xref: /netbsd-src/usr.sbin/traceroute6/traceroute6.c (revision 95d875fb90b1458e4f1de6950286ddcd6644bc61)
1 /*	$NetBSD: traceroute6.c,v 1.5 1999/09/03 01:49:16 itojun Exp $	*/
2 
3 /*
4  * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
5  * All rights reserved.
6  *
7  * Redistribution and use in source and binary forms, with or without
8  * modification, are permitted provided that the following conditions
9  * are met:
10  * 1. Redistributions of source code must retain the above copyright
11  *    notice, this list of conditions and the following disclaimer.
12  * 2. Redistributions in binary form must reproduce the above copyright
13  *    notice, this list of conditions and the following disclaimer in the
14  *    documentation and/or other materials provided with the distribution.
15  * 3. Neither the name of the project nor the names of its contributors
16  *    may be used to endorse or promote products derived from this software
17  *    without specific prior written permission.
18  *
19  * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
20  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22  * ARE DISCLAIMED.  IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
23  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
24  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
25  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
26  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
28  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29  * SUCH DAMAGE.
30  */
31 
32 /*-
33  * Copyright (c) 1990, 1993
34  *	The Regents of the University of California.  All rights reserved.
35  *
36  * This code is derived from software contributed to Berkeley by
37  * Van Jacobson.
38  *
39  * Redistribution and use in source and binary forms, with or without
40  * modification, are permitted provided that the following conditions
41  * are met:
42  * 1. Redistributions of source code must retain the above copyright
43  *    notice, this list of conditions and the following disclaimer.
44  * 2. Redistributions in binary form must reproduce the above copyright
45  *    notice, this list of conditions and the following disclaimer in the
46  *    documentation and/or other materials provided with the distribution.
47  * 3. All advertising materials mentioning features or use of this software
48  *    must display the following acknowledgement:
49  *	This product includes software developed by the University of
50  *	California, Berkeley and its contributors.
51  * 4. Neither the name of the University nor the names of its contributors
52  *    may be used to endorse or promote products derived from this software
53  *    without specific prior written permission.
54  *
55  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
56  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
57  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
58  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
59  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
60  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
61  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
62  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
63  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
64  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
65  * SUCH DAMAGE.
66  */
67 
68 #if 0
69 #ifndef lint
70 static char copyright[] =
71 "@(#) Copyright (c) 1990, 1993\n\
72 	The Regents of the University of California.  All rights reserved.\n";
73 #endif /* not lint */
74 
75 #ifndef lint
76 static char sccsid[] = "@(#)traceroute.c	8.1 (Berkeley) 6/6/93";
77 #endif /* not lint */
78 #else
79 #include <sys/cdefs.h>
80 #ifndef lint
81 __RCSID("$NetBSD: traceroute6.c,v 1.5 1999/09/03 01:49:16 itojun Exp $");
82 #endif
83 #endif
84 
85 /*
86  * traceroute host  - trace the route ip packets follow going to "host".
87  *
88  * Attempt to trace the route an ip packet would follow to some
89  * internet host.  We find out intermediate hops by launching probe
90  * packets with a small ttl (time to live) then listening for an
91  * icmp "time exceeded" reply from a gateway.  We start our probes
92  * with a ttl of one and increase by one until we get an icmp "port
93  * unreachable" (which means we got to "host") or hit a max (which
94  * defaults to 30 hops & can be changed with the -m flag).  Three
95  * probes (change with -q flag) are sent at each ttl setting and a
96  * line is printed showing the ttl, address of the gateway and
97  * round trip time of each probe.  If the probe answers come from
98  * different gateways, the address of each responding system will
99  * be printed.  If there is no response within a 5 sec. timeout
100  * interval (changed with the -w flag), a "*" is printed for that
101  * probe.
102  *
103  * Probe packets are UDP format.  We don't want the destination
104  * host to process them so the destination port is set to an
105  * unlikely value (if some clod on the destination is using that
106  * value, it can be changed with the -p flag).
107  *
108  * A sample use might be:
109  *
110  *     [yak 71]% traceroute nis.nsf.net.
111  *     traceroute to nis.nsf.net (35.1.1.48), 30 hops max, 56 byte packet
112  *      1  helios.ee.lbl.gov (128.3.112.1)  19 ms  19 ms  0 ms
113  *      2  lilac-dmc.Berkeley.EDU (128.32.216.1)  39 ms  39 ms  19 ms
114  *      3  lilac-dmc.Berkeley.EDU (128.32.216.1)  39 ms  39 ms  19 ms
115  *      4  ccngw-ner-cc.Berkeley.EDU (128.32.136.23)  39 ms  40 ms  39 ms
116  *      5  ccn-nerif22.Berkeley.EDU (128.32.168.22)  39 ms  39 ms  39 ms
117  *      6  128.32.197.4 (128.32.197.4)  40 ms  59 ms  59 ms
118  *      7  131.119.2.5 (131.119.2.5)  59 ms  59 ms  59 ms
119  *      8  129.140.70.13 (129.140.70.13)  99 ms  99 ms  80 ms
120  *      9  129.140.71.6 (129.140.71.6)  139 ms  239 ms  319 ms
121  *     10  129.140.81.7 (129.140.81.7)  220 ms  199 ms  199 ms
122  *     11  nic.merit.edu (35.1.1.48)  239 ms  239 ms  239 ms
123  *
124  * Note that lines 2 & 3 are the same.  This is due to a buggy
125  * kernel on the 2nd hop system -- lbl-csam.arpa -- that forwards
126  * packets with a zero ttl.
127  *
128  * A more interesting example is:
129  *
130  *     [yak 72]% traceroute allspice.lcs.mit.edu.
131  *     traceroute to allspice.lcs.mit.edu (18.26.0.115), 30 hops max
132  *      1  helios.ee.lbl.gov (128.3.112.1)  0 ms  0 ms  0 ms
133  *      2  lilac-dmc.Berkeley.EDU (128.32.216.1)  19 ms  19 ms  19 ms
134  *      3  lilac-dmc.Berkeley.EDU (128.32.216.1)  39 ms  19 ms  19 ms
135  *      4  ccngw-ner-cc.Berkeley.EDU (128.32.136.23)  19 ms  39 ms  39 ms
136  *      5  ccn-nerif22.Berkeley.EDU (128.32.168.22)  20 ms  39 ms  39 ms
137  *      6  128.32.197.4 (128.32.197.4)  59 ms  119 ms  39 ms
138  *      7  131.119.2.5 (131.119.2.5)  59 ms  59 ms  39 ms
139  *      8  129.140.70.13 (129.140.70.13)  80 ms  79 ms  99 ms
140  *      9  129.140.71.6 (129.140.71.6)  139 ms  139 ms  159 ms
141  *     10  129.140.81.7 (129.140.81.7)  199 ms  180 ms  300 ms
142  *     11  129.140.72.17 (129.140.72.17)  300 ms  239 ms  239 ms
143  *     12  * * *
144  *     13  128.121.54.72 (128.121.54.72)  259 ms  499 ms  279 ms
145  *     14  * * *
146  *     15  * * *
147  *     16  * * *
148  *     17  * * *
149  *     18  ALLSPICE.LCS.MIT.EDU (18.26.0.115)  339 ms  279 ms  279 ms
150  *
151  * (I start to see why I'm having so much trouble with mail to
152  * MIT.)  Note that the gateways 12, 14, 15, 16 & 17 hops away
153  * either don't send ICMP "time exceeded" messages or send them
154  * with a ttl too small to reach us.  14 - 17 are running the
155  * MIT C Gateway code that doesn't send "time exceeded"s.  God
156  * only knows what's going on with 12.
157  *
158  * The silent gateway 12 in the above may be the result of a bug in
159  * the 4.[23]BSD network code (and its derivatives):  4.x (x <= 3)
160  * sends an unreachable message using whatever ttl remains in the
161  * original datagram.  Since, for gateways, the remaining ttl is
162  * zero, the icmp "time exceeded" is guaranteed to not make it back
163  * to us.  The behavior of this bug is slightly more interesting
164  * when it appears on the destination system:
165  *
166  *      1  helios.ee.lbl.gov (128.3.112.1)  0 ms  0 ms  0 ms
167  *      2  lilac-dmc.Berkeley.EDU (128.32.216.1)  39 ms  19 ms  39 ms
168  *      3  lilac-dmc.Berkeley.EDU (128.32.216.1)  19 ms  39 ms  19 ms
169  *      4  ccngw-ner-cc.Berkeley.EDU (128.32.136.23)  39 ms  40 ms  19 ms
170  *      5  ccn-nerif35.Berkeley.EDU (128.32.168.35)  39 ms  39 ms  39 ms
171  *      6  csgw.Berkeley.EDU (128.32.133.254)  39 ms  59 ms  39 ms
172  *      7  * * *
173  *      8  * * *
174  *      9  * * *
175  *     10  * * *
176  *     11  * * *
177  *     12  * * *
178  *     13  rip.Berkeley.EDU (128.32.131.22)  59 ms !  39 ms !  39 ms !
179  *
180  * Notice that there are 12 "gateways" (13 is the final
181  * destination) and exactly the last half of them are "missing".
182  * What's really happening is that rip (a Sun-3 running Sun OS3.5)
183  * is using the ttl from our arriving datagram as the ttl in its
184  * icmp reply.  So, the reply will time out on the return path
185  * (with no notice sent to anyone since icmp's aren't sent for
186  * icmp's) until we probe with a ttl that's at least twice the path
187  * length.  I.e., rip is really only 7 hops away.  A reply that
188  * returns with a ttl of 1 is a clue this problem exists.
189  * Traceroute prints a "!" after the time if the ttl is <= 1.
190  * Since vendors ship a lot of obsolete (DEC's Ultrix, Sun 3.x) or
191  * non-standard (HPUX) software, expect to see this problem
192  * frequently and/or take care picking the target host of your
193  * probes.
194  *
195  * Other possible annotations after the time are !H, !N, !P (got a host,
196  * network or protocol unreachable, respectively), !S or !F (source
197  * route failed or fragmentation needed -- neither of these should
198  * ever occur and the associated gateway is busted if you see one).  If
199  * almost all the probes result in some kind of unreachable, traceroute
200  * will give up and exit.
201  *
202  * Notes
203  * -----
204  * This program must be run by root or be setuid.  (I suggest that
205  * you *don't* make it setuid -- casual use could result in a lot
206  * of unnecessary traffic on our poor, congested nets.)
207  *
208  * This program requires a kernel mod that does not appear in any
209  * system available from Berkeley:  A raw ip socket using proto
210  * IPPROTO_RAW must interpret the data sent as an ip datagram (as
211  * opposed to data to be wrapped in a ip datagram).  See the README
212  * file that came with the source to this program for a description
213  * of the mods I made to /sys/netinet/raw_ip.c.  Your mileage may
214  * vary.  But, again, ANY 4.x (x < 4) BSD KERNEL WILL HAVE TO BE
215  * MODIFIED TO RUN THIS PROGRAM.
216  *
217  * The udp port usage may appear bizarre (well, ok, it is bizarre).
218  * The problem is that an icmp message only contains 8 bytes of
219  * data from the original datagram.  8 bytes is the size of a udp
220  * header so, if we want to associate replies with the original
221  * datagram, the necessary information must be encoded into the
222  * udp header (the ip id could be used but there's no way to
223  * interlock with the kernel's assignment of ip id's and, anyway,
224  * it would have taken a lot more kernel hacking to allow this
225  * code to set the ip id).  So, to allow two or more users to
226  * use traceroute simultaneously, we use this task's pid as the
227  * source port (the high bit is set to move the port number out
228  * of the "likely" range).  To keep track of which probe is being
229  * replied to (so times and/or hop counts don't get confused by a
230  * reply that was delayed in transit), we increment the destination
231  * port number before each probe.
232  *
233  * Don't use this as a coding example.  I was trying to find a
234  * routing problem and this code sort-of popped out after 48 hours
235  * without sleep.  I was amazed it ever compiled, much less ran.
236  *
237  * I stole the idea for this program from Steve Deering.  Since
238  * the first release, I've learned that had I attended the right
239  * IETF working group meetings, I also could have stolen it from Guy
240  * Almes or Matt Mathis.  I don't know (or care) who came up with
241  * the idea first.  I envy the originators' perspicacity and I'm
242  * glad they didn't keep the idea a secret.
243  *
244  * Tim Seaver, Ken Adelman and C. Philip Wood provided bug fixes and/or
245  * enhancements to the original distribution.
246  *
247  * I've hacked up a round-trip-route version of this that works by
248  * sending a loose-source-routed udp datagram through the destination
249  * back to yourself.  Unfortunately, SO many gateways botch source
250  * routing, the thing is almost worthless.  Maybe one day...
251  *
252  *  -- Van Jacobson (van@helios.ee.lbl.gov)
253  *     Tue Dec 20 03:50:13 PST 1988
254  */
255 
256 #include <sys/param.h>
257 #include <sys/time.h>
258 #include <sys/socket.h>
259 #include <sys/uio.h>
260 #include <sys/file.h>
261 #include <sys/ioctl.h>
262 
263 #include <netinet/in.h>
264 
265 #include <arpa/inet.h>
266 
267 #include <netdb.h>
268 #include <stdio.h>
269 #include <err.h>
270 #include <errno.h>
271 #include <stdlib.h>
272 #include <string.h>
273 #include <unistd.h>
274 
275 #include <netinet/ip6.h>
276 #include <netinet/icmp6.h>
277 #include <netinet/udp.h>
278 
279 #ifdef IPSEC
280 #include <net/route.h>
281 #include <netinet6/ipsec.h>
282 #endif
283 
284 #define	freehostent(hp)
285 #define DUMMY_PORT 10010
286 
287 #define	MAXPACKET	65535	/* max ip packet size */
288 #ifndef MAXHOSTNAMELEN
289 #define MAXHOSTNAMELEN	64
290 #endif
291 
292 #ifndef FD_SET
293 #define NFDBITS         (8*sizeof(fd_set))
294 #define FD_SETSIZE      NFDBITS
295 #define FD_SET(n, p)    ((p)->fds_bits[(n)/NFDBITS] |= (1 << ((n) % NFDBITS)))
296 #define FD_CLR(n, p)    ((p)->fds_bits[(n)/NFDBITS] &= ~(1 << ((n) % NFDBITS)))
297 #define FD_ISSET(n, p)  ((p)->fds_bits[(n)/NFDBITS] & (1 << ((n) % NFDBITS)))
298 #define FD_ZERO(p)      bzero((char *)(p), sizeof(*(p)))
299 #endif
300 
301 #define Fprintf (void)fprintf
302 #define Sprintf (void)sprintf
303 #define Printf (void)printf
304 
305 /*
306  * format of a (udp) probe packet.
307  */
308 struct opacket {
309 	u_char seq;		/* sequence number of this packet */
310 	u_char hops;		/* hop limit of the packet */
311 	struct timeval tv;	/* time packet left */
312 };
313 
314 u_char	packet[512];		/* last inbound (icmp) packet */
315 struct opacket	*outpacket;	/* last output (udp) packet */
316 
317 int	main __P((int, char *[]));
318 int	wait_for_reply __P((int, struct msghdr *));
319 void	send_probe __P((int, int));
320 struct udphdr *get_udphdr __P((struct ip6_hdr *, u_char *));
321 int	get_hoplim __P((struct msghdr *));
322 double	deltaT __P((struct timeval *, struct timeval *));
323 char	*pr_type __P((int));
324 int	packet_ok __P((struct msghdr *, int, int));
325 void	print __P((struct msghdr *, int));
326 void	tvsub __P((struct timeval *, struct timeval *));
327 char	*inetname __P((struct in6_addr *));
328 void	usage __P((void));
329 
330 int rcvsock;			/* receive (icmp) socket file descriptor */
331 int sndsock;			/* send (udp) socket file descriptor */
332 struct timezone tz;		/* leftover */
333 
334 struct msghdr rcvmhdr;
335 struct iovec rcviov[2];
336 int rcvhlim;
337 struct in6_pktinfo *rcvpktinfo;
338 
339 struct sockaddr_in6 Src, Dst, Rcv;
340 struct sockaddr_in6 *src = &Src, *dst = &Dst, *rcv = &Rcv;
341 int datalen;			/* How much data */
342 char rtbuf[1024];	/*XXX*/
343 struct cmsghdr *cmsg;
344 
345 char *source = 0;
346 char *hostname;
347 
348 int nprobes = 3;
349 int max_hops = 30;
350 u_short ident;
351 u_short port = 32768+666;	/* start udp dest port # for probe packets */
352 int options;			/* socket options */
353 int verbose;
354 int waittime = 5;		/* time to wait for response (in seconds) */
355 int nflag;			/* print addresses numerically */
356 int lflag;			/* print both numerical address & hostname */
357 
358 char ntop_buf[INET6_ADDRSTRLEN]; /* for inet_ntop() */
359 
360 int
361 main(argc, argv)
362 	int argc;
363 	char *argv[];
364 {
365 	extern char *optarg;
366 	extern int optind;
367 	struct hostent *hp;
368 	int ch, i, on, probe, seq, hops;
369 	static u_char rcvcmsgbuf[CMSG_SPACE(sizeof(struct in6_pktinfo))
370 				+ CMSG_SPACE(sizeof(int))];
371 
372 	on = 1;
373 	seq = 0;
374 
375 	while ((ch = getopt(argc, argv, "dlm:np:q:rs:w:vg:")) != EOF)
376 		switch(ch) {
377 		case 'd':
378 			options |= SO_DEBUG;
379 			break;
380 		case 'l':
381 			lflag++;
382 			break;
383 		case 'g':
384 #if 0
385 			hp = getipnodebyname(optarg, AF_INET6, 0, &h_errno);
386 #else
387 			hp = gethostbyname2(optarg, AF_INET6);
388 #endif
389 			if (hp == NULL) {
390 				Fprintf(stderr,
391 				    "traceroute6: unknown host %s\n", optarg);
392 				exit(1);
393 			}
394 			if (cmsg == NULL)
395 				cmsg = inet6_rthdr_init(rtbuf, IPV6_RTHDR_TYPE_0);
396 			inet6_rthdr_add(cmsg, (struct in6_addr *)hp->h_addr, IPV6_RTHDR_LOOSE);
397 			break;
398 		case 'm':
399 			max_hops = atoi(optarg);
400 			if (max_hops <= 1) {
401 				Fprintf(stderr,
402 				    "traceroute6: max hoplimit must be >1.\n");
403 				exit(1);
404 			}
405 			break;
406 		case 'n':
407 			nflag++;
408 			break;
409 		case 'p':
410 			port = atoi(optarg);
411 			if (port < 1) {
412 				Fprintf(stderr,
413 				    "traceroute6: port must be >0.\n");
414 				exit(1);
415 			}
416 			break;
417 		case 'q':
418 			nprobes = atoi(optarg);
419 			if (nprobes < 1) {
420 				Fprintf(stderr,
421 				    "traceroute6: nprobes must be >0.\n");
422 				exit(1);
423 			}
424 			break;
425 		case 'r':
426 			options |= SO_DONTROUTE;
427 			break;
428 		case 's':
429 			/*
430 			 * set the ip source address of the outbound
431 			 * probe (e.g., on a multi-homed host).
432 			 */
433 			source = optarg;
434 			break;
435 		case 'v':
436 			verbose++;
437 			break;
438 		case 'w':
439 			waittime = atoi(optarg);
440 			if (waittime <= 1) {
441 				Fprintf(stderr,
442 				    "traceroute6: wait must be >1 sec.\n");
443 				exit(1);
444 			}
445 			break;
446 		default:
447 			usage();
448 		}
449 	argc -= optind;
450 	argv += optind;
451 
452 	if (argc < 1)
453 		usage();
454 
455 #if 1
456 	setvbuf(stdout, NULL, _IOLBF, BUFSIZ);
457 #else
458 	setlinebuf (stdout);
459 #endif
460 
461 	(void) bzero((char *)dst, sizeof(Dst));
462 	Dst.sin6_family = AF_INET6;
463 
464 	hp = (struct hostent *)gethostbyname2(*argv, AF_INET6);
465 	if (hp == NULL) {
466 		if (inet_pton(AF_INET6, *argv, &Dst.sin6_addr) != 1) {
467 			(void)fprintf(stderr,
468 				      "traceroute6: unknown host %s\n", *argv);
469 			exit(1);
470 		}
471 		hostname = *argv;
472 	} else {
473 		bcopy(hp->h_addr, (caddr_t)&Dst.sin6_addr, hp->h_length);
474 		hostname = strdup(hp->h_name);
475 	}
476 	freehostent(hp);
477 
478 	if (*++argv)
479 		datalen = atoi(*argv);
480 	if (datalen < 0 || datalen >= MAXPACKET - sizeof(struct opacket)) {
481 		Fprintf(stderr,
482 		    "traceroute6: packet size must be 0 <= s < %ld.\n",
483 		    (long)(MAXPACKET - sizeof(struct opacket)));
484 		exit(1);
485 	}
486 	datalen += sizeof(struct opacket);
487 	outpacket = (struct opacket *)malloc((unsigned)datalen);
488 	if (! outpacket) {
489 		perror("traceroute6: malloc");
490 		exit(1);
491 	}
492 	(void) bzero((char *)outpacket, datalen);
493 
494 	/*
495 	 * Receive ICMP
496 	 */
497 	if ((rcvsock = socket(AF_INET6, SOCK_RAW, IPPROTO_ICMPV6)) < 0) {
498 		perror("traceroute6: icmp socket");
499 		exit(5);
500 	}
501 	/* initialize msghdr for receiving packets */
502 	rcviov[0].iov_base = (caddr_t)packet;
503 	rcviov[0].iov_len = sizeof(packet);
504 	rcvmhdr.msg_name = (caddr_t)rcv;
505 	rcvmhdr.msg_namelen = sizeof(*rcv);
506 	rcvmhdr.msg_iov = rcviov;
507 	rcvmhdr.msg_iovlen = 1;
508 	rcvmhdr.msg_control = (caddr_t) rcvcmsgbuf;
509 	rcvmhdr.msg_controllen = sizeof(rcvcmsgbuf);
510 
511 	/* specify to tell receiving interface */
512 	if (setsockopt(rcvsock, IPPROTO_IPV6, IPV6_PKTINFO, &on,
513 		       sizeof(on)) < 0)
514 		err(1, "setsockopt(IPV6_PKTINFO)");
515 
516 	/* specify to tell value of hoplimit field of received IP6 hdr */
517 	if (setsockopt(rcvsock, IPPROTO_IPV6, IPV6_HOPLIMIT, &on,
518 		       sizeof(on)) < 0)
519 		err(1, "setsockopt(IPV6_HOPLIMIT)");
520 
521 	if (options & SO_DEBUG)
522 		(void) setsockopt(rcvsock, SOL_SOCKET, SO_DEBUG,
523 				  (char *)&on, sizeof(on));
524 	if (options & SO_DONTROUTE)
525 		(void) setsockopt(rcvsock, SOL_SOCKET, SO_DONTROUTE,
526 				  (char *)&on, sizeof(on));
527 #ifdef IPSEC
528 #ifdef IPSEC_POLICY_IPSEC
529     {
530 	int len;
531 	char buf[16];
532 
533 	/*
534 	 * do not raise error even if setsockopt fails, kernel may have ipsec
535 	 * turned off.
536 	 */
537 	if ((len = ipsec_set_policy(buf, sizeof(buf), "bypass")) < 0)
538 		errx(1, ipsec_strerror());
539 	(void)setsockopt(rcvsock, IPPROTO_IPV6, IPV6_IPSEC_POLICY, buf, len);
540     }
541 #else
542     {
543 	int level = IPSEC_LEVEL_NONE;
544 
545 	(void)setsockopt(rcvsock, IPPROTO_IPV6, IPV6_ESP_TRANS_LEVEL, &level,
546 		sizeof(level));
547 	(void)setsockopt(rcvsock, IPPROTO_IPV6, IPV6_ESP_NETWORK_LEVEL, &level,
548 		sizeof(level));
549 #ifdef IP_AUTH_TRANS_LEVEL
550 	(void)setsockopt(rcvsock, IPPROTO_IPV6, IPV6_AUTH_TRANS_LEVEL, &level,
551 		sizeof(level));
552 #else
553 	(void)setsockopt(rcvsock, IPPROTO_IPV6, IPV6_AUTH_LEVEL, &level,
554 		sizeof(level));
555 #endif
556 #ifdef IP_AUTH_NETWORK_LEVEL
557 	(void)setsockopt(rcvsock, IPPROTO_IPV6, IPV6_AUTH_NETWORK_LEVEL, &level,
558 		sizeof(level));
559 #endif
560     }
561 #endif /*IPSEC_POLICY_IPSEC*/
562 #endif /*IPSEC*/
563 
564 	/*
565 	 * Send UDP
566 	 */
567 	if ((sndsock = socket(AF_INET6, SOCK_DGRAM, 0)) < 0) {
568 		perror("traceroute6: udp socket");
569 		exit(5);
570 	}
571 #ifdef SO_SNDBUF
572 	if (setsockopt(sndsock, SOL_SOCKET, SO_SNDBUF, (char *)&datalen,
573 		       sizeof(datalen)) < 0) {
574 		perror("traceroute6: SO_SNDBUF");
575 		exit(6);
576 	}
577 #endif /* SO_SNDBUF */
578 	if (options & SO_DEBUG)
579 		(void) setsockopt(sndsock, SOL_SOCKET, SO_DEBUG,
580 				  (char *)&on, sizeof(on));
581 	if (options & SO_DONTROUTE)
582 		(void) setsockopt(sndsock, SOL_SOCKET, SO_DONTROUTE,
583 				  (char *)&on, sizeof(on));
584 	if (cmsg != NULL) {
585 		inet6_rthdr_lasthop(cmsg, IPV6_RTHDR_LOOSE);
586 		(void) setsockopt(sndsock, IPPROTO_IPV6, IPV6_PKTOPTIONS,
587 				  rtbuf, cmsg->cmsg_len);
588 	}
589 #ifdef IPSEC
590 #ifdef IPSEC_POLICY_IPSEC
591     {
592 	int len;
593 	char buf[16];
594 
595 	/*
596 	 * do not raise error even if setsockopt fails, kernel may have ipsec
597 	 * turned off.
598 	 */
599 	if ((len = ipsec_set_policy(buf, sizeof(buf), "bypass")) < 0)
600 		errx(1, ipsec_strerror());
601 	(void)setsockopt(sndsock, IPPROTO_IPV6, IPV6_IPSEC_POLICY, buf, len);
602     }
603 #else
604     {
605 	int level = IPSEC_LEVEL_BYPASS;
606 
607 	(void)setsockopt(sndsock, IPPROTO_IPV6, IPV6_ESP_TRANS_LEVEL, &level,
608 		sizeof(level));
609 	(void)setsockopt(sndsock, IPPROTO_IPV6, IPV6_ESP_NETWORK_LEVEL, &level,
610 		sizeof(level));
611 #ifdef IP_AUTH_TRANS_LEVEL
612 	(void)setsockopt(sndsock, IPPROTO_IPV6, IPV6_AUTH_TRANS_LEVEL, &level,
613 		sizeof(level));
614 #else
615 	(void)setsockopt(sndsock, IPPROTO_IPV6, IPV6_AUTH_LEVEL, &level,
616 		sizeof(level));
617 #endif
618 #ifdef IP_AUTH_NETWORK_LEVEL
619 	(void)setsockopt(sndsock, IPPROTO_IPV6, IPV6_AUTH_NETWORK_LEVEL, &level,
620 		sizeof(level));
621 #endif
622     }
623 #endif /*IPSEC_POLICY_IPSEC*/
624 #endif /*IPSEC*/
625 
626 	/*
627 	 * Source selection
628 	 */
629 	bzero((char *)src, sizeof(Src));
630 	if (source) {
631 		if (inet_pton(AF_INET6, source, &Src.sin6_addr) != 1) {
632 			Printf("traceroute6: unknown host %s\n", source);
633 			exit(1);
634 		}
635 	} else {
636 		struct sockaddr_in6 Nxt;
637 		int dummy, len;
638 
639 		len = sizeof(Src);
640 		Nxt = Dst;
641 		Nxt.sin6_port = htons(DUMMY_PORT);
642 		if (cmsg != NULL)
643 			bcopy(inet6_rthdr_getaddr(cmsg, 1), &Nxt.sin6_addr,
644 			      sizeof(Nxt.sin6_addr));
645 		if ((dummy = socket(AF_INET6, SOCK_DGRAM, 0)) < 0) {
646 			perror("socket") ;
647 		}
648 		if(-1 == connect(dummy, (struct sockaddr *)&Nxt, sizeof(Nxt)))
649 			perror("connect");
650 		if(-1 == getsockname(dummy, (struct sockaddr *)src, &len)) {
651 			perror("getsockname");
652 			printf("%d\n", errno);
653 		}
654 		close(dummy) ;
655 	}
656 	ident = (getpid() & 0xffff) | 0x8000;
657 	Src.sin6_family = AF_INET6;
658 	Src.sin6_port = htons(ident);
659 	if (bind(sndsock, (struct sockaddr *)src, sizeof(Src))  < 0){
660 		perror ("traceroute6: bind:");
661 		exit (1);
662 	}
663 
664 	/*
665 	 * Message to users
666 	 */
667 	Fprintf(stderr, "traceroute to %s (%s)", hostname,
668 		inet_ntop(AF_INET6, &Dst.sin6_addr,
669 			  ntop_buf, sizeof(ntop_buf)));
670 	if (source)
671 		Fprintf(stderr, " from %s", source);
672 	Fprintf(stderr, ", %d hops max, %d byte packets\n", max_hops, datalen);
673 	(void) fflush(stderr);
674 
675 	/*
676 	 * Main loop
677 	 */
678 	for (hops = 1; hops <= max_hops; ++hops) {
679 		struct in6_addr lastaddr;
680 		int got_there = 0;
681 		int unreachable = 0;
682 
683 		Printf("%2d ", hops);
684 		bzero(&lastaddr, sizeof(lastaddr));
685 		for (probe = 0; probe < nprobes; ++probe) {
686 			int cc;
687 			struct timeval t1, t2;
688 			struct timezone tz;
689 
690 			(void) gettimeofday(&t1, &tz);
691 			send_probe(++seq, hops);
692 			while ((cc = wait_for_reply(rcvsock, &rcvmhdr))) {
693 				(void) gettimeofday(&t2, &tz);
694 				if ((i = packet_ok(&rcvmhdr, cc, seq))) {
695 					if (! IN6_ARE_ADDR_EQUAL(&Rcv.sin6_addr,
696 							    &lastaddr)) {
697 						print(&rcvmhdr, cc);
698 						lastaddr = Rcv.sin6_addr;
699 					}
700 					Printf("  %g ms", deltaT(&t1, &t2));
701 					switch(i - 1) {
702 					case ICMP6_DST_UNREACH_NOROUTE:
703 						++unreachable;
704 						Printf(" !N");
705 						break;
706 					case ICMP6_DST_UNREACH_ADMIN:
707 						++unreachable;
708 						Printf(" !P");
709 						break;
710 					case ICMP6_DST_UNREACH_NOTNEIGHBOR:
711 						++unreachable;
712 						Printf(" !S");
713 						break;
714 					case ICMP6_DST_UNREACH_ADDR:
715 						++unreachable;
716 						Printf(" !A");
717 						break;
718 					case ICMP6_DST_UNREACH_NOPORT:
719 						if (rcvhlim >= 0 &&
720 						    rcvhlim <= 1)
721 							Printf(" !");
722 						++got_there;
723 						break;
724 					}
725 					break;
726 				}
727 			}
728 			if (cc == 0)
729 				Printf(" *");
730 			(void) fflush(stdout);
731 		}
732 		putchar('\n');
733 		if (got_there ||
734 		    (unreachable > 0 && unreachable >= ((nprobes + 1) / 2))) {
735 			exit(0);
736 		}
737 	}
738 
739 	exit(0);
740 }
741 
742 int
743 wait_for_reply(sock, mhdr)
744 	int sock;
745 	struct msghdr *mhdr;
746 {
747 	fd_set fds;
748 	struct timeval wait;
749 	int cc = 0;
750 
751 	FD_ZERO(&fds);
752 	FD_SET(sock, &fds);
753 	wait.tv_sec = waittime; wait.tv_usec = 0;
754 
755 	if (select(sock+1, &fds, (fd_set *)0, (fd_set *)0, &wait) > 0)
756 		cc = recvmsg(rcvsock, mhdr, 0);
757 
758 	return(cc);
759 }
760 
761 
762 void
763 send_probe(seq, hops)
764 	int seq, hops;
765 {
766 	struct opacket *op = outpacket;
767 	int i;
768 
769 	if(setsockopt(sndsock, IPPROTO_IPV6, IPV6_UNICAST_HOPS,
770 		      (char *)&hops, sizeof(hops)) < 0) {
771 		perror("setsockopt IPV6_UNICAST_HOPS");
772 	}
773 
774 	Dst.sin6_port = htons(port + seq);
775 
776 	op->seq = seq;
777 	op->hops = hops;
778 	(void) gettimeofday(&op->tv, &tz);
779 
780 	i = sendto(sndsock, (char *)outpacket, datalen , 0,
781 		   (struct sockaddr *)dst, sizeof(Dst));
782 	if (i < 0 || i != datalen)  {
783 		if (i<0)
784 			perror("sendto");
785 		Printf("traceroute6: wrote %s %d chars, ret=%d\n", hostname,
786 		       datalen, i);
787 		(void) fflush(stdout);
788 	}
789 }
790 
791 int
792 get_hoplim(mhdr)
793 	struct msghdr *mhdr;
794 {
795 	struct cmsghdr *cm;
796 
797 	for (cm = (struct cmsghdr *)CMSG_FIRSTHDR(mhdr); cm;
798 	     cm = (struct cmsghdr *)CMSG_NXTHDR(mhdr, cm)) {
799 		if (cm->cmsg_level == IPPROTO_IPV6 &&
800 		    cm->cmsg_type == IPV6_HOPLIMIT &&
801 		    cm->cmsg_len == CMSG_LEN(sizeof(int)))
802 			return(*(int *)CMSG_DATA(cm));
803 	}
804 
805 	return(-1);
806 }
807 
808 double
809 deltaT(t1p, t2p)
810 	struct timeval *t1p, *t2p;
811 {
812 	register double dt;
813 
814 	dt = (double)(t2p->tv_sec - t1p->tv_sec) * 1000.0 +
815 	     (double)(t2p->tv_usec - t1p->tv_usec) / 1000.0;
816 	return (dt);
817 }
818 
819 
820 /*
821  * Convert an ICMP "type" field to a printable string.
822  */
823 char *
824 pr_type(t0)
825 	int t0;
826 {
827 	u_char t = t0 & 0xff;
828 	char *cp;
829 
830 	switch (t) {
831 	case ICMP6_DST_UNREACH:
832 		cp = "Destination Unreachable";
833 		break;
834 	case ICMP6_PACKET_TOO_BIG:
835 		cp = "Pakcet Too Big";
836 		break;
837 	case ICMP6_TIME_EXCEEDED:
838 		cp = "Time Exceeded";
839 		break;
840 	case ICMP6_PARAM_PROB:
841 		cp = "Parameter Problem";
842 		break;
843 	case ICMP6_ECHO_REQUEST:
844 		cp = "Echo Request";
845 		break;
846 	case ICMP6_ECHO_REPLY:
847 		cp = "Echo Reply";
848 		break;
849 	case ICMP6_MEMBERSHIP_QUERY:
850 		cp = "Group Membership Query";
851 		break;
852 	case ICMP6_MEMBERSHIP_REPORT:
853 		cp = "Group Membership Report";
854 		break;
855 	case ICMP6_MEMBERSHIP_REDUCTION:
856 		cp = "Group Membership Reduction";
857 		break;
858 	case ND_ROUTER_SOLICIT:
859 		cp = "Router Solicitation";
860 		break;
861 	case ND_ROUTER_ADVERT:
862 		cp = "Router Advertisement";
863 		break;
864 	case ND_NEIGHBOR_SOLICIT:
865 		cp = "Neighbor Solicitation";
866 		break;
867 	case ND_NEIGHBOR_ADVERT:
868 		cp = "Neighbor Advertisement";
869 		break;
870 	case ND_REDIRECT:
871 		cp = "Ridirect";
872 		break;
873 	default:
874 		cp = "Unknown";
875 		break;
876 	}
877 	return cp;
878 }
879 
880 
881 int
882 packet_ok(mhdr, cc, seq)
883 	struct msghdr *mhdr;
884 	int cc;
885 	int seq;
886 {
887 	register struct icmp6_hdr *icp;
888 	struct sockaddr_in6 *from = (struct sockaddr_in6 *)mhdr->msg_name;
889 	u_char type, code;
890 	char *buf = (char *)mhdr->msg_iov[0].iov_base;
891 	struct cmsghdr *cm;
892 	int *hlimp;
893 
894 #ifdef OLDRAWSOCKET
895 	int hlen;
896 	struct ip6_hdr *ip;
897 #endif
898 
899 #ifdef OLDRAWSOCKET
900 	ip = (struct ip6_hdr *) buf;
901 	hlen = sizeof(struct ip6_hdr);
902 	if (cc < hlen + sizeof(struct icmp6_hdr)) {
903 		if (verbose)
904 			Printf("packet too short (%d bytes) from %s\n", cc,
905 				inet_ntop(AF_INET6, &from->sin6_addr,
906 					   ntop_buf, sizeof(ntop_buf)));
907 		return (0);
908 	}
909 	cc -= hlen;
910 	icp = (struct icmp6_hdr *)(buf + hlen);
911 #else
912 	if (cc < sizeof(struct icmp6_hdr)) {
913 		if (verbose)
914 			Printf("data too short (%d bytes) from %s\n", cc,
915 			       inet_ntop(AF_INET6, &from->sin6_addr,
916 					 ntop_buf, sizeof(ntop_buf)));
917 		return(0);
918 	}
919 	icp = (struct icmp6_hdr *)buf;
920 #endif
921 	/* get optional information via advanced API */
922 	rcvpktinfo = NULL;
923 	hlimp = NULL;
924 	for (cm = (struct cmsghdr *)CMSG_FIRSTHDR(mhdr); cm;
925 	     cm = (struct cmsghdr *)CMSG_NXTHDR(mhdr, cm)) {
926 		if (cm->cmsg_level == IPPROTO_IPV6 &&
927 		    cm->cmsg_type == IPV6_PKTINFO &&
928 		    cm->cmsg_len ==
929 		    CMSG_LEN(sizeof(struct in6_pktinfo)))
930 			rcvpktinfo = (struct in6_pktinfo *)(CMSG_DATA(cm));
931 
932 		if (cm->cmsg_level == IPPROTO_IPV6 &&
933 		    cm->cmsg_type == IPV6_HOPLIMIT &&
934 		    cm->cmsg_len == CMSG_LEN(sizeof(int)))
935 			hlimp = (int *)CMSG_DATA(cm);
936 	}
937 	if (rcvpktinfo == NULL || hlimp == NULL) {
938 		warnx("failed to get received hop limit or packet info");
939 		return(0);
940 	}
941 	rcvhlim = *hlimp;
942 
943 	type = icp->icmp6_type;
944 	code = icp->icmp6_code;
945 	if ((type == ICMP6_TIME_EXCEEDED && code == ICMP6_TIME_EXCEED_TRANSIT)
946 	 || type == ICMP6_DST_UNREACH) {
947 		struct ip6_hdr *hip;
948 		struct udphdr *up;
949 
950 		hip = (struct ip6_hdr *)(icp + 1);
951 		if ((up = get_udphdr(hip, (u_char *)(buf + cc))) == NULL) {
952 			if (verbose)
953 				warnx("failed to get upper layer header");
954 			return(0);
955 		}
956 		if (up->uh_sport == htons(ident) &&
957 		    up->uh_dport == htons(port+seq))
958 			return (type == ICMP6_TIME_EXCEEDED ? -1 : code + 1);
959 	}
960 	if (verbose) {
961 		int i;
962 		u_long *lp = (u_long *)(icp + 1);
963 		char sbuf[INET6_ADDRSTRLEN+1], dbuf[INET6_ADDRSTRLEN];
964 
965 		Printf("\n%d bytes from %s to %s", cc,
966 			inet_ntop(AF_INET6, &from->sin6_addr,
967 				   sbuf, sizeof(sbuf)),
968 			inet_ntop(AF_INET6, &rcvpktinfo->ipi6_addr,
969 				   dbuf, sizeof(dbuf)));
970 		Printf(": icmp type %d (%s) code %d\n", type, pr_type(type),
971 		       icp->icmp6_code);
972 		for (i = 4; i < cc ; i += sizeof(long))
973 			Printf("%2d: %8.8x\n", i, (u_int32_t)ntohl(*lp++));
974 	}
975 	return(0);
976 }
977 
978 /*
979  * Increment pointer until find the UDP header.
980  */
981 struct udphdr *
982 get_udphdr(ip6, lim)
983 	struct ip6_hdr *ip6;
984 	u_char *lim;
985 {
986 	u_char *cp = (u_char *)ip6, nh;
987 	int hlen;
988 
989 	if (cp + sizeof(*ip6) >= lim)
990 		return(NULL);
991 
992 	nh = ip6->ip6_nxt;
993 	cp += sizeof(struct ip6_hdr);
994 
995 	while(lim - cp >= 8) {
996 		switch(nh) {
997 		 case IPPROTO_ESP:
998 		 case IPPROTO_TCP:
999 		 case IPPROTO_ICMPV6:
1000 			 return(NULL);
1001 		 case IPPROTO_UDP:
1002 			 return((struct udphdr *)cp);
1003 		 case IPPROTO_FRAGMENT:
1004 			 hlen = sizeof(struct ip6_frag);
1005 			 nh = ((struct ip6_frag *)cp)->ip6f_nxt;
1006 			 break;
1007 		 case IPPROTO_AH:
1008 			 hlen = (((struct ip6_ext *)cp)->ip6e_len + 2) << 2;
1009 			 nh = ((struct ip6_ext *)cp)->ip6e_nxt;
1010 			 break;
1011 		  default:
1012 			  hlen = (((struct ip6_ext *)cp)->ip6e_len + 1) << 3;
1013 			  nh = ((struct ip6_ext *)cp)->ip6e_nxt;
1014 			  break;
1015 		}
1016 
1017 		cp += hlen;
1018 	}
1019 
1020 	return(NULL);
1021 }
1022 
1023 void
1024 print(mhdr, cc)
1025 	struct msghdr *mhdr;
1026 	int cc;
1027 {
1028 	struct sockaddr_in6 *from = (struct sockaddr_in6 *)mhdr->msg_name;
1029 
1030 	if (nflag) {
1031 		Printf(" %s", inet_ntop(AF_INET6, &from->sin6_addr,
1032 					 ntop_buf, sizeof(ntop_buf)));
1033 	}
1034 	else if (lflag) {
1035 		Printf(" %s (%s)", inetname(&from->sin6_addr),
1036 		       inet_ntop(AF_INET6, &from->sin6_addr,
1037 				  ntop_buf, sizeof(ntop_buf)));
1038 	}
1039 	else {
1040 		Printf(" %s", inetname(&from->sin6_addr));
1041 	}
1042 
1043 	if (verbose) {
1044 #ifdef OLDRAWSOCKET
1045 		Printf(" %d bytes to %s", cc,
1046 		       inet_ntop(AF_INET6, &rcvpktinfo->ipi6_addr,
1047 				 ntop_buf, sizeof(ntop_buf)));
1048 #else
1049 		Printf(" %d bytes of data to %s", cc,
1050 		       inet_ntop(AF_INET6, &rcvpktinfo->ipi6_addr,
1051 				 ntop_buf, sizeof(ntop_buf)));
1052 #endif
1053 	}
1054 }
1055 
1056 /*
1057  * Subtract 2 timeval structs:  out = out - in.
1058  * Out is assumed to be >= in.
1059  */
1060 void
1061 tvsub(out, in)
1062 	register struct timeval *out, *in;
1063 {
1064 	if ((out->tv_usec -= in->tv_usec) < 0)   {
1065 		out->tv_sec--;
1066 		out->tv_usec += 1000000;
1067 	}
1068 	out->tv_sec -= in->tv_sec;
1069 }
1070 
1071 
1072 /*
1073  * Construct an Internet address representation.
1074  * If the nflag has been supplied, give
1075  * numeric value, otherwise try for symbolic name.
1076  */
1077 char *
1078 inetname(in)
1079 	struct in6_addr *in;
1080 {
1081 	register char *cp;
1082 	static char line[50];
1083 	struct hostent *hp;
1084 	static char domain[MAXHOSTNAMELEN + 1];
1085 	static int first = 1;
1086 
1087 	if (first && !nflag) {
1088 		first = 0;
1089 		if (gethostname(domain, MAXHOSTNAMELEN) == 0 &&
1090 		    (cp = index(domain, '.')))
1091 			(void) strcpy(domain, cp + 1);
1092 		else
1093 			domain[0] = 0;
1094 	}
1095 	cp = 0;
1096 	if (!nflag) {
1097 		/* hp = (struct hostent *)addr2hostname(in, sizeof(*in), AF_INET6, &herr); */
1098 	  hp = (struct hostent *)gethostbyaddr((const char *)in, sizeof(*in), AF_INET6);
1099 		if (hp) {
1100 			if ((cp = index(hp->h_name, '.')) &&
1101 			    !strcmp(cp + 1, domain))
1102 				*cp = 0;
1103 			cp = hp->h_name;
1104 			freehostent(hp);
1105 		}
1106 	}
1107 	if (cp)
1108 		(void) strcpy(line, cp);
1109 	else {
1110 		(void)inet_ntop(AF_INET6, in, line, sizeof(line));
1111 	}
1112 	return (line);
1113 }
1114 
1115 void
1116 usage()
1117 {
1118 	(void)fprintf(stderr,
1119 "usage: traceroute6 [-dlnrv] [-m max_hops] [-p port#] [-q nqueries]\n\t\
1120 [-s src_addr] [-g gateway] [-w wait] host [data size]\n");
1121 	exit(1);
1122 }
1123