1/* $NetBSD: ip6_input.c,v 1.170 2016/11/01 10:32:57 ozaki-r 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 1.170 2016/11/01 10:32:57 ozaki-r Exp $");
66
67#ifdef _KERNEL_OPT
68#include "opt_gateway.h"
69#include "opt_inet.h"
70#include "opt_inet6.h"
71#include "opt_ipsec.h"
72#include "opt_compat_netbsd.h"
73#include "opt_net_mpsafe.h"
74#endif
75
76#include <sys/param.h>
77#include <sys/systm.h>
78#include <sys/malloc.h>
79#include <sys/mbuf.h>
80#include <sys/domain.h>
81#include <sys/protosw.h>
82#include <sys/socket.h>
83#include <sys/socketvar.h>
84#include <sys/errno.h>
85#include <sys/time.h>
86#include <sys/kernel.h>
87#include <sys/syslog.h>
88#include <sys/proc.h>
89#include <sys/sysctl.h>
90#include <sys/cprng.h>
91#include <sys/percpu.h>
92
93#include <net/if.h>
94#include <net/if_types.h>
95#include <net/if_dl.h>
96#include <net/route.h>
97#include <net/pktqueue.h>
98#include <net/pfil.h>
99
100#include <netinet/in.h>
101#include <netinet/in_systm.h>
102#ifdef INET
103#include <netinet/ip.h>
104#include <netinet/ip_var.h>
105#include <netinet/ip_icmp.h>
106#endif /* INET */
107#include <netinet/ip6.h>
108#include <netinet/portalgo.h>
109#include <netinet6/in6_var.h>
110#include <netinet6/ip6_var.h>
111#include <netinet6/ip6_private.h>
112#include <netinet6/in6_pcb.h>
113#include <netinet/icmp6.h>
114#include <netinet6/scope6_var.h>
115#include <netinet6/in6_ifattach.h>
116#include <netinet6/nd6.h>
117
118#ifdef IPSEC
119#include <netipsec/ipsec.h>
120#include <netipsec/ipsec6.h>
121#include <netipsec/key.h>
122#endif /* IPSEC */
123
124#ifdef COMPAT_50
125#include <compat/sys/time.h>
126#include <compat/sys/socket.h>
127#endif
128
129#include <netinet6/ip6protosw.h>
130
131#include "faith.h"
132
133#include <net/net_osdep.h>
134
135extern struct domain inet6domain;
136
137u_char ip6_protox[IPPROTO_MAX];
138pktqueue_t *ip6_pktq __read_mostly;
139
140int ip6_forward_srcrt; /* XXX */
141int ip6_sourcecheck; /* XXX */
142int ip6_sourcecheck_interval; /* XXX */
143
144pfil_head_t *inet6_pfil_hook;
145
146percpu_t *ip6stat_percpu;
147
148percpu_t *ip6_forward_rt_percpu __cacheline_aligned;
149
150static void ip6_init2(void);
151static void ip6intr(void *);
152static struct m_tag *ip6_setdstifaddr(struct mbuf *, const struct in6_ifaddr *);
153
154static int ip6_process_hopopts(struct mbuf *, u_int8_t *, int, u_int32_t *,
155 u_int32_t *);
156static struct mbuf *ip6_pullexthdr(struct mbuf *, size_t, int);
157static void sysctl_net_inet6_ip6_setup(struct sysctllog **);
158
159#ifdef NET_MPSAFE
160#define SOFTNET_LOCK() mutex_enter(softnet_lock)
161#define SOFTNET_UNLOCK() mutex_exit(softnet_lock)
162#else
163#define SOFTNET_LOCK() KASSERT(mutex_owned(softnet_lock))
164#define SOFTNET_UNLOCK() KASSERT(mutex_owned(softnet_lock))
165#endif
166
167/*
168 * IP6 initialization: fill in IP6 protocol switch table.
169 * All protocols not implemented in kernel go to raw IP6 protocol handler.
170 */
171void
172ip6_init(void)
173{
174 const struct ip6protosw *pr;
175 int i;
176
177 in6_init();
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
191 ip6_pktq = pktq_create(IFQ_MAXLEN, ip6intr, NULL);
192 KASSERT(ip6_pktq != NULL);
193
194 scope6_init();
195 addrsel_policy_init();
196 nd6_init();
197 frag6_init();
198 ip6_desync_factor = cprng_fast32() % MAX_TEMP_DESYNC_FACTOR;
199
200 ip6_init2();
201#ifdef GATEWAY
202 ip6flow_init(ip6_hashsize);
203#endif
204 /* Register our Packet Filter hook. */
205 inet6_pfil_hook = pfil_head_create(PFIL_TYPE_AF, (void *)AF_INET6);
206 KASSERT(inet6_pfil_hook != NULL);
207
208 ip6stat_percpu = percpu_alloc(sizeof(uint64_t) * IP6_NSTATS);
209
210 ip6_forward_rt_percpu = percpu_alloc(sizeof(struct route));
211 if (ip6_forward_rt_percpu == NULL)
212 panic("failed to alllocate ip6_forward_rt_percpu");
213}
214
215static void
216ip6_init2(void)
217{
218
219 /* timer for regeneranation of temporary addresses randomize ID */
220 callout_init(&in6_tmpaddrtimer_ch, CALLOUT_MPSAFE);
221 callout_reset(&in6_tmpaddrtimer_ch,
222 (ip6_temp_preferred_lifetime - ip6_desync_factor -
223 ip6_temp_regen_advance) * hz,
224 in6_tmpaddrtimer, NULL);
225}
226
227/*
228 * IP6 input interrupt handling. Just pass the packet to ip6_input.
229 */
230static void
231ip6intr(void *arg __unused)
232{
233 struct mbuf *m;
234
235#ifndef NET_MPSAFE
236 mutex_enter(softnet_lock);
237#endif
238 while ((m = pktq_dequeue(ip6_pktq)) != NULL) {
239 struct psref psref;
240 struct ifnet *rcvif = m_get_rcvif_psref(m, &psref);
241
242 if (rcvif == NULL) {
243 m_freem(m);
244 continue;
245 }
246 /*
247 * Drop the packet if IPv6 is disabled on the interface.
248 */
249 if ((ND_IFINFO(rcvif)->flags & ND6_IFF_IFDISABLED)) {
250 m_put_rcvif_psref(rcvif, &psref);
251 m_freem(m);
252 continue;
253 }
254 ip6_input(m, rcvif);
255 m_put_rcvif_psref(rcvif, &psref);
256 }
257#ifndef NET_MPSAFE
258 mutex_exit(softnet_lock);
259#endif
260}
261
262void
263ip6_input(struct mbuf *m, struct ifnet *rcvif)
264{
265 struct ip6_hdr *ip6;
266 int hit, off = sizeof(struct ip6_hdr), nest;
267 u_int32_t plen;
268 u_int32_t rtalert = ~0;
269 int nxt, ours = 0, rh_present = 0;
270 struct ifnet *deliverifp = NULL;
271 int srcrt = 0;
272 const struct rtentry *rt;
273 union {
274 struct sockaddr dst;
275 struct sockaddr_in6 dst6;
276 } u;
277 struct route *ro;
278
279 /*
280 * make sure we don't have onion peering information into m_tag.
281 */
282 ip6_delaux(m);
283
284 /*
285 * mbuf statistics
286 */
287 if (m->m_flags & M_EXT) {
288 if (m->m_next)
289 IP6_STATINC(IP6_STAT_MEXT2M);
290 else
291 IP6_STATINC(IP6_STAT_MEXT1);
292 } else {
293#define M2MMAX 32
294 if (m->m_next) {
295 if (m->m_flags & M_LOOP)
296 /*XXX*/ IP6_STATINC(IP6_STAT_M2M + lo0ifp->if_index);
297 else if (rcvif->if_index < M2MMAX)
298 IP6_STATINC(IP6_STAT_M2M + rcvif->if_index);
299 else
300 IP6_STATINC(IP6_STAT_M2M);
301 } else
302 IP6_STATINC(IP6_STAT_M1);
303#undef M2MMAX
304 }
305
306 in6_ifstat_inc(rcvif, ifs6_in_receive);
307 IP6_STATINC(IP6_STAT_TOTAL);
308
309 /*
310 * If the IPv6 header is not aligned, slurp it up into a new
311 * mbuf with space for link headers, in the event we forward
312 * it. Otherwise, if it is aligned, make sure the entire base
313 * IPv6 header is in the first mbuf of the chain.
314 */
315 if (IP6_HDR_ALIGNED_P(mtod(m, void *)) == 0) {
316 if ((m = m_copyup(m, sizeof(struct ip6_hdr),
317 (max_linkhdr + 3) & ~3)) == NULL) {
318 /* XXXJRT new stat, please */
319 IP6_STATINC(IP6_STAT_TOOSMALL);
320 in6_ifstat_inc(rcvif, ifs6_in_hdrerr);
321 return;
322 }
323 } else if (__predict_false(m->m_len < sizeof(struct ip6_hdr))) {
324 if ((m = m_pullup(m, sizeof(struct ip6_hdr))) == NULL) {
325 IP6_STATINC(IP6_STAT_TOOSMALL);
326 in6_ifstat_inc(rcvif, ifs6_in_hdrerr);
327 return;
328 }
329 }
330
331 ip6 = mtod(m, struct ip6_hdr *);
332
333 if ((ip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION) {
334 IP6_STATINC(IP6_STAT_BADVERS);
335 in6_ifstat_inc(rcvif, ifs6_in_hdrerr);
336 goto bad;
337 }
338
339 /*
340 * Assume that we can create a fast-forward IP flow entry
341 * based on this packet.
342 */
343 m->m_flags |= M_CANFASTFWD;
344
345 /*
346 * Run through list of hooks for input packets. If there are any
347 * filters which require that additional packets in the flow are
348 * not fast-forwarded, they must clear the M_CANFASTFWD flag.
349 * Note that filters must _never_ set this flag, as another filter
350 * in the list may have previously cleared it.
351 */
352 /*
353 * let ipfilter look at packet on the wire,
354 * not the decapsulated packet.
355 */
356#if defined(IPSEC)
357 if (!ipsec_used || !ipsec_indone(m))
358#else
359 if (1)
360#endif
361 {
362 struct in6_addr odst;
363
364 odst = ip6->ip6_dst;
365 if (pfil_run_hooks(inet6_pfil_hook, &m, rcvif, PFIL_IN) != 0)
366 return;
367 if (m == NULL)
368 return;
369 ip6 = mtod(m, struct ip6_hdr *);
370 srcrt = !IN6_ARE_ADDR_EQUAL(&odst, &ip6->ip6_dst);
371 }
372
373 IP6_STATINC(IP6_STAT_NXTHIST + ip6->ip6_nxt);
374
375#ifdef ALTQ
376 if (altq_input != NULL) {
377 SOFTNET_LOCK();
378 if ((*altq_input)(m, AF_INET6) == 0) {
379 SOFTNET_UNLOCK();
380 /* packet is dropped by traffic conditioner */
381 return;
382 }
383 SOFTNET_UNLOCK();
384 }
385#endif
386
387 /*
388 * Check against address spoofing/corruption.
389 */
390 if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_src) ||
391 IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_dst)) {
392 /*
393 * XXX: "badscope" is not very suitable for a multicast source.
394 */
395 IP6_STATINC(IP6_STAT_BADSCOPE);
396 in6_ifstat_inc(rcvif, ifs6_in_addrerr);
397 goto bad;
398 }
399 /*
400 * The following check is not documented in specs. A malicious
401 * party may be able to use IPv4 mapped addr to confuse tcp/udp stack
402 * and bypass security checks (act as if it was from 127.0.0.1 by using
403 * IPv6 src ::ffff:127.0.0.1). Be cautious.
404 *
405 * This check chokes if we are in an SIIT cloud. As none of BSDs
406 * support IPv4-less kernel compilation, we cannot support SIIT
407 * environment at all. So, it makes more sense for us to reject any
408 * malicious packets for non-SIIT environment, than try to do a
409 * partial support for SIIT environment.
410 */
411 if (IN6_IS_ADDR_V4MAPPED(&ip6->ip6_src) ||
412 IN6_IS_ADDR_V4MAPPED(&ip6->ip6_dst)) {
413 IP6_STATINC(IP6_STAT_BADSCOPE);
414 in6_ifstat_inc(rcvif, ifs6_in_addrerr);
415 goto bad;
416 }
417#if 0
418 /*
419 * Reject packets with IPv4 compatible addresses (auto tunnel).
420 *
421 * The code forbids auto tunnel relay case in RFC1933 (the check is
422 * stronger than RFC1933). We may want to re-enable it if mech-xx
423 * is revised to forbid relaying case.
424 */
425 if (IN6_IS_ADDR_V4COMPAT(&ip6->ip6_src) ||
426 IN6_IS_ADDR_V4COMPAT(&ip6->ip6_dst)) {
427 IP6_STATINC(IP6_STAT_BADSCOPE);
428 in6_ifstat_inc(rcvif, ifs6_in_addrerr);
429 goto bad;
430 }
431#endif
432
433 /*
434 * Disambiguate address scope zones (if there is ambiguity).
435 * We first make sure that the original source or destination address
436 * is not in our internal form for scoped addresses. Such addresses
437 * are not necessarily invalid spec-wise, but we cannot accept them due
438 * to the usage conflict.
439 * in6_setscope() then also checks and rejects the cases where src or
440 * dst are the loopback address and the receiving interface
441 * is not loopback.
442 */
443 if (__predict_false(
444 m_makewritable(&m, 0, sizeof(struct ip6_hdr), M_DONTWAIT)))
445 goto bad;
446 ip6 = mtod(m, struct ip6_hdr *);
447 if (in6_clearscope(&ip6->ip6_src) || in6_clearscope(&ip6->ip6_dst)) {
448 IP6_STATINC(IP6_STAT_BADSCOPE); /* XXX */
449 goto bad;
450 }
451 if (in6_setscope(&ip6->ip6_src, rcvif, NULL) ||
452 in6_setscope(&ip6->ip6_dst, rcvif, NULL)) {
453 IP6_STATINC(IP6_STAT_BADSCOPE);
454 goto bad;
455 }
456
457 /*
458 * Multicast check
459 */
460 if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) {
461 struct in6_multi *in6m = 0;
462
463 in6_ifstat_inc(rcvif, ifs6_in_mcast);
464 /*
465 * See if we belong to the destination multicast group on the
466 * arrival interface.
467 */
468 IN6_LOOKUP_MULTI(ip6->ip6_dst, rcvif, in6m);
469 if (in6m)
470 ours = 1;
471 else if (!ip6_mrouter) {
472 uint64_t *ip6s = IP6_STAT_GETREF();
473 ip6s[IP6_STAT_NOTMEMBER]++;
474 ip6s[IP6_STAT_CANTFORWARD]++;
475 IP6_STAT_PUTREF();
476 in6_ifstat_inc(rcvif, ifs6_in_discard);
477 goto bad;
478 }
479 deliverifp = rcvif;
480 goto hbhcheck;
481 }
482
483 sockaddr_in6_init(&u.dst6, &ip6->ip6_dst, 0, 0, 0);
484
485 /*
486 * Unicast check
487 */
488 ro = percpu_getref(ip6_forward_rt_percpu);
489 rt = rtcache_lookup2(ro, &u.dst, 1, &hit);
490 percpu_putref(ip6_forward_rt_percpu);
491 if (hit)
492 IP6_STATINC(IP6_STAT_FORWARD_CACHEHIT);
493 else
494 IP6_STATINC(IP6_STAT_FORWARD_CACHEMISS);
495
496#define rt6_getkey(__rt) satocsin6(rt_getkey(__rt))
497
498 /*
499 * Accept the packet if the forwarding interface to the destination
500 * according to the routing table is the loopback interface,
501 * unless the associated route has a gateway.
502 * Note that this approach causes to accept a packet if there is a
503 * route to the loopback interface for the destination of the packet.
504 * But we think it's even useful in some situations, e.g. when using
505 * a special daemon which wants to intercept the packet.
506 */
507 if (rt != NULL &&
508 (rt->rt_flags & (RTF_HOST|RTF_GATEWAY)) == RTF_HOST &&
509#if 0
510 /*
511 * The check below is redundant since the comparison of
512 * the destination and the key of the rtentry has
513 * already done through looking up the routing table.
514 */
515 IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst, &rt6_getkey(rt)->sin6_addr) &&
516#endif
517 rt->rt_ifp->if_type == IFT_LOOP) {
518 struct in6_ifaddr *ia6 = (struct in6_ifaddr *)rt->rt_ifa;
519 if (ia6->ia6_flags & IN6_IFF_ANYCAST)
520 m->m_flags |= M_ANYCAST6;
521 /*
522 * packets to a tentative, duplicated, or somehow invalid
523 * address must not be accepted.
524 */
525 if (!(ia6->ia6_flags & (IN6_IFF_NOTREADY | IN6_IFF_DETACHED))) {
526 /* this address is ready */
527 ours = 1;
528 deliverifp = ia6->ia_ifp; /* correct? */
529 goto hbhcheck;
530 } else {
531 /* address is not ready, so discard the packet. */
532 nd6log(LOG_INFO, "packet to an unready address %s->%s\n",
533 ip6_sprintf(&ip6->ip6_src),
534 ip6_sprintf(&ip6->ip6_dst));
535
536 goto bad;
537 }
538 }
539
540 /*
541 * FAITH (Firewall Aided Internet Translator)
542 */
543#if defined(NFAITH) && 0 < NFAITH
544 if (ip6_keepfaith) {
545 if (rt != NULL && rt->rt_ifp != NULL &&
546 rt->rt_ifp->if_type == IFT_FAITH) {
547 /* XXX do we need more sanity checks? */
548 ours = 1;
549 deliverifp = rt->rt_ifp; /* faith */
550 goto hbhcheck;
551 }
552 }
553#endif
554
555#if 0
556 {
557 /*
558 * Last resort: check in6_ifaddr for incoming interface.
559 * The code is here until I update the "goto ours hack" code above
560 * working right.
561 */
562 struct ifaddr *ifa;
563 IFADDR_READER_FOREACH(ifa, rcvif) {
564 if (ifa->ifa_addr->sa_family != AF_INET6)
565 continue;
566 if (IN6_ARE_ADDR_EQUAL(IFA_IN6(ifa), &ip6->ip6_dst)) {
567 ours = 1;
568 deliverifp = ifa->ifa_ifp;
569 goto hbhcheck;
570 }
571 }
572 }
573#endif
574
575 /*
576 * Now there is no reason to process the packet if it's not our own
577 * and we're not a router.
578 */
579 if (!ip6_forwarding) {
580 IP6_STATINC(IP6_STAT_CANTFORWARD);
581 in6_ifstat_inc(rcvif, ifs6_in_discard);
582 goto bad;
583 }
584
585 hbhcheck:
586 /*
587 * record address information into m_tag, if we don't have one yet.
588 * note that we are unable to record it, if the address is not listed
589 * as our interface address (e.g. multicast addresses, addresses
590 * within FAITH prefixes and such).
591 */
592 if (deliverifp && ip6_getdstifaddr(m) == NULL) {
593 struct in6_ifaddr *ia6;
594 int s = pserialize_read_enter();
595
596 ia6 = in6_ifawithifp(deliverifp, &ip6->ip6_dst);
597 /* Depends on ip6_setdstifaddr never sleep */
598 if (ia6 != NULL && ip6_setdstifaddr(m, ia6) == NULL) {
599 /*
600 * XXX maybe we should drop the packet here,
601 * as we could not provide enough information
602 * to the upper layers.
603 */
604 }
605 pserialize_read_exit(s);
606 }
607
608 /*
609 * Process Hop-by-Hop options header if it's contained.
610 * m may be modified in ip6_hopopts_input().
611 * If a JumboPayload option is included, plen will also be modified.
612 */
613 plen = (u_int32_t)ntohs(ip6->ip6_plen);
614 if (ip6->ip6_nxt == IPPROTO_HOPOPTS) {
615 struct ip6_hbh *hbh;
616
617 if (ip6_hopopts_input(&plen, &rtalert, &m, &off)) {
618#if 0 /*touches NULL pointer*/
619 in6_ifstat_inc(rcvif, ifs6_in_discard);
620#endif
621 return; /* m have already been freed */
622 }
623
624 /* adjust pointer */
625 ip6 = mtod(m, struct ip6_hdr *);
626
627 /*
628 * if the payload length field is 0 and the next header field
629 * indicates Hop-by-Hop Options header, then a Jumbo Payload
630 * option MUST be included.
631 */
632 if (ip6->ip6_plen == 0 && plen == 0) {
633 /*
634 * Note that if a valid jumbo payload option is
635 * contained, ip6_hopopts_input() must set a valid
636 * (non-zero) payload length to the variable plen.
637 */
638 IP6_STATINC(IP6_STAT_BADOPTIONS);
639 in6_ifstat_inc(rcvif, ifs6_in_discard);
640 in6_ifstat_inc(rcvif, ifs6_in_hdrerr);
641 icmp6_error(m, ICMP6_PARAM_PROB,
642 ICMP6_PARAMPROB_HEADER,
643 (char *)&ip6->ip6_plen - (char *)ip6);
644 return;
645 }
646 IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m, sizeof(struct ip6_hdr),
647 sizeof(struct ip6_hbh));
648 if (hbh == NULL) {
649 IP6_STATINC(IP6_STAT_TOOSHORT);
650 return;
651 }
652 KASSERT(IP6_HDR_ALIGNED_P(hbh));
653 nxt = hbh->ip6h_nxt;
654
655 /*
656 * accept the packet if a router alert option is included
657 * and we act as an IPv6 router.
658 */
659 if (rtalert != ~0 && ip6_forwarding)
660 ours = 1;
661 } else
662 nxt = ip6->ip6_nxt;
663
664 /*
665 * Check that the amount of data in the buffers
666 * is as at least much as the IPv6 header would have us expect.
667 * Trim mbufs if longer than we expect.
668 * Drop packet if shorter than we expect.
669 */
670 if (m->m_pkthdr.len - sizeof(struct ip6_hdr) < plen) {
671 IP6_STATINC(IP6_STAT_TOOSHORT);
672 in6_ifstat_inc(rcvif, ifs6_in_truncated);
673 goto bad;
674 }
675 if (m->m_pkthdr.len > sizeof(struct ip6_hdr) + plen) {
676 if (m->m_len == m->m_pkthdr.len) {
677 m->m_len = sizeof(struct ip6_hdr) + plen;
678 m->m_pkthdr.len = sizeof(struct ip6_hdr) + plen;
679 } else
680 m_adj(m, sizeof(struct ip6_hdr) + plen - m->m_pkthdr.len);
681 }
682
683 /*
684 * Forward if desirable.
685 */
686 if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) {
687 /*
688 * If we are acting as a multicast router, all
689 * incoming multicast packets are passed to the
690 * kernel-level multicast forwarding function.
691 * The packet is returned (relatively) intact; if
692 * ip6_mforward() returns a non-zero value, the packet
693 * must be discarded, else it may be accepted below.
694 */
695 if (ip6_mrouter != NULL) {
696 int error;
697
698 SOFTNET_LOCK();
699 error = ip6_mforward(ip6, rcvif, m);
700 SOFTNET_UNLOCK();
701
702 if (error != 0) {
703 IP6_STATINC(IP6_STAT_CANTFORWARD);
704 goto bad;
705 }
706 }
707 if (!ours)
708 goto bad;
709 } else if (!ours) {
710 ip6_forward(m, srcrt);
711 return;
712 }
713
714 ip6 = mtod(m, struct ip6_hdr *);
715
716 /*
717 * Malicious party may be able to use IPv4 mapped addr to confuse
718 * tcp/udp stack and bypass security checks (act as if it was from
719 * 127.0.0.1 by using IPv6 src ::ffff:127.0.0.1). Be cautious.
720 *
721 * For SIIT end node behavior, you may want to disable the check.
722 * However, you will become vulnerable to attacks using IPv4 mapped
723 * source.
724 */
725 if (IN6_IS_ADDR_V4MAPPED(&ip6->ip6_src) ||
726 IN6_IS_ADDR_V4MAPPED(&ip6->ip6_dst)) {
727 IP6_STATINC(IP6_STAT_BADSCOPE);
728 in6_ifstat_inc(rcvif, ifs6_in_addrerr);
729 goto bad;
730 }
731
732 /*
733 * Tell launch routine the next header
734 */
735#ifdef IFA_STATS
736 if (deliverifp != NULL) {
737 struct in6_ifaddr *ia6;
738 int s = pserialize_read_enter();
739 ia6 = in6_ifawithifp(deliverifp, &ip6->ip6_dst);
740 if (ia6)
741 ia6->ia_ifa.ifa_data.ifad_inbytes += m->m_pkthdr.len;
742 pserialize_read_exit(s);
743 }
744#endif
745 IP6_STATINC(IP6_STAT_DELIVERED);
746 in6_ifstat_inc(deliverifp, ifs6_in_deliver);
747 nest = 0;
748
749 rh_present = 0;
750 while (nxt != IPPROTO_DONE) {
751 if (ip6_hdrnestlimit && (++nest > ip6_hdrnestlimit)) {
752 IP6_STATINC(IP6_STAT_TOOMANYHDR);
753 in6_ifstat_inc(rcvif, ifs6_in_hdrerr);
754 goto bad;
755 }
756
757 /*
758 * protection against faulty packet - there should be
759 * more sanity checks in header chain processing.
760 */
761 if (m->m_pkthdr.len < off) {
762 IP6_STATINC(IP6_STAT_TOOSHORT);
763 in6_ifstat_inc(rcvif, ifs6_in_truncated);
764 goto bad;
765 }
766
767 if (nxt == IPPROTO_ROUTING) {
768 if (rh_present++) {
769 in6_ifstat_inc(rcvif, ifs6_in_hdrerr);
770 IP6_STATINC(IP6_STAT_BADOPTIONS);
771 goto bad;
772 }
773 }
774
775#ifdef IPSEC
776 if (ipsec_used) {
777 /*
778 * enforce IPsec policy checking if we are seeing last
779 * header. note that we do not visit this with
780 * protocols with pcb layer code - like udp/tcp/raw ip.
781 */
782 if ((inet6sw[ip_protox[nxt]].pr_flags
783 & PR_LASTHDR) != 0) {
784 int error;
785
786 SOFTNET_LOCK();
787 error = ipsec6_input(m);
788 SOFTNET_UNLOCK();
789 if (error)
790 goto bad;
791 }
792 }
793#endif /* IPSEC */
794
795 SOFTNET_LOCK();
796 nxt = (*inet6sw[ip6_protox[nxt]].pr_input)(&m, &off, nxt);
797 SOFTNET_UNLOCK();
798 }
799 return;
800 bad:
801 m_freem(m);
802}
803
804/*
805 * set/grab in6_ifaddr correspond to IPv6 destination address.
806 */
807static struct m_tag *
808ip6_setdstifaddr(struct mbuf *m, const struct in6_ifaddr *ia)
809{
810 struct m_tag *mtag;
811 struct ip6aux *ip6a;
812
813 mtag = ip6_addaux(m);
814 if (mtag == NULL)
815 return NULL;
816
817 ip6a = (struct ip6aux *)(mtag + 1);
818 if (in6_setscope(&ip6a->ip6a_src, ia->ia_ifp, &ip6a->ip6a_scope_id)) {
819 IP6_STATINC(IP6_STAT_BADSCOPE);
820 return NULL;
821 }
822
823 ip6a->ip6a_src = ia->ia_addr.sin6_addr;
824 ip6a->ip6a_flags = ia->ia6_flags;
825 return mtag;
826}
827
828const struct ip6aux *
829ip6_getdstifaddr(struct mbuf *m)
830{
831 struct m_tag *mtag;
832
833 mtag = ip6_findaux(m);
834 if (mtag != NULL)
835 return (struct ip6aux *)(mtag + 1);
836 else
837 return NULL;
838}
839
840/*
841 * Hop-by-Hop options header processing. If a valid jumbo payload option is
842 * included, the real payload length will be stored in plenp.
843 *
844 * rtalertp - XXX: should be stored more smart way
845 */
846int
847ip6_hopopts_input(u_int32_t *plenp, u_int32_t *rtalertp,
848 struct mbuf **mp, int *offp)
849{
850 struct mbuf *m = *mp;
851 int off = *offp, hbhlen;
852 struct ip6_hbh *hbh;
853
854 /* validation of the length of the header */
855 IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m,
856 sizeof(struct ip6_hdr), sizeof(struct ip6_hbh));
857 if (hbh == NULL) {
858 IP6_STATINC(IP6_STAT_TOOSHORT);
859 return -1;
860 }
861 hbhlen = (hbh->ip6h_len + 1) << 3;
862 IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m, sizeof(struct ip6_hdr),
863 hbhlen);
864 if (hbh == NULL) {
865 IP6_STATINC(IP6_STAT_TOOSHORT);
866 return -1;
867 }
868 KASSERT(IP6_HDR_ALIGNED_P(hbh));
869 off += hbhlen;
870 hbhlen -= sizeof(struct ip6_hbh);
871
872 if (ip6_process_hopopts(m, (u_int8_t *)hbh + sizeof(struct ip6_hbh),
873 hbhlen, rtalertp, plenp) < 0)
874 return (-1);
875
876 *offp = off;
877 *mp = m;
878 return (0);
879}
880
881/*
882 * Search header for all Hop-by-hop options and process each option.
883 * This function is separate from ip6_hopopts_input() in order to
884 * handle a case where the sending node itself process its hop-by-hop
885 * options header. In such a case, the function is called from ip6_output().
886 *
887 * The function assumes that hbh header is located right after the IPv6 header
888 * (RFC2460 p7), opthead is pointer into data content in m, and opthead to
889 * opthead + hbhlen is located in continuous memory region.
890 */
891static int
892ip6_process_hopopts(struct mbuf *m, u_int8_t *opthead, int hbhlen,
893 u_int32_t *rtalertp, u_int32_t *plenp)
894{
895 struct ip6_hdr *ip6;
896 int optlen = 0;
897 u_int8_t *opt = opthead;
898 u_int16_t rtalert_val;
899 u_int32_t jumboplen;
900 const int erroff = sizeof(struct ip6_hdr) + sizeof(struct ip6_hbh);
901
902 for (; hbhlen > 0; hbhlen -= optlen, opt += optlen) {
903 switch (*opt) {
904 case IP6OPT_PAD1:
905 optlen = 1;
906 break;
907 case IP6OPT_PADN:
908 if (hbhlen < IP6OPT_MINLEN) {
909 IP6_STATINC(IP6_STAT_TOOSMALL);
910 goto bad;
911 }
912 optlen = *(opt + 1) + 2;
913 break;
914 case IP6OPT_RTALERT:
915 /* XXX may need check for alignment */
916 if (hbhlen < IP6OPT_RTALERT_LEN) {
917 IP6_STATINC(IP6_STAT_TOOSMALL);
918 goto bad;
919 }
920 if (*(opt + 1) != IP6OPT_RTALERT_LEN - 2) {
921 /* XXX stat */
922 icmp6_error(m, ICMP6_PARAM_PROB,
923 ICMP6_PARAMPROB_HEADER,
924 erroff + opt + 1 - opthead);
925 return (-1);
926 }
927 optlen = IP6OPT_RTALERT_LEN;
928 memcpy((void *)&rtalert_val, (void *)(opt + 2), 2);
929 *rtalertp = ntohs(rtalert_val);
930 break;
931 case IP6OPT_JUMBO:
932 /* XXX may need check for alignment */
933 if (hbhlen < IP6OPT_JUMBO_LEN) {
934 IP6_STATINC(IP6_STAT_TOOSMALL);
935 goto bad;
936 }
937 if (*(opt + 1) != IP6OPT_JUMBO_LEN - 2) {
938 /* XXX stat */
939 icmp6_error(m, ICMP6_PARAM_PROB,
940 ICMP6_PARAMPROB_HEADER,
941 erroff + opt + 1 - opthead);
942 return (-1);
943 }
944 optlen = IP6OPT_JUMBO_LEN;
945
946 /*
947 * IPv6 packets that have non 0 payload length
948 * must not contain a jumbo payload option.
949 */
950 ip6 = mtod(m, struct ip6_hdr *);
951 if (ip6->ip6_plen) {
952 IP6_STATINC(IP6_STAT_BADOPTIONS);
953 icmp6_error(m, ICMP6_PARAM_PROB,
954 ICMP6_PARAMPROB_HEADER,
955 erroff + opt - opthead);
956 return (-1);
957 }
958
959 /*
960 * We may see jumbolen in unaligned location, so
961 * we'd need to perform bcopy().
962 */
963 memcpy(&jumboplen, opt + 2, sizeof(jumboplen));
964 jumboplen = (u_int32_t)htonl(jumboplen);
965
966#if 1
967 /*
968 * if there are multiple jumbo payload options,
969 * *plenp will be non-zero and the packet will be
970 * rejected.
971 * the behavior may need some debate in ipngwg -
972 * multiple options does not make sense, however,
973 * there's no explicit mention in specification.
974 */
975 if (*plenp != 0) {
976 IP6_STATINC(IP6_STAT_BADOPTIONS);
977 icmp6_error(m, ICMP6_PARAM_PROB,
978 ICMP6_PARAMPROB_HEADER,
979 erroff + opt + 2 - opthead);
980 return (-1);
981 }
982#endif
983
984 /*
985 * jumbo payload length must be larger than 65535.
986 */
987 if (jumboplen <= IPV6_MAXPACKET) {
988 IP6_STATINC(IP6_STAT_BADOPTIONS);
989 icmp6_error(m, ICMP6_PARAM_PROB,
990 ICMP6_PARAMPROB_HEADER,
991 erroff + opt + 2 - opthead);
992 return (-1);
993 }
994 *plenp = jumboplen;
995
996 break;
997 default: /* unknown option */
998 if (hbhlen < IP6OPT_MINLEN) {
999 IP6_STATINC(IP6_STAT_TOOSMALL);
1000 goto bad;
1001 }
1002 optlen = ip6_unknown_opt(opt, m,
1003 erroff + opt - opthead);
1004 if (optlen == -1)
1005 return (-1);
1006 optlen += 2;
1007 break;
1008 }
1009 }
1010
1011 return (0);
1012
1013 bad:
1014 m_freem(m);
1015 return (-1);
1016}
1017
1018/*
1019 * Unknown option processing.
1020 * The third argument `off' is the offset from the IPv6 header to the option,
1021 * which is necessary if the IPv6 header the and option header and IPv6 header
1022 * is not continuous in order to return an ICMPv6 error.
1023 */
1024int
1025ip6_unknown_opt(u_int8_t *optp, struct mbuf *m, int off)
1026{
1027 struct ip6_hdr *ip6;
1028
1029 switch (IP6OPT_TYPE(*optp)) {
1030 case IP6OPT_TYPE_SKIP: /* ignore the option */
1031 return ((int)*(optp + 1));
1032 case IP6OPT_TYPE_DISCARD: /* silently discard */
1033 m_freem(m);
1034 return (-1);
1035 case IP6OPT_TYPE_FORCEICMP: /* send ICMP even if multicasted */
1036 IP6_STATINC(IP6_STAT_BADOPTIONS);
1037 icmp6_error(m, ICMP6_PARAM_PROB, ICMP6_PARAMPROB_OPTION, off);
1038 return (-1);
1039 case IP6OPT_TYPE_ICMP: /* send ICMP if not multicasted */
1040 IP6_STATINC(IP6_STAT_BADOPTIONS);
1041 ip6 = mtod(m, struct ip6_hdr *);
1042 if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) ||
1043 (m->m_flags & (M_BCAST|M_MCAST)))
1044 m_freem(m);
1045 else
1046 icmp6_error(m, ICMP6_PARAM_PROB,
1047 ICMP6_PARAMPROB_OPTION, off);
1048 return (-1);
1049 }
1050
1051 m_freem(m); /* XXX: NOTREACHED */
1052 return (-1);
1053}
1054
1055/*
1056 * Create the "control" list for this pcb.
1057 *
1058 * The routine will be called from upper layer handlers like tcp6_input().
1059 * Thus the routine assumes that the caller (tcp6_input) have already
1060 * called IP6_EXTHDR_CHECK() and all the extension headers are located in the
1061 * very first mbuf on the mbuf chain.
1062 * We may want to add some infinite loop prevention or sanity checks for safety.
1063 * (This applies only when you are using KAME mbuf chain restriction, i.e.
1064 * you are using IP6_EXTHDR_CHECK() not m_pulldown())
1065 */
1066void
1067ip6_savecontrol(struct in6pcb *in6p, struct mbuf **mp,
1068 struct ip6_hdr *ip6, struct mbuf *m)
1069{
1070#ifdef RFC2292
1071#define IS2292(x, y) ((in6p->in6p_flags & IN6P_RFC2292) ? (x) : (y))
1072#else
1073#define IS2292(x, y) (y)
1074#endif
1075
1076 if (in6p->in6p_socket->so_options & SO_TIMESTAMP
1077#ifdef SO_OTIMESTAMP
1078 || in6p->in6p_socket->so_options & SO_OTIMESTAMP
1079#endif
1080 ) {
1081 struct timeval tv;
1082
1083 microtime(&tv);
1084#ifdef SO_OTIMESTAMP
1085 if (in6p->in6p_socket->so_options & SO_OTIMESTAMP) {
1086 struct timeval50 tv50;
1087 timeval_to_timeval50(&tv, &tv50);
1088 *mp = sbcreatecontrol((void *) &tv50, sizeof(tv50),
1089 SCM_OTIMESTAMP, SOL_SOCKET);
1090 } else
1091#endif
1092 *mp = sbcreatecontrol((void *) &tv, sizeof(tv),
1093 SCM_TIMESTAMP, SOL_SOCKET);
1094 if (*mp)
1095 mp = &(*mp)->m_next;
1096 }
1097
1098 /* some OSes call this logic with IPv4 packet, for SO_TIMESTAMP */
1099 if ((ip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION)
1100 return;
1101
1102 /* RFC 2292 sec. 5 */
1103 if ((in6p->in6p_flags & IN6P_PKTINFO) != 0) {
1104 struct in6_pktinfo pi6;
1105
1106 memcpy(&pi6.ipi6_addr, &ip6->ip6_dst, sizeof(struct in6_addr));
1107 in6_clearscope(&pi6.ipi6_addr); /* XXX */
1108 pi6.ipi6_ifindex = m->m_pkthdr.rcvif_index;
1109 *mp = sbcreatecontrol((void *) &pi6,
1110 sizeof(struct in6_pktinfo),
1111 IS2292(IPV6_2292PKTINFO, IPV6_PKTINFO), IPPROTO_IPV6);
1112 if (*mp)
1113 mp = &(*mp)->m_next;
1114 }
1115
1116 if (in6p->in6p_flags & IN6P_HOPLIMIT) {
1117 int hlim = ip6->ip6_hlim & 0xff;
1118
1119 *mp = sbcreatecontrol((void *) &hlim, sizeof(int),
1120 IS2292(IPV6_2292HOPLIMIT, IPV6_HOPLIMIT), IPPROTO_IPV6);
1121 if (*mp)
1122 mp = &(*mp)->m_next;
1123 }
1124
1125 if ((in6p->in6p_flags & IN6P_TCLASS) != 0) {
1126 u_int32_t flowinfo;
1127 int tclass;
1128
1129 flowinfo = (u_int32_t)ntohl(ip6->ip6_flow & IPV6_FLOWINFO_MASK);
1130 flowinfo >>= 20;
1131
1132 tclass = flowinfo & 0xff;
1133 *mp = sbcreatecontrol((void *)&tclass, sizeof(tclass),
1134 IPV6_TCLASS, IPPROTO_IPV6);
1135
1136 if (*mp)
1137 mp = &(*mp)->m_next;
1138 }
1139
1140 /*
1141 * IPV6_HOPOPTS socket option. Recall that we required super-user
1142 * privilege for the option (see ip6_ctloutput), but it might be too
1143 * strict, since there might be some hop-by-hop options which can be
1144 * returned to normal user.
1145 * See also RFC3542 section 8 (or RFC2292 section 6).
1146 */
1147 if ((in6p->in6p_flags & IN6P_HOPOPTS) != 0) {
1148 /*
1149 * Check if a hop-by-hop options header is contatined in the
1150 * received packet, and if so, store the options as ancillary
1151 * data. Note that a hop-by-hop options header must be
1152 * just after the IPv6 header, which fact is assured through
1153 * the IPv6 input processing.
1154 */
1155 struct ip6_hdr *xip6 = mtod(m, struct ip6_hdr *);
1156 if (xip6->ip6_nxt == IPPROTO_HOPOPTS) {
1157 struct ip6_hbh *hbh;
1158 int hbhlen;
1159 struct mbuf *ext;
1160
1161 ext = ip6_pullexthdr(m, sizeof(struct ip6_hdr),
1162 xip6->ip6_nxt);
1163 if (ext == NULL) {
1164 IP6_STATINC(IP6_STAT_TOOSHORT);
1165 return;
1166 }
1167 hbh = mtod(ext, struct ip6_hbh *);
1168 hbhlen = (hbh->ip6h_len + 1) << 3;
1169 if (hbhlen != ext->m_len) {
1170 m_freem(ext);
1171 IP6_STATINC(IP6_STAT_TOOSHORT);
1172 return;
1173 }
1174
1175 /*
1176 * XXX: We copy whole the header even if a jumbo
1177 * payload option is included, which option is to
1178 * be removed before returning in the RFC 2292.
1179 * Note: this constraint is removed in RFC3542.
1180 */
1181 *mp = sbcreatecontrol((void *)hbh, hbhlen,
1182 IS2292(IPV6_2292HOPOPTS, IPV6_HOPOPTS),
1183 IPPROTO_IPV6);
1184 if (*mp)
1185 mp = &(*mp)->m_next;
1186 m_freem(ext);
1187 }
1188 }
1189
1190 /* IPV6_DSTOPTS and IPV6_RTHDR socket options */
1191 if (in6p->in6p_flags & (IN6P_DSTOPTS | IN6P_RTHDR)) {
1192 struct ip6_hdr *xip6 = mtod(m, struct ip6_hdr *);
1193 int nxt = xip6->ip6_nxt, off = sizeof(struct ip6_hdr);
1194
1195 /*
1196 * Search for destination options headers or routing
1197 * header(s) through the header chain, and stores each
1198 * header as ancillary data.
1199 * Note that the order of the headers remains in
1200 * the chain of ancillary data.
1201 */
1202 for (;;) { /* is explicit loop prevention necessary? */
1203 struct ip6_ext *ip6e = NULL;
1204 int elen;
1205 struct mbuf *ext = NULL;
1206
1207 /*
1208 * if it is not an extension header, don't try to
1209 * pull it from the chain.
1210 */
1211 switch (nxt) {
1212 case IPPROTO_DSTOPTS:
1213 case IPPROTO_ROUTING:
1214 case IPPROTO_HOPOPTS:
1215 case IPPROTO_AH: /* is it possible? */
1216 break;
1217 default:
1218 goto loopend;
1219 }
1220
1221 ext = ip6_pullexthdr(m, off, nxt);
1222 if (ext == NULL) {
1223 IP6_STATINC(IP6_STAT_TOOSHORT);
1224 return;
1225 }
1226 ip6e = mtod(ext, struct ip6_ext *);
1227 if (nxt == IPPROTO_AH)
1228 elen = (ip6e->ip6e_len + 2) << 2;
1229 else
1230 elen = (ip6e->ip6e_len + 1) << 3;
1231 if (elen != ext->m_len) {
1232 m_freem(ext);
1233 IP6_STATINC(IP6_STAT_TOOSHORT);
1234 return;
1235 }
1236 KASSERT(IP6_HDR_ALIGNED_P(ip6e));
1237
1238 switch (nxt) {
1239 case IPPROTO_DSTOPTS:
1240 if (!(in6p->in6p_flags & IN6P_DSTOPTS))
1241 break;
1242
1243 *mp = sbcreatecontrol((void *)ip6e, elen,
1244 IS2292(IPV6_2292DSTOPTS, IPV6_DSTOPTS),
1245 IPPROTO_IPV6);
1246 if (*mp)
1247 mp = &(*mp)->m_next;
1248 break;
1249
1250 case IPPROTO_ROUTING:
1251 if (!(in6p->in6p_flags & IN6P_RTHDR))
1252 break;
1253
1254 *mp = sbcreatecontrol((void *)ip6e, elen,
1255 IS2292(IPV6_2292RTHDR, IPV6_RTHDR),
1256 IPPROTO_IPV6);
1257 if (*mp)
1258 mp = &(*mp)->m_next;
1259 break;
1260
1261 case IPPROTO_HOPOPTS:
1262 case IPPROTO_AH: /* is it possible? */
1263 break;
1264
1265 default:
1266 /*
1267 * other cases have been filtered in the above.
1268 * none will visit this case. here we supply
1269 * the code just in case (nxt overwritten or
1270 * other cases).
1271 */
1272 m_freem(ext);
1273 goto loopend;
1274
1275 }
1276
1277 /* proceed with the next header. */
1278 off += elen;
1279 nxt = ip6e->ip6e_nxt;
1280 ip6e = NULL;
1281 m_freem(ext);
1282 ext = NULL;
1283 }
1284 loopend:
1285 ;
1286 }
1287}
1288#undef IS2292
1289
1290
1291void
1292ip6_notify_pmtu(struct in6pcb *in6p, const struct sockaddr_in6 *dst,
1293 uint32_t *mtu)
1294{
1295 struct socket *so;
1296 struct mbuf *m_mtu;
1297 struct ip6_mtuinfo mtuctl;
1298
1299 so = in6p->in6p_socket;
1300
1301 if (mtu == NULL)
1302 return;
1303
1304#ifdef DIAGNOSTIC
1305 if (so == NULL) /* I believe this is impossible */
1306 panic("ip6_notify_pmtu: socket is NULL");
1307#endif
1308
1309 memset(&mtuctl, 0, sizeof(mtuctl)); /* zero-clear for safety */
1310 mtuctl.ip6m_mtu = *mtu;
1311 mtuctl.ip6m_addr = *dst;
1312 if (sa6_recoverscope(&mtuctl.ip6m_addr))
1313 return;
1314
1315 if ((m_mtu = sbcreatecontrol((void *)&mtuctl, sizeof(mtuctl),
1316 IPV6_PATHMTU, IPPROTO_IPV6)) == NULL)
1317 return;
1318
1319 if (sbappendaddr(&so->so_rcv, (const struct sockaddr *)dst, NULL, m_mtu)
1320 == 0) {
1321 m_freem(m_mtu);
1322 /* XXX: should count statistics */
1323 } else
1324 sorwakeup(so);
1325
1326 return;
1327}
1328
1329/*
1330 * pull single extension header from mbuf chain. returns single mbuf that
1331 * contains the result, or NULL on error.
1332 */
1333static struct mbuf *
1334ip6_pullexthdr(struct mbuf *m, size_t off, int nxt)
1335{
1336 struct ip6_ext ip6e;
1337 size_t elen;
1338 struct mbuf *n;
1339
1340#ifdef DIAGNOSTIC
1341 switch (nxt) {
1342 case IPPROTO_DSTOPTS:
1343 case IPPROTO_ROUTING:
1344 case IPPROTO_HOPOPTS:
1345 case IPPROTO_AH: /* is it possible? */
1346 break;
1347 default:
1348 printf("ip6_pullexthdr: invalid nxt=%d\n", nxt);
1349 }
1350#endif
1351
1352 m_copydata(m, off, sizeof(ip6e), (void *)&ip6e);
1353 if (nxt == IPPROTO_AH)
1354 elen = (ip6e.ip6e_len + 2) << 2;
1355 else
1356 elen = (ip6e.ip6e_len + 1) << 3;
1357
1358 MGET(n, M_DONTWAIT, MT_DATA);
1359 if (n && elen >= MLEN) {
1360 MCLGET(n, M_DONTWAIT);
1361 if ((n->m_flags & M_EXT) == 0) {
1362 m_free(n);
1363 n = NULL;
1364 }
1365 }
1366 if (!n)
1367 return NULL;
1368
1369 n->m_len = 0;
1370 if (elen >= M_TRAILINGSPACE(n)) {
1371 m_free(n);
1372 return NULL;
1373 }
1374
1375 m_copydata(m, off, elen, mtod(n, void *));
1376 n->m_len = elen;
1377 return n;
1378}
1379
1380/*
1381 * Get pointer to the previous header followed by the header
1382 * currently processed.
1383 * XXX: This function supposes that
1384 * M includes all headers,
1385 * the next header field and the header length field of each header
1386 * are valid, and
1387 * the sum of each header length equals to OFF.
1388 * Because of these assumptions, this function must be called very
1389 * carefully. Moreover, it will not be used in the near future when
1390 * we develop `neater' mechanism to process extension headers.
1391 */
1392u_int8_t *
1393ip6_get_prevhdr(struct mbuf *m, int off)
1394{
1395 struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *);
1396
1397 if (off == sizeof(struct ip6_hdr))
1398 return (&ip6->ip6_nxt);
1399 else {
1400 int len, nxt;
1401 struct ip6_ext *ip6e = NULL;
1402
1403 nxt = ip6->ip6_nxt;
1404 len = sizeof(struct ip6_hdr);
1405 while (len < off) {
1406 ip6e = (struct ip6_ext *)(mtod(m, char *) + len);
1407
1408 switch (nxt) {
1409 case IPPROTO_FRAGMENT:
1410 len += sizeof(struct ip6_frag);
1411 break;
1412 case IPPROTO_AH:
1413 len += (ip6e->ip6e_len + 2) << 2;
1414 break;
1415 default:
1416 len += (ip6e->ip6e_len + 1) << 3;
1417 break;
1418 }
1419 nxt = ip6e->ip6e_nxt;
1420 }
1421 if (ip6e)
1422 return (&ip6e->ip6e_nxt);
1423 else
1424 return NULL;
1425 }
1426}
1427
1428/*
1429 * get next header offset. m will be retained.
1430 */
1431int
1432ip6_nexthdr(struct mbuf *m, int off, int proto, int *nxtp)
1433{
1434 struct ip6_hdr ip6;
1435 struct ip6_ext ip6e;
1436 struct ip6_frag fh;
1437
1438 /* just in case */
1439 if (m == NULL)
1440 panic("ip6_nexthdr: m == NULL");
1441 if ((m->m_flags & M_PKTHDR) == 0 || m->m_pkthdr.len < off)
1442 return -1;
1443
1444 switch (proto) {
1445 case IPPROTO_IPV6:
1446 /* do not chase beyond intermediate IPv6 headers */
1447 if (off != 0)
1448 return -1;
1449 if (m->m_pkthdr.len < off + sizeof(ip6))
1450 return -1;
1451 m_copydata(m, off, sizeof(ip6), (void *)&ip6);
1452 if (nxtp)
1453 *nxtp = ip6.ip6_nxt;
1454 off += sizeof(ip6);
1455 return off;
1456
1457 case IPPROTO_FRAGMENT:
1458 /*
1459 * terminate parsing if it is not the first fragment,
1460 * it does not make sense to parse through it.
1461 */
1462 if (m->m_pkthdr.len < off + sizeof(fh))
1463 return -1;
1464 m_copydata(m, off, sizeof(fh), (void *)&fh);
1465 if ((fh.ip6f_offlg & IP6F_OFF_MASK) != 0)
1466 return -1;
1467 if (nxtp)
1468 *nxtp = fh.ip6f_nxt;
1469 off += sizeof(struct ip6_frag);
1470 return off;
1471
1472 case IPPROTO_AH:
1473 if (m->m_pkthdr.len < off + sizeof(ip6e))
1474 return -1;
1475 m_copydata(m, off, sizeof(ip6e), (void *)&ip6e);
1476 if (nxtp)
1477 *nxtp = ip6e.ip6e_nxt;
1478 off += (ip6e.ip6e_len + 2) << 2;
1479 if (m->m_pkthdr.len < off)
1480 return -1;
1481 return off;
1482
1483 case IPPROTO_HOPOPTS:
1484 case IPPROTO_ROUTING:
1485 case IPPROTO_DSTOPTS:
1486 if (m->m_pkthdr.len < off + sizeof(ip6e))
1487 return -1;
1488 m_copydata(m, off, sizeof(ip6e), (void *)&ip6e);
1489 if (nxtp)
1490 *nxtp = ip6e.ip6e_nxt;
1491 off += (ip6e.ip6e_len + 1) << 3;
1492 if (m->m_pkthdr.len < off)
1493 return -1;
1494 return off;
1495
1496 case IPPROTO_NONE:
1497 case IPPROTO_ESP:
1498 case IPPROTO_IPCOMP:
1499 /* give up */
1500 return -1;
1501
1502 default:
1503 return -1;
1504 }
1505}
1506
1507/*
1508 * get offset for the last header in the chain. m will be kept untainted.
1509 */
1510int
1511ip6_lasthdr(struct mbuf *m, int off, int proto, int *nxtp)
1512{
1513 int newoff;
1514 int nxt;
1515
1516 if (!nxtp) {
1517 nxt = -1;
1518 nxtp = &nxt;
1519 }
1520 for (;;) {
1521 newoff = ip6_nexthdr(m, off, proto, nxtp);
1522 if (newoff < 0)
1523 return off;
1524 else if (newoff < off)
1525 return -1; /* invalid */
1526 else if (newoff == off)
1527 return newoff;
1528
1529 off = newoff;
1530 proto = *nxtp;
1531 }
1532}
1533
1534struct m_tag *
1535ip6_addaux(struct mbuf *m)
1536{
1537 struct m_tag *mtag;
1538
1539 mtag = m_tag_find(m, PACKET_TAG_INET6, NULL);
1540 if (!mtag) {
1541 mtag = m_tag_get(PACKET_TAG_INET6, sizeof(struct ip6aux),
1542 M_NOWAIT);
1543 if (mtag) {
1544 m_tag_prepend(m, mtag);
1545 memset(mtag + 1, 0, sizeof(struct ip6aux));
1546 }
1547 }
1548 return mtag;
1549}
1550
1551struct m_tag *
1552ip6_findaux(struct mbuf *m)
1553{
1554 struct m_tag *mtag;
1555
1556 mtag = m_tag_find(m, PACKET_TAG_INET6, NULL);
1557 return mtag;
1558}
1559
1560void
1561ip6_delaux(struct mbuf *m)
1562{
1563 struct m_tag *mtag;
1564
1565 mtag = m_tag_find(m, PACKET_TAG_INET6, NULL);
1566 if (mtag)
1567 m_tag_delete(m, mtag);
1568}
1569
1570/*
1571 * System control for IP6
1572 */
1573
1574const u_char inet6ctlerrmap[PRC_NCMDS] = {
1575 0, 0, 0, 0,
1576 0, EMSGSIZE, EHOSTDOWN, EHOSTUNREACH,
1577 EHOSTUNREACH, EHOSTUNREACH, ECONNREFUSED, ECONNREFUSED,
1578 EMSGSIZE, EHOSTUNREACH, 0, 0,
1579 0, 0, 0, 0,
1580 ENOPROTOOPT
1581};
1582
1583extern int sysctl_net_inet6_addrctlpolicy(SYSCTLFN_ARGS);
1584
1585static int
1586sysctl_net_inet6_ip6_stats(SYSCTLFN_ARGS)
1587{
1588
1589 return (NETSTAT_SYSCTL(ip6stat_percpu, IP6_NSTATS));
1590}
1591
1592static void
1593sysctl_net_inet6_ip6_setup(struct sysctllog **clog)
1594{
1595#ifdef RFC2292
1596#define IS2292(x, y) ((in6p->in6p_flags & IN6P_RFC2292) ? (x) : (y))
1597#else
1598#define IS2292(x, y) (y)
1599#endif
1600
1601 sysctl_createv(clog, 0, NULL, NULL,
1602 CTLFLAG_PERMANENT,
1603 CTLTYPE_NODE, "inet6",
1604 SYSCTL_DESCR("PF_INET6 related settings"),
1605 NULL, 0, NULL, 0,
1606 CTL_NET, PF_INET6, CTL_EOL);
1607 sysctl_createv(clog, 0, NULL, NULL,
1608 CTLFLAG_PERMANENT,
1609 CTLTYPE_NODE, "ip6",
1610 SYSCTL_DESCR("IPv6 related settings"),
1611 NULL, 0, NULL, 0,
1612 CTL_NET, PF_INET6, IPPROTO_IPV6, CTL_EOL);
1613
1614 sysctl_createv(clog, 0, NULL, NULL,
1615 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1616 CTLTYPE_INT, "forwarding",
1617 SYSCTL_DESCR("Enable forwarding of INET6 datagrams"),
1618 NULL, 0, &ip6_forwarding, 0,
1619 CTL_NET, PF_INET6, IPPROTO_IPV6,
1620 IPV6CTL_FORWARDING, CTL_EOL);
1621 sysctl_createv(clog, 0, NULL, NULL,
1622 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1623 CTLTYPE_INT, "redirect",
1624 SYSCTL_DESCR("Enable sending of ICMPv6 redirect messages"),
1625 NULL, 0, &ip6_sendredirects, 0,
1626 CTL_NET, PF_INET6, IPPROTO_IPV6,
1627 IPV6CTL_SENDREDIRECTS, CTL_EOL);
1628 sysctl_createv(clog, 0, NULL, NULL,
1629 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1630 CTLTYPE_INT, "hlim",
1631 SYSCTL_DESCR("Hop limit for an INET6 datagram"),
1632 NULL, 0, &ip6_defhlim, 0,
1633 CTL_NET, PF_INET6, IPPROTO_IPV6,
1634 IPV6CTL_DEFHLIM, CTL_EOL);
1635#ifdef notyet
1636 sysctl_createv(clog, 0, NULL, NULL,
1637 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1638 CTLTYPE_INT, "mtu", NULL,
1639 NULL, 0, &, 0,
1640 CTL_NET, PF_INET6, IPPROTO_IPV6,
1641 IPV6CTL_DEFMTU, CTL_EOL);
1642#endif
1643#ifdef __no_idea__
1644 sysctl_createv(clog, 0, NULL, NULL,
1645 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1646 CTLTYPE_INT, "forwsrcrt", NULL,
1647 NULL, 0, &?, 0,
1648 CTL_NET, PF_INET6, IPPROTO_IPV6,
1649 IPV6CTL_FORWSRCRT, CTL_EOL);
1650 sysctl_createv(clog, 0, NULL, NULL,
1651 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1652 CTLTYPE_STRUCT, "mrtstats", NULL,
1653 NULL, 0, &?, sizeof(?),
1654 CTL_NET, PF_INET6, IPPROTO_IPV6,
1655 IPV6CTL_MRTSTATS, CTL_EOL);
1656 sysctl_createv(clog, 0, NULL, NULL,
1657 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1658 CTLTYPE_?, "mrtproto", NULL,
1659 NULL, 0, &?, sizeof(?),
1660 CTL_NET, PF_INET6, IPPROTO_IPV6,
1661 IPV6CTL_MRTPROTO, CTL_EOL);
1662#endif
1663 sysctl_createv(clog, 0, NULL, NULL,
1664 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1665 CTLTYPE_INT, "maxfragpackets",
1666 SYSCTL_DESCR("Maximum number of fragments to buffer "
1667 "for reassembly"),
1668 NULL, 0, &ip6_maxfragpackets, 0,
1669 CTL_NET, PF_INET6, IPPROTO_IPV6,
1670 IPV6CTL_MAXFRAGPACKETS, CTL_EOL);
1671#ifdef __no_idea__
1672 sysctl_createv(clog, 0, NULL, NULL,
1673 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1674 CTLTYPE_INT, "sourcecheck", NULL,
1675 NULL, 0, &?, 0,
1676 CTL_NET, PF_INET6, IPPROTO_IPV6,
1677 IPV6CTL_SOURCECHECK, CTL_EOL);
1678 sysctl_createv(clog, 0, NULL, NULL,
1679 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1680 CTLTYPE_INT, "sourcecheck_logint", NULL,
1681 NULL, 0, &?, 0,
1682 CTL_NET, PF_INET6, IPPROTO_IPV6,
1683 IPV6CTL_SOURCECHECK_LOGINT, CTL_EOL);
1684#endif
1685 sysctl_createv(clog, 0, NULL, NULL,
1686 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1687 CTLTYPE_INT, "accept_rtadv",
1688 SYSCTL_DESCR("Accept router advertisements"),
1689 NULL, 0, &ip6_accept_rtadv, 0,
1690 CTL_NET, PF_INET6, IPPROTO_IPV6,
1691 IPV6CTL_ACCEPT_RTADV, CTL_EOL);
1692 sysctl_createv(clog, 0, NULL, NULL,
1693 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1694 CTLTYPE_INT, "rtadv_maxroutes",
1695 SYSCTL_DESCR("Maximum number of routes accepted via router advertisements"),
1696 NULL, 0, &ip6_rtadv_maxroutes, 0,
1697 CTL_NET, PF_INET6, IPPROTO_IPV6,
1698 IPV6CTL_RTADV_MAXROUTES, CTL_EOL);
1699 sysctl_createv(clog, 0, NULL, NULL,
1700 CTLFLAG_PERMANENT,
1701 CTLTYPE_INT, "rtadv_numroutes",
1702 SYSCTL_DESCR("Current number of routes accepted via router advertisements"),
1703 NULL, 0, &nd6_numroutes, 0,
1704 CTL_NET, PF_INET6, IPPROTO_IPV6,
1705 IPV6CTL_RTADV_NUMROUTES, CTL_EOL);
1706 sysctl_createv(clog, 0, NULL, NULL,
1707 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1708 CTLTYPE_INT, "keepfaith",
1709 SYSCTL_DESCR("Activate faith interface"),
1710 NULL, 0, &ip6_keepfaith, 0,
1711 CTL_NET, PF_INET6, IPPROTO_IPV6,
1712 IPV6CTL_KEEPFAITH, CTL_EOL);
1713 sysctl_createv(clog, 0, NULL, NULL,
1714 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1715 CTLTYPE_INT, "log_interval",
1716 SYSCTL_DESCR("Minumum interval between logging "
1717 "unroutable packets"),
1718 NULL, 0, &ip6_log_interval, 0,
1719 CTL_NET, PF_INET6, IPPROTO_IPV6,
1720 IPV6CTL_LOG_INTERVAL, CTL_EOL);
1721 sysctl_createv(clog, 0, NULL, NULL,
1722 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1723 CTLTYPE_INT, "hdrnestlimit",
1724 SYSCTL_DESCR("Maximum number of nested IPv6 headers"),
1725 NULL, 0, &ip6_hdrnestlimit, 0,
1726 CTL_NET, PF_INET6, IPPROTO_IPV6,
1727 IPV6CTL_HDRNESTLIMIT, CTL_EOL);
1728 sysctl_createv(clog, 0, NULL, NULL,
1729 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1730 CTLTYPE_INT, "dad_count",
1731 SYSCTL_DESCR("Number of Duplicate Address Detection "
1732 "probes to send"),
1733 NULL, 0, &ip6_dad_count, 0,
1734 CTL_NET, PF_INET6, IPPROTO_IPV6,
1735 IPV6CTL_DAD_COUNT, CTL_EOL);
1736 sysctl_createv(clog, 0, NULL, NULL,
1737 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1738 CTLTYPE_INT, "auto_flowlabel",
1739 SYSCTL_DESCR("Assign random IPv6 flow labels"),
1740 NULL, 0, &ip6_auto_flowlabel, 0,
1741 CTL_NET, PF_INET6, IPPROTO_IPV6,
1742 IPV6CTL_AUTO_FLOWLABEL, CTL_EOL);
1743 sysctl_createv(clog, 0, NULL, NULL,
1744 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1745 CTLTYPE_INT, "defmcasthlim",
1746 SYSCTL_DESCR("Default multicast hop limit"),
1747 NULL, 0, &ip6_defmcasthlim, 0,
1748 CTL_NET, PF_INET6, IPPROTO_IPV6,
1749 IPV6CTL_DEFMCASTHLIM, CTL_EOL);
1750 sysctl_createv(clog, 0, NULL, NULL,
1751 CTLFLAG_PERMANENT,
1752 CTLTYPE_STRING, "kame_version",
1753 SYSCTL_DESCR("KAME Version"),
1754 NULL, 0, __UNCONST(__KAME_VERSION), 0,
1755 CTL_NET, PF_INET6, IPPROTO_IPV6,
1756 IPV6CTL_KAME_VERSION, CTL_EOL);
1757 sysctl_createv(clog, 0, NULL, NULL,
1758 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1759 CTLTYPE_INT, "use_deprecated",
1760 SYSCTL_DESCR("Allow use of deprecated addresses as "
1761 "source addresses"),
1762 NULL, 0, &ip6_use_deprecated, 0,
1763 CTL_NET, PF_INET6, IPPROTO_IPV6,
1764 IPV6CTL_USE_DEPRECATED, CTL_EOL);
1765 sysctl_createv(clog, 0, NULL, NULL,
1766 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1767 CTLTYPE_INT, "rr_prune", NULL,
1768 NULL, 0, &ip6_rr_prune, 0,
1769 CTL_NET, PF_INET6, IPPROTO_IPV6,
1770 IPV6CTL_RR_PRUNE, CTL_EOL);
1771 sysctl_createv(clog, 0, NULL, NULL,
1772 CTLFLAG_PERMANENT
1773#ifndef INET6_BINDV6ONLY
1774 |CTLFLAG_READWRITE,
1775#endif
1776 CTLTYPE_INT, "v6only",
1777 SYSCTL_DESCR("Disallow PF_INET6 sockets from connecting "
1778 "to PF_INET sockets"),
1779 NULL, 0, &ip6_v6only, 0,
1780 CTL_NET, PF_INET6, IPPROTO_IPV6,
1781 IPV6CTL_V6ONLY, CTL_EOL);
1782 sysctl_createv(clog, 0, NULL, NULL,
1783 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1784 CTLTYPE_INT, "anonportmin",
1785 SYSCTL_DESCR("Lowest ephemeral port number to assign"),
1786 sysctl_net_inet_ip_ports, 0, &ip6_anonportmin, 0,
1787 CTL_NET, PF_INET6, IPPROTO_IPV6,
1788 IPV6CTL_ANONPORTMIN, CTL_EOL);
1789 sysctl_createv(clog, 0, NULL, NULL,
1790 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1791 CTLTYPE_INT, "anonportmax",
1792 SYSCTL_DESCR("Highest ephemeral port number to assign"),
1793 sysctl_net_inet_ip_ports, 0, &ip6_anonportmax, 0,
1794 CTL_NET, PF_INET6, IPPROTO_IPV6,
1795 IPV6CTL_ANONPORTMAX, CTL_EOL);
1796#ifndef IPNOPRIVPORTS
1797 sysctl_createv(clog, 0, NULL, NULL,
1798 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1799 CTLTYPE_INT, "lowportmin",
1800 SYSCTL_DESCR("Lowest privileged ephemeral port number "
1801 "to assign"),
1802 sysctl_net_inet_ip_ports, 0, &ip6_lowportmin, 0,
1803 CTL_NET, PF_INET6, IPPROTO_IPV6,
1804 IPV6CTL_LOWPORTMIN, CTL_EOL);
1805 sysctl_createv(clog, 0, NULL, NULL,
1806 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1807 CTLTYPE_INT, "lowportmax",
1808 SYSCTL_DESCR("Highest privileged ephemeral port number "
1809 "to assign"),
1810 sysctl_net_inet_ip_ports, 0, &ip6_lowportmax, 0,
1811 CTL_NET, PF_INET6, IPPROTO_IPV6,
1812 IPV6CTL_LOWPORTMAX, CTL_EOL);
1813#endif /* IPNOPRIVPORTS */
1814 sysctl_createv(clog, 0, NULL, NULL,
1815 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1816 CTLTYPE_INT, "auto_linklocal",
1817 SYSCTL_DESCR("Default value of per-interface flag for "
1818 "adding an IPv6 link-local address to "
1819 "interfaces when attached"),
1820 NULL, 0, &ip6_auto_linklocal, 0,
1821 CTL_NET, PF_INET6, IPPROTO_IPV6,
1822 IPV6CTL_AUTO_LINKLOCAL, CTL_EOL);
1823 sysctl_createv(clog, 0, NULL, NULL,
1824 CTLFLAG_PERMANENT|CTLFLAG_READONLY,
1825 CTLTYPE_STRUCT, "addctlpolicy",
1826 SYSCTL_DESCR("Return the current address control"
1827 " policy"),
1828 sysctl_net_inet6_addrctlpolicy, 0, NULL, 0,
1829 CTL_NET, PF_INET6, IPPROTO_IPV6,
1830 IPV6CTL_ADDRCTLPOLICY, CTL_EOL);
1831 sysctl_createv(clog, 0, NULL, NULL,
1832 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1833 CTLTYPE_INT, "use_tempaddr",
1834 SYSCTL_DESCR("Use temporary address"),
1835 NULL, 0, &ip6_use_tempaddr, 0,
1836 CTL_NET, PF_INET6, IPPROTO_IPV6,
1837 CTL_CREATE, CTL_EOL);
1838 sysctl_createv(clog, 0, NULL, NULL,
1839 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1840 CTLTYPE_INT, "prefer_tempaddr",
1841 SYSCTL_DESCR("Prefer temporary address as source "
1842 "address"),
1843 NULL, 0, &ip6_prefer_tempaddr, 0,
1844 CTL_NET, PF_INET6, IPPROTO_IPV6,
1845 CTL_CREATE, CTL_EOL);
1846 sysctl_createv(clog, 0, NULL, NULL,
1847 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1848 CTLTYPE_INT, "temppltime",
1849 SYSCTL_DESCR("preferred lifetime of a temporary address"),
1850 NULL, 0, &ip6_temp_preferred_lifetime, 0,
1851 CTL_NET, PF_INET6, IPPROTO_IPV6,
1852 CTL_CREATE, CTL_EOL);
1853 sysctl_createv(clog, 0, NULL, NULL,
1854 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1855 CTLTYPE_INT, "tempvltime",
1856 SYSCTL_DESCR("valid lifetime of a temporary address"),
1857 NULL, 0, &ip6_temp_valid_lifetime, 0,
1858 CTL_NET, PF_INET6, IPPROTO_IPV6,
1859 CTL_CREATE, CTL_EOL);
1860 sysctl_createv(clog, 0, NULL, NULL,
1861 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1862 CTLTYPE_INT, "maxfrags",
1863 SYSCTL_DESCR("Maximum fragments in reassembly queue"),
1864 NULL, 0, &ip6_maxfrags, 0,
1865 CTL_NET, PF_INET6, IPPROTO_IPV6,
1866 IPV6CTL_MAXFRAGS, CTL_EOL);
1867 sysctl_createv(clog, 0, NULL, NULL,
1868 CTLFLAG_PERMANENT,
1869 CTLTYPE_STRUCT, "stats",
1870 SYSCTL_DESCR("IPv6 statistics"),
1871 sysctl_net_inet6_ip6_stats, 0, NULL, 0,
1872 CTL_NET, PF_INET6, IPPROTO_IPV6,
1873 IPV6CTL_STATS, CTL_EOL);
1874 sysctl_createv(clog, 0, NULL, NULL,
1875 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1876 CTLTYPE_INT, "use_defaultzone",
1877 SYSCTL_DESCR("Whether to use the default scope zones"),
1878 NULL, 0, &ip6_use_defzone, 0,
1879 CTL_NET, PF_INET6, IPPROTO_IPV6,
1880 IPV6CTL_USE_DEFAULTZONE, CTL_EOL);
1881 sysctl_createv(clog, 0, NULL, NULL,
1882 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1883 CTLTYPE_INT, "mcast_pmtu",
1884 SYSCTL_DESCR("Enable pMTU discovery for multicast packet"),
1885 NULL, 0, &ip6_mcast_pmtu, 0,
1886 CTL_NET, PF_INET6, IPPROTO_IPV6,
1887 CTL_CREATE, CTL_EOL);
1888 /* anonportalgo RFC6056 subtree */
1889 const struct sysctlnode *portalgo_node;
1890 sysctl_createv(clog, 0, NULL, &portalgo_node,
1891 CTLFLAG_PERMANENT,
1892 CTLTYPE_NODE, "anonportalgo",
1893 SYSCTL_DESCR("Anonymous port algorithm selection (RFC 6056)"),
1894 NULL, 0, NULL, 0,
1895 CTL_NET, PF_INET6, IPPROTO_IPV6, CTL_CREATE, CTL_EOL);
1896 sysctl_createv(clog, 0, &portalgo_node, NULL,
1897 CTLFLAG_PERMANENT,
1898 CTLTYPE_STRING, "available",
1899 SYSCTL_DESCR("available algorithms"),
1900 sysctl_portalgo_available, 0, NULL, PORTALGO_MAXLEN,
1901 CTL_CREATE, CTL_EOL);
1902 sysctl_createv(clog, 0, &portalgo_node, NULL,
1903 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1904 CTLTYPE_STRING, "selected",
1905 SYSCTL_DESCR("selected algorithm"),
1906 sysctl_portalgo_selected6, 0, NULL, PORTALGO_MAXLEN,
1907 CTL_CREATE, CTL_EOL);
1908 sysctl_createv(clog, 0, &portalgo_node, NULL,
1909 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1910 CTLTYPE_STRUCT, "reserve",
1911 SYSCTL_DESCR("bitmap of reserved ports"),
1912 sysctl_portalgo_reserve6, 0, NULL, 0,
1913 CTL_CREATE, CTL_EOL);
1914 sysctl_createv(clog, 0, NULL, NULL,
1915 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1916 CTLTYPE_INT, "neighborgcthresh",
1917 SYSCTL_DESCR("Maximum number of entries in neighbor"
1918 " cache"),
1919 NULL, 1, &ip6_neighborgcthresh, 0,
1920 CTL_NET, PF_INET6, IPPROTO_IPV6,
1921 CTL_CREATE, CTL_EOL);
1922 sysctl_createv(clog, 0, NULL, NULL,
1923 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1924 CTLTYPE_INT, "maxifprefixes",
1925 SYSCTL_DESCR("Maximum number of prefixes created by"
1926 " route advertisement per interface"),
1927 NULL, 1, &ip6_maxifprefixes, 0,
1928 CTL_NET, PF_INET6, IPPROTO_IPV6,
1929 CTL_CREATE, CTL_EOL);
1930 sysctl_createv(clog, 0, NULL, NULL,
1931 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1932 CTLTYPE_INT, "maxifdefrouters",
1933 SYSCTL_DESCR("Maximum number of default routers created"
1934 " by route advertisement per interface"),
1935 NULL, 1, &ip6_maxifdefrouters, 0,
1936 CTL_NET, PF_INET6, IPPROTO_IPV6,
1937 CTL_CREATE, CTL_EOL);
1938 sysctl_createv(clog, 0, NULL, NULL,
1939 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1940 CTLTYPE_INT, "maxdynroutes",
1941 SYSCTL_DESCR("Maximum number of routes created via"
1942 " redirect"),
1943 NULL, 1, &ip6_maxdynroutes, 0,
1944 CTL_NET, PF_INET6, IPPROTO_IPV6,
1945 CTL_CREATE, CTL_EOL);
1946}
1947
1948void
1949ip6_statinc(u_int stat)
1950{
1951
1952 KASSERT(stat < IP6_NSTATS);
1953 IP6_STATINC(stat);
1954}
1955