summaryrefslogtreecommitdiff
path: root/net/ipv4/udp_tunnel_nic.c
blob: f0dbd9905a5316ad5071c7e6476de91fa026fa86 (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
// SPDX-License-Identifier: GPL-2.0-only
// Copyright (c) 2020 Facebook Inc.

#include <linux/ethtool_netlink.h>
#include <linux/netdevice.h>
#include <linux/slab.h>
#include <linux/types.h>
#include <linux/workqueue.h>
#include <net/udp_tunnel.h>

enum udp_tunnel_nic_table_entry_flags {
	UDP_TUNNEL_NIC_ENTRY_ADD	= BIT(0),
	UDP_TUNNEL_NIC_ENTRY_DEL	= BIT(1),
	UDP_TUNNEL_NIC_ENTRY_OP_FAIL	= BIT(2),
	UDP_TUNNEL_NIC_ENTRY_FROZEN	= BIT(3),
};

struct udp_tunnel_nic_table_entry {
	__be16 port;
	u8 type;
	u8 use_cnt;
	u8 flags;
	u8 hw_priv;
};

/**
 * struct udp_tunnel_nic - UDP tunnel port offload state
 * @work:	async work for talking to hardware from process context
 * @dev:	netdev pointer
 * @need_sync:	at least one port start changed
 * @need_replay: space was freed, we need a replay of all ports
 * @work_pending: @work is currently scheduled
 * @n_tables:	number of tables under @entries
 * @missed:	bitmap of tables which overflown
 * @entries:	table of tables of ports currently offloaded
 */
struct udp_tunnel_nic {
	struct work_struct work;

	struct net_device *dev;

	u8 need_sync:1;
	u8 need_replay:1;
	u8 work_pending:1;

	unsigned int n_tables;
	unsigned long missed;
	struct udp_tunnel_nic_table_entry **entries;
};

/* We ensure all work structs are done using driver state, but not the code.
 * We need a workqueue we can flush before module gets removed.
 */
static struct workqueue_struct *udp_tunnel_nic_workqueue;

static const char *udp_tunnel_nic_tunnel_type_name(unsigned int type)
{
	switch (type) {
	case UDP_TUNNEL_TYPE_VXLAN:
		return "vxlan";
	case UDP_TUNNEL_TYPE_GENEVE:
		return "geneve";
	case UDP_TUNNEL_TYPE_VXLAN_GPE:
		return "vxlan-gpe";
	default:
		return "unknown";
	}
}

static bool
udp_tunnel_nic_entry_is_free(struct udp_tunnel_nic_table_entry *entry)
{
	return entry->use_cnt == 0 && !entry->flags;
}

static bool
udp_tunnel_nic_entry_is_present(struct udp_tunnel_nic_table_entry *entry)
{
	return entry->use_cnt && !(entry->flags & ~UDP_TUNNEL_NIC_ENTRY_FROZEN);
}

static bool
udp_tunnel_nic_entry_is_frozen(struct udp_tunnel_nic_table_entry *entry)
{
	return entry->flags & UDP_TUNNEL_NIC_ENTRY_FROZEN;
}

static void
udp_tunnel_nic_entry_freeze_used(struct udp_tunnel_nic_table_entry *entry)
{
	if (!udp_tunnel_nic_entry_is_free(entry))
		entry->flags |= UDP_TUNNEL_NIC_ENTRY_FROZEN;
}

static void
udp_tunnel_nic_entry_unfreeze(struct udp_tunnel_nic_table_entry *entry)
{
	entry->flags &= ~UDP_TUNNEL_NIC_ENTRY_FROZEN;
}

static bool
udp_tunnel_nic_entry_is_queued(struct udp_tunnel_nic_table_entry *entry)
{
	return entry->flags & (UDP_TUNNEL_NIC_ENTRY_ADD |
			       UDP_TUNNEL_NIC_ENTRY_DEL);
}

static void
udp_tunnel_nic_entry_queue(struct udp_tunnel_nic *utn,
			   struct udp_tunnel_nic_table_entry *entry,
			   unsigned int flag)
{
	entry->flags |= flag;
	utn->need_sync = 1;
}

static void
udp_tunnel_nic_ti_from_entry(struct udp_tunnel_nic_table_entry *entry,
			     struct udp_tunnel_info *ti)
{
	memset(ti, 0, sizeof(*ti));
	ti->port = entry->port;
	ti->type = entry->type;
	ti->hw_priv = entry->hw_priv;
}

static bool
udp_tunnel_nic_is_empty(struct net_device *dev, struct udp_tunnel_nic *utn)
{
	const struct udp_tunnel_nic_info *info = dev->udp_tunnel_nic_info;
	unsigned int i, j;

	for (i = 0; i < utn->n_tables; i++)
		for (j = 0; j < info->tables[i].n_entries; j++)
			if (!udp_tunnel_nic_entry_is_free(&utn->entries[i][j]))
				return false;
	return true;
}

static bool
udp_tunnel_nic_should_replay(struct net_device *dev, struct udp_tunnel_nic *utn)
{
	const struct udp_tunnel_nic_table_info *table;
	unsigned int i, j;

	if (!utn->missed)
		return false;

	for (i = 0; i < utn->n_tables; i++) {
		table = &dev->udp_tunnel_nic_info->tables[i];
		if (!test_bit(i, &utn->missed))
			continue;

		for (j = 0; j < table->n_entries; j++)
			if (udp_tunnel_nic_entry_is_free(&utn->entries[i][j]))
				return true;
	}

	return false;
}

static void
__udp_tunnel_nic_get_port(struct net_device *dev, unsigned int table,
			  unsigned int idx, struct udp_tunnel_info *ti)
{
	struct udp_tunnel_nic_table_entry *entry;
	struct udp_tunnel_nic *utn;

	utn = dev->udp_tunnel_nic;
	entry = &utn->entries[table][idx];

	if (entry->use_cnt)
		udp_tunnel_nic_ti_from_entry(entry, ti);
}

static void
__udp_tunnel_nic_set_port_priv(struct net_device *dev, unsigned int table,
			       unsigned int idx, u8 priv)
{
	dev->udp_tunnel_nic->entries[table][idx].hw_priv = priv;
}

static void
udp_tunnel_nic_entry_update_done(struct udp_tunnel_nic_table_entry *entry,
				 int err)
{
	bool dodgy = entry->flags & UDP_TUNNEL_NIC_ENTRY_OP_FAIL;

	WARN_ON_ONCE(entry->flags & UDP_TUNNEL_NIC_ENTRY_ADD &&
		     entry->flags & UDP_TUNNEL_NIC_ENTRY_DEL);

	if (entry->flags & UDP_TUNNEL_NIC_ENTRY_ADD &&
	    (!err || (err == -EEXIST && dodgy)))
		entry->flags &= ~UDP_TUNNEL_NIC_ENTRY_ADD;

	if (entry->flags & UDP_TUNNEL_NIC_ENTRY_DEL &&
	    (!err || (err == -ENOENT && dodgy)))
		entry->flags &= ~UDP_TUNNEL_NIC_ENTRY_DEL;

	if (!err)
		entry->flags &= ~UDP_TUNNEL_NIC_ENTRY_OP_FAIL;
	else
		entry->flags |= UDP_TUNNEL_NIC_ENTRY_OP_FAIL;
}

static void
udp_tunnel_nic_device_sync_one(struct net_device *dev,
			       struct udp_tunnel_nic *utn,
			       unsigned int table, unsigned int idx)
{
	struct udp_tunnel_nic_table_entry *entry;
	struct udp_tunnel_info ti;
	int err;

	entry = &utn->entries[table][idx];
	if (!udp_tunnel_nic_entry_is_queued(entry))
		return;

	udp_tunnel_nic_ti_from_entry(entry, &ti);
	if (entry->flags & UDP_TUNNEL_NIC_ENTRY_ADD)
		err = dev->udp_tunnel_nic_info->set_port(dev, table, idx, &ti);
	else
		err = dev->udp_tunnel_nic_info->unset_port(dev, table, idx,
							   &ti);
	udp_tunnel_nic_entry_update_done(entry, err);

	if (err)
		netdev_warn(dev,
			    "UDP tunnel port sync failed port %d type %s: %d\n",
			    be16_to_cpu(entry->port),
			    udp_tunnel_nic_tunnel_type_name(entry->type),
			    err);
}

static void
udp_tunnel_nic_device_sync_by_port(struct net_device *dev,
				   struct udp_tunnel_nic *utn)
{
	const struct udp_tunnel_nic_info *info = dev->udp_tunnel_nic_info;
	unsigned int i, j;

	for (i = 0; i < utn->n_tables; i++)
		for (j = 0; j < info->tables[i].n_entries; j++)
			udp_tunnel_nic_device_sync_one(dev, utn, i, j);
}

static void
udp_tunnel_nic_device_sync_by_table(struct net_device *dev,
				    struct udp_tunnel_nic *utn)
{
	const struct udp_tunnel_nic_info *info = dev->udp_tunnel_nic_info;
	unsigned int i, j;
	int err;

	for (i = 0; i < utn->n_tables; i++) {
		/* Find something that needs sync in this table */
		for (j = 0; j < info->tables[i].n_entries; j++)
			if (udp_tunnel_nic_entry_is_queued(&utn->entries[i][j]))
				break;
		if (j == info->tables[i].n_entries)
			continue;

		err = info->sync_table(dev, i);
		if (err)
			netdev_warn(dev, "UDP tunnel port sync failed for table %d: %d\n",
				    i, err);

		for (j = 0; j < info->tables[i].n_entries; j++) {
			struct udp_tunnel_nic_table_entry *entry;

			entry = &utn->entries[i][j];
			if (udp_tunnel_nic_entry_is_queued(entry))
				udp_tunnel_nic_entry_update_done(entry, err);
		}
	}
}

static void
__udp_tunnel_nic_device_sync(struct net_device *dev, struct udp_tunnel_nic *utn)
{
	if (!utn->need_sync)
		return;

	if (dev->udp_tunnel_nic_info->sync_table)
		udp_tunnel_nic_device_sync_by_table(dev, utn);
	else
		udp_tunnel_nic_device_sync_by_port(dev, utn);

	utn->need_sync = 0;
	/* Can't replay directly here, in case we come from the tunnel driver's
	 * notification - trying to replay may deadlock inside tunnel driver.
	 */
	utn->need_replay = udp_tunnel_nic_should_replay(dev, utn);
}

static void
udp_tunnel_nic_device_sync(struct net_device *dev, struct udp_tunnel_nic *utn)
{
	const struct udp_tunnel_nic_info *info = dev->udp_tunnel_nic_info;
	bool may_sleep;

	if (!utn->need_sync)
		return;

	/* Drivers which sleep in the callback need to update from
	 * the workqueue, if we come from the tunnel driver's notification.
	 */
	may_sleep = info->flags & UDP_TUNNEL_NIC_INFO_MAY_SLEEP;
	if (!may_sleep)
		__udp_tunnel_nic_device_sync(dev, utn);
	if (may_sleep || utn->need_replay) {
		queue_work(udp_tunnel_nic_workqueue, &utn->work);
		utn->work_pending = 1;
	}
}

static bool
udp_tunnel_nic_table_is_capable(const struct udp_tunnel_nic_table_info *table,
				struct udp_tunnel_info *ti)
{
	return table->tunnel_types & ti->type;
}

static bool
udp_tunnel_nic_is_capable(struct net_device *dev, struct udp_tunnel_nic *utn,
			  struct udp_tunnel_info *ti)
{
	const struct udp_tunnel_nic_info *info = dev->udp_tunnel_nic_info;
	unsigned int i;

	/* Special case IPv4-only NICs */
	if (info->flags & UDP_TUNNEL_NIC_INFO_IPV4_ONLY &&
	    ti->sa_family != AF_INET)
		return false;

	for (i = 0; i < utn->n_tables; i++)
		if (udp_tunnel_nic_table_is_capable(&info->tables[i], ti))
			return true;
	return false;
}

static int
udp_tunnel_nic_has_collision(struct net_device *dev, struct udp_tunnel_nic *utn,
			     struct udp_tunnel_info *ti)
{
	const struct udp_tunnel_nic_info *info = dev->udp_tunnel_nic_info;
	struct udp_tunnel_nic_table_entry *entry;
	unsigned int i, j;

	for (i = 0; i < utn->n_tables; i++)
		for (j = 0; j < info->tables[i].n_entries; j++) {
			entry =	&utn->entries[i][j];

			if (!udp_tunnel_nic_entry_is_free(entry) &&
			    entry->port == ti->port &&
			    entry->type != ti->type) {
				__set_bit(i, &utn->missed);
				return true;
			}
		}
	return false;
}

static void
udp_tunnel_nic_entry_adj(struct udp_tunnel_nic *utn,
			 unsigned int table, unsigned int idx, int use_cnt_adj)
{
	struct udp_tunnel_nic_table_entry *entry =  &utn->entries[table][idx];
	bool dodgy = entry->flags & UDP_TUNNEL_NIC_ENTRY_OP_FAIL;
	unsigned int from, to;

	/* If not going from used to unused or vice versa - all done.
	 * For dodgy entries make sure we try to sync again (queue the entry).
	 */
	entry->use_cnt += use_cnt_adj;
	if (!dodgy && !entry->use_cnt == !(entry->use_cnt - use_cnt_adj))
		return;

	/* Cancel the op before it was sent to the device, if possible,
	 * otherwise we'd need to take special care to issue commands
	 * in the same order the ports arrived.
	 */
	if (use_cnt_adj < 0) {
		from = UDP_TUNNEL_NIC_ENTRY_ADD;
		to = UDP_TUNNEL_NIC_ENTRY_DEL;
	} else {
		from = UDP_TUNNEL_NIC_ENTRY_DEL;
		to = UDP_TUNNEL_NIC_ENTRY_ADD;
	}

	if (entry->flags & from) {
		entry->flags &= ~from;
		if (!dodgy)
			return;
	}

	udp_tunnel_nic_entry_queue(utn, entry, to);
}

static bool
udp_tunnel_nic_entry_try_adj(struct udp_tunnel_nic *utn,
			     unsigned int table, unsigned int idx,
			     struct udp_tunnel_info *ti, int use_cnt_adj)
{
	struct udp_tunnel_nic_table_entry *entry =  &utn->entries[table][idx];

	if (udp_tunnel_nic_entry_is_free(entry) ||
	    entry->port != ti->port ||
	    entry->type != ti->type)
		return false;

	if (udp_tunnel_nic_entry_is_frozen(entry))
		return true;

	udp_tunnel_nic_entry_adj(utn, table, idx, use_cnt_adj);
	return true;
}

/* Try to find existing matching entry and adjust its use count, instead of
 * adding a new one. Returns true if entry was found. In case of delete the
 * entry may have gotten removed in the process, in which case it will be
 * queued for removal.
 */
static bool
udp_tunnel_nic_try_existing(struct net_device *dev, struct udp_tunnel_nic *utn,
			    struct udp_tunnel_info *ti, int use_cnt_adj)
{
	const struct udp_tunnel_nic_table_info *table;
	unsigned int i, j;

	for (i = 0; i < utn->n_tables; i++) {
		table = &dev->udp_tunnel_nic_info->tables[i];
		if (!udp_tunnel_nic_table_is_capable(table, ti))
			continue;

		for (j = 0; j < table->n_entries; j++)
			if (udp_tunnel_nic_entry_try_adj(utn, i, j, ti,
							 use_cnt_adj))
				return true;
	}

	return false;
}

static bool
udp_tunnel_nic_add_existing(struct net_device *dev, struct udp_tunnel_nic *utn,
			    struct udp_tunnel_info *ti)
{
	return udp_tunnel_nic_try_existing(dev, utn, ti, +1);
}

static bool
udp_tunnel_nic_del_existing(struct net_device *dev, struct udp_tunnel_nic *utn,
			    struct udp_tunnel_info *ti)
{
	return udp_tunnel_nic_try_existing(dev, utn, ti, -1);
}

static bool
udp_tunnel_nic_add_new(struct net_device *dev, struct udp_tunnel_nic *utn,
		       struct udp_tunnel_info *ti)
{
	const struct udp_tunnel_nic_table_info *table;
	unsigned int i, j;

	for (i = 0; i < utn->n_tables; i++) {
		table = &dev->udp_tunnel_nic_info->tables[i];
		if (!udp_tunnel_nic_table_is_capable(table, ti))
			continue;

		for (j = 0; j < table->n_entries; j++) {
			struct udp_tunnel_nic_table_entry *entry;

			entry = &utn->entries[i][j];
			if (!udp_tunnel_nic_entry_is_free(entry))
				continue;

			entry->port = ti->port;
			entry->type = ti->type;
			entry->use_cnt = 1;
			udp_tunnel_nic_entry_queue(utn, entry,
						   UDP_TUNNEL_NIC_ENTRY_ADD);
			return true;
		}

		/* The different table may still fit this port in, but there
		 * are no devices currently which have multiple tables accepting
		 * the same tunnel type, and false positives are okay.
		 */
		__set_bit(i, &utn->missed);
	}

	return false;
}

static void
__udp_tunnel_nic_add_port(struct net_device *dev, struct udp_tunnel_info *ti)
{
	const struct udp_tunnel_nic_info *info = dev->udp_tunnel_nic_info;
	struct udp_tunnel_nic *utn;

	utn = dev->udp_tunnel_nic;
	if (!utn)
		return;
	if (!netif_running(dev) && info->flags & UDP_TUNNEL_NIC_INFO_OPEN_ONLY)
		return;

	if (!udp_tunnel_nic_is_capable(dev, utn, ti))
		return;

	/* It may happen that a tunnel of one type is removed and different
	 * tunnel type tries to reuse its port before the device was informed.
	 * Rely on utn->missed to re-add this port later.
	 */
	if (udp_tunnel_nic_has_collision(dev, utn, ti))
		return;

	if (!udp_tunnel_nic_add_existing(dev, utn, ti))
		udp_tunnel_nic_add_new(dev, utn, ti);

	udp_tunnel_nic_device_sync(dev, utn);
}

static void
__udp_tunnel_nic_del_port(struct net_device *dev, struct udp_tunnel_info *ti)
{
	struct udp_tunnel_nic *utn;

	utn = dev->udp_tunnel_nic;
	if (!utn)
		return;

	if (!udp_tunnel_nic_is_capable(dev, utn, ti))
		return;

	udp_tunnel_nic_del_existing(dev, utn, ti);

	udp_tunnel_nic_device_sync(dev, utn);
}

static void __udp_tunnel_nic_reset_ntf(struct net_device *dev)
{
	const struct udp_tunnel_nic_info *info = dev->udp_tunnel_nic_info;
	struct udp_tunnel_nic *utn;
	unsigned int i, j;

	ASSERT_RTNL();

	utn = dev->udp_tunnel_nic;
	if (!utn)
		return;

	utn->need_sync = false;
	for (i = 0; i < utn->n_tables; i++)
		for (j = 0; j < info->tables[i].n_entries; j++) {
			struct udp_tunnel_nic_table_entry *entry;

			entry = &utn->entries[i][j];

			entry->flags &= ~(UDP_TUNNEL_NIC_ENTRY_DEL |
					  UDP_TUNNEL_NIC_ENTRY_OP_FAIL);
			/* We don't release rtnl across ops */
			WARN_ON(entry->flags & UDP_TUNNEL_NIC_ENTRY_FROZEN);
			if (!entry->use_cnt)
				continue;

			udp_tunnel_nic_entry_queue(utn, entry,
						   UDP_TUNNEL_NIC_ENTRY_ADD);
		}

	__udp_tunnel_nic_device_sync(dev, utn);
}

static size_t
__udp_tunnel_nic_dump_size(struct net_device *dev, unsigned int table)
{
	const struct udp_tunnel_nic_info *info = dev->udp_tunnel_nic_info;
	struct udp_tunnel_nic *utn;
	unsigned int j;
	size_t size;

	utn = dev->udp_tunnel_nic;
	if (!utn)
		return 0;

	size = 0;
	for (j = 0; j < info->tables[table].n_entries; j++) {
		if (!udp_tunnel_nic_entry_is_present(&utn->entries[table][j]))
			continue;

		size += nla_total_size(0) +		 /* _TABLE_ENTRY */
			nla_total_size(sizeof(__be16)) + /* _ENTRY_PORT */
			nla_total_size(sizeof(u32));	 /* _ENTRY_TYPE */
	}

	return size;
}

static int
__udp_tunnel_nic_dump_write(struct net_device *dev, unsigned int table,
			    struct sk_buff *skb)
{
	const struct udp_tunnel_nic_info *info = dev->udp_tunnel_nic_info;
	struct udp_tunnel_nic *utn;
	struct nlattr *nest;
	unsigned int j;

	utn = dev->udp_tunnel_nic;
	if (!utn)
		return 0;

	for (j = 0; j < info->tables[table].n_entries; j++) {
		if (!udp_tunnel_nic_entry_is_present(&utn->entries[table][j]))
			continue;

		nest = nla_nest_start(skb, ETHTOOL_A_TUNNEL_UDP_TABLE_ENTRY);

		if (nla_put_be16(skb, ETHTOOL_A_TUNNEL_UDP_ENTRY_PORT,
				 utn->entries[table][j].port) ||
		    nla_put_u32(skb, ETHTOOL_A_TUNNEL_UDP_ENTRY_TYPE,
				ilog2(utn->entries[table][j].type)))
			goto err_cancel;

		nla_nest_end(skb, nest);
	}

	return 0;

err_cancel:
	nla_nest_cancel(skb, nest);
	return -EMSGSIZE;
}

static const struct udp_tunnel_nic_ops __udp_tunnel_nic_ops = {
	.get_port	= __udp_tunnel_nic_get_port,
	.set_port_priv	= __udp_tunnel_nic_set_port_priv,
	.add_port	= __udp_tunnel_nic_add_port,
	.del_port	= __udp_tunnel_nic_del_port,
	.reset_ntf	= __udp_tunnel_nic_reset_ntf,
	.dump_size	= __udp_tunnel_nic_dump_size,
	.dump_write	= __udp_tunnel_nic_dump_write,
};

static void
udp_tunnel_nic_flush(struct net_device *dev, struct udp_tunnel_nic *utn)
{
	const struct udp_tunnel_nic_info *info = dev->udp_tunnel_nic_info;
	unsigned int i, j;

	for (i = 0; i < utn->n_tables; i++)
		for (j = 0; j < info->tables[i].n_entries; j++) {
			int adj_cnt = -utn->entries[i][j].use_cnt;

			if (adj_cnt)
				udp_tunnel_nic_entry_adj(utn, i, j, adj_cnt);
		}

	__udp_tunnel_nic_device_sync(dev, utn);

	for (i = 0; i < utn->n_tables; i++)
		memset(utn->entries[i], 0, array_size(info->tables[i].n_entries,
						      sizeof(**utn->entries)));
	WARN_ON(utn->need_sync);
	utn->need_replay = 0;
}

static void
udp_tunnel_nic_replay(struct net_device *dev, struct udp_tunnel_nic *utn)
{
	const struct udp_tunnel_nic_info *info = dev->udp_tunnel_nic_info;
	unsigned int i, j;

	/* Freeze all the ports we are already tracking so that the replay
	 * does not double up the refcount.
	 */
	for (i = 0; i < utn->n_tables; i++)
		for (j = 0; j < info->tables[i].n_entries; j++)
			udp_tunnel_nic_entry_freeze_used(&utn->entries[i][j]);
	utn->missed = 0;
	utn->need_replay = 0;

	udp_tunnel_get_rx_info(dev);

	for (i = 0; i < utn->n_tables; i++)
		for (j = 0; j < info->tables[i].n_entries; j++)
			udp_tunnel_nic_entry_unfreeze(&utn->entries[i][j]);
}

static void udp_tunnel_nic_device_sync_work(struct work_struct *work)
{
	struct udp_tunnel_nic *utn =
		container_of(work, struct udp_tunnel_nic, work);

	rtnl_lock();
	utn->work_pending = 0;
	__udp_tunnel_nic_device_sync(utn->dev, utn);

	if (utn->need_replay)
		udp_tunnel_nic_replay(utn->dev, utn);
	rtnl_unlock();
}

static struct udp_tunnel_nic *
udp_tunnel_nic_alloc(const struct udp_tunnel_nic_info *info,
		     unsigned int n_tables)
{
	struct udp_tunnel_nic *utn;
	unsigned int i;

	utn = kzalloc(sizeof(*utn), GFP_KERNEL);
	if (!utn)
		return NULL;
	utn->n_tables = n_tables;
	INIT_WORK(&utn->work, udp_tunnel_nic_device_sync_work);

	utn->entries = kmalloc_array(n_tables, sizeof(void *), GFP_KERNEL);
	if (!utn->entries)
		goto err_free_utn;

	for (i = 0; i < n_tables; i++) {
		utn->entries[i] = kcalloc(info->tables[i].n_entries,
					  sizeof(*utn->entries[i]), GFP_KERNEL);
		if (!utn->entries[i])
			goto err_free_prev_entries;
	}

	return utn;

err_free_prev_entries:
	while (i--)
		kfree(utn->entries[i]);
	kfree(utn->entries);
err_free_utn:
	kfree(utn);
	return NULL;
}

static int udp_tunnel_nic_register(struct net_device *dev)
{
	const struct udp_tunnel_nic_info *info = dev->udp_tunnel_nic_info;
	struct udp_tunnel_nic *utn;
	unsigned int n_tables, i;

	BUILD_BUG_ON(sizeof(utn->missed) * BITS_PER_BYTE <
		     UDP_TUNNEL_NIC_MAX_TABLES);

	if (WARN_ON(!info->set_port != !info->unset_port) ||
	    WARN_ON(!info->set_port == !info->sync_table) ||
	    WARN_ON(!info->tables[0].n_entries))
		return -EINVAL;

	n_tables = 1;
	for (i = 1; i < UDP_TUNNEL_NIC_MAX_TABLES; i++) {
		if (!info->tables[i].n_entries)
			continue;

		n_tables++;
		if (WARN_ON(!info->tables[i - 1].n_entries))
			return -EINVAL;
	}

	utn = udp_tunnel_nic_alloc(info, n_tables);
	if (!utn)
		return -ENOMEM;

	utn->dev = dev;
	dev_hold(dev);
	dev->udp_tunnel_nic = utn;

	if (!(info->flags & UDP_TUNNEL_NIC_INFO_OPEN_ONLY))
		udp_tunnel_get_rx_info(dev);

	return 0;
}

static void
udp_tunnel_nic_unregister(struct net_device *dev, struct udp_tunnel_nic *utn)
{
	unsigned int i;

	/* Flush before we check work, so we don't waste time adding entries
	 * from the work which we will boot immediately.
	 */
	udp_tunnel_nic_flush(dev, utn);

	/* Wait for the work to be done using the state, netdev core will
	 * retry unregister until we give up our reference on this device.
	 */
	if (utn->work_pending)
		return;

	for (i = 0; i < utn->n_tables; i++)
		kfree(utn->entries[i]);
	kfree(utn->entries);
	kfree(utn);
	dev->udp_tunnel_nic = NULL;
	dev_put(dev);
}

static int
udp_tunnel_nic_netdevice_event(struct notifier_block *unused,
			       unsigned long event, void *ptr)
{
	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
	const struct udp_tunnel_nic_info *info;
	struct udp_tunnel_nic *utn;

	info = dev->udp_tunnel_nic_info;
	if (!info)
		return NOTIFY_DONE;

	if (event == NETDEV_REGISTER) {
		int err;

		err = udp_tunnel_nic_register(dev);
		if (err)
			netdev_WARN(dev, "failed to register for UDP tunnel offloads: %d", err);
		return notifier_from_errno(err);
	}
	/* All other events will need the udp_tunnel_nic state */
	utn = dev->udp_tunnel_nic;
	if (!utn)
		return NOTIFY_DONE;

	if (event == NETDEV_UNREGISTER) {
		udp_tunnel_nic_unregister(dev, utn);
		return NOTIFY_OK;
	}

	/* All other events only matter if NIC has to be programmed open */
	if (!(info->flags & UDP_TUNNEL_NIC_INFO_OPEN_ONLY))
		return NOTIFY_DONE;

	if (event == NETDEV_UP) {
		WARN_ON(!udp_tunnel_nic_is_empty(dev, utn));
		udp_tunnel_get_rx_info(dev);
		return NOTIFY_OK;
	}
	if (event == NETDEV_GOING_DOWN) {
		udp_tunnel_nic_flush(dev, utn);
		return NOTIFY_OK;
	}

	return NOTIFY_DONE;
}

static struct notifier_block udp_tunnel_nic_notifier_block __read_mostly = {
	.notifier_call = udp_tunnel_nic_netdevice_event,
};

static int __init udp_tunnel_nic_init_module(void)
{
	int err;

	udp_tunnel_nic_workqueue = alloc_workqueue("udp_tunnel_nic", 0, 0);
	if (!udp_tunnel_nic_workqueue)
		return -ENOMEM;

	rtnl_lock();
	udp_tunnel_nic_ops = &__udp_tunnel_nic_ops;
	rtnl_unlock();

	err = register_netdevice_notifier(&udp_tunnel_nic_notifier_block);
	if (err)
		goto err_unset_ops;

	return 0;

err_unset_ops:
	rtnl_lock();
	udp_tunnel_nic_ops = NULL;
	rtnl_unlock();
	destroy_workqueue(udp_tunnel_nic_workqueue);
	return err;
}
late_initcall(udp_tunnel_nic_init_module);

static void __exit udp_tunnel_nic_cleanup_module(void)
{
	unregister_netdevice_notifier(&udp_tunnel_nic_notifier_block);

	rtnl_lock();
	udp_tunnel_nic_ops = NULL;
	rtnl_unlock();

	destroy_workqueue(udp_tunnel_nic_workqueue);
}
module_exit(udp_tunnel_nic_cleanup_module);

MODULE_LICENSE("GPL");