summaryrefslogtreecommitdiffstats
path: root/proto/bgp/bgp.c
blob: 59468c3f14286e78b1be8b4e33e6229118fe5875 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
/*
 *	BIRD -- The Border Gateway Protocol
 *
 *	(c) 2000 Martin Mares <mj@ucw.cz>
 *
 *	Can be freely distributed and used under the terms of the GNU GPL.
 */

/**
 * DOC: Border Gateway Protocol
 *
 * The BGP protocol is implemented in three parts: |bgp.c| which takes care of the
 * connection and most of the interface with BIRD core, |packets.c| handling
 * both incoming and outgoing BGP packets and |attrs.c| containing functions for
 * manipulation with BGP attribute lists.
 *
 * As opposed to the other existing routing daemons, BIRD has a sophisticated core
 * architecture which is able to keep all the information needed by BGP in the
 * primary routing table, therefore no complex data structures like a central
 * BGP table are needed. This increases memory footprint of a BGP router with
 * many connections, but not too much and, which is more important, it makes
 * BGP much easier to implement.
 *
 * Each instance of BGP (corresponding to a single BGP peer) is described by a &bgp_proto
 * structure to which are attached individual connections represented by &bgp_connection
 * (usually, there exists only one connection, but during BGP session setup, there
 * can be more of them). The connections are handled according to the BGP state machine
 * defined in the RFC with all the timers and all the parameters configurable.
 *
 * In incoming direction, we listen on the connection's socket and each time we receive
 * some input, we pass it to bgp_rx(). It decodes packet headers and the markers and
 * passes complete packets to bgp_rx_packet() which distributes the packet according
 * to its type.
 *
 * In outgoing direction, we gather all the routing updates and sort them to buckets
 * (&bgp_bucket) according to their attributes (we keep a hash table for fast comparison
 * of &rta's and a &fib which helps us to find if we already have another route for
 * the same destination queued for sending, so that we can replace it with the new one
 * immediately instead of sending both updates). There also exists a special bucket holding
 * all the route withdrawals which cannot be queued anywhere else as they don't have any
 * attributes. If we have any packet to send (due to either new routes or the connection
 * tracking code wanting to send a Open, Keepalive or Notification message), we call
 * bgp_schedule_packet() which sets the corresponding bit in a @packet_to_send
 * bit field in &bgp_conn and as soon as the transmit socket buffer becomes empty,
 * we call bgp_fire_tx(). It inspects state of all the packet type bits and calls
 * the corresponding bgp_create_xx() functions, eventually rescheduling the same packet
 * type if we have more data of the same type to send.
 *
 * The processing of attributes consists of two functions: bgp_decode_attrs() for checking
 * of the attribute blocks and translating them to the language of BIRD's extended attributes
 * and bgp_encode_attrs() which does the converse. Both functions are built around a
 * @bgp_attr_table array describing all important characteristics of all known attributes.
 * Unknown transitive attributes are attached to the route as %EAF_TYPE_OPAQUE byte streams.
 */

#undef LOCAL_DEBUG

#include "nest/bird.h"
#include "nest/iface.h"
#include "nest/protocol.h"
#include "nest/route.h"
#include "nest/locks.h"
#include "nest/cli.h"
#include "conf/conf.h"
#include "lib/socket.h"
#include "lib/resource.h"
#include "lib/string.h"

#include "bgp.h"

struct linpool *bgp_linpool;		/* Global temporary pool */
static sock *bgp_listen_sk;		/* Global listening socket */
static int bgp_counter;			/* Number of protocol instances using the listening socket */

static void bgp_close(struct bgp_proto *p, int apply_md5);
static void bgp_connect(struct bgp_proto *p);
static void bgp_active(struct bgp_proto *p);
static sock *bgp_setup_listen_sk(ip_addr addr, unsigned port, u32 flags);


/**
 * bgp_open - open a BGP instance
 * @p: BGP instance
 *
 * This function allocates and configures shared BGP resources.
 * Should be called as the last step during initialization
 * (when lock is acquired and neighbor is ready).
 * When error, state changed to PS_DOWN, -1 is returned and caller
 * should return immediately.
 */
static int
bgp_open(struct bgp_proto *p)
{
  struct config *cfg = p->cf->c.global;
  int errcode;

  bgp_counter++;

  if (!bgp_listen_sk)
    bgp_listen_sk = bgp_setup_listen_sk(cfg->listen_bgp_addr, cfg->listen_bgp_port, cfg->listen_bgp_flags);

  if (!bgp_listen_sk)
    {
      bgp_counter--;
      errcode = BEM_NO_SOCKET;
      goto err;
    }

  if (!bgp_linpool)
    bgp_linpool = lp_new(&root_pool, 4080);

  if (p->cf->password)
    {
      int rv = sk_set_md5_auth(bgp_listen_sk, p->cf->remote_ip, p->cf->password);
      if (rv < 0)
	{
	  bgp_close(p, 0);
	  errcode = BEM_INVALID_MD5;
	  goto err;
	}
    }

  return 0;

err:
  p->p.disabled = 1;
  bgp_store_error(p, NULL, BE_MISC, errcode);
  proto_notify_state(&p->p, PS_DOWN);
  return -1;
}

static void
bgp_startup(struct bgp_proto *p)
{
  BGP_TRACE(D_EVENTS, "Started");
  p->start_state = p->cf->capabilities ? BSS_CONNECT : BSS_CONNECT_NOCAP;

  if (!p->cf->passive)
    bgp_active(p);
}

static void
bgp_startup_timeout(timer *t)
{
  bgp_startup(t->data);
}


static void
bgp_initiate(struct bgp_proto *p)
{
  int rv = bgp_open(p);
  if (rv < 0)
    return;

  if (p->startup_delay)
    {
      BGP_TRACE(D_EVENTS, "Startup delayed by %d seconds", p->startup_delay);
      bgp_start_timer(p->startup_timer, p->startup_delay);
    }
  else
    bgp_startup(p);
}

/**
 * bgp_close - close a BGP instance
 * @p: BGP instance
 * @apply_md5: 0 to disable unsetting MD5 auth
 *
 * This function frees and deconfigures shared BGP resources.
 * @apply_md5 is set to 0 when bgp_close is called as a cleanup
 * from failed bgp_open().
 */
static void
bgp_close(struct bgp_proto *p, int apply_md5)
{
  ASSERT(bgp_counter);
  bgp_counter--;

  if (p->cf->password && apply_md5)
    sk_set_md5_auth(bgp_listen_sk, p->cf->remote_ip, NULL);

  if (!bgp_counter)
    {
      rfree(bgp_listen_sk);
      bgp_listen_sk = NULL;
      rfree(bgp_linpool);
      bgp_linpool = NULL;
    }
}

/**
 * bgp_start_timer - start a BGP timer
 * @t: timer
 * @value: time to fire (0 to disable the timer)
 *
 * This functions calls tm_start() on @t with time @value and the
 * amount of randomization suggested by the BGP standard. Please use
 * it for all BGP timers.
 */
void
bgp_start_timer(timer *t, int value)
{
  if (value)
    {
      /* The randomization procedure is specified in RFC 1771: 9.2.3.3 */
      t->randomize = value / 4;
      tm_start(t, value - t->randomize);
    }
  else
    tm_stop(t);
}

/**
 * bgp_close_conn - close a BGP connection
 * @conn: connection to close
 *
 * This function takes a connection described by the &bgp_conn structure,
 * closes its socket and frees all resources associated with it.
 */
void
bgp_close_conn(struct bgp_conn *conn)
{
  // struct bgp_proto *p = conn->bgp;

  DBG("BGP: Closing connection\n");
  conn->packets_to_send = 0;
  rfree(conn->connect_retry_timer);
  conn->connect_retry_timer = NULL;
  rfree(conn->keepalive_timer);
  conn->keepalive_timer = NULL;
  rfree(conn->hold_timer);
  conn->hold_timer = NULL;
  rfree(conn->sk);
  conn->sk = NULL;
  rfree(conn->tx_ev);
  conn->tx_ev = NULL;
}


/**
 * bgp_update_startup_delay - update a startup delay
 * @p: BGP instance
 *
 * This function updates a startup delay that is used to postpone next BGP connect.
 * It also handles disable_after_error and might stop BGP instance when error
 * happened and disable_after_error is on.
 *
 * It should be called when BGP protocol error happened.
 */
void
bgp_update_startup_delay(struct bgp_proto *p)
{
  struct bgp_config *cf = p->cf;

  DBG("BGP: Updating startup delay\n");

  if (p->last_proto_error && ((now - p->last_proto_error) >= (int) cf->error_amnesia_time))
    p->startup_delay = 0;

  p->last_proto_error = now;

  if (cf->disable_after_error)
    {
      p->startup_delay = 0;
      p->p.disabled = 1;
      return;
    }

  if (!p->startup_delay)
    p->startup_delay = cf->error_delay_time_min;
  else
    p->startup_delay = MIN(2 * p->startup_delay, cf->error_delay_time_max);
}

static void
bgp_graceful_close_conn(struct bgp_conn *conn, unsigned subcode)
{
  switch (conn->state)
    {
    case BS_IDLE:
    case BS_CLOSE:
      return;
    case BS_CONNECT:
    case BS_ACTIVE:
      bgp_conn_enter_idle_state(conn);
      return;
    case BS_OPENSENT:
    case BS_OPENCONFIRM:
    case BS_ESTABLISHED:
      bgp_error(conn, 6, subcode, NULL, 0);
      return;
    default:
      bug("bgp_graceful_close_conn: Unknown state %d", conn->state);
    }
}

static void
bgp_down(struct bgp_proto *p)
{
  if (p->start_state > BSS_PREPARE)
    bgp_close(p, 1);

  BGP_TRACE(D_EVENTS, "Down");
  proto_notify_state(&p->p, PS_DOWN);
}

static void
bgp_decision(void *vp)
{
  struct bgp_proto *p = vp;

  DBG("BGP: Decision start\n");
  if ((p->p.proto_state == PS_START)
      && (p->outgoing_conn.state == BS_IDLE)
      && (!p->cf->passive))
    bgp_active(p);

  if ((p->p.proto_state == PS_STOP)
      && (p->outgoing_conn.state == BS_IDLE)
      && (p->incoming_conn.state == BS_IDLE))
    bgp_down(p);
}

void
bgp_stop(struct bgp_proto *p, unsigned subcode)
{
  proto_notify_state(&p->p, PS_STOP);
  bgp_graceful_close_conn(&p->outgoing_conn, subcode);
  bgp_graceful_close_conn(&p->incoming_conn, subcode);
  ev_schedule(p->event);
}

static inline void
bgp_conn_set_state(struct bgp_conn *conn, unsigned new_state)
{
  if (conn->bgp->p.mrtdump & MD_STATES)
    mrt_dump_bgp_state_change(conn, conn->state, new_state);

  conn->state = new_state;
}

void
bgp_conn_enter_openconfirm_state(struct bgp_conn *conn)
{
  /* Really, most of the work is done in bgp_rx_open(). */
  bgp_conn_set_state(conn, BS_OPENCONFIRM);
}

void
bgp_conn_enter_established_state(struct bgp_conn *conn)
{
  struct bgp_proto *p = conn->bgp;
 
  BGP_TRACE(D_EVENTS, "BGP session established");
  DBG("BGP: UP!!!\n");

  /* For multi-hop BGP sessions */
  if (ipa_zero(p->source_addr))
    p->source_addr = conn->sk->saddr; 

  p->conn = conn;
  p->last_error_class = 0;
  p->last_error_code = 0;
  bgp_attr_init(conn->bgp);
  bgp_conn_set_state(conn, BS_ESTABLISHED);
  proto_notify_state(&p->p, PS_UP);
}

static void
bgp_conn_leave_established_state(struct bgp_proto *p)
{
  BGP_TRACE(D_EVENTS, "BGP session closed");
  p->conn = NULL;

  if (p->p.proto_state == PS_UP)
    bgp_stop(p, 0);
}

void
bgp_conn_enter_close_state(struct bgp_conn *conn)
{
  struct bgp_proto *p = conn->bgp;
  int os = conn->state;

  bgp_conn_set_state(conn, BS_CLOSE);
  tm_stop(conn->hold_timer);
  tm_stop(conn->keepalive_timer);
  conn->sk->rx_hook = NULL;

  if (os == BS_ESTABLISHED)
    bgp_conn_leave_established_state(p);
}

void
bgp_conn_enter_idle_state(struct bgp_conn *conn)
{
  struct bgp_proto *p = conn->bgp;
  int os = conn->state;

  bgp_close_conn(conn);
  bgp_conn_set_state(conn, BS_IDLE);
  ev_schedule(p->event);

  if (os == BS_ESTABLISHED)
    bgp_conn_leave_established_state(p);
}

static void
bgp_send_open(struct bgp_conn *conn)
{
  conn->start_state = conn->bgp->start_state;
  conn->want_as4_support = conn->bgp->cf->enable_as4 && (conn->start_state != BSS_CONNECT_NOCAP);
  conn->peer_as4_support = 0;	// Default value, possibly changed by receiving capability.
  conn->advertised_as = 0;

  DBG("BGP: Sending open\n");
  conn->sk->rx_hook = bgp_rx;
  conn->sk->tx_hook = bgp_tx;
  tm_stop(conn->connect_retry_timer);
  bgp_schedule_packet(conn, PKT_OPEN);
  bgp_conn_set_state(conn, BS_OPENSENT);
  bgp_start_timer(conn->hold_timer, conn->bgp->cf->initial_hold_time);
}

static void
bgp_connected(sock *sk)
{
  struct bgp_conn *conn = sk->data;
  struct bgp_proto *p = conn->bgp;

  BGP_TRACE(D_EVENTS, "Connected");
  bgp_send_open(conn);
}

static void
bgp_connect_timeout(timer *t)
{
  struct bgp_conn *conn = t->data;
  struct bgp_proto *p = conn->bgp;

  DBG("BGP: connect_timeout\n");
  if (p->p.proto_state == PS_START)
    {
      bgp_close_conn(conn);
      bgp_connect(p);
    }
  else
    bgp_conn_enter_idle_state(conn);
}

static void
bgp_sock_err(sock *sk, int err)
{
  struct bgp_conn *conn = sk->data;
  struct bgp_proto *p = conn->bgp;

  /*
   * This error hook may be called either asynchronously from main
   * loop, or synchronously from sk_send().  But sk_send() is called
   * only from bgp_tx() and bgp_kick_tx(), which are both called
   * asynchronously from main loop. Moreover, they end if err hook is
   * called. Therefore, we could suppose that it is always called
   * asynchronously.
   */

  bgp_store_error(p, conn, BE_SOCKET, err);

  if (err)
    BGP_TRACE(D_EVENTS, "Connection lost (%M)", err);
  else
    BGP_TRACE(D_EVENTS, "Connection closed");

  bgp_conn_enter_idle_state(conn);
}

static void
bgp_hold_timeout(timer *t)
{
  struct bgp_conn *conn = t->data;

  DBG("BGP: Hold timeout\n");

  /* If there is something in input queue, we are probably congested
     and perhaps just not processed BGP packets in time. */

  if (sk_rx_ready(conn->sk) > 0)
    bgp_start_timer(conn->hold_timer, 10);
  else
    bgp_error(conn, 4, 0, NULL, 0);
}

static void
bgp_keepalive_timeout(timer *t)
{
  struct bgp_conn *conn = t->data;

  DBG("BGP: Keepalive timer\n");
  bgp_schedule_packet(conn, PKT_KEEPALIVE);
}

static void
bgp_setup_conn(struct bgp_proto *p, struct bgp_conn *conn)
{
  timer *t;

  conn->sk = NULL;
  conn->bgp = p;
  conn->packets_to_send = 0;

  t = conn->connect_retry_timer = tm_new(p->p.pool);
  t->hook = bgp_connect_timeout;
  t->data = conn;
  t = conn->hold_timer = tm_new(p->p.pool);
  t->hook = bgp_hold_timeout;
  t->data = conn;
  t = conn->keepalive_timer = tm_new(p->p.pool);
  t->hook = bgp_keepalive_timeout;
  t->data = conn;
  conn->tx_ev = ev_new(p->p.pool);
  conn->tx_ev->hook = bgp_kick_tx;
  conn->tx_ev->data = conn;
}

static void
bgp_setup_sk(struct bgp_conn *conn, sock *s)
{
  s->data = conn;
  s->err_hook = bgp_sock_err;
  conn->sk = s;
}

static void
bgp_active(struct bgp_proto *p)
{
  int delay = MAX(1, p->cf->start_delay_time);
  struct bgp_conn *conn = &p->outgoing_conn;

  BGP_TRACE(D_EVENTS, "Connect delayed by %d seconds", delay);
  bgp_setup_conn(p, conn);
  bgp_conn_set_state(conn, BS_ACTIVE);
  bgp_start_timer(conn->connect_retry_timer, delay);
}

int
bgp_apply_limits(struct bgp_proto *p)
{
  if (p->cf->route_limit && (p->p.stats.imp_routes > p->cf->route_limit))
    {
      log(L_WARN "%s: Route limit exceeded, shutting down", p->p.name);
      bgp_store_error(p, NULL, BE_AUTO_DOWN, BEA_ROUTE_LIMIT_EXCEEDED);
      bgp_update_startup_delay(p);
      bgp_stop(p, 1); // Errcode 6, 1 - max number of prefixes reached
      return -1;
    }

  return 0;
}


/**
 * bgp_connect - initiate an outgoing connection
 * @p: BGP instance
 *
 * The bgp_connect() function creates a new &bgp_conn and initiates
 * a TCP connection to the peer. The rest of connection setup is governed
 * by the BGP state machine as described in the standard.
 */
static void
bgp_connect(struct bgp_proto *p)	/* Enter Connect state and start establishing connection */
{
  sock *s;
  struct bgp_conn *conn = &p->outgoing_conn;
  int hops = p->cf->multihop ? : 1;

  DBG("BGP: Connecting\n");
  s = sk_new(p->p.pool);
  s->type = SK_TCP_ACTIVE;
  s->saddr = p->source_addr;
  s->daddr = p->cf->remote_ip;
  s->dport = BGP_PORT;
  s->ttl = p->cf->ttl_security ? 255 : hops;
  s->rbsize = BGP_RX_BUFFER_SIZE;
  s->tbsize = BGP_TX_BUFFER_SIZE;
  s->tos = IP_PREC_INTERNET_CONTROL;
  s->password = p->cf->password;
  s->tx_hook = bgp_connected;

  /* Can only be NULL for multihop instances */
  if (p->neigh && *p->cf->interface)
    s->iface = p->neigh->iface;

  BGP_TRACE(D_EVENTS, "Connecting to %I from local address %I", s->daddr, s->saddr);
  bgp_setup_conn(p, conn);
  bgp_setup_sk(conn, s);
  bgp_conn_set_state(conn, BS_CONNECT);

  if (sk_open(s) < 0)
    {
      bgp_sock_err(s, 0);
      return;
    }

  /* Set minimal receive TTL if needed */
  if (p->cf->ttl_security)
  {
    DBG("Setting minimum received TTL to %d", 256 - hops);
    if (sk_set_min_ttl(s, 256 - hops) < 0)
    {
      log(L_ERR "TTL security configuration failed, closing session");
      bgp_sock_err(s, 0);
      return;
    }
  }

  DBG("BGP: Waiting for connect success\n");
  bgp_start_timer(conn->connect_retry_timer, p->cf->connect_retry_time);
}

/**
 * bgp_incoming_connection - handle an incoming connection
 * @sk: TCP socket
 * @dummy: unused
 *
 * This function serves as a socket hook for accepting of new BGP
 * connections. It searches a BGP instance corresponding to the peer
 * which has connected and if such an instance exists, it creates a
 * &bgp_conn structure, attaches it to the instance and either sends
 * an Open message or (if there already is an active connection) it
 * closes the new connection by sending a Notification message.
 */
static int
bgp_incoming_connection(sock *sk, int dummy UNUSED)
{
  struct proto_config *pc;

  DBG("BGP: Incoming connection from %I port %d\n", sk->daddr, sk->dport);
  WALK_LIST(pc, config->protos)
    if (pc->protocol == &proto_bgp && pc->proto)
      {
	struct bgp_proto *p = (struct bgp_proto *) pc->proto;
	if (ipa_equal(p->cf->remote_ip, sk->daddr))
	  {
	    /* We are in proper state and there is no other incoming connection */
	    int acc = (p->p.proto_state == PS_START || p->p.proto_state == PS_UP) &&
	      (p->start_state >= BSS_CONNECT) && (!p->incoming_conn.sk);

	    BGP_TRACE(D_EVENTS, "Incoming connection from %I (port %d) %s",
		      sk->daddr, sk->dport, acc ? "accepted" : "rejected");

	    if (!acc)
	      goto err;

	    int hops = p->cf->multihop ? : 1;
	    if (p->cf->ttl_security)
	    {
	      /* TTL security support */
	      if ((sk_set_ttl(sk, 255) < 0) ||
		  (sk_set_min_ttl(sk, 256 - hops) < 0))
	      {
		log(L_ERR "TTL security configuration failed, closing session");
		goto err;
	      }
	    }
	    else
	      sk_set_ttl(sk, hops);

	    bgp_setup_conn(p, &p->incoming_conn);
	    bgp_setup_sk(&p->incoming_conn, sk);
	    bgp_send_open(&p->incoming_conn);
	    return 0;
	  }
      }

  log(L_WARN "BGP: Unexpected connect from unknown address %I (port %d)", sk->daddr, sk->dport);
 err:
  rfree(sk);
  return 0;
}

static void
bgp_listen_sock_err(sock *sk UNUSED, int err)
{
  if (err == ECONNABORTED)
    log(L_WARN "BGP: Incoming connection aborted");
  else
    log(L_ERR "BGP: Error on listening socket: %M", err);
}

static sock *
bgp_setup_listen_sk(ip_addr addr, unsigned port, u32 flags)
{
  sock *s = sk_new(&root_pool);
  DBG("BGP: Creating listening socket\n");
  s->type = SK_TCP_PASSIVE;
  s->ttl = 255;
  s->saddr = addr;
  s->sport = port ? port : BGP_PORT;
  s->flags = flags ? 0 : SKF_V6ONLY;
  s->tos = IP_PREC_INTERNET_CONTROL;
  s->rbsize = BGP_RX_BUFFER_SIZE;
  s->tbsize = BGP_TX_BUFFER_SIZE;
  s->rx_hook = bgp_incoming_connection;
  s->err_hook = bgp_listen_sock_err;

  if (sk_open(s) < 0)
    {
      log(L_ERR "BGP: Unable to open listening socket");
      rfree(s);
      return NULL;
    }

  return s;
}

static void
bgp_start_neighbor(struct bgp_proto *p)
{
  /* Called only for single-hop BGP sessions */

  if (ipa_zero(p->source_addr))
    p->source_addr = p->neigh->iface->addr->ip; 

#ifdef IPV6
  {
    struct ifa *a;
    p->local_link = IPA_NONE;
    WALK_LIST(a, p->neigh->iface->addrs)
      if (a->scope == SCOPE_LINK)
        {
	  p->local_link = a->ip;
	  break;
	}

    if (! ipa_nonzero(p->local_link))
      log(L_WARN "%s: Missing link local address on interface %s", p->p.name,  p->neigh->iface->name);

    DBG("BGP: Selected link-level address %I\n", p->local_link);
  }
#endif

  bgp_initiate(p);
}

static void
bgp_neigh_notify(neighbor *n)
{
  struct bgp_proto *p = (struct bgp_proto *) n->proto;

  if (n->iface)
    {
      if ((p->p.proto_state == PS_START) && (p->start_state == BSS_PREPARE))
	{
	  BGP_TRACE(D_EVENTS, "Neighbor found");
	  bgp_start_neighbor(p);
	}
    }
  else
    {
      if ((p->p.proto_state == PS_START) || (p->p.proto_state == PS_UP))
	{
	  BGP_TRACE(D_EVENTS, "Neighbor lost");
	  bgp_store_error(p, NULL, BE_MISC, BEM_NEIGHBOR_LOST);
	  bgp_stop(p, 0);
	}
    }
}

static int
bgp_reload_routes(struct proto *P)
{
  struct bgp_proto *p = (struct bgp_proto *) P;
  if (!p->conn || !p->conn->peer_refresh_support)
    return 0;

  bgp_schedule_packet(p->conn, PKT_ROUTE_REFRESH);
  return 1;
}

static void
bgp_start_locked(struct object_lock *lock)
{
  struct bgp_proto *p = lock->data;
  struct bgp_config *cf = p->cf;

  if (p->p.proto_state != PS_START)
    {
      DBG("BGP: Got lock in different state %d\n", p->p.proto_state);
      return;
    }

  DBG("BGP: Got lock\n");

  if (cf->multihop)
    {
      /* Multi-hop sessions do not use neighbor entries */
      bgp_initiate(p);
      return;
    }

  p->neigh = neigh_find_ifname(&p->p, &cf->remote_ip, (*cf->interface) ? cf->interface : NULL, NEF_STICKY);
  if (!p->neigh || (p->neigh->scope == SCOPE_HOST))
    {
      log(L_ERR "%s: Invalid remote address %I", p->p.name, cf->remote_ip);
      /* As we do not start yet, we can just disable protocol */
      p->p.disabled = 1;
      bgp_store_error(p, NULL, BE_MISC, BEM_INVALID_NEXT_HOP);
      proto_notify_state(&p->p, PS_DOWN);
      return;
    }
  
  if (p->neigh->iface)
    bgp_start_neighbor(p);
  else
    BGP_TRACE(D_EVENTS, "Waiting for %I to become my neighbor%s%s", cf->remote_ip, (*cf->interface) ? " on " : "", cf->interface);
}

static int
bgp_start(struct proto *P)
{
  struct bgp_proto *p = (struct bgp_proto *) P;
  struct object_lock *lock;

  DBG("BGP: Startup.\n");
  p->start_state = BSS_PREPARE;
  p->outgoing_conn.state = BS_IDLE;
  p->incoming_conn.state = BS_IDLE;
  p->neigh = NULL;

  rt_lock_table(p->igp_table);

  p->event = ev_new(p->p.pool);
  p->event->hook = bgp_decision;
  p->event->data = p;

  p->startup_timer = tm_new(p->p.pool);
  p->startup_timer->hook = bgp_startup_timeout;
  p->startup_timer->data = p;

  p->local_id = proto_get_router_id(P->cf);
  if (p->rr_client)
    p->rr_cluster_id = p->cf->rr_cluster_id ? p->cf->rr_cluster_id : p->local_id;

  p->remote_id = 0;
  p->source_addr = p->cf->source_addr;

  /*
   *  Before attempting to create the connection, we need to lock the
   *  port, so that are sure we're the only instance attempting to talk
   *  with that neighbor.
   */

  lock = p->lock = olock_new(P->pool);
  lock->addr = p->cf->remote_ip;
  lock->type = OBJLOCK_TCP;
  lock->port = BGP_PORT;
  lock->iface = NULL;
  lock->hook = bgp_start_locked;
  lock->data = p;

  olock_acquire(lock);

  return PS_START;
}

static int
bgp_shutdown(struct proto *P)
{
  struct bgp_proto *p = (struct bgp_proto *) P;
  unsigned subcode;

  BGP_TRACE(D_EVENTS, "Shutdown requested");
  bgp_store_error(p, NULL, BE_MAN_DOWN, 0);

  if (P->reconfiguring)
    {
      if (P->cf_new)
	subcode = 6; // Errcode 6, 6 - other configuration change
      else
	subcode = 3; // Errcode 6, 3 - peer de-configured
    }
  else
    subcode = 2; // Errcode 6, 2 - administrative shutdown

  p->startup_delay = 0;
  bgp_stop(p, subcode);

  return p->p.proto_state;
}

static void
bgp_cleanup(struct proto *P)
{
  struct bgp_proto *p = (struct bgp_proto *) P;
  rt_unlock_table(p->igp_table);
}

static rtable *
get_igp_table(struct bgp_config *cf)
{
  return cf->igp_table ? cf->igp_table->table : cf->c.table->table;
}

static struct proto *
bgp_init(struct proto_config *C)
{
  struct bgp_config *c = (struct bgp_config *) C;
  struct proto *P = proto_new(C, sizeof(struct bgp_proto));
  struct bgp_proto *p = (struct bgp_proto *) P;

  P->accept_ra_types = RA_OPTIMAL;
  P->rt_notify = bgp_rt_notify;
  P->rte_better = bgp_rte_better;
  P->import_control = bgp_import_control;
  P->neigh_notify = bgp_neigh_notify;
  P->reload_routes = bgp_reload_routes;

  if (c->deterministic_med)
    P->rte_recalculate = bgp_rte_recalculate;

  p->cf = c;
  p->local_as = c->local_as;
  p->remote_as = c->remote_as;
  p->is_internal = (c->local_as == c->remote_as);
  p->rs_client = c->rs_client;
  p->rr_client = c->rr_client;
  p->igp_table = get_igp_table(c);

  return P;
}


void
bgp_check_config(struct bgp_config *c)
{
  int internal = (c->local_as == c->remote_as);

  /* Do not check templates at all */
  if (c->c.class == SYM_TEMPLATE)
    return;

  if (!c->local_as)
    cf_error("Local AS number must be set");

  if (!c->remote_as)
    cf_error("Neighbor must be configured");

  if (!(c->capabilities && c->enable_as4) && (c->remote_as > 0xFFFF))
    cf_error("Neighbor AS number out of range (AS4 not available)");

  if (!internal && c->rr_client)
    cf_error("Only internal neighbor can be RR client");

  if (internal && c->rs_client)
    cf_error("Only external neighbor can be RS client");

  if (c->multihop && (c->gw_mode == GW_DIRECT))
    cf_error("Multihop BGP cannot use direct gateway mode");

  /* Different default based on rs_client */
  if (!c->missing_lladdr)
    c->missing_lladdr = c->rs_client ? MLL_IGNORE : MLL_SELF;

  /* Different default for gw_mode */
  if (!c->gw_mode)
    c->gw_mode = (c->multihop || internal) ? GW_RECURSIVE : GW_DIRECT;
}

static int
bgp_reconfigure(struct proto *P, struct proto_config *C)
{
  struct bgp_config *new = (struct bgp_config *) C;
  struct bgp_proto *p = (struct bgp_proto *) P;
  struct bgp_config *old = p->cf;

  int same = !memcmp(((byte *) old) + sizeof(struct proto_config),
		     ((byte *) new) + sizeof(struct proto_config),
		     // password item is last and must be checked separately
		     OFFSETOF(struct bgp_config, password) - sizeof(struct proto_config))
    && ((!old->password && !new->password)
	|| (old->password && new->password && !strcmp(old->password, new->password)))
    && (get_igp_table(old) == get_igp_table(new));

  /* We should update our copy of configuration ptr as old configuration will be freed */
  if (same)
    p->cf = new;

  return same;
}

static void
bgp_copy_config(struct proto_config *dest, struct proto_config *src)
{
  /* Just a shallow copy */
  proto_copy_rest(dest, src, sizeof(struct bgp_config));
}


/**
 * bgp_error - report a protocol error
 * @c: connection
 * @code: error code (according to the RFC)
 * @subcode: error sub-code
 * @data: data to be passed in the Notification message
 * @len: length of the data
 *
 * bgp_error() sends a notification packet to tell the other side that a protocol
 * error has occurred (including the data considered erroneous if possible) and
 * closes the connection.
 */
void
bgp_error(struct bgp_conn *c, unsigned code, unsigned subcode, byte *data, int len)
{
  struct bgp_proto *p = c->bgp;

  if (c->state == BS_CLOSE)
    return;

  bgp_log_error(p, BE_BGP_TX, "Error", code, subcode, data, (len > 0) ? len : -len);
  bgp_store_error(p, c, BE_BGP_TX, (code << 16) | subcode);
  bgp_conn_enter_close_state(c);

  c->notify_code = code;
  c->notify_subcode = subcode;
  c->notify_data = data;
  c->notify_size = (len > 0) ? len : 0;
  bgp_schedule_packet(c, PKT_NOTIFICATION);

  if (code != 6)
    {
      bgp_update_startup_delay(p);
      bgp_stop(p, 0);
    }
}

/**
 * bgp_store_error - store last error for status report
 * @p: BGP instance
 * @c: connection
 * @class: error class (BE_xxx constants)
 * @code: error code (class specific)
 *
 * bgp_store_error() decides whether given error is interesting enough
 * and store that error to last_error variables of @p
 */
void
bgp_store_error(struct bgp_proto *p, struct bgp_conn *c, u8 class, u32 code)
{
  /* During PS_UP, we ignore errors on secondary connection */
  if ((p->p.proto_state == PS_UP) && c && (c != p->conn))
    return;

  /* During PS_STOP, we ignore any errors, as we want to report
   * the error that caused transition to PS_STOP
   */
  if (p->p.proto_state == PS_STOP)
    return;

  p->last_error_class = class;
  p->last_error_code = code;
}

static char *bgp_state_names[] = { "Idle", "Connect", "Active", "OpenSent", "OpenConfirm", "Established", "Close" };
static char *bgp_err_classes[] = { "", "Error: ", "Socket: ", "Received: ", "BGP Error: ", "Automatic shutdown: ", ""};
static char *bgp_misc_errors[] = { "", "Neighbor lost", "Invalid next hop", "Kernel MD5 auth failed", "No listening socket" };
static char *bgp_auto_errors[] = { "", "Route limit exceeded"};

static const char *
bgp_last_errmsg(struct bgp_proto *p)
{
  switch (p->last_error_class)
    {
    case BE_MISC:
      return bgp_misc_errors[p->last_error_code];
    case BE_SOCKET:
      return (p->last_error_code == 0) ? "Connection closed" : strerror(p->last_error_code);
    case BE_BGP_RX:
    case BE_BGP_TX:
      return bgp_error_dsc(p->last_error_code >> 16, p->last_error_code & 0xFF);
    case BE_AUTO_DOWN:
      return bgp_auto_errors[p->last_error_code];
    default:
      return "";
    }
}

static const char *
bgp_state_dsc(struct bgp_proto *p)
{
  if (p->p.proto_state == PS_DOWN)
    return "Down";

  int state = MAX(p->incoming_conn.state, p->outgoing_conn.state);
  if ((state == BS_IDLE) && (p->start_state >= BSS_CONNECT) && p->cf->passive)
    return "Passive";

  return bgp_state_names[state];
}

static void
bgp_get_status(struct proto *P, byte *buf)
{
  struct bgp_proto *p = (struct bgp_proto *) P;

  const char *err1 = bgp_err_classes[p->last_error_class];
  const char *err2 = bgp_last_errmsg(p);

  if (P->proto_state == PS_DOWN)
    bsprintf(buf, "%s%s", err1, err2);
  else
    bsprintf(buf, "%-14s%s%s", bgp_state_dsc(p), err1, err2);
}

static inline bird_clock_t tm_remains(timer *t)
{ return t->expires ? t->expires - now : 0; }

static void
bgp_show_proto_info(struct proto *P)
{
  struct bgp_proto *p = (struct bgp_proto *) P;
  struct bgp_conn *c = p->conn;

  cli_msg(-1006, "  BGP state:          %s", bgp_state_dsc(p));
  cli_msg(-1006, "    Neighbor address: %I", p->cf->remote_ip);
  cli_msg(-1006, "    Neighbor AS:      %u", p->remote_as);

  if (P->proto_state == PS_START)
    {
      struct bgp_conn *oc = &p->outgoing_conn;

      if ((p->start_state < BSS_CONNECT) &&
	  (p->startup_timer->expires))
	cli_msg(-1006, "    Error wait:       %d/%d", 
		p->startup_timer->expires - now, p->startup_delay);

      if ((oc->state == BS_ACTIVE) &&
	  (oc->connect_retry_timer->expires))
	cli_msg(-1006, "    Start delay:      %d/%d", 
		oc->connect_retry_timer->expires - now, p->cf->start_delay_time);
    }
  else if (P->proto_state == PS_UP)
    {
      cli_msg(-1006, "    Neighbor ID:      %R", p->remote_id);
      cli_msg(-1006, "    Neighbor caps:   %s%s",
	      c->peer_refresh_support ? " refresh" : "",
	      c->peer_as4_support ? " AS4" : "");
      cli_msg(-1006, "    Session:          %s%s%s%s%s",
	      p->is_internal ? "internal" : "external",
	      p->cf->multihop ? " multihop" : "",
	      p->rr_client ? " route-reflector" : "",
	      p->rs_client ? " route-server" : "",
	      p->as4_session ? " AS4" : "");
      cli_msg(-1006, "    Source address:   %I", p->source_addr);
      if (p->neigh && p->neigh->iface)
	cli_msg(-1006, "    Interface:        %s%s",
		p->neigh->iface->name, (*p->cf->interface) ? " (fixed)" : "");
      if (p->cf->route_limit)
	cli_msg(-1006, "    Route limit:      %d/%d",
		p->p.stats.imp_routes, p->cf->route_limit);
      cli_msg(-1006, "    Hold timer:       %d/%d",
	      tm_remains(c->hold_timer), c->hold_time);
      cli_msg(-1006, "    Keepalive timer:  %d/%d",
	      tm_remains(c->keepalive_timer), c->keepalive_time);
    }

  if ((p->last_error_class != BE_NONE) && 
      (p->last_error_class != BE_MAN_DOWN))
    {
      const char *err1 = bgp_err_classes[p->last_error_class];
      const char *err2 = bgp_last_errmsg(p);
      cli_msg(-1006, "    Last error:       %s%s", err1, err2);
    }
}

struct protocol proto_bgp = {
  name:			"BGP",
  template:		"bgp%d",
  attr_class:		EAP_BGP,
  init:			bgp_init,
  start:		bgp_start,
  shutdown:		bgp_shutdown,
  cleanup:		bgp_cleanup,
  reconfigure:		bgp_reconfigure,
  copy_config:		bgp_copy_config,
  get_status:		bgp_get_status,
  get_attr:		bgp_get_attr,
  get_route_info:	bgp_get_route_info,
  show_proto_info:	bgp_show_proto_info
};