xref: /freebsd/usr.bin/whois/whois.c (revision f6a3b357e9be4c6423c85eff9a847163a0d307c8)
1 /*-
2  * SPDX-License-Identifier: BSD-3-Clause
3  *
4  * Copyright (c) 1980, 1993
5  *	The Regents of the University of California.  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 University 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 REGENTS 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 REGENTS 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 #ifndef lint
33 static const char copyright[] =
34 "@(#) Copyright (c) 1980, 1993\n\
35 	The Regents of the University of California.  All rights reserved.\n";
36 #endif /* not lint */
37 
38 #if 0
39 #ifndef lint
40 static char sccsid[] = "@(#)whois.c	8.1 (Berkeley) 6/6/93";
41 #endif /* not lint */
42 #endif
43 
44 #include <sys/cdefs.h>
45 __FBSDID("$FreeBSD$");
46 
47 #include <sys/types.h>
48 #include <sys/socket.h>
49 #include <sys/poll.h>
50 #include <netinet/in.h>
51 #include <arpa/inet.h>
52 #include <ctype.h>
53 #include <err.h>
54 #include <netdb.h>
55 #include <stdarg.h>
56 #include <stdio.h>
57 #include <stdlib.h>
58 #include <string.h>
59 #include <sysexits.h>
60 #include <unistd.h>
61 #include <fcntl.h>
62 #include <errno.h>
63 
64 #define	ABUSEHOST	"whois.abuse.net"
65 #define	ANICHOST	"whois.arin.net"
66 #define	DENICHOST	"whois.denic.de"
67 #define	DKNICHOST	"whois.dk-hostmaster.dk"
68 #define	FNICHOST	"whois.afrinic.net"
69 #define	GNICHOST	"whois.nic.gov"
70 #define	IANAHOST	"whois.iana.org"
71 #define	INICHOST	"whois.internic.net"
72 #define	KNICHOST	"whois.krnic.net"
73 #define	LNICHOST	"whois.lacnic.net"
74 #define	MNICHOST	"whois.ra.net"
75 #define	PDBHOST		"whois.peeringdb.com"
76 #define	PNICHOST	"whois.apnic.net"
77 #define	QNICHOST_TAIL	".whois-servers.net"
78 #define	RNICHOST	"whois.ripe.net"
79 #define	VNICHOST	"whois.verisign-grs.com"
80 
81 #define	DEFAULT_PORT	"whois"
82 
83 #define WHOIS_RECURSE	0x01
84 #define WHOIS_QUICK	0x02
85 #define WHOIS_SPAM_ME	0x04
86 
87 #define CHOPSPAM	">>> Last update of WHOIS database:"
88 
89 #define ishost(h) (isalnum((unsigned char)h) || h == '.' || h == '-')
90 
91 #define SCAN(p, end, check)					\
92 	while ((p) < (end))					\
93 		if (check) ++(p);				\
94 		else break
95 
96 static struct {
97 	const char *suffix, *server;
98 } whoiswhere[] = {
99 	/* Various handles */
100 	{ "-ARIN", ANICHOST },
101 	{ "-NICAT", "at" QNICHOST_TAIL },
102 	{ "-NORID", "no" QNICHOST_TAIL },
103 	{ "-RIPE", RNICHOST },
104 	/* Nominet's whois server doesn't return referrals to JANET */
105 	{ ".ac.uk", "ac.uk" QNICHOST_TAIL },
106 	{ ".gov.uk", "ac.uk" QNICHOST_TAIL },
107 	{ "", IANAHOST }, /* default */
108 	{ NULL, NULL } /* safety belt */
109 };
110 
111 #define WHOIS_REFERRAL(s) { s, sizeof(s) - 1 }
112 static struct {
113 	const char *prefix;
114 	size_t len;
115 } whois_referral[] = {
116 	WHOIS_REFERRAL("whois:"), /* IANA */
117 	WHOIS_REFERRAL("Whois Server:"),
118 	WHOIS_REFERRAL("Registrar WHOIS Server:"), /* corporatedomains.com */
119 	WHOIS_REFERRAL("ReferralServer:  whois://"), /* ARIN */
120 	WHOIS_REFERRAL("descr:          region. Please query"), /* AfriNIC */
121 	{ NULL, 0 }
122 };
123 
124 /*
125  * We have a list of patterns for RIRs that assert ignorance rather than
126  * providing referrals. If that happens, we guess that ARIN will be more
127  * helpful. But, before following a referral to an RIR, we check if we have
128  * asked that RIR already, and if so we make another guess.
129  */
130 static const char *actually_arin[] = {
131 	"netname:        ERX-NETBLOCK\n", /* APNIC */
132 	"netname:        NON-RIPE-NCC-MANAGED-ADDRESS-BLOCK\n",
133 	NULL
134 };
135 
136 static struct {
137 	int loop;
138 	const char *host;
139 } try_rir[] = {
140 	{ 0, ANICHOST },
141 	{ 0, RNICHOST },
142 	{ 0, PNICHOST },
143 	{ 0, FNICHOST },
144 	{ 0, LNICHOST },
145 	{ 0, NULL }
146 };
147 
148 static void
149 reset_rir(void) {
150 	int i;
151 
152 	for (i = 0; try_rir[i].host != NULL; i++)
153 		try_rir[i].loop = 0;
154 }
155 
156 static const char *port = DEFAULT_PORT;
157 
158 static const char *choose_server(char *);
159 static struct addrinfo *gethostinfo(char const *host, int exitnoname);
160 static void s_asprintf(char **ret, const char *format, ...) __printflike(2, 3);
161 static void usage(void);
162 static void whois(const char *, const char *, int);
163 
164 int
165 main(int argc, char *argv[])
166 {
167 	const char *country, *host;
168 	int ch, flags;
169 
170 #ifdef	SOCKS
171 	SOCKSinit(argv[0]);
172 #endif
173 
174 	country = host = NULL;
175 	flags = 0;
176 	while ((ch = getopt(argc, argv, "aAbc:fgh:iIklmp:PQrRS")) != -1) {
177 		switch (ch) {
178 		case 'a':
179 			host = ANICHOST;
180 			break;
181 		case 'A':
182 			host = PNICHOST;
183 			break;
184 		case 'b':
185 			host = ABUSEHOST;
186 			break;
187 		case 'c':
188 			country = optarg;
189 			break;
190 		case 'f':
191 			host = FNICHOST;
192 			break;
193 		case 'g':
194 			host = GNICHOST;
195 			break;
196 		case 'h':
197 			host = optarg;
198 			break;
199 		case 'i':
200 			host = INICHOST;
201 			break;
202 		case 'I':
203 			host = IANAHOST;
204 			break;
205 		case 'k':
206 			host = KNICHOST;
207 			break;
208 		case 'l':
209 			host = LNICHOST;
210 			break;
211 		case 'm':
212 			host = MNICHOST;
213 			break;
214 		case 'p':
215 			port = optarg;
216 			break;
217 		case 'P':
218 			host = PDBHOST;
219 			break;
220 		case 'Q':
221 			flags |= WHOIS_QUICK;
222 			break;
223 		case 'r':
224 			host = RNICHOST;
225 			break;
226 		case 'R':
227 			flags |= WHOIS_RECURSE;
228 			break;
229 		case 'S':
230 			flags |= WHOIS_SPAM_ME;
231 			break;
232 		case '?':
233 		default:
234 			usage();
235 			/* NOTREACHED */
236 		}
237 	}
238 	argc -= optind;
239 	argv += optind;
240 
241 	if (!argc || (country != NULL && host != NULL))
242 		usage();
243 
244 	/*
245 	 * If no host or country is specified, rely on referrals from IANA.
246 	 */
247 	if (host == NULL && country == NULL) {
248 		if ((host = getenv("WHOIS_SERVER")) == NULL &&
249 		    (host = getenv("RA_SERVER")) == NULL) {
250 			if (!(flags & WHOIS_QUICK))
251 				flags |= WHOIS_RECURSE;
252 		}
253 	}
254 	while (argc-- > 0) {
255 		if (country != NULL) {
256 			char *qnichost;
257 			s_asprintf(&qnichost, "%s%s", country, QNICHOST_TAIL);
258 			whois(*argv, qnichost, flags);
259 			free(qnichost);
260 		} else
261 			whois(*argv, host != NULL ? host :
262 			      choose_server(*argv), flags);
263 		reset_rir();
264 		argv++;
265 	}
266 	exit(0);
267 }
268 
269 static const char *
270 choose_server(char *domain)
271 {
272 	size_t len = strlen(domain);
273 	int i;
274 
275 	for (i = 0; whoiswhere[i].suffix != NULL; i++) {
276 		size_t suffix_len = strlen(whoiswhere[i].suffix);
277 		if (len > suffix_len &&
278 		    strcasecmp(domain + len - suffix_len,
279 			       whoiswhere[i].suffix) == 0)
280 			return (whoiswhere[i].server);
281 	}
282 	errx(EX_SOFTWARE, "no default whois server");
283 }
284 
285 static struct addrinfo *
286 gethostinfo(char const *host, int exit_on_noname)
287 {
288 	struct addrinfo hints, *res;
289 	int error;
290 
291 	memset(&hints, 0, sizeof(hints));
292 	hints.ai_flags = AI_CANONNAME;
293 	hints.ai_family = AF_UNSPEC;
294 	hints.ai_socktype = SOCK_STREAM;
295 	res = NULL;
296 	error = getaddrinfo(host, port, &hints, &res);
297 	if (error && (exit_on_noname || error != EAI_NONAME))
298 		err(EX_NOHOST, "%s: %s", host, gai_strerror(error));
299 	return (res);
300 }
301 
302 /*
303  * Wrapper for asprintf(3) that exits on error.
304  */
305 static void
306 s_asprintf(char **ret, const char *format, ...)
307 {
308 	va_list ap;
309 
310 	va_start(ap, format);
311 	if (vasprintf(ret, format, ap) == -1) {
312 		va_end(ap);
313 		err(EX_OSERR, "vasprintf()");
314 	}
315 	va_end(ap);
316 }
317 
318 static int
319 connect_to_any_host(struct addrinfo *hostres)
320 {
321 	struct addrinfo *res;
322 	nfds_t i, j;
323 	size_t count;
324 	struct pollfd *fds;
325 	int timeout = 180, s = -1;
326 
327 	for (res = hostres, count = 0; res; res = res->ai_next)
328 		count++;
329 	fds = calloc(count, sizeof(*fds));
330 	if (fds == NULL)
331 		err(EX_OSERR, "calloc()");
332 
333 	/*
334 	 * Traverse the result list elements and make non-block
335 	 * connection attempts.
336 	 */
337 	count = i = 0;
338 	for (res = hostres; res != NULL; res = res->ai_next) {
339 		s = socket(res->ai_family, res->ai_socktype | SOCK_NONBLOCK,
340 		    res->ai_protocol);
341 		if (s < 0)
342 			continue;
343 		if (connect(s, res->ai_addr, res->ai_addrlen) < 0) {
344 			if (errno == EINPROGRESS) {
345 				/* Add the socket to poll list */
346 				fds[i].fd = s;
347 				fds[i].events = POLLERR | POLLHUP |
348 						POLLIN | POLLOUT;
349 				/*
350 				 * From here until a socket connects, the
351 				 * socket fd is owned by the fds[] poll array.
352 				 */
353 				s = -1;
354 				count++;
355 				i++;
356 			} else {
357 				close(s);
358 				s = -1;
359 
360 				/*
361 				 * Poll only if we have something to poll,
362 				 * otherwise just go ahead and try next
363 				 * address
364 				 */
365 				if (count == 0)
366 					continue;
367 			}
368 		} else
369 			goto done;
370 
371 		/*
372 		 * If we are at the last address, poll until a connection is
373 		 * established or we failed all connection attempts.
374 		 */
375 		if (res->ai_next == NULL)
376 			timeout = INFTIM;
377 
378 		/*
379 		 * Poll the watched descriptors for successful connections:
380 		 * if we still have more untried resolved addresses, poll only
381 		 * once; otherwise, poll until all descriptors have errors,
382 		 * which will be considered as ETIMEDOUT later.
383 		 */
384 		do {
385 			int n;
386 
387 			n = poll(fds, i, timeout);
388 			if (n == 0) {
389 				/*
390 				 * No event reported in time.  Try with a
391 				 * smaller timeout (but cap at 2-3ms)
392 				 * after a new host have been added.
393 				 */
394 				if (timeout >= 3)
395 					timeout >>= 1;
396 
397 				break;
398 			} else if (n < 0) {
399 				/*
400 				 * errno here can only be EINTR which we would
401 				 * want to clean up and bail out.
402 				 */
403 				s = -1;
404 				goto done;
405 			}
406 
407 			/*
408 			 * Check for the event(s) we have seen.
409 			 */
410 			for (j = 0; j < i; j++) {
411 				if (fds[j].fd == -1 || fds[j].events == 0 ||
412 				    fds[j].revents == 0)
413 					continue;
414 				if (fds[j].revents & ~(POLLIN | POLLOUT)) {
415 					close(fds[j].fd);
416 					fds[j].fd = -1;
417 					fds[j].events = 0;
418 					count--;
419 					continue;
420 				} else if (fds[j].revents & (POLLIN | POLLOUT)) {
421 					/* Connect succeeded. */
422 					s = fds[j].fd;
423 					fds[j].fd = -1;
424 
425 					goto done;
426 				}
427 
428 			}
429 		} while (timeout == INFTIM && count != 0);
430 	}
431 
432 	/* All attempts were failed */
433 	s = -1;
434 	if (count == 0)
435 		errno = ETIMEDOUT;
436 
437 done:
438 	/* Close all watched fds except the succeeded one */
439 	for (j = 0; j < i; j++)
440 		if (fds[j].fd != -1)
441 			close(fds[j].fd);
442 	free(fds);
443 	return (s);
444 }
445 
446 static void
447 whois(const char *query, const char *hostname, int flags)
448 {
449 	FILE *fp;
450 	struct addrinfo *hostres;
451 	char *buf, *host, *nhost, *p;
452 	int comment, s, f;
453 	size_t len, i;
454 
455 	hostres = gethostinfo(hostname, 1);
456 	s = connect_to_any_host(hostres);
457 	if (s == -1)
458 		err(EX_OSERR, "connect()");
459 
460 	/* Restore default blocking behavior.  */
461 	if ((f = fcntl(s, F_GETFL)) == -1)
462 		err(EX_OSERR, "fcntl()");
463 	f &= ~O_NONBLOCK;
464 	if (fcntl(s, F_SETFL, f) == -1)
465 		err(EX_OSERR, "fcntl()");
466 
467 	fp = fdopen(s, "r+");
468 	if (fp == NULL)
469 		err(EX_OSERR, "fdopen()");
470 
471 	if (!(flags & WHOIS_SPAM_ME) &&
472 	    (strcasecmp(hostname, DENICHOST) == 0 ||
473 	     strcasecmp(hostname, "de" QNICHOST_TAIL) == 0)) {
474 		const char *q;
475 		int idn = 0;
476 		for (q = query; *q != '\0'; q++)
477 			if (!isascii(*q))
478 				idn = 1;
479 		fprintf(fp, "-T dn%s %s\r\n", idn ? "" : ",ace", query);
480 	} else if (!(flags & WHOIS_SPAM_ME) &&
481 		   (strcasecmp(hostname, DKNICHOST) == 0 ||
482 		    strcasecmp(hostname, "dk" QNICHOST_TAIL) == 0))
483 		fprintf(fp, "--show-handles %s\r\n", query);
484 	else if ((flags & WHOIS_SPAM_ME) ||
485 		 strchr(query, ' ') != NULL)
486 		fprintf(fp, "%s\r\n", query);
487 	else if (strcasecmp(hostname, ANICHOST) == 0) {
488 		if (strncasecmp(query, "AS", 2) == 0 &&
489 		    strspn(query+2, "0123456789") == strlen(query+2))
490 			fprintf(fp, "+ a %s\r\n", query+2);
491 		else
492 			fprintf(fp, "+ %s\r\n", query);
493 	} else if (strcasecmp(hostres->ai_canonname, VNICHOST) == 0)
494 		fprintf(fp, "domain %s\r\n", query);
495 	else
496 		fprintf(fp, "%s\r\n", query);
497 	fflush(fp);
498 
499 	comment = 0;
500 	if (!(flags & WHOIS_SPAM_ME) &&
501 	    (strcasecmp(hostname, ANICHOST) == 0 ||
502 	     strcasecmp(hostname, RNICHOST) == 0)) {
503 		comment = 2;
504 	}
505 
506 	nhost = NULL;
507 	while ((buf = fgetln(fp, &len)) != NULL) {
508 		/* Nominet */
509 		if (!(flags & WHOIS_SPAM_ME) &&
510 		    len == 5 && strncmp(buf, "-- \r\n", 5) == 0)
511 			break;
512 		/* RIRs */
513 		if (comment == 1 && buf[0] == '#')
514 			break;
515 		else if (comment == 2) {
516 			if (strchr("#%\r\n", buf[0]) != NULL)
517 				continue;
518 			else
519 				comment = 1;
520 		}
521 
522 		printf("%.*s", (int)len, buf);
523 
524 		if ((flags & WHOIS_RECURSE) && nhost == NULL) {
525 			for (i = 0; whois_referral[i].prefix != NULL; i++) {
526 				p = buf;
527 				SCAN(p, buf+len, *p == ' ');
528 				if (strncasecmp(p, whois_referral[i].prefix,
529 					           whois_referral[i].len) != 0)
530 					continue;
531 				p += whois_referral[i].len;
532 				SCAN(p, buf+len, *p == ' ');
533 				host = p;
534 				SCAN(p, buf+len, ishost(*p));
535 				if (p > host)
536 					s_asprintf(&nhost, "%.*s",
537 						   (int)(p - host), host);
538 				break;
539 			}
540 			for (i = 0; actually_arin[i] != NULL; i++) {
541 				if (strncmp(buf, actually_arin[i], len) == 0) {
542 					s_asprintf(&nhost, "%s", ANICHOST);
543 					break;
544 				}
545 			}
546 		}
547 		/* Verisign etc. */
548 		if (!(flags & WHOIS_SPAM_ME) &&
549 		    len >= sizeof(CHOPSPAM)-1 &&
550 		    (strncasecmp(buf, CHOPSPAM, sizeof(CHOPSPAM)-1) == 0 ||
551 		     strncasecmp(buf, CHOPSPAM+4, sizeof(CHOPSPAM)-5) == 0)) {
552 			printf("\n");
553 			break;
554 		}
555 	}
556 	fclose(fp);
557 	freeaddrinfo(hostres);
558 
559 	f = 0;
560 	for (i = 0; try_rir[i].host != NULL; i++) {
561 		/* Remember visits to RIRs */
562 		if (try_rir[i].loop == 0 &&
563 		    strcasecmp(try_rir[i].host, hostname) == 0)
564 			try_rir[i].loop = 1;
565 		/* Do we need to find an alternative RIR? */
566 		if (try_rir[i].loop != 0 && nhost != NULL &&
567 		    strcasecmp(try_rir[i].host, nhost) == 0) {
568 			    free(nhost);
569 			    nhost = NULL;
570 			    f = 1;
571 		}
572 	}
573 	if (f) {
574 		/* Find a replacement RIR */
575 		for (i = 0; try_rir[i].host != NULL; i++) {
576 			if (try_rir[i].loop == 0) {
577 				s_asprintf(&nhost, "%s",
578 					try_rir[i].host);
579 				break;
580 			}
581 		}
582 	}
583 	if (nhost != NULL) {
584 		/* Ignore self-referrals */
585 		if (strcasecmp(hostname, nhost) != 0) {
586 			printf("# %s\n\n", nhost);
587 			whois(query, nhost, flags);
588 		}
589 		free(nhost);
590 	}
591 }
592 
593 static void
594 usage(void)
595 {
596 	fprintf(stderr,
597 	    "usage: whois [-aAbfgiIklmPQrRS] [-c country-code | -h hostname] "
598 	    "[-p port] name ...\n");
599 	exit(EX_USAGE);
600 }
601