in6_src.c revision 1.16 1 /* $NetBSD: in6_src.c,v 1.16 2003/08/07 16:33:26 agc Exp $ */
2 /* $KAME: in6_src.c,v 1.36 2001/02/06 04:08:17 itojun Exp $ */
3
4 /*
5 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
6 * All rights reserved.
7 *
8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions
10 * are met:
11 * 1. Redistributions of source code must retain the above copyright
12 * notice, this list of conditions and the following disclaimer.
13 * 2. Redistributions in binary form must reproduce the above copyright
14 * notice, this list of conditions and the following disclaimer in the
15 * documentation and/or other materials provided with the distribution.
16 * 3. Neither the name of the project nor the names of its contributors
17 * may be used to endorse or promote products derived from this software
18 * without specific prior written permission.
19 *
20 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
21 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
22 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
23 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
24 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
25 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
26 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
28 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
29 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
30 * SUCH DAMAGE.
31 */
32
33 /*
34 * Copyright (c) 1982, 1986, 1991, 1993
35 * The Regents of the University of California. All rights reserved.
36 *
37 * Redistribution and use in source and binary forms, with or without
38 * modification, are permitted provided that the following conditions
39 * are met:
40 * 1. Redistributions of source code must retain the above copyright
41 * notice, this list of conditions and the following disclaimer.
42 * 2. Redistributions in binary form must reproduce the above copyright
43 * notice, this list of conditions and the following disclaimer in the
44 * documentation and/or other materials provided with the distribution.
45 * 3. Neither the name of the University nor the names of its contributors
46 * may be used to endorse or promote products derived from this software
47 * without specific prior written permission.
48 *
49 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
50 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
51 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
52 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
53 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
54 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
55 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
56 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
57 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
58 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
59 * SUCH DAMAGE.
60 *
61 * @(#)in_pcb.c 8.2 (Berkeley) 1/4/94
62 */
63
64 #include <sys/cdefs.h>
65 __KERNEL_RCSID(0, "$NetBSD: in6_src.c,v 1.16 2003/08/07 16:33:26 agc Exp $");
66
67 #include "opt_inet.h"
68
69 #include <sys/param.h>
70 #include <sys/systm.h>
71 #include <sys/malloc.h>
72 #include <sys/mbuf.h>
73 #include <sys/protosw.h>
74 #include <sys/socket.h>
75 #include <sys/socketvar.h>
76 #include <sys/ioctl.h>
77 #include <sys/errno.h>
78 #include <sys/time.h>
79 #include <sys/proc.h>
80
81 #include <net/if.h>
82 #include <net/route.h>
83
84 #include <netinet/in.h>
85 #include <netinet/in_var.h>
86 #include <netinet/in_systm.h>
87 #include <netinet/ip.h>
88 #include <netinet/in_pcb.h>
89 #include <netinet6/in6_var.h>
90 #include <netinet/ip6.h>
91 #include <netinet6/in6_pcb.h>
92 #include <netinet6/ip6_var.h>
93 #include <netinet6/nd6.h>
94 #ifdef ENABLE_DEFAULT_SCOPE
95 #include <netinet6/scope6_var.h>
96 #endif
97
98 #include <net/net_osdep.h>
99
100 #include "loop.h"
101 extern struct ifnet loif[NLOOP];
102
103 /*
104 * Return an IPv6 address, which is the most appropriate for a given
105 * destination and user specified options.
106 * If necessary, this function lookups the routing table and returns
107 * an entry to the caller for later use.
108 */
109 struct in6_addr *
110 in6_selectsrc(dstsock, opts, mopts, ro, laddr, errorp)
111 struct sockaddr_in6 *dstsock;
112 struct ip6_pktopts *opts;
113 struct ip6_moptions *mopts;
114 struct route_in6 *ro;
115 struct in6_addr *laddr;
116 int *errorp;
117 {
118 struct in6_addr *dst;
119 struct in6_ifaddr *ia6 = 0;
120 struct in6_pktinfo *pi = NULL;
121
122 dst = &dstsock->sin6_addr;
123 *errorp = 0;
124
125 /*
126 * If the source address is explicitly specified by the caller,
127 * use it.
128 */
129 if (opts && (pi = opts->ip6po_pktinfo) &&
130 !IN6_IS_ADDR_UNSPECIFIED(&pi->ipi6_addr))
131 return (&pi->ipi6_addr);
132
133 /*
134 * If the source address is not specified but the socket(if any)
135 * is already bound, use the bound address.
136 */
137 if (laddr && !IN6_IS_ADDR_UNSPECIFIED(laddr))
138 return (laddr);
139
140 /*
141 * If the caller doesn't specify the source address but
142 * the outgoing interface, use an address associated with
143 * the interface.
144 */
145 if (pi && pi->ipi6_ifindex) {
146 /* XXX boundary check is assumed to be already done. */
147 ia6 = in6_ifawithscope(ifindex2ifnet[pi->ipi6_ifindex],
148 dst);
149 if (ia6 == 0) {
150 *errorp = EADDRNOTAVAIL;
151 return (0);
152 }
153 return (&satosin6(&ia6->ia_addr)->sin6_addr);
154 }
155
156 /*
157 * If the destination address is a link-local unicast address or
158 * a multicast address, and if the outgoing interface is specified
159 * by the sin6_scope_id filed, use an address associated with the
160 * interface.
161 * XXX: We're now trying to define more specific semantics of
162 * sin6_scope_id field, so this part will be rewritten in
163 * the near future.
164 */
165 if ((IN6_IS_ADDR_LINKLOCAL(dst) || IN6_IS_ADDR_MULTICAST(dst)) &&
166 dstsock->sin6_scope_id) {
167 /*
168 * I'm not sure if boundary check for scope_id is done
169 * somewhere...
170 */
171 if (dstsock->sin6_scope_id < 0 ||
172 if_index < dstsock->sin6_scope_id) {
173 *errorp = ENXIO; /* XXX: better error? */
174 return (0);
175 }
176 ia6 = in6_ifawithscope(ifindex2ifnet[dstsock->sin6_scope_id],
177 dst);
178 if (ia6 == 0) {
179 *errorp = EADDRNOTAVAIL;
180 return (0);
181 }
182 return (&satosin6(&ia6->ia_addr)->sin6_addr);
183 }
184
185 /*
186 * If the destination address is a multicast address and
187 * the outgoing interface for the address is specified
188 * by the caller, use an address associated with the interface.
189 * There is a sanity check here; if the destination has node-local
190 * scope, the outgoing interfacde should be a loopback address.
191 * Even if the outgoing interface is not specified, we also
192 * choose a loopback interface as the outgoing interface.
193 */
194 if (IN6_IS_ADDR_MULTICAST(dst)) {
195 struct ifnet *ifp = mopts ? mopts->im6o_multicast_ifp : NULL;
196
197 if (ifp == NULL && IN6_IS_ADDR_MC_NODELOCAL(dst)) {
198 ifp = &loif[0];
199 }
200
201 if (ifp) {
202 ia6 = in6_ifawithscope(ifp, dst);
203 if (ia6 == 0) {
204 *errorp = EADDRNOTAVAIL;
205 return (0);
206 }
207 return (&satosin6(&ia6->ia_addr)->sin6_addr);
208 }
209 }
210
211 /*
212 * If the next hop address for the packet is specified
213 * by caller, use an address associated with the route
214 * to the next hop.
215 */
216 {
217 struct sockaddr_in6 *sin6_next;
218 struct rtentry *rt;
219
220 if (opts && opts->ip6po_nexthop) {
221 sin6_next = satosin6(opts->ip6po_nexthop);
222 rt = nd6_lookup(&sin6_next->sin6_addr, 1, NULL);
223 if (rt) {
224 ia6 = in6_ifawithscope(rt->rt_ifp, dst);
225 if (ia6 == 0)
226 ia6 = ifatoia6(rt->rt_ifa);
227 }
228 if (ia6 == 0) {
229 *errorp = EADDRNOTAVAIL;
230 return (0);
231 }
232 return (&satosin6(&ia6->ia_addr)->sin6_addr);
233 }
234 }
235
236 /*
237 * If route is known or can be allocated now,
238 * our src addr is taken from the i/f, else punt.
239 * Note that we should check the address family of the
240 * cached destination, in case of sharing the cache with IPv4.
241 */
242 if (ro) {
243 if (ro->ro_rt &&
244 (ro->ro_dst.sin6_family != AF_INET6 ||
245 !IN6_ARE_ADDR_EQUAL(&satosin6(&ro->ro_dst)->sin6_addr, dst))) {
246 RTFREE(ro->ro_rt);
247 ro->ro_rt = (struct rtentry *)0;
248 }
249 if (ro->ro_rt == (struct rtentry *)0 ||
250 ro->ro_rt->rt_ifp == (struct ifnet *)0) {
251 struct sockaddr_in6 *sa6;
252
253 /* No route yet, so try to acquire one */
254 bzero(&ro->ro_dst, sizeof(struct sockaddr_in6));
255 sa6 = (struct sockaddr_in6 *)&ro->ro_dst;
256 sa6->sin6_family = AF_INET6;
257 sa6->sin6_len = sizeof(struct sockaddr_in6);
258 sa6->sin6_addr = *dst;
259 sa6->sin6_scope_id = dstsock->sin6_scope_id;
260 if (IN6_IS_ADDR_MULTICAST(dst)) {
261 ro->ro_rt = rtalloc1(&((struct route *)ro)
262 ->ro_dst, 0);
263 } else {
264 rtalloc((struct route *)ro);
265 }
266 }
267
268 /*
269 * in_pcbconnect() checks out IFF_LOOPBACK to skip using
270 * the address. But we don't know why it does so.
271 * It is necessary to ensure the scope even for lo0
272 * so doesn't check out IFF_LOOPBACK.
273 */
274
275 if (ro->ro_rt) {
276 ia6 = in6_ifawithscope(ro->ro_rt->rt_ifa->ifa_ifp, dst);
277 if (ia6 == 0) /* xxx scope error ?*/
278 ia6 = ifatoia6(ro->ro_rt->rt_ifa);
279 }
280 #if 0
281 /*
282 * xxx The followings are necessary? (kazu)
283 * I don't think so.
284 * It's for SO_DONTROUTE option in IPv4.(jinmei)
285 */
286 if (ia6 == 0) {
287 struct sockaddr_in6 sin6 = {sizeof(sin6), AF_INET6, 0};
288
289 sin6->sin6_addr = *dst;
290
291 ia6 = ifatoia6(ifa_ifwithdstaddr(sin6tosa(&sin6)));
292 if (ia6 == 0)
293 ia6 = ifatoia6(ifa_ifwithnet(sin6tosa(&sin6)));
294 if (ia6 == 0)
295 return (0);
296 return (&satosin6(&ia6->ia_addr)->sin6_addr);
297 }
298 #endif /* 0 */
299 if (ia6 == 0) {
300 *errorp = EHOSTUNREACH; /* no route */
301 return (0);
302 }
303 return (&satosin6(&ia6->ia_addr)->sin6_addr);
304 }
305
306 *errorp = EADDRNOTAVAIL;
307 return (0);
308 }
309
310 /*
311 * Default hop limit selection. The precedence is as follows:
312 * 1. Hoplimit value specified via ioctl.
313 * 2. (If the outgoing interface is detected) the current
314 * hop limit of the interface specified by router advertisement.
315 * 3. The system default hoplimit.
316 */
317 int
318 in6_selecthlim(in6p, ifp)
319 struct in6pcb *in6p;
320 struct ifnet *ifp;
321 {
322 if (in6p && in6p->in6p_hops >= 0)
323 return (in6p->in6p_hops);
324 else if (ifp)
325 return (ND_IFINFO(ifp)->chlim);
326 else
327 return (ip6_defhlim);
328 }
329
330 /*
331 * Find an empty port and set it to the specified PCB.
332 */
333 int
334 in6_pcbsetport(laddr, in6p, p)
335 struct in6_addr *laddr;
336 struct in6pcb *in6p;
337 struct proc *p;
338 {
339 struct socket *so = in6p->in6p_socket;
340 struct in6pcb *head = in6p->in6p_head;
341 u_int16_t last_port, lport = 0;
342 int wild = 0;
343 void *t;
344 u_int16_t min, max;
345
346 /* XXX: this is redundant when called from in6_pcbbind */
347 if ((so->so_options & (SO_REUSEADDR|SO_REUSEPORT)) == 0 &&
348 ((so->so_proto->pr_flags & PR_CONNREQUIRED) == 0 ||
349 (so->so_options & SO_ACCEPTCONN) == 0))
350 wild = IN6PLOOKUP_WILDCARD;
351
352 if (in6p->in6p_flags & IN6P_LOWPORT) {
353 #ifndef IPNOPRIVPORTS
354 if (p == 0 || (suser(p->p_ucred, &p->p_acflag) != 0))
355 return (EACCES);
356 #endif
357 min = ip6_lowportmin;
358 max = ip6_lowportmax;
359 } else {
360 min = ip6_anonportmin;
361 max = ip6_anonportmax;
362 }
363
364 /* value out of range */
365 if (head->in6p_lport < min)
366 head->in6p_lport = min;
367 else if (head->in6p_lport > max)
368 head->in6p_lport = min;
369 last_port = head->in6p_lport;
370 goto startover; /*to randomize*/
371 for (;;) {
372 lport = htons(head->in6p_lport);
373 if (IN6_IS_ADDR_V4MAPPED(laddr)) {
374 #if 0
375 t = in_pcblookup_bind(&tcbtable,
376 (struct in_addr *)&in6p->in6p_laddr.s6_addr32[3],
377 lport);
378 #else
379 t = NULL;
380 #endif
381 } else {
382 t = in6_pcblookup(head, &zeroin6_addr, 0, laddr,
383 lport, wild);
384 }
385 if (t == 0)
386 break;
387 startover:
388 if (head->in6p_lport >= max)
389 head->in6p_lport = min;
390 else
391 head->in6p_lport++;
392 if (head->in6p_lport == last_port)
393 return (EADDRINUSE);
394 }
395
396 in6p->in6p_lport = lport;
397 return (0); /* success */
398 }
399
400 /*
401 * generate kernel-internal form (scopeid embedded into s6_addr16[1]).
402 * If the address scope of is link-local, embed the interface index in the
403 * address. The routine determines our precedence
404 * between advanced API scope/interface specification and basic API
405 * specification.
406 *
407 * this function should be nuked in the future, when we get rid of
408 * embedded scopeid thing.
409 *
410 * XXX actually, it is over-specification to return ifp against sin6_scope_id.
411 * there can be multiple interfaces that belong to a particular scope zone
412 * (in specification, we have 1:N mapping between a scope zone and interfaces).
413 * we may want to change the function to return something other than ifp.
414 */
415 int
416 in6_embedscope(in6, sin6, in6p, ifpp)
417 struct in6_addr *in6;
418 const struct sockaddr_in6 *sin6;
419 struct in6pcb *in6p;
420 struct ifnet **ifpp;
421 {
422 struct ifnet *ifp = NULL;
423 u_int32_t scopeid;
424
425 *in6 = sin6->sin6_addr;
426 scopeid = sin6->sin6_scope_id;
427 if (ifpp)
428 *ifpp = NULL;
429
430 /*
431 * don't try to read sin6->sin6_addr beyond here, since the caller may
432 * ask us to overwrite existing sockaddr_in6
433 */
434
435 #ifdef ENABLE_DEFAULT_SCOPE
436 if (scopeid == 0)
437 scopeid = scope6_addr2default(in6);
438 #endif
439
440 if (IN6_IS_SCOPE_LINKLOCAL(in6)) {
441 struct in6_pktinfo *pi;
442
443 /*
444 * KAME assumption: link id == interface id
445 */
446
447 if (in6p && in6p->in6p_outputopts &&
448 (pi = in6p->in6p_outputopts->ip6po_pktinfo) &&
449 pi->ipi6_ifindex) {
450 ifp = ifindex2ifnet[pi->ipi6_ifindex];
451 in6->s6_addr16[1] = htons(pi->ipi6_ifindex);
452 } else if (in6p && IN6_IS_ADDR_MULTICAST(in6) &&
453 in6p->in6p_moptions &&
454 in6p->in6p_moptions->im6o_multicast_ifp) {
455 ifp = in6p->in6p_moptions->im6o_multicast_ifp;
456 in6->s6_addr16[1] = htons(ifp->if_index);
457 } else if (scopeid) {
458 /* boundary check */
459 if (scopeid < 0 || if_index < scopeid)
460 return ENXIO; /* XXX EINVAL? */
461 ifp = ifindex2ifnet[scopeid];
462 /* XXX assignment to 16bit from 32bit variable */
463 in6->s6_addr16[1] = htons(scopeid & 0xffff);
464 }
465
466 if (ifpp)
467 *ifpp = ifp;
468 }
469
470 return 0;
471 }
472
473 /*
474 * generate standard sockaddr_in6 from embedded form.
475 * touches sin6_addr and sin6_scope_id only.
476 *
477 * this function should be nuked in the future, when we get rid of
478 * embedded scopeid thing.
479 */
480 int
481 in6_recoverscope(sin6, in6, ifp)
482 struct sockaddr_in6 *sin6;
483 const struct in6_addr *in6;
484 struct ifnet *ifp;
485 {
486 u_int32_t scopeid;
487
488 sin6->sin6_addr = *in6;
489
490 /*
491 * don't try to read *in6 beyond here, since the caller may
492 * ask us to overwrite existing sockaddr_in6
493 */
494
495 sin6->sin6_scope_id = 0;
496 if (IN6_IS_SCOPE_LINKLOCAL(in6)) {
497 /*
498 * KAME assumption: link id == interface id
499 */
500 scopeid = ntohs(sin6->sin6_addr.s6_addr16[1]);
501 if (scopeid) {
502 /* sanity check */
503 if (scopeid < 0 || if_index < scopeid)
504 return ENXIO;
505 if (ifp && ifp->if_index != scopeid)
506 return ENXIO;
507 sin6->sin6_addr.s6_addr16[1] = 0;
508 sin6->sin6_scope_id = scopeid;
509 }
510 }
511
512 return 0;
513 }
514
515 /*
516 * just clear the embedded scope identifer.
517 */
518 void
519 in6_clearscope(addr)
520 struct in6_addr *addr;
521 {
522 if (IN6_IS_SCOPE_LINKLOCAL(addr))
523 addr->s6_addr16[1] = 0;
524 }
525