in6_src.c revision 1.5 1 /* $NetBSD: in6_src.c,v 1.5 2001/02/08 14:56:16 itojun 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. All advertising materials mentioning features or use of this software
46 * must display the following acknowledgement:
47 * This product includes software developed by the University of
48 * California, Berkeley and its contributors.
49 * 4. Neither the name of the University nor the names of its contributors
50 * may be used to endorse or promote products derived from this software
51 * without specific prior written permission.
52 *
53 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
54 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
55 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
56 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
57 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
58 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
59 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
60 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
61 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
62 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
63 * SUCH DAMAGE.
64 *
65 * @(#)in_pcb.c 8.2 (Berkeley) 1/4/94
66 */
67
68 #include "opt_inet.h"
69
70 #include <sys/param.h>
71 #include <sys/systm.h>
72 #include <sys/malloc.h>
73 #include <sys/mbuf.h>
74 #include <sys/protosw.h>
75 #include <sys/socket.h>
76 #include <sys/socketvar.h>
77 #include <sys/ioctl.h>
78 #include <sys/errno.h>
79 #include <sys/time.h>
80 #include <sys/proc.h>
81
82 #include <net/if.h>
83 #include <net/route.h>
84
85 #include <netinet/in.h>
86 #include <netinet/in_var.h>
87 #include <netinet/in_systm.h>
88 #include <netinet/ip.h>
89 #include <netinet/in_pcb.h>
90 #include <netinet6/in6_var.h>
91 #include <netinet/ip6.h>
92 #include <netinet6/in6_pcb.h>
93 #include <netinet6/ip6_var.h>
94 #include <netinet6/nd6.h>
95 #ifdef ENABLE_DEFAULT_SCOPE
96 #include <netinet6/scope6_var.h>
97 #endif
98
99 #include <net/net_osdep.h>
100
101 #include "loop.h"
102 extern struct ifnet loif[NLOOP];
103
104 /*
105 * Return an IPv6 address, which is the most appropriate for a given
106 * destination and user specified options.
107 * If necessary, this function lookups the routing table and returns
108 * an entry to the caller for later use.
109 */
110 struct in6_addr *
111 in6_selectsrc(dstsock, opts, mopts, ro, laddr, errorp)
112 struct sockaddr_in6 *dstsock;
113 struct ip6_pktopts *opts;
114 struct ip6_moptions *mopts;
115 struct route_in6 *ro;
116 struct in6_addr *laddr;
117 int *errorp;
118 {
119 struct in6_addr *dst;
120 struct in6_ifaddr *ia6 = 0;
121 struct in6_pktinfo *pi = NULL;
122
123 dst = &dstsock->sin6_addr;
124 *errorp = 0;
125
126 /*
127 * If the source address is explicitly specified by the caller,
128 * use it.
129 */
130 if (opts && (pi = opts->ip6po_pktinfo) &&
131 !IN6_IS_ADDR_UNSPECIFIED(&pi->ipi6_addr))
132 return(&pi->ipi6_addr);
133
134 /*
135 * If the source address is not specified but the socket(if any)
136 * is already bound, use the bound address.
137 */
138 if (laddr && !IN6_IS_ADDR_UNSPECIFIED(laddr))
139 return(laddr);
140
141 /*
142 * If the caller doesn't specify the source address but
143 * the outgoing interface, use an address associated with
144 * the interface.
145 */
146 if (pi && pi->ipi6_ifindex) {
147 /* XXX boundary check is assumed to be already done. */
148 ia6 = in6_ifawithscope(ifindex2ifnet[pi->ipi6_ifindex],
149 dst);
150 if (ia6 == 0) {
151 *errorp = EADDRNOTAVAIL;
152 return(0);
153 }
154 return(&satosin6(&ia6->ia_addr)->sin6_addr);
155 }
156
157 /*
158 * If the destination address is a link-local unicast address or
159 * a multicast address, and if the outgoing interface is specified
160 * by the sin6_scope_id filed, use an address associated with the
161 * interface.
162 * XXX: We're now trying to define more specific semantics of
163 * sin6_scope_id field, so this part will be rewritten in
164 * the near future.
165 */
166 if ((IN6_IS_ADDR_LINKLOCAL(dst) || IN6_IS_ADDR_MULTICAST(dst)) &&
167 dstsock->sin6_scope_id) {
168 /*
169 * I'm not sure if boundary check for scope_id is done
170 * somewhere...
171 */
172 if (dstsock->sin6_scope_id < 0 ||
173 if_index < dstsock->sin6_scope_id) {
174 *errorp = ENXIO; /* XXX: better error? */
175 return(0);
176 }
177 ia6 = in6_ifawithscope(ifindex2ifnet[dstsock->sin6_scope_id],
178 dst);
179 if (ia6 == 0) {
180 *errorp = EADDRNOTAVAIL;
181 return(0);
182 }
183 return(&satosin6(&ia6->ia_addr)->sin6_addr);
184 }
185
186 /*
187 * If the destination address is a multicast address and
188 * the outgoing interface for the address is specified
189 * by the caller, use an address associated with the interface.
190 * There is a sanity check here; if the destination has node-local
191 * scope, the outgoing interfacde should be a loopback address.
192 * Even if the outgoing interface is not specified, we also
193 * choose a loopback interface as the outgoing interface.
194 */
195 if (IN6_IS_ADDR_MULTICAST(dst)) {
196 struct ifnet *ifp = mopts ? mopts->im6o_multicast_ifp : NULL;
197
198 if (ifp == NULL && IN6_IS_ADDR_MC_NODELOCAL(dst)) {
199 ifp = &loif[0];
200 }
201
202 if (ifp) {
203 ia6 = in6_ifawithscope(ifp, dst);
204 if (ia6 == 0) {
205 *errorp = EADDRNOTAVAIL;
206 return(0);
207 }
208 return(&satosin6(&ia6->ia_addr)->sin6_addr);
209 }
210 }
211
212 /*
213 * If the next hop address for the packet is specified
214 * by caller, use an address associated with the route
215 * to the next hop.
216 */
217 {
218 struct sockaddr_in6 *sin6_next;
219 struct rtentry *rt;
220
221 if (opts && opts->ip6po_nexthop) {
222 sin6_next = satosin6(opts->ip6po_nexthop);
223 rt = nd6_lookup(&sin6_next->sin6_addr, 1, NULL);
224 if (rt) {
225 ia6 = in6_ifawithscope(rt->rt_ifp, dst);
226 if (ia6 == 0)
227 ia6 = ifatoia6(rt->rt_ifa);
228 }
229 if (ia6 == 0) {
230 *errorp = EADDRNOTAVAIL;
231 return(0);
232 }
233 return(&satosin6(&ia6->ia_addr)->sin6_addr);
234 }
235 }
236
237 /*
238 * If route is known or can be allocated now,
239 * our src addr is taken from the i/f, else punt.
240 */
241 if (ro) {
242 if (ro->ro_rt &&
243 !IN6_ARE_ADDR_EQUAL(&satosin6(&ro->ro_dst)->sin6_addr, dst)) {
244 RTFREE(ro->ro_rt);
245 ro->ro_rt = (struct rtentry *)0;
246 }
247 if (ro->ro_rt == (struct rtentry *)0 ||
248 ro->ro_rt->rt_ifp == (struct ifnet *)0) {
249 struct sockaddr_in6 *sa6;
250
251 /* No route yet, so try to acquire one */
252 bzero(&ro->ro_dst, sizeof(struct sockaddr_in6));
253 sa6 = (struct sockaddr_in6 *)&ro->ro_dst;
254 sa6->sin6_family = AF_INET6;
255 sa6->sin6_len = sizeof(struct sockaddr_in6);
256 sa6->sin6_addr = *dst;
257 sa6->sin6_scope_id = dstsock->sin6_scope_id;
258 if (IN6_IS_ADDR_MULTICAST(dst)) {
259 ro->ro_rt = rtalloc1(&((struct route *)ro)
260 ->ro_dst, 0);
261 } else {
262 rtalloc((struct route *)ro);
263 }
264 }
265
266 /*
267 * in_pcbconnect() checks out IFF_LOOPBACK to skip using
268 * the address. But we don't know why it does so.
269 * It is necessary to ensure the scope even for lo0
270 * so doesn't check out IFF_LOOPBACK.
271 */
272
273 if (ro->ro_rt) {
274 ia6 = in6_ifawithscope(ro->ro_rt->rt_ifa->ifa_ifp, dst);
275 if (ia6 == 0) /* xxx scope error ?*/
276 ia6 = ifatoia6(ro->ro_rt->rt_ifa);
277 }
278 #if 0
279 /*
280 * xxx The followings are necessary? (kazu)
281 * I don't think so.
282 * It's for SO_DONTROUTE option in IPv4.(jinmei)
283 */
284 if (ia6 == 0) {
285 struct sockaddr_in6 sin6 = {sizeof(sin6), AF_INET6, 0};
286
287 sin6->sin6_addr = *dst;
288
289 ia6 = ifatoia6(ifa_ifwithdstaddr(sin6tosa(&sin6)));
290 if (ia6 == 0)
291 ia6 = ifatoia6(ifa_ifwithnet(sin6tosa(&sin6)));
292 if (ia6 == 0)
293 return(0);
294 return(&satosin6(&ia6->ia_addr)->sin6_addr);
295 }
296 #endif /* 0 */
297 if (ia6 == 0) {
298 *errorp = EHOSTUNREACH; /* no route */
299 return(0);
300 }
301 return(&satosin6(&ia6->ia_addr)->sin6_addr);
302 }
303
304 *errorp = EADDRNOTAVAIL;
305 return(0);
306 }
307
308 /*
309 * Default hop limit selection. The precedence is as follows:
310 * 1. Hoplimit value specified via ioctl.
311 * 2. (If the outgoing interface is detected) the current
312 * hop limit of the interface specified by router advertisement.
313 * 3. The system default hoplimit.
314 */
315 int
316 in6_selecthlim(in6p, ifp)
317 struct in6pcb *in6p;
318 struct ifnet *ifp;
319 {
320 if (in6p && in6p->in6p_hops >= 0)
321 return(in6p->in6p_hops);
322 else if (ifp)
323 return(nd_ifinfo[ifp->if_index].chlim);
324 else
325 return(ip6_defhlim);
326 }
327
328 /*
329 * Find an empty port and set it to the specified PCB.
330 */
331 int
332 in6_pcbsetport(laddr, in6p)
333 struct in6_addr *laddr;
334 struct in6pcb *in6p;
335 {
336 struct socket *so = in6p->in6p_socket;
337 struct in6pcb *head = in6p->in6p_head;
338 u_int16_t last_port, lport = 0;
339 int wild = 0;
340 void *t;
341 u_int16_t min, max;
342 struct proc *p = curproc; /* XXX */
343
344 /* XXX: this is redundant when called from in6_pcbbind */
345 if ((so->so_options & (SO_REUSEADDR|SO_REUSEPORT)) == 0 &&
346 ((so->so_proto->pr_flags & PR_CONNREQUIRED) == 0 ||
347 (so->so_options & SO_ACCEPTCONN) == 0))
348 wild = IN6PLOOKUP_WILDCARD;
349
350 if (in6p->in6p_flags & IN6P_LOWPORT) {
351 #ifndef IPNOPRIVPORTS
352 if (p == 0 || (suser(p->p_ucred, &p->p_acflag) != 0))
353 return (EACCES);
354 #endif
355 min = ip6_lowportmin;
356 max = ip6_lowportmax;
357 } else {
358 min = ip6_anonportmin;
359 max = ip6_anonportmax;
360 }
361
362 /* value out of range */
363 if (head->in6p_lport < min)
364 head->in6p_lport = min;
365 else if (head->in6p_lport > max)
366 head->in6p_lport = min;
367 last_port = head->in6p_lport;
368 goto startover; /*to randomize*/
369 for (;;) {
370 lport = htons(head->in6p_lport);
371 if (IN6_IS_ADDR_V4MAPPED(laddr)) {
372 #if 0
373 t = in_pcblookup_bind(&tcbtable,
374 (struct in_addr *)&in6p->in6p_laddr.s6_addr32[3],
375 lport);
376 #else
377 t = NULL;
378 #endif
379 } else {
380 t = in6_pcblookup(head, &zeroin6_addr, 0, laddr,
381 lport, wild);
382 }
383 if (t == 0)
384 break;
385 startover:
386 if (head->in6p_lport >= max)
387 head->in6p_lport = min;
388 else
389 head->in6p_lport++;
390 if (head->in6p_lport == last_port)
391 return (EADDRINUSE);
392 }
393
394 in6p->in6p_lport = lport;
395 return(0); /* success */
396 }
397
398 /*
399 * generate kernel-internal form (scopeid embedded into s6_addr16[1]).
400 * If the address scope of is link-local, embed the interface index in the
401 * address. The routine determines our precedence
402 * between advanced API scope/interface specification and basic API
403 * specification.
404 *
405 * this function should be nuked in the future, when we get rid of
406 * embedded scopeid thing.
407 *
408 * XXX actually, it is over-specification to return ifp against sin6_scope_id.
409 * there can be multiple interfaces that belong to a particular scope zone
410 * (in specification, we have 1:N mapping between a scope zone and interfaces).
411 * we may want to change the function to return something other than ifp.
412 */
413 int
414 in6_embedscope(in6, sin6, in6p, ifpp)
415 struct in6_addr *in6;
416 const struct sockaddr_in6 *sin6;
417 struct in6pcb *in6p;
418 struct ifnet **ifpp;
419 {
420 struct ifnet *ifp = NULL;
421 u_int32_t scopeid;
422
423 *in6 = sin6->sin6_addr;
424 scopeid = sin6->sin6_scope_id;
425 if (ifpp)
426 *ifpp = NULL;
427
428 /*
429 * don't try to read sin6->sin6_addr beyond here, since the caller may
430 * ask us to overwrite existing sockaddr_in6
431 */
432
433 #ifdef ENABLE_DEFAULT_SCOPE
434 if (scopeid == 0)
435 scopeid = scope6_addr2default(in6);
436 #endif
437
438 if (IN6_IS_SCOPE_LINKLOCAL(in6)) {
439 struct in6_pktinfo *pi;
440
441 /*
442 * KAME assumption: link id == interface id
443 */
444
445 if (in6p && in6p->in6p_outputopts &&
446 (pi = in6p->in6p_outputopts->ip6po_pktinfo) &&
447 pi->ipi6_ifindex) {
448 ifp = ifindex2ifnet[pi->ipi6_ifindex];
449 in6->s6_addr16[1] = htons(pi->ipi6_ifindex);
450 } else if (in6p && IN6_IS_ADDR_MULTICAST(in6) &&
451 in6p->in6p_moptions &&
452 in6p->in6p_moptions->im6o_multicast_ifp) {
453 ifp = in6p->in6p_moptions->im6o_multicast_ifp;
454 in6->s6_addr16[1] = htons(ifp->if_index);
455 } else if (scopeid) {
456 /* boundary check */
457 if (scopeid < 0 || if_index < scopeid)
458 return ENXIO; /* XXX EINVAL? */
459 ifp = ifindex2ifnet[scopeid];
460 /*XXX assignment to 16bit from 32bit variable */
461 in6->s6_addr16[1] = htons(scopeid & 0xffff);
462 }
463
464 if (ifpp)
465 *ifpp = ifp;
466 }
467
468 return 0;
469 }
470
471 /*
472 * generate standard sockaddr_in6 from embedded form.
473 * touches sin6_addr and sin6_scope_id only.
474 *
475 * this function should be nuked in the future, when we get rid of
476 * embedded scopeid thing.
477 */
478 int
479 in6_recoverscope(sin6, in6, ifp)
480 struct sockaddr_in6 *sin6;
481 const struct in6_addr *in6;
482 struct ifnet *ifp;
483 {
484 u_int32_t scopeid;
485
486 sin6->sin6_addr = *in6;
487
488 /*
489 * don't try to read *in6 beyond here, since the caller may
490 * ask us to overwrite existing sockaddr_in6
491 */
492
493 sin6->sin6_scope_id = 0;
494 if (IN6_IS_SCOPE_LINKLOCAL(in6)) {
495 /*
496 * KAME assumption: link id == interface id
497 */
498 scopeid = ntohs(sin6->sin6_addr.s6_addr16[1]);
499 if (scopeid) {
500 /* sanity check */
501 if (scopeid < 0 || if_index < scopeid)
502 return ENXIO;
503 #ifndef FAKE_LOOPBACK_IF
504 if (ifp && (ifp->if_flags & IFF_LOOPBACK) == 0 &&
505 ifp->if_index != scopeid) {
506 return ENXIO;
507 }
508 #else
509 if (ifp && ifp->if_index != scopeid)
510 return ENXIO;
511 #endif
512 sin6->sin6_addr.s6_addr16[1] = 0;
513 sin6->sin6_scope_id = scopeid;
514 }
515 }
516
517 return 0;
518 }
519
520 /*
521 * just clear the embedded scope identifer.
522 * XXX: currently used for bsdi4 only as a supplement function.
523 */
524 void
525 in6_clearscope(addr)
526 struct in6_addr *addr;
527 {
528 if (IN6_IS_SCOPE_LINKLOCAL(addr))
529 addr->s6_addr16[1] = 0;
530 }
531