in6_pcb.c revision 1.69.2.2 1 /* $NetBSD: in6_pcb.c,v 1.69.2.2 2006/09/09 02:58:55 rpaulo Exp $ */
2 /* $KAME: in6_pcb.c,v 1.84 2001/02/08 18:02:08 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_pcb.c,v 1.69.2.2 2006/09/09 02:58:55 rpaulo Exp $");
66
67 #include "opt_inet.h"
68 #include "opt_ipsec.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 #include <sys/kauth.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 <netinet/ip6.h>
92 #include <netinet6/ip6_var.h>
93 #include <netinet6/scope6_var.h>
94 #include <netinet6/nd6.h>
95
96 #include "faith.h"
97
98 #ifdef IPSEC
99 #include <netinet6/ipsec.h>
100 #include <netkey/key.h>
101 #endif /* IPSEC */
102
103 #ifdef FAST_IPSEC
104 #include <netipsec/ipsec.h>
105 #include <netipsec/ipsec6.h>
106 #include <netipsec/key.h>
107 #endif /* FAST_IPSEC */
108
109 struct in6_addr zeroin6_addr;
110
111 #define IN6PCBHASH_PORT(table, lport) \
112 &(table)->inpt_porthashtbl[ntohs(lport) & (table)->inpt_porthash]
113 #define IN6PCBHASH_BIND(table, laddr, lport) \
114 &(table)->inpt_bindhashtbl[ \
115 (((laddr)->s6_addr32[0] ^ (laddr)->s6_addr32[1] ^ \
116 (laddr)->s6_addr32[2] ^ (laddr)->s6_addr32[3]) + ntohs(lport)) & \
117 (table)->inpt_bindhash]
118 #define IN6PCBHASH_CONNECT(table, faddr, fport, laddr, lport) \
119 &(table)->inpt_bindhashtbl[ \
120 ((((faddr)->s6_addr32[0] ^ (faddr)->s6_addr32[1] ^ \
121 (faddr)->s6_addr32[2] ^ (faddr)->s6_addr32[3]) + ntohs(fport)) + \
122 (((laddr)->s6_addr32[0] ^ (laddr)->s6_addr32[1] ^ \
123 (laddr)->s6_addr32[2] ^ (laddr)->s6_addr32[3]) + \
124 ntohs(lport))) & (table)->inpt_bindhash]
125
126 int ip6_anonportmin = IPV6PORT_ANONMIN;
127 int ip6_anonportmax = IPV6PORT_ANONMAX;
128 int ip6_lowportmin = IPV6PORT_RESERVEDMIN;
129 int ip6_lowportmax = IPV6PORT_RESERVEDMAX;
130
131 POOL_INIT(in6pcb_pool, sizeof(struct in6pcb), 0, 0, 0, "in6pcbpl", NULL);
132
133 void
134 in6_pcbinit(table, bindhashsize, connecthashsize)
135 struct inpcbtable *table;
136 int bindhashsize, connecthashsize;
137 {
138
139 in_pcbinit(table, bindhashsize, connecthashsize);
140 table->inpt_lastport = (u_int16_t)ip6_anonportmax;
141 }
142
143 int
144 in6_pcballoc(so, v)
145 struct socket *so;
146 void *v;
147 {
148 struct inpcbtable *table = v;
149 struct in6pcb *in6p;
150 int s;
151 #if defined(IPSEC) || defined(FAST_IPSEC)
152 int error;
153 #endif
154
155 in6p = pool_get(&in6pcb_pool, PR_NOWAIT);
156 if (in6p == NULL)
157 return (ENOBUFS);
158 bzero((caddr_t)in6p, sizeof(*in6p));
159 in6p->in6p_af = AF_INET6;
160 in6p->in6p_table = table;
161 in6p->in6p_socket = so;
162 in6p->in6p_hops = -1; /* use kernel default */
163 in6p->in6p_icmp6filt = NULL;
164 #if defined(IPSEC) || defined(FAST_IPSEC)
165 error = ipsec_init_pcbpolicy(so, &in6p->in6p_sp);
166 if (error != 0) {
167 pool_put(&in6pcb_pool, in6p);
168 return error;
169 }
170 #endif /* IPSEC */
171 s = splnet();
172 CIRCLEQ_INSERT_HEAD(&table->inpt_queue, (struct inpcb_hdr*)in6p,
173 inph_queue);
174 LIST_INSERT_HEAD(IN6PCBHASH_PORT(table, in6p->in6p_lport),
175 &in6p->in6p_head, inph_lhash);
176 in6_pcbstate(in6p, IN6P_ATTACHED);
177 splx(s);
178 if (ip6_v6only)
179 in6p->in6p_flags |= IN6P_IPV6_V6ONLY;
180 so->so_pcb = (caddr_t)in6p;
181 return (0);
182 }
183
184 int
185 in6_pcbbind(v, nam, l)
186 void *v;
187 struct mbuf *nam;
188 struct lwp *l;
189 {
190 struct in6pcb *in6p = v;
191 struct socket *so = in6p->in6p_socket;
192 struct inpcbtable *table = in6p->in6p_table;
193 struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)NULL;
194 u_int16_t lport = 0;
195 int wild = 0, reuseport = (so->so_options & SO_REUSEPORT);
196
197 if (in6p->in6p_af != AF_INET6)
198 return (EINVAL);
199
200 if (in6p->in6p_lport || !IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_laddr))
201 return (EINVAL);
202 if ((so->so_options & (SO_REUSEADDR|SO_REUSEPORT)) == 0 &&
203 ((so->so_proto->pr_flags & PR_CONNREQUIRED) == 0 ||
204 (so->so_options & SO_ACCEPTCONN) == 0))
205 wild = 1;
206 if (nam) {
207 int error;
208
209 sin6 = mtod(nam, struct sockaddr_in6 *);
210 if (nam->m_len != sizeof(*sin6))
211 return (EINVAL);
212 /*
213 * We should check the family, but old programs
214 * incorrectly fail to intialize it.
215 */
216 if (sin6->sin6_family != AF_INET6)
217 return (EAFNOSUPPORT);
218
219 #ifndef INET
220 if (IN6_IS_ADDR_V4MAPPED(&sin6->sin6_addr))
221 return (EADDRNOTAVAIL);
222 #endif
223
224 if ((error = sa6_embedscope(sin6, ip6_use_defzone)) != 0)
225 return (error);
226
227 lport = sin6->sin6_port;
228 if (IN6_IS_ADDR_MULTICAST(&sin6->sin6_addr)) {
229 /*
230 * Treat SO_REUSEADDR as SO_REUSEPORT for multicast;
231 * allow compepte duplication of binding if
232 * SO_REUSEPORT is set, or if SO_REUSEADDR is set
233 * and a multicast address is bound on both
234 * new and duplicated sockets.
235 */
236 if (so->so_options & SO_REUSEADDR)
237 reuseport = SO_REUSEADDR|SO_REUSEPORT;
238 } else if (IN6_IS_ADDR_V4MAPPED(&sin6->sin6_addr)) {
239 if ((in6p->in6p_flags & IN6P_IPV6_V6ONLY) != 0)
240 return (EINVAL);
241 if (sin6->sin6_addr.s6_addr32[3]) {
242 struct sockaddr_in sin;
243
244 bzero(&sin, sizeof(sin));
245 sin.sin_len = sizeof(sin);
246 sin.sin_family = AF_INET;
247 bcopy(&sin6->sin6_addr.s6_addr32[3],
248 &sin.sin_addr, sizeof(sin.sin_addr));
249 if (ifa_ifwithaddr((struct sockaddr *)&sin) == 0)
250 return EADDRNOTAVAIL;
251 }
252 } else if (!IN6_IS_ADDR_UNSPECIFIED(&sin6->sin6_addr)) {
253 struct ifaddr *ia = NULL;
254
255 sin6->sin6_port = 0; /* yech... */
256 if ((in6p->in6p_flags & IN6P_FAITH) == 0 &&
257 (ia = ifa_ifwithaddr((struct sockaddr *)sin6)) == 0)
258 return (EADDRNOTAVAIL);
259
260 /*
261 * bind to an anycast address might accidentally
262 * cause sending a packet with an anycast source
263 * address, so we forbid it.
264 *
265 * We should allow to bind to a deprecated address,
266 * since the application dare to use it.
267 * But, can we assume that they are careful enough
268 * to check if the address is deprecated or not?
269 * Maybe, as a safeguard, we should have a setsockopt
270 * flag to control the bind(2) behavior against
271 * deprecated addresses (default: forbid bind(2)).
272 */
273 if (ia &&
274 ((struct in6_ifaddr *)ia)->ia6_flags &
275 (IN6_IFF_ANYCAST|IN6_IFF_NOTREADY|IN6_IFF_DETACHED))
276 return (EADDRNOTAVAIL);
277 }
278 if (lport) {
279 #ifndef IPNOPRIVPORTS
280 int priv;
281
282 /*
283 * NOTE: all operating systems use suser() for
284 * privilege check! do not rewrite it into SS_PRIV.
285 */
286 priv = (l && !kauth_authorize_generic(l->l_cred,
287 KAUTH_GENERIC_ISSUSER, &l->l_acflag)) ? 1 : 0;
288 /* GROSS */
289 if (ntohs(lport) < IPV6PORT_RESERVED && !priv)
290 return (EACCES);
291 #endif
292
293 if (IN6_IS_ADDR_V4MAPPED(&sin6->sin6_addr)) {
294 #ifdef INET
295 struct inpcb *t;
296
297 t = in_pcblookup_port(table,
298 *(struct in_addr *)&sin6->sin6_addr.s6_addr32[3],
299 lport, wild);
300 if (t && (reuseport & t->inp_socket->so_options) == 0)
301 return (EADDRINUSE);
302 #else
303 return (EADDRNOTAVAIL);
304 #endif
305 }
306
307 {
308 struct in6pcb *t;
309
310 t = in6_pcblookup_port(table, &sin6->sin6_addr,
311 lport, wild);
312 if (t && (reuseport & t->in6p_socket->so_options) == 0)
313 return (EADDRINUSE);
314 }
315 }
316 in6p->in6p_laddr = sin6->sin6_addr;
317 }
318
319 if (lport == 0) {
320 int e;
321 e = in6_pcbsetport(&in6p->in6p_laddr, in6p, l);
322 if (e != 0)
323 return (e);
324 } else {
325 in6p->in6p_lport = lport;
326 in6_pcbstate(in6p, IN6P_BOUND);
327 }
328
329 LIST_REMOVE(&in6p->in6p_head, inph_lhash);
330 LIST_INSERT_HEAD(IN6PCBHASH_PORT(table, in6p->in6p_lport),
331 &in6p->in6p_head, inph_lhash);
332
333 #if 0
334 in6p->in6p_flowinfo = 0; /* XXX */
335 #endif
336 return (0);
337 }
338
339 /*
340 * Connect from a socket to a specified address.
341 * Both address and port must be specified in argument sin6.
342 * If don't have a local address for this socket yet,
343 * then pick one.
344 */
345 int
346 in6_pcbconnect(v, nam, l)
347 void *v;
348 struct mbuf *nam;
349 struct lwp *l;
350 {
351 struct in6pcb *in6p = v;
352 struct in6_addr *in6a = NULL;
353 struct sockaddr_in6 *sin6 = mtod(nam, struct sockaddr_in6 *);
354 struct ifnet *ifp = NULL; /* outgoing interface */
355 int error = 0;
356 int scope_ambiguous = 0;
357 #ifdef INET
358 struct in6_addr mapped;
359 #endif
360 struct sockaddr_in6 tmp;
361
362 (void)&in6a; /* XXX fool gcc */
363
364 if (in6p->in6p_af != AF_INET6)
365 return (EINVAL);
366
367 if (nam->m_len != sizeof(*sin6))
368 return (EINVAL);
369 if (sin6->sin6_family != AF_INET6)
370 return (EAFNOSUPPORT);
371 if (sin6->sin6_port == 0)
372 return (EADDRNOTAVAIL);
373
374 if (sin6->sin6_scope_id == 0 && !ip6_use_defzone)
375 scope_ambiguous = 1;
376 if ((error = sa6_embedscope(sin6, ip6_use_defzone)) != 0)
377 return(error);
378
379 /* sanity check for mapped address case */
380 if (IN6_IS_ADDR_V4MAPPED(&sin6->sin6_addr)) {
381 if ((in6p->in6p_flags & IN6P_IPV6_V6ONLY) != 0)
382 return EINVAL;
383 if (IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_laddr))
384 in6p->in6p_laddr.s6_addr16[5] = htons(0xffff);
385 if (!IN6_IS_ADDR_V4MAPPED(&in6p->in6p_laddr))
386 return EINVAL;
387 } else
388 {
389 if (IN6_IS_ADDR_V4MAPPED(&in6p->in6p_laddr))
390 return EINVAL;
391 }
392
393 /* protect *sin6 from overwrites */
394 tmp = *sin6;
395 sin6 = &tmp;
396
397 /* Source address selection. */
398 if (IN6_IS_ADDR_V4MAPPED(&in6p->in6p_laddr) &&
399 in6p->in6p_laddr.s6_addr32[3] == 0) {
400 #ifdef INET
401 struct sockaddr_in sin, *sinp;
402
403 bzero(&sin, sizeof(sin));
404 sin.sin_len = sizeof(sin);
405 sin.sin_family = AF_INET;
406 bcopy(&sin6->sin6_addr.s6_addr32[3], &sin.sin_addr,
407 sizeof(sin.sin_addr));
408 sinp = in_selectsrc(&sin, (struct route *)&in6p->in6p_route,
409 in6p->in6p_socket->so_options, NULL, &error);
410 if (sinp == 0) {
411 if (error == 0)
412 error = EADDRNOTAVAIL;
413 return (error);
414 }
415 bzero(&mapped, sizeof(mapped));
416 mapped.s6_addr16[5] = htons(0xffff);
417 bcopy(&sinp->sin_addr, &mapped.s6_addr32[3], sizeof(sinp->sin_addr));
418 in6a = &mapped;
419 #else
420 return EADDRNOTAVAIL;
421 #endif
422 } else {
423 /*
424 * XXX: in6_selectsrc might replace the bound local address
425 * with the address specified by setsockopt(IPV6_PKTINFO).
426 * Is it the intended behavior?
427 */
428 in6a = in6_selectsrc(sin6, in6p->in6p_outputopts,
429 in6p->in6p_moptions,
430 &in6p->in6p_route,
431 &in6p->in6p_laddr, &ifp, &error);
432 if (ifp && scope_ambiguous &&
433 (error = in6_setscope(&sin6->sin6_addr, ifp, NULL)) != 0) {
434 return(error);
435 }
436
437 if (in6a == 0) {
438 if (error == 0)
439 error = EADDRNOTAVAIL;
440 return (error);
441 }
442 }
443 if (ifp == NULL && in6p->in6p_route.ro_rt)
444 ifp = in6p->in6p_route.ro_rt->rt_ifp;
445
446 in6p->in6p_ip6.ip6_hlim = (u_int8_t)in6_selecthlim(in6p, ifp);
447
448 if (in6_pcblookup_connect(in6p->in6p_table, &sin6->sin6_addr,
449 sin6->sin6_port,
450 IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_laddr) ? in6a : &in6p->in6p_laddr,
451 in6p->in6p_lport, 0))
452 return (EADDRINUSE);
453 if (IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_laddr) ||
454 (IN6_IS_ADDR_V4MAPPED(&in6p->in6p_laddr) &&
455 in6p->in6p_laddr.s6_addr32[3] == 0))
456 {
457 if (in6p->in6p_lport == 0) {
458 error = in6_pcbbind(in6p, (struct mbuf *)0, l);
459 if (error != 0)
460 return error;
461 }
462 in6p->in6p_laddr = *in6a;
463 }
464 in6p->in6p_faddr = sin6->sin6_addr;
465 in6p->in6p_fport = sin6->sin6_port;
466 in6_pcbstate(in6p, IN6P_CONNECTED);
467 in6p->in6p_flowinfo &= ~IPV6_FLOWLABEL_MASK;
468 if (ip6_auto_flowlabel)
469 in6p->in6p_flowinfo |=
470 (htonl(ip6_randomflowlabel()) & IPV6_FLOWLABEL_MASK);
471 #if defined(IPSEC) || defined(FAST_IPSEC)
472 if (in6p->in6p_socket->so_type == SOCK_STREAM)
473 ipsec_pcbconn(in6p->in6p_sp);
474 #endif
475 return (0);
476 }
477
478 void
479 in6_pcbdisconnect(in6p)
480 struct in6pcb *in6p;
481 {
482 bzero((caddr_t)&in6p->in6p_faddr, sizeof(in6p->in6p_faddr));
483 in6p->in6p_fport = 0;
484 in6_pcbstate(in6p, IN6P_BOUND);
485 in6p->in6p_flowinfo &= ~IPV6_FLOWLABEL_MASK;
486 #if defined(IPSEC) || defined(FAST_IPSEC)
487 ipsec_pcbdisconn(in6p->in6p_sp);
488 #endif
489 if (in6p->in6p_socket->so_state & SS_NOFDREF)
490 in6_pcbdetach(in6p);
491 }
492
493 void
494 in6_pcbdetach(in6p)
495 struct in6pcb *in6p;
496 {
497 struct socket *so = in6p->in6p_socket;
498 int s;
499
500 if (in6p->in6p_af != AF_INET6)
501 return;
502
503 #if defined(IPSEC) || defined(FAST_IPSEC)
504 ipsec6_delete_pcbpolicy(in6p);
505 #endif /* IPSEC */
506 so->so_pcb = 0;
507 sofree(so);
508 if (in6p->in6p_options)
509 m_freem(in6p->in6p_options);
510 if (in6p->in6p_outputopts) {
511 if (in6p->in6p_outputopts->ip6po_rthdr &&
512 in6p->in6p_outputopts->ip6po_route.ro_rt)
513 RTFREE(in6p->in6p_outputopts->ip6po_route.ro_rt);
514 if (in6p->in6p_outputopts->ip6po_m)
515 (void)m_free(in6p->in6p_outputopts->ip6po_m);
516 free(in6p->in6p_outputopts, M_IP6OPT);
517 }
518 if (in6p->in6p_route.ro_rt)
519 rtfree(in6p->in6p_route.ro_rt);
520 ip6_freemoptions(in6p->in6p_moptions);
521 s = splnet();
522 in6_pcbstate(in6p, IN6P_ATTACHED);
523 LIST_REMOVE(&in6p->in6p_head, inph_lhash);
524 CIRCLEQ_REMOVE(&in6p->in6p_table->inpt_queue, &in6p->in6p_head,
525 inph_queue);
526 splx(s);
527 pool_put(&in6pcb_pool, in6p);
528 }
529
530 void
531 in6_setsockaddr(in6p, nam)
532 struct in6pcb *in6p;
533 struct mbuf *nam;
534 {
535 struct sockaddr_in6 *sin6;
536
537 if (in6p->in6p_af != AF_INET6)
538 return;
539
540 nam->m_len = sizeof(*sin6);
541 sin6 = mtod(nam, struct sockaddr_in6 *);
542 bzero((caddr_t)sin6, sizeof(*sin6));
543 sin6->sin6_family = AF_INET6;
544 sin6->sin6_len = sizeof(struct sockaddr_in6);
545 sin6->sin6_port = in6p->in6p_lport;
546 sin6->sin6_addr = in6p->in6p_laddr;
547 (void)sa6_recoverscope(sin6); /* XXX: should catch errors */
548 }
549
550 void
551 in6_setpeeraddr(in6p, nam)
552 struct in6pcb *in6p;
553 struct mbuf *nam;
554 {
555 struct sockaddr_in6 *sin6;
556
557 if (in6p->in6p_af != AF_INET6)
558 return;
559
560 nam->m_len = sizeof(*sin6);
561 sin6 = mtod(nam, struct sockaddr_in6 *);
562 bzero((caddr_t)sin6, sizeof(*sin6));
563 sin6->sin6_family = AF_INET6;
564 sin6->sin6_len = sizeof(struct sockaddr_in6);
565 sin6->sin6_port = in6p->in6p_fport;
566 sin6->sin6_addr = in6p->in6p_faddr;
567 (void)sa6_recoverscope(sin6); /* XXX: should catch errors */
568 }
569
570 /*
571 * Pass some notification to all connections of a protocol
572 * associated with address dst. The local address and/or port numbers
573 * may be specified to limit the search. The "usual action" will be
574 * taken, depending on the ctlinput cmd. The caller must filter any
575 * cmds that are uninteresting (e.g., no error in the map).
576 * Call the protocol specific routine (if any) to report
577 * any errors for each matching socket.
578 *
579 * Must be called at splsoftnet.
580 *
581 * Note: src (4th arg) carries the flowlabel value on the original IPv6
582 * header, in sin6_flowinfo member.
583 */
584 int
585 in6_pcbnotify(table, dst, fport_arg, src, lport_arg, cmd, cmdarg, notify)
586 struct inpcbtable *table;
587 struct sockaddr *dst;
588 const struct sockaddr *src;
589 u_int fport_arg, lport_arg;
590 int cmd;
591 void *cmdarg;
592 void (*notify) __P((struct in6pcb *, int));
593 {
594 struct in6pcb *in6p, *nin6p;
595 struct sockaddr_in6 sa6_src, *sa6_dst;
596 u_int16_t fport = fport_arg, lport = lport_arg;
597 int errno;
598 int nmatch = 0;
599 u_int32_t flowinfo;
600
601 if ((unsigned)cmd >= PRC_NCMDS || dst->sa_family != AF_INET6)
602 return 0;
603
604 sa6_dst = (struct sockaddr_in6 *)dst;
605 if (IN6_IS_ADDR_UNSPECIFIED(&sa6_dst->sin6_addr))
606 return 0;
607
608 /*
609 * note that src can be NULL when we get notify by local fragmentation.
610 */
611 sa6_src = (src == NULL) ? sa6_any : *(const struct sockaddr_in6 *)src;
612 flowinfo = sa6_src.sin6_flowinfo;
613
614 /*
615 * Redirects go to all references to the destination,
616 * and use in6_rtchange to invalidate the route cache.
617 * Dead host indications: also use in6_rtchange to invalidate
618 * the cache, and deliver the error to all the sockets.
619 * Otherwise, if we have knowledge of the local port and address,
620 * deliver only to that socket.
621 */
622 if (PRC_IS_REDIRECT(cmd) || cmd == PRC_HOSTDEAD) {
623 fport = 0;
624 lport = 0;
625 bzero((caddr_t)&sa6_src.sin6_addr, sizeof(sa6_src.sin6_addr));
626
627 if (cmd != PRC_HOSTDEAD)
628 notify = in6_rtchange;
629 }
630
631 errno = inet6ctlerrmap[cmd];
632 for (in6p = (struct in6pcb *)CIRCLEQ_FIRST(&table->inpt_queue);
633 in6p != (void *)&table->inpt_queue;
634 in6p = nin6p) {
635 nin6p = (struct in6pcb *)CIRCLEQ_NEXT(in6p, in6p_queue);
636
637 if (in6p->in6p_af != AF_INET6)
638 continue;
639
640 /*
641 * Under the following condition, notify of redirects
642 * to the pcb, without making address matches against inpcb.
643 * - redirect notification is arrived.
644 * - the inpcb is unconnected.
645 * - the inpcb is caching !RTF_HOST routing entry.
646 * - the ICMPv6 notification is from the gateway cached in the
647 * inpcb. i.e. ICMPv6 notification is from nexthop gateway
648 * the inpcb used very recently.
649 *
650 * This is to improve interaction between netbsd/openbsd
651 * redirect handling code, and inpcb route cache code.
652 * without the clause, !RTF_HOST routing entry (which carries
653 * gateway used by inpcb right before the ICMPv6 redirect)
654 * will be cached forever in unconnected inpcb.
655 *
656 * There still is a question regarding to what is TRT:
657 * - On bsdi/freebsd, RTF_HOST (cloned) routing entry will be
658 * generated on packet output. inpcb will always cache
659 * RTF_HOST routing entry so there's no need for the clause
660 * (ICMPv6 redirect will update RTF_HOST routing entry,
661 * and inpcb is caching it already).
662 * However, bsdi/freebsd are vulnerable to local DoS attacks
663 * due to the cloned routing entries.
664 * - Specwise, "destination cache" is mentioned in RFC2461.
665 * Jinmei says that it implies bsdi/freebsd behavior, itojun
666 * is not really convinced.
667 * - Having hiwat/lowat on # of cloned host route (redirect/
668 * pmtud) may be a good idea. netbsd/openbsd has it. see
669 * icmp6_mtudisc_update().
670 */
671 if ((PRC_IS_REDIRECT(cmd) || cmd == PRC_HOSTDEAD) &&
672 IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_laddr) &&
673 in6p->in6p_route.ro_rt &&
674 !(in6p->in6p_route.ro_rt->rt_flags & RTF_HOST)) {
675 struct sockaddr_in6 *dst6;
676
677 dst6 = (struct sockaddr_in6 *)&in6p->in6p_route.ro_dst;
678 if (IN6_ARE_ADDR_EQUAL(&dst6->sin6_addr,
679 &sa6_dst->sin6_addr))
680 goto do_notify;
681 }
682
683 /*
684 * If the error designates a new path MTU for a destination
685 * and the application (associated with this socket) wanted to
686 * know the value, notify. Note that we notify for all
687 * disconnected sockets if the corresponding application
688 * wanted. This is because some UDP applications keep sending
689 * sockets disconnected.
690 * XXX: should we avoid to notify the value to TCP sockets?
691 */
692 if (cmd == PRC_MSGSIZE && (in6p->in6p_flags & IN6P_MTU) != 0 &&
693 (IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_faddr) ||
694 IN6_ARE_ADDR_EQUAL(&in6p->in6p_faddr, &sa6_dst->sin6_addr))) {
695 ip6_notify_pmtu(in6p, (struct sockaddr_in6 *)dst,
696 (u_int32_t *)cmdarg);
697 }
698
699 /*
700 * Detect if we should notify the error. If no source and
701 * destination ports are specified, but non-zero flowinfo and
702 * local address match, notify the error. This is the case
703 * when the error is delivered with an encrypted buffer
704 * by ESP. Otherwise, just compare addresses and ports
705 * as usual.
706 */
707 if (lport == 0 && fport == 0 && flowinfo &&
708 in6p->in6p_socket != NULL &&
709 flowinfo == (in6p->in6p_flowinfo & IPV6_FLOWLABEL_MASK) &&
710 IN6_ARE_ADDR_EQUAL(&in6p->in6p_laddr, &sa6_src.sin6_addr))
711 goto do_notify;
712 else if (!IN6_ARE_ADDR_EQUAL(&in6p->in6p_faddr,
713 &sa6_dst->sin6_addr) ||
714 in6p->in6p_socket == 0 ||
715 (lport && in6p->in6p_lport != lport) ||
716 (!IN6_IS_ADDR_UNSPECIFIED(&sa6_src.sin6_addr) &&
717 !IN6_ARE_ADDR_EQUAL(&in6p->in6p_laddr,
718 &sa6_src.sin6_addr)) ||
719 (fport && in6p->in6p_fport != fport))
720 continue;
721
722 do_notify:
723 if (notify)
724 (*notify)(in6p, errno);
725 nmatch++;
726 }
727 return nmatch;
728 }
729
730 void
731 in6_pcbpurgeif0(table, ifp)
732 struct inpcbtable *table;
733 struct ifnet *ifp;
734 {
735 struct in6pcb *in6p, *nin6p;
736 struct ip6_moptions *im6o;
737 struct in6_multi_mship *imm, *nimm;
738
739 for (in6p = (struct in6pcb *)CIRCLEQ_FIRST(&table->inpt_queue);
740 in6p != (void *)&table->inpt_queue;
741 in6p = nin6p) {
742 nin6p = (struct in6pcb *)CIRCLEQ_NEXT(in6p, in6p_queue);
743 if (in6p->in6p_af != AF_INET6)
744 continue;
745
746 im6o = in6p->in6p_moptions;
747 if (im6o) {
748 /*
749 * Unselect the outgoing interface if it is being
750 * detached.
751 */
752 if (im6o->im6o_multicast_ifp == ifp)
753 im6o->im6o_multicast_ifp = NULL;
754
755 /*
756 * Drop multicast group membership if we joined
757 * through the interface being detached.
758 * XXX controversial - is it really legal for kernel
759 * to force this?
760 */
761 for (imm = im6o->im6o_memberships.lh_first;
762 imm != NULL; imm = nimm) {
763 nimm = imm->i6mm_chain.le_next;
764 if (imm->i6mm_maddr->in6m_ifp == ifp) {
765 LIST_REMOVE(imm, i6mm_chain);
766 in6_leavegroup(imm);
767 }
768 }
769 }
770 }
771 }
772
773 void
774 in6_pcbpurgeif(table, ifp)
775 struct inpcbtable *table;
776 struct ifnet *ifp;
777 {
778 struct in6pcb *in6p, *nin6p;
779
780 for (in6p = (struct in6pcb *)CIRCLEQ_FIRST(&table->inpt_queue);
781 in6p != (void *)&table->inpt_queue;
782 in6p = nin6p) {
783 nin6p = (struct in6pcb *)CIRCLEQ_NEXT(in6p, in6p_queue);
784 if (in6p->in6p_af != AF_INET6)
785 continue;
786 if (in6p->in6p_route.ro_rt != NULL &&
787 in6p->in6p_route.ro_rt->rt_ifp == ifp)
788 in6_rtchange(in6p, 0);
789 }
790 }
791
792 /*
793 * Check for alternatives when higher level complains
794 * about service problems. For now, invalidate cached
795 * routing information. If the route was created dynamically
796 * (by a redirect), time to try a default gateway again.
797 */
798 void
799 in6_losing(in6p)
800 struct in6pcb *in6p;
801 {
802 struct rtentry *rt;
803 struct rt_addrinfo info;
804
805 if (in6p->in6p_af != AF_INET6)
806 return;
807
808 if ((rt = in6p->in6p_route.ro_rt) != NULL) {
809 in6p->in6p_route.ro_rt = 0;
810 bzero((caddr_t)&info, sizeof(info));
811 info.rti_info[RTAX_DST] =
812 (struct sockaddr *)&in6p->in6p_route.ro_dst;
813 info.rti_info[RTAX_GATEWAY] = rt->rt_gateway;
814 info.rti_info[RTAX_NETMASK] = rt_mask(rt);
815 rt_missmsg(RTM_LOSING, &info, rt->rt_flags, 0);
816 if (rt->rt_flags & RTF_DYNAMIC) {
817 (void)rtrequest(RTM_DELETE, rt_key(rt),
818 rt->rt_gateway, rt_mask(rt), rt->rt_flags,
819 (struct rtentry **)0);
820 } else {
821 /*
822 * A new route can be allocated
823 * the next time output is attempted.
824 */
825 rtfree(rt);
826 }
827 }
828 }
829
830 /*
831 * After a routing change, flush old routing
832 * and allocate a (hopefully) better one.
833 */
834 void
835 in6_rtchange(in6p, errno)
836 struct in6pcb *in6p;
837 int errno;
838 {
839 if (in6p->in6p_af != AF_INET6)
840 return;
841
842 if (in6p->in6p_route.ro_rt) {
843 rtfree(in6p->in6p_route.ro_rt);
844 in6p->in6p_route.ro_rt = 0;
845 /*
846 * A new route can be allocated the next time
847 * output is attempted.
848 */
849 }
850 }
851
852 struct in6pcb *
853 in6_pcblookup_port(table, laddr6, lport_arg, lookup_wildcard)
854 struct inpcbtable *table;
855 struct in6_addr *laddr6;
856 u_int lport_arg;
857 int lookup_wildcard;
858 {
859 struct inpcbhead *head;
860 struct inpcb_hdr *inph;
861 struct in6pcb *in6p, *match = 0;
862 int matchwild = 3, wildcard;
863 u_int16_t lport = lport_arg;
864
865 head = IN6PCBHASH_PORT(table, lport);
866 LIST_FOREACH(inph, head, inph_lhash) {
867 in6p = (struct in6pcb *)inph;
868 if (in6p->in6p_af != AF_INET6)
869 continue;
870
871 if (in6p->in6p_lport != lport)
872 continue;
873 wildcard = 0;
874 if (IN6_IS_ADDR_V4MAPPED(&in6p->in6p_faddr)) {
875 if ((in6p->in6p_flags & IN6P_IPV6_V6ONLY) != 0)
876 continue;
877 }
878 if (!IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_faddr))
879 wildcard++;
880 if (IN6_IS_ADDR_V4MAPPED(&in6p->in6p_laddr)) {
881 if ((in6p->in6p_flags & IN6P_IPV6_V6ONLY) != 0)
882 continue;
883 if (!IN6_IS_ADDR_V4MAPPED(laddr6))
884 continue;
885
886 /* duplicate of IPv4 logic */
887 wildcard = 0;
888 if (IN6_IS_ADDR_V4MAPPED(&in6p->in6p_faddr) &&
889 in6p->in6p_faddr.s6_addr32[3])
890 wildcard++;
891 if (!in6p->in6p_laddr.s6_addr32[3]) {
892 if (laddr6->s6_addr32[3])
893 wildcard++;
894 } else {
895 if (!laddr6->s6_addr32[3])
896 wildcard++;
897 else {
898 if (in6p->in6p_laddr.s6_addr32[3] !=
899 laddr6->s6_addr32[3])
900 continue;
901 }
902 }
903 } else if (IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_laddr)) {
904 if (IN6_IS_ADDR_V4MAPPED(laddr6)) {
905 if ((in6p->in6p_flags & IN6P_IPV6_V6ONLY) != 0)
906 continue;
907 }
908 if (!IN6_IS_ADDR_UNSPECIFIED(laddr6))
909 wildcard++;
910 } else {
911 if (IN6_IS_ADDR_V4MAPPED(laddr6)) {
912 if ((in6p->in6p_flags & IN6P_IPV6_V6ONLY) != 0)
913 continue;
914 }
915 if (IN6_IS_ADDR_UNSPECIFIED(laddr6))
916 wildcard++;
917 else {
918 if (!IN6_ARE_ADDR_EQUAL(&in6p->in6p_laddr,
919 laddr6))
920 continue;
921 }
922 }
923 if (wildcard && !lookup_wildcard)
924 continue;
925 if (wildcard < matchwild) {
926 match = in6p;
927 matchwild = wildcard;
928 if (matchwild == 0)
929 break;
930 }
931 }
932 return (match);
933 }
934 #undef continue
935
936 /*
937 * WARNING: return value (rtentry) could be IPv4 one if in6pcb is connected to
938 * IPv4 mapped address.
939 */
940 struct rtentry *
941 in6_pcbrtentry(in6p)
942 struct in6pcb *in6p;
943 {
944 struct route_in6 *ro;
945 struct sockaddr_in6 *dst6;
946
947 ro = &in6p->in6p_route;
948 dst6 = (struct sockaddr_in6 *)&ro->ro_dst;
949
950 if (in6p->in6p_af != AF_INET6)
951 return (NULL);
952
953 if (ro->ro_rt && ((ro->ro_rt->rt_flags & RTF_UP) == 0 ||
954 !IN6_ARE_ADDR_EQUAL(&dst6->sin6_addr, &in6p->in6p_faddr))) {
955 RTFREE(ro->ro_rt);
956 ro->ro_rt = (struct rtentry *)NULL;
957 }
958 #ifdef INET
959 if (ro->ro_rt == (struct rtentry *)NULL &&
960 IN6_IS_ADDR_V4MAPPED(&in6p->in6p_faddr)) {
961 struct sockaddr_in *dst = (struct sockaddr_in *)&ro->ro_dst;
962
963 bzero(dst, sizeof(*dst));
964 dst->sin_family = AF_INET;
965 dst->sin_len = sizeof(struct sockaddr_in);
966 bcopy(&in6p->in6p_faddr.s6_addr32[3], &dst->sin_addr,
967 sizeof(dst->sin_addr));
968 rtalloc((struct route *)ro);
969 } else
970 #endif
971 if (ro->ro_rt == (struct rtentry *)NULL &&
972 !IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_faddr)) {
973 bzero(dst6, sizeof(*dst6));
974 dst6->sin6_family = AF_INET6;
975 dst6->sin6_len = sizeof(struct sockaddr_in6);
976 dst6->sin6_addr = in6p->in6p_faddr;
977 rtalloc((struct route *)ro);
978 }
979 return (ro->ro_rt);
980 }
981
982 struct in6pcb *
983 in6_pcblookup_connect(table, faddr6, fport_arg, laddr6, lport_arg, faith)
984 struct inpcbtable *table;
985 struct in6_addr *faddr6;
986 const struct in6_addr *laddr6;
987 u_int fport_arg, lport_arg;
988 int faith;
989 {
990 struct inpcbhead *head;
991 struct inpcb_hdr *inph;
992 struct in6pcb *in6p;
993 u_int16_t fport = fport_arg, lport = lport_arg;
994
995 head = IN6PCBHASH_CONNECT(table, faddr6, fport, laddr6, lport);
996 LIST_FOREACH(inph, head, inph_hash) {
997 in6p = (struct in6pcb *)inph;
998 if (in6p->in6p_af != AF_INET6)
999 continue;
1000
1001 /* find exact match on both source and dest */
1002 if (in6p->in6p_fport != fport)
1003 continue;
1004 if (in6p->in6p_lport != lport)
1005 continue;
1006 if (IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_faddr))
1007 continue;
1008 if (!IN6_ARE_ADDR_EQUAL(&in6p->in6p_faddr, faddr6))
1009 continue;
1010 if (IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_laddr))
1011 continue;
1012 if (!IN6_ARE_ADDR_EQUAL(&in6p->in6p_laddr, laddr6))
1013 continue;
1014 if ((IN6_IS_ADDR_V4MAPPED(laddr6) ||
1015 IN6_IS_ADDR_V4MAPPED(faddr6)) &&
1016 (in6p->in6p_flags & IN6P_IPV6_V6ONLY))
1017 continue;
1018 return in6p;
1019 }
1020 return NULL;
1021 }
1022
1023 struct in6pcb *
1024 in6_pcblookup_bind(table, laddr6, lport_arg, faith)
1025 struct inpcbtable *table;
1026 struct in6_addr *laddr6;
1027 u_int lport_arg;
1028 int faith;
1029 {
1030 struct inpcbhead *head;
1031 struct inpcb_hdr *inph;
1032 struct in6pcb *in6p;
1033 u_int16_t lport = lport_arg;
1034 #ifdef INET
1035 struct in6_addr zero_mapped;
1036 #endif
1037
1038 head = IN6PCBHASH_BIND(table, laddr6, lport);
1039 LIST_FOREACH(inph, head, inph_hash) {
1040 in6p = (struct in6pcb *)inph;
1041 if (in6p->in6p_af != AF_INET6)
1042 continue;
1043
1044 if (faith && (in6p->in6p_flags & IN6P_FAITH) == 0)
1045 continue;
1046 if (in6p->in6p_fport != 0)
1047 continue;
1048 if (in6p->in6p_lport != lport)
1049 continue;
1050 if (IN6_IS_ADDR_V4MAPPED(laddr6) &&
1051 (in6p->in6p_flags & IN6P_IPV6_V6ONLY) != 0)
1052 continue;
1053 if (IN6_ARE_ADDR_EQUAL(&in6p->in6p_laddr, laddr6))
1054 goto out;
1055 }
1056 #ifdef INET
1057 if (IN6_IS_ADDR_V4MAPPED(laddr6)) {
1058 memset(&zero_mapped, 0, sizeof(zero_mapped));
1059 zero_mapped.s6_addr16[5] = 0xffff;
1060 head = IN6PCBHASH_BIND(table, &zero_mapped, lport);
1061 LIST_FOREACH(inph, head, inph_hash) {
1062 in6p = (struct in6pcb *)inph;
1063 if (in6p->in6p_af != AF_INET6)
1064 continue;
1065
1066 if (faith && (in6p->in6p_flags & IN6P_FAITH) == 0)
1067 continue;
1068 if (in6p->in6p_fport != 0)
1069 continue;
1070 if (in6p->in6p_lport != lport)
1071 continue;
1072 if ((in6p->in6p_flags & IN6P_IPV6_V6ONLY) != 0)
1073 continue;
1074 if (IN6_ARE_ADDR_EQUAL(&in6p->in6p_laddr, &zero_mapped))
1075 goto out;
1076 }
1077 }
1078 #endif
1079 head = IN6PCBHASH_BIND(table, &zeroin6_addr, lport);
1080 LIST_FOREACH(inph, head, inph_hash) {
1081 in6p = (struct in6pcb *)inph;
1082 if (in6p->in6p_af != AF_INET6)
1083 continue;
1084
1085 if (faith && (in6p->in6p_flags & IN6P_FAITH) == 0)
1086 continue;
1087 if (in6p->in6p_fport != 0)
1088 continue;
1089 if (in6p->in6p_lport != lport)
1090 continue;
1091 if (IN6_IS_ADDR_V4MAPPED(laddr6) &&
1092 (in6p->in6p_flags & IN6P_IPV6_V6ONLY) != 0)
1093 continue;
1094 if (IN6_ARE_ADDR_EQUAL(&in6p->in6p_laddr, &zeroin6_addr))
1095 goto out;
1096 }
1097 return (NULL);
1098
1099 out:
1100 inph = &in6p->in6p_head;
1101 if (inph != LIST_FIRST(head)) {
1102 LIST_REMOVE(inph, inph_hash);
1103 LIST_INSERT_HEAD(head, inph, inph_hash);
1104 }
1105 return in6p;
1106 }
1107
1108 void
1109 in6_pcbstate(in6p, state)
1110 struct in6pcb *in6p;
1111 int state;
1112 {
1113
1114 if (in6p->in6p_af != AF_INET6)
1115 return;
1116
1117 if (in6p->in6p_state > IN6P_ATTACHED)
1118 LIST_REMOVE(&in6p->in6p_head, inph_hash);
1119
1120 switch (state) {
1121 case IN6P_BOUND:
1122 LIST_INSERT_HEAD(IN6PCBHASH_BIND(in6p->in6p_table,
1123 &in6p->in6p_laddr, in6p->in6p_lport), &in6p->in6p_head,
1124 inph_hash);
1125 break;
1126 case IN6P_CONNECTED:
1127 LIST_INSERT_HEAD(IN6PCBHASH_CONNECT(in6p->in6p_table,
1128 &in6p->in6p_faddr, in6p->in6p_fport,
1129 &in6p->in6p_laddr, in6p->in6p_lport), &in6p->in6p_head,
1130 inph_hash);
1131 break;
1132 }
1133
1134 in6p->in6p_state = state;
1135 }
1136