1116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch/*
2116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch *      Modified for NRL 4.4BSD IPv6 release.
3116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch *      07/31/96 bgp
4116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch *
5116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch *      Search for "#ifdef NRL" to find the changes.
6116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch */
7116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch
8116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch/*
9116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch *	Modified for Linux IPv6 by Pedro Roque <roque@di.fc.ul.pt>
10116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch *	31/07/1996
11116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch *
12116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch *	As ICMP error messages for IPv6 now include more than 8 bytes
13116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch *	UDP datagrams are now sent via an UDP socket instead of magic
14116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch *	RAW socket tricks.
15116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch *
16116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch *	Original copyright and comments left intact. They might not
17116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch *	match the code anymore.
18116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch */
19116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch
20116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch/*-
21116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch * Copyright (c) 1990, 1993
22116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch *	The Regents of the University of California.  All rights reserved.
23116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch *
24116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch * This code is derived from software contributed to Berkeley by
25116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch * Van Jacobson.
26116680a4aac90f2aa7413d9095a592090648e557Ben Murdoch *
27 * Redistribution and use in source and binary forms, with or without
28 * modification, are permitted provided that the following conditions
29 * are met:
30 * 1. Redistributions of source code must retain the above copyright
31 *    notice, this list of conditions and the following disclaimer.
32 * 2. Redistributions in binary form must reproduce the above copyright
33 *    notice, this list of conditions and the following disclaimer in the
34 *    documentation and/or other materials provided with the distribution.
35 * 3. All advertising materials mentioning features or use of this software
36 *    must display the following acknowledgement:
37 *	This product includes software developed by the University of
38 *	California, Berkeley and its contributors.
39 * 4. Neither the name of the University nor the names of its contributors
40 *    may be used to endorse or promote products derived from this software
41 *    without specific prior written permission.
42 *
43 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
44 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
45 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
46 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
47 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
48 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
49 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
50 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
51 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
52 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
53 * SUCH DAMAGE.
54 */
55
56#ifndef lint
57char copyright[] =
58"@(#) Copyright (c) 1990, 1993\n\
59	The Regents of the University of California.  All rights reserved.\n";
60#endif /* not lint */
61
62/*
63 * traceroute host  - trace the route ip packets follow going to "host".
64 *
65 * Attempt to trace the route an ip packet would follow to some
66 * internet host.  We find out intermediate hops by launching probe
67 * packets with a small ttl (time to live) then listening for an
68 * icmp "time exceeded" reply from a gateway.  We start our probes
69 * with a ttl of one and increase by one until we get an icmp "port
70 * unreachable" (which means we got to "host") or hit a max (which
71 * defaults to 30 hops & can be changed with the -m flag).  Three
72 * probes (change with -q flag) are sent at each ttl setting and a
73 * line is printed showing the ttl, address of the gateway and
74 * round trip time of each probe.  If the probe answers come from
75 * different gateways, the address of each responding system will
76 * be printed.  If there is no response within a 5 sec. timeout
77 * interval (changed with the -w flag), a "*" is printed for that
78 * probe.
79 *
80 * Probe packets are UDP format.  We don't want the destination
81 * host to process them so the destination port is set to an
82 * unlikely value (if some clod on the destination is using that
83 * value, it can be changed with the -p flag).
84 *
85 * A sample use might be:
86 *
87 *     [yak 71]% traceroute nis.nsf.net.
88 *     traceroute to nis.nsf.net (35.1.1.48), 30 hops max, 56 byte packet
89 *      1  helios.ee.lbl.gov (128.3.112.1)  19 ms  19 ms  0 ms
90 *      2  lilac-dmc.Berkeley.EDU (128.32.216.1)  39 ms  39 ms  19 ms
91 *      3  lilac-dmc.Berkeley.EDU (128.32.216.1)  39 ms  39 ms  19 ms
92 *      4  ccngw-ner-cc.Berkeley.EDU (128.32.136.23)  39 ms  40 ms  39 ms
93 *      5  ccn-nerif22.Berkeley.EDU (128.32.168.22)  39 ms  39 ms  39 ms
94 *      6  128.32.197.4 (128.32.197.4)  40 ms  59 ms  59 ms
95 *      7  131.119.2.5 (131.119.2.5)  59 ms  59 ms  59 ms
96 *      8  129.140.70.13 (129.140.70.13)  99 ms  99 ms  80 ms
97 *      9  129.140.71.6 (129.140.71.6)  139 ms  239 ms  319 ms
98 *     10  129.140.81.7 (129.140.81.7)  220 ms  199 ms  199 ms
99 *     11  nic.merit.edu (35.1.1.48)  239 ms  239 ms  239 ms
100 *
101 * Note that lines 2 & 3 are the same.  This is due to a buggy
102 * kernel on the 2nd hop system -- lbl-csam.arpa -- that forwards
103 * packets with a zero ttl.
104 *
105 * A more interesting example is:
106 *
107 *     [yak 72]% traceroute allspice.lcs.mit.edu.
108 *     traceroute to allspice.lcs.mit.edu (18.26.0.115), 30 hops max
109 *      1  helios.ee.lbl.gov (128.3.112.1)  0 ms  0 ms  0 ms
110 *      2  lilac-dmc.Berkeley.EDU (128.32.216.1)  19 ms  19 ms  19 ms
111 *      3  lilac-dmc.Berkeley.EDU (128.32.216.1)  39 ms  19 ms  19 ms
112 *      4  ccngw-ner-cc.Berkeley.EDU (128.32.136.23)  19 ms  39 ms  39 ms
113 *      5  ccn-nerif22.Berkeley.EDU (128.32.168.22)  20 ms  39 ms  39 ms
114 *      6  128.32.197.4 (128.32.197.4)  59 ms  119 ms  39 ms
115 *      7  131.119.2.5 (131.119.2.5)  59 ms  59 ms  39 ms
116 *      8  129.140.70.13 (129.140.70.13)  80 ms  79 ms  99 ms
117 *      9  129.140.71.6 (129.140.71.6)  139 ms  139 ms  159 ms
118 *     10  129.140.81.7 (129.140.81.7)  199 ms  180 ms  300 ms
119 *     11  129.140.72.17 (129.140.72.17)  300 ms  239 ms  239 ms
120 *     12  * * *
121 *     13  128.121.54.72 (128.121.54.72)  259 ms  499 ms  279 ms
122 *     14  * * *
123 *     15  * * *
124 *     16  * * *
125 *     17  * * *
126 *     18  ALLSPICE.LCS.MIT.EDU (18.26.0.115)  339 ms  279 ms  279 ms
127 *
128 * (I start to see why I'm having so much trouble with mail to
129 * MIT.)  Note that the gateways 12, 14, 15, 16 & 17 hops away
130 * either don't send ICMP "time exceeded" messages or send them
131 * with a ttl too small to reach us.  14 - 17 are running the
132 * MIT C Gateway code that doesn't send "time exceeded"s.  God
133 * only knows what's going on with 12.
134 *
135 * The silent gateway 12 in the above may be the result of a bug in
136 * the 4.[23]BSD network code (and its derivatives):  4.x (x <= 3)
137 * sends an unreachable message using whatever ttl remains in the
138 * original datagram.  Since, for gateways, the remaining ttl is
139 * zero, the icmp "time exceeded" is guaranteed to not make it back
140 * to us.  The behavior of this bug is slightly more interesting
141 * when it appears on the destination system:
142 *
143 *      1  helios.ee.lbl.gov (128.3.112.1)  0 ms  0 ms  0 ms
144 *      2  lilac-dmc.Berkeley.EDU (128.32.216.1)  39 ms  19 ms  39 ms
145 *      3  lilac-dmc.Berkeley.EDU (128.32.216.1)  19 ms  39 ms  19 ms
146 *      4  ccngw-ner-cc.Berkeley.EDU (128.32.136.23)  39 ms  40 ms  19 ms
147 *      5  ccn-nerif35.Berkeley.EDU (128.32.168.35)  39 ms  39 ms  39 ms
148 *      6  csgw.Berkeley.EDU (128.32.133.254)  39 ms  59 ms  39 ms
149 *      7  * * *
150 *      8  * * *
151 *      9  * * *
152 *     10  * * *
153 *     11  * * *
154 *     12  * * *
155 *     13  rip.Berkeley.EDU (128.32.131.22)  59 ms !  39 ms !  39 ms !
156 *
157 * Notice that there are 12 "gateways" (13 is the final
158 * destination) and exactly the last half of them are "missing".
159 * What's really happening is that rip (a Sun-3 running Sun OS3.5)
160 * is using the ttl from our arriving datagram as the ttl in its
161 * icmp reply.  So, the reply will time out on the return path
162 * (with no notice sent to anyone since icmp's aren't sent for
163 * icmp's) until we probe with a ttl that's at least twice the path
164 * length.  I.e., rip is really only 7 hops away.  A reply that
165 * returns with a ttl of 1 is a clue this problem exists.
166 * Traceroute prints a "!" after the time if the ttl is <= 1.
167 * Since vendors ship a lot of obsolete (DEC's Ultrix, Sun 3.x) or
168 * non-standard (HPUX) software, expect to see this problem
169 * frequently and/or take care picking the target host of your
170 * probes.
171 *
172 * Other possible annotations after the time are !H, !N, !P (got a host,
173 * network or protocol unreachable, respectively), !S or !F (source
174 * route failed or fragmentation needed -- neither of these should
175 * ever occur and the associated gateway is busted if you see one).  If
176 * almost all the probes result in some kind of unreachable, traceroute
177 * will give up and exit.
178 *
179 * Notes
180 * -----
181 * This program must be run by root or be setuid.  (I suggest that
182 * you *don't* make it setuid -- casual use could result in a lot
183 * of unnecessary traffic on our poor, congested nets.)
184 *
185 * This program requires a kernel mod that does not appear in any
186 * system available from Berkeley:  A raw ip socket using proto
187 * IPPROTO_RAW must interpret the data sent as an ip datagram (as
188 * opposed to data to be wrapped in a ip datagram).  See the README
189 * file that came with the source to this program for a description
190 * of the mods I made to /sys/netinet/raw_ip.c.  Your mileage may
191 * vary.  But, again, ANY 4.x (x < 4) BSD KERNEL WILL HAVE TO BE
192 * MODIFIED TO RUN THIS PROGRAM.
193 *
194 * The udp port usage may appear bizarre (well, ok, it is bizarre).
195 * The problem is that an icmp message only contains 8 bytes of
196 * data from the original datagram.  8 bytes is the size of a udp
197 * header so, if we want to associate replies with the original
198 * datagram, the necessary information must be encoded into the
199 * udp header (the ip id could be used but there's no way to
200 * interlock with the kernel's assignment of ip id's and, anyway,
201 * it would have taken a lot more kernel hacking to allow this
202 * code to set the ip id).  So, to allow two or more users to
203 * use traceroute simultaneously, we use this task's pid as the
204 * source port (the high bit is set to move the port number out
205 * of the "likely" range).  To keep track of which probe is being
206 * replied to (so times and/or hop counts don't get confused by a
207 * reply that was delayed in transit), we increment the destination
208 * port number before each probe.
209 *
210 * Don't use this as a coding example.  I was trying to find a
211 * routing problem and this code sort-of popped out after 48 hours
212 * without sleep.  I was amazed it ever compiled, much less ran.
213 *
214 * I stole the idea for this program from Steve Deering.  Since
215 * the first release, I've learned that had I attended the right
216 * IETF working group meetings, I also could have stolen it from Guy
217 * Almes or Matt Mathis.  I don't know (or care) who came up with
218 * the idea first.  I envy the originators' perspicacity and I'm
219 * glad they didn't keep the idea a secret.
220 *
221 * Tim Seaver, Ken Adelman and C. Philip Wood provided bug fixes and/or
222 * enhancements to the original distribution.
223 *
224 * I've hacked up a round-trip-route version of this that works by
225 * sending a loose-source-routed udp datagram through the destination
226 * back to yourself.  Unfortunately, SO many gateways botch source
227 * routing, the thing is almost worthless.  Maybe one day...
228 *
229 *  -- Van Jacobson (van@helios.ee.lbl.gov)
230 *     Tue Dec 20 03:50:13 PST 1988
231 */
232
233#include <sys/param.h>
234#include <sys/time.h>
235#include <sys/socket.h>
236#include <sys/file.h>
237#include <sys/ioctl.h>
238#include <net/if.h>
239
240#if __linux__
241#include <endian.h>
242#endif
243#include <netinet/in_systm.h>
244#include <netinet/in.h>
245#include <netinet/ip.h>
246#include <netinet/ip_icmp.h>
247#include <netinet/udp.h>
248
249#include <netinet/ip6.h>
250#include <netinet/icmp6.h>
251#include <linux/types.h>
252#ifdef CAPABILITIES
253#include <sys/capability.h>
254#endif
255
256#ifdef USE_IDN
257#include <idna.h>
258#include <locale.h>
259#endif
260
261#include <arpa/inet.h>
262
263#include <netdb.h>
264#include <stdio.h>
265#include <errno.h>
266#include <stdlib.h>
267#include <string.h>
268#include <unistd.h>
269
270#include "SNAPSHOT.h"
271
272#ifndef SOL_IPV6
273#define SOL_IPV6 IPPROTO_IPV6
274#endif
275
276#define	MAXPACKET	65535
277#define MAX_HOSTNAMELEN	NI_MAXHOST
278
279#ifndef FD_SET
280#define NFDBITS         (8*sizeof(fd_set))
281#define FD_SETSIZE      NFDBITS
282#define FD_SET(n, p)    ((p)->fds_bits[(n)/NFDBITS] |= (1 << ((n) % NFDBITS)))
283#define FD_CLR(n, p)    ((p)->fds_bits[(n)/NFDBITS] &= ~(1 << ((n) % NFDBITS)))
284#define FD_ISSET(n, p)  ((p)->fds_bits[(n)/NFDBITS] & (1 << ((n) % NFDBITS)))
285#define FD_ZERO(p)      memset((char *)(p), 0, sizeof(*(p)))
286#endif
287
288#define Fprintf (void)fprintf
289#define Printf (void)printf
290
291u_char	packet[512];		/* last inbound (icmp) packet */
292
293int	wait_for_reply(int, struct sockaddr_in6 *, struct in6_addr *, int);
294int	packet_ok(u_char *buf, int cc, struct sockaddr_in6 *from,
295		  struct in6_addr *to, int seq, struct timeval *);
296void	send_probe(int seq, int ttl);
297double	deltaT (struct timeval *, struct timeval *);
298void	print(unsigned char *buf, int cc, struct sockaddr_in6 *from);
299void	tvsub (struct timeval *, struct timeval *);
300void	usage(void);
301
302int icmp_sock;			/* receive (icmp) socket file descriptor */
303int sndsock;			/* send (udp) socket file descriptor */
304struct timezone tz;		/* leftover */
305
306struct sockaddr_in6 whereto;	/* Who to try to reach */
307
308struct sockaddr_in6 saddr;
309struct sockaddr_in6 firsthop;
310char *source = NULL;
311char *device = NULL;
312char *hostname;
313
314int nprobes = 3;
315int max_ttl = 30;
316pid_t ident;
317u_short port = 32768+666;	/* start udp dest port # for probe packets */
318int options;			/* socket options */
319int verbose;
320int waittime = 5;		/* time to wait for response (in seconds) */
321int nflag;			/* print addresses numerically */
322
323
324struct pkt_format
325{
326	__u32 ident;
327	__u32 seq;
328	struct timeval tv;
329};
330
331char *sendbuff;
332int datalen = sizeof(struct pkt_format);
333
334
335
336int main(int argc, char *argv[])
337{
338	char pa[MAX_HOSTNAMELEN];
339	extern char *optarg;
340	extern int optind;
341	struct hostent *hp;
342	struct sockaddr_in6 from, *to;
343	int ch, i, on, probe, seq, tos, ttl;
344	int socket_errno;
345
346	icmp_sock = socket(AF_INET6, SOCK_RAW, IPPROTO_ICMPV6);
347	socket_errno = errno;
348
349	if (setuid(getuid())) {
350		perror("traceroute6: setuid");
351		exit(-1);
352	}
353#ifdef CAPABILITIES
354	{
355		cap_t caps = cap_init();
356		if (cap_set_proc(caps)) {
357			perror("traceroute6: cap_set_proc");
358			exit(-1);
359		}
360		cap_free(caps);
361	}
362#endif
363
364#ifdef USE_IDN
365	setlocale(LC_ALL, "");
366#endif
367
368	on = 1;
369	seq = tos = 0;
370	to = (struct sockaddr_in6 *)&whereto;
371	while ((ch = getopt(argc, argv, "dm:np:q:rs:t:w:vi:g:V")) != EOF) {
372		switch(ch) {
373		case 'd':
374			options |= SO_DEBUG;
375			break;
376		case 'm':
377			max_ttl = atoi(optarg);
378			if (max_ttl <= 1) {
379				Fprintf(stderr,
380				    "traceroute: max ttl must be >1.\n");
381				exit(1);
382			}
383			break;
384		case 'n':
385			nflag++;
386			break;
387		case 'p':
388			port = atoi(optarg);
389			if (port < 1) {
390				Fprintf(stderr,
391				    "traceroute: port must be >0.\n");
392				exit(1);
393			}
394			break;
395		case 'q':
396			nprobes = atoi(optarg);
397			if (nprobes < 1) {
398				Fprintf(stderr,
399				    "traceroute: nprobes must be >0.\n");
400				exit(1);
401			}
402			break;
403		case 'r':
404			options |= SO_DONTROUTE;
405			break;
406		case 's':
407			/*
408			 * set the ip source address of the outbound
409			 * probe (e.g., on a multi-homed host).
410			 */
411			source = optarg;
412			break;
413		case 'i':
414			device = optarg;
415			break;
416		case 'g':
417			Fprintf(stderr, "Sorry, rthdr is not yet supported\n");
418			break;
419		case 'v':
420			verbose++;
421			break;
422		case 'w':
423			waittime = atoi(optarg);
424			if (waittime <= 1) {
425				Fprintf(stderr,
426				    "traceroute: wait must be >1 sec.\n");
427				exit(1);
428			}
429			break;
430		case 'V':
431			printf("traceroute6 utility, iputils-%s\n", SNAPSHOT);
432			exit(0);
433		default:
434			usage();
435		}
436	}
437	argc -= optind;
438	argv += optind;
439
440	if (argc < 1)
441		usage();
442
443	setlinebuf (stdout);
444
445	(void) memset((char *)&whereto, 0, sizeof(whereto));
446
447	to->sin6_family = AF_INET6;
448	to->sin6_port = htons(port);
449
450	if (inet_pton(AF_INET6, *argv, &to->sin6_addr) > 0) {
451		hostname = *argv;
452	} else {
453		char *idn = NULL;
454#ifdef USE_IDN
455		if (idna_to_ascii_lz(*argv, &idn, 0) != IDNA_SUCCESS)
456			idn = NULL;
457#endif
458		hp = gethostbyname2(idn ? idn : *argv, AF_INET6);
459		if (hp) {
460			memmove((caddr_t)&to->sin6_addr, hp->h_addr, sizeof(to->sin6_addr));
461			hostname = (char *)hp->h_name;
462		} else {
463			(void)fprintf(stderr,
464			    "traceroute: unknown host %s\n", *argv);
465			exit(1);
466		}
467	}
468	firsthop = *to;
469	if (*++argv) {
470		datalen = atoi(*argv);
471		/* Message for rpm maintainers: have _shame_. If you want
472		 * to fix something send the patch to me for sanity checking.
473		 * "datalen" patch is a shit. */
474		if (datalen == 0)
475			datalen = sizeof(struct pkt_format);
476		else if (datalen < (int)sizeof(struct pkt_format) ||
477			 datalen >= MAXPACKET) {
478			Fprintf(stderr,
479			    "traceroute: packet size must be %d <= s < %d.\n",
480				(int)sizeof(struct pkt_format), MAXPACKET);
481			exit(1);
482		}
483	}
484
485	ident = getpid();
486
487	sendbuff = malloc(datalen);
488	if (sendbuff == NULL) {
489		fprintf(stderr, "malloc failed\n");
490		exit(1);
491	}
492
493	if (icmp_sock < 0) {
494		errno = socket_errno;
495		perror("traceroute6: icmp socket");
496		exit(1);
497	}
498
499#ifdef IPV6_RECVPKTINFO
500	setsockopt(icmp_sock, SOL_IPV6, IPV6_RECVPKTINFO, &on, sizeof(on));
501	setsockopt(icmp_sock, SOL_IPV6, IPV6_2292PKTINFO, &on, sizeof(on));
502#else
503	setsockopt(icmp_sock, SOL_IPV6, IPV6_PKTINFO, &on, sizeof(on));
504#endif
505
506	if (options & SO_DEBUG)
507		setsockopt(icmp_sock, SOL_SOCKET, SO_DEBUG,
508			   (char *)&on, sizeof(on));
509	if (options & SO_DONTROUTE)
510		setsockopt(icmp_sock, SOL_SOCKET, SO_DONTROUTE,
511			   (char *)&on, sizeof(on));
512
513#ifdef __linux__
514	on = 2;
515	if (setsockopt(icmp_sock, SOL_RAW, IPV6_CHECKSUM, &on, sizeof(on)) < 0) {
516		/* checksum should be enabled by default and setting this
517		 * option might fail anyway.
518		 */
519		fprintf(stderr, "setsockopt(RAW_CHECKSUM) failed - try to continue.");
520	}
521#endif
522
523	if ((sndsock = socket(AF_INET6, SOCK_DGRAM, 0)) < 0) {
524		perror("traceroute: UDP socket");
525		exit(5);
526	}
527#ifdef SO_SNDBUF
528	if (setsockopt(sndsock, SOL_SOCKET, SO_SNDBUF, (char *)&datalen,
529		       sizeof(datalen)) < 0) {
530		perror("traceroute: SO_SNDBUF");
531		exit(6);
532	}
533#endif /* SO_SNDBUF */
534
535	if (options & SO_DEBUG)
536		(void) setsockopt(sndsock, SOL_SOCKET, SO_DEBUG,
537				  (char *)&on, sizeof(on));
538	if (options & SO_DONTROUTE)
539		(void) setsockopt(sndsock, SOL_SOCKET, SO_DONTROUTE,
540				  (char *)&on, sizeof(on));
541
542	if (source == NULL) {
543		socklen_t alen;
544		int probe_fd = socket(AF_INET6, SOCK_DGRAM, 0);
545
546		if (probe_fd < 0) {
547			perror("socket");
548			exit(1);
549		}
550		if (device) {
551			if (setsockopt(probe_fd, SOL_SOCKET, SO_BINDTODEVICE, device, strlen(device)+1) == -1)
552				perror("WARNING: interface is ignored");
553		}
554		firsthop.sin6_port = htons(1025);
555		if (connect(probe_fd, (struct sockaddr*)&firsthop, sizeof(firsthop)) == -1) {
556			perror("connect");
557			exit(1);
558		}
559		alen = sizeof(saddr);
560		if (getsockname(probe_fd, (struct sockaddr*)&saddr, &alen) == -1) {
561			perror("getsockname");
562			exit(1);
563		}
564		saddr.sin6_port = 0;
565		close(probe_fd);
566	} else {
567		(void) memset((char *)&saddr, 0, sizeof(saddr));
568		saddr.sin6_family = AF_INET6;
569		if (inet_pton(AF_INET6, source, &saddr.sin6_addr) <= 0)
570		{
571			Printf("traceroute: unknown addr %s\n", source);
572			exit(1);
573		}
574	}
575
576	if (bind(sndsock, (struct sockaddr *)&saddr, sizeof(saddr)) < 0) {
577		perror ("traceroute: bind sending socket");
578		exit (1);
579	}
580	if (bind(icmp_sock, (struct sockaddr *)&saddr, sizeof(saddr)) < 0) {
581		perror ("traceroute: bind icmp6 socket");
582		exit (1);
583	}
584
585	Fprintf(stderr, "traceroute to %s (%s)", hostname,
586		inet_ntop(AF_INET6, &to->sin6_addr, pa, sizeof(pa)));
587
588	Fprintf(stderr, " from %s",
589		inet_ntop(AF_INET6, &saddr.sin6_addr, pa, sizeof(pa)));
590	Fprintf(stderr, ", %d hops max, %d byte packets\n", max_ttl, datalen);
591	(void) fflush(stderr);
592
593	for (ttl = 1; ttl <= max_ttl; ++ttl) {
594		struct in6_addr lastaddr = {{{0,}}};
595		int got_there = 0;
596		int unreachable = 0;
597
598		Printf("%2d ", ttl);
599		for (probe = 0; probe < nprobes; ++probe) {
600			int cc, reset_timer;
601			struct timeval t1, t2;
602			struct timezone tz;
603			struct in6_addr to;
604
605			gettimeofday(&t1, &tz);
606			send_probe(++seq, ttl);
607			reset_timer = 1;
608
609			while ((cc = wait_for_reply(icmp_sock, &from, &to, reset_timer)) != 0) {
610				gettimeofday(&t2, &tz);
611				if ((i = packet_ok(packet, cc, &from, &to, seq, &t1))) {
612					reset_timer = 1;
613					if (memcmp(&from.sin6_addr, &lastaddr, sizeof(from.sin6_addr))) {
614						print(packet, cc, &from);
615						memcpy(&lastaddr,
616						       &from.sin6_addr,
617						       sizeof(lastaddr));
618					}
619					Printf("  %g ms", deltaT(&t1, &t2));
620					switch(i - 1) {
621					case ICMP6_DST_UNREACH_NOPORT:
622						++got_there;
623						break;
624
625					case ICMP6_DST_UNREACH_NOROUTE:
626						++unreachable;
627						Printf(" !N");
628						break;
629					case ICMP6_DST_UNREACH_ADDR:
630						++unreachable;
631						Printf(" !H");
632						break;
633
634					case ICMP6_DST_UNREACH_ADMIN:
635						++unreachable;
636						Printf(" !S");
637						break;
638					}
639					break;
640				} else
641					reset_timer = 0;
642			}
643			if (cc <= 0)
644				Printf(" *");
645			(void) fflush(stdout);
646		}
647		putchar('\n');
648		if (got_there ||
649		    (unreachable > 0 && unreachable >= nprobes-1))
650			exit(0);
651	}
652
653	return 0;
654}
655
656int
657wait_for_reply(sock, from, to, reset_timer)
658	int sock;
659	struct sockaddr_in6 *from;
660	struct in6_addr *to;
661	int reset_timer;
662{
663	fd_set fds;
664	static struct timeval wait;
665	int cc = 0;
666	char cbuf[512];
667
668	FD_ZERO(&fds);
669	FD_SET(sock, &fds);
670	if (reset_timer) {
671		/*
672		 * traceroute could hang if someone else has a ping
673		 * running and our ICMP reply gets dropped but we don't
674		 * realize it because we keep waking up to handle those
675		 * other ICMP packets that keep coming in.  To fix this,
676		 * "reset_timer" will only be true if the last packet that
677		 * came in was for us or if this is the first time we're
678		 * waiting for a reply since sending out a probe.  Note
679		 * that this takes advantage of the select() feature on
680		 * Linux where the remaining timeout is written to the
681		 * struct timeval area.
682		 */
683		wait.tv_sec = waittime;
684		wait.tv_usec = 0;
685	}
686
687	if (select(sock+1, &fds, (fd_set *)0, (fd_set *)0, &wait) > 0) {
688		struct iovec iov;
689		struct msghdr msg;
690		iov.iov_base = packet;
691		iov.iov_len = sizeof(packet);
692		msg.msg_name = (void *)from;
693		msg.msg_namelen = sizeof(*from);
694		msg.msg_iov = &iov;
695		msg.msg_iovlen = 1;
696		msg.msg_flags = 0;
697		msg.msg_control = cbuf;
698		msg.msg_controllen = sizeof(cbuf);
699
700		cc = recvmsg(icmp_sock, &msg, 0);
701		if (cc >= 0) {
702			struct cmsghdr *cmsg;
703			struct in6_pktinfo *ipi;
704
705			for (cmsg = CMSG_FIRSTHDR(&msg);
706			     cmsg;
707			     cmsg = CMSG_NXTHDR(&msg, cmsg)) {
708				if (cmsg->cmsg_level != SOL_IPV6)
709					continue;
710				switch (cmsg->cmsg_type) {
711				case IPV6_PKTINFO:
712#ifdef IPV6_2292PKTINFO
713				case IPV6_2292PKTINFO:
714#endif
715					ipi = (struct in6_pktinfo *)CMSG_DATA(cmsg);
716					memcpy(to, ipi, sizeof(*to));
717				}
718			}
719		}
720	}
721
722	return(cc);
723}
724
725
726void send_probe(int seq, int ttl)
727{
728	struct pkt_format *pkt = (struct pkt_format *) sendbuff;
729	int i;
730
731	pkt->ident = htonl(ident);
732	pkt->seq = htonl(seq);
733	gettimeofday(&pkt->tv, &tz);
734
735	i = setsockopt(sndsock, SOL_IPV6, IPV6_UNICAST_HOPS, &ttl, sizeof(ttl));
736	if (i < 0)
737	{
738		perror("setsockopt");
739		exit(1);
740	}
741
742	do {
743		i = sendto(sndsock, sendbuff, datalen, 0,
744			   (struct sockaddr *)&whereto, sizeof(whereto));
745	} while (i<0 && errno == ECONNREFUSED);
746
747	if (i < 0 || i != datalen)  {
748		if (i<0)
749			perror("sendto");
750		Printf("traceroute: wrote %s %d chars, ret=%d\n", hostname,
751			datalen, i);
752		(void) fflush(stdout);
753	}
754}
755
756
757double deltaT(struct timeval *t1p, struct timeval *t2p)
758{
759	register double dt;
760
761	dt = (double)(t2p->tv_sec - t1p->tv_sec) * 1000.0 +
762	     (double)(t2p->tv_usec - t1p->tv_usec) / 1000.0;
763	return (dt);
764}
765
766
767/*
768 * Convert an ICMP "type" field to a printable string.
769 */
770char * pr_type(unsigned char t)
771{
772	switch(t) {
773	/* Unknown */
774	case 0:
775		return "Error";
776	case 1:
777		/* ICMP6_DST_UNREACH: */
778		return "Destination Unreachable";
779	case 2:
780		/* ICMP6_PACKET_TOO_BIG: */
781		return "Packet Too Big";
782	case 3:
783		/* ICMP6_TIME_EXCEEDED */
784		return "Time Exceeded in Transit";
785	case 4:
786		/* ICMP6_PARAM_PROB */
787		return "Parameter Problem";
788	case 128:
789		/* ICMP6_ECHO_REQUEST */
790		return "Echo Request";
791	case 129:
792		/* ICMP6_ECHO_REPLY */
793		return "Echo Reply";
794	case 130:
795		/* ICMP6_MEMBERSHIP_QUERY */
796		return "Membership Query";
797	case 131:
798		/* ICMP6_MEMBERSHIP_REPORT */
799		return "Membership Report";
800	case 132:
801		/* ICMP6_MEMBERSHIP_REDUCTION */
802		return "Membership Reduction";
803	case 133:
804		/* ND_ROUTER_SOLICIT */
805		return "Router Solicitation";
806	case 134:
807		/* ND_ROUTER_ADVERT */
808		return "Router Advertisement";
809	case 135:
810		/* ND_NEIGHBOR_SOLICIT */
811		return "Neighbor Solicitation";
812	case 136:
813		/* ND_NEIGHBOR_ADVERT */
814		return "Neighbor Advertisement";
815	case 137:
816		/* ND_REDIRECT */
817		return "Redirect";
818	}
819
820	return("OUT-OF-RANGE");
821}
822
823
824int packet_ok(u_char *buf, int cc, struct sockaddr_in6 *from,
825	      struct in6_addr *to, int seq,
826	      struct timeval *tv)
827{
828	struct icmp6_hdr *icp;
829	u_char type, code;
830
831	icp = (struct icmp6_hdr *) buf;
832
833	type = icp->icmp6_type;
834	code = icp->icmp6_code;
835
836	if ((type == ICMP6_TIME_EXCEEDED && code == ICMP6_TIME_EXCEED_TRANSIT) ||
837	    type == ICMP6_DST_UNREACH)
838	{
839		struct ip6_hdr *hip;
840		struct udphdr *up;
841		int nexthdr;
842
843		hip = (struct ip6_hdr *) (icp + 1);
844		up = (struct udphdr *)(hip+1);
845		nexthdr = hip->ip6_nxt;
846
847		if (nexthdr == 44) {
848			nexthdr = *(unsigned char*)up;
849			up++;
850		}
851		if (nexthdr == IPPROTO_UDP)
852		{
853			struct pkt_format *pkt;
854
855			pkt = (struct pkt_format *) (up + 1);
856
857			if (ntohl(pkt->ident) == ident &&
858			    ntohl(pkt->seq) == seq)
859			{
860				*tv = pkt->tv;
861				return (type == ICMP6_TIME_EXCEEDED ? -1 : code+1);
862			}
863		}
864
865	}
866
867	if (verbose) {
868		unsigned char *p;
869		char pa1[MAX_HOSTNAMELEN];
870		char pa2[MAX_HOSTNAMELEN];
871		int i;
872
873		p = (unsigned char *) (icp + 1);
874
875		Printf("\n%d bytes from %s to %s", cc,
876		       inet_ntop(AF_INET6, &from->sin6_addr, pa1, sizeof(pa1)),
877		       inet_ntop(AF_INET6, to, pa2, sizeof(pa2)));
878
879		Printf(": icmp type %d (%s) code %d\n", type, pr_type(type),
880		       icp->icmp6_code);
881
882		cc -= sizeof(struct icmp6_hdr);
883		for (i = 0; i < cc ; i++) {
884			if (i % 16 == 0)
885				Printf("%04x:", i);
886			if (i % 4 == 0)
887				Printf(" ");
888			Printf("%02x", 0xff & (unsigned)p[i]);
889			if (i % 16 == 15 && i + 1 < cc)
890				Printf("\n");
891		}
892		Printf("\n");
893	}
894
895	return(0);
896}
897
898
899void print(unsigned char *buf, int cc, struct sockaddr_in6 *from)
900{
901	char pa[MAX_HOSTNAMELEN];
902
903	if (nflag)
904		Printf(" %s", inet_ntop(AF_INET6, &from->sin6_addr,
905					pa, sizeof(pa)));
906	else
907	{
908		const char *hostname;
909		struct hostent *hp;
910		char *s = NULL;
911
912		hostname = inet_ntop(AF_INET6, &from->sin6_addr, pa, sizeof(pa));
913
914		if ((hp = gethostbyaddr((char *)&from->sin6_addr,
915					sizeof(from->sin6_addr), AF_INET6))) {
916#ifdef USE_IDN
917			if (idna_to_unicode_lzlz(hp->h_name, &s, 0) != IDNA_SUCCESS)
918				s = NULL;
919#endif
920		}
921
922		Printf(" %s (%s)", hp ? (s ? s : hp->h_name) : hostname, pa);
923
924		free(s);
925	}
926}
927
928
929/*
930 * Subtract 2 timeval structs:  out = out - in.
931 * Out is assumed to be >= in.
932 */
933void
934tvsub(out, in)
935	register struct timeval *out, *in;
936{
937	if ((out->tv_usec -= in->tv_usec) < 0)   {
938		out->tv_sec--;
939		out->tv_usec += 1000000;
940	}
941	out->tv_sec -= in->tv_sec;
942}
943
944void usage(void)
945{
946	fprintf(stderr,
947"Usage: traceroute6 [-dnrvV] [-m max_ttl] [-p port#] [-q nqueries]\n\t\
948[-s src_addr] [-t tos] [-w wait] host [data size]\n");
949	exit(1);
950}
951