ_MSIX);
	if (ret < 0) {
		ret = pci_alloc_irq_vectors(pdev, 1, 1,
				PCI_IRQ_MSI | PCI_IRQ_LEGACY);
		if (ret < 0)
			return ret;
	}
	dev->nr_vectors = ret;

	ret = request_irq(pci_irq_vector(dev->pdev, 0), pvrdma_intr0_handler,
			pdev->msix_enabled ? 0 : IRQF_SHARED, DRV_NAME, dev);
	if (ret) {
		dev_err(&dev->pdev->dev,
			"failed to request interrupt 0\n");
		goto out_free_vectors;
	}

	for (i = 1; i < dev->nr_vectors; i++) {
		ret = request_irq(pci_irq_vector(dev->pdev, i),
				i == 1 ? pvrdma_intr1_handler :
					 pvrdma_intrx_handler,
				0, DRV_NAME, dev);
		if (ret) {
			dev_err(&dev->pdev->dev,
				"failed to request interrupt %d\n", i);
			goto free_irqs;
		}
	}

	return 0;

free_irqs:
	while (--i >= 0)
		free_irq(pci_irq_vector(dev->pdev, i), dev);
out_free_vectors:
	pci_free_irq_vectors(pdev);
	return ret;
}

static void pvrdma_free_slots(struct pvrdma_dev *dev)
{
	struct pci_dev *pdev = dev->pdev;

	if (dev->resp_slot)
		dma_free_coherent(&pdev->dev, PAGE_SIZE, dev->resp_slot,
				  dev->dsr->resp_slot_dma);
	if (dev->cmd_slot)
		dma_free_coherent(&pdev->dev, PAGE_SIZE, dev->cmd_slot,
				  dev->dsr->cmd_slot_dma);
}

static int pvrdma_add_gid_at_index(struct pvrdma_dev *dev,
				   const union ib_gid *gid,
				   u8 gid_type,
				   int index)
{
	int ret;
	union pvrdma_cmd_req req;
	struct pvrdma_cmd_create_bind *cmd_bind = &req.create_bind;

	if (!dev->sgid_tbl) {
		dev_warn(&dev->pdev->dev, "sgid table not initialized\n");
		return -EINVAL;
	}

	memset(cmd_bind, 0, sizeof(*cmd_bind));
	cmd_bind->hdr.cmd = PVRDMA_CMD_CREATE_BIND;
	memcpy(cmd_bind->new_gid, gid->raw, 16);
	cmd_bind->mtu = ib_mtu_enum_to_int(IB_MTU_1024);
	cmd_bind->vlan = 0xfff;
	cmd_bind->index = index;
	cmd_bind->gid_type = gid_type;

	ret = pvrdma_cmd_post(dev, &req, NULL, 0);
	if (ret < 0) {
		dev_warn(&dev->pdev->dev,
			 "could not create binding, error: %d\n", ret);
		return -EFAULT;
	}
	memcpy(&dev->sgid_tbl[index], gid, sizeof(*gid));
	return 0;
}

static int pvrdma_add_gid(const struct ib_gid_attr *attr, void **context)
{
	struct pvrdma_dev *dev = to_vdev(attr->device);

	return pvrdma_add_gid_at_index(dev, &attr->gid,
				       ib_gid_type_to_pvrdma(attr->gid_type),
				       attr->index);
}

static int pvrdma_del_gid_at_index(struct pvrdma_dev *dev, int index)
{
	int ret;
	union pvrdma_cmd_req req;
	struct pvrdma_cmd_destroy_bind *cmd_dest = &req.destroy_bind;

	/* Update sgid table. */
	if (!dev->sgid_tbl) {
		dev_warn(&dev->pdev->dev, "sgid table not initialized\n");
		return -EINVAL;
	}

	memset(cmd_dest, 0, sizeof(*cmd_dest));
	cmd_dest->hdr.cmd = PVRDMA_CMD_DESTROY_BIND;
	memcpy(cmd_dest->dest_gid, &dev->sgid_tbl[index], 16);
	cmd_dest->index = index;

	ret = pvrdma_cmd_post(dev, &req, NULL, 0);
	if (ret < 0) {
		dev_warn(&dev->pdev->dev,
			 "could not destroy binding, error: %d\n", ret);
		return ret;
	}
	memset(&dev->sgid_tbl[index], 0, 16);
	return 0;
}

static int pvrdma_del_gid(const struct ib_gid_attr *attr, void **context)
{
	struct pvrdma_dev *dev = to_vdev(attr->device);

	dev_dbg(&dev->pdev->dev, "removing gid at index %u from %s",
		attr->index, dev->netdev->name);

	return pvrdma_del_gid_at_index(dev, attr->index);
}

static void pvrdma_netdevice_event_handle(struct pvrdma_dev *dev,
					  struct net_device *ndev,
					  unsigned long event)
{
	struct pci_dev *pdev_net;
	unsigned int slot;

	switch (event) {
	case NETDEV_REBOOT:
	case NETDEV_DOWN:
		pvrdma_dispatch_event(dev, 1, IB_EVENT_PORT_ERR);
		break;
	case NETDEV_UP:
		pvrdma_write_reg(dev, PVRDMA_REG_CTL,
				 PVRDMA_DEVICE_CTL_UNQUIESCE);

		mb();

		if (pvrdma_read_reg(dev, PVRDMA_REG_ERR))
			dev_err(&dev->pdev->dev,
				"failed to activate device during link up\n");
		else
			pvrdma_dispatch_event(dev, 1, IB_EVENT_PORT_ACTIVE);
		break;
	case NETDEV_UNREGISTER:
		dev_put(dev->netdev);
		dev->netdev = NULL;
		break;
	case NETDEV_REGISTER:
		/* vmxnet3 will have same bus, slot. But func will be 0 */
		slot = PCI_SLOT(dev->pdev->devfn);
		pdev_net = pci_get_slot(dev->pdev->bus,
					PCI_DEVFN(slot, 0));
		if ((dev->netdev == NULL) &&
		    (pci_get_drvdata(pdev_net) == ndev)) {
			/* this is our netdev */
			dev->netdev = ndev;
			dev_hold(ndev);
		}
		pci_dev_put(pdev_net);
		break;

	default:
		dev_dbg(&dev->pdev->dev, "ignore netdevice event %ld on %s\n",
			event, dev_name(&dev->ib_dev.dev));
		break;
	}
}

static void pvrdma_netdevice_event_work(struct work_struct *work)
{
	struct pvrdma_netdevice_work *netdev_work;
	struct pvrdma_dev *dev;

	netdev_work = container_of(work, struct pvrdma_netdevice_work, work);

	mutex_lock(&pvrdma_device_list_lock);
	list_for_each_entry(dev, &pvrdma_device_list, device_link) {
		if ((netdev_work->event == NETDEV_REGISTER) ||
		    (dev->netdev == netdev_work->event_netdev)) {
			pvrdma_netdevice_event_handle(dev,
						      netdev_work->event_netdev,
						      netdev_work->event);
			break;
		}
	}
	mutex_unlock(&pvrdma_device_list_lock);

	kfree(netdev_work);
}

static int pvrdma_netdevice_event(struct notifier_block *this,
				  unsigned long event, void *ptr)
{
	struct net_device *event_netdev = netdev_notifier_info_to_dev(ptr);
	struct pvrdma_netdevice_work *netdev_work;

	netdev_work = kmalloc(sizeof(*netdev_work), GFP_ATOMIC);
	if (!netdev_work)
		return NOTIFY_BAD;

	INIT_WORK(&netdev_work->work, pvrdma_netdevice_event_work);
	netdev_work->event_netdev = event_netdev;
	netdev_work->event = event;
	queue_work(event_wq, &netdev_work->work);

	return NOTIFY_DONE;
}

static int pvrdma_pci_probe(struct pci_dev *pdev,
			    const struct pci_device_id *id)
{
	struct pci_dev *pdev_net;
	struct pvrdma_dev *dev;
	int ret;
	unsigned long start;
	unsigned long len;
	dma_addr_t slot_dma = 0;

	dev_dbg(&pdev->dev, "initializing driver %s\n", pci_name(pdev));

	/* Allocate zero-out device */
	dev = ib_alloc_device(pvrdma_dev, ib_dev);
	if (!dev) {
		dev_err(&pdev->dev, "failed to allocate IB device\n");
		return -ENOMEM;
	}

	mutex_lock(&pvrdma_device_list_lock);
	list_add(&dev->device_link, &pvrdma_device_list);
	mutex_unlock(&pvrdma_device_list_lock);

	ret = pvrdma_init_device(dev);
	if (ret)
		goto err_free_device;

	dev->pdev = pdev;
	pci_set_drvdata(pdev, dev);

	ret = pci_enable_device(pdev);
	if (ret) {
		dev_err(&pdev->dev, "cannot enable PCI device\n");
		goto err_free_device;
	}

	dev_dbg(&pdev->dev, "PCI resource flags BAR0 %#lx\n",
		pci_resource_flags(pdev, 0));
	dev_dbg(&pdev->dev, "PCI resource len %#llx\n",
		(unsigned long long)pci_resource_len(pdev, 0));
	dev_dbg(&pdev->dev, "PCI resource start %#llx\n",
		(unsigned long long)pci_resource_start(pdev, 0));
	dev_dbg(&pdev->dev, "PCI resource flags BAR1 %#lx\n",
		pci_resource_flags(pdev, 1));
	dev_dbg(&pdev->dev, "PCI resource len %#llx\n",
		(unsigned long long)pci_resource_len(pdev, 1));
	dev_dbg(&pdev->dev, "PCI resource start %#llx\n",
		(unsigned long long)pci_resource_start(pdev, 1));

	if (!(pci_resource_flags(pdev, 0) & IORESOURCE_MEM) ||
	    !(pci_resource_flags(pdev, 1) & IORESOURCE_MEM)) {
		dev_err(&pdev->dev, "PCI BAR region not MMIO\n");
		ret = -ENOMEM;
		goto err_free_device;
	}

	ret = pci_request_regions(pdev, DRV_NAME);
	if (ret) {
		dev_err(&pdev->dev, "cannot request PCI resources\n");
		goto err_disable_pdev;
	}

	/* Enable 64-Bit DMA */
	if (pci_set_dma_mask(pdev, DMA_BIT_MASK(64)) == 0) {
		ret = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(64));
		if (ret != 0) {
			dev_err(&pdev->dev,
				"pci_set_consistent_dma_mask failed\n");
			goto err_free_resource;
		}
	} else {
		ret = pci_set_dma_mask(pdev, DMA_BIT_MASK(32));
		if (ret != 0) {
			dev_err(&pdev->dev,
				"pci_set_dma_mask failed\n");
			goto err_free_resource;
		}
	}

	pci_set_master(pdev);

	/* Map register space */
	start = pci_resource_start(dev->pdev, PVRDMA_PCI_RESOURCE_REG);
	len = pci_resource_len(dev->pdev, PVRDMA_PCI_RESOURCE_REG);
	dev->regs = ioremap(start, len);
	if (!dev->regs) {
		dev_err(&pdev->dev, "register mapping failed\n");
		ret = -ENOMEM;
		goto err_free_resource;
	}

	/* Setup per-device UAR. */
	dev->driver_uar.index = 0;
	dev->driver_uar.pfn =
		pci_resource_start(dev->pdev, PVRDMA_PCI_RESOURCE_UAR) >>
		PAGE_SHIFT;
	dev->driver_uar.map =
		ioremap(dev->driver_uar.pfn << PAGE_SHIFT, PAGE_SIZE);
	if (!dev->driver_uar.map) {
		dev_err(&pdev->dev, "failed to remap UAR pages\n");
		ret = -ENOMEM;
		goto err_unmap_regs;
	}

	dev->dsr_version = pvrdma_read_reg(dev, PVRDMA_REG_VERSION);
	dev_info(&pdev->dev, "device version %d, driver version %d\n",
		 dev->dsr_version, PVRDMA_VERSION);

	dev->dsr = dma_alloc_coherent(&pdev->dev, sizeof(*dev->dsr),
				      &dev->dsrbase, GFP_KERNEL);
	if (!dev->dsr) {
		dev_err(&pdev->dev, "failed to allocate shared region\n");
		ret = -ENOMEM;
		goto err_uar_unmap;
	}

	/* Setup the shared region */
	dev->dsr->driver_version = PVRDMA_VERSION;
	dev->dsr->gos_info.gos_bits = sizeof(void *) == 4 ?
		PVRDMA_GOS_BITS_32 :
		PVRDMA_GOS_BITS_64;
	dev->dsr->gos_info.gos_type = PVRDMA_GOS_TYPE_LINUX;
	dev->dsr->gos_info.gos_ver = 1;

	if (dev->dsr_version < PVRDMA_PPN64_VERSION)
		dev->dsr->uar_pfn = dev->driver_uar.pfn;
	else
		dev->dsr->uar_pfn64 = dev->driver_uar.pfn;

	/* Command slot. */
	dev->cmd_slot = dma_alloc_coherent(&pdev->dev, PAGE_SIZE,
					   &slot_dma, GFP_KERNEL);
	if (!dev->cmd_slot) {
		ret = -ENOMEM;
		goto err_free_dsr;
	}

	dev->dsr->cmd_slot_dma = (u64)slot_dma;

	/* Response slot. */
	dev->resp_slot = dma_alloc_coherent(&pdev->dev, PAGE_SIZE,
					    &slot_dma, GFP_KERNEL);
	if (!dev->resp_slot) {
		ret = -ENOMEM;
		goto err_free_slots;
	}

	dev->dsr->resp_slot_dma = (u64)slot_dma;

	/* Async event ring */
	dev->dsr->async_ring_pages.num_pages = PVRDMA_NUM_RING_PAGES;
	ret = pvrdma_page_dir_init(dev, &dev->async_pdir,
				   dev->dsr->async_ring_pages.num_pages, true);
	if (ret)
		goto err_free_slots;
	dev->async_ring_state = dev->async_pdir.pages[0];
	dev->dsr->async_ring_pages.pdir_dma = dev->async_pdir.dir_dma;

	/* CQ notification ring */
	dev->dsr->cq_ring_pages.num_pages = PVRDMA_NUM_RING_PAGES;
	ret = pvrdma_page_dir_init(dev, &dev->cq_pdir,
				   dev->dsr->cq_ring_pages.num_pages, true);
	if (ret)
		goto err_free_async_ring;
	dev->cq_ring_state = dev->cq_pdir.pages[0];
	dev->dsr->cq_ring_pages.pdir_dma = dev->cq_pdir.dir_dma;

	/*
	 * Write the PA of the shared region to the device. The writes must be
	 * ordered such that the high bits are written last. When the writes
	 * complete, the device will have filled out the capabilities.
	 */

	pvrdma_write_reg(dev, PVRDMA_REG_DSRLOW, (u32)dev->dsrbase);
	pvrdma_write_reg(dev, PVRDMA_REG_DSRHIGH,
			 (u32)((u64)(dev->dsrbase) >> 32));

	/* Make sure the write is complete before reading status. */
	mb();

	/* The driver supports RoCE V1 and V2. */
	if (!PVRDMA_SUPPORTED(dev)) {
		dev_err(&pdev->dev, "driver needs RoCE v1 or v2 support\n");
		ret = -EFAULT;
		goto err_free_cq_ring;
	}

	/* Paired vmxnet3 will have same bus, slot. But func will be 0 */
	pdev_net = pci_get_slot(pdev->bus, PCI_DEVFN(PCI_SLOT(pdev->devfn), 0));
	if (!pdev_net) {
		dev_err(&pdev->dev, "failed to find paired net device\n");
		ret = -ENODEV;
		goto err_free_cq_ring;
	}

	if (pdev_net->vendor != PCI_VENDOR_ID_VMWARE ||
	    pdev_net->device != PCI_DEVICE_ID_VMWARE_VMXNET3) {
		dev_err(&pdev->dev, "failed to find paired vmxnet3 device\n");
		pci_dev_put(pdev_net);
		ret = -ENODEV;
		goto err_free_cq_ring;
	}

	dev->netdev = pci_get_drvdata(pdev_net);
	pci_dev_put(pdev_net);
	if (!dev->netdev) {
		dev_err(&pdev->dev, "failed to get vmxnet3 device\n");
		ret = -ENODEV;
		goto err_free_cq_ring;
	}
	dev_hold(dev->netdev);

	dev_info(&pdev->dev, "paired device to %s\n", dev->netdev->name);

	/* Interrupt setup */
	ret = pvrdma_alloc_intrs(dev);
	if (ret) {
		dev_err(&pdev->dev, "failed to allocate interrupts\n");
		ret = -ENOMEM;
		goto err_free_cq_ring;
	}

	/* Allocate UAR table. */
	ret = pvrdma_uar_table_init(dev);
	if (ret) {
		dev_err(&pdev->dev, "failed to allocate UAR table\n");
		ret = -ENOMEM;
		goto err_free_intrs;
	}

	/* Allocate GID table */
	dev->sgid_tbl = kcalloc(dev->dsr->caps.gid_tbl_len,
				sizeof(union ib_gid), GFP_KERNEL);
	if (!dev->sgid_tbl) {
		ret = -ENOMEM;
		goto err_free_uar_table;
	}
	dev_dbg(&pdev->dev, "gid table len %d\n", dev->dsr->caps.gid_tbl_len);

	pvrdma_enable_intrs(dev);

	/* Activate pvrdma device */
	pvrdma_write_reg(dev, PVRDMA_REG_CTL, PVRDMA_DEVICE_CTL_ACTIVATE);

	/* Make sure the write is complete before reading status. */
	mb();

	/* Check if device was successfully activated */
	ret = pvrdma_read_reg(dev, PVRDMA_REG_ERR);
	if (ret != 0) {
		dev_err(&pdev->dev, "failed to activate device\n");
		ret = -EFAULT;
		goto err_disable_intr;
	}

	/* Register IB device */
	ret = pvrdma_register_device(dev);
	if (ret) {
		dev_err(&pdev->dev, "failed to register IB device\n");
		goto err_disable_intr;
	}

	dev->nb_netdev.notifier_call = pvrdma_netdevice_event;
	ret = register_netdevice_notifier(&dev->nb_netdev);
	if (ret) {
		dev_err(&pdev->dev, "failed to register netdevice events\n");
		goto err_unreg_ibdev;
	}

	dev_info(&pdev->dev, "attached to device\n");
	return 0;

err_unreg_ibdev:
	ib_unregister_device(&dev->ib_dev);
err_disable_intr:
	pvrdma_disable_intrs(dev);
	kfree(dev->sgid_tbl);
err_free_uar_table:
	pvrdma_uar_table_cleanup(dev);
err_free_intrs:
	pvrdma_free_irq(dev);
	pci_free_irq_vectors(pdev);
err_free_cq_ring:
	if (dev->netdev) {
		dev_put(dev->netdev);
		dev->netdev = NULL;
	}
	pvrdma_page_dir_cleanup(dev, &dev->cq_pdir);
err_free_async_ring:
	pvrdma_page_dir_cleanup(dev, &dev->async_pdir);
err_free_slots:
	pvrdma_free_slots(dev);
err_free_dsr:
	dma_free_coherent(&pdev->dev, sizeof(*dev->dsr), dev->dsr,
			  dev->dsrbase);
err_uar_unmap:
	iounmap(dev->driver_uar.map);
err_unmap_regs:
	iounmap(dev->regs);
err_free_resource:
	pci_release_regions(pdev);
err_disable_pdev:
	pci_disable_device(pdev);
	pci_set_drvdata(pdev, NULL);
err_free_device:
	mutex_lock(&pvrdma_device_list_lock);
	list_del(&dev->device_link);
	mutex_unlock(&pvrdma_device_list_lock);
	ib_dealloc_device(&dev->ib_dev);
	return ret;
}

static void pvrdma_pci_remove(struct pci_dev *pdev)
{
	struct pvrdma_dev *dev = pci_get_drvdata(pdev);

	if (!dev)
		return;

	dev_info(&pdev->dev, "detaching from device\n");

	unregister_netdevice_notifier(&dev->nb_netdev);
	dev->nb_netdev.notifier_call = NULL;

	flush_workqueue(event_wq);

	if (dev->netdev) {
		dev_put(dev->netdev);
		dev->netdev = NULL;
	}

	/* Unregister ib device */
	ib_unregister_device(&dev->ib_dev);

	mutex_lock(&pvrdma_device_list_lock);
	list_del(&dev->device_link);
	mutex_unlock(&pvrdma_device_list_lock);

	pvrdma_disable_intrs(dev);
	pvrdma_free_irq(dev);
	pci_free_irq_vectors(pdev);

	/* Deactivate pvrdma device */
	pvrdma_write_reg(dev, PVRDMA_REG_CTL, PVRDMA_DEVICE_CTL_RESET);
	pvrdma_page_dir_cleanup(dev, &dev->cq_pdir);
	pvrdma_page_dir_cleanup(dev, &dev->async_pdir);
	pvrdma_free_slots(dev);

	iounmap(dev->regs);
	kfree(dev->sgid_tbl);
	kfree(dev->cq_tbl);
	kfree(dev->srq_tbl);
	kfree(dev->qp_tbl);
	pvrdma_uar_table_cleanup(dev);
	iounmap(dev->driver_uar.map);

	ib_dealloc_device(&dev->ib_dev);

	/* Free pci resources */
	pci_release_regions(pdev);
	pci_disable_device(pdev);
	pci_set_drvdata(pdev, NULL);
}

static const struct pci_device_id pvrdma_pci_table[] = {
	{ PCI_DEVICE(PCI_VENDOR_ID_VMWARE, PCI_DEVICE_ID_VMWARE_PVRDMA), },
	{ 0 },
};

MODULE_DEVICE_TABLE(pci, pvrdma_pci_table);

static struct pci_driver pvrdma_driver = {
	.name		= DRV_NAME,
	.id_table	= pvrdma_pci_table,
	.probe		= pvrdma_pci_probe,
	.remove		= pvrdma_pci_remove,
};

static int __init pvrdma_init(void)
{
	int err;

	event_wq = alloc_ordered_workqueue("pvrdma_event_wq", WQ_MEM_RECLAIM);
	if (!event_wq)
		return -ENOMEM;

	err = pci_register_driver(&pvrdma_driver);
	if (err)
		destroy_workqueue(event_wq);

	return err;
}

static void __exit pvrdma_cleanup(void)
{
	pci_unregister_driver(&pvrdma_driver);

	destroy_workqueue(event_wq);
}

module_init(pvrdma_init);
module_exit(pvrdma_cleanup);

MODULE_AUTHOR("VMware, Inc");
MODULE_DESCRIPTION("VMware Paravirtual RDMA driver");
MODULE_LICENSE("Dual BSD/GPL");