summaryrefslogtreecommitdiffstats
path: root/ldpd/packet.c
blob: 6770ab73ef3ec4f69904d6272d607a3b7cc561d8 (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
// SPDX-License-Identifier: ISC
/*	$OpenBSD$ */

/*
 * Copyright (c) 2013, 2016 Renato Westphal <renato@openbsd.org>
 * Copyright (c) 2009 Michele Marchetto <michele@openbsd.org>
 * Copyright (c) 2004, 2005, 2008 Esben Norby <norby@openbsd.org>
 */

#include <zebra.h>

#include "ldpd.h"
#include "ldpe.h"
#include "log.h"

#include "sockopt.h"

static struct iface		*disc_find_iface(unsigned int, int,
				    union ldpd_addr *);
static void session_read(struct event *thread);
static void session_write(struct event *thread);
static ssize_t			 session_get_pdu(struct ibuf_read *, char **);
static void			 tcp_close(struct tcp_conn *);
static struct pending_conn	*pending_conn_new(int, int, union ldpd_addr *);
static void pending_conn_timeout(struct event *thread);

int
gen_ldp_hdr(struct ibuf *buf, uint16_t size)
{
	struct ldp_hdr	ldp_hdr;

	memset(&ldp_hdr, 0, sizeof(ldp_hdr));
	ldp_hdr.version = htons(LDP_VERSION);
	/* exclude the 'Version' and 'PDU Length' fields from the total */
	ldp_hdr.length = htons(size - LDP_HDR_DEAD_LEN);
	ldp_hdr.lsr_id = ldp_rtr_id_get(leconf);
	ldp_hdr.lspace_id = 0;

	return (ibuf_add(buf, &ldp_hdr, LDP_HDR_SIZE));
}

int
gen_msg_hdr(struct ibuf *buf, uint16_t type, uint16_t size)
{
	static int	msgcnt = 0;
	struct ldp_msg	msg;

	memset(&msg, 0, sizeof(msg));
	msg.type = htons(type);
	/* exclude the 'Type' and 'Length' fields from the total */
	msg.length = htons(size - LDP_MSG_DEAD_LEN);
	msg.id = htonl(++msgcnt);

	return (ibuf_add(buf, &msg, sizeof(msg)));
}

/* send packets */
int
send_packet(int fd, int af, union ldpd_addr *dst, struct iface_af *ia,
    void *pkt, size_t len)
{
	union sockunion su;

	switch (af) {
	case AF_INET:
		if (ia && IN_MULTICAST(ntohl(dst->v4.s_addr))) {
			/* set outgoing interface for multicast traffic */
			if (sock_set_ipv4_mcast(ia->iface) == -1) {
				log_debug("%s: error setting multicast interface, %s", __func__, ia->iface->name);
				return (-1);
			}
		}
		break;
	case AF_INET6:
		if (ia && IN6_IS_ADDR_MULTICAST(&dst->v6)) {
			/* set outgoing interface for multicast traffic */
			if (sock_set_ipv6_mcast(ia->iface) == -1) {
				log_debug("%s: error setting multicast interface, %s", __func__, ia->iface->name);
				return (-1);
			}
		}
		break;
	default:
		fatalx("send_packet: unknown af");
	}

	addr2sa(af, dst, LDP_PORT, &su);
	if (sendto(fd, pkt, len, 0, &su.sa, sockaddr_len(&su.sa)) == -1) {
		log_warn("%s: error sending packet to %s", __func__,
			 log_sockaddr(&su.sa));
		return (-1);
	}

	return (0);
}

/* Discovery functions */
void disc_recv_packet(struct event *thread)
{
	int			 fd = THREAD_FD(thread);
	struct event **threadp = THREAD_ARG(thread);

	union {
		struct	cmsghdr hdr;
#ifdef HAVE_STRUCT_SOCKADDR_DL
		char	buf[CMSG_SPACE(sizeof(struct sockaddr_dl))];
#else
		char	buf[CMSG_SPACE(sizeof(struct in6_pktinfo))];
#endif
	} cmsgbuf;
	struct msghdr		 m;
	struct sockaddr_storage	 from;
	struct iovec		 iov;
	char			*buf;
#ifndef MSG_MCAST
	struct cmsghdr		*cmsg;
#endif
	ssize_t			 r;
	int			 multicast;
	int			 af;
	union ldpd_addr		 src;
	unsigned int		 ifindex = 0;
	struct iface		*iface = NULL;
	uint16_t		 len;
	struct ldp_hdr		 ldp_hdr;
	uint16_t		 pdu_len;
	struct ldp_msg		 msg;
	uint16_t		 msg_len;
	struct in_addr		 lsr_id;

	/* reschedule read */
	thread_add_read(master, disc_recv_packet, threadp, fd, threadp);

	/* setup buffer */
	memset(&m, 0, sizeof(m));
	iov.iov_base = buf = pkt_ptr;
	iov.iov_len = IBUF_READ_SIZE;
	m.msg_name = &from;
	m.msg_namelen = sizeof(from);
	m.msg_iov = &iov;
	m.msg_iovlen = 1;
	m.msg_control = &cmsgbuf.buf;
	m.msg_controllen = sizeof(cmsgbuf.buf);

	if ((r = recvmsg(fd, &m, 0)) == -1) {
		if (errno != EAGAIN && errno != EINTR)
			log_debug("%s: read error: %s", __func__,
			    strerror(errno));
		return;
	}

	sa2addr((struct sockaddr *)&from, &af, &src, NULL);
#ifdef MSG_MCAST
	multicast = (m.msg_flags & MSG_MCAST) ? 1 : 0;
#else
	multicast = 0;
	for (cmsg = CMSG_FIRSTHDR(&m); cmsg != NULL;
	    cmsg = CMSG_NXTHDR(&m, cmsg)) {
#if defined(HAVE_IP_PKTINFO)
		if (af == AF_INET && cmsg->cmsg_level == IPPROTO_IP &&
		    cmsg->cmsg_type == IP_PKTINFO) {
			struct in_pktinfo	*pktinfo;

			pktinfo = (struct in_pktinfo *)CMSG_DATA(cmsg);
			if (IN_MULTICAST(ntohl(pktinfo->ipi_addr.s_addr)))
				multicast = 1;
			break;
		}
#elif defined(HAVE_IP_RECVDSTADDR)
		if (af == AF_INET && cmsg->cmsg_level == IPPROTO_IP &&
		    cmsg->cmsg_type == IP_RECVDSTADDR) {
			struct in_addr		*addr;

			addr = (struct in_addr *)CMSG_DATA(cmsg);
			if (IN_MULTICAST(ntohl(addr->s_addr)))
				multicast = 1;
			break;
		}
#else
#error "Unsupported socket API"
#endif
		if (af == AF_INET6 && cmsg->cmsg_level == IPPROTO_IPV6 &&
		    cmsg->cmsg_type == IPV6_PKTINFO) {
			struct in6_pktinfo	*pktinfo;

			pktinfo = (struct in6_pktinfo *)CMSG_DATA(cmsg);
			if (IN6_IS_ADDR_MULTICAST(&pktinfo->ipi6_addr))
				multicast = 1;
			break;
		}
	}
#endif /* MSG_MCAST */
	if (bad_addr(af, &src)) {
		log_debug("%s: invalid source address: %s", __func__,
		    log_addr(af, &src));
		return;
	}
	ifindex = getsockopt_ifindex(af, &m);

	/* find a matching interface */
	if (multicast) {
		iface = disc_find_iface(ifindex, af, &src);
		if (iface == NULL)
			return;
	}

	/* check packet size */
	len = (uint16_t)r;
	if (len < (LDP_HDR_SIZE + LDP_MSG_SIZE) || len > LDP_MAX_LEN) {
		log_debug("%s: bad packet size, source %s", __func__,
		    log_addr(af, &src));
		return;
	}

	/* LDP header sanity checks */
	memcpy(&ldp_hdr, buf, sizeof(ldp_hdr));
	if (ntohs(ldp_hdr.version) != LDP_VERSION) {
		log_debug("%s: invalid LDP version %d, source %s", __func__,
		    ntohs(ldp_hdr.version), log_addr(af, &src));
		return;
	}
	if (ntohs(ldp_hdr.lspace_id) != 0) {
		log_debug("%s: invalid label space %u, source %s", __func__,
		    ntohs(ldp_hdr.lspace_id), log_addr(af, &src));
		return;
	}
	/* check "PDU Length" field */
	pdu_len = ntohs(ldp_hdr.length);
	if ((pdu_len < (LDP_HDR_PDU_LEN + LDP_MSG_SIZE)) ||
	    (pdu_len > (len - LDP_HDR_DEAD_LEN))) {
		log_debug("%s: invalid LDP packet length %u, source %s",
		    __func__, ntohs(ldp_hdr.length), log_addr(af, &src));
		return;
	}
	buf += LDP_HDR_SIZE;
	len -= LDP_HDR_SIZE;

	lsr_id.s_addr = ldp_hdr.lsr_id;

	/*
	 * For UDP, we process only the first message of each packet. This does
	 * not impose any restrictions since LDP uses UDP only for sending Hello
	 * packets.
	 */
	memcpy(&msg, buf, sizeof(msg));

	/* check "Message Length" field */
	msg_len = ntohs(msg.length);
	if (msg_len < LDP_MSG_LEN || ((msg_len + LDP_MSG_DEAD_LEN) > pdu_len)) {
		log_debug("%s: invalid LDP message length %u, source %s",
		    __func__, ntohs(msg.length), log_addr(af, &src));
		return;
	}
	buf += LDP_MSG_SIZE;
	len -= LDP_MSG_SIZE;

	/* switch LDP packet type */
	switch (ntohs(msg.type)) {
	case MSG_TYPE_HELLO:
		recv_hello(lsr_id, &msg, af, &src, iface, multicast, buf, len);
		break;
	default:
		log_debug("%s: unknown LDP packet type, source %s", __func__,
		    log_addr(af, &src));
	}
}

static struct iface *
disc_find_iface(unsigned int ifindex, int af, union ldpd_addr *src)
{
	struct iface	*iface;
	struct iface_af	*ia;

	iface = if_lookup(leconf, ifindex);
	if (iface == NULL)
		return (NULL);

	ia = iface_af_get(iface, af);
	if (!ia->enabled)
		return (NULL);

	/*
	 * RFC 7552 - Section 5.1:
	 * "Link-local IPv6 address MUST be used as the source IP address in
	 * IPv6 LDP Link Hellos".
	 */
	if (af == AF_INET6 && !IN6_IS_ADDR_LINKLOCAL(&src->v6))
		return (NULL);

	return (iface);
}

void session_accept(struct event *thread)
{
	int			 fd = THREAD_FD(thread);
	struct sockaddr_storage	 src;
	socklen_t		 len = sizeof(src);
	int			 newfd;
	int			 af;
	union ldpd_addr		 addr;
	struct nbr		*nbr;
	struct pending_conn	*pconn;

	newfd = accept(fd, (struct sockaddr *)&src, &len);
	if (newfd == -1) {
		/*
		 * Pause accept if we are out of file descriptors, or
		 * libevent will haunt us here too.
		 */
		if (errno == ENFILE || errno == EMFILE) {
			accept_pause();
		} else if (errno != EWOULDBLOCK && errno != EINTR &&
		    errno != ECONNABORTED)
			log_debug("%s: accept error: %s", __func__,
			    strerror(errno));
		return;
	}
	sock_set_nonblock(newfd);

	sa2addr((struct sockaddr *)&src, &af, &addr, NULL);

	/*
	 * Since we don't support label spaces, we can identify this neighbor
	 * just by its source address. This way we don't need to wait for its
	 * Initialization message to know who we are talking to.
	 */
	nbr = nbr_find_addr(af, &addr);
	if (nbr == NULL) {
		/*
		 * According to RFC 5036, we would need to send a No Hello
		 * Error Notification message and close this TCP connection
		 * right now. But doing so would trigger the backoff exponential
		 * timer in the remote peer, which would considerably slow down
		 * the session establishment process. The trick here is to wait
		 * five seconds before sending the Notification Message. There's
		 * a good chance that the remote peer will send us a Hello
		 * message within this interval, so it's worth waiting before
		 * taking a more drastic measure.
		 */
		pconn = pending_conn_find(af, &addr);
		if (pconn)
			close(newfd);
		else
			pending_conn_new(newfd, af, &addr);
		return;
	}
	/* protection against buggy implementations */
	if (nbr_session_active_role(nbr)) {
		close(newfd);
		return;
	}
	if (nbr->state != NBR_STA_PRESENT) {
		log_debug("%s: lsr-id %pI4: rejecting additional transport connection", __func__, &nbr->id);
		close(newfd);
		return;
	}

	session_accept_nbr(nbr, newfd);
}

void
session_accept_nbr(struct nbr *nbr, int fd)
{
#ifdef __OpenBSD__
	struct nbr_params	*nbrp;
	int			 opt;
	socklen_t		 len;

	nbrp = nbr_params_find(leconf, nbr->id);
	if (nbr_gtsm_check(fd, nbr, nbrp)) {
		close(fd);
		return;
	}

	if (nbrp && nbrp->auth.method == AUTH_MD5SIG) {
		if (sysdep.no_pfkey || sysdep.no_md5sig) {
			log_warnx("md5sig configured but not available");
			close(fd);
			return;
		}

		len = sizeof(opt);
		if (getsockopt(fd, IPPROTO_TCP, TCP_MD5SIG, &opt, &len) == -1)
			fatal("getsockopt TCP_MD5SIG");
		if (!opt) {	/* non-md5'd connection! */
			log_warnx("connection attempt without md5 signature");
			close(fd);
			return;
		}
	}
#endif

	nbr->tcp = tcp_new(fd, nbr);
	nbr_fsm(nbr, NBR_EVT_MATCH_ADJ);
}

static void session_read(struct event *thread)
{
	int		 fd = THREAD_FD(thread);
	struct nbr	*nbr = THREAD_ARG(thread);
	struct tcp_conn	*tcp = nbr->tcp;
	struct ldp_hdr	*ldp_hdr;
	struct ldp_msg	*msg;
	char		*buf = NULL, *pdu;
	ssize_t		 n, len;
	uint16_t	 pdu_len, msg_len, msg_size, max_pdu_len;
	int		 ret;

	thread_add_read(master, session_read, nbr, fd, &tcp->rev);

	if ((n = read(fd, tcp->rbuf->buf + tcp->rbuf->wpos,
	    sizeof(tcp->rbuf->buf) - tcp->rbuf->wpos)) == -1) {
		if (errno != EINTR && errno != EAGAIN) {
			log_warn("%s: read error", __func__);
			nbr_fsm(nbr, NBR_EVT_CLOSE_SESSION);
			return;
		}
		/* retry read */
		return;
	}
	if (n == 0) {
		/* connection closed */
		log_debug("%s: connection closed by remote end", __func__);
		nbr_fsm(nbr, NBR_EVT_CLOSE_SESSION);
		return;
	}
	tcp->rbuf->wpos += n;

	while ((len = session_get_pdu(tcp->rbuf, &buf)) > 0) {
		pdu = buf;
		ldp_hdr = (struct ldp_hdr *)pdu;
		if (ntohs(ldp_hdr->version) != LDP_VERSION) {
			session_shutdown(nbr, S_BAD_PROTO_VER, 0, 0);
			free(buf);
			return;
		}

		pdu_len = ntohs(ldp_hdr->length);
		/*
	 	 * RFC 5036 - Section 3.5.3:
		 * "Prior to completion of the negotiation, the maximum
		 * allowable length is 4096 bytes".
		 */
		if (nbr->state == NBR_STA_OPER)
			max_pdu_len = nbr->max_pdu_len;
		else
			max_pdu_len = LDP_MAX_LEN;
		if (pdu_len < (LDP_HDR_PDU_LEN + LDP_MSG_SIZE) ||
		    pdu_len > max_pdu_len) {
			session_shutdown(nbr, S_BAD_PDU_LEN, 0, 0);
			free(buf);
			return;
		}
		pdu_len -= LDP_HDR_PDU_LEN;
		if (ldp_hdr->lsr_id != nbr->id.s_addr ||
		    ldp_hdr->lspace_id != 0) {
			session_shutdown(nbr, S_BAD_LDP_ID, 0, 0);
			free(buf);
			return;
		}
		pdu += LDP_HDR_SIZE;
		len -= LDP_HDR_SIZE;

		nbr_fsm(nbr, NBR_EVT_PDU_RCVD);

		while (len >= LDP_MSG_SIZE) {
			uint16_t type;

			msg = (struct ldp_msg *)pdu;
			type = ntohs(msg->type);
			msg_len = ntohs(msg->length);
			if (msg_len < LDP_MSG_LEN ||
			    (msg_len + LDP_MSG_DEAD_LEN) > pdu_len) {
				session_shutdown(nbr, S_BAD_MSG_LEN, msg->id,
				    msg->type);
				free(buf);
				return;
			}
			msg_size = msg_len + LDP_MSG_DEAD_LEN;
			pdu_len -= msg_size;

			/* check for error conditions earlier */
			switch (type) {
			case MSG_TYPE_INIT:
				if ((nbr->state != NBR_STA_INITIAL) &&
				    (nbr->state != NBR_STA_OPENSENT)) {
					session_shutdown(nbr, S_SHUTDOWN,
					    msg->id, msg->type);
					free(buf);
					return;
				}
				break;
			case MSG_TYPE_KEEPALIVE:
				if ((nbr->state == NBR_STA_INITIAL) ||
				    (nbr->state == NBR_STA_OPENSENT)) {
					session_shutdown(nbr, S_SHUTDOWN,
					    msg->id, msg->type);
					free(buf);
					return;
				}
				break;
			case MSG_TYPE_NOTIFICATION:
				break;
			default:
				if (nbr->state != NBR_STA_OPER) {
					session_shutdown(nbr, S_SHUTDOWN,
					    msg->id, msg->type);
					free(buf);
					return;
				}
				break;
			}

			/* switch LDP packet type */
			switch (type) {
			case MSG_TYPE_NOTIFICATION:
				ret = recv_notification(nbr, pdu, msg_size);
				break;
			case MSG_TYPE_INIT:
				ret = recv_init(nbr, pdu, msg_size);
				break;
			case MSG_TYPE_KEEPALIVE:
				ret = recv_keepalive(nbr, pdu, msg_size);
				break;
			case MSG_TYPE_CAPABILITY:
				ret = recv_capability(nbr, pdu, msg_size);
				break;
			case MSG_TYPE_ADDR:
			case MSG_TYPE_ADDRWITHDRAW:
				ret = recv_address(nbr, pdu, msg_size);
				break;
			case MSG_TYPE_LABELMAPPING:
			case MSG_TYPE_LABELREQUEST:
			case MSG_TYPE_LABELWITHDRAW:
			case MSG_TYPE_LABELRELEASE:
			case MSG_TYPE_LABELABORTREQ:
				ret = recv_labelmessage(nbr, pdu, msg_size,
				    type);
				break;
			default:
				log_debug("%s: unknown LDP message from nbr %pI4",
				    __func__, &nbr->id);
				if (!(ntohs(msg->type) & UNKNOWN_FLAG)) {
					nbr->stats.unknown_msg++;
					send_notification(nbr->tcp,
					    S_UNKNOWN_MSG, msg->id, msg->type);
				}
				/* ignore the message */
				ret = 0;
				break;
			}

			if (ret == -1) {
				/* parser failed, giving up */
				free(buf);
				return;
			}

			/* no errors - update per neighbor message counters */
			switch (type) {
			case MSG_TYPE_NOTIFICATION:
				nbr->stats.notif_rcvd++;
				break;
			case MSG_TYPE_KEEPALIVE:
				nbr->stats.kalive_rcvd++;
				break;
			case MSG_TYPE_CAPABILITY:
				nbr->stats.capability_rcvd++;
				break;
			case MSG_TYPE_ADDR:
				nbr->stats.addr_rcvd++;
				break;
			case MSG_TYPE_ADDRWITHDRAW:
				nbr->stats.addrwdraw_rcvd++;
				break;
			case MSG_TYPE_LABELMAPPING:
				nbr->stats.labelmap_rcvd++;
				break;
			case MSG_TYPE_LABELREQUEST:
				nbr->stats.labelreq_rcvd++;
				break;
			case MSG_TYPE_LABELWITHDRAW:
				nbr->stats.labelwdraw_rcvd++;
				break;
			case MSG_TYPE_LABELRELEASE:
				nbr->stats.labelrel_rcvd++;
				break;
			case MSG_TYPE_LABELABORTREQ:
				nbr->stats.labelabreq_rcvd++;
				break;
			default:
				break;
			}

			/* Analyse the next message */
			pdu += msg_size;
			len -= msg_size;
		}
		free(buf);
		buf = NULL;
		if (len != 0) {
			session_shutdown(nbr, S_BAD_PDU_LEN, 0, 0);
			return;
		}
	}

	/* shouldn't happen, session_get_pdu should be > 0 if buf was
	 * allocated - but let's get rid of the SA warning.
	 */
	free(buf);
}

static void session_write(struct event *thread)
{
	struct tcp_conn *tcp = THREAD_ARG(thread);
	struct nbr	*nbr = tcp->nbr;

	tcp->wbuf.ev = NULL;

	if (msgbuf_write(&tcp->wbuf.wbuf) <= 0)
		if (errno != EAGAIN && nbr)
			nbr_fsm(nbr, NBR_EVT_CLOSE_SESSION);

	if (nbr == NULL && !tcp->wbuf.wbuf.queued) {
		/*
		 * We are done sending the notification message, now we can
		 * close the socket.
		 */
		tcp_close(tcp);
		return;
	}

	evbuf_event_add(&tcp->wbuf);
}

void
session_shutdown(struct nbr *nbr, uint32_t status, uint32_t msg_id,
    uint32_t msg_type)
{
	switch (nbr->state) {
	case NBR_STA_PRESENT:
		if (nbr_pending_connect(nbr))
			THREAD_OFF(nbr->ev_connect);
		break;
	case NBR_STA_INITIAL:
	case NBR_STA_OPENREC:
	case NBR_STA_OPENSENT:
		/* update SNMP session counters during initialization */
		leconf->stats.session_attempts++;
		send_notification(nbr->tcp, status, msg_id, msg_type);

		nbr_fsm(nbr, NBR_EVT_CLOSE_SESSION);
		break;
	case NBR_STA_OPER:
		send_notification(nbr->tcp, status, msg_id, msg_type);

		nbr_fsm(nbr, NBR_EVT_CLOSE_SESSION);
		break;
	default:
		fatalx("session_shutdown: unknown neighbor state");
	}
}

void
session_close(struct nbr *nbr)
{
	log_debug("%s: closing session with lsr-id %pI4", __func__,
	    &nbr->id);

	ldp_sync_fsm_nbr_event(nbr, LDP_SYNC_EVT_SESSION_CLOSE);

	tcp_close(nbr->tcp);
	nbr_stop_ktimer(nbr);
	nbr_stop_ktimeout(nbr);
	nbr_stop_itimeout(nbr);
}

static ssize_t
session_get_pdu(struct ibuf_read *r, char **b)
{
	struct ldp_hdr	l;
	size_t		av, dlen, left;

	av = r->wpos;
	if (av < sizeof(l))
		return (0);

	memcpy(&l, r->buf, sizeof(l));
	dlen = ntohs(l.length) + LDP_HDR_DEAD_LEN;
	if (dlen > av)
		return (0);

	if ((*b = malloc(dlen)) == NULL)
		return (-1);

	memcpy(*b, r->buf, dlen);
	if (dlen < av) {
		left = av - dlen;
		memmove(r->buf, r->buf + dlen, left);
		r->wpos = left;
	} else
		r->wpos = 0;

	return (dlen);
}

struct tcp_conn *
tcp_new(int fd, struct nbr *nbr)
{
	struct tcp_conn		*tcp;
	struct sockaddr_storage	 ss;
	socklen_t		 len = sizeof(ss);

	if ((tcp = calloc(1, sizeof(*tcp))) == NULL)
		fatal(__func__);

	tcp->fd = fd;
	evbuf_init(&tcp->wbuf, tcp->fd, session_write, tcp);

	if (nbr) {
		if ((tcp->rbuf = calloc(1, sizeof(struct ibuf_read))) == NULL)
			fatal(__func__);

		thread_add_read(master, session_read, nbr, tcp->fd, &tcp->rev);
		tcp->nbr = nbr;
	}

	if (getsockname(fd, (struct sockaddr *)&ss, &len) != 0)
		log_warn("%s: getsockname", __func__);
	else
		sa2addr((struct sockaddr *)&ss, NULL, NULL, &tcp->lport);
	if (getpeername(fd, (struct sockaddr *)&ss, &len) != 0)
		log_warn("%s: getpeername", __func__);
	else
		sa2addr((struct sockaddr *)&ss, NULL, NULL, &tcp->rport);

	return (tcp);
}

static void
tcp_close(struct tcp_conn *tcp)
{
	/* try to flush write buffer */
	msgbuf_write(&tcp->wbuf.wbuf);
	evbuf_clear(&tcp->wbuf);

	if (tcp->nbr) {
		THREAD_OFF(tcp->rev);
		free(tcp->rbuf);
		tcp->nbr->tcp = NULL;
	}

	close(tcp->fd);
	accept_unpause();
	free(tcp);
}

static struct pending_conn *
pending_conn_new(int fd, int af, union ldpd_addr *addr)
{
	struct pending_conn	*pconn;

	if ((pconn = calloc(1, sizeof(*pconn))) == NULL)
		fatal(__func__);

	pconn->fd = fd;
	pconn->af = af;
	pconn->addr = *addr;
	TAILQ_INSERT_TAIL(&global.pending_conns, pconn, entry);
	pconn->ev_timeout = NULL;
	thread_add_timer(master, pending_conn_timeout, pconn, PENDING_CONN_TIMEOUT,
			 &pconn->ev_timeout);

	return (pconn);
}

void
pending_conn_del(struct pending_conn *pconn)
{
	THREAD_OFF(pconn->ev_timeout);
	TAILQ_REMOVE(&global.pending_conns, pconn, entry);
	free(pconn);
}

struct pending_conn *
pending_conn_find(int af, union ldpd_addr *addr)
{
	struct pending_conn	*pconn;

	TAILQ_FOREACH(pconn, &global.pending_conns, entry)
		if (af == pconn->af &&
		    ldp_addrcmp(af, addr, &pconn->addr) == 0)
			return (pconn);

	return (NULL);
}

static void pending_conn_timeout(struct event *thread)
{
	struct pending_conn	*pconn = THREAD_ARG(thread);
	struct tcp_conn		*tcp;

	pconn->ev_timeout = NULL;

	log_debug("%s: no adjacency with remote end: %s", __func__,
	    log_addr(pconn->af, &pconn->addr));

	/*
	 * Create a write buffer detached from any neighbor to send a
	 * notification message reliably.
	 */
	tcp = tcp_new(pconn->fd, NULL);
	send_notification(tcp, S_NO_HELLO, 0, 0);
	msgbuf_write(&tcp->wbuf.wbuf);

	pending_conn_del(pconn);
}