1/*	$NetBSD: ip6_input.c,v 1.136 2012/01/10 20:01:56 drochner 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 1.136 2012/01/10 20:01:56 drochner Exp $");
66
67#include "opt_gateway.h"
68#include "opt_inet.h"
69#include "opt_inet6.h"
70#include "opt_ipsec.h"
71#include "opt_pfil_hooks.h"
72#include "opt_compat_netbsd.h"
73
74#include <sys/param.h>
75#include <sys/systm.h>
76#include <sys/malloc.h>
77#include <sys/mbuf.h>
78#include <sys/domain.h>
79#include <sys/protosw.h>
80#include <sys/socket.h>
81#include <sys/socketvar.h>
82#include <sys/errno.h>
83#include <sys/time.h>
84#include <sys/kernel.h>
85#include <sys/syslog.h>
86#include <sys/proc.h>
87#include <sys/sysctl.h>
88#include <sys/cprng.h>
89
90#include <net/if.h>
91#include <net/if_types.h>
92#include <net/if_dl.h>
93#include <net/route.h>
94#include <net/netisr.h>
95#ifdef PFIL_HOOKS
96#include <net/pfil.h>
97#endif
98
99#include <netinet/in.h>
100#include <netinet/in_systm.h>
101#ifdef INET
102#include <netinet/ip.h>
103#include <netinet/ip_icmp.h>
104#endif /* INET */
105#include <netinet/ip6.h>
106#include <netinet6/in6_var.h>
107#include <netinet6/ip6_var.h>
108#include <netinet6/ip6_private.h>
109#include <netinet6/in6_pcb.h>
110#include <netinet/icmp6.h>
111#include <netinet6/scope6_var.h>
112#include <netinet6/in6_ifattach.h>
113#include <netinet6/nd6.h>
114
115#ifdef KAME_IPSEC
116#include <netinet6/ipsec.h>
117#include <netinet6/ipsec_private.h>
118#endif
119
120#ifdef FAST_IPSEC
121#include <netipsec/ipsec.h>
122#include <netipsec/ipsec6.h>
123#include <netipsec/key.h>
124#endif /* FAST_IPSEC */
125
126#ifdef COMPAT_50
127#include <compat/sys/time.h>
128#include <compat/sys/socket.h>
129#endif
130
131#include <netinet6/ip6protosw.h>
132
133#include "faith.h"
134#include "gif.h"
135
136#if NGIF > 0
137#include <netinet6/in6_gif.h>
138#endif
139
140#include <net/net_osdep.h>
141
142extern struct domain inet6domain;
143
144u_char ip6_protox[IPPROTO_MAX];
145static int ip6qmaxlen = IFQ_MAXLEN;
146struct in6_ifaddr *in6_ifaddr;
147struct ifqueue ip6intrq;
148
149extern callout_t in6_tmpaddrtimer_ch;
150
151int ip6_forward_srcrt;			/* XXX */
152int ip6_sourcecheck;			/* XXX */
153int ip6_sourcecheck_interval;		/* XXX */
154
155#ifdef PFIL_HOOKS
156struct pfil_head inet6_pfil_hook;
157#endif
158
159percpu_t *ip6stat_percpu;
160
161static void ip6_init2(void *);
162static struct m_tag *ip6_setdstifaddr(struct mbuf *, const struct in6_ifaddr *);
163
164static int ip6_process_hopopts(struct mbuf *, u_int8_t *, int, u_int32_t *,
165	u_int32_t *);
166static struct mbuf *ip6_pullexthdr(struct mbuf *, size_t, int);
167static void sysctl_net_inet6_ip6_setup(struct sysctllog **);
168
169/*
170 * IP6 initialization: fill in IP6 protocol switch table.
171 * All protocols not implemented in kernel go to raw IP6 protocol handler.
172 */
173void
174ip6_init(void)
175{
176	const struct ip6protosw *pr;
177	int i;
178
179	sysctl_net_inet6_ip6_setup(NULL);
180	pr = (const struct ip6protosw *)pffindproto(PF_INET6, IPPROTO_RAW, SOCK_RAW);
181	if (pr == 0)
182		panic("ip6_init");
183	for (i = 0; i < IPPROTO_MAX; i++)
184		ip6_protox[i] = pr - inet6sw;
185	for (pr = (const struct ip6protosw *)inet6domain.dom_protosw;
186	    pr < (const struct ip6protosw *)inet6domain.dom_protoswNPROTOSW; pr++)
187		if (pr->pr_domain->dom_family == PF_INET6 &&
188		    pr->pr_protocol && pr->pr_protocol != IPPROTO_RAW)
189			ip6_protox[pr->pr_protocol] = pr - inet6sw;
190	ip6intrq.ifq_maxlen = ip6qmaxlen;
191	scope6_init();
192	addrsel_policy_init();
193	nd6_init();
194	frag6_init();
195	ip6_desync_factor = cprng_fast32() % MAX_TEMP_DESYNC_FACTOR;
196
197	ip6_init2(NULL);
198#ifdef GATEWAY
199	ip6flow_init(ip6_hashsize);
200#endif
201
202#ifdef PFIL_HOOKS
203	/* Register our Packet Filter hook. */
204	inet6_pfil_hook.ph_type = PFIL_TYPE_AF;
205	inet6_pfil_hook.ph_af   = AF_INET6;
206	i = pfil_head_register(&inet6_pfil_hook);
207	if (i != 0)
208		printf("ip6_init: WARNING: unable to register pfil hook, "
209		    "error %d\n", i);
210#endif /* PFIL_HOOKS */
211
212	ip6stat_percpu = percpu_alloc(sizeof(uint64_t) * IP6_NSTATS);
213}
214
215static void
216ip6_init2(void *dummy)
217{
218
219	/* nd6_timer_init */
220	callout_init(&nd6_timer_ch, CALLOUT_MPSAFE);
221	callout_reset(&nd6_timer_ch, hz, nd6_timer, NULL);
222
223	/* timer for regeneranation of temporary addresses randomize ID */
224	callout_init(&in6_tmpaddrtimer_ch, CALLOUT_MPSAFE);
225	callout_reset(&in6_tmpaddrtimer_ch,
226		      (ip6_temp_preferred_lifetime - ip6_desync_factor -
227		       ip6_temp_regen_advance) * hz,
228		      in6_tmpaddrtimer, NULL);
229}
230
231/*
232 * IP6 input interrupt handling. Just pass the packet to ip6_input.
233 */
234void
235ip6intr(void)
236{
237	int s;
238	struct mbuf *m;
239
240	mutex_enter(softnet_lock);
241	KERNEL_LOCK(1, NULL);
242	for (;;) {
243		s = splnet();
244		IF_DEQUEUE(&ip6intrq, m);
245		splx(s);
246		if (m == 0)
247			break;
248		/* drop the packet if IPv6 operation is disabled on the IF */
249		if ((ND_IFINFO(m->m_pkthdr.rcvif)->flags & ND6_IFF_IFDISABLED)) {
250			m_freem(m);
251			break;
252		}
253		ip6_input(m);
254	}
255	KERNEL_UNLOCK_ONE(NULL);
256	mutex_exit(softnet_lock);
257}
258
259extern struct	route ip6_forward_rt;
260
261void
262ip6_input(struct mbuf *m)
263{
264	struct ip6_hdr *ip6;
265	int hit, off = sizeof(struct ip6_hdr), nest;
266	u_int32_t plen;
267	u_int32_t rtalert = ~0;
268	int nxt, ours = 0, rh_present = 0;
269	struct ifnet *deliverifp = NULL;
270	int srcrt = 0;
271	const struct rtentry *rt;
272	union {
273		struct sockaddr		dst;
274		struct sockaddr_in6	dst6;
275	} u;
276#ifdef FAST_IPSEC
277	struct m_tag *mtag;
278	struct tdb_ident *tdbi;
279	struct secpolicy *sp;
280	int s, error;
281#endif
282
283#ifdef KAME_IPSEC
284	/*
285	 * should the inner packet be considered authentic?
286	 * see comment in ah4_input().
287	 */
288	m->m_flags &= ~M_AUTHIPHDR;
289	m->m_flags &= ~M_AUTHIPDGM;
290#endif
291
292	/*
293	 * make sure we don't have onion peering information into m_tag.
294	 */
295	ip6_delaux(m);
296
297	/*
298	 * mbuf statistics
299	 */
300	if (m->m_flags & M_EXT) {
301		if (m->m_next)
302			IP6_STATINC(IP6_STAT_MEXT2M);
303		else
304			IP6_STATINC(IP6_STAT_MEXT1);
305	} else {
306#define M2MMAX	32
307		if (m->m_next) {
308			if (m->m_flags & M_LOOP) {
309			/*XXX*/	IP6_STATINC(IP6_STAT_M2M + lo0ifp->if_index);
310			} else if (m->m_pkthdr.rcvif->if_index < M2MMAX) {
311				IP6_STATINC(IP6_STAT_M2M +
312					    m->m_pkthdr.rcvif->if_index);
313			} else
314				IP6_STATINC(IP6_STAT_M2M);
315		} else
316			IP6_STATINC(IP6_STAT_M1);
317#undef M2MMAX
318	}
319
320	in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_receive);
321	IP6_STATINC(IP6_STAT_TOTAL);
322
323	/*
324	 * If the IPv6 header is not aligned, slurp it up into a new
325	 * mbuf with space for link headers, in the event we forward
326	 * it.  Otherwise, if it is aligned, make sure the entire base
327	 * IPv6 header is in the first mbuf of the chain.
328	 */
329	if (IP6_HDR_ALIGNED_P(mtod(m, void *)) == 0) {
330		struct ifnet *inifp = m->m_pkthdr.rcvif;
331		if ((m = m_copyup(m, sizeof(struct ip6_hdr),
332				  (max_linkhdr + 3) & ~3)) == NULL) {
333			/* XXXJRT new stat, please */
334			IP6_STATINC(IP6_STAT_TOOSMALL);
335			in6_ifstat_inc(inifp, ifs6_in_hdrerr);
336			return;
337		}
338	} else if (__predict_false(m->m_len < sizeof(struct ip6_hdr))) {
339		struct ifnet *inifp = m->m_pkthdr.rcvif;
340		if ((m = m_pullup(m, sizeof(struct ip6_hdr))) == NULL) {
341			IP6_STATINC(IP6_STAT_TOOSMALL);
342			in6_ifstat_inc(inifp, ifs6_in_hdrerr);
343			return;
344		}
345	}
346
347	ip6 = mtod(m, struct ip6_hdr *);
348
349	if ((ip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION) {
350		IP6_STATINC(IP6_STAT_BADVERS);
351		in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_hdrerr);
352		goto bad;
353	}
354
355#if defined(KAME_IPSEC)
356	/* IPv6 fast forwarding is not compatible with IPsec. */
357	m->m_flags &= ~M_CANFASTFWD;
358#else
359	/*
360	 * Assume that we can create a fast-forward IP flow entry
361	 * based on this packet.
362	 */
363	m->m_flags |= M_CANFASTFWD;
364#endif
365
366#ifdef PFIL_HOOKS
367	/*
368	 * Run through list of hooks for input packets.  If there are any
369	 * filters which require that additional packets in the flow are
370	 * not fast-forwarded, they must clear the M_CANFASTFWD flag.
371	 * Note that filters must _never_ set this flag, as another filter
372	 * in the list may have previously cleared it.
373	 */
374	/*
375	 * let ipfilter look at packet on the wire,
376	 * not the decapsulated packet.
377	 */
378#ifdef KAME_IPSEC
379	if (!ipsec_getnhist(m))
380#elif defined(FAST_IPSEC)
381	if (!ipsec_indone(m))
382#else
383	if (1)
384#endif
385	{
386		struct in6_addr odst;
387
388		odst = ip6->ip6_dst;
389		if (pfil_run_hooks(&inet6_pfil_hook, &m, m->m_pkthdr.rcvif,
390				   PFIL_IN) != 0)
391			return;
392		if (m == NULL)
393			return;
394		ip6 = mtod(m, struct ip6_hdr *);
395		srcrt = !IN6_ARE_ADDR_EQUAL(&odst, &ip6->ip6_dst);
396	}
397#endif /* PFIL_HOOKS */
398
399	IP6_STATINC(IP6_STAT_NXTHIST + ip6->ip6_nxt);
400
401#ifdef ALTQ
402	if (altq_input != NULL && (*altq_input)(m, AF_INET6) == 0) {
403		/* packet is dropped by traffic conditioner */
404		return;
405	}
406#endif
407
408	/*
409	 * Check against address spoofing/corruption.
410	 */
411	if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_src) ||
412	    IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_dst)) {
413		/*
414		 * XXX: "badscope" is not very suitable for a multicast source.
415		 */
416		IP6_STATINC(IP6_STAT_BADSCOPE);
417		in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr);
418		goto bad;
419	}
420	/*
421	 * The following check is not documented in specs.  A malicious
422	 * party may be able to use IPv4 mapped addr to confuse tcp/udp stack
423	 * and bypass security checks (act as if it was from 127.0.0.1 by using
424	 * IPv6 src ::ffff:127.0.0.1).  Be cautious.
425	 *
426	 * This check chokes if we are in an SIIT cloud.  As none of BSDs
427	 * support IPv4-less kernel compilation, we cannot support SIIT
428	 * environment at all.  So, it makes more sense for us to reject any
429	 * malicious packets for non-SIIT environment, than try to do a
430	 * partial support for SIIT environment.
431	 */
432	if (IN6_IS_ADDR_V4MAPPED(&ip6->ip6_src) ||
433	    IN6_IS_ADDR_V4MAPPED(&ip6->ip6_dst)) {
434		IP6_STATINC(IP6_STAT_BADSCOPE);
435		in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr);
436		goto bad;
437	}
438#if 0
439	/*
440	 * Reject packets with IPv4 compatible addresses (auto tunnel).
441	 *
442	 * The code forbids auto tunnel relay case in RFC1933 (the check is
443	 * stronger than RFC1933).  We may want to re-enable it if mech-xx
444	 * is revised to forbid relaying case.
445	 */
446	if (IN6_IS_ADDR_V4COMPAT(&ip6->ip6_src) ||
447	    IN6_IS_ADDR_V4COMPAT(&ip6->ip6_dst)) {
448		IP6_STATINC(IP6_STAT_BADSCOPE);
449		in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr);
450		goto bad;
451	}
452#endif
453
454	/*
455	 * Disambiguate address scope zones (if there is ambiguity).
456	 * We first make sure that the original source or destination address
457	 * is not in our internal form for scoped addresses.  Such addresses
458	 * are not necessarily invalid spec-wise, but we cannot accept them due
459	 * to the usage conflict.
460	 * in6_setscope() then also checks and rejects the cases where src or
461	 * dst are the loopback address and the receiving interface
462	 * is not loopback.
463	 */
464	if (__predict_false(
465	    m_makewritable(&m, 0, sizeof(struct ip6_hdr), M_DONTWAIT)))
466		goto bad;
467	ip6 = mtod(m, struct ip6_hdr *);
468	if (in6_clearscope(&ip6->ip6_src) || in6_clearscope(&ip6->ip6_dst)) {
469		IP6_STATINC(IP6_STAT_BADSCOPE);	/* XXX */
470		goto bad;
471	}
472	if (in6_setscope(&ip6->ip6_src, m->m_pkthdr.rcvif, NULL) ||
473	    in6_setscope(&ip6->ip6_dst, m->m_pkthdr.rcvif, NULL)) {
474		IP6_STATINC(IP6_STAT_BADSCOPE);
475		goto bad;
476	}
477
478	/*
479	 * Multicast check
480	 */
481	if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) {
482	  	struct	in6_multi *in6m = 0;
483
484		in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_mcast);
485		/*
486		 * See if we belong to the destination multicast group on the
487		 * arrival interface.
488		 */
489		IN6_LOOKUP_MULTI(ip6->ip6_dst, m->m_pkthdr.rcvif, in6m);
490		if (in6m)
491			ours = 1;
492		else if (!ip6_mrouter) {
493			uint64_t *ip6s = IP6_STAT_GETREF();
494			ip6s[IP6_STAT_NOTMEMBER]++;
495			ip6s[IP6_STAT_CANTFORWARD]++;
496			IP6_STAT_PUTREF();
497			in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_discard);
498			goto bad;
499		}
500		deliverifp = m->m_pkthdr.rcvif;
501		goto hbhcheck;
502	}
503
504	sockaddr_in6_init(&u.dst6, &ip6->ip6_dst, 0, 0, 0);
505
506	/*
507	 *  Unicast check
508	 */
509	rt = rtcache_lookup2(&ip6_forward_rt, &u.dst, 1, &hit);
510	if (hit)
511		IP6_STATINC(IP6_STAT_FORWARD_CACHEHIT);
512	else
513		IP6_STATINC(IP6_STAT_FORWARD_CACHEMISS);
514
515#define rt6_getkey(__rt) satocsin6(rt_getkey(__rt))
516
517	/*
518	 * Accept the packet if the forwarding interface to the destination
519	 * according to the routing table is the loopback interface,
520	 * unless the associated route has a gateway.
521	 * Note that this approach causes to accept a packet if there is a
522	 * route to the loopback interface for the destination of the packet.
523	 * But we think it's even useful in some situations, e.g. when using
524	 * a special daemon which wants to intercept the packet.
525	 */
526	if (rt != NULL &&
527	    (rt->rt_flags & (RTF_HOST|RTF_GATEWAY)) == RTF_HOST &&
528	    !(rt->rt_flags & RTF_CLONED) &&
529#if 0
530	    /*
531	     * The check below is redundant since the comparison of
532	     * the destination and the key of the rtentry has
533	     * already done through looking up the routing table.
534	     */
535	    IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst, &rt6_getkey(rt)->sin6_addr) &&
536#endif
537	    rt->rt_ifp->if_type == IFT_LOOP) {
538		struct in6_ifaddr *ia6 = (struct in6_ifaddr *)rt->rt_ifa;
539		if (ia6->ia6_flags & IN6_IFF_ANYCAST)
540			m->m_flags |= M_ANYCAST6;
541		/*
542		 * packets to a tentative, duplicated, or somehow invalid
543		 * address must not be accepted.
544		 */
545		if (!(ia6->ia6_flags & IN6_IFF_NOTREADY)) {
546			/* this address is ready */
547			ours = 1;
548			deliverifp = ia6->ia_ifp;	/* correct? */
549			goto hbhcheck;
550		} else {
551			/* address is not ready, so discard the packet. */
552			nd6log((LOG_INFO,
553			    "ip6_input: packet to an unready address %s->%s\n",
554			    ip6_sprintf(&ip6->ip6_src),
555			    ip6_sprintf(&ip6->ip6_dst)));
556
557			goto bad;
558		}
559	}
560
561	/*
562	 * FAITH (Firewall Aided Internet Translator)
563	 */
564#if defined(NFAITH) && 0 < NFAITH
565	if (ip6_keepfaith) {
566		if (rt != NULL && rt->rt_ifp != NULL &&
567		    rt->rt_ifp->if_type == IFT_FAITH) {
568			/* XXX do we need more sanity checks? */
569			ours = 1;
570			deliverifp = rt->rt_ifp; /* faith */
571			goto hbhcheck;
572		}
573	}
574#endif
575
576#if 0
577    {
578	/*
579	 * Last resort: check in6_ifaddr for incoming interface.
580	 * The code is here until I update the "goto ours hack" code above
581	 * working right.
582	 */
583	struct ifaddr *ifa;
584	IFADDR_FOREACH(ifa, m->m_pkthdr.rcvif) {
585		if (ifa->ifa_addr == NULL)
586			continue;	/* just for safety */
587		if (ifa->ifa_addr->sa_family != AF_INET6)
588			continue;
589		if (IN6_ARE_ADDR_EQUAL(IFA_IN6(ifa), &ip6->ip6_dst)) {
590			ours = 1;
591			deliverifp = ifa->ifa_ifp;
592			goto hbhcheck;
593		}
594	}
595    }
596#endif
597
598	/*
599	 * Now there is no reason to process the packet if it's not our own
600	 * and we're not a router.
601	 */
602	if (!ip6_forwarding) {
603		IP6_STATINC(IP6_STAT_CANTFORWARD);
604		in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_discard);
605		goto bad;
606	}
607
608  hbhcheck:
609	/*
610	 * record address information into m_tag, if we don't have one yet.
611	 * note that we are unable to record it, if the address is not listed
612	 * as our interface address (e.g. multicast addresses, addresses
613	 * within FAITH prefixes and such).
614	 */
615	if (deliverifp && ip6_getdstifaddr(m) == NULL) {
616		struct in6_ifaddr *ia6;
617
618		ia6 = in6_ifawithifp(deliverifp, &ip6->ip6_dst);
619		if (ia6 != NULL && ip6_setdstifaddr(m, ia6) == NULL) {
620			/*
621			 * XXX maybe we should drop the packet here,
622			 * as we could not provide enough information
623			 * to the upper layers.
624			 */
625		}
626	}
627
628	/*
629	 * Process Hop-by-Hop options header if it's contained.
630	 * m may be modified in ip6_hopopts_input().
631	 * If a JumboPayload option is included, plen will also be modified.
632	 */
633	plen = (u_int32_t)ntohs(ip6->ip6_plen);
634	if (ip6->ip6_nxt == IPPROTO_HOPOPTS) {
635		struct ip6_hbh *hbh;
636
637		if (ip6_hopopts_input(&plen, &rtalert, &m, &off)) {
638#if 0	/*touches NULL pointer*/
639			in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_discard);
640#endif
641			return;	/* m have already been freed */
642		}
643
644		/* adjust pointer */
645		ip6 = mtod(m, struct ip6_hdr *);
646
647		/*
648		 * if the payload length field is 0 and the next header field
649		 * indicates Hop-by-Hop Options header, then a Jumbo Payload
650		 * option MUST be included.
651		 */
652		if (ip6->ip6_plen == 0 && plen == 0) {
653			/*
654			 * Note that if a valid jumbo payload option is
655			 * contained, ip6_hopopts_input() must set a valid
656			 * (non-zero) payload length to the variable plen.
657			 */
658			IP6_STATINC(IP6_STAT_BADOPTIONS);
659			in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_discard);
660			in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_hdrerr);
661			icmp6_error(m, ICMP6_PARAM_PROB,
662				    ICMP6_PARAMPROB_HEADER,
663				    (char *)&ip6->ip6_plen - (char *)ip6);
664			return;
665		}
666		IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m, sizeof(struct ip6_hdr),
667			sizeof(struct ip6_hbh));
668		if (hbh == NULL) {
669			IP6_STATINC(IP6_STAT_TOOSHORT);
670			return;
671		}
672		KASSERT(IP6_HDR_ALIGNED_P(hbh));
673		nxt = hbh->ip6h_nxt;
674
675		/*
676		 * accept the packet if a router alert option is included
677		 * and we act as an IPv6 router.
678		 */
679		if (rtalert != ~0 && ip6_forwarding)
680			ours = 1;
681	} else
682		nxt = ip6->ip6_nxt;
683
684	/*
685	 * Check that the amount of data in the buffers
686	 * is as at least much as the IPv6 header would have us expect.
687	 * Trim mbufs if longer than we expect.
688	 * Drop packet if shorter than we expect.
689	 */
690	if (m->m_pkthdr.len - sizeof(struct ip6_hdr) < plen) {
691		IP6_STATINC(IP6_STAT_TOOSHORT);
692		in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_truncated);
693		goto bad;
694	}
695	if (m->m_pkthdr.len > sizeof(struct ip6_hdr) + plen) {
696		if (m->m_len == m->m_pkthdr.len) {
697			m->m_len = sizeof(struct ip6_hdr) + plen;
698			m->m_pkthdr.len = sizeof(struct ip6_hdr) + plen;
699		} else
700			m_adj(m, sizeof(struct ip6_hdr) + plen - m->m_pkthdr.len);
701	}
702
703	/*
704	 * Forward if desirable.
705	 */
706	if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) {
707		/*
708		 * If we are acting as a multicast router, all
709		 * incoming multicast packets are passed to the
710		 * kernel-level multicast forwarding function.
711		 * The packet is returned (relatively) intact; if
712		 * ip6_mforward() returns a non-zero value, the packet
713		 * must be discarded, else it may be accepted below.
714		 */
715		if (ip6_mrouter && ip6_mforward(ip6, m->m_pkthdr.rcvif, m)) {
716			IP6_STATINC(IP6_STAT_CANTFORWARD);
717			m_freem(m);
718			return;
719		}
720		if (!ours) {
721			m_freem(m);
722			return;
723		}
724	} else if (!ours) {
725		ip6_forward(m, srcrt);
726		return;
727	}
728
729	ip6 = mtod(m, struct ip6_hdr *);
730
731	/*
732	 * Malicious party may be able to use IPv4 mapped addr to confuse
733	 * tcp/udp stack and bypass security checks (act as if it was from
734	 * 127.0.0.1 by using IPv6 src ::ffff:127.0.0.1).  Be cautious.
735	 *
736	 * For SIIT end node behavior, you may want to disable the check.
737	 * However, you will  become vulnerable to attacks using IPv4 mapped
738	 * source.
739	 */
740	if (IN6_IS_ADDR_V4MAPPED(&ip6->ip6_src) ||
741	    IN6_IS_ADDR_V4MAPPED(&ip6->ip6_dst)) {
742		IP6_STATINC(IP6_STAT_BADSCOPE);
743		in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr);
744		goto bad;
745	}
746
747	/*
748	 * Tell launch routine the next header
749	 */
750#ifdef IFA_STATS
751	if (deliverifp != NULL) {
752		struct in6_ifaddr *ia6;
753		ia6 = in6_ifawithifp(deliverifp, &ip6->ip6_dst);
754		if (ia6)
755			ia6->ia_ifa.ifa_data.ifad_inbytes += m->m_pkthdr.len;
756	}
757#endif
758	IP6_STATINC(IP6_STAT_DELIVERED);
759	in6_ifstat_inc(deliverifp, ifs6_in_deliver);
760	nest = 0;
761
762	rh_present = 0;
763	while (nxt != IPPROTO_DONE) {
764		if (ip6_hdrnestlimit && (++nest > ip6_hdrnestlimit)) {
765			IP6_STATINC(IP6_STAT_TOOMANYHDR);
766			in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_hdrerr);
767			goto bad;
768		}
769
770		/*
771		 * protection against faulty packet - there should be
772		 * more sanity checks in header chain processing.
773		 */
774		if (m->m_pkthdr.len < off) {
775			IP6_STATINC(IP6_STAT_TOOSHORT);
776			in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_truncated);
777			goto bad;
778		}
779
780		if (nxt == IPPROTO_ROUTING) {
781			if (rh_present++) {
782				in6_ifstat_inc(m->m_pkthdr.rcvif,
783				    ifs6_in_hdrerr);
784				IP6_STATINC(IP6_STAT_BADOPTIONS);
785				goto bad;
786			}
787		}
788
789#ifdef KAME_IPSEC
790		/*
791		 * enforce IPsec policy checking if we are seeing last header.
792		 * note that we do not visit this with protocols with pcb layer
793		 * code - like udp/tcp/raw ip.
794		 */
795		if ((inet6sw[ip6_protox[nxt]].pr_flags & PR_LASTHDR) != 0 &&
796		    ipsec6_in_reject(m, NULL)) {
797			IPSEC6_STATINC(IPSEC_STAT_IN_POLVIO);
798			goto bad;
799		}
800#endif
801#ifdef FAST_IPSEC
802	/*
803	 * enforce IPsec policy checking if we are seeing last header.
804	 * note that we do not visit this with protocols with pcb layer
805	 * code - like udp/tcp/raw ip.
806	 */
807	if ((inet6sw[ip_protox[nxt]].pr_flags & PR_LASTHDR) != 0) {
808		/*
809		 * Check if the packet has already had IPsec processing
810		 * done.  If so, then just pass it along.  This tag gets
811		 * set during AH, ESP, etc. input handling, before the
812		 * packet is returned to the ip input queue for delivery.
813		 */
814		mtag = m_tag_find(m, PACKET_TAG_IPSEC_IN_DONE, NULL);
815		s = splsoftnet();
816		if (mtag != NULL) {
817			tdbi = (struct tdb_ident *)(mtag + 1);
818			sp = ipsec_getpolicy(tdbi, IPSEC_DIR_INBOUND);
819		} else {
820			sp = ipsec_getpolicybyaddr(m, IPSEC_DIR_INBOUND,
821									IP_FORWARDING, &error);
822		}
823		if (sp != NULL) {
824			/*
825			 * Check security policy against packet attributes.
826			 */
827			error = ipsec_in_reject(sp, m);
828			KEY_FREESP(&sp);
829		} else {
830			/* XXX error stat??? */
831			error = EINVAL;
832			DPRINTF(("ip6_input: no SP, packet discarded\n"));/*XXX*/
833		}
834		splx(s);
835		if (error)
836			goto bad;
837	}
838#endif /* FAST_IPSEC */
839
840
841		nxt = (*inet6sw[ip6_protox[nxt]].pr_input)(&m, &off, nxt);
842	}
843	return;
844 bad:
845	m_freem(m);
846}
847
848/*
849 * set/grab in6_ifaddr correspond to IPv6 destination address.
850 */
851static struct m_tag *
852ip6_setdstifaddr(struct mbuf *m, const struct in6_ifaddr *ia)
853{
854	struct m_tag *mtag;
855
856	mtag = ip6_addaux(m);
857	if (mtag != NULL) {
858		struct ip6aux *ip6a;
859
860		ip6a = (struct ip6aux *)(mtag + 1);
861		in6_setscope(&ip6a->ip6a_src, ia->ia_ifp, &ip6a->ip6a_scope_id);
862		ip6a->ip6a_src = ia->ia_addr.sin6_addr;
863		ip6a->ip6a_flags = ia->ia6_flags;
864	}
865	return mtag;	/* NULL if failed to set */
866}
867
868const struct ip6aux *
869ip6_getdstifaddr(struct mbuf *m)
870{
871	struct m_tag *mtag;
872
873	mtag = ip6_findaux(m);
874	if (mtag != NULL)
875		return (struct ip6aux *)(mtag + 1);
876	else
877		return NULL;
878}
879
880/*
881 * Hop-by-Hop options header processing. If a valid jumbo payload option is
882 * included, the real payload length will be stored in plenp.
883 *
884 * rtalertp - XXX: should be stored more smart way
885 */
886int
887ip6_hopopts_input(u_int32_t *plenp, u_int32_t *rtalertp,
888	struct mbuf **mp, int *offp)
889{
890	struct mbuf *m = *mp;
891	int off = *offp, hbhlen;
892	struct ip6_hbh *hbh;
893
894	/* validation of the length of the header */
895	IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m,
896		sizeof(struct ip6_hdr), sizeof(struct ip6_hbh));
897	if (hbh == NULL) {
898		IP6_STATINC(IP6_STAT_TOOSHORT);
899		return -1;
900	}
901	hbhlen = (hbh->ip6h_len + 1) << 3;
902	IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m, sizeof(struct ip6_hdr),
903		hbhlen);
904	if (hbh == NULL) {
905		IP6_STATINC(IP6_STAT_TOOSHORT);
906		return -1;
907	}
908	KASSERT(IP6_HDR_ALIGNED_P(hbh));
909	off += hbhlen;
910	hbhlen -= sizeof(struct ip6_hbh);
911
912	if (ip6_process_hopopts(m, (u_int8_t *)hbh + sizeof(struct ip6_hbh),
913				hbhlen, rtalertp, plenp) < 0)
914		return (-1);
915
916	*offp = off;
917	*mp = m;
918	return (0);
919}
920
921/*
922 * Search header for all Hop-by-hop options and process each option.
923 * This function is separate from ip6_hopopts_input() in order to
924 * handle a case where the sending node itself process its hop-by-hop
925 * options header. In such a case, the function is called from ip6_output().
926 *
927 * The function assumes that hbh header is located right after the IPv6 header
928 * (RFC2460 p7), opthead is pointer into data content in m, and opthead to
929 * opthead + hbhlen is located in continuous memory region.
930 */
931static int
932ip6_process_hopopts(struct mbuf *m, u_int8_t *opthead, int hbhlen,
933	u_int32_t *rtalertp, u_int32_t *plenp)
934{
935	struct ip6_hdr *ip6;
936	int optlen = 0;
937	u_int8_t *opt = opthead;
938	u_int16_t rtalert_val;
939	u_int32_t jumboplen;
940	const int erroff = sizeof(struct ip6_hdr) + sizeof(struct ip6_hbh);
941
942	for (; hbhlen > 0; hbhlen -= optlen, opt += optlen) {
943		switch (*opt) {
944		case IP6OPT_PAD1:
945			optlen = 1;
946			break;
947		case IP6OPT_PADN:
948			if (hbhlen < IP6OPT_MINLEN) {
949				IP6_STATINC(IP6_STAT_TOOSMALL);
950				goto bad;
951			}
952			optlen = *(opt + 1) + 2;
953			break;
954		case IP6OPT_RTALERT:
955			/* XXX may need check for alignment */
956			if (hbhlen < IP6OPT_RTALERT_LEN) {
957				IP6_STATINC(IP6_STAT_TOOSMALL);
958				goto bad;
959			}
960			if (*(opt + 1) != IP6OPT_RTALERT_LEN - 2) {
961				/* XXX stat */
962				icmp6_error(m, ICMP6_PARAM_PROB,
963				    ICMP6_PARAMPROB_HEADER,
964				    erroff + opt + 1 - opthead);
965				return (-1);
966			}
967			optlen = IP6OPT_RTALERT_LEN;
968			memcpy((void *)&rtalert_val, (void *)(opt + 2), 2);
969			*rtalertp = ntohs(rtalert_val);
970			break;
971		case IP6OPT_JUMBO:
972			/* XXX may need check for alignment */
973			if (hbhlen < IP6OPT_JUMBO_LEN) {
974				IP6_STATINC(IP6_STAT_TOOSMALL);
975				goto bad;
976			}
977			if (*(opt + 1) != IP6OPT_JUMBO_LEN - 2) {
978				/* XXX stat */
979				icmp6_error(m, ICMP6_PARAM_PROB,
980				    ICMP6_PARAMPROB_HEADER,
981				    erroff + opt + 1 - opthead);
982				return (-1);
983			}
984			optlen = IP6OPT_JUMBO_LEN;
985
986			/*
987			 * IPv6 packets that have non 0 payload length
988			 * must not contain a jumbo payload option.
989			 */
990			ip6 = mtod(m, struct ip6_hdr *);
991			if (ip6->ip6_plen) {
992				IP6_STATINC(IP6_STAT_BADOPTIONS);
993				icmp6_error(m, ICMP6_PARAM_PROB,
994				    ICMP6_PARAMPROB_HEADER,
995				    erroff + opt - opthead);
996				return (-1);
997			}
998
999			/*
1000			 * We may see jumbolen in unaligned location, so
1001			 * we'd need to perform bcopy().
1002			 */
1003			memcpy(&jumboplen, opt + 2, sizeof(jumboplen));
1004			jumboplen = (u_int32_t)htonl(jumboplen);
1005
1006#if 1
1007			/*
1008			 * if there are multiple jumbo payload options,
1009			 * *plenp will be non-zero and the packet will be
1010			 * rejected.
1011			 * the behavior may need some debate in ipngwg -
1012			 * multiple options does not make sense, however,
1013			 * there's no explicit mention in specification.
1014			 */
1015			if (*plenp != 0) {
1016				IP6_STATINC(IP6_STAT_BADOPTIONS);
1017				icmp6_error(m, ICMP6_PARAM_PROB,
1018				    ICMP6_PARAMPROB_HEADER,
1019				    erroff + opt + 2 - opthead);
1020				return (-1);
1021			}
1022#endif
1023
1024			/*
1025			 * jumbo payload length must be larger than 65535.
1026			 */
1027			if (jumboplen <= IPV6_MAXPACKET) {
1028				IP6_STATINC(IP6_STAT_BADOPTIONS);
1029				icmp6_error(m, ICMP6_PARAM_PROB,
1030				    ICMP6_PARAMPROB_HEADER,
1031				    erroff + opt + 2 - opthead);
1032				return (-1);
1033			}
1034			*plenp = jumboplen;
1035
1036			break;
1037		default:		/* unknown option */
1038			if (hbhlen < IP6OPT_MINLEN) {
1039				IP6_STATINC(IP6_STAT_TOOSMALL);
1040				goto bad;
1041			}
1042			optlen = ip6_unknown_opt(opt, m,
1043			    erroff + opt - opthead);
1044			if (optlen == -1)
1045				return (-1);
1046			optlen += 2;
1047			break;
1048		}
1049	}
1050
1051	return (0);
1052
1053  bad:
1054	m_freem(m);
1055	return (-1);
1056}
1057
1058/*
1059 * Unknown option processing.
1060 * The third argument `off' is the offset from the IPv6 header to the option,
1061 * which is necessary if the IPv6 header the and option header and IPv6 header
1062 * is not continuous in order to return an ICMPv6 error.
1063 */
1064int
1065ip6_unknown_opt(u_int8_t *optp, struct mbuf *m, int off)
1066{
1067	struct ip6_hdr *ip6;
1068
1069	switch (IP6OPT_TYPE(*optp)) {
1070	case IP6OPT_TYPE_SKIP: /* ignore the option */
1071		return ((int)*(optp + 1));
1072	case IP6OPT_TYPE_DISCARD:	/* silently discard */
1073		m_freem(m);
1074		return (-1);
1075	case IP6OPT_TYPE_FORCEICMP: /* send ICMP even if multicasted */
1076		IP6_STATINC(IP6_STAT_BADOPTIONS);
1077		icmp6_error(m, ICMP6_PARAM_PROB, ICMP6_PARAMPROB_OPTION, off);
1078		return (-1);
1079	case IP6OPT_TYPE_ICMP: /* send ICMP if not multicasted */
1080		IP6_STATINC(IP6_STAT_BADOPTIONS);
1081		ip6 = mtod(m, struct ip6_hdr *);
1082		if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) ||
1083		    (m->m_flags & (M_BCAST|M_MCAST)))
1084			m_freem(m);
1085		else
1086			icmp6_error(m, ICMP6_PARAM_PROB,
1087				    ICMP6_PARAMPROB_OPTION, off);
1088		return (-1);
1089	}
1090
1091	m_freem(m);		/* XXX: NOTREACHED */
1092	return (-1);
1093}
1094
1095/*
1096 * Create the "control" list for this pcb.
1097 *
1098 * The routine will be called from upper layer handlers like tcp6_input().
1099 * Thus the routine assumes that the caller (tcp6_input) have already
1100 * called IP6_EXTHDR_CHECK() and all the extension headers are located in the
1101 * very first mbuf on the mbuf chain.
1102 * We may want to add some infinite loop prevention or sanity checks for safety.
1103 * (This applies only when you are using KAME mbuf chain restriction, i.e.
1104 * you are using IP6_EXTHDR_CHECK() not m_pulldown())
1105 */
1106void
1107ip6_savecontrol(struct in6pcb *in6p, struct mbuf **mp,
1108	struct ip6_hdr *ip6, struct mbuf *m)
1109{
1110#ifdef RFC2292
1111#define IS2292(x, y)	((in6p->in6p_flags & IN6P_RFC2292) ? (x) : (y))
1112#else
1113#define IS2292(x, y)	(y)
1114#endif
1115
1116	if (in6p->in6p_socket->so_options & SO_TIMESTAMP
1117#ifdef SO_OTIMESTAMP
1118	    || in6p->in6p_socket->so_options & SO_OTIMESTAMP
1119#endif
1120	) {
1121		struct timeval tv;
1122
1123		microtime(&tv);
1124#ifdef SO_OTIMESTAMP
1125		if (in6p->in6p_socket->so_options & SO_OTIMESTAMP) {
1126			struct timeval50 tv50;
1127			timeval_to_timeval50(&tv, &tv50);
1128			*mp = sbcreatecontrol((void *) &tv50, sizeof(tv50),
1129			    SCM_OTIMESTAMP, SOL_SOCKET);
1130		} else
1131#endif
1132		*mp = sbcreatecontrol((void *) &tv, sizeof(tv),
1133		    SCM_TIMESTAMP, SOL_SOCKET);
1134		if (*mp)
1135			mp = &(*mp)->m_next;
1136	}
1137
1138	/* some OSes call this logic with IPv4 packet, for SO_TIMESTAMP */
1139	if ((ip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION)
1140		return;
1141
1142	/* RFC 2292 sec. 5 */
1143	if ((in6p->in6p_flags & IN6P_PKTINFO) != 0) {
1144		struct in6_pktinfo pi6;
1145
1146		memcpy(&pi6.ipi6_addr, &ip6->ip6_dst, sizeof(struct in6_addr));
1147		in6_clearscope(&pi6.ipi6_addr);	/* XXX */
1148		pi6.ipi6_ifindex = m->m_pkthdr.rcvif ?
1149		    m->m_pkthdr.rcvif->if_index : 0;
1150		*mp = sbcreatecontrol((void *) &pi6,
1151		    sizeof(struct in6_pktinfo),
1152		    IS2292(IPV6_2292PKTINFO, IPV6_PKTINFO), IPPROTO_IPV6);
1153		if (*mp)
1154			mp = &(*mp)->m_next;
1155	}
1156
1157	if (in6p->in6p_flags & IN6P_HOPLIMIT) {
1158		int hlim = ip6->ip6_hlim & 0xff;
1159
1160		*mp = sbcreatecontrol((void *) &hlim, sizeof(int),
1161		    IS2292(IPV6_2292HOPLIMIT, IPV6_HOPLIMIT), IPPROTO_IPV6);
1162		if (*mp)
1163			mp = &(*mp)->m_next;
1164	}
1165
1166	if ((in6p->in6p_flags & IN6P_TCLASS) != 0) {
1167		u_int32_t flowinfo;
1168		int tclass;
1169
1170		flowinfo = (u_int32_t)ntohl(ip6->ip6_flow & IPV6_FLOWINFO_MASK);
1171		flowinfo >>= 20;
1172
1173		tclass = flowinfo & 0xff;
1174		*mp = sbcreatecontrol((void *)&tclass, sizeof(tclass),
1175		    IPV6_TCLASS, IPPROTO_IPV6);
1176
1177		if (*mp)
1178			mp = &(*mp)->m_next;
1179	}
1180
1181	/*
1182	 * IPV6_HOPOPTS socket option.  Recall that we required super-user
1183	 * privilege for the option (see ip6_ctloutput), but it might be too
1184	 * strict, since there might be some hop-by-hop options which can be
1185	 * returned to normal user.
1186	 * See also RFC3542 section 8 (or RFC2292 section 6).
1187	 */
1188	if ((in6p->in6p_flags & IN6P_HOPOPTS) != 0) {
1189		/*
1190		 * Check if a hop-by-hop options header is contatined in the
1191		 * received packet, and if so, store the options as ancillary
1192		 * data. Note that a hop-by-hop options header must be
1193		 * just after the IPv6 header, which fact is assured through
1194		 * the IPv6 input processing.
1195		 */
1196		struct ip6_hdr *xip6 = mtod(m, struct ip6_hdr *);
1197		if (xip6->ip6_nxt == IPPROTO_HOPOPTS) {
1198			struct ip6_hbh *hbh;
1199			int hbhlen;
1200			struct mbuf *ext;
1201
1202			ext = ip6_pullexthdr(m, sizeof(struct ip6_hdr),
1203			    xip6->ip6_nxt);
1204			if (ext == NULL) {
1205				IP6_STATINC(IP6_STAT_TOOSHORT);
1206				return;
1207			}
1208			hbh = mtod(ext, struct ip6_hbh *);
1209			hbhlen = (hbh->ip6h_len + 1) << 3;
1210			if (hbhlen != ext->m_len) {
1211				m_freem(ext);
1212				IP6_STATINC(IP6_STAT_TOOSHORT);
1213				return;
1214			}
1215
1216			/*
1217			 * XXX: We copy whole the header even if a jumbo
1218			 * payload option is included, which option is to
1219			 * be removed before returning in the RFC 2292.
1220			 * Note: this constraint is removed in RFC3542.
1221			 */
1222			*mp = sbcreatecontrol((void *)hbh, hbhlen,
1223			    IS2292(IPV6_2292HOPOPTS, IPV6_HOPOPTS),
1224			    IPPROTO_IPV6);
1225			if (*mp)
1226				mp = &(*mp)->m_next;
1227			m_freem(ext);
1228		}
1229	}
1230
1231	/* IPV6_DSTOPTS and IPV6_RTHDR socket options */
1232	if (in6p->in6p_flags & (IN6P_DSTOPTS | IN6P_RTHDR)) {
1233		struct ip6_hdr *xip6 = mtod(m, struct ip6_hdr *);
1234		int nxt = xip6->ip6_nxt, off = sizeof(struct ip6_hdr);
1235
1236		/*
1237		 * Search for destination options headers or routing
1238		 * header(s) through the header chain, and stores each
1239		 * header as ancillary data.
1240		 * Note that the order of the headers remains in
1241		 * the chain of ancillary data.
1242		 */
1243		for (;;) {	/* is explicit loop prevention necessary? */
1244			struct ip6_ext *ip6e = NULL;
1245			int elen;
1246			struct mbuf *ext = NULL;
1247
1248			/*
1249			 * if it is not an extension header, don't try to
1250			 * pull it from the chain.
1251			 */
1252			switch (nxt) {
1253			case IPPROTO_DSTOPTS:
1254			case IPPROTO_ROUTING:
1255			case IPPROTO_HOPOPTS:
1256			case IPPROTO_AH: /* is it possible? */
1257				break;
1258			default:
1259				goto loopend;
1260			}
1261
1262			ext = ip6_pullexthdr(m, off, nxt);
1263			if (ext == NULL) {
1264				IP6_STATINC(IP6_STAT_TOOSHORT);
1265				return;
1266			}
1267			ip6e = mtod(ext, struct ip6_ext *);
1268			if (nxt == IPPROTO_AH)
1269				elen = (ip6e->ip6e_len + 2) << 2;
1270			else
1271				elen = (ip6e->ip6e_len + 1) << 3;
1272			if (elen != ext->m_len) {
1273				m_freem(ext);
1274				IP6_STATINC(IP6_STAT_TOOSHORT);
1275				return;
1276			}
1277			KASSERT(IP6_HDR_ALIGNED_P(ip6e));
1278
1279			switch (nxt) {
1280			case IPPROTO_DSTOPTS:
1281				if (!(in6p->in6p_flags & IN6P_DSTOPTS))
1282					break;
1283
1284				*mp = sbcreatecontrol((void *)ip6e, elen,
1285				    IS2292(IPV6_2292DSTOPTS, IPV6_DSTOPTS),
1286				    IPPROTO_IPV6);
1287				if (*mp)
1288					mp = &(*mp)->m_next;
1289				break;
1290
1291			case IPPROTO_ROUTING:
1292				if (!(in6p->in6p_flags & IN6P_RTHDR))
1293					break;
1294
1295				*mp = sbcreatecontrol((void *)ip6e, elen,
1296				    IS2292(IPV6_2292RTHDR, IPV6_RTHDR),
1297				    IPPROTO_IPV6);
1298				if (*mp)
1299					mp = &(*mp)->m_next;
1300				break;
1301
1302			case IPPROTO_HOPOPTS:
1303			case IPPROTO_AH: /* is it possible? */
1304				break;
1305
1306			default:
1307				/*
1308			 	 * other cases have been filtered in the above.
1309				 * none will visit this case.  here we supply
1310				 * the code just in case (nxt overwritten or
1311				 * other cases).
1312				 */
1313				m_freem(ext);
1314				goto loopend;
1315
1316			}
1317
1318			/* proceed with the next header. */
1319			off += elen;
1320			nxt = ip6e->ip6e_nxt;
1321			ip6e = NULL;
1322			m_freem(ext);
1323			ext = NULL;
1324		}
1325	  loopend:
1326	  	;
1327	}
1328}
1329#undef IS2292
1330
1331
1332void
1333ip6_notify_pmtu(struct in6pcb *in6p, const struct sockaddr_in6 *dst,
1334    uint32_t *mtu)
1335{
1336	struct socket *so;
1337	struct mbuf *m_mtu;
1338	struct ip6_mtuinfo mtuctl;
1339
1340	so = in6p->in6p_socket;
1341
1342	if (mtu == NULL)
1343		return;
1344
1345#ifdef DIAGNOSTIC
1346	if (so == NULL)		/* I believe this is impossible */
1347		panic("ip6_notify_pmtu: socket is NULL");
1348#endif
1349
1350	memset(&mtuctl, 0, sizeof(mtuctl));	/* zero-clear for safety */
1351	mtuctl.ip6m_mtu = *mtu;
1352	mtuctl.ip6m_addr = *dst;
1353	if (sa6_recoverscope(&mtuctl.ip6m_addr))
1354		return;
1355
1356	if ((m_mtu = sbcreatecontrol((void *)&mtuctl, sizeof(mtuctl),
1357	    IPV6_PATHMTU, IPPROTO_IPV6)) == NULL)
1358		return;
1359
1360	if (sbappendaddr(&so->so_rcv, (const struct sockaddr *)dst, NULL, m_mtu)
1361	    == 0) {
1362		m_freem(m_mtu);
1363		/* XXX: should count statistics */
1364	} else
1365		sorwakeup(so);
1366
1367	return;
1368}
1369
1370/*
1371 * pull single extension header from mbuf chain.  returns single mbuf that
1372 * contains the result, or NULL on error.
1373 */
1374static struct mbuf *
1375ip6_pullexthdr(struct mbuf *m, size_t off, int nxt)
1376{
1377	struct ip6_ext ip6e;
1378	size_t elen;
1379	struct mbuf *n;
1380
1381#ifdef DIAGNOSTIC
1382	switch (nxt) {
1383	case IPPROTO_DSTOPTS:
1384	case IPPROTO_ROUTING:
1385	case IPPROTO_HOPOPTS:
1386	case IPPROTO_AH: /* is it possible? */
1387		break;
1388	default:
1389		printf("ip6_pullexthdr: invalid nxt=%d\n", nxt);
1390	}
1391#endif
1392
1393	m_copydata(m, off, sizeof(ip6e), (void *)&ip6e);
1394	if (nxt == IPPROTO_AH)
1395		elen = (ip6e.ip6e_len + 2) << 2;
1396	else
1397		elen = (ip6e.ip6e_len + 1) << 3;
1398
1399	MGET(n, M_DONTWAIT, MT_DATA);
1400	if (n && elen >= MLEN) {
1401		MCLGET(n, M_DONTWAIT);
1402		if ((n->m_flags & M_EXT) == 0) {
1403			m_free(n);
1404			n = NULL;
1405		}
1406	}
1407	if (!n)
1408		return NULL;
1409
1410	n->m_len = 0;
1411	if (elen >= M_TRAILINGSPACE(n)) {
1412		m_free(n);
1413		return NULL;
1414	}
1415
1416	m_copydata(m, off, elen, mtod(n, void *));
1417	n->m_len = elen;
1418	return n;
1419}
1420
1421/*
1422 * Get pointer to the previous header followed by the header
1423 * currently processed.
1424 * XXX: This function supposes that
1425 *	M includes all headers,
1426 *	the next header field and the header length field of each header
1427 *	are valid, and
1428 *	the sum of each header length equals to OFF.
1429 * Because of these assumptions, this function must be called very
1430 * carefully. Moreover, it will not be used in the near future when
1431 * we develop `neater' mechanism to process extension headers.
1432 */
1433u_int8_t *
1434ip6_get_prevhdr(struct mbuf *m, int off)
1435{
1436	struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *);
1437
1438	if (off == sizeof(struct ip6_hdr))
1439		return (&ip6->ip6_nxt);
1440	else {
1441		int len, nxt;
1442		struct ip6_ext *ip6e = NULL;
1443
1444		nxt = ip6->ip6_nxt;
1445		len = sizeof(struct ip6_hdr);
1446		while (len < off) {
1447			ip6e = (struct ip6_ext *)(mtod(m, char *) + len);
1448
1449			switch (nxt) {
1450			case IPPROTO_FRAGMENT:
1451				len += sizeof(struct ip6_frag);
1452				break;
1453			case IPPROTO_AH:
1454				len += (ip6e->ip6e_len + 2) << 2;
1455				break;
1456			default:
1457				len += (ip6e->ip6e_len + 1) << 3;
1458				break;
1459			}
1460			nxt = ip6e->ip6e_nxt;
1461		}
1462		if (ip6e)
1463			return (&ip6e->ip6e_nxt);
1464		else
1465			return NULL;
1466	}
1467}
1468
1469/*
1470 * get next header offset.  m will be retained.
1471 */
1472int
1473ip6_nexthdr(struct mbuf *m, int off, int proto, int *nxtp)
1474{
1475	struct ip6_hdr ip6;
1476	struct ip6_ext ip6e;
1477	struct ip6_frag fh;
1478
1479	/* just in case */
1480	if (m == NULL)
1481		panic("ip6_nexthdr: m == NULL");
1482	if ((m->m_flags & M_PKTHDR) == 0 || m->m_pkthdr.len < off)
1483		return -1;
1484
1485	switch (proto) {
1486	case IPPROTO_IPV6:
1487		/* do not chase beyond intermediate IPv6 headers */
1488		if (off != 0)
1489			return -1;
1490		if (m->m_pkthdr.len < off + sizeof(ip6))
1491			return -1;
1492		m_copydata(m, off, sizeof(ip6), (void *)&ip6);
1493		if (nxtp)
1494			*nxtp = ip6.ip6_nxt;
1495		off += sizeof(ip6);
1496		return off;
1497
1498	case IPPROTO_FRAGMENT:
1499		/*
1500		 * terminate parsing if it is not the first fragment,
1501		 * it does not make sense to parse through it.
1502		 */
1503		if (m->m_pkthdr.len < off + sizeof(fh))
1504			return -1;
1505		m_copydata(m, off, sizeof(fh), (void *)&fh);
1506		if ((fh.ip6f_offlg & IP6F_OFF_MASK) != 0)
1507			return -1;
1508		if (nxtp)
1509			*nxtp = fh.ip6f_nxt;
1510		off += sizeof(struct ip6_frag);
1511		return off;
1512
1513	case IPPROTO_AH:
1514		if (m->m_pkthdr.len < off + sizeof(ip6e))
1515			return -1;
1516		m_copydata(m, off, sizeof(ip6e), (void *)&ip6e);
1517		if (nxtp)
1518			*nxtp = ip6e.ip6e_nxt;
1519		off += (ip6e.ip6e_len + 2) << 2;
1520		if (m->m_pkthdr.len < off)
1521			return -1;
1522		return off;
1523
1524	case IPPROTO_HOPOPTS:
1525	case IPPROTO_ROUTING:
1526	case IPPROTO_DSTOPTS:
1527		if (m->m_pkthdr.len < off + sizeof(ip6e))
1528			return -1;
1529		m_copydata(m, off, sizeof(ip6e), (void *)&ip6e);
1530		if (nxtp)
1531			*nxtp = ip6e.ip6e_nxt;
1532		off += (ip6e.ip6e_len + 1) << 3;
1533		if (m->m_pkthdr.len < off)
1534			return -1;
1535		return off;
1536
1537	case IPPROTO_NONE:
1538	case IPPROTO_ESP:
1539	case IPPROTO_IPCOMP:
1540		/* give up */
1541		return -1;
1542
1543	default:
1544		return -1;
1545	}
1546}
1547
1548/*
1549 * get offset for the last header in the chain.  m will be kept untainted.
1550 */
1551int
1552ip6_lasthdr(struct mbuf *m, int off, int proto, int *nxtp)
1553{
1554	int newoff;
1555	int nxt;
1556
1557	if (!nxtp) {
1558		nxt = -1;
1559		nxtp = &nxt;
1560	}
1561	for (;;) {
1562		newoff = ip6_nexthdr(m, off, proto, nxtp);
1563		if (newoff < 0)
1564			return off;
1565		else if (newoff < off)
1566			return -1;	/* invalid */
1567		else if (newoff == off)
1568			return newoff;
1569
1570		off = newoff;
1571		proto = *nxtp;
1572	}
1573}
1574
1575struct m_tag *
1576ip6_addaux(struct mbuf *m)
1577{
1578	struct m_tag *mtag;
1579
1580	mtag = m_tag_find(m, PACKET_TAG_INET6, NULL);
1581	if (!mtag) {
1582		mtag = m_tag_get(PACKET_TAG_INET6, sizeof(struct ip6aux),
1583		    M_NOWAIT);
1584		if (mtag) {
1585			m_tag_prepend(m, mtag);
1586			memset(mtag + 1, 0, sizeof(struct ip6aux));
1587		}
1588	}
1589	return mtag;
1590}
1591
1592struct m_tag *
1593ip6_findaux(struct mbuf *m)
1594{
1595	struct m_tag *mtag;
1596
1597	mtag = m_tag_find(m, PACKET_TAG_INET6, NULL);
1598	return mtag;
1599}
1600
1601void
1602ip6_delaux(struct mbuf *m)
1603{
1604	struct m_tag *mtag;
1605
1606	mtag = m_tag_find(m, PACKET_TAG_INET6, NULL);
1607	if (mtag)
1608		m_tag_delete(m, mtag);
1609}
1610
1611#ifdef GATEWAY
1612/*
1613 * sysctl helper routine for net.inet.ip6.maxflows. Since
1614 * we could reduce this value, call ip6flow_reap();
1615 */
1616static int
1617sysctl_net_inet6_ip6_maxflows(SYSCTLFN_ARGS)
1618{
1619	int error;
1620
1621	error = sysctl_lookup(SYSCTLFN_CALL(rnode));
1622	if (error || newp == NULL)
1623		return (error);
1624
1625	mutex_enter(softnet_lock);
1626	KERNEL_LOCK(1, NULL);
1627
1628	ip6flow_reap(0);
1629
1630	KERNEL_UNLOCK_ONE(NULL);
1631	mutex_exit(softnet_lock);
1632
1633	return (0);
1634}
1635
1636static int
1637sysctl_net_inet6_ip6_hashsize(SYSCTLFN_ARGS)
1638{
1639	int error, tmp;
1640	struct sysctlnode node;
1641
1642	node = *rnode;
1643	tmp = ip6_hashsize;
1644	node.sysctl_data = &tmp;
1645	error = sysctl_lookup(SYSCTLFN_CALL(&node));
1646	if (error || newp == NULL)
1647		return (error);
1648
1649	if ((tmp & (tmp - 1)) == 0 && tmp != 0) {
1650		/*
1651		 * Can only fail due to malloc()
1652		 */
1653		mutex_enter(softnet_lock);
1654		KERNEL_LOCK(1, NULL);
1655
1656		error = ip6flow_invalidate_all(tmp);
1657
1658		KERNEL_UNLOCK_ONE(NULL);
1659		mutex_exit(softnet_lock);
1660	} else {
1661		/*
1662		 * EINVAL if not a power of 2
1663		 */
1664		error = EINVAL;
1665	}
1666
1667	return error;
1668}
1669#endif /* GATEWAY */
1670
1671/*
1672 * System control for IP6
1673 */
1674
1675const u_char inet6ctlerrmap[PRC_NCMDS] = {
1676	0,		0,		0,		0,
1677	0,		EMSGSIZE,	EHOSTDOWN,	EHOSTUNREACH,
1678	EHOSTUNREACH,	EHOSTUNREACH,	ECONNREFUSED,	ECONNREFUSED,
1679	EMSGSIZE,	EHOSTUNREACH,	0,		0,
1680	0,		0,		0,		0,
1681	ENOPROTOOPT
1682};
1683
1684static int
1685sysctl_net_inet6_ip6_stats(SYSCTLFN_ARGS)
1686{
1687
1688	return (NETSTAT_SYSCTL(ip6stat_percpu, IP6_NSTATS));
1689}
1690
1691static void
1692sysctl_net_inet6_ip6_setup(struct sysctllog **clog)
1693{
1694#ifdef RFC2292
1695#define IS2292(x, y)	((in6p->in6p_flags & IN6P_RFC2292) ? (x) : (y))
1696#else
1697#define IS2292(x, y)	(y)
1698#endif
1699
1700	sysctl_createv(clog, 0, NULL, NULL,
1701		       CTLFLAG_PERMANENT,
1702		       CTLTYPE_NODE, "net", NULL,
1703		       NULL, 0, NULL, 0,
1704		       CTL_NET, CTL_EOL);
1705	sysctl_createv(clog, 0, NULL, NULL,
1706		       CTLFLAG_PERMANENT,
1707		       CTLTYPE_NODE, "inet6",
1708		       SYSCTL_DESCR("PF_INET6 related settings"),
1709		       NULL, 0, NULL, 0,
1710		       CTL_NET, PF_INET6, CTL_EOL);
1711	sysctl_createv(clog, 0, NULL, NULL,
1712		       CTLFLAG_PERMANENT,
1713		       CTLTYPE_NODE, "ip6",
1714		       SYSCTL_DESCR("IPv6 related settings"),
1715		       NULL, 0, NULL, 0,
1716		       CTL_NET, PF_INET6, IPPROTO_IPV6, CTL_EOL);
1717
1718	sysctl_createv(clog, 0, NULL, NULL,
1719		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1720		       CTLTYPE_INT, "forwarding",
1721		       SYSCTL_DESCR("Enable forwarding of INET6 datagrams"),
1722		       NULL, 0, &ip6_forwarding, 0,
1723		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1724		       IPV6CTL_FORWARDING, CTL_EOL);
1725	sysctl_createv(clog, 0, NULL, NULL,
1726		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1727		       CTLTYPE_INT, "redirect",
1728		       SYSCTL_DESCR("Enable sending of ICMPv6 redirect messages"),
1729		       NULL, 0, &ip6_sendredirects, 0,
1730		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1731		       IPV6CTL_SENDREDIRECTS, CTL_EOL);
1732	sysctl_createv(clog, 0, NULL, NULL,
1733		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1734		       CTLTYPE_INT, "hlim",
1735		       SYSCTL_DESCR("Hop limit for an INET6 datagram"),
1736		       NULL, 0, &ip6_defhlim, 0,
1737		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1738		       IPV6CTL_DEFHLIM, CTL_EOL);
1739#ifdef notyet
1740	sysctl_createv(clog, 0, NULL, NULL,
1741		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1742		       CTLTYPE_INT, "mtu", NULL,
1743		       NULL, 0, &, 0,
1744		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1745		       IPV6CTL_DEFMTU, CTL_EOL);
1746#endif
1747#ifdef __no_idea__
1748	sysctl_createv(clog, 0, NULL, NULL,
1749		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1750		       CTLTYPE_INT, "forwsrcrt", NULL,
1751		       NULL, 0, &?, 0,
1752		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1753		       IPV6CTL_FORWSRCRT, CTL_EOL);
1754	sysctl_createv(clog, 0, NULL, NULL,
1755		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1756		       CTLTYPE_STRUCT, "mrtstats", NULL,
1757		       NULL, 0, &?, sizeof(?),
1758		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1759		       IPV6CTL_MRTSTATS, CTL_EOL);
1760	sysctl_createv(clog, 0, NULL, NULL,
1761		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1762		       CTLTYPE_?, "mrtproto", NULL,
1763		       NULL, 0, &?, sizeof(?),
1764		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1765		       IPV6CTL_MRTPROTO, CTL_EOL);
1766#endif
1767	sysctl_createv(clog, 0, NULL, NULL,
1768		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1769		       CTLTYPE_INT, "maxfragpackets",
1770		       SYSCTL_DESCR("Maximum number of fragments to buffer "
1771				    "for reassembly"),
1772		       NULL, 0, &ip6_maxfragpackets, 0,
1773		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1774		       IPV6CTL_MAXFRAGPACKETS, CTL_EOL);
1775#ifdef __no_idea__
1776	sysctl_createv(clog, 0, NULL, NULL,
1777		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1778		       CTLTYPE_INT, "sourcecheck", NULL,
1779		       NULL, 0, &?, 0,
1780		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1781		       IPV6CTL_SOURCECHECK, CTL_EOL);
1782	sysctl_createv(clog, 0, NULL, NULL,
1783		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1784		       CTLTYPE_INT, "sourcecheck_logint", NULL,
1785		       NULL, 0, &?, 0,
1786		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1787		       IPV6CTL_SOURCECHECK_LOGINT, CTL_EOL);
1788#endif
1789	sysctl_createv(clog, 0, NULL, NULL,
1790		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1791		       CTLTYPE_INT, "accept_rtadv",
1792		       SYSCTL_DESCR("Accept router advertisements"),
1793		       NULL, 0, &ip6_accept_rtadv, 0,
1794		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1795		       IPV6CTL_ACCEPT_RTADV, CTL_EOL);
1796	sysctl_createv(clog, 0, NULL, NULL,
1797		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1798		       CTLTYPE_INT, "rtadv_maxroutes",
1799		       SYSCTL_DESCR("Maximum number of routes accepted via router advertisements"),
1800		       NULL, 0, &ip6_rtadv_maxroutes, 0,
1801		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1802		       IPV6CTL_RTADV_MAXROUTES, CTL_EOL);
1803	sysctl_createv(clog, 0, NULL, NULL,
1804		       CTLFLAG_PERMANENT,
1805		       CTLTYPE_INT, "rtadv_numroutes",
1806		       SYSCTL_DESCR("Current number of routes accepted via router advertisements"),
1807		       NULL, 0, &nd6_numroutes, 0,
1808		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1809		       IPV6CTL_RTADV_NUMROUTES, CTL_EOL);
1810	sysctl_createv(clog, 0, NULL, NULL,
1811		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1812		       CTLTYPE_INT, "keepfaith",
1813		       SYSCTL_DESCR("Activate faith interface"),
1814		       NULL, 0, &ip6_keepfaith, 0,
1815		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1816		       IPV6CTL_KEEPFAITH, CTL_EOL);
1817	sysctl_createv(clog, 0, NULL, NULL,
1818		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1819		       CTLTYPE_INT, "log_interval",
1820		       SYSCTL_DESCR("Minumum interval between logging "
1821				    "unroutable packets"),
1822		       NULL, 0, &ip6_log_interval, 0,
1823		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1824		       IPV6CTL_LOG_INTERVAL, CTL_EOL);
1825	sysctl_createv(clog, 0, NULL, NULL,
1826		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1827		       CTLTYPE_INT, "hdrnestlimit",
1828		       SYSCTL_DESCR("Maximum number of nested IPv6 headers"),
1829		       NULL, 0, &ip6_hdrnestlimit, 0,
1830		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1831		       IPV6CTL_HDRNESTLIMIT, CTL_EOL);
1832	sysctl_createv(clog, 0, NULL, NULL,
1833		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1834		       CTLTYPE_INT, "dad_count",
1835		       SYSCTL_DESCR("Number of Duplicate Address Detection "
1836				    "probes to send"),
1837		       NULL, 0, &ip6_dad_count, 0,
1838		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1839		       IPV6CTL_DAD_COUNT, CTL_EOL);
1840	sysctl_createv(clog, 0, NULL, NULL,
1841		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1842		       CTLTYPE_INT, "auto_flowlabel",
1843		       SYSCTL_DESCR("Assign random IPv6 flow labels"),
1844		       NULL, 0, &ip6_auto_flowlabel, 0,
1845		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1846		       IPV6CTL_AUTO_FLOWLABEL, CTL_EOL);
1847	sysctl_createv(clog, 0, NULL, NULL,
1848		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1849		       CTLTYPE_INT, "defmcasthlim",
1850		       SYSCTL_DESCR("Default multicast hop limit"),
1851		       NULL, 0, &ip6_defmcasthlim, 0,
1852		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1853		       IPV6CTL_DEFMCASTHLIM, CTL_EOL);
1854#if NGIF > 0
1855	sysctl_createv(clog, 0, NULL, NULL,
1856		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1857		       CTLTYPE_INT, "gifhlim",
1858		       SYSCTL_DESCR("Default hop limit for a gif tunnel datagram"),
1859		       NULL, 0, &ip6_gif_hlim, 0,
1860		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1861		       IPV6CTL_GIF_HLIM, CTL_EOL);
1862#endif /* NGIF */
1863	sysctl_createv(clog, 0, NULL, NULL,
1864		       CTLFLAG_PERMANENT,
1865		       CTLTYPE_STRING, "kame_version",
1866		       SYSCTL_DESCR("KAME Version"),
1867		       NULL, 0, __UNCONST(__KAME_VERSION), 0,
1868		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1869		       IPV6CTL_KAME_VERSION, CTL_EOL);
1870	sysctl_createv(clog, 0, NULL, NULL,
1871		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1872		       CTLTYPE_INT, "use_deprecated",
1873		       SYSCTL_DESCR("Allow use of deprecated addresses as "
1874				    "source addresses"),
1875		       NULL, 0, &ip6_use_deprecated, 0,
1876		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1877		       IPV6CTL_USE_DEPRECATED, CTL_EOL);
1878	sysctl_createv(clog, 0, NULL, NULL,
1879		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1880		       CTLTYPE_INT, "rr_prune", NULL,
1881		       NULL, 0, &ip6_rr_prune, 0,
1882		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1883		       IPV6CTL_RR_PRUNE, CTL_EOL);
1884	sysctl_createv(clog, 0, NULL, NULL,
1885		       CTLFLAG_PERMANENT
1886#ifndef INET6_BINDV6ONLY
1887		       |CTLFLAG_READWRITE,
1888#endif
1889		       CTLTYPE_INT, "v6only",
1890		       SYSCTL_DESCR("Disallow PF_INET6 sockets from connecting "
1891				    "to PF_INET sockets"),
1892		       NULL, 0, &ip6_v6only, 0,
1893		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1894		       IPV6CTL_V6ONLY, CTL_EOL);
1895	sysctl_createv(clog, 0, NULL, NULL,
1896		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1897		       CTLTYPE_INT, "anonportmin",
1898		       SYSCTL_DESCR("Lowest ephemeral port number to assign"),
1899		       sysctl_net_inet_ip_ports, 0, &ip6_anonportmin, 0,
1900		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1901		       IPV6CTL_ANONPORTMIN, CTL_EOL);
1902	sysctl_createv(clog, 0, NULL, NULL,
1903		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1904		       CTLTYPE_INT, "anonportmax",
1905		       SYSCTL_DESCR("Highest ephemeral port number to assign"),
1906		       sysctl_net_inet_ip_ports, 0, &ip6_anonportmax, 0,
1907		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1908		       IPV6CTL_ANONPORTMAX, CTL_EOL);
1909#ifndef IPNOPRIVPORTS
1910	sysctl_createv(clog, 0, NULL, NULL,
1911		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1912		       CTLTYPE_INT, "lowportmin",
1913		       SYSCTL_DESCR("Lowest privileged ephemeral port number "
1914				    "to assign"),
1915		       sysctl_net_inet_ip_ports, 0, &ip6_lowportmin, 0,
1916		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1917		       IPV6CTL_LOWPORTMIN, CTL_EOL);
1918	sysctl_createv(clog, 0, NULL, NULL,
1919		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1920		       CTLTYPE_INT, "lowportmax",
1921		       SYSCTL_DESCR("Highest privileged ephemeral port number "
1922				    "to assign"),
1923		       sysctl_net_inet_ip_ports, 0, &ip6_lowportmax, 0,
1924		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1925		       IPV6CTL_LOWPORTMAX, CTL_EOL);
1926#endif /* IPNOPRIVPORTS */
1927	sysctl_createv(clog, 0, NULL, NULL,
1928		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1929		       CTLTYPE_INT, "use_tempaddr",
1930		       SYSCTL_DESCR("Use temporary address"),
1931		       NULL, 0, &ip6_use_tempaddr, 0,
1932		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1933		       CTL_CREATE, CTL_EOL);
1934	sysctl_createv(clog, 0, NULL, NULL,
1935		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1936		       CTLTYPE_INT, "temppltime",
1937		       SYSCTL_DESCR("preferred lifetime of a temporary address"),
1938		       NULL, 0, &ip6_temp_preferred_lifetime, 0,
1939		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1940		       CTL_CREATE, CTL_EOL);
1941	sysctl_createv(clog, 0, NULL, NULL,
1942		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1943		       CTLTYPE_INT, "tempvltime",
1944		       SYSCTL_DESCR("valid lifetime of a temporary address"),
1945		       NULL, 0, &ip6_temp_valid_lifetime, 0,
1946		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1947		       CTL_CREATE, CTL_EOL);
1948	sysctl_createv(clog, 0, NULL, NULL,
1949		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1950		       CTLTYPE_INT, "maxfrags",
1951		       SYSCTL_DESCR("Maximum fragments in reassembly queue"),
1952		       NULL, 0, &ip6_maxfrags, 0,
1953		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1954		       IPV6CTL_MAXFRAGS, CTL_EOL);
1955	sysctl_createv(clog, 0, NULL, NULL,
1956		       CTLFLAG_PERMANENT,
1957		       CTLTYPE_STRUCT, "stats",
1958		       SYSCTL_DESCR("IPv6 statistics"),
1959		       sysctl_net_inet6_ip6_stats, 0, NULL, 0,
1960		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1961		       IPV6CTL_STATS, CTL_EOL);
1962	sysctl_createv(clog, 0, NULL, NULL,
1963		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1964		       CTLTYPE_INT, "use_defaultzone",
1965		       SYSCTL_DESCR("Whether to use the default scope zones"),
1966		       NULL, 0, &ip6_use_defzone, 0,
1967		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1968		       IPV6CTL_USE_DEFAULTZONE, CTL_EOL);
1969	sysctl_createv(clog, 0, NULL, NULL,
1970		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1971		       CTLTYPE_INT, "mcast_pmtu",
1972		       SYSCTL_DESCR("Enable pMTU discovery for multicast packet"),
1973		       NULL, 0, &ip6_mcast_pmtu, 0,
1974		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1975		       CTL_CREATE, CTL_EOL);
1976#ifdef GATEWAY
1977	sysctl_createv(clog, 0, NULL, NULL,
1978			CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1979			CTLTYPE_INT, "maxflows",
1980			SYSCTL_DESCR("Number of flows for fast forwarding (IPv6)"),
1981			sysctl_net_inet6_ip6_maxflows, 0, &ip6_maxflows, 0,
1982			CTL_NET, PF_INET6, IPPROTO_IPV6,
1983			CTL_CREATE, CTL_EOL);
1984	sysctl_createv(clog, 0, NULL, NULL,
1985			CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1986			CTLTYPE_INT, "hashsize",
1987			SYSCTL_DESCR("Size of hash table for fast forwarding (IPv6)"),
1988			sysctl_net_inet6_ip6_hashsize, 0, &ip6_hashsize, 0,
1989			CTL_NET, PF_INET6, IPPROTO_IPV6,
1990			CTL_CREATE, CTL_EOL);
1991#endif
1992	sysctl_createv(clog, 0, NULL, NULL,
1993		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1994		       CTLTYPE_INT, "neighborgcthresh",
1995		       SYSCTL_DESCR("Maximum number of entries in neighbor"
1996			 " cache"),
1997		       NULL, 1, &ip6_neighborgcthresh, 0,
1998		       CTL_NET, PF_INET6, IPPROTO_IPV6,
1999		       CTL_CREATE, CTL_EOL);
2000	sysctl_createv(clog, 0, NULL, NULL,
2001		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
2002		       CTLTYPE_INT, "maxifprefixes",
2003		       SYSCTL_DESCR("Maximum number of prefixes created by"
2004			   " route advertisement per interface"),
2005		       NULL, 1, &ip6_maxifprefixes, 0,
2006		       CTL_NET, PF_INET6, IPPROTO_IPV6,
2007		       CTL_CREATE, CTL_EOL);
2008	sysctl_createv(clog, 0, NULL, NULL,
2009		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
2010		       CTLTYPE_INT, "maxifdefrouters",
2011		       SYSCTL_DESCR("Maximum number of default routers created"
2012			   " by route advertisement per interface"),
2013		       NULL, 1, &ip6_maxifdefrouters, 0,
2014		       CTL_NET, PF_INET6, IPPROTO_IPV6,
2015		       CTL_CREATE, CTL_EOL);
2016	sysctl_createv(clog, 0, NULL, NULL,
2017		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
2018		       CTLTYPE_INT, "maxdynroutes",
2019		       SYSCTL_DESCR("Maximum number of routes created via"
2020			   " redirect"),
2021		       NULL, 1, &ip6_maxdynroutes, 0,
2022		       CTL_NET, PF_INET6, IPPROTO_IPV6,
2023		       CTL_CREATE, CTL_EOL);
2024}
2025
2026void
2027ip6_statinc(u_int stat)
2028{
2029
2030	KASSERT(stat < IP6_NSTATS);
2031	IP6_STATINC(stat);
2032}
2033