summaryrefslogtreecommitdiffstats
path: root/sysdep/linux/netlink/netlink.c
blob: 38d00afe7343e0c8edef32f2d32280f080125656 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
/*
 *	BIRD -- Linux Netlink Interface
 *
 *	(c) 1999--2000 Martin Mares <mj@ucw.cz>
 *
 *	Can be freely distributed and used under the terms of the GNU GPL.
 */

#include <stdio.h>
#include <fcntl.h>
#include <net/if.h>
#include <sys/socket.h>
#include <sys/uio.h>
#include <errno.h>

#undef LOCAL_DEBUG

#include "nest/bird.h"
#include "nest/route.h"
#include "nest/protocol.h"
#include "nest/iface.h"
#include "lib/timer.h"
#include "lib/unix.h"
#include "lib/krt.h"
#include "lib/socket.h"
#include "lib/string.h"
#include "conf/conf.h"

#include <asm/types.h>
#include <linux/netlink.h>
#include <linux/rtnetlink.h>

#ifndef MSG_TRUNC			/* Hack: Several versions of glibc miss this one :( */
#define MSG_TRUNC 0x20
#endif

/*
 *	Synchronous Netlink interface
 */

static int nl_sync_fd = -1;		/* Unix socket for synchronous netlink actions */
static u32 nl_sync_seq;			/* Sequence number of last request sent */

static byte *nl_rx_buffer;		/* Receive buffer */
#define NL_RX_SIZE 8192

static struct nlmsghdr *nl_last_hdr;	/* Recently received packet */
static unsigned int nl_last_size;

static void
nl_open(void)
{
  if (nl_sync_fd < 0)
    {
      nl_sync_fd = socket(PF_NETLINK, SOCK_RAW, NETLINK_ROUTE);
      if (nl_sync_fd < 0)
	die("Unable to open rtnetlink socket: %m");
      nl_sync_seq = now;
      nl_rx_buffer = xmalloc(NL_RX_SIZE);
    }
}

static void
nl_send(struct nlmsghdr *nh)
{
  struct sockaddr_nl sa;

  memset(&sa, 0, sizeof(sa));
  sa.nl_family = AF_NETLINK;
  nh->nlmsg_pid = 0;
  nh->nlmsg_seq = ++nl_sync_seq;
  if (sendto(nl_sync_fd, nh, nh->nlmsg_len, 0, (struct sockaddr *)&sa, sizeof(sa)) < 0)
    die("rtnetlink sendto: %m");
  nl_last_hdr = NULL;
}

static void
nl_request_dump(int cmd)
{
  struct {
    struct nlmsghdr nh;
    struct rtgenmsg g;
  } req;
  req.nh.nlmsg_type = cmd;
  req.nh.nlmsg_len = sizeof(req);
  req.nh.nlmsg_flags = NLM_F_REQUEST | NLM_F_DUMP;
  req.g.rtgen_family = BIRD_PF;
  nl_send(&req.nh);
}

static struct nlmsghdr *
nl_get_reply(void)
{
  for(;;)
    {
      if (!nl_last_hdr)
	{
	  struct iovec iov = { nl_rx_buffer, NL_RX_SIZE };
	  struct sockaddr_nl sa;
	  struct msghdr m = { (struct sockaddr *) &sa, sizeof(sa), &iov, 1, NULL, 0, 0 };
	  int x = recvmsg(nl_sync_fd, &m, 0);
	  if (x < 0)
	    die("nl_get_reply: %m");
	  if (sa.nl_pid)		/* It isn't from the kernel */
	    {
	      DBG("Non-kernel packet\n");
	      continue;
	    }
	  nl_last_size = x;
	  nl_last_hdr = (void *) nl_rx_buffer;
	  if (m.msg_flags & MSG_TRUNC)
	    bug("nl_get_reply: got truncated reply which should be impossible");
	}
      if (NLMSG_OK(nl_last_hdr, nl_last_size))
	{
	  struct nlmsghdr *h = nl_last_hdr;
	  nl_last_hdr = NLMSG_NEXT(h, nl_last_size);
	  if (h->nlmsg_seq != nl_sync_seq)
	    {
	      log(L_WARN "nl_get_reply: Ignoring out of sequence netlink packet (%x != %x)",
		  h->nlmsg_seq, nl_sync_seq);
	      continue;
	    }
	  return h;
	}
      if (nl_last_size)
	log(L_WARN "nl_get_reply: Found packet remnant of size %d", nl_last_size);
      nl_last_hdr = NULL;
    }
}

static int
nl_error(struct nlmsghdr *h)
{
  struct nlmsgerr *e;
  int ec;

  if (h->nlmsg_len < NLMSG_LENGTH(sizeof(struct nlmsgerr)))
    {
      log(L_WARN "Netlink: Truncated error message received");
      return ENOBUFS;
    }
  e = (struct nlmsgerr *) NLMSG_DATA(h);
  ec = -e->error;
  if (ec)
    log(L_WARN "Netlink: %s", strerror(ec));
  return ec;
}

static struct nlmsghdr *
nl_get_scan(void)
{
  struct nlmsghdr *h = nl_get_reply();

  if (h->nlmsg_type == NLMSG_DONE)
    return NULL;
  if (h->nlmsg_type == NLMSG_ERROR)
    {
      nl_error(h);
      return NULL;
    }
  return h;
}

static int
nl_exchange(struct nlmsghdr *pkt)
{
  struct nlmsghdr *h;

  nl_send(pkt);
  for(;;)
    {
      h = nl_get_reply();
      if (h->nlmsg_type == NLMSG_ERROR)
	break;
      log(L_WARN "nl_exchange: Unexpected reply received");
    }
  return nl_error(h);
}

/*
 *	Netlink attributes
 */

static int nl_attr_len;

static void *
nl_checkin(struct nlmsghdr *h, int lsize)
{
  nl_attr_len = h->nlmsg_len - NLMSG_LENGTH(lsize);
  if (nl_attr_len < 0)
    {
      log(L_ERR "nl_checkin: underrun by %d bytes", -nl_attr_len);
      return NULL;
    }
  return NLMSG_DATA(h);
}

static int
nl_parse_attrs(struct rtattr *a, struct rtattr **k, int ksize)
{
  int max = ksize / sizeof(struct rtattr *);
  bzero(k, ksize);
  while (RTA_OK(a, nl_attr_len))
    {
      if (a->rta_type < max)
	k[a->rta_type] = a;
      a = RTA_NEXT(a, nl_attr_len);
    }
  if (nl_attr_len)
    {
      log(L_ERR "nl_parse_attrs: remnant of size %d", nl_attr_len);
      return 0;
    }
  else
    return 1;
}

static void
nl_add_attr_u32(struct nlmsghdr *h, unsigned maxsize, int code, u32 data)
{
  unsigned len = RTA_LENGTH(4);
  struct rtattr *a;

  if (NLMSG_ALIGN(h->nlmsg_len) + len > maxsize)
    bug("nl_add_attr32: packet buffer overflow");
  a = (struct rtattr *)((char *)h + NLMSG_ALIGN(h->nlmsg_len));
  a->rta_type = code;
  a->rta_len = len;
  memcpy(RTA_DATA(a), &data, 4);
  h->nlmsg_len = NLMSG_ALIGN(h->nlmsg_len) + len;
}

static void
nl_add_attr_ipa(struct nlmsghdr *h, unsigned maxsize, int code, ip_addr ipa)
{
  unsigned len = RTA_LENGTH(sizeof(ipa));
  struct rtattr *a;

  if (NLMSG_ALIGN(h->nlmsg_len) + len > maxsize)
    bug("nl_add_attr_ipa: packet buffer overflow");
  a = (struct rtattr *)((char *)h + NLMSG_ALIGN(h->nlmsg_len));
  a->rta_type = code;
  a->rta_len = len;
  ipa_hton(ipa);
  memcpy(RTA_DATA(a), &ipa, sizeof(ipa));
  h->nlmsg_len = NLMSG_ALIGN(h->nlmsg_len) + len;
}

/*
 *	Scanning of interfaces
 */

static void
nl_parse_link(struct nlmsghdr *h, int scan)
{
  struct ifinfomsg *i;
  struct rtattr *a[IFLA_STATS+1];
  int new = h->nlmsg_type == RTM_NEWLINK;
  struct iface f;
  struct iface *ifi;
  char *name;
  u32 mtu;
  unsigned int fl;

  if (!(i = nl_checkin(h, sizeof(*i))) || !nl_parse_attrs(IFLA_RTA(i), a, sizeof(a)))
    return;
  if (!a[IFLA_IFNAME] || RTA_PAYLOAD(a[IFLA_IFNAME]) < 2 ||
      !a[IFLA_MTU] || RTA_PAYLOAD(a[IFLA_MTU]) != 4)
    {
      if (scan)
        log(L_ERR "nl_parse_link: Malformed message received");
      return;
    }
  name = RTA_DATA(a[IFLA_IFNAME]);
  memcpy(&mtu, RTA_DATA(a[IFLA_MTU]), sizeof(u32));

  ifi = if_find_by_index(i->ifi_index);
  if (!new)
    {
      DBG("KIF: IF%d(%s) goes down\n", i->ifi_index, name);
      if (ifi && !scan)
	{
	  memcpy(&f, ifi, sizeof(struct iface));
	  f.flags |= IF_ADMIN_DOWN;
	  if_update(&f);
	}
    }
  else
    {
      DBG("KIF: IF%d(%s) goes up (mtu=%d,flg=%x)\n", i->ifi_index, name, mtu, i->ifi_flags);
      if (ifi)
	memcpy(&f, ifi, sizeof(f));
      else
	{
	  bzero(&f, sizeof(f));
	  f.index = i->ifi_index;
	}
      strncpy(f.name, RTA_DATA(a[IFLA_IFNAME]), sizeof(f.name)-1);
      f.mtu = mtu;
      f.flags = 0;
      fl = i->ifi_flags;
      if (fl & IFF_UP)
	f.flags |= IF_LINK_UP;
      if (fl & IFF_LOOPBACK)		/* Loopback */
	f.flags |= IF_MULTIACCESS | IF_LOOPBACK | IF_IGNORE;
      else if (fl & IFF_POINTOPOINT)	/* PtP */
	f.flags |= IF_MULTICAST;
      else if (fl & IFF_BROADCAST)	/* Broadcast */
	f.flags |= IF_MULTIACCESS | IF_BROADCAST | IF_MULTICAST;
      else
	f.flags |= IF_MULTIACCESS;	/* NBMA */
      if_update(&f);
    }
}

static void
nl_parse_addr(struct nlmsghdr *h)
{
  struct ifaddrmsg *i;
  struct rtattr *a[IFA_ANYCAST+1];
  int new = h->nlmsg_type == RTM_NEWADDR;
  struct ifa ifa;
  struct iface *ifi;
  int scope;

  if (!(i = nl_checkin(h, sizeof(*i))) || !nl_parse_attrs(IFA_RTA(i), a, sizeof(a)))
    return;
  if (i->ifa_family != BIRD_AF)
    return;
  if (!a[IFA_ADDRESS] || RTA_PAYLOAD(a[IFA_ADDRESS]) != sizeof(ip_addr)
#ifdef IPV6
      || a[IFA_LOCAL] && RTA_PAYLOAD(a[IFA_LOCAL]) != sizeof(ip_addr)
#else
      || !a[IFA_LOCAL] || RTA_PAYLOAD(a[IFA_LOCAL]) != sizeof(ip_addr)
      || (a[IFA_BROADCAST] && RTA_PAYLOAD(a[IFA_BROADCAST]) != sizeof(ip_addr))
#endif
      )
    {
      log(L_ERR "nl_parse_addr: Malformed message received");
      return;
    }

  ifi = if_find_by_index(i->ifa_index);
  if (!ifi)
    {
      log(L_ERR "KIF: Received address message for unknown interface %d", i->ifa_index);
      return;
    }

  bzero(&ifa, sizeof(ifa));
  ifa.iface = ifi;
  if (i->ifa_flags & IFA_F_SECONDARY)
    ifa.flags |= IA_SECONDARY;

  /* IFA_LOCAL can be unset for IPv6 interfaces */
  memcpy(&ifa.ip, RTA_DATA(a[IFA_LOCAL] ? : a[IFA_ADDRESS]), sizeof(ifa.ip));
  ipa_ntoh(ifa.ip);
  ifa.pxlen = i->ifa_prefixlen;
  if (i->ifa_prefixlen > BITS_PER_IP_ADDRESS ||
      i->ifa_prefixlen == BITS_PER_IP_ADDRESS - 1)
    {
      log(L_ERR "KIF: Invalid prefix length for interface %s: %d", ifi->name, i->ifa_prefixlen);
      new = 0;
    }
  if (i->ifa_prefixlen == BITS_PER_IP_ADDRESS)
    {
      ifa.flags |= IA_UNNUMBERED;
      memcpy(&ifa.opposite, RTA_DATA(a[IFA_ADDRESS]), sizeof(ifa.opposite));
      ipa_ntoh(ifa.opposite);
      ifa.prefix = ifa.brd = ifa.opposite;
    }
  else
    {
      ip_addr netmask = ipa_mkmask(ifa.pxlen);
      ip_addr xbrd;
      ifa.prefix = ipa_and(ifa.ip, netmask);
      ifa.brd = ipa_or(ifa.ip, ipa_not(netmask));
#ifndef IPV6
      if (i->ifa_prefixlen == BITS_PER_IP_ADDRESS - 2)
	ifa.opposite = ipa_opposite(ifa.ip, i->ifa_prefixlen);
      if ((ifi->flags & IF_BROADCAST) && a[IFA_BROADCAST])
	{
	  memcpy(&xbrd, RTA_DATA(a[IFA_BROADCAST]), sizeof(xbrd));
	  ipa_ntoh(xbrd);
	  if (ipa_equal(xbrd, ifa.prefix) || ipa_equal(xbrd, ifa.brd))
	    ifa.brd = xbrd;
	  else if (ifi->flags & IF_TMP_DOWN) /* Complain only during the first scan */
	    log(L_ERR "KIF: Invalid broadcast address %I for %s", xbrd, ifi->name);
	}
#endif
    }

  scope = ipa_classify(ifa.ip);
  if (scope < 0)
    {
      log(L_ERR "KIF: Invalid interface address %I for %s", ifa.ip, ifi->name);
      return;
    }
  ifa.scope = scope & IADDR_SCOPE_MASK;

  DBG("KIF: IF%d(%s): %s IPA %I, flg %x, net %I/%d, brd %I, opp %I\n",
      ifi->index, ifi->name,
      new ? "added" : "removed",
      ifa.ip, ifa.flags, ifa.prefix, ifa.pxlen, ifa.brd, ifa.opposite);
  if (new)
    ifa_update(&ifa);
  else
    ifa_delete(&ifa);
}

void
krt_if_scan(struct kif_proto *p UNUSED)
{
  struct nlmsghdr *h;

  if_start_update();

  nl_request_dump(RTM_GETLINK);
  while (h = nl_get_scan())
    if (h->nlmsg_type == RTM_NEWLINK || h->nlmsg_type == RTM_DELLINK)
      nl_parse_link(h, 1);
    else
      log(L_DEBUG "nl_scan_ifaces: Unknown packet received (type=%d)", h->nlmsg_type);

  nl_request_dump(RTM_GETADDR);
  while (h = nl_get_scan())
    if (h->nlmsg_type == RTM_NEWADDR || h->nlmsg_type == RTM_DELADDR)
      nl_parse_addr(h);
    else
      log(L_DEBUG "nl_scan_ifaces: Unknown packet received (type=%d)", h->nlmsg_type);

  if_end_update();
}

/*
 *	Routes
 */

static struct krt_proto *nl_table_map[NL_NUM_TABLES];

int
krt_capable(rte *e)
{
  rta *a = e->attrs;

  if (a->cast != RTC_UNICAST
#if 0
      && a->cast != RTC_ANYCAST
#endif
      )
    return 0;
  if (a->source == RTS_DEVICE)	/* Kernel takes care of device routes itself */
    return 0;
  switch (a->dest)
    {
    case RTD_ROUTER:
    case RTD_DEVICE:
    case RTD_BLACKHOLE:
    case RTD_UNREACHABLE:
    case RTD_PROHIBIT:
      break;
    default:
      return 0;
    }
  return 1;
}

static void
nl_send_route(struct krt_proto *p, rte *e, int new)
{
  net *net = e->net;
  rta *a = e->attrs;
  struct {
    struct nlmsghdr h;
    struct rtmsg r;
    char buf[128];
  } r;

  DBG("nl_send_route(%I/%d,new=%d)\n", net->n.prefix, net->n.pxlen, new);

  bzero(&r.h, sizeof(r.h));
  bzero(&r.r, sizeof(r.r));
  r.h.nlmsg_type = new ? RTM_NEWROUTE : RTM_DELROUTE;
  r.h.nlmsg_len = NLMSG_LENGTH(sizeof(struct rtmsg));
  r.h.nlmsg_flags = NLM_F_REQUEST | NLM_F_ACK | (new ? NLM_F_CREATE|NLM_F_REPLACE : 0);

  r.r.rtm_family = BIRD_AF;
  r.r.rtm_dst_len = net->n.pxlen;
  r.r.rtm_tos = 0;
  r.r.rtm_table = KRT_CF->scan.table_id;
  r.r.rtm_protocol = RTPROT_BIRD;
  r.r.rtm_scope = RT_SCOPE_UNIVERSE;
  nl_add_attr_ipa(&r.h, sizeof(r), RTA_DST, net->n.prefix);
  switch (a->dest)
    {
    case RTD_ROUTER:
      r.r.rtm_type = RTN_UNICAST;
      nl_add_attr_ipa(&r.h, sizeof(r), RTA_GATEWAY, a->gw);
      break;
    case RTD_DEVICE:
      if (!a->iface)
	return;
      r.r.rtm_type = RTN_UNICAST;
      nl_add_attr_u32(&r.h, sizeof(r), RTA_OIF, a->iface->index);
      break;
    case RTD_BLACKHOLE:
      r.r.rtm_type = RTN_BLACKHOLE;
      break;
    case RTD_UNREACHABLE:
      r.r.rtm_type = RTN_UNREACHABLE;
      break;
    case RTD_PROHIBIT:
      r.r.rtm_type = RTN_PROHIBIT;
      break;
    default:
      bug("krt_capable inconsistent with nl_send_route");
    }

  nl_exchange(&r.h);
}

void
krt_set_notify(struct krt_proto *p, net *n UNUSED, rte *new, rte *old)
{
  if (old && new)
    {
      /*
       *  We should check whether priority and TOS is identical as well,
       *  but we don't use these and default value is always equal to default value. :-)
       */
      nl_send_route(p, new, 1);
    }
  else
    {
      if (old)
	nl_send_route(p, old, 0);

      if (new)
	nl_send_route(p, new, 1);
    }
}

static struct iface *
krt_temp_iface(struct krt_proto *p, unsigned index)
{
  struct iface *i, *j;

  WALK_LIST(i, p->scan.temp_ifs)
    if (i->index == index)
      return i;
  i = mb_allocz(p->p.pool, sizeof(struct iface));
  if (j = if_find_by_index(index))
    strcpy(i->name, j->name);
  else
    strcpy(i->name, "?");
  i->index = index;
  add_tail(&p->scan.temp_ifs, &i->n);
  return i;
}

static void
nl_parse_route(struct nlmsghdr *h, int scan)
{
  struct krt_proto *p;
  struct rtmsg *i;
  struct rtattr *a[RTA_CACHEINFO+1];
  int new = h->nlmsg_type == RTM_NEWROUTE;
  ip_addr dst;
  rta ra;
  rte *e;
  net *net;
  u32 oif;
  int src;

  if (!(i = nl_checkin(h, sizeof(*i))) || !nl_parse_attrs(RTM_RTA(i), a, sizeof(a)))
    return;
  if (i->rtm_family != BIRD_AF)
    return;
  if ((a[RTA_DST] && RTA_PAYLOAD(a[RTA_DST]) != sizeof(ip_addr)) ||
      (a[RTA_OIF] && RTA_PAYLOAD(a[RTA_OIF]) != 4) ||
      (a[RTA_PRIORITY] && RTA_PAYLOAD(a[RTA_PRIORITY]) != 4) ||
#ifdef IPV6
      (a[RTA_IIF] && RTA_PAYLOAD(a[RTA_IIF]) != 4) ||
#endif
      (a[RTA_GATEWAY] && RTA_PAYLOAD(a[RTA_GATEWAY]) != sizeof(ip_addr)))
    {
      log(L_ERR "nl_parse_route: Malformed message received");
      return;
    }

  p = nl_table_map[i->rtm_table];	/* Do we know this table? */
  if (!p)
    return;

#ifdef IPV6
  if (a[RTA_IIF])
    {
      DBG("KRT: Ignoring route with IIF set\n");
      return;
    }
#else
  if (i->rtm_tos != 0)			/* We don't support TOS */
    {
      DBG("KRT: Ignoring route with TOS %02x\n", i->rtm_tos);
      return;
    }
#endif

  if (scan && !new)
    {
      DBG("KRT: Ignoring route deletion\n");
      return;
    }

  if (a[RTA_DST])
    {
      memcpy(&dst, RTA_DATA(a[RTA_DST]), sizeof(dst));
      ipa_ntoh(dst);
    }
  else
    dst = IPA_NONE;
  if (a[RTA_OIF])
    memcpy(&oif, RTA_DATA(a[RTA_OIF]), sizeof(oif));
  else
    oif = ~0;

  DBG("Got %I/%d, type=%d, oif=%d, table=%d, prid=%d, proto=%s\n", dst, i->rtm_dst_len, i->rtm_type, oif, i->rtm_table, i->rtm_protocol, p->p.name);

  switch (i->rtm_protocol)
    {
    case RTPROT_REDIRECT:
      src = KRT_SRC_REDIRECT;
      break;
    case RTPROT_KERNEL:
      DBG("Route originated in kernel, ignoring\n");
      return;
    case RTPROT_BIRD:
#ifdef IPV6
    case RTPROT_BOOT:
      /* Current Linux kernels don't remember rtm_protocol for IPv6 routes and supply RTPROT_BOOT instead */
#endif
      if (!scan)
	{
	  DBG("Echo of our own route, ignoring\n");
	  return;
	}
      src = KRT_SRC_BIRD;
      break;
    default:
      src = KRT_SRC_ALIEN;
    }

  net = net_get(p->p.table, dst, i->rtm_dst_len);
  ra.proto = &p->p;
  ra.source = RTS_INHERIT;
  ra.scope = SCOPE_UNIVERSE;
  ra.cast = RTC_UNICAST;
  ra.flags = ra.aflags = 0;
  ra.from = IPA_NONE;
  ra.gw = IPA_NONE;
  ra.iface = NULL;
  ra.eattrs = NULL;

  switch (i->rtm_type)
    {
    case RTN_UNICAST:
      if (oif == ~0U)
	{
	  log(L_ERR "KRT: Mysterious route with no OIF (%I/%d)", net->n.prefix, net->n.pxlen);
	  return;
	}
      if (a[RTA_GATEWAY])
	{
	  neighbor *ng;
	  ra.dest = RTD_ROUTER;
	  memcpy(&ra.gw, RTA_DATA(a[RTA_GATEWAY]), sizeof(ra.gw));
	  ipa_ntoh(ra.gw);
	  ng = neigh_find(&p->p, &ra.gw, 0);
	  if (ng && ng->scope)
	    ra.iface = ng->iface;
	  else
	    /* FIXME: Remove this warning? Handle it somehow... */
	    log(L_WARN "Kernel told us to use non-neighbor %I for %I/%d", ra.gw, net->n.prefix, net->n.pxlen);
	}
      else
	{
	  ra.dest = RTD_DEVICE;
	  ra.iface = krt_temp_iface(p, oif);
	}
      break;
    case RTN_BLACKHOLE:
      ra.dest = RTD_BLACKHOLE;
      break;
    case RTN_UNREACHABLE:
      ra.dest = RTD_UNREACHABLE;
      break;
    case RTN_PROHIBIT:
      ra.dest = RTD_PROHIBIT;
      break;
    /* FIXME: What about RTN_THROW? */
    default:
      DBG("KRT: Ignoring route with type=%d\n", i->rtm_type);
      return;
    }

  if (i->rtm_scope != RT_SCOPE_UNIVERSE)
    {
      DBG("KRT: Ignoring route with scope=%d\n", i->rtm_scope);
      return;
    }

  e = rte_get_temp(&ra);
  e->net = net;
  e->u.krt.src = src;
  e->u.krt.proto = i->rtm_protocol;
  e->u.krt.type = i->rtm_type;
  if (a[RTA_PRIORITY])
    memcpy(&e->u.krt.metric, RTA_DATA(a[RTA_PRIORITY]), sizeof(e->u.krt.metric));
  else
    e->u.krt.metric = 0;
  if (scan)
    krt_got_route(p, e);
  else
    krt_got_route_async(p, e, new);
}

void
krt_scan_fire(struct krt_proto *p UNUSED)	/* CONFIG_ALL_TABLES_AT_ONCE => p is NULL */
{
  struct nlmsghdr *h;

  nl_request_dump(RTM_GETROUTE);
  while (h = nl_get_scan())
    if (h->nlmsg_type == RTM_NEWROUTE || h->nlmsg_type == RTM_DELROUTE)
      nl_parse_route(h, 1);
    else
      log(L_DEBUG "nl_scan_fire: Unknown packet received (type=%d)", h->nlmsg_type);
}

/*
 *	Asynchronous Netlink interface
 */

static sock *nl_async_sk;		/* BIRD socket for asynchronous notifications */
static byte *nl_async_rx_buffer;	/* Receive buffer */

static void
nl_async_msg(struct nlmsghdr *h)
{
  switch (h->nlmsg_type)
    {
    case RTM_NEWROUTE:
    case RTM_DELROUTE:
      DBG("KRT: Received async route notification (%d)\n", h->nlmsg_type);
      nl_parse_route(h, 0);
      break;
    case RTM_NEWLINK:
    case RTM_DELLINK:
      DBG("KRT: Received async link notification (%d)\n", h->nlmsg_type);
      nl_parse_link(h, 0);
      break;
    case RTM_NEWADDR:
    case RTM_DELADDR:
      DBG("KRT: Received async address notification (%d)\n", h->nlmsg_type);
      nl_parse_addr(h);
      break;
    default:
      DBG("KRT: Received unknown async notification (%d)\n", h->nlmsg_type);
    }
}

static int
nl_async_hook(sock *sk, int size UNUSED)
{
  struct iovec iov = { nl_async_rx_buffer, NL_RX_SIZE };
  struct sockaddr_nl sa;
  struct msghdr m = { (struct sockaddr *) &sa, sizeof(sa), &iov, 1, NULL, 0, 0 };
  struct nlmsghdr *h;
  int x;
  unsigned int len;

  nl_last_hdr = NULL;		/* Discard packets accidentally remaining in the rxbuf */
  x = recvmsg(sk->fd, &m, 0);
  if (x < 0)
    {
      if (errno == ENOBUFS)
	{
	  /*
	   *  Netlink reports some packets have been thrown away.
	   *  One day we might react to it by asking for route table
	   *  scan in near future.
	   */
	  return 1;	/* More data are likely to be ready */
	}
      else if (errno != EWOULDBLOCK)
	log(L_ERR "Netlink recvmsg: %m");
      return 0;
    }
  if (sa.nl_pid)		/* It isn't from the kernel */
    {
      DBG("Non-kernel packet\n");
      return 1;
    }
  h = (void *) nl_async_rx_buffer;
  len = x;
  if (m.msg_flags & MSG_TRUNC)
    {
      log(L_WARN "Netlink got truncated asynchronous message");
      return 1;
    }
  while (NLMSG_OK(h, len))
    {
      nl_async_msg(h);
      h = NLMSG_NEXT(h, len);
    }
  if (len)
    log(L_WARN "nl_async_hook: Found packet remnant of size %d", len);
  return 1;
}

static void
nl_open_async(void)
{
  sock *sk;
  struct sockaddr_nl sa;
  int fd;
  static int nl_open_tried = 0;

  if (nl_open_tried)
    return;
  nl_open_tried = 1;

  DBG("KRT: Opening async netlink socket\n");

  fd = socket(PF_NETLINK, SOCK_RAW, NETLINK_ROUTE);
  if (fd < 0)
    {
      log(L_ERR "Unable to open asynchronous rtnetlink socket: %m");
      return;
    }

  bzero(&sa, sizeof(sa));
  sa.nl_family = AF_NETLINK;
#ifdef IPV6
  sa.nl_groups = RTMGRP_LINK | RTMGRP_IPV6_IFADDR | RTMGRP_IPV6_ROUTE;
#else
  sa.nl_groups = RTMGRP_LINK | RTMGRP_IPV4_IFADDR | RTMGRP_IPV4_ROUTE;
#endif
  if (bind(fd, (struct sockaddr *) &sa, sizeof(sa)) < 0)
    {
      log(L_ERR "Unable to bind asynchronous rtnetlink socket: %m");
      return;
    }

  sk = nl_async_sk = sk_new(krt_pool);
  sk->type = SK_MAGIC;
  sk->rx_hook = nl_async_hook;
  sk->fd = fd;
  if (sk_open(sk))
    bug("Netlink: sk_open failed");

  if (!nl_async_rx_buffer)
    nl_async_rx_buffer = xmalloc(NL_RX_SIZE);
}

/*
 *	Interface to the UNIX krt module
 */

static u8 nl_cf_table[(NL_NUM_TABLES+7) / 8];

void
krt_scan_preconfig(struct config *c UNUSED)
{
  bzero(&nl_cf_table, sizeof(nl_cf_table));
}

void
krt_scan_postconfig(struct krt_config *x)
{
  int id = x->scan.table_id;

  if (nl_cf_table[id/8] & (1 << (id%8)))
    cf_error("Multiple kernel syncers defined for table #%d", id);
  nl_cf_table[id/8] |= (1 << (id%8));
}

void
krt_scan_construct(struct krt_config *x)
{
#ifndef IPV6
  x->scan.table_id = RT_TABLE_MAIN;
#else
  x->scan.table_id = 254;
#endif
}

void
krt_scan_start(struct krt_proto *p, int first)
{
  init_list(&p->scan.temp_ifs);
  nl_table_map[KRT_CF->scan.table_id] = p;
  if (first)
    {
      nl_open();
      nl_open_async();
    }
}

void
krt_scan_shutdown(struct krt_proto *p UNUSED, int last UNUSED)
{
}

void
krt_if_start(struct kif_proto *p UNUSED)
{
  nl_open();
  nl_open_async();
}