1 /* $NetBSD: ip6_input.c,v 1.73.2.3 2007/06/04 19:24:35 bouyer Exp $ */
2 /* $KAME: ip6_input.c,v 1.188 2001/03/29 05:34:31 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, 1988, 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 * @(#)ip_input.c 8.2 (Berkeley) 1/4/94
62 */
63
64 #include <sys/cdefs.h>
65 __KERNEL_RCSID(0, "$NetBSD: ip6_input.c,v 1.73.2.3 2007/06/04 19:24:35 bouyer Exp $");
66
67 #include "opt_inet.h"
68 #include "opt_ipsec.h"
69 #include "opt_pfil_hooks.h"
70
71 #include <sys/param.h>
72 #include <sys/systm.h>
73 #include <sys/malloc.h>
74 #include <sys/mbuf.h>
75 #include <sys/domain.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/kernel.h>
82 #include <sys/syslog.h>
83 #include <sys/proc.h>
84 #include <sys/sysctl.h>
85
86 #include <net/if.h>
87 #include <net/if_types.h>
88 #include <net/if_dl.h>
89 #include <net/route.h>
90 #include <net/netisr.h>
91 #ifdef PFIL_HOOKS
92 #include <net/pfil.h>
93 #endif
94
95 #include <netinet/in.h>
96 #include <netinet/in_systm.h>
97 #ifdef INET
98 #include <netinet/ip.h>
99 #include <netinet/ip_icmp.h>
100 #endif /* INET */
101 #include <netinet/ip6.h>
102 #include <netinet6/in6_var.h>
103 #include <netinet6/ip6_var.h>
104 #include <netinet6/in6_pcb.h>
105 #include <netinet/icmp6.h>
106 #include <netinet6/in6_ifattach.h>
107 #include <netinet6/nd6.h>
108
109 #ifdef IPSEC
110 #include <netinet6/ipsec.h>
111 #endif
112
113 #include <netinet6/ip6protosw.h>
114
115 /* we need it for NLOOP. */
116 #include "loop.h"
117 #include "faith.h"
118 #include "gif.h"
119 #include "bpfilter.h"
120
121 #if NGIF > 0
122 #include <netinet6/in6_gif.h>
123 #endif
124
125 #include <net/net_osdep.h>
126
127 extern struct domain inet6domain;
128
129 u_char ip6_protox[IPPROTO_MAX];
130 static int ip6qmaxlen = IFQ_MAXLEN;
131 struct in6_ifaddr *in6_ifaddr;
132 struct ifqueue ip6intrq;
133
134 extern struct ifnet loif[NLOOP];
135 int ip6_forward_srcrt; /* XXX */
136 int ip6_sourcecheck; /* XXX */
137 int ip6_sourcecheck_interval; /* XXX */
138
139 #ifdef PFIL_HOOKS
140 struct pfil_head inet6_pfil_hook;
141 #endif
142
143 struct ip6stat ip6stat;
144
145 static void ip6_init2 __P((void *));
146
147 static int ip6_hopopts_input __P((u_int32_t *, u_int32_t *, struct mbuf **, int *));
148 static struct mbuf *ip6_pullexthdr __P((struct mbuf *, size_t, int));
149
150 /*
151 * IP6 initialization: fill in IP6 protocol switch table.
152 * All protocols not implemented in kernel go to raw IP6 protocol handler.
153 */
154 void
155 ip6_init()
156 {
157 struct ip6protosw *pr;
158 int i;
159
160 pr = (struct ip6protosw *)pffindproto(PF_INET6, IPPROTO_RAW, SOCK_RAW);
161 if (pr == 0)
162 panic("ip6_init");
163 for (i = 0; i < IPPROTO_MAX; i++)
164 ip6_protox[i] = pr - inet6sw;
165 for (pr = (struct ip6protosw *)inet6domain.dom_protosw;
166 pr < (struct ip6protosw *)inet6domain.dom_protoswNPROTOSW; pr++)
167 if (pr->pr_domain->dom_family == PF_INET6 &&
168 pr->pr_protocol && pr->pr_protocol != IPPROTO_RAW)
169 ip6_protox[pr->pr_protocol] = pr - inet6sw;
170 ip6intrq.ifq_maxlen = ip6qmaxlen;
171 nd6_init();
172 frag6_init();
173
174 ip6_init2((void *)0);
175
176 #ifdef PFIL_HOOKS
177 /* Register our Packet Filter hook. */
178 inet6_pfil_hook.ph_type = PFIL_TYPE_AF;
179 inet6_pfil_hook.ph_af = AF_INET6;
180 i = pfil_head_register(&inet6_pfil_hook);
181 if (i != 0)
182 printf("ip6_init: WARNING: unable to register pfil hook, "
183 "error %d\n", i);
184 #endif /* PFIL_HOOKS */
185 }
186
187 static void
188 ip6_init2(dummy)
189 void *dummy;
190 {
191
192 /* nd6_timer_init */
193 callout_init(&nd6_timer_ch);
194 callout_reset(&nd6_timer_ch, hz, nd6_timer, NULL);
195 }
196
197 /*
198 * IP6 input interrupt handling. Just pass the packet to ip6_input.
199 */
200 void
201 ip6intr()
202 {
203 int s;
204 struct mbuf *m;
205
206 for (;;) {
207 s = splnet();
208 IF_DEQUEUE(&ip6intrq, m);
209 splx(s);
210 if (m == 0)
211 return;
212 ip6_input(m);
213 }
214 }
215
216 extern struct route_in6 ip6_forward_rt;
217
218 void
219 ip6_input(m)
220 struct mbuf *m;
221 {
222 struct ip6_hdr *ip6;
223 int off = sizeof(struct ip6_hdr), nest;
224 u_int32_t plen;
225 u_int32_t rtalert = ~0;
226 int nxt, ours = 0;
227 struct ifnet *deliverifp = NULL;
228 int srcrt = 0;
229
230 #ifdef IPSEC
231 /*
232 * should the inner packet be considered authentic?
233 * see comment in ah4_input().
234 */
235 m->m_flags &= ~M_AUTHIPHDR;
236 m->m_flags &= ~M_AUTHIPDGM;
237 #endif
238
239 /*
240 * mbuf statistics
241 */
242 if (m->m_flags & M_EXT) {
243 if (m->m_next)
244 ip6stat.ip6s_mext2m++;
245 else
246 ip6stat.ip6s_mext1++;
247 } else {
248 #define M2MMAX (sizeof(ip6stat.ip6s_m2m)/sizeof(ip6stat.ip6s_m2m[0]))
249 if (m->m_next) {
250 if (m->m_flags & M_LOOP) {
251 ip6stat.ip6s_m2m[loif[0].if_index]++; /* XXX */
252 } else if (m->m_pkthdr.rcvif->if_index < M2MMAX)
253 ip6stat.ip6s_m2m[m->m_pkthdr.rcvif->if_index]++;
254 else
255 ip6stat.ip6s_m2m[0]++;
256 } else
257 ip6stat.ip6s_m1++;
258 #undef M2MMAX
259 }
260
261 in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_receive);
262 ip6stat.ip6s_total++;
263
264 /*
265 * If the IPv6 header is not aligned, slurp it up into a new
266 * mbuf with space for link headers, in the event we forward
267 * it. OTherwise, if it is aligned, make sure the entire base
268 * IPv6 header is in the first mbuf of the chain.
269 */
270 if (IP6_HDR_ALIGNED_P(mtod(m, caddr_t)) == 0) {
271 struct ifnet *inifp = m->m_pkthdr.rcvif;
272 if ((m = m_copyup(m, sizeof(struct ip6_hdr),
273 (max_linkhdr + 3) & ~3)) == NULL) {
274 /* XXXJRT new stat, please */
275 ip6stat.ip6s_toosmall++;
276 in6_ifstat_inc(inifp, ifs6_in_hdrerr);
277 return;
278 }
279 } else if (__predict_false(m->m_len < sizeof(struct ip6_hdr))) {
280 struct ifnet *inifp = m->m_pkthdr.rcvif;
281 if ((m = m_pullup(m, sizeof(struct ip6_hdr))) == NULL) {
282 ip6stat.ip6s_toosmall++;
283 in6_ifstat_inc(inifp, ifs6_in_hdrerr);
284 return;
285 }
286 }
287
288 ip6 = mtod(m, struct ip6_hdr *);
289
290 if ((ip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION) {
291 ip6stat.ip6s_badvers++;
292 in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_hdrerr);
293 goto bad;
294 }
295
296 #ifdef PFIL_HOOKS
297 /*
298 * Run through list of hooks for input packets. If there are any
299 * filters which require that additional packets in the flow are
300 * not fast-forwarded, they must clear the M_CANFASTFWD flag.
301 * Note that filters must _never_ set this flag, as another filter
302 * in the list may have previously cleared it.
303 */
304 /*
305 * let ipfilter look at packet on the wire,
306 * not the decapsulated packet.
307 */
308 #ifdef IPSEC
309 if (!ipsec_getnhist(m))
310 #else
311 if (1)
312 #endif
313 {
314 struct in6_addr odst;
315
316 odst = ip6->ip6_dst;
317 if (pfil_run_hooks(&inet6_pfil_hook, &m, m->m_pkthdr.rcvif,
318 PFIL_IN) != 0)
319 return;
320 if (m == NULL)
321 return;
322 ip6 = mtod(m, struct ip6_hdr *);
323 srcrt = !IN6_ARE_ADDR_EQUAL(&odst, &ip6->ip6_dst);
324 }
325 #endif /* PFIL_HOOKS */
326
327 ip6stat.ip6s_nxthist[ip6->ip6_nxt]++;
328
329 #ifdef ALTQ
330 if (altq_input != NULL && (*altq_input)(m, AF_INET6) == 0) {
331 /* packet is dropped by traffic conditioner */
332 return;
333 }
334 #endif
335
336 /*
337 * Check against address spoofing/corruption.
338 */
339 if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_src) ||
340 IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_dst)) {
341 /*
342 * XXX: "badscope" is not very suitable for a multicast source.
343 */
344 ip6stat.ip6s_badscope++;
345 in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr);
346 goto bad;
347 }
348 /*
349 * The following check is not documented in specs. A malicious
350 * party may be able to use IPv4 mapped addr to confuse tcp/udp stack
351 * and bypass security checks (act as if it was from 127.0.0.1 by using
352 * IPv6 src ::ffff:127.0.0.1). Be cautious.
353 *
354 * This check chokes if we are in an SIIT cloud. As none of BSDs
355 * support IPv4-less kernel compilation, we cannot support SIIT
356 * environment at all. So, it makes more sense for us to reject any
357 * malicious packets for non-SIIT environment, than try to do a
358 * partial support for SIIT environment.
359 */
360 if (IN6_IS_ADDR_V4MAPPED(&ip6->ip6_src) ||
361 IN6_IS_ADDR_V4MAPPED(&ip6->ip6_dst)) {
362 ip6stat.ip6s_badscope++;
363 in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr);
364 goto bad;
365 }
366 #if 0
367 /*
368 * Reject packets with IPv4 compatible addresses (auto tunnel).
369 *
370 * The code forbids auto tunnel relay case in RFC1933 (the check is
371 * stronger than RFC1933). We may want to re-enable it if mech-xx
372 * is revised to forbid relaying case.
373 */
374 if (IN6_IS_ADDR_V4COMPAT(&ip6->ip6_src) ||
375 IN6_IS_ADDR_V4COMPAT(&ip6->ip6_dst)) {
376 ip6stat.ip6s_badscope++;
377 in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr);
378 goto bad;
379 }
380 #endif
381
382 if (IN6_IS_ADDR_LOOPBACK(&ip6->ip6_src) ||
383 IN6_IS_ADDR_LOOPBACK(&ip6->ip6_dst)) {
384 if (m->m_pkthdr.rcvif->if_flags & IFF_LOOPBACK) {
385 ours = 1;
386 deliverifp = m->m_pkthdr.rcvif;
387 goto hbhcheck;
388 } else {
389 ip6stat.ip6s_badscope++;
390 in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr);
391 goto bad;
392 }
393 }
394
395 /* drop packets if interface ID portion is already filled */
396 if ((m->m_pkthdr.rcvif->if_flags & IFF_LOOPBACK) == 0) {
397 if (IN6_IS_SCOPE_LINKLOCAL(&ip6->ip6_src) &&
398 ip6->ip6_src.s6_addr16[1]) {
399 ip6stat.ip6s_badscope++;
400 goto bad;
401 }
402 if (IN6_IS_SCOPE_LINKLOCAL(&ip6->ip6_dst) &&
403 ip6->ip6_dst.s6_addr16[1]) {
404 ip6stat.ip6s_badscope++;
405 goto bad;
406 }
407 }
408
409 if (IN6_IS_SCOPE_LINKLOCAL(&ip6->ip6_src))
410 ip6->ip6_src.s6_addr16[1]
411 = htons(m->m_pkthdr.rcvif->if_index);
412 if (IN6_IS_SCOPE_LINKLOCAL(&ip6->ip6_dst))
413 ip6->ip6_dst.s6_addr16[1]
414 = htons(m->m_pkthdr.rcvif->if_index);
415
416 /*
417 * We use rt->rt_ifp to determine if the address is ours or not.
418 * If rt_ifp is lo0, the address is ours.
419 * The problem here is, rt->rt_ifp for fe80::%lo0/64 is set to lo0,
420 * so any address under fe80::%lo0/64 will be mistakenly considered
421 * local. The special case is supplied to handle the case properly
422 * by actually looking at interface addresses
423 * (using in6ifa_ifpwithaddr).
424 */
425 if ((m->m_pkthdr.rcvif->if_flags & IFF_LOOPBACK) != 0 &&
426 IN6_IS_ADDR_LINKLOCAL(&ip6->ip6_dst)) {
427 if (!in6ifa_ifpwithaddr(m->m_pkthdr.rcvif, &ip6->ip6_dst)) {
428 icmp6_error(m, ICMP6_DST_UNREACH,
429 ICMP6_DST_UNREACH_ADDR, 0);
430 /* m is already freed */
431 return;
432 }
433
434 ours = 1;
435 deliverifp = m->m_pkthdr.rcvif;
436 goto hbhcheck;
437 }
438
439 /*
440 * Multicast check
441 */
442 if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) {
443 struct in6_multi *in6m = 0;
444
445 in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_mcast);
446 /*
447 * See if we belong to the destination multicast group on the
448 * arrival interface.
449 */
450 IN6_LOOKUP_MULTI(ip6->ip6_dst, m->m_pkthdr.rcvif, in6m);
451 if (in6m)
452 ours = 1;
453 else if (!ip6_mrouter) {
454 ip6stat.ip6s_notmember++;
455 ip6stat.ip6s_cantforward++;
456 in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_discard);
457 goto bad;
458 }
459 deliverifp = m->m_pkthdr.rcvif;
460 goto hbhcheck;
461 }
462
463 /*
464 * Unicast check
465 */
466 if (ip6_forward_rt.ro_rt != NULL &&
467 (ip6_forward_rt.ro_rt->rt_flags & RTF_UP) != 0 &&
468 IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst,
469 &((struct sockaddr_in6 *)(&ip6_forward_rt.ro_dst))->sin6_addr))
470 ip6stat.ip6s_forward_cachehit++;
471 else {
472 struct sockaddr_in6 *dst6;
473
474 if (ip6_forward_rt.ro_rt) {
475 /* route is down or destination is different */
476 ip6stat.ip6s_forward_cachemiss++;
477 RTFREE(ip6_forward_rt.ro_rt);
478 ip6_forward_rt.ro_rt = 0;
479 }
480
481 bzero(&ip6_forward_rt.ro_dst, sizeof(struct sockaddr_in6));
482 dst6 = (struct sockaddr_in6 *)&ip6_forward_rt.ro_dst;
483 dst6->sin6_len = sizeof(struct sockaddr_in6);
484 dst6->sin6_family = AF_INET6;
485 dst6->sin6_addr = ip6->ip6_dst;
486
487 rtalloc((struct route *)&ip6_forward_rt);
488 }
489
490 #define rt6_key(r) ((struct sockaddr_in6 *)((r)->rt_nodes->rn_key))
491
492 /*
493 * Accept the packet if the forwarding interface to the destination
494 * according to the routing table is the loopback interface,
495 * unless the associated route has a gateway.
496 * Note that this approach causes to accept a packet if there is a
497 * route to the loopback interface for the destination of the packet.
498 * But we think it's even useful in some situations, e.g. when using
499 * a special daemon which wants to intercept the packet.
500 */
501 if (ip6_forward_rt.ro_rt &&
502 (ip6_forward_rt.ro_rt->rt_flags &
503 (RTF_HOST|RTF_GATEWAY)) == RTF_HOST &&
504 !(ip6_forward_rt.ro_rt->rt_flags & RTF_CLONED) &&
505 #if 0
506 /*
507 * The check below is redundant since the comparison of
508 * the destination and the key of the rtentry has
509 * already done through looking up the routing table.
510 */
511 IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst,
512 &rt6_key(ip6_forward_rt.ro_rt)->sin6_addr) &&
513 #endif
514 ip6_forward_rt.ro_rt->rt_ifp->if_type == IFT_LOOP) {
515 struct in6_ifaddr *ia6 =
516 (struct in6_ifaddr *)ip6_forward_rt.ro_rt->rt_ifa;
517 if (ia6->ia6_flags & IN6_IFF_ANYCAST)
518 m->m_flags |= M_ANYCAST6;
519 /*
520 * packets to a tentative, duplicated, or somehow invalid
521 * address must not be accepted.
522 */
523 if (!(ia6->ia6_flags & IN6_IFF_NOTREADY)) {
524 /* this address is ready */
525 ours = 1;
526 deliverifp = ia6->ia_ifp; /* correct? */
527 goto hbhcheck;
528 } else {
529 /* address is not ready, so discard the packet. */
530 nd6log((LOG_INFO,
531 "ip6_input: packet to an unready address %s->%s\n",
532 ip6_sprintf(&ip6->ip6_src),
533 ip6_sprintf(&ip6->ip6_dst)));
534
535 goto bad;
536 }
537 }
538
539 /*
540 * FAITH (Firewall Aided Internet Translator)
541 */
542 #if defined(NFAITH) && 0 < NFAITH
543 if (ip6_keepfaith) {
544 if (ip6_forward_rt.ro_rt && ip6_forward_rt.ro_rt->rt_ifp &&
545 ip6_forward_rt.ro_rt->rt_ifp->if_type == IFT_FAITH) {
546 /* XXX do we need more sanity checks? */
547 ours = 1;
548 deliverifp = ip6_forward_rt.ro_rt->rt_ifp; /* faith */
549 goto hbhcheck;
550 }
551 }
552 #endif
553
554 #if 0
555 {
556 /*
557 * Last resort: check in6_ifaddr for incoming interface.
558 * The code is here until I update the "goto ours hack" code above
559 * working right.
560 */
561 struct ifaddr *ifa;
562 for (ifa = m->m_pkthdr.rcvif->if_addrlist.tqh_first;
563 ifa;
564 ifa = ifa->ifa_list.tqe_next) {
565 if (ifa->ifa_addr == NULL)
566 continue; /* just for safety */
567 if (ifa->ifa_addr->sa_family != AF_INET6)
568 continue;
569 if (IN6_ARE_ADDR_EQUAL(IFA_IN6(ifa), &ip6->ip6_dst)) {
570 ours = 1;
571 deliverifp = ifa->ifa_ifp;
572 goto hbhcheck;
573 }
574 }
575 }
576 #endif
577
578 /*
579 * Now there is no reason to process the packet if it's not our own
580 * and we're not a router.
581 */
582 if (!ip6_forwarding) {
583 ip6stat.ip6s_cantforward++;
584 in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_discard);
585 goto bad;
586 }
587
588 hbhcheck:
589 /*
590 * Process Hop-by-Hop options header if it's contained.
591 * m may be modified in ip6_hopopts_input().
592 * If a JumboPayload option is included, plen will also be modified.
593 */
594 plen = (u_int32_t)ntohs(ip6->ip6_plen);
595 if (ip6->ip6_nxt == IPPROTO_HOPOPTS) {
596 struct ip6_hbh *hbh;
597
598 if (ip6_hopopts_input(&plen, &rtalert, &m, &off)) {
599 #if 0 /*touches NULL pointer*/
600 in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_discard);
601 #endif
602 return; /* m have already been freed */
603 }
604
605 /* adjust pointer */
606 ip6 = mtod(m, struct ip6_hdr *);
607
608 /*
609 * if the payload length field is 0 and the next header field
610 * indicates Hop-by-Hop Options header, then a Jumbo Payload
611 * option MUST be included.
612 */
613 if (ip6->ip6_plen == 0 && plen == 0) {
614 /*
615 * Note that if a valid jumbo payload option is
616 * contained, ip6_hoptops_input() must set a valid
617 * (non-zero) payload length to the variable plen.
618 */
619 ip6stat.ip6s_badoptions++;
620 in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_discard);
621 in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_hdrerr);
622 icmp6_error(m, ICMP6_PARAM_PROB,
623 ICMP6_PARAMPROB_HEADER,
624 (caddr_t)&ip6->ip6_plen - (caddr_t)ip6);
625 return;
626 }
627 IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m, sizeof(struct ip6_hdr),
628 sizeof(struct ip6_hbh));
629 if (hbh == NULL) {
630 ip6stat.ip6s_tooshort++;
631 return;
632 }
633 KASSERT(IP6_HDR_ALIGNED_P(hbh));
634 nxt = hbh->ip6h_nxt;
635
636 /*
637 * accept the packet if a router alert option is included
638 * and we act as an IPv6 router.
639 */
640 if (rtalert != ~0 && ip6_forwarding)
641 ours = 1;
642 } else
643 nxt = ip6->ip6_nxt;
644
645 /*
646 * Check that the amount of data in the buffers
647 * is as at least much as the IPv6 header would have us expect.
648 * Trim mbufs if longer than we expect.
649 * Drop packet if shorter than we expect.
650 */
651 if (m->m_pkthdr.len - sizeof(struct ip6_hdr) < plen) {
652 ip6stat.ip6s_tooshort++;
653 in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_truncated);
654 goto bad;
655 }
656 if (m->m_pkthdr.len > sizeof(struct ip6_hdr) + plen) {
657 if (m->m_len == m->m_pkthdr.len) {
658 m->m_len = sizeof(struct ip6_hdr) + plen;
659 m->m_pkthdr.len = sizeof(struct ip6_hdr) + plen;
660 } else
661 m_adj(m, sizeof(struct ip6_hdr) + plen - m->m_pkthdr.len);
662 }
663
664 /*
665 * Forward if desirable.
666 */
667 if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) {
668 /*
669 * If we are acting as a multicast router, all
670 * incoming multicast packets are passed to the
671 * kernel-level multicast forwarding function.
672 * The packet is returned (relatively) intact; if
673 * ip6_mforward() returns a non-zero value, the packet
674 * must be discarded, else it may be accepted below.
675 */
676 if (ip6_mrouter && ip6_mforward(ip6, m->m_pkthdr.rcvif, m)) {
677 ip6stat.ip6s_cantforward++;
678 m_freem(m);
679 return;
680 }
681 if (!ours) {
682 m_freem(m);
683 return;
684 }
685 } else if (!ours) {
686 ip6_forward(m, srcrt);
687 return;
688 }
689
690 ip6 = mtod(m, struct ip6_hdr *);
691
692 /*
693 * Malicious party may be able to use IPv4 mapped addr to confuse
694 * tcp/udp stack and bypass security checks (act as if it was from
695 * 127.0.0.1 by using IPv6 src ::ffff:127.0.0.1). Be cautious.
696 *
697 * For SIIT end node behavior, you may want to disable the check.
698 * However, you will become vulnerable to attacks using IPv4 mapped
699 * source.
700 */
701 if (IN6_IS_ADDR_V4MAPPED(&ip6->ip6_src) ||
702 IN6_IS_ADDR_V4MAPPED(&ip6->ip6_dst)) {
703 ip6stat.ip6s_badscope++;
704 in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr);
705 goto bad;
706 }
707
708 /*
709 * Tell launch routine the next header
710 */
711 #ifdef IFA_STATS
712 if (deliverifp != NULL) {
713 struct in6_ifaddr *ia6;
714 ia6 = in6_ifawithifp(deliverifp, &ip6->ip6_dst);
715 if (ia6)
716 ia6->ia_ifa.ifa_data.ifad_inbytes += m->m_pkthdr.len;
717 }
718 #endif
719 ip6stat.ip6s_delivered++;
720 in6_ifstat_inc(deliverifp, ifs6_in_deliver);
721 nest = 0;
722
723 while (nxt != IPPROTO_DONE) {
724 if (ip6_hdrnestlimit && (++nest > ip6_hdrnestlimit)) {
725 ip6stat.ip6s_toomanyhdr++;
726 goto bad;
727 }
728
729 /*
730 * protection against faulty packet - there should be
731 * more sanity checks in header chain processing.
732 */
733 if (m->m_pkthdr.len < off) {
734 ip6stat.ip6s_tooshort++;
735 in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_truncated);
736 goto bad;
737 }
738
739 #ifdef IPSEC
740 /*
741 * enforce IPsec policy checking if we are seeing last header.
742 * note that we do not visit this with protocols with pcb layer
743 * code - like udp/tcp/raw ip.
744 */
745 if ((inet6sw[ip6_protox[nxt]].pr_flags & PR_LASTHDR) != 0 &&
746 ipsec6_in_reject(m, NULL)) {
747 ipsec6stat.in_polvio++;
748 goto bad;
749 }
750 #endif
751
752 nxt = (*inet6sw[ip6_protox[nxt]].pr_input)(&m, &off, nxt);
753 }
754 return;
755 bad:
756 m_freem(m);
757 }
758
759 /*
760 * Hop-by-Hop options header processing. If a valid jumbo payload option is
761 * included, the real payload length will be stored in plenp.
762 */
763 static int
764 ip6_hopopts_input(plenp, rtalertp, mp, offp)
765 u_int32_t *plenp;
766 u_int32_t *rtalertp; /* XXX: should be stored more smart way */
767 struct mbuf **mp;
768 int *offp;
769 {
770 struct mbuf *m = *mp;
771 int off = *offp, hbhlen;
772 struct ip6_hbh *hbh;
773
774 /* validation of the length of the header */
775 IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m,
776 sizeof(struct ip6_hdr), sizeof(struct ip6_hbh));
777 if (hbh == NULL) {
778 ip6stat.ip6s_tooshort++;
779 return -1;
780 }
781 hbhlen = (hbh->ip6h_len + 1) << 3;
782 IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m, sizeof(struct ip6_hdr),
783 hbhlen);
784 if (hbh == NULL) {
785 ip6stat.ip6s_tooshort++;
786 return -1;
787 }
788 KASSERT(IP6_HDR_ALIGNED_P(hbh));
789 off += hbhlen;
790 hbhlen -= sizeof(struct ip6_hbh);
791
792 if (ip6_process_hopopts(m, (u_int8_t *)hbh + sizeof(struct ip6_hbh),
793 hbhlen, rtalertp, plenp) < 0)
794 return (-1);
795
796 *offp = off;
797 *mp = m;
798 return (0);
799 }
800
801 /*
802 * Search header for all Hop-by-hop options and process each option.
803 * This function is separate from ip6_hopopts_input() in order to
804 * handle a case where the sending node itself process its hop-by-hop
805 * options header. In such a case, the function is called from ip6_output().
806 *
807 * The function assumes that hbh header is located right after the IPv6 header
808 * (RFC2460 p7), opthead is pointer into data content in m, and opthead to
809 * opthead + hbhlen is located in continuous memory region.
810 */
811 int
812 ip6_process_hopopts(m, opthead, hbhlen, rtalertp, plenp)
813 struct mbuf *m;
814 u_int8_t *opthead;
815 int hbhlen;
816 u_int32_t *rtalertp;
817 u_int32_t *plenp;
818 {
819 struct ip6_hdr *ip6;
820 int optlen = 0;
821 u_int8_t *opt = opthead;
822 u_int16_t rtalert_val;
823 u_int32_t jumboplen;
824 const int erroff = sizeof(struct ip6_hdr) + sizeof(struct ip6_hbh);
825
826 for (; hbhlen > 0; hbhlen -= optlen, opt += optlen) {
827 switch (*opt) {
828 case IP6OPT_PAD1:
829 optlen = 1;
830 break;
831 case IP6OPT_PADN:
832 if (hbhlen < IP6OPT_MINLEN) {
833 ip6stat.ip6s_toosmall++;
834 goto bad;
835 }
836 optlen = *(opt + 1) + 2;
837 break;
838 case IP6OPT_RTALERT:
839 /* XXX may need check for alignment */
840 if (hbhlen < IP6OPT_RTALERT_LEN) {
841 ip6stat.ip6s_toosmall++;
842 goto bad;
843 }
844 if (*(opt + 1) != IP6OPT_RTALERT_LEN - 2) {
845 /* XXX stat */
846 icmp6_error(m, ICMP6_PARAM_PROB,
847 ICMP6_PARAMPROB_HEADER,
848 erroff + opt + 1 - opthead);
849 return (-1);
850 }
851 optlen = IP6OPT_RTALERT_LEN;
852 bcopy((caddr_t)(opt + 2), (caddr_t)&rtalert_val, 2);
853 *rtalertp = ntohs(rtalert_val);
854 break;
855 case IP6OPT_JUMBO:
856 /* XXX may need check for alignment */
857 if (hbhlen < IP6OPT_JUMBO_LEN) {
858 ip6stat.ip6s_toosmall++;
859 goto bad;
860 }
861 if (*(opt + 1) != IP6OPT_JUMBO_LEN - 2) {
862 /* XXX stat */
863 icmp6_error(m, ICMP6_PARAM_PROB,
864 ICMP6_PARAMPROB_HEADER,
865 erroff + opt + 1 - opthead);
866 return (-1);
867 }
868 optlen = IP6OPT_JUMBO_LEN;
869
870 /*
871 * IPv6 packets that have non 0 payload length
872 * must not contain a jumbo payload option.
873 */
874 ip6 = mtod(m, struct ip6_hdr *);
875 if (ip6->ip6_plen) {
876 ip6stat.ip6s_badoptions++;
877 icmp6_error(m, ICMP6_PARAM_PROB,
878 ICMP6_PARAMPROB_HEADER,
879 erroff + opt - opthead);
880 return (-1);
881 }
882
883 /*
884 * We may see jumbolen in unaligned location, so
885 * we'd need to perform bcopy().
886 */
887 bcopy(opt + 2, &jumboplen, sizeof(jumboplen));
888 jumboplen = (u_int32_t)htonl(jumboplen);
889
890 #if 1
891 /*
892 * if there are multiple jumbo payload options,
893 * *plenp will be non-zero and the packet will be
894 * rejected.
895 * the behavior may need some debate in ipngwg -
896 * multiple options does not make sense, however,
897 * there's no explicit mention in specification.
898 */
899 if (*plenp != 0) {
900 ip6stat.ip6s_badoptions++;
901 icmp6_error(m, ICMP6_PARAM_PROB,
902 ICMP6_PARAMPROB_HEADER,
903 erroff + opt + 2 - opthead);
904 return (-1);
905 }
906 #endif
907
908 /*
909 * jumbo payload length must be larger than 65535.
910 */
911 if (jumboplen <= IPV6_MAXPACKET) {
912 ip6stat.ip6s_badoptions++;
913 icmp6_error(m, ICMP6_PARAM_PROB,
914 ICMP6_PARAMPROB_HEADER,
915 erroff + opt + 2 - opthead);
916 return (-1);
917 }
918 *plenp = jumboplen;
919
920 break;
921 default: /* unknown option */
922 if (hbhlen < IP6OPT_MINLEN) {
923 ip6stat.ip6s_toosmall++;
924 goto bad;
925 }
926 optlen = ip6_unknown_opt(opt, m,
927 erroff + opt - opthead);
928 if (optlen == -1)
929 return (-1);
930 optlen += 2;
931 break;
932 }
933 }
934
935 return (0);
936
937 bad:
938 m_freem(m);
939 return (-1);
940 }
941
942 /*
943 * Unknown option processing.
944 * The third argument `off' is the offset from the IPv6 header to the option,
945 * which is necessary if the IPv6 header the and option header and IPv6 header
946 * is not continuous in order to return an ICMPv6 error.
947 */
948 int
949 ip6_unknown_opt(optp, m, off)
950 u_int8_t *optp;
951 struct mbuf *m;
952 int off;
953 {
954 struct ip6_hdr *ip6;
955
956 switch (IP6OPT_TYPE(*optp)) {
957 case IP6OPT_TYPE_SKIP: /* ignore the option */
958 return ((int)*(optp + 1));
959 case IP6OPT_TYPE_DISCARD: /* silently discard */
960 m_freem(m);
961 return (-1);
962 case IP6OPT_TYPE_FORCEICMP: /* send ICMP even if multicasted */
963 ip6stat.ip6s_badoptions++;
964 icmp6_error(m, ICMP6_PARAM_PROB, ICMP6_PARAMPROB_OPTION, off);
965 return (-1);
966 case IP6OPT_TYPE_ICMP: /* send ICMP if not multicasted */
967 ip6stat.ip6s_badoptions++;
968 ip6 = mtod(m, struct ip6_hdr *);
969 if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) ||
970 (m->m_flags & (M_BCAST|M_MCAST)))
971 m_freem(m);
972 else
973 icmp6_error(m, ICMP6_PARAM_PROB,
974 ICMP6_PARAMPROB_OPTION, off);
975 return (-1);
976 }
977
978 m_freem(m); /* XXX: NOTREACHED */
979 return (-1);
980 }
981
982 /*
983 * Create the "control" list for this pcb.
984 *
985 * The routine will be called from upper layer handlers like tcp6_input().
986 * Thus the routine assumes that the caller (tcp6_input) have already
987 * called IP6_EXTHDR_CHECK() and all the extension headers are located in the
988 * very first mbuf on the mbuf chain.
989 * We may want to add some infinite loop prevention or sanity checks for safety.
990 * (This applies only when you are using KAME mbuf chain restriction, i.e.
991 * you are using IP6_EXTHDR_CHECK() not m_pulldown())
992 */
993 void
994 ip6_savecontrol(in6p, mp, ip6, m)
995 struct in6pcb *in6p;
996 struct mbuf **mp;
997 struct ip6_hdr *ip6;
998 struct mbuf *m;
999 {
1000 struct proc *p = curproc; /* XXX */
1001 int privileged;
1002
1003 privileged = 0;
1004 if (p && !suser(p->p_ucred, &p->p_acflag))
1005 privileged++;
1006
1007 #ifdef SO_TIMESTAMP
1008 if (in6p->in6p_socket->so_options & SO_TIMESTAMP) {
1009 struct timeval tv;
1010
1011 microtime(&tv);
1012 *mp = sbcreatecontrol((caddr_t) &tv, sizeof(tv),
1013 SCM_TIMESTAMP, SOL_SOCKET);
1014 if (*mp)
1015 mp = &(*mp)->m_next;
1016 }
1017 #endif
1018
1019 /* some OSes call this logic with IPv4 packet, for SO_TIMESTAMP */
1020 if ((ip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION)
1021 return;
1022
1023 if (in6p->in6p_flags & IN6P_RECVDSTADDR) {
1024 *mp = sbcreatecontrol((caddr_t) &ip6->ip6_dst,
1025 sizeof(struct in6_addr), IPV6_RECVDSTADDR, IPPROTO_IPV6);
1026 if (*mp)
1027 mp = &(*mp)->m_next;
1028 }
1029
1030 #ifdef noyet
1031 /* options were tossed above */
1032 if (in6p->in6p_flags & IN6P_RECVOPTS)
1033 /* broken */
1034 /* ip6_srcroute doesn't do what we want here, need to fix */
1035 if (in6p->in6p_flags & IPV6P_RECVRETOPTS)
1036 /* broken */
1037 #endif
1038
1039 /* RFC 2292 sec. 5 */
1040 if ((in6p->in6p_flags & IN6P_PKTINFO) != 0) {
1041 struct in6_pktinfo pi6;
1042 bcopy(&ip6->ip6_dst, &pi6.ipi6_addr, sizeof(struct in6_addr));
1043 if (IN6_IS_SCOPE_LINKLOCAL(&pi6.ipi6_addr))
1044 pi6.ipi6_addr.s6_addr16[1] = 0;
1045 pi6.ipi6_ifindex = (m && m->m_pkthdr.rcvif)
1046 ? m->m_pkthdr.rcvif->if_index
1047 : 0;
1048 *mp = sbcreatecontrol((caddr_t) &pi6,
1049 sizeof(struct in6_pktinfo), IPV6_PKTINFO, IPPROTO_IPV6);
1050 if (*mp)
1051 mp = &(*mp)->m_next;
1052 }
1053 if (in6p->in6p_flags & IN6P_HOPLIMIT) {
1054 int hlim = ip6->ip6_hlim & 0xff;
1055 *mp = sbcreatecontrol((caddr_t) &hlim, sizeof(int),
1056 IPV6_HOPLIMIT, IPPROTO_IPV6);
1057 if (*mp)
1058 mp = &(*mp)->m_next;
1059 }
1060 /* IN6P_NEXTHOP - for outgoing packet only */
1061
1062 /*
1063 * IPV6_HOPOPTS socket option. We require super-user privilege
1064 * for the option, but it might be too strict, since there might
1065 * be some hop-by-hop options which can be returned to normal user.
1066 * See RFC 2292 section 6.
1067 */
1068 if ((in6p->in6p_flags & IN6P_HOPOPTS) != 0 && privileged) {
1069 /*
1070 * Check if a hop-by-hop options header is contatined in the
1071 * received packet, and if so, store the options as ancillary
1072 * data. Note that a hop-by-hop options header must be
1073 * just after the IPv6 header, which fact is assured through
1074 * the IPv6 input processing.
1075 */
1076 struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *);
1077 if (ip6->ip6_nxt == IPPROTO_HOPOPTS) {
1078 struct ip6_hbh *hbh;
1079 int hbhlen;
1080 struct mbuf *ext;
1081
1082 ext = ip6_pullexthdr(m, sizeof(struct ip6_hdr),
1083 ip6->ip6_nxt);
1084 if (ext == NULL) {
1085 ip6stat.ip6s_tooshort++;
1086 return;
1087 }
1088 hbh = mtod(ext, struct ip6_hbh *);
1089 hbhlen = (hbh->ip6h_len + 1) << 3;
1090 if (hbhlen != ext->m_len) {
1091 m_freem(ext);
1092 ip6stat.ip6s_tooshort++;
1093 return;
1094 }
1095
1096 /*
1097 * XXX: We copy whole the header even if a jumbo
1098 * payload option is included, which option is to
1099 * be removed before returning in the RFC 2292.
1100 * But it's too painful operation...
1101 */
1102 *mp = sbcreatecontrol((caddr_t)hbh, hbhlen,
1103 IPV6_HOPOPTS, IPPROTO_IPV6);
1104 if (*mp)
1105 mp = &(*mp)->m_next;
1106 m_freem(ext);
1107 }
1108 }
1109
1110 /* IPV6_DSTOPTS and IPV6_RTHDR socket options */
1111 if (in6p->in6p_flags & (IN6P_DSTOPTS | IN6P_RTHDR)) {
1112 struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *);
1113 int nxt = ip6->ip6_nxt, off = sizeof(struct ip6_hdr);
1114
1115 /*
1116 * Search for destination options headers or routing
1117 * header(s) through the header chain, and stores each
1118 * header as ancillary data.
1119 * Note that the order of the headers remains in
1120 * the chain of ancillary data.
1121 */
1122 while (1) { /* is explicit loop prevention necessary? */
1123 struct ip6_ext *ip6e = NULL;
1124 int elen;
1125 struct mbuf *ext = NULL;
1126
1127 /*
1128 * if it is not an extension header, don't try to
1129 * pull it from the chain.
1130 */
1131 switch (nxt) {
1132 case IPPROTO_DSTOPTS:
1133 case IPPROTO_ROUTING:
1134 case IPPROTO_HOPOPTS:
1135 case IPPROTO_AH: /* is it possible? */
1136 break;
1137 default:
1138 goto loopend;
1139 }
1140
1141 ext = ip6_pullexthdr(m, off, nxt);
1142 if (ext == NULL) {
1143 ip6stat.ip6s_tooshort++;
1144 return;
1145 }
1146 ip6e = mtod(ext, struct ip6_ext *);
1147 if (nxt == IPPROTO_AH)
1148 elen = (ip6e->ip6e_len + 2) << 2;
1149 else
1150 elen = (ip6e->ip6e_len + 1) << 3;
1151 if (elen != ext->m_len) {
1152 m_freem(ext);
1153 ip6stat.ip6s_tooshort++;
1154 return;
1155 }
1156 KASSERT(IP6_HDR_ALIGNED_P(ip6e));
1157
1158 switch (nxt) {
1159 case IPPROTO_DSTOPTS:
1160 if (!in6p->in6p_flags & IN6P_DSTOPTS)
1161 break;
1162
1163 /*
1164 * We also require super-user privilege for
1165 * the option.
1166 * See the comments on IN6_HOPOPTS.
1167 */
1168 if (!privileged)
1169 break;
1170
1171 *mp = sbcreatecontrol((caddr_t)ip6e, elen,
1172 IPV6_DSTOPTS, IPPROTO_IPV6);
1173 if (*mp)
1174 mp = &(*mp)->m_next;
1175 break;
1176
1177 case IPPROTO_ROUTING:
1178 if (!in6p->in6p_flags & IN6P_RTHDR)
1179 break;
1180
1181 *mp = sbcreatecontrol((caddr_t)ip6e, elen,
1182 IPV6_RTHDR, IPPROTO_IPV6);
1183 if (*mp)
1184 mp = &(*mp)->m_next;
1185 break;
1186
1187 case IPPROTO_HOPOPTS:
1188 case IPPROTO_AH: /* is it possible? */
1189 break;
1190
1191 default:
1192 /*
1193 * other cases have been filtered in the above.
1194 * none will visit this case. here we supply
1195 * the code just in case (nxt overwritten or
1196 * other cases).
1197 */
1198 m_freem(ext);
1199 goto loopend;
1200
1201 }
1202
1203 /* proceed with the next header. */
1204 off += elen;
1205 nxt = ip6e->ip6e_nxt;
1206 ip6e = NULL;
1207 m_freem(ext);
1208 ext = NULL;
1209 }
1210 loopend:
1211 ;
1212 }
1213 }
1214
1215 /*
1216 * pull single extension header from mbuf chain. returns single mbuf that
1217 * contains the result, or NULL on error.
1218 */
1219 static struct mbuf *
1220 ip6_pullexthdr(m, off, nxt)
1221 struct mbuf *m;
1222 size_t off;
1223 int nxt;
1224 {
1225 struct ip6_ext ip6e;
1226 size_t elen;
1227 struct mbuf *n;
1228
1229 #ifdef DIAGNOSTIC
1230 switch (nxt) {
1231 case IPPROTO_DSTOPTS:
1232 case IPPROTO_ROUTING:
1233 case IPPROTO_HOPOPTS:
1234 case IPPROTO_AH: /* is it possible? */
1235 break;
1236 default:
1237 printf("ip6_pullexthdr: invalid nxt=%d\n", nxt);
1238 }
1239 #endif
1240
1241 m_copydata(m, off, sizeof(ip6e), (caddr_t)&ip6e);
1242 if (nxt == IPPROTO_AH)
1243 elen = (ip6e.ip6e_len + 2) << 2;
1244 else
1245 elen = (ip6e.ip6e_len + 1) << 3;
1246
1247 MGET(n, M_DONTWAIT, MT_DATA);
1248 if (n && elen >= MLEN) {
1249 MCLGET(n, M_DONTWAIT);
1250 if ((n->m_flags & M_EXT) == 0) {
1251 m_free(n);
1252 n = NULL;
1253 }
1254 }
1255 if (!n)
1256 return NULL;
1257
1258 n->m_len = 0;
1259 if (elen >= M_TRAILINGSPACE(n)) {
1260 m_free(n);
1261 return NULL;
1262 }
1263
1264 m_copydata(m, off, elen, mtod(n, caddr_t));
1265 n->m_len = elen;
1266 return n;
1267 }
1268
1269 /*
1270 * Get pointer to the previous header followed by the header
1271 * currently processed.
1272 * XXX: This function supposes that
1273 * M includes all headers,
1274 * the next header field and the header length field of each header
1275 * are valid, and
1276 * the sum of each header length equals to OFF.
1277 * Because of these assumptions, this function must be called very
1278 * carefully. Moreover, it will not be used in the near future when
1279 * we develop `neater' mechanism to process extension headers.
1280 */
1281 u_int8_t *
1282 ip6_get_prevhdr(m, off)
1283 struct mbuf *m;
1284 int off;
1285 {
1286 struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *);
1287
1288 if (off == sizeof(struct ip6_hdr))
1289 return (&ip6->ip6_nxt);
1290 else {
1291 int len, nxt;
1292 struct ip6_ext *ip6e = NULL;
1293
1294 nxt = ip6->ip6_nxt;
1295 len = sizeof(struct ip6_hdr);
1296 while (len < off) {
1297 ip6e = (struct ip6_ext *)(mtod(m, caddr_t) + len);
1298
1299 switch (nxt) {
1300 case IPPROTO_FRAGMENT:
1301 len += sizeof(struct ip6_frag);
1302 break;
1303 case IPPROTO_AH:
1304 len += (ip6e->ip6e_len + 2) << 2;
1305 break;
1306 default:
1307 len += (ip6e->ip6e_len + 1) << 3;
1308 break;
1309 }
1310 nxt = ip6e->ip6e_nxt;
1311 }
1312 if (ip6e)
1313 return (&ip6e->ip6e_nxt);
1314 else
1315 return NULL;
1316 }
1317 }
1318
1319 /*
1320 * get next header offset. m will be retained.
1321 */
1322 int
1323 ip6_nexthdr(m, off, proto, nxtp)
1324 struct mbuf *m;
1325 int off;
1326 int proto;
1327 int *nxtp;
1328 {
1329 struct ip6_hdr ip6;
1330 struct ip6_ext ip6e;
1331 struct ip6_frag fh;
1332
1333 /* just in case */
1334 if (m == NULL)
1335 panic("ip6_nexthdr: m == NULL");
1336 if ((m->m_flags & M_PKTHDR) == 0 || m->m_pkthdr.len < off)
1337 return -1;
1338
1339 switch (proto) {
1340 case IPPROTO_IPV6:
1341 if (m->m_pkthdr.len < off + sizeof(ip6))
1342 return -1;
1343 m_copydata(m, off, sizeof(ip6), (caddr_t)&ip6);
1344 if (nxtp)
1345 *nxtp = ip6.ip6_nxt;
1346 off += sizeof(ip6);
1347 return off;
1348
1349 case IPPROTO_FRAGMENT:
1350 /*
1351 * terminate parsing if it is not the first fragment,
1352 * it does not make sense to parse through it.
1353 */
1354 if (m->m_pkthdr.len < off + sizeof(fh))
1355 return -1;
1356 m_copydata(m, off, sizeof(fh), (caddr_t)&fh);
1357 if ((fh.ip6f_offlg & IP6F_OFF_MASK) != 0)
1358 return -1;
1359 if (nxtp)
1360 *nxtp = fh.ip6f_nxt;
1361 off += sizeof(struct ip6_frag);
1362 return off;
1363
1364 case IPPROTO_AH:
1365 if (m->m_pkthdr.len < off + sizeof(ip6e))
1366 return -1;
1367 m_copydata(m, off, sizeof(ip6e), (caddr_t)&ip6e);
1368 if (nxtp)
1369 *nxtp = ip6e.ip6e_nxt;
1370 off += (ip6e.ip6e_len + 2) << 2;
1371 if (m->m_pkthdr.len < off)
1372 return -1;
1373 return off;
1374
1375 case IPPROTO_HOPOPTS:
1376 case IPPROTO_ROUTING:
1377 case IPPROTO_DSTOPTS:
1378 if (m->m_pkthdr.len < off + sizeof(ip6e))
1379 return -1;
1380 m_copydata(m, off, sizeof(ip6e), (caddr_t)&ip6e);
1381 if (nxtp)
1382 *nxtp = ip6e.ip6e_nxt;
1383 off += (ip6e.ip6e_len + 1) << 3;
1384 if (m->m_pkthdr.len < off)
1385 return -1;
1386 return off;
1387
1388 case IPPROTO_NONE:
1389 case IPPROTO_ESP:
1390 case IPPROTO_IPCOMP:
1391 /* give up */
1392 return -1;
1393
1394 default:
1395 return -1;
1396 }
1397 }
1398
1399 /*
1400 * get offset for the last header in the chain. m will be kept untainted.
1401 */
1402 int
1403 ip6_lasthdr(m, off, proto, nxtp)
1404 struct mbuf *m;
1405 int off;
1406 int proto;
1407 int *nxtp;
1408 {
1409 int newoff;
1410 int nxt;
1411
1412 if (!nxtp) {
1413 nxt = -1;
1414 nxtp = &nxt;
1415 }
1416 while (1) {
1417 newoff = ip6_nexthdr(m, off, proto, nxtp);
1418 if (newoff < 0)
1419 return off;
1420 else if (newoff < off)
1421 return -1; /* invalid */
1422 else if (newoff == off)
1423 return newoff;
1424
1425 off = newoff;
1426 proto = *nxtp;
1427 }
1428 }
1429
1430 /*
1431 * System control for IP6
1432 */
1433
1434 u_char inet6ctlerrmap[PRC_NCMDS] = {
1435 0, 0, 0, 0,
1436 0, EMSGSIZE, EHOSTDOWN, EHOSTUNREACH,
1437 EHOSTUNREACH, EHOSTUNREACH, ECONNREFUSED, ECONNREFUSED,
1438 EMSGSIZE, EHOSTUNREACH, 0, 0,
1439 0, 0, 0, 0,
1440 ENOPROTOOPT
1441 };
1442
1443 static int
1444 sysctl_net_inet6_ip6_rht0(SYSCTLFN_ARGS)
1445 {
1446 int error, tmp;
1447 struct sysctlnode node;
1448
1449 node = *rnode;
1450 tmp = ip6_rht0;
1451 node.sysctl_data = &tmp;
1452 error = sysctl_lookup(SYSCTLFN_CALL(&node));
1453 if (error || newp == NULL)
1454 return error;
1455
1456 switch (tmp) {
1457 case -1: /* disable processing */
1458 case 0: /* disable for host, enable for router */
1459 case 1: /* enable for all */
1460 break;
1461 default:
1462 return EINVAL;
1463 }
1464 ip6_rht0 = tmp;
1465 return 0;
1466 }
1467
1468 SYSCTL_SETUP(sysctl_net_inet6_ip6_setup, "sysctl net.inet6.ip6 subtree setup")
1469 {
1470
1471 sysctl_createv(clog, 0, NULL, NULL,
1472 CTLFLAG_PERMANENT,
1473 CTLTYPE_NODE, "net", NULL,
1474 NULL, 0, NULL, 0,
1475 CTL_NET, CTL_EOL);
1476 sysctl_createv(clog, 0, NULL, NULL,
1477 CTLFLAG_PERMANENT,
1478 CTLTYPE_NODE, "inet6",
1479 SYSCTL_DESCR("PF_INET6 related settings"),
1480 NULL, 0, NULL, 0,
1481 CTL_NET, PF_INET6, CTL_EOL);
1482 sysctl_createv(clog, 0, NULL, NULL,
1483 CTLFLAG_PERMANENT,
1484 CTLTYPE_NODE, "ip6",
1485 SYSCTL_DESCR("IPv6 related settings"),
1486 NULL, 0, NULL, 0,
1487 CTL_NET, PF_INET6, IPPROTO_IPV6, CTL_EOL);
1488
1489 sysctl_createv(clog, 0, NULL, NULL,
1490 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1491 CTLTYPE_INT, "forwarding",
1492 SYSCTL_DESCR("Enable forwarding of INET6 datagrams"),
1493 NULL, 0, &ip6_forwarding, 0,
1494 CTL_NET, PF_INET6, IPPROTO_IPV6,
1495 IPV6CTL_FORWARDING, CTL_EOL);
1496 sysctl_createv(clog, 0, NULL, NULL,
1497 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1498 CTLTYPE_INT, "redirect",
1499 SYSCTL_DESCR("Enable sending of ICMPv6 redirect messages"),
1500 NULL, 0, &ip6_sendredirects, 0,
1501 CTL_NET, PF_INET6, IPPROTO_IPV6,
1502 IPV6CTL_SENDREDIRECTS, CTL_EOL);
1503 sysctl_createv(clog, 0, NULL, NULL,
1504 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1505 CTLTYPE_INT, "hlim",
1506 SYSCTL_DESCR("Hop limit for an INET6 datagram"),
1507 NULL, 0, &ip6_defhlim, 0,
1508 CTL_NET, PF_INET6, IPPROTO_IPV6,
1509 IPV6CTL_DEFHLIM, CTL_EOL);
1510 #ifdef notyet
1511 sysctl_createv(clog, 0, NULL, NULL,
1512 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1513 CTLTYPE_INT, "mtu", NULL,
1514 NULL, 0, &, 0,
1515 CTL_NET, PF_INET6, IPPROTO_IPV6,
1516 IPV6CTL_DEFMTU, CTL_EOL);
1517 #endif
1518 #ifdef __no_idea__
1519 sysctl_createv(clog, 0, NULL, NULL,
1520 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1521 CTLTYPE_INT, "forwsrcrt", NULL,
1522 NULL, 0, &?, 0,
1523 CTL_NET, PF_INET6, IPPROTO_IPV6,
1524 IPV6CTL_FORWSRCRT, CTL_EOL);
1525 sysctl_createv(clog, 0, NULL, NULL,
1526 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1527 CTLTYPE_STRUCT, "stats", NULL,
1528 NULL, 0, &?, sizeof(?),
1529 CTL_NET, PF_INET6, IPPROTO_IPV6,
1530 IPV6CTL_STATS, CTL_EOL);
1531 sysctl_createv(clog, 0, NULL, NULL,
1532 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1533 CTLTYPE_STRUCT, "mrtstats", NULL,
1534 NULL, 0, &?, sizeof(?),
1535 CTL_NET, PF_INET6, IPPROTO_IPV6,
1536 IPV6CTL_MRTSTATS, CTL_EOL);
1537 sysctl_createv(clog, 0, NULL, NULL,
1538 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1539 CTLTYPE_?, "mrtproto", NULL,
1540 NULL, 0, &?, sizeof(?),
1541 CTL_NET, PF_INET6, IPPROTO_IPV6,
1542 IPV6CTL_MRTPROTO, CTL_EOL);
1543 #endif
1544 sysctl_createv(clog, 0, NULL, NULL,
1545 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1546 CTLTYPE_INT, "maxfragpackets",
1547 SYSCTL_DESCR("Maximum number of fragments to buffer "
1548 "for reassembly"),
1549 NULL, 0, &ip6_maxfragpackets, 0,
1550 CTL_NET, PF_INET6, IPPROTO_IPV6,
1551 IPV6CTL_MAXFRAGPACKETS, CTL_EOL);
1552 #ifdef __no_idea__
1553 sysctl_createv(clog, 0, NULL, NULL,
1554 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1555 CTLTYPE_INT, "sourcecheck", NULL,
1556 NULL, 0, &?, 0,
1557 CTL_NET, PF_INET6, IPPROTO_IPV6,
1558 IPV6CTL_SOURCECHECK, CTL_EOL);
1559 sysctl_createv(clog, 0, NULL, NULL,
1560 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1561 CTLTYPE_INT, "sourcecheck_logint", NULL,
1562 NULL, 0, &?, 0,
1563 CTL_NET, PF_INET6, IPPROTO_IPV6,
1564 IPV6CTL_SOURCECHECK_LOGINT, CTL_EOL);
1565 #endif
1566 sysctl_createv(clog, 0, NULL, NULL,
1567 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1568 CTLTYPE_INT, "accept_rtadv",
1569 SYSCTL_DESCR("Accept router advertisements"),
1570 NULL, 0, &ip6_accept_rtadv, 0,
1571 CTL_NET, PF_INET6, IPPROTO_IPV6,
1572 IPV6CTL_ACCEPT_RTADV, CTL_EOL);
1573 sysctl_createv(clog, 0, NULL, NULL,
1574 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1575 CTLTYPE_INT, "keepfaith",
1576 SYSCTL_DESCR("Activate faith interface"),
1577 NULL, 0, &ip6_keepfaith, 0,
1578 CTL_NET, PF_INET6, IPPROTO_IPV6,
1579 IPV6CTL_KEEPFAITH, CTL_EOL);
1580 sysctl_createv(clog, 0, NULL, NULL,
1581 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1582 CTLTYPE_INT, "log_interval",
1583 SYSCTL_DESCR("Minumum interval between logging "
1584 "unroutable packets"),
1585 NULL, 0, &ip6_log_interval, 0,
1586 CTL_NET, PF_INET6, IPPROTO_IPV6,
1587 IPV6CTL_LOG_INTERVAL, CTL_EOL);
1588 sysctl_createv(clog, 0, NULL, NULL,
1589 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1590 CTLTYPE_INT, "hdrnestlimit",
1591 SYSCTL_DESCR("Maximum number of nested IPv6 headers"),
1592 NULL, 0, &ip6_hdrnestlimit, 0,
1593 CTL_NET, PF_INET6, IPPROTO_IPV6,
1594 IPV6CTL_HDRNESTLIMIT, CTL_EOL);
1595 sysctl_createv(clog, 0, NULL, NULL,
1596 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1597 CTLTYPE_INT, "dad_count",
1598 SYSCTL_DESCR("Number of Duplicate Address Detection "
1599 "probes to send"),
1600 NULL, 0, &ip6_dad_count, 0,
1601 CTL_NET, PF_INET6, IPPROTO_IPV6,
1602 IPV6CTL_DAD_COUNT, CTL_EOL);
1603 sysctl_createv(clog, 0, NULL, NULL,
1604 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1605 CTLTYPE_INT, "auto_flowlabel",
1606 SYSCTL_DESCR("Assign random IPv6 flow labels"),
1607 NULL, 0, &ip6_auto_flowlabel, 0,
1608 CTL_NET, PF_INET6, IPPROTO_IPV6,
1609 IPV6CTL_AUTO_FLOWLABEL, CTL_EOL);
1610 sysctl_createv(clog, 0, NULL, NULL,
1611 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1612 CTLTYPE_INT, "defmcasthlim",
1613 SYSCTL_DESCR("Default multicast hop limit"),
1614 NULL, 0, &ip6_defmcasthlim, 0,
1615 CTL_NET, PF_INET6, IPPROTO_IPV6,
1616 IPV6CTL_DEFMCASTHLIM, CTL_EOL);
1617 #if NGIF > 0
1618 sysctl_createv(clog, 0, NULL, NULL,
1619 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1620 CTLTYPE_INT, "gifhlim",
1621 SYSCTL_DESCR("Default hop limit for a gif tunnel datagram"),
1622 NULL, 0, &ip6_gif_hlim, 0,
1623 CTL_NET, PF_INET6, IPPROTO_IPV6,
1624 IPV6CTL_GIF_HLIM, CTL_EOL);
1625 #endif /* NGIF */
1626 sysctl_createv(clog, 0, NULL, NULL,
1627 CTLFLAG_PERMANENT,
1628 CTLTYPE_STRING, "kame_version",
1629 SYSCTL_DESCR("KAME Version"),
1630 NULL, 0, __KAME_VERSION, 0,
1631 CTL_NET, PF_INET6, IPPROTO_IPV6,
1632 IPV6CTL_KAME_VERSION, CTL_EOL);
1633 sysctl_createv(clog, 0, NULL, NULL,
1634 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1635 CTLTYPE_INT, "use_deprecated",
1636 SYSCTL_DESCR("Allow use of deprecated addresses as "
1637 "source addresses"),
1638 NULL, 0, &ip6_use_deprecated, 0,
1639 CTL_NET, PF_INET6, IPPROTO_IPV6,
1640 IPV6CTL_USE_DEPRECATED, CTL_EOL);
1641 sysctl_createv(clog, 0, NULL, NULL,
1642 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1643 CTLTYPE_INT, "rr_prune", NULL,
1644 NULL, 0, &ip6_rr_prune, 0,
1645 CTL_NET, PF_INET6, IPPROTO_IPV6,
1646 IPV6CTL_RR_PRUNE, CTL_EOL);
1647 sysctl_createv(clog, 0, NULL, NULL,
1648 CTLFLAG_PERMANENT
1649 #ifndef INET6_BINDV6ONLY
1650 |CTLFLAG_READWRITE,
1651 #endif
1652 CTLTYPE_INT, "v6only",
1653 SYSCTL_DESCR("Disallow PF_INET6 sockets from connecting "
1654 "to PF_INET sockets"),
1655 NULL, 0, &ip6_v6only, 0,
1656 CTL_NET, PF_INET6, IPPROTO_IPV6,
1657 IPV6CTL_V6ONLY, CTL_EOL);
1658 sysctl_createv(clog, 0, NULL, NULL,
1659 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1660 CTLTYPE_INT, "anonportmin",
1661 SYSCTL_DESCR("Lowest ephemeral port number to assign"),
1662 sysctl_net_inet_ip_ports, 0, &ip6_anonportmin, 0,
1663 CTL_NET, PF_INET6, IPPROTO_IPV6,
1664 IPV6CTL_ANONPORTMIN, CTL_EOL);
1665 sysctl_createv(clog, 0, NULL, NULL,
1666 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1667 CTLTYPE_INT, "anonportmax",
1668 SYSCTL_DESCR("Highest ephemeral port number to assign"),
1669 sysctl_net_inet_ip_ports, 0, &ip6_anonportmax, 0,
1670 CTL_NET, PF_INET6, IPPROTO_IPV6,
1671 IPV6CTL_ANONPORTMAX, CTL_EOL);
1672 #ifndef IPNOPRIVPORTS
1673 sysctl_createv(clog, 0, NULL, NULL,
1674 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1675 CTLTYPE_INT, "lowportmin",
1676 SYSCTL_DESCR("Lowest privileged ephemeral port number "
1677 "to assign"),
1678 sysctl_net_inet_ip_ports, 0, &ip6_lowportmin, 0,
1679 CTL_NET, PF_INET6, IPPROTO_IPV6,
1680 IPV6CTL_LOWPORTMIN, CTL_EOL);
1681 sysctl_createv(clog, 0, NULL, NULL,
1682 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1683 CTLTYPE_INT, "lowportmax",
1684 SYSCTL_DESCR("Highest privileged ephemeral port number "
1685 "to assign"),
1686 sysctl_net_inet_ip_ports, 0, &ip6_lowportmax, 0,
1687 CTL_NET, PF_INET6, IPPROTO_IPV6,
1688 IPV6CTL_LOWPORTMAX, CTL_EOL);
1689 #endif /* IPNOPRIVPORTS */
1690 sysctl_createv(clog, 0, NULL, NULL,
1691 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1692 CTLTYPE_INT, "maxfrags",
1693 SYSCTL_DESCR("Maximum fragments in reassembly queue"),
1694 NULL, 0, &ip6_maxfrags, 0,
1695 CTL_NET, PF_INET6, IPPROTO_IPV6,
1696 IPV6CTL_MAXFRAGS, CTL_EOL);
1697 sysctl_createv(clog, 0, NULL, NULL,
1698 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1699 CTLTYPE_INT, "rht0",
1700 SYSCTL_DESCR("Processing of routing header type 0 (IPv6)"),
1701 sysctl_net_inet6_ip6_rht0, 0, &ip6_rht0, 0,
1702 CTL_NET, PF_INET6, IPPROTO_IPV6,
1703 CTL_CREATE, CTL_EOL);
1704 }
Cache object: 988cdec893f65ba8a5b774df3d7862e7
|