proc->thread stage 4: rework the VFS and DEVICE subsystems to take thread
[dragonfly.git] / sys / netinet6 / in6_src.c
CommitLineData
984263bc 1/* $FreeBSD: src/sys/netinet6/in6_src.c,v 1.1.2.3 2002/02/26 18:02:06 ume Exp $ */
dadab5e9 2/* $DragonFly: src/sys/netinet6/in6_src.c,v 1.4 2003/06/25 03:56:04 dillon Exp $ */
984263bc
MD
3/* $KAME: in6_src.c,v 1.37 2001/03/29 05:34:31 itojun Exp $ */
4
5/*
6 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
7 * All rights reserved.
8 *
9 * Redistribution and use in source and binary forms, with or without
10 * modification, are permitted provided that the following conditions
11 * are met:
12 * 1. Redistributions of source code must retain the above copyright
13 * notice, this list of conditions and the following disclaimer.
14 * 2. Redistributions in binary form must reproduce the above copyright
15 * notice, this list of conditions and the following disclaimer in the
16 * documentation and/or other materials provided with the distribution.
17 * 3. Neither the name of the project nor the names of its contributors
18 * may be used to endorse or promote products derived from this software
19 * without specific prior written permission.
20 *
21 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
31 * SUCH DAMAGE.
32 */
33
34/*
35 * Copyright (c) 1982, 1986, 1991, 1993
36 * The Regents of the University of California. All rights reserved.
37 *
38 * Redistribution and use in source and binary forms, with or without
39 * modification, are permitted provided that the following conditions
40 * are met:
41 * 1. Redistributions of source code must retain the above copyright
42 * notice, this list of conditions and the following disclaimer.
43 * 2. Redistributions in binary form must reproduce the above copyright
44 * notice, this list of conditions and the following disclaimer in the
45 * documentation and/or other materials provided with the distribution.
46 * 3. All advertising materials mentioning features or use of this software
47 * must display the following acknowledgement:
48 * This product includes software developed by the University of
49 * California, Berkeley and its contributors.
50 * 4. Neither the name of the University nor the names of its contributors
51 * may be used to endorse or promote products derived from this software
52 * without specific prior written permission.
53 *
54 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
55 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
56 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
57 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
58 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
59 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
60 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
61 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
62 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
63 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
64 * SUCH DAMAGE.
65 *
66 * @(#)in_pcb.c 8.2 (Berkeley) 1/4/94
67 */
68
69#include "opt_inet.h"
70#include "opt_inet6.h"
71
72#include <sys/param.h>
73#include <sys/systm.h>
74#include <sys/malloc.h>
75#include <sys/mbuf.h>
76#include <sys/protosw.h>
77#include <sys/socket.h>
78#include <sys/socketvar.h>
79#include <sys/errno.h>
80#include <sys/time.h>
81#include <sys/proc.h>
82
83#include <net/if.h>
84#include <net/route.h>
85
86#include <netinet/in.h>
87#include <netinet/in_var.h>
88#include <netinet/in_systm.h>
89#include <netinet/ip.h>
90#include <netinet/in_pcb.h>
91#include <netinet6/in6_var.h>
92#include <netinet/ip6.h>
93#include <netinet6/in6_pcb.h>
94#include <netinet6/ip6_var.h>
95#include <netinet6/nd6.h>
96#ifdef ENABLE_DEFAULT_SCOPE
97#include <netinet6/scope6_var.h>
98#endif
99
100#include <net/net_osdep.h>
101
102#include "loop.h"
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 */
110struct in6_addr *
111in6_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 (!(ro->ro_rt->rt_flags & RTF_UP) ||
244 satosin6(&ro->ro_dst)->sin6_family != AF_INET6 ||
245 !IN6_ARE_ADDR_EQUAL(&satosin6(&ro->ro_dst)->sin6_addr,
246 dst))) {
247 RTFREE(ro->ro_rt);
248 ro->ro_rt = (struct rtentry *)0;
249 }
250 if (ro->ro_rt == (struct rtentry *)0 ||
251 ro->ro_rt->rt_ifp == (struct ifnet *)0) {
252 struct sockaddr_in6 *sa6;
253
254 /* No route yet, so try to acquire one */
255 bzero(&ro->ro_dst, sizeof(struct sockaddr_in6));
256 sa6 = (struct sockaddr_in6 *)&ro->ro_dst;
257 sa6->sin6_family = AF_INET6;
258 sa6->sin6_len = sizeof(struct sockaddr_in6);
259 sa6->sin6_addr = *dst;
260 sa6->sin6_scope_id = dstsock->sin6_scope_id;
261 if (IN6_IS_ADDR_MULTICAST(dst)) {
262 ro->ro_rt = rtalloc1(&((struct route *)ro)
263 ->ro_dst, 0, 0UL);
264 } else {
265 rtalloc((struct route *)ro);
266 }
267 }
268
269 /*
270 * in_pcbconnect() checks out IFF_LOOPBACK to skip using
271 * the address. But we don't know why it does so.
272 * It is necessary to ensure the scope even for lo0
273 * so doesn't check out IFF_LOOPBACK.
274 */
275
276 if (ro->ro_rt) {
277 ia6 = in6_ifawithscope(ro->ro_rt->rt_ifa->ifa_ifp, dst);
278 if (ia6 == 0) /* xxx scope error ?*/
279 ia6 = ifatoia6(ro->ro_rt->rt_ifa);
280 }
281#if 0
282 /*
283 * xxx The followings are necessary? (kazu)
284 * I don't think so.
285 * It's for SO_DONTROUTE option in IPv4.(jinmei)
286 */
287 if (ia6 == 0) {
288 struct sockaddr_in6 sin6 = {sizeof(sin6), AF_INET6, 0};
289
290 sin6->sin6_addr = *dst;
291
292 ia6 = ifatoia6(ifa_ifwithdstaddr(sin6tosa(&sin6)));
293 if (ia6 == 0)
294 ia6 = ifatoia6(ifa_ifwithnet(sin6tosa(&sin6)));
295 if (ia6 == 0)
296 return(0);
297 return(&satosin6(&ia6->ia_addr)->sin6_addr);
298 }
299#endif /* 0 */
300 if (ia6 == 0) {
301 *errorp = EHOSTUNREACH; /* no route */
302 return(0);
303 }
304 return(&satosin6(&ia6->ia_addr)->sin6_addr);
305 }
306
307 *errorp = EADDRNOTAVAIL;
308 return(0);
309}
310
311/*
312 * Default hop limit selection. The precedence is as follows:
313 * 1. Hoplimit value specified via ioctl.
314 * 2. (If the outgoing interface is detected) the current
315 * hop limit of the interface specified by router advertisement.
316 * 3. The system default hoplimit.
317*/
318int
319in6_selecthlim(in6p, ifp)
320 struct in6pcb *in6p;
321 struct ifnet *ifp;
322{
323 if (in6p && in6p->in6p_hops >= 0)
324 return(in6p->in6p_hops);
325 else if (ifp)
326 return(nd_ifinfo[ifp->if_index].chlim);
327 else
328 return(ip6_defhlim);
329}
330
331/*
332 * XXX: this is borrowed from in6_pcbbind(). If possible, we should
333 * share this function by all *bsd*...
334 */
335int
dadab5e9 336in6_pcbsetport(struct in6_addr *laddr, struct inpcb *inp, struct thread *td)
984263bc
MD
337{
338 struct socket *so = inp->inp_socket;
339 u_int16_t lport = 0, first, last, *lastport;
340 int count, error = 0, wild = 0;
341 struct inpcbinfo *pcbinfo = inp->inp_pcbinfo;
342
343 /* XXX: this is redundant when called from in6_pcbbind */
344 if ((so->so_options & (SO_REUSEADDR|SO_REUSEPORT)) == 0)
345 wild = INPLOOKUP_WILDCARD;
346
347 inp->inp_flags |= INP_ANONPORT;
348
349 if (inp->inp_flags & INP_HIGHPORT) {
350 first = ipport_hifirstauto; /* sysctl */
351 last = ipport_hilastauto;
352 lastport = &pcbinfo->lasthi;
353 } else if (inp->inp_flags & INP_LOWPORT) {
dadab5e9 354 if ((error = suser(td)) != 0)
984263bc
MD
355 return error;
356 first = ipport_lowfirstauto; /* 1023 */
357 last = ipport_lowlastauto; /* 600 */
358 lastport = &pcbinfo->lastlow;
359 } else {
360 first = ipport_firstauto; /* sysctl */
361 last = ipport_lastauto;
362 lastport = &pcbinfo->lastport;
363 }
364 /*
365 * Simple check to ensure all ports are not used up causing
366 * a deadlock here.
367 *
368 * We split the two cases (up and down) so that the direction
369 * is not being tested on each round of the loop.
370 */
371 if (first > last) {
372 /*
373 * counting down
374 */
375 count = first - last;
376
377 do {
378 if (count-- < 0) { /* completely used? */
379 /*
380 * Undo any address bind that may have
381 * occurred above.
382 */
383 inp->in6p_laddr = in6addr_any;
384 return (EAGAIN);
385 }
386 --*lastport;
387 if (*lastport > first || *lastport < last)
388 *lastport = first;
389 lport = htons(*lastport);
390 } while (in6_pcblookup_local(pcbinfo,
391 &inp->in6p_laddr, lport, wild));
392 } else {
393 /*
394 * counting up
395 */
396 count = last - first;
397
398 do {
399 if (count-- < 0) { /* completely used? */
400 /*
401 * Undo any address bind that may have
402 * occurred above.
403 */
404 inp->in6p_laddr = in6addr_any;
405 return (EAGAIN);
406 }
407 ++*lastport;
408 if (*lastport < first || *lastport > last)
409 *lastport = first;
410 lport = htons(*lastport);
411 } while (in6_pcblookup_local(pcbinfo,
412 &inp->in6p_laddr, lport, wild));
413 }
414
415 inp->inp_lport = lport;
416 if (in_pcbinshash(inp) != 0) {
417 inp->in6p_laddr = in6addr_any;
418 inp->inp_lport = 0;
419 return (EAGAIN);
420 }
421
422 return(0);
423}
424
425/*
426 * generate kernel-internal form (scopeid embedded into s6_addr16[1]).
427 * If the address scope of is link-local, embed the interface index in the
428 * address. The routine determines our precedence
429 * between advanced API scope/interface specification and basic API
430 * specification.
431 *
432 * this function should be nuked in the future, when we get rid of
433 * embedded scopeid thing.
434 *
435 * XXX actually, it is over-specification to return ifp against sin6_scope_id.
436 * there can be multiple interfaces that belong to a particular scope zone
437 * (in specification, we have 1:N mapping between a scope zone and interfaces).
438 * we may want to change the function to return something other than ifp.
439 */
440int
441in6_embedscope(in6, sin6, in6p, ifpp)
442 struct in6_addr *in6;
443 const struct sockaddr_in6 *sin6;
444#ifdef HAVE_NRL_INPCB
445 struct inpcb *in6p;
446#define in6p_outputopts inp_outputopts6
447#define in6p_moptions inp_moptions6
448#else
449 struct in6pcb *in6p;
450#endif
451 struct ifnet **ifpp;
452{
453 struct ifnet *ifp = NULL;
454 u_int32_t scopeid;
455
456 *in6 = sin6->sin6_addr;
457 scopeid = sin6->sin6_scope_id;
458 if (ifpp)
459 *ifpp = NULL;
460
461 /*
462 * don't try to read sin6->sin6_addr beyond here, since the caller may
463 * ask us to overwrite existing sockaddr_in6
464 */
465
466#ifdef ENABLE_DEFAULT_SCOPE
467 if (scopeid == 0)
468 scopeid = scope6_addr2default(in6);
469#endif
470
471 if (IN6_IS_SCOPE_LINKLOCAL(in6)) {
472 struct in6_pktinfo *pi;
473
474 /*
475 * KAME assumption: link id == interface id
476 */
477
478 if (in6p && in6p->in6p_outputopts &&
479 (pi = in6p->in6p_outputopts->ip6po_pktinfo) &&
480 pi->ipi6_ifindex) {
481 ifp = ifindex2ifnet[pi->ipi6_ifindex];
482 in6->s6_addr16[1] = htons(pi->ipi6_ifindex);
483 } else if (in6p && IN6_IS_ADDR_MULTICAST(in6) &&
484 in6p->in6p_moptions &&
485 in6p->in6p_moptions->im6o_multicast_ifp) {
486 ifp = in6p->in6p_moptions->im6o_multicast_ifp;
487 in6->s6_addr16[1] = htons(ifp->if_index);
488 } else if (scopeid) {
489 /* boundary check */
490 if (scopeid < 0 || if_index < scopeid)
491 return ENXIO; /* XXX EINVAL? */
492 ifp = ifindex2ifnet[scopeid];
493 /*XXX assignment to 16bit from 32bit variable */
494 in6->s6_addr16[1] = htons(scopeid & 0xffff);
495 }
496
497 if (ifpp)
498 *ifpp = ifp;
499 }
500
501 return 0;
502}
503#ifdef HAVE_NRL_INPCB
504#undef in6p_outputopts
505#undef in6p_moptions
506#endif
507
508/*
509 * generate standard sockaddr_in6 from embedded form.
510 * touches sin6_addr and sin6_scope_id only.
511 *
512 * this function should be nuked in the future, when we get rid of
513 * embedded scopeid thing.
514 */
515int
516in6_recoverscope(sin6, in6, ifp)
517 struct sockaddr_in6 *sin6;
518 const struct in6_addr *in6;
519 struct ifnet *ifp;
520{
521 u_int32_t scopeid;
522
523 sin6->sin6_addr = *in6;
524
525 /*
526 * don't try to read *in6 beyond here, since the caller may
527 * ask us to overwrite existing sockaddr_in6
528 */
529
530 sin6->sin6_scope_id = 0;
531 if (IN6_IS_SCOPE_LINKLOCAL(in6)) {
532 /*
533 * KAME assumption: link id == interface id
534 */
535 scopeid = ntohs(sin6->sin6_addr.s6_addr16[1]);
536 if (scopeid) {
537 /* sanity check */
538 if (scopeid < 0 || if_index < scopeid)
539 return ENXIO;
540 if (ifp && ifp->if_index != scopeid)
541 return ENXIO;
542 sin6->sin6_addr.s6_addr16[1] = 0;
543 sin6->sin6_scope_id = scopeid;
544 }
545 }
546
547 return 0;
548}
549
550/*
551 * just clear the embedded scope identifer.
552 * XXX: currently used for bsdi4 only as a supplement function.
553 */
554void
555in6_clearscope(addr)
556 struct in6_addr *addr;
557{
558 if (IN6_IS_SCOPE_LINKLOCAL(addr))
559 addr->s6_addr16[1] = 0;
560}