ip6_input.c revision 1.186 1 /* $NetBSD: ip6_input.c,v 1.186 2018/01/29 10:57:13 maxv 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 1.186 2018/01/29 10:57:13 maxv 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_net_mpsafe.h"
73 #endif
74
75 #include <sys/param.h>
76 #include <sys/systm.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 #include <sys/percpu.h>
90
91 #include <net/if.h>
92 #include <net/if_types.h>
93 #include <net/if_dl.h>
94 #include <net/route.h>
95 #include <net/pktqueue.h>
96 #include <net/pfil.h>
97
98 #include <netinet/in.h>
99 #include <netinet/in_systm.h>
100 #ifdef INET
101 #include <netinet/ip.h>
102 #include <netinet/ip_var.h>
103 #include <netinet/ip_icmp.h>
104 #endif /* INET */
105 #include <netinet/ip6.h>
106 #include <netinet/portalgo.h>
107 #include <netinet6/in6_var.h>
108 #include <netinet6/ip6_var.h>
109 #include <netinet6/ip6_private.h>
110 #include <netinet6/in6_pcb.h>
111 #include <netinet/icmp6.h>
112 #include <netinet6/scope6_var.h>
113 #include <netinet6/in6_ifattach.h>
114 #include <netinet6/nd6.h>
115
116 #ifdef IPSEC
117 #include <netipsec/ipsec.h>
118 #include <netipsec/ipsec6.h>
119 #include <netipsec/key.h>
120 #endif /* IPSEC */
121
122 #include <netinet6/ip6protosw.h>
123
124 #include "faith.h"
125
126 #include <net/net_osdep.h>
127
128 extern struct domain inet6domain;
129
130 u_char ip6_protox[IPPROTO_MAX];
131 pktqueue_t *ip6_pktq __read_mostly;
132
133 pfil_head_t *inet6_pfil_hook;
134
135 percpu_t *ip6stat_percpu;
136
137 percpu_t *ip6_forward_rt_percpu __cacheline_aligned;
138
139 static void ip6_init2(void);
140 static void ip6intr(void *);
141 static struct m_tag *ip6_setdstifaddr(struct mbuf *, const struct in6_ifaddr *);
142
143 static int ip6_process_hopopts(struct mbuf *, u_int8_t *, int, u_int32_t *,
144 u_int32_t *);
145 static struct mbuf *ip6_pullexthdr(struct mbuf *, size_t, int);
146 static void sysctl_net_inet6_ip6_setup(struct sysctllog **);
147
148 #ifdef NET_MPSAFE
149 #define SOFTNET_LOCK() mutex_enter(softnet_lock)
150 #define SOFTNET_UNLOCK() mutex_exit(softnet_lock)
151 #else
152 #define SOFTNET_LOCK() KASSERT(mutex_owned(softnet_lock))
153 #define SOFTNET_UNLOCK() KASSERT(mutex_owned(softnet_lock))
154 #endif
155
156 /*
157 * IP6 initialization: fill in IP6 protocol switch table.
158 * All protocols not implemented in kernel go to raw IP6 protocol handler.
159 */
160 void
161 ip6_init(void)
162 {
163 const struct ip6protosw *pr;
164 int i;
165
166 in6_init();
167
168 sysctl_net_inet6_ip6_setup(NULL);
169 pr = (const struct ip6protosw *)pffindproto(PF_INET6, IPPROTO_RAW, SOCK_RAW);
170 if (pr == 0)
171 panic("ip6_init");
172 for (i = 0; i < IPPROTO_MAX; i++)
173 ip6_protox[i] = pr - inet6sw;
174 for (pr = (const struct ip6protosw *)inet6domain.dom_protosw;
175 pr < (const struct ip6protosw *)inet6domain.dom_protoswNPROTOSW; pr++)
176 if (pr->pr_domain->dom_family == PF_INET6 &&
177 pr->pr_protocol && pr->pr_protocol != IPPROTO_RAW)
178 ip6_protox[pr->pr_protocol] = pr - inet6sw;
179
180 ip6_pktq = pktq_create(IFQ_MAXLEN, ip6intr, NULL);
181 KASSERT(ip6_pktq != NULL);
182
183 scope6_init();
184 addrsel_policy_init();
185 nd6_init();
186 frag6_init();
187 ip6_desync_factor = cprng_fast32() % MAX_TEMP_DESYNC_FACTOR;
188
189 ip6_init2();
190 #ifdef GATEWAY
191 ip6flow_init(ip6_hashsize);
192 #endif
193 /* Register our Packet Filter hook. */
194 inet6_pfil_hook = pfil_head_create(PFIL_TYPE_AF, (void *)AF_INET6);
195 KASSERT(inet6_pfil_hook != NULL);
196
197 ip6stat_percpu = percpu_alloc(sizeof(uint64_t) * IP6_NSTATS);
198 ip6_forward_rt_percpu = percpu_alloc(sizeof(struct route));
199 }
200
201 static void
202 ip6_init2(void)
203 {
204
205 /* timer for regeneration of temporary addresses randomize ID */
206 callout_init(&in6_tmpaddrtimer_ch, CALLOUT_MPSAFE);
207 callout_reset(&in6_tmpaddrtimer_ch,
208 (ip6_temp_preferred_lifetime - ip6_desync_factor -
209 ip6_temp_regen_advance) * hz,
210 in6_tmpaddrtimer, NULL);
211 }
212
213 /*
214 * IP6 input interrupt handling. Just pass the packet to ip6_input.
215 */
216 static void
217 ip6intr(void *arg __unused)
218 {
219 struct mbuf *m;
220
221 SOFTNET_LOCK_UNLESS_NET_MPSAFE();
222 while ((m = pktq_dequeue(ip6_pktq)) != NULL) {
223 struct psref psref;
224 struct ifnet *rcvif = m_get_rcvif_psref(m, &psref);
225
226 if (rcvif == NULL) {
227 m_freem(m);
228 continue;
229 }
230 /*
231 * Drop the packet if IPv6 is disabled on the interface.
232 */
233 if ((ND_IFINFO(rcvif)->flags & ND6_IFF_IFDISABLED)) {
234 m_put_rcvif_psref(rcvif, &psref);
235 m_freem(m);
236 continue;
237 }
238 ip6_input(m, rcvif);
239 m_put_rcvif_psref(rcvif, &psref);
240 }
241 SOFTNET_UNLOCK_UNLESS_NET_MPSAFE();
242 }
243
244 void
245 ip6_input(struct mbuf *m, struct ifnet *rcvif)
246 {
247 struct ip6_hdr *ip6;
248 int hit, off = sizeof(struct ip6_hdr), nest;
249 u_int32_t plen;
250 u_int32_t rtalert = ~0;
251 int nxt, ours = 0, rh_present = 0;
252 struct ifnet *deliverifp = NULL;
253 int srcrt = 0;
254 struct rtentry *rt = NULL;
255 union {
256 struct sockaddr dst;
257 struct sockaddr_in6 dst6;
258 } u;
259 struct route *ro;
260
261 KASSERT(rcvif != NULL);
262
263 /*
264 * make sure we don't have onion peering information into m_tag.
265 */
266 ip6_delaux(m);
267
268 /*
269 * mbuf statistics
270 */
271 if (m->m_flags & M_EXT) {
272 if (m->m_next)
273 IP6_STATINC(IP6_STAT_MEXT2M);
274 else
275 IP6_STATINC(IP6_STAT_MEXT1);
276 } else {
277 #define M2MMAX 32
278 if (m->m_next) {
279 if (m->m_flags & M_LOOP)
280 /*XXX*/ IP6_STATINC(IP6_STAT_M2M + lo0ifp->if_index);
281 else if (rcvif->if_index < M2MMAX)
282 IP6_STATINC(IP6_STAT_M2M + rcvif->if_index);
283 else
284 IP6_STATINC(IP6_STAT_M2M);
285 } else
286 IP6_STATINC(IP6_STAT_M1);
287 #undef M2MMAX
288 }
289
290 in6_ifstat_inc(rcvif, ifs6_in_receive);
291 IP6_STATINC(IP6_STAT_TOTAL);
292
293 /*
294 * If the IPv6 header is not aligned, slurp it up into a new
295 * mbuf with space for link headers, in the event we forward
296 * it. Otherwise, if it is aligned, make sure the entire base
297 * IPv6 header is in the first mbuf of the chain.
298 */
299 if (IP6_HDR_ALIGNED_P(mtod(m, void *)) == 0) {
300 if ((m = m_copyup(m, sizeof(struct ip6_hdr),
301 (max_linkhdr + 3) & ~3)) == NULL) {
302 /* XXXJRT new stat, please */
303 IP6_STATINC(IP6_STAT_TOOSMALL);
304 in6_ifstat_inc(rcvif, ifs6_in_hdrerr);
305 return;
306 }
307 } else if (__predict_false(m->m_len < sizeof(struct ip6_hdr))) {
308 if ((m = m_pullup(m, sizeof(struct ip6_hdr))) == NULL) {
309 IP6_STATINC(IP6_STAT_TOOSMALL);
310 in6_ifstat_inc(rcvif, ifs6_in_hdrerr);
311 return;
312 }
313 }
314
315 ip6 = mtod(m, struct ip6_hdr *);
316
317 if ((ip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION) {
318 IP6_STATINC(IP6_STAT_BADVERS);
319 in6_ifstat_inc(rcvif, ifs6_in_hdrerr);
320 goto bad;
321 }
322
323 /*
324 * Assume that we can create a fast-forward IP flow entry
325 * based on this packet.
326 */
327 m->m_flags |= M_CANFASTFWD;
328
329 /*
330 * Run through list of hooks for input packets. If there are any
331 * filters which require that additional packets in the flow are
332 * not fast-forwarded, they must clear the M_CANFASTFWD flag.
333 * Note that filters must _never_ set this flag, as another filter
334 * in the list may have previously cleared it.
335 */
336 /*
337 * let ipfilter look at packet on the wire,
338 * not the decapsulated packet.
339 */
340 #if defined(IPSEC)
341 if (!ipsec_used || !ipsec_indone(m))
342 #else
343 if (1)
344 #endif
345 {
346 struct in6_addr odst;
347
348 odst = ip6->ip6_dst;
349 if (pfil_run_hooks(inet6_pfil_hook, &m, rcvif, PFIL_IN) != 0)
350 return;
351 if (m == NULL)
352 return;
353 ip6 = mtod(m, struct ip6_hdr *);
354 srcrt = !IN6_ARE_ADDR_EQUAL(&odst, &ip6->ip6_dst);
355 }
356
357 IP6_STATINC(IP6_STAT_NXTHIST + ip6->ip6_nxt);
358
359 #ifdef ALTQ
360 if (altq_input != NULL) {
361 SOFTNET_LOCK();
362 if ((*altq_input)(m, AF_INET6) == 0) {
363 SOFTNET_UNLOCK();
364 /* packet is dropped by traffic conditioner */
365 return;
366 }
367 SOFTNET_UNLOCK();
368 }
369 #endif
370
371 /*
372 * Check against address spoofing/corruption.
373 */
374 if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_src) ||
375 IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_dst)) {
376 /*
377 * XXX: "badscope" is not very suitable for a multicast source.
378 */
379 IP6_STATINC(IP6_STAT_BADSCOPE);
380 in6_ifstat_inc(rcvif, ifs6_in_addrerr);
381 goto bad;
382 }
383
384 /*
385 * The following check is not documented in specs. A malicious
386 * party may be able to use IPv4 mapped addr to confuse tcp/udp stack
387 * and bypass security checks (act as if it was from 127.0.0.1 by using
388 * IPv6 src ::ffff:127.0.0.1). Be cautious.
389 *
390 * This check chokes if we are in an SIIT cloud. As none of BSDs
391 * support IPv4-less kernel compilation, we cannot support SIIT
392 * environment at all. So, it makes more sense for us to reject any
393 * malicious packets for non-SIIT environment, than try to do a
394 * partial support for SIIT environment.
395 */
396 if (IN6_IS_ADDR_V4MAPPED(&ip6->ip6_src) ||
397 IN6_IS_ADDR_V4MAPPED(&ip6->ip6_dst)) {
398 IP6_STATINC(IP6_STAT_BADSCOPE);
399 in6_ifstat_inc(rcvif, ifs6_in_addrerr);
400 goto bad;
401 }
402
403 #if 0
404 /*
405 * Reject packets with IPv4 compatible addresses (auto tunnel).
406 *
407 * The code forbids auto tunnel relay case in RFC1933 (the check is
408 * stronger than RFC1933). We may want to re-enable it if mech-xx
409 * is revised to forbid relaying case.
410 */
411 if (IN6_IS_ADDR_V4COMPAT(&ip6->ip6_src) ||
412 IN6_IS_ADDR_V4COMPAT(&ip6->ip6_dst)) {
413 IP6_STATINC(IP6_STAT_BADSCOPE);
414 in6_ifstat_inc(rcvif, ifs6_in_addrerr);
415 goto bad;
416 }
417 #endif
418
419 /*
420 * Disambiguate address scope zones (if there is ambiguity).
421 * We first make sure that the original source or destination address
422 * is not in our internal form for scoped addresses. Such addresses
423 * are not necessarily invalid spec-wise, but we cannot accept them due
424 * to the usage conflict.
425 * in6_setscope() then also checks and rejects the cases where src or
426 * dst are the loopback address and the receiving interface
427 * is not loopback.
428 */
429 if (__predict_false(
430 m_makewritable(&m, 0, sizeof(struct ip6_hdr), M_DONTWAIT)))
431 goto bad;
432 ip6 = mtod(m, struct ip6_hdr *);
433 if (in6_clearscope(&ip6->ip6_src) || in6_clearscope(&ip6->ip6_dst)) {
434 IP6_STATINC(IP6_STAT_BADSCOPE); /* XXX */
435 goto bad;
436 }
437 if (in6_setscope(&ip6->ip6_src, rcvif, NULL) ||
438 in6_setscope(&ip6->ip6_dst, rcvif, NULL)) {
439 IP6_STATINC(IP6_STAT_BADSCOPE);
440 goto bad;
441 }
442
443 ro = percpu_getref(ip6_forward_rt_percpu);
444
445 /*
446 * Multicast check
447 */
448 if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) {
449 bool ingroup;
450
451 in6_ifstat_inc(rcvif, ifs6_in_mcast);
452 /*
453 * See if we belong to the destination multicast group on the
454 * arrival interface.
455 */
456 ingroup = in6_multi_group(&ip6->ip6_dst, rcvif);
457 if (ingroup) {
458 ours = 1;
459 } else if (!ip6_mrouter) {
460 uint64_t *ip6s = IP6_STAT_GETREF();
461 ip6s[IP6_STAT_NOTMEMBER]++;
462 ip6s[IP6_STAT_CANTFORWARD]++;
463 IP6_STAT_PUTREF();
464 in6_ifstat_inc(rcvif, ifs6_in_discard);
465 goto bad_unref;
466 }
467 deliverifp = rcvif;
468 goto hbhcheck;
469 }
470
471 sockaddr_in6_init(&u.dst6, &ip6->ip6_dst, 0, 0, 0);
472
473 /*
474 * Unicast check
475 */
476 rt = rtcache_lookup2(ro, &u.dst, 1, &hit);
477 if (hit)
478 IP6_STATINC(IP6_STAT_FORWARD_CACHEHIT);
479 else
480 IP6_STATINC(IP6_STAT_FORWARD_CACHEMISS);
481
482 /*
483 * Accept the packet if the forwarding interface to the destination
484 * (according to the routing table) is the loopback interface,
485 * unless the associated route has a gateway.
486 *
487 * We don't explicitly match ip6_dst against an interface here. It
488 * is already done in rtcache_lookup2: rt->rt_ifp->if_type will be
489 * IFT_LOOP if the packet is for us.
490 *
491 * Note that this approach causes to accept a packet if there is a
492 * route to the loopback interface for the destination of the packet.
493 * But we think it's even useful in some situations, e.g. when using
494 * a special daemon which wants to intercept the packet.
495 */
496 if (rt != NULL &&
497 (rt->rt_flags & (RTF_HOST|RTF_GATEWAY)) == RTF_HOST &&
498 rt->rt_ifp->if_type == IFT_LOOP) {
499 struct in6_ifaddr *ia6 = (struct in6_ifaddr *)rt->rt_ifa;
500 int addrok;
501
502 if (ia6->ia6_flags & IN6_IFF_ANYCAST)
503 m->m_flags |= M_ANYCAST6;
504 /*
505 * packets to a tentative, duplicated, or somehow invalid
506 * address must not be accepted.
507 */
508 if (ia6->ia6_flags & IN6_IFF_NOTREADY)
509 addrok = 0;
510 else if (ia6->ia6_flags & IN6_IFF_DETACHED &&
511 !IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_src))
512 {
513 /* Allow internal traffic to DETACHED addresses */
514 struct sockaddr_in6 sin6;
515 int s;
516
517 memset(&sin6, 0, sizeof(sin6));
518 sin6.sin6_family = AF_INET6;
519 sin6.sin6_len = sizeof(sin6);
520 sin6.sin6_addr = ip6->ip6_src;
521 s = pserialize_read_enter();
522 addrok = (ifa_ifwithaddr(sin6tosa(&sin6)) != NULL);
523 pserialize_read_exit(s);
524 } else
525 addrok = 1;
526 if (addrok) {
527 /* this address is ready */
528 ours = 1;
529 deliverifp = ia6->ia_ifp; /* correct? */
530 goto hbhcheck;
531 } else {
532 /* address is not ready, so discard the packet. */
533 char ip6bufs[INET6_ADDRSTRLEN];
534 char ip6bufd[INET6_ADDRSTRLEN];
535 nd6log(LOG_INFO, "packet to an unready address %s->%s\n",
536 IN6_PRINT(ip6bufs, &ip6->ip6_src),
537 IN6_PRINT(ip6bufd, &ip6->ip6_dst));
538
539 goto bad_unref;
540 }
541 }
542
543 /*
544 * FAITH (Firewall Aided Internet Translator)
545 */
546 #if defined(NFAITH) && 0 < NFAITH
547 if (ip6_keepfaith) {
548 if (rt != NULL && rt->rt_ifp != NULL &&
549 rt->rt_ifp->if_type == IFT_FAITH) {
550 /* XXX do we need more sanity checks? */
551 ours = 1;
552 deliverifp = rt->rt_ifp; /* faith */
553 goto hbhcheck;
554 }
555 }
556 #endif
557
558 /*
559 * Now there is no reason to process the packet if it's not our own
560 * and we're not a router.
561 */
562 if (!ip6_forwarding) {
563 IP6_STATINC(IP6_STAT_CANTFORWARD);
564 in6_ifstat_inc(rcvif, ifs6_in_discard);
565 goto bad_unref;
566 }
567
568 hbhcheck:
569 /*
570 * Record address information into m_tag, if we don't have one yet.
571 * Note that we are unable to record it, if the address is not listed
572 * as our interface address (e.g. multicast addresses, addresses
573 * within FAITH prefixes and such).
574 */
575 if (deliverifp && ip6_getdstifaddr(m) == NULL) {
576 struct in6_ifaddr *ia6;
577 int s = pserialize_read_enter();
578
579 ia6 = in6_ifawithifp(deliverifp, &ip6->ip6_dst);
580 /* Depends on ip6_setdstifaddr never sleep */
581 if (ia6 != NULL && ip6_setdstifaddr(m, ia6) == NULL) {
582 /*
583 * XXX maybe we should drop the packet here,
584 * as we could not provide enough information
585 * to the upper layers.
586 */
587 }
588 pserialize_read_exit(s);
589 }
590
591 /*
592 * Process Hop-by-Hop options header if it's contained.
593 * m may be modified in ip6_hopopts_input().
594 * If a JumboPayload option is included, plen will also be modified.
595 */
596 plen = (u_int32_t)ntohs(ip6->ip6_plen);
597 if (ip6->ip6_nxt == IPPROTO_HOPOPTS) {
598 struct ip6_hbh *hbh;
599
600 if (ip6_hopopts_input(&plen, &rtalert, &m, &off)) {
601 /* m already freed */
602 in6_ifstat_inc(rcvif, ifs6_in_discard);
603 rtcache_unref(rt, ro);
604 percpu_putref(ip6_forward_rt_percpu);
605 return;
606 }
607
608 /* adjust pointer */
609 ip6 = mtod(m, struct ip6_hdr *);
610
611 /*
612 * if the payload length field is 0 and the next header field
613 * indicates Hop-by-Hop Options header, then a Jumbo Payload
614 * option MUST be included.
615 */
616 if (ip6->ip6_plen == 0 && plen == 0) {
617 /*
618 * Note that if a valid jumbo payload option is
619 * contained, ip6_hopopts_input() must set a valid
620 * (non-zero) payload length to the variable plen.
621 */
622 IP6_STATINC(IP6_STAT_BADOPTIONS);
623 in6_ifstat_inc(rcvif, ifs6_in_discard);
624 in6_ifstat_inc(rcvif, ifs6_in_hdrerr);
625 icmp6_error(m, ICMP6_PARAM_PROB,
626 ICMP6_PARAMPROB_HEADER,
627 (char *)&ip6->ip6_plen - (char *)ip6);
628 rtcache_unref(rt, ro);
629 percpu_putref(ip6_forward_rt_percpu);
630 return;
631 }
632 IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m, sizeof(struct ip6_hdr),
633 sizeof(struct ip6_hbh));
634 if (hbh == NULL) {
635 IP6_STATINC(IP6_STAT_TOOSHORT);
636 rtcache_unref(rt, ro);
637 percpu_putref(ip6_forward_rt_percpu);
638 return;
639 }
640 KASSERT(IP6_HDR_ALIGNED_P(hbh));
641 nxt = hbh->ip6h_nxt;
642
643 /*
644 * accept the packet if a router alert option is included
645 * and we act as an IPv6 router.
646 */
647 if (rtalert != ~0 && ip6_forwarding)
648 ours = 1;
649 } else
650 nxt = ip6->ip6_nxt;
651
652 /*
653 * Check that the amount of data in the buffers is at least much as
654 * the IPv6 header would have us expect. Trim mbufs if longer than we
655 * expect. Drop packet if shorter than we expect.
656 */
657 if (m->m_pkthdr.len - sizeof(struct ip6_hdr) < plen) {
658 IP6_STATINC(IP6_STAT_TOOSHORT);
659 in6_ifstat_inc(rcvif, ifs6_in_truncated);
660 goto bad_unref;
661 }
662 if (m->m_pkthdr.len > sizeof(struct ip6_hdr) + plen) {
663 if (m->m_len == m->m_pkthdr.len) {
664 m->m_len = sizeof(struct ip6_hdr) + plen;
665 m->m_pkthdr.len = sizeof(struct ip6_hdr) + plen;
666 } else
667 m_adj(m, sizeof(struct ip6_hdr) + plen - m->m_pkthdr.len);
668 }
669
670 /*
671 * Forward if desirable.
672 */
673 if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) {
674 /*
675 * If we are acting as a multicast router, all
676 * incoming multicast packets are passed to the
677 * kernel-level multicast forwarding function.
678 * The packet is returned (relatively) intact; if
679 * ip6_mforward() returns a non-zero value, the packet
680 * must be discarded, else it may be accepted below.
681 */
682 if (ip6_mrouter != NULL) {
683 int error;
684
685 SOFTNET_LOCK();
686 error = ip6_mforward(ip6, rcvif, m);
687 SOFTNET_UNLOCK();
688
689 if (error != 0) {
690 rtcache_unref(rt, ro);
691 percpu_putref(ip6_forward_rt_percpu);
692 IP6_STATINC(IP6_STAT_CANTFORWARD);
693 goto bad;
694 }
695 }
696 if (!ours)
697 goto bad_unref;
698 } else if (!ours) {
699 rtcache_unref(rt, ro);
700 percpu_putref(ip6_forward_rt_percpu);
701 ip6_forward(m, srcrt);
702 return;
703 }
704
705 ip6 = mtod(m, struct ip6_hdr *);
706
707 /*
708 * Malicious party may be able to use IPv4 mapped addr to confuse
709 * tcp/udp stack and bypass security checks (act as if it was from
710 * 127.0.0.1 by using IPv6 src ::ffff:127.0.0.1). Be cautious.
711 *
712 * For SIIT end node behavior, you may want to disable the check.
713 * However, you will become vulnerable to attacks using IPv4 mapped
714 * source.
715 */
716 if (IN6_IS_ADDR_V4MAPPED(&ip6->ip6_src) ||
717 IN6_IS_ADDR_V4MAPPED(&ip6->ip6_dst)) {
718 IP6_STATINC(IP6_STAT_BADSCOPE);
719 in6_ifstat_inc(rcvif, ifs6_in_addrerr);
720 goto bad_unref;
721 }
722
723 #ifdef IFA_STATS
724 if (deliverifp != NULL) {
725 struct in6_ifaddr *ia6;
726 int s = pserialize_read_enter();
727 ia6 = in6_ifawithifp(deliverifp, &ip6->ip6_dst);
728 if (ia6)
729 ia6->ia_ifa.ifa_data.ifad_inbytes += m->m_pkthdr.len;
730 pserialize_read_exit(s);
731 }
732 #endif
733 IP6_STATINC(IP6_STAT_DELIVERED);
734 in6_ifstat_inc(deliverifp, ifs6_in_deliver);
735 nest = 0;
736
737 if (rt != NULL) {
738 rtcache_unref(rt, ro);
739 rt = NULL;
740 }
741 percpu_putref(ip6_forward_rt_percpu);
742
743 rh_present = 0;
744 while (nxt != IPPROTO_DONE) {
745 if (ip6_hdrnestlimit && (++nest > ip6_hdrnestlimit)) {
746 IP6_STATINC(IP6_STAT_TOOMANYHDR);
747 in6_ifstat_inc(rcvif, ifs6_in_hdrerr);
748 goto bad;
749 }
750
751 /*
752 * protection against faulty packet - there should be
753 * more sanity checks in header chain processing.
754 */
755 if (m->m_pkthdr.len < off) {
756 IP6_STATINC(IP6_STAT_TOOSHORT);
757 in6_ifstat_inc(rcvif, ifs6_in_truncated);
758 goto bad;
759 }
760
761 if (nxt == IPPROTO_ROUTING) {
762 if (rh_present++) {
763 in6_ifstat_inc(rcvif, ifs6_in_hdrerr);
764 IP6_STATINC(IP6_STAT_BADOPTIONS);
765 goto bad;
766 }
767 }
768
769 #ifdef IPSEC
770 if (ipsec_used) {
771 /*
772 * Enforce IPsec policy checking if we are seeing last
773 * header. Note that we do not visit this with
774 * protocols with pcb layer code - like udp/tcp/raw ip.
775 */
776 if ((inet6sw[ip_protox[nxt]].pr_flags
777 & PR_LASTHDR) != 0) {
778 int error;
779
780 error = ipsec6_input(m);
781 if (error)
782 goto bad;
783 }
784 }
785 #endif
786
787 nxt = (*inet6sw[ip6_protox[nxt]].pr_input)(&m, &off, nxt);
788 }
789 return;
790
791 bad_unref:
792 rtcache_unref(rt, ro);
793 percpu_putref(ip6_forward_rt_percpu);
794 bad:
795 m_freem(m);
796 return;
797 }
798
799 /*
800 * set/grab in6_ifaddr correspond to IPv6 destination address.
801 */
802 static struct m_tag *
803 ip6_setdstifaddr(struct mbuf *m, const struct in6_ifaddr *ia)
804 {
805 struct m_tag *mtag;
806 struct ip6aux *ip6a;
807
808 mtag = ip6_addaux(m);
809 if (mtag == NULL)
810 return NULL;
811
812 ip6a = (struct ip6aux *)(mtag + 1);
813 if (in6_setscope(&ip6a->ip6a_src, ia->ia_ifp, &ip6a->ip6a_scope_id)) {
814 IP6_STATINC(IP6_STAT_BADSCOPE);
815 return NULL;
816 }
817
818 ip6a->ip6a_src = ia->ia_addr.sin6_addr;
819 ip6a->ip6a_flags = ia->ia6_flags;
820 return mtag;
821 }
822
823 const struct ip6aux *
824 ip6_getdstifaddr(struct mbuf *m)
825 {
826 struct m_tag *mtag;
827
828 mtag = ip6_findaux(m);
829 if (mtag != NULL)
830 return (struct ip6aux *)(mtag + 1);
831 else
832 return NULL;
833 }
834
835 /*
836 * Hop-by-Hop options header processing. If a valid jumbo payload option is
837 * included, the real payload length will be stored in plenp.
838 *
839 * rtalertp - XXX: should be stored more smart way
840 */
841 int
842 ip6_hopopts_input(u_int32_t *plenp, u_int32_t *rtalertp,
843 struct mbuf **mp, int *offp)
844 {
845 struct mbuf *m = *mp;
846 int off = *offp, hbhlen;
847 struct ip6_hbh *hbh;
848
849 /* validation of the length of the header */
850 IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m,
851 sizeof(struct ip6_hdr), sizeof(struct ip6_hbh));
852 if (hbh == NULL) {
853 IP6_STATINC(IP6_STAT_TOOSHORT);
854 return -1;
855 }
856 hbhlen = (hbh->ip6h_len + 1) << 3;
857 IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m, sizeof(struct ip6_hdr),
858 hbhlen);
859 if (hbh == NULL) {
860 IP6_STATINC(IP6_STAT_TOOSHORT);
861 return -1;
862 }
863 KASSERT(IP6_HDR_ALIGNED_P(hbh));
864 off += hbhlen;
865 hbhlen -= sizeof(struct ip6_hbh);
866
867 if (ip6_process_hopopts(m, (u_int8_t *)hbh + sizeof(struct ip6_hbh),
868 hbhlen, rtalertp, plenp) < 0)
869 return (-1);
870
871 *offp = off;
872 *mp = m;
873 return (0);
874 }
875
876 /*
877 * Search header for all Hop-by-hop options and process each option.
878 * This function is separate from ip6_hopopts_input() in order to
879 * handle a case where the sending node itself process its hop-by-hop
880 * options header. In such a case, the function is called from ip6_output().
881 *
882 * The function assumes that hbh header is located right after the IPv6 header
883 * (RFC2460 p7), opthead is pointer into data content in m, and opthead to
884 * opthead + hbhlen is located in continuous memory region.
885 */
886 static int
887 ip6_process_hopopts(struct mbuf *m, u_int8_t *opthead, int hbhlen,
888 u_int32_t *rtalertp, u_int32_t *plenp)
889 {
890 struct ip6_hdr *ip6;
891 int optlen = 0;
892 u_int8_t *opt = opthead;
893 u_int16_t rtalert_val;
894 u_int32_t jumboplen;
895 const int erroff = sizeof(struct ip6_hdr) + sizeof(struct ip6_hbh);
896
897 for (; hbhlen > 0; hbhlen -= optlen, opt += optlen) {
898 switch (*opt) {
899 case IP6OPT_PAD1:
900 optlen = 1;
901 break;
902 case IP6OPT_PADN:
903 if (hbhlen < IP6OPT_MINLEN) {
904 IP6_STATINC(IP6_STAT_TOOSMALL);
905 goto bad;
906 }
907 optlen = *(opt + 1) + 2;
908 break;
909 case IP6OPT_RTALERT:
910 /* XXX may need check for alignment */
911 if (hbhlen < IP6OPT_RTALERT_LEN) {
912 IP6_STATINC(IP6_STAT_TOOSMALL);
913 goto bad;
914 }
915 if (*(opt + 1) != IP6OPT_RTALERT_LEN - 2) {
916 /* XXX stat */
917 icmp6_error(m, ICMP6_PARAM_PROB,
918 ICMP6_PARAMPROB_HEADER,
919 erroff + opt + 1 - opthead);
920 return (-1);
921 }
922 optlen = IP6OPT_RTALERT_LEN;
923 memcpy((void *)&rtalert_val, (void *)(opt + 2), 2);
924 *rtalertp = ntohs(rtalert_val);
925 break;
926 case IP6OPT_JUMBO:
927 /* XXX may need check for alignment */
928 if (hbhlen < IP6OPT_JUMBO_LEN) {
929 IP6_STATINC(IP6_STAT_TOOSMALL);
930 goto bad;
931 }
932 if (*(opt + 1) != IP6OPT_JUMBO_LEN - 2) {
933 /* XXX stat */
934 icmp6_error(m, ICMP6_PARAM_PROB,
935 ICMP6_PARAMPROB_HEADER,
936 erroff + opt + 1 - opthead);
937 return (-1);
938 }
939 optlen = IP6OPT_JUMBO_LEN;
940
941 /*
942 * IPv6 packets that have non 0 payload length
943 * must not contain a jumbo payload option.
944 */
945 ip6 = mtod(m, struct ip6_hdr *);
946 if (ip6->ip6_plen) {
947 IP6_STATINC(IP6_STAT_BADOPTIONS);
948 icmp6_error(m, ICMP6_PARAM_PROB,
949 ICMP6_PARAMPROB_HEADER,
950 erroff + opt - opthead);
951 return (-1);
952 }
953
954 /*
955 * We may see jumbolen in unaligned location, so
956 * we'd need to perform bcopy().
957 */
958 memcpy(&jumboplen, opt + 2, sizeof(jumboplen));
959 jumboplen = (u_int32_t)htonl(jumboplen);
960
961 #if 1
962 /*
963 * if there are multiple jumbo payload options,
964 * *plenp will be non-zero and the packet will be
965 * rejected.
966 * the behavior may need some debate in ipngwg -
967 * multiple options does not make sense, however,
968 * there's no explicit mention in specification.
969 */
970 if (*plenp != 0) {
971 IP6_STATINC(IP6_STAT_BADOPTIONS);
972 icmp6_error(m, ICMP6_PARAM_PROB,
973 ICMP6_PARAMPROB_HEADER,
974 erroff + opt + 2 - opthead);
975 return (-1);
976 }
977 #endif
978
979 /*
980 * jumbo payload length must be larger than 65535.
981 */
982 if (jumboplen <= IPV6_MAXPACKET) {
983 IP6_STATINC(IP6_STAT_BADOPTIONS);
984 icmp6_error(m, ICMP6_PARAM_PROB,
985 ICMP6_PARAMPROB_HEADER,
986 erroff + opt + 2 - opthead);
987 return (-1);
988 }
989 *plenp = jumboplen;
990
991 break;
992 default: /* unknown option */
993 if (hbhlen < IP6OPT_MINLEN) {
994 IP6_STATINC(IP6_STAT_TOOSMALL);
995 goto bad;
996 }
997 optlen = ip6_unknown_opt(opt, m,
998 erroff + opt - opthead);
999 if (optlen == -1)
1000 return (-1);
1001 optlen += 2;
1002 break;
1003 }
1004 }
1005
1006 return (0);
1007
1008 bad:
1009 m_freem(m);
1010 return (-1);
1011 }
1012
1013 /*
1014 * Unknown option processing.
1015 * The third argument `off' is the offset from the IPv6 header to the option,
1016 * which is necessary if the IPv6 header the and option header and IPv6 header
1017 * is not continuous in order to return an ICMPv6 error.
1018 */
1019 int
1020 ip6_unknown_opt(u_int8_t *optp, struct mbuf *m, int off)
1021 {
1022 struct ip6_hdr *ip6;
1023
1024 switch (IP6OPT_TYPE(*optp)) {
1025 case IP6OPT_TYPE_SKIP: /* ignore the option */
1026 return ((int)*(optp + 1));
1027 case IP6OPT_TYPE_DISCARD: /* silently discard */
1028 m_freem(m);
1029 return (-1);
1030 case IP6OPT_TYPE_FORCEICMP: /* send ICMP even if multicasted */
1031 IP6_STATINC(IP6_STAT_BADOPTIONS);
1032 icmp6_error(m, ICMP6_PARAM_PROB, ICMP6_PARAMPROB_OPTION, off);
1033 return (-1);
1034 case IP6OPT_TYPE_ICMP: /* send ICMP if not multicasted */
1035 IP6_STATINC(IP6_STAT_BADOPTIONS);
1036 ip6 = mtod(m, struct ip6_hdr *);
1037 if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) ||
1038 (m->m_flags & (M_BCAST|M_MCAST)))
1039 m_freem(m);
1040 else
1041 icmp6_error(m, ICMP6_PARAM_PROB,
1042 ICMP6_PARAMPROB_OPTION, off);
1043 return (-1);
1044 }
1045
1046 m_freem(m); /* XXX: NOTREACHED */
1047 return (-1);
1048 }
1049
1050 /*
1051 * Create the "control" list for this pcb.
1052 *
1053 * The routine will be called from upper layer handlers like tcp6_input().
1054 * Thus the routine assumes that the caller (tcp6_input) have already
1055 * called IP6_EXTHDR_CHECK() and all the extension headers are located in the
1056 * very first mbuf on the mbuf chain.
1057 * We may want to add some infinite loop prevention or sanity checks for safety.
1058 * (This applies only when you are using KAME mbuf chain restriction, i.e.
1059 * you are using IP6_EXTHDR_CHECK() not m_pulldown())
1060 */
1061 void
1062 ip6_savecontrol(struct in6pcb *in6p, struct mbuf **mp,
1063 struct ip6_hdr *ip6, struct mbuf *m)
1064 {
1065 struct socket *so = in6p->in6p_socket;
1066 #ifdef RFC2292
1067 #define IS2292(x, y) ((in6p->in6p_flags & IN6P_RFC2292) ? (x) : (y))
1068 #else
1069 #define IS2292(x, y) (y)
1070 #endif
1071
1072 if (SOOPT_TIMESTAMP(so->so_options))
1073 mp = sbsavetimestamp(so->so_options, m, mp);
1074
1075 /* some OSes call this logic with IPv4 packet, for SO_TIMESTAMP */
1076 if ((ip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION)
1077 return;
1078
1079 /* RFC 2292 sec. 5 */
1080 if ((in6p->in6p_flags & IN6P_PKTINFO) != 0) {
1081 struct in6_pktinfo pi6;
1082
1083 memcpy(&pi6.ipi6_addr, &ip6->ip6_dst, sizeof(struct in6_addr));
1084 in6_clearscope(&pi6.ipi6_addr); /* XXX */
1085 pi6.ipi6_ifindex = m->m_pkthdr.rcvif_index;
1086 *mp = sbcreatecontrol(&pi6, sizeof(pi6),
1087 IS2292(IPV6_2292PKTINFO, IPV6_PKTINFO), IPPROTO_IPV6);
1088 if (*mp)
1089 mp = &(*mp)->m_next;
1090 }
1091
1092 if (in6p->in6p_flags & IN6P_HOPLIMIT) {
1093 int hlim = ip6->ip6_hlim & 0xff;
1094
1095 *mp = sbcreatecontrol(&hlim, sizeof(hlim),
1096 IS2292(IPV6_2292HOPLIMIT, IPV6_HOPLIMIT), IPPROTO_IPV6);
1097 if (*mp)
1098 mp = &(*mp)->m_next;
1099 }
1100
1101 if ((in6p->in6p_flags & IN6P_TCLASS) != 0) {
1102 u_int32_t flowinfo;
1103 int tclass;
1104
1105 flowinfo = (u_int32_t)ntohl(ip6->ip6_flow & IPV6_FLOWINFO_MASK);
1106 flowinfo >>= 20;
1107
1108 tclass = flowinfo & 0xff;
1109 *mp = sbcreatecontrol(&tclass, sizeof(tclass),
1110 IPV6_TCLASS, IPPROTO_IPV6);
1111
1112 if (*mp)
1113 mp = &(*mp)->m_next;
1114 }
1115
1116 /*
1117 * IPV6_HOPOPTS socket option. Recall that we required super-user
1118 * privilege for the option (see ip6_ctloutput), but it might be too
1119 * strict, since there might be some hop-by-hop options which can be
1120 * returned to normal user.
1121 * See also RFC3542 section 8 (or RFC2292 section 6).
1122 */
1123 if ((in6p->in6p_flags & IN6P_HOPOPTS) != 0) {
1124 /*
1125 * Check if a hop-by-hop options header is contatined in the
1126 * received packet, and if so, store the options as ancillary
1127 * data. Note that a hop-by-hop options header must be
1128 * just after the IPv6 header, which fact is assured through
1129 * the IPv6 input processing.
1130 */
1131 struct ip6_hdr *xip6 = mtod(m, struct ip6_hdr *);
1132 if (xip6->ip6_nxt == IPPROTO_HOPOPTS) {
1133 struct ip6_hbh *hbh;
1134 int hbhlen;
1135 struct mbuf *ext;
1136
1137 ext = ip6_pullexthdr(m, sizeof(struct ip6_hdr),
1138 xip6->ip6_nxt);
1139 if (ext == NULL) {
1140 IP6_STATINC(IP6_STAT_TOOSHORT);
1141 return;
1142 }
1143 hbh = mtod(ext, struct ip6_hbh *);
1144 hbhlen = (hbh->ip6h_len + 1) << 3;
1145 if (hbhlen != ext->m_len) {
1146 m_freem(ext);
1147 IP6_STATINC(IP6_STAT_TOOSHORT);
1148 return;
1149 }
1150
1151 /*
1152 * XXX: We copy whole the header even if a jumbo
1153 * payload option is included, which option is to
1154 * be removed before returning in the RFC 2292.
1155 * Note: this constraint is removed in RFC3542.
1156 */
1157 *mp = sbcreatecontrol(hbh, hbhlen,
1158 IS2292(IPV6_2292HOPOPTS, IPV6_HOPOPTS),
1159 IPPROTO_IPV6);
1160 if (*mp)
1161 mp = &(*mp)->m_next;
1162 m_freem(ext);
1163 }
1164 }
1165
1166 /* IPV6_DSTOPTS and IPV6_RTHDR socket options */
1167 if (in6p->in6p_flags & (IN6P_DSTOPTS | IN6P_RTHDR)) {
1168 struct ip6_hdr *xip6 = mtod(m, struct ip6_hdr *);
1169 int nxt = xip6->ip6_nxt, off = sizeof(struct ip6_hdr);
1170
1171 /*
1172 * Search for destination options headers or routing
1173 * header(s) through the header chain, and stores each
1174 * header as ancillary data.
1175 * Note that the order of the headers remains in
1176 * the chain of ancillary data.
1177 */
1178 for (;;) { /* is explicit loop prevention necessary? */
1179 struct ip6_ext *ip6e = NULL;
1180 int elen;
1181 struct mbuf *ext = NULL;
1182
1183 /*
1184 * if it is not an extension header, don't try to
1185 * pull it from the chain.
1186 */
1187 switch (nxt) {
1188 case IPPROTO_DSTOPTS:
1189 case IPPROTO_ROUTING:
1190 case IPPROTO_HOPOPTS:
1191 case IPPROTO_AH: /* is it possible? */
1192 break;
1193 default:
1194 goto loopend;
1195 }
1196
1197 ext = ip6_pullexthdr(m, off, nxt);
1198 if (ext == NULL) {
1199 IP6_STATINC(IP6_STAT_TOOSHORT);
1200 return;
1201 }
1202 ip6e = mtod(ext, struct ip6_ext *);
1203 if (nxt == IPPROTO_AH)
1204 elen = (ip6e->ip6e_len + 2) << 2;
1205 else
1206 elen = (ip6e->ip6e_len + 1) << 3;
1207 if (elen != ext->m_len) {
1208 m_freem(ext);
1209 IP6_STATINC(IP6_STAT_TOOSHORT);
1210 return;
1211 }
1212 KASSERT(IP6_HDR_ALIGNED_P(ip6e));
1213
1214 switch (nxt) {
1215 case IPPROTO_DSTOPTS:
1216 if (!(in6p->in6p_flags & IN6P_DSTOPTS))
1217 break;
1218
1219 *mp = sbcreatecontrol(ip6e, elen,
1220 IS2292(IPV6_2292DSTOPTS, IPV6_DSTOPTS),
1221 IPPROTO_IPV6);
1222 if (*mp)
1223 mp = &(*mp)->m_next;
1224 break;
1225
1226 case IPPROTO_ROUTING:
1227 if (!(in6p->in6p_flags & IN6P_RTHDR))
1228 break;
1229
1230 *mp = sbcreatecontrol(ip6e, elen,
1231 IS2292(IPV6_2292RTHDR, IPV6_RTHDR),
1232 IPPROTO_IPV6);
1233 if (*mp)
1234 mp = &(*mp)->m_next;
1235 break;
1236
1237 case IPPROTO_HOPOPTS:
1238 case IPPROTO_AH: /* is it possible? */
1239 break;
1240
1241 default:
1242 /*
1243 * other cases have been filtered in the above.
1244 * none will visit this case. here we supply
1245 * the code just in case (nxt overwritten or
1246 * other cases).
1247 */
1248 m_freem(ext);
1249 goto loopend;
1250
1251 }
1252
1253 /* proceed with the next header. */
1254 off += elen;
1255 nxt = ip6e->ip6e_nxt;
1256 ip6e = NULL;
1257 m_freem(ext);
1258 ext = NULL;
1259 }
1260 loopend:
1261 ;
1262 }
1263 }
1264 #undef IS2292
1265
1266
1267 void
1268 ip6_notify_pmtu(struct in6pcb *in6p, const struct sockaddr_in6 *dst,
1269 uint32_t *mtu)
1270 {
1271 struct socket *so;
1272 struct mbuf *m_mtu;
1273 struct ip6_mtuinfo mtuctl;
1274
1275 so = in6p->in6p_socket;
1276
1277 if (mtu == NULL)
1278 return;
1279
1280 KASSERT(so != NULL);
1281
1282 memset(&mtuctl, 0, sizeof(mtuctl)); /* zero-clear for safety */
1283 mtuctl.ip6m_mtu = *mtu;
1284 mtuctl.ip6m_addr = *dst;
1285 if (sa6_recoverscope(&mtuctl.ip6m_addr))
1286 return;
1287
1288 if ((m_mtu = sbcreatecontrol(&mtuctl, sizeof(mtuctl),
1289 IPV6_PATHMTU, IPPROTO_IPV6)) == NULL)
1290 return;
1291
1292 if (sbappendaddr(&so->so_rcv, (const struct sockaddr *)dst, NULL, m_mtu)
1293 == 0) {
1294 m_freem(m_mtu);
1295 /* XXX: should count statistics */
1296 } else
1297 sorwakeup(so);
1298
1299 return;
1300 }
1301
1302 /*
1303 * pull single extension header from mbuf chain. returns single mbuf that
1304 * contains the result, or NULL on error.
1305 */
1306 static struct mbuf *
1307 ip6_pullexthdr(struct mbuf *m, size_t off, int nxt)
1308 {
1309 struct ip6_ext ip6e;
1310 size_t elen;
1311 struct mbuf *n;
1312
1313 #ifdef DIAGNOSTIC
1314 switch (nxt) {
1315 case IPPROTO_DSTOPTS:
1316 case IPPROTO_ROUTING:
1317 case IPPROTO_HOPOPTS:
1318 case IPPROTO_AH: /* is it possible? */
1319 break;
1320 default:
1321 printf("ip6_pullexthdr: invalid nxt=%d\n", nxt);
1322 }
1323 #endif
1324
1325 m_copydata(m, off, sizeof(ip6e), (void *)&ip6e);
1326 if (nxt == IPPROTO_AH)
1327 elen = (ip6e.ip6e_len + 2) << 2;
1328 else
1329 elen = (ip6e.ip6e_len + 1) << 3;
1330
1331 MGET(n, M_DONTWAIT, MT_DATA);
1332 if (n && elen >= MLEN) {
1333 MCLGET(n, M_DONTWAIT);
1334 if ((n->m_flags & M_EXT) == 0) {
1335 m_free(n);
1336 n = NULL;
1337 }
1338 }
1339 if (!n)
1340 return NULL;
1341
1342 n->m_len = 0;
1343 if (elen >= M_TRAILINGSPACE(n)) {
1344 m_free(n);
1345 return NULL;
1346 }
1347
1348 m_copydata(m, off, elen, mtod(n, void *));
1349 n->m_len = elen;
1350 return n;
1351 }
1352
1353 /*
1354 * Get pointer to the previous header followed by the header
1355 * currently processed.
1356 * XXX: This function supposes that
1357 * M includes all headers,
1358 * the next header field and the header length field of each header
1359 * are valid, and
1360 * the sum of each header length equals to OFF.
1361 * Because of these assumptions, this function must be called very
1362 * carefully. Moreover, it will not be used in the near future when
1363 * we develop `neater' mechanism to process extension headers.
1364 */
1365 u_int8_t *
1366 ip6_get_prevhdr(struct mbuf *m, int off)
1367 {
1368 struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *);
1369
1370 if (off == sizeof(struct ip6_hdr))
1371 return (&ip6->ip6_nxt);
1372 else {
1373 int len, nxt;
1374 struct ip6_ext *ip6e = NULL;
1375
1376 nxt = ip6->ip6_nxt;
1377 len = sizeof(struct ip6_hdr);
1378 while (len < off) {
1379 ip6e = (struct ip6_ext *)(mtod(m, char *) + len);
1380
1381 switch (nxt) {
1382 case IPPROTO_FRAGMENT:
1383 len += sizeof(struct ip6_frag);
1384 break;
1385 case IPPROTO_AH:
1386 len += (ip6e->ip6e_len + 2) << 2;
1387 break;
1388 default:
1389 len += (ip6e->ip6e_len + 1) << 3;
1390 break;
1391 }
1392 nxt = ip6e->ip6e_nxt;
1393 }
1394 if (ip6e)
1395 return (&ip6e->ip6e_nxt);
1396 else
1397 return NULL;
1398 }
1399 }
1400
1401 /*
1402 * get next header offset. m will be retained.
1403 */
1404 int
1405 ip6_nexthdr(struct mbuf *m, int off, int proto, int *nxtp)
1406 {
1407 struct ip6_hdr ip6;
1408 struct ip6_ext ip6e;
1409 struct ip6_frag fh;
1410
1411 /* just in case */
1412 if (m == NULL)
1413 panic("ip6_nexthdr: m == NULL");
1414 if ((m->m_flags & M_PKTHDR) == 0 || m->m_pkthdr.len < off)
1415 return -1;
1416
1417 switch (proto) {
1418 case IPPROTO_IPV6:
1419 /* do not chase beyond intermediate IPv6 headers */
1420 if (off != 0)
1421 return -1;
1422 if (m->m_pkthdr.len < off + sizeof(ip6))
1423 return -1;
1424 m_copydata(m, off, sizeof(ip6), (void *)&ip6);
1425 if (nxtp)
1426 *nxtp = ip6.ip6_nxt;
1427 off += sizeof(ip6);
1428 return off;
1429
1430 case IPPROTO_FRAGMENT:
1431 /*
1432 * terminate parsing if it is not the first fragment,
1433 * it does not make sense to parse through it.
1434 */
1435 if (m->m_pkthdr.len < off + sizeof(fh))
1436 return -1;
1437 m_copydata(m, off, sizeof(fh), (void *)&fh);
1438 if ((fh.ip6f_offlg & IP6F_OFF_MASK) != 0)
1439 return -1;
1440 if (nxtp)
1441 *nxtp = fh.ip6f_nxt;
1442 off += sizeof(struct ip6_frag);
1443 return off;
1444
1445 case IPPROTO_AH:
1446 if (m->m_pkthdr.len < off + sizeof(ip6e))
1447 return -1;
1448 m_copydata(m, off, sizeof(ip6e), (void *)&ip6e);
1449 if (nxtp)
1450 *nxtp = ip6e.ip6e_nxt;
1451 off += (ip6e.ip6e_len + 2) << 2;
1452 if (m->m_pkthdr.len < off)
1453 return -1;
1454 return off;
1455
1456 case IPPROTO_HOPOPTS:
1457 case IPPROTO_ROUTING:
1458 case IPPROTO_DSTOPTS:
1459 if (m->m_pkthdr.len < off + sizeof(ip6e))
1460 return -1;
1461 m_copydata(m, off, sizeof(ip6e), (void *)&ip6e);
1462 if (nxtp)
1463 *nxtp = ip6e.ip6e_nxt;
1464 off += (ip6e.ip6e_len + 1) << 3;
1465 if (m->m_pkthdr.len < off)
1466 return -1;
1467 return off;
1468
1469 case IPPROTO_NONE:
1470 case IPPROTO_ESP:
1471 case IPPROTO_IPCOMP:
1472 /* give up */
1473 return -1;
1474
1475 default:
1476 return -1;
1477 }
1478 }
1479
1480 /*
1481 * get offset for the last header in the chain. m will be kept untainted.
1482 */
1483 int
1484 ip6_lasthdr(struct mbuf *m, int off, int proto, int *nxtp)
1485 {
1486 int newoff;
1487 int nxt;
1488
1489 if (!nxtp) {
1490 nxt = -1;
1491 nxtp = &nxt;
1492 }
1493 for (;;) {
1494 newoff = ip6_nexthdr(m, off, proto, nxtp);
1495 if (newoff < 0)
1496 return off;
1497 else if (newoff < off)
1498 return -1; /* invalid */
1499 else if (newoff == off)
1500 return newoff;
1501
1502 off = newoff;
1503 proto = *nxtp;
1504 }
1505 }
1506
1507 struct m_tag *
1508 ip6_addaux(struct mbuf *m)
1509 {
1510 struct m_tag *mtag;
1511
1512 mtag = m_tag_find(m, PACKET_TAG_INET6, NULL);
1513 if (!mtag) {
1514 mtag = m_tag_get(PACKET_TAG_INET6, sizeof(struct ip6aux),
1515 M_NOWAIT);
1516 if (mtag) {
1517 m_tag_prepend(m, mtag);
1518 memset(mtag + 1, 0, sizeof(struct ip6aux));
1519 }
1520 }
1521 return mtag;
1522 }
1523
1524 struct m_tag *
1525 ip6_findaux(struct mbuf *m)
1526 {
1527 struct m_tag *mtag;
1528
1529 mtag = m_tag_find(m, PACKET_TAG_INET6, NULL);
1530 return mtag;
1531 }
1532
1533 void
1534 ip6_delaux(struct mbuf *m)
1535 {
1536 struct m_tag *mtag;
1537
1538 mtag = m_tag_find(m, PACKET_TAG_INET6, NULL);
1539 if (mtag)
1540 m_tag_delete(m, mtag);
1541 }
1542
1543 /*
1544 * System control for IP6
1545 */
1546
1547 const u_char inet6ctlerrmap[PRC_NCMDS] = {
1548 0, 0, 0, 0,
1549 0, EMSGSIZE, EHOSTDOWN, EHOSTUNREACH,
1550 EHOSTUNREACH, EHOSTUNREACH, ECONNREFUSED, ECONNREFUSED,
1551 EMSGSIZE, EHOSTUNREACH, 0, 0,
1552 0, 0, 0, 0,
1553 ENOPROTOOPT
1554 };
1555
1556 extern int sysctl_net_inet6_addrctlpolicy(SYSCTLFN_ARGS);
1557
1558 static int
1559 sysctl_net_inet6_ip6_stats(SYSCTLFN_ARGS)
1560 {
1561
1562 return (NETSTAT_SYSCTL(ip6stat_percpu, IP6_NSTATS));
1563 }
1564
1565 static void
1566 sysctl_net_inet6_ip6_setup(struct sysctllog **clog)
1567 {
1568 #ifdef RFC2292
1569 #define IS2292(x, y) ((in6p->in6p_flags & IN6P_RFC2292) ? (x) : (y))
1570 #else
1571 #define IS2292(x, y) (y)
1572 #endif
1573
1574 sysctl_createv(clog, 0, NULL, NULL,
1575 CTLFLAG_PERMANENT,
1576 CTLTYPE_NODE, "inet6",
1577 SYSCTL_DESCR("PF_INET6 related settings"),
1578 NULL, 0, NULL, 0,
1579 CTL_NET, PF_INET6, CTL_EOL);
1580 sysctl_createv(clog, 0, NULL, NULL,
1581 CTLFLAG_PERMANENT,
1582 CTLTYPE_NODE, "ip6",
1583 SYSCTL_DESCR("IPv6 related settings"),
1584 NULL, 0, NULL, 0,
1585 CTL_NET, PF_INET6, IPPROTO_IPV6, CTL_EOL);
1586
1587 sysctl_createv(clog, 0, NULL, NULL,
1588 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1589 CTLTYPE_INT, "forwarding",
1590 SYSCTL_DESCR("Enable forwarding of INET6 datagrams"),
1591 NULL, 0, &ip6_forwarding, 0,
1592 CTL_NET, PF_INET6, IPPROTO_IPV6,
1593 IPV6CTL_FORWARDING, CTL_EOL);
1594 sysctl_createv(clog, 0, NULL, NULL,
1595 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1596 CTLTYPE_INT, "redirect",
1597 SYSCTL_DESCR("Enable sending of ICMPv6 redirect messages"),
1598 NULL, 0, &ip6_sendredirects, 0,
1599 CTL_NET, PF_INET6, IPPROTO_IPV6,
1600 IPV6CTL_SENDREDIRECTS, CTL_EOL);
1601 sysctl_createv(clog, 0, NULL, NULL,
1602 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1603 CTLTYPE_INT, "hlim",
1604 SYSCTL_DESCR("Hop limit for an INET6 datagram"),
1605 NULL, 0, &ip6_defhlim, 0,
1606 CTL_NET, PF_INET6, IPPROTO_IPV6,
1607 IPV6CTL_DEFHLIM, CTL_EOL);
1608 #ifdef notyet
1609 sysctl_createv(clog, 0, NULL, NULL,
1610 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1611 CTLTYPE_INT, "mtu", NULL,
1612 NULL, 0, &, 0,
1613 CTL_NET, PF_INET6, IPPROTO_IPV6,
1614 IPV6CTL_DEFMTU, CTL_EOL);
1615 #endif
1616 #ifdef __no_idea__
1617 sysctl_createv(clog, 0, NULL, NULL,
1618 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1619 CTLTYPE_INT, "forwsrcrt", NULL,
1620 NULL, 0, &?, 0,
1621 CTL_NET, PF_INET6, IPPROTO_IPV6,
1622 IPV6CTL_FORWSRCRT, CTL_EOL);
1623 sysctl_createv(clog, 0, NULL, NULL,
1624 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1625 CTLTYPE_STRUCT, "mrtstats", NULL,
1626 NULL, 0, &?, sizeof(?),
1627 CTL_NET, PF_INET6, IPPROTO_IPV6,
1628 IPV6CTL_MRTSTATS, CTL_EOL);
1629 sysctl_createv(clog, 0, NULL, NULL,
1630 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1631 CTLTYPE_?, "mrtproto", NULL,
1632 NULL, 0, &?, sizeof(?),
1633 CTL_NET, PF_INET6, IPPROTO_IPV6,
1634 IPV6CTL_MRTPROTO, CTL_EOL);
1635 #endif
1636 sysctl_createv(clog, 0, NULL, NULL,
1637 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1638 CTLTYPE_INT, "maxfragpackets",
1639 SYSCTL_DESCR("Maximum number of fragments to buffer "
1640 "for reassembly"),
1641 NULL, 0, &ip6_maxfragpackets, 0,
1642 CTL_NET, PF_INET6, IPPROTO_IPV6,
1643 IPV6CTL_MAXFRAGPACKETS, CTL_EOL);
1644 #ifdef __no_idea__
1645 sysctl_createv(clog, 0, NULL, NULL,
1646 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1647 CTLTYPE_INT, "sourcecheck", NULL,
1648 NULL, 0, &?, 0,
1649 CTL_NET, PF_INET6, IPPROTO_IPV6,
1650 IPV6CTL_SOURCECHECK, CTL_EOL);
1651 sysctl_createv(clog, 0, NULL, NULL,
1652 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1653 CTLTYPE_INT, "sourcecheck_logint", NULL,
1654 NULL, 0, &?, 0,
1655 CTL_NET, PF_INET6, IPPROTO_IPV6,
1656 IPV6CTL_SOURCECHECK_LOGINT, CTL_EOL);
1657 #endif
1658 sysctl_createv(clog, 0, NULL, NULL,
1659 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1660 CTLTYPE_INT, "accept_rtadv",
1661 SYSCTL_DESCR("Accept router advertisements"),
1662 NULL, 0, &ip6_accept_rtadv, 0,
1663 CTL_NET, PF_INET6, IPPROTO_IPV6,
1664 IPV6CTL_ACCEPT_RTADV, CTL_EOL);
1665 sysctl_createv(clog, 0, NULL, NULL,
1666 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1667 CTLTYPE_INT, "rtadv_maxroutes",
1668 SYSCTL_DESCR("Maximum number of routes accepted via router advertisements"),
1669 NULL, 0, &ip6_rtadv_maxroutes, 0,
1670 CTL_NET, PF_INET6, IPPROTO_IPV6,
1671 IPV6CTL_RTADV_MAXROUTES, CTL_EOL);
1672 sysctl_createv(clog, 0, NULL, NULL,
1673 CTLFLAG_PERMANENT,
1674 CTLTYPE_INT, "rtadv_numroutes",
1675 SYSCTL_DESCR("Current number of routes accepted via router advertisements"),
1676 NULL, 0, &nd6_numroutes, 0,
1677 CTL_NET, PF_INET6, IPPROTO_IPV6,
1678 IPV6CTL_RTADV_NUMROUTES, CTL_EOL);
1679 sysctl_createv(clog, 0, NULL, NULL,
1680 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1681 CTLTYPE_INT, "keepfaith",
1682 SYSCTL_DESCR("Activate faith interface"),
1683 NULL, 0, &ip6_keepfaith, 0,
1684 CTL_NET, PF_INET6, IPPROTO_IPV6,
1685 IPV6CTL_KEEPFAITH, CTL_EOL);
1686 sysctl_createv(clog, 0, NULL, NULL,
1687 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1688 CTLTYPE_INT, "log_interval",
1689 SYSCTL_DESCR("Minumum interval between logging "
1690 "unroutable packets"),
1691 NULL, 0, &ip6_log_interval, 0,
1692 CTL_NET, PF_INET6, IPPROTO_IPV6,
1693 IPV6CTL_LOG_INTERVAL, CTL_EOL);
1694 sysctl_createv(clog, 0, NULL, NULL,
1695 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1696 CTLTYPE_INT, "hdrnestlimit",
1697 SYSCTL_DESCR("Maximum number of nested IPv6 headers"),
1698 NULL, 0, &ip6_hdrnestlimit, 0,
1699 CTL_NET, PF_INET6, IPPROTO_IPV6,
1700 IPV6CTL_HDRNESTLIMIT, CTL_EOL);
1701 sysctl_createv(clog, 0, NULL, NULL,
1702 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1703 CTLTYPE_INT, "dad_count",
1704 SYSCTL_DESCR("Number of Duplicate Address Detection "
1705 "probes to send"),
1706 NULL, 0, &ip6_dad_count, 0,
1707 CTL_NET, PF_INET6, IPPROTO_IPV6,
1708 IPV6CTL_DAD_COUNT, CTL_EOL);
1709 sysctl_createv(clog, 0, NULL, NULL,
1710 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1711 CTLTYPE_INT, "auto_flowlabel",
1712 SYSCTL_DESCR("Assign random IPv6 flow labels"),
1713 NULL, 0, &ip6_auto_flowlabel, 0,
1714 CTL_NET, PF_INET6, IPPROTO_IPV6,
1715 IPV6CTL_AUTO_FLOWLABEL, CTL_EOL);
1716 sysctl_createv(clog, 0, NULL, NULL,
1717 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1718 CTLTYPE_INT, "defmcasthlim",
1719 SYSCTL_DESCR("Default multicast hop limit"),
1720 NULL, 0, &ip6_defmcasthlim, 0,
1721 CTL_NET, PF_INET6, IPPROTO_IPV6,
1722 IPV6CTL_DEFMCASTHLIM, CTL_EOL);
1723 sysctl_createv(clog, 0, NULL, NULL,
1724 CTLFLAG_PERMANENT,
1725 CTLTYPE_STRING, "kame_version",
1726 SYSCTL_DESCR("KAME Version"),
1727 NULL, 0, __UNCONST(__KAME_VERSION), 0,
1728 CTL_NET, PF_INET6, IPPROTO_IPV6,
1729 IPV6CTL_KAME_VERSION, CTL_EOL);
1730 sysctl_createv(clog, 0, NULL, NULL,
1731 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1732 CTLTYPE_INT, "use_deprecated",
1733 SYSCTL_DESCR("Allow use of deprecated addresses as "
1734 "source addresses"),
1735 NULL, 0, &ip6_use_deprecated, 0,
1736 CTL_NET, PF_INET6, IPPROTO_IPV6,
1737 IPV6CTL_USE_DEPRECATED, CTL_EOL);
1738 sysctl_createv(clog, 0, NULL, NULL,
1739 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1740 CTLTYPE_INT, "rr_prune", NULL,
1741 NULL, 0, &ip6_rr_prune, 0,
1742 CTL_NET, PF_INET6, IPPROTO_IPV6,
1743 IPV6CTL_RR_PRUNE, CTL_EOL);
1744 sysctl_createv(clog, 0, NULL, NULL,
1745 CTLFLAG_PERMANENT
1746 #ifndef INET6_BINDV6ONLY
1747 |CTLFLAG_READWRITE,
1748 #endif
1749 CTLTYPE_INT, "v6only",
1750 SYSCTL_DESCR("Disallow PF_INET6 sockets from connecting "
1751 "to PF_INET sockets"),
1752 NULL, 0, &ip6_v6only, 0,
1753 CTL_NET, PF_INET6, IPPROTO_IPV6,
1754 IPV6CTL_V6ONLY, CTL_EOL);
1755 sysctl_createv(clog, 0, NULL, NULL,
1756 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1757 CTLTYPE_INT, "anonportmin",
1758 SYSCTL_DESCR("Lowest ephemeral port number to assign"),
1759 sysctl_net_inet_ip_ports, 0, &ip6_anonportmin, 0,
1760 CTL_NET, PF_INET6, IPPROTO_IPV6,
1761 IPV6CTL_ANONPORTMIN, CTL_EOL);
1762 sysctl_createv(clog, 0, NULL, NULL,
1763 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1764 CTLTYPE_INT, "anonportmax",
1765 SYSCTL_DESCR("Highest ephemeral port number to assign"),
1766 sysctl_net_inet_ip_ports, 0, &ip6_anonportmax, 0,
1767 CTL_NET, PF_INET6, IPPROTO_IPV6,
1768 IPV6CTL_ANONPORTMAX, CTL_EOL);
1769 #ifndef IPNOPRIVPORTS
1770 sysctl_createv(clog, 0, NULL, NULL,
1771 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1772 CTLTYPE_INT, "lowportmin",
1773 SYSCTL_DESCR("Lowest privileged ephemeral port number "
1774 "to assign"),
1775 sysctl_net_inet_ip_ports, 0, &ip6_lowportmin, 0,
1776 CTL_NET, PF_INET6, IPPROTO_IPV6,
1777 IPV6CTL_LOWPORTMIN, CTL_EOL);
1778 sysctl_createv(clog, 0, NULL, NULL,
1779 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1780 CTLTYPE_INT, "lowportmax",
1781 SYSCTL_DESCR("Highest privileged ephemeral port number "
1782 "to assign"),
1783 sysctl_net_inet_ip_ports, 0, &ip6_lowportmax, 0,
1784 CTL_NET, PF_INET6, IPPROTO_IPV6,
1785 IPV6CTL_LOWPORTMAX, CTL_EOL);
1786 #endif /* IPNOPRIVPORTS */
1787 sysctl_createv(clog, 0, NULL, NULL,
1788 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1789 CTLTYPE_INT, "auto_linklocal",
1790 SYSCTL_DESCR("Default value of per-interface flag for "
1791 "adding an IPv6 link-local address to "
1792 "interfaces when attached"),
1793 NULL, 0, &ip6_auto_linklocal, 0,
1794 CTL_NET, PF_INET6, IPPROTO_IPV6,
1795 IPV6CTL_AUTO_LINKLOCAL, CTL_EOL);
1796 sysctl_createv(clog, 0, NULL, NULL,
1797 CTLFLAG_PERMANENT|CTLFLAG_READONLY,
1798 CTLTYPE_STRUCT, "addctlpolicy",
1799 SYSCTL_DESCR("Return the current address control"
1800 " policy"),
1801 sysctl_net_inet6_addrctlpolicy, 0, NULL, 0,
1802 CTL_NET, PF_INET6, IPPROTO_IPV6,
1803 IPV6CTL_ADDRCTLPOLICY, CTL_EOL);
1804 sysctl_createv(clog, 0, NULL, NULL,
1805 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1806 CTLTYPE_INT, "use_tempaddr",
1807 SYSCTL_DESCR("Use temporary address"),
1808 NULL, 0, &ip6_use_tempaddr, 0,
1809 CTL_NET, PF_INET6, IPPROTO_IPV6,
1810 CTL_CREATE, CTL_EOL);
1811 sysctl_createv(clog, 0, NULL, NULL,
1812 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1813 CTLTYPE_INT, "prefer_tempaddr",
1814 SYSCTL_DESCR("Prefer temporary address as source "
1815 "address"),
1816 NULL, 0, &ip6_prefer_tempaddr, 0,
1817 CTL_NET, PF_INET6, IPPROTO_IPV6,
1818 CTL_CREATE, CTL_EOL);
1819 sysctl_createv(clog, 0, NULL, NULL,
1820 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1821 CTLTYPE_INT, "temppltime",
1822 SYSCTL_DESCR("preferred lifetime of a temporary address"),
1823 NULL, 0, &ip6_temp_preferred_lifetime, 0,
1824 CTL_NET, PF_INET6, IPPROTO_IPV6,
1825 CTL_CREATE, CTL_EOL);
1826 sysctl_createv(clog, 0, NULL, NULL,
1827 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1828 CTLTYPE_INT, "tempvltime",
1829 SYSCTL_DESCR("valid lifetime of a temporary address"),
1830 NULL, 0, &ip6_temp_valid_lifetime, 0,
1831 CTL_NET, PF_INET6, IPPROTO_IPV6,
1832 CTL_CREATE, CTL_EOL);
1833 sysctl_createv(clog, 0, NULL, NULL,
1834 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1835 CTLTYPE_INT, "maxfrags",
1836 SYSCTL_DESCR("Maximum fragments in reassembly queue"),
1837 NULL, 0, &ip6_maxfrags, 0,
1838 CTL_NET, PF_INET6, IPPROTO_IPV6,
1839 IPV6CTL_MAXFRAGS, CTL_EOL);
1840 sysctl_createv(clog, 0, NULL, NULL,
1841 CTLFLAG_PERMANENT,
1842 CTLTYPE_STRUCT, "stats",
1843 SYSCTL_DESCR("IPv6 statistics"),
1844 sysctl_net_inet6_ip6_stats, 0, NULL, 0,
1845 CTL_NET, PF_INET6, IPPROTO_IPV6,
1846 IPV6CTL_STATS, CTL_EOL);
1847 sysctl_createv(clog, 0, NULL, NULL,
1848 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1849 CTLTYPE_INT, "use_defaultzone",
1850 SYSCTL_DESCR("Whether to use the default scope zones"),
1851 NULL, 0, &ip6_use_defzone, 0,
1852 CTL_NET, PF_INET6, IPPROTO_IPV6,
1853 IPV6CTL_USE_DEFAULTZONE, CTL_EOL);
1854 sysctl_createv(clog, 0, NULL, NULL,
1855 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1856 CTLTYPE_INT, "mcast_pmtu",
1857 SYSCTL_DESCR("Enable pMTU discovery for multicast packet"),
1858 NULL, 0, &ip6_mcast_pmtu, 0,
1859 CTL_NET, PF_INET6, IPPROTO_IPV6,
1860 CTL_CREATE, CTL_EOL);
1861 /* anonportalgo RFC6056 subtree */
1862 const struct sysctlnode *portalgo_node;
1863 sysctl_createv(clog, 0, NULL, &portalgo_node,
1864 CTLFLAG_PERMANENT,
1865 CTLTYPE_NODE, "anonportalgo",
1866 SYSCTL_DESCR("Anonymous port algorithm selection (RFC 6056)"),
1867 NULL, 0, NULL, 0,
1868 CTL_NET, PF_INET6, IPPROTO_IPV6, CTL_CREATE, CTL_EOL);
1869 sysctl_createv(clog, 0, &portalgo_node, NULL,
1870 CTLFLAG_PERMANENT,
1871 CTLTYPE_STRING, "available",
1872 SYSCTL_DESCR("available algorithms"),
1873 sysctl_portalgo_available, 0, NULL, PORTALGO_MAXLEN,
1874 CTL_CREATE, CTL_EOL);
1875 sysctl_createv(clog, 0, &portalgo_node, NULL,
1876 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1877 CTLTYPE_STRING, "selected",
1878 SYSCTL_DESCR("selected algorithm"),
1879 sysctl_portalgo_selected6, 0, NULL, PORTALGO_MAXLEN,
1880 CTL_CREATE, CTL_EOL);
1881 sysctl_createv(clog, 0, &portalgo_node, NULL,
1882 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1883 CTLTYPE_STRUCT, "reserve",
1884 SYSCTL_DESCR("bitmap of reserved ports"),
1885 sysctl_portalgo_reserve6, 0, NULL, 0,
1886 CTL_CREATE, CTL_EOL);
1887 sysctl_createv(clog, 0, NULL, NULL,
1888 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1889 CTLTYPE_INT, "neighborgcthresh",
1890 SYSCTL_DESCR("Maximum number of entries in neighbor"
1891 " cache"),
1892 NULL, 1, &ip6_neighborgcthresh, 0,
1893 CTL_NET, PF_INET6, IPPROTO_IPV6,
1894 CTL_CREATE, CTL_EOL);
1895 sysctl_createv(clog, 0, NULL, NULL,
1896 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1897 CTLTYPE_INT, "maxifprefixes",
1898 SYSCTL_DESCR("Maximum number of prefixes created by"
1899 " route advertisement per interface"),
1900 NULL, 1, &ip6_maxifprefixes, 0,
1901 CTL_NET, PF_INET6, IPPROTO_IPV6,
1902 CTL_CREATE, CTL_EOL);
1903 sysctl_createv(clog, 0, NULL, NULL,
1904 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1905 CTLTYPE_INT, "maxifdefrouters",
1906 SYSCTL_DESCR("Maximum number of default routers created"
1907 " by route advertisement per interface"),
1908 NULL, 1, &ip6_maxifdefrouters, 0,
1909 CTL_NET, PF_INET6, IPPROTO_IPV6,
1910 CTL_CREATE, CTL_EOL);
1911 sysctl_createv(clog, 0, NULL, NULL,
1912 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1913 CTLTYPE_INT, "maxdynroutes",
1914 SYSCTL_DESCR("Maximum number of routes created via"
1915 " redirect"),
1916 NULL, 1, &ip6_maxdynroutes, 0,
1917 CTL_NET, PF_INET6, IPPROTO_IPV6,
1918 CTL_CREATE, CTL_EOL);
1919 }
1920
1921 void
1922 ip6_statinc(u_int stat)
1923 {
1924
1925 KASSERT(stat < IP6_NSTATS);
1926 IP6_STATINC(stat);
1927 }
1928