1 /* $NetBSD: ip6_input.c,v 1.164 2016/07/07 09:32:03 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 1.164 2016/07/07 09:32:03 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 #endif
74
75 #include <sys/param.h>
76 #include <sys/systm.h>
77 #include <sys/malloc.h>
78 #include <sys/mbuf.h>
79 #include <sys/domain.h>
80 #include <sys/protosw.h>
81 #include <sys/socket.h>
82 #include <sys/socketvar.h>
83 #include <sys/errno.h>
84 #include <sys/time.h>
85 #include <sys/kernel.h>
86 #include <sys/syslog.h>
87 #include <sys/proc.h>
88 #include <sys/sysctl.h>
89 #include <sys/cprng.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 #ifdef COMPAT_50
123 #include <compat/sys/time.h>
124 #include <compat/sys/socket.h>
125 #endif
126
127 #include <netinet6/ip6protosw.h>
128
129 #include "faith.h"
130
131 #include <net/net_osdep.h>
132
133 extern struct domain inet6domain;
134
135 u_char ip6_protox[IPPROTO_MAX];
136 pktqueue_t *ip6_pktq __read_mostly;
137
138 int ip6_forward_srcrt; /* XXX */
139 int ip6_sourcecheck; /* XXX */
140 int ip6_sourcecheck_interval; /* XXX */
141
142 pfil_head_t *inet6_pfil_hook;
143
144 percpu_t *ip6stat_percpu;
145
146 static void ip6_init2(void);
147 static void ip6intr(void *);
148 static struct m_tag *ip6_setdstifaddr(struct mbuf *, const struct in6_ifaddr *);
149
150 static int ip6_process_hopopts(struct mbuf *, u_int8_t *, int, u_int32_t *,
151 u_int32_t *);
152 static struct mbuf *ip6_pullexthdr(struct mbuf *, size_t, int);
153 static void sysctl_net_inet6_ip6_setup(struct sysctllog **);
154
155 /*
156 * IP6 initialization: fill in IP6 protocol switch table.
157 * All protocols not implemented in kernel go to raw IP6 protocol handler.
158 */
159 void
ip6_init(void)160 ip6_init(void)
161 {
162 const struct ip6protosw *pr;
163 int i;
164
165 in6_init();
166
167 sysctl_net_inet6_ip6_setup(NULL);
168 pr = (const struct ip6protosw *)pffindproto(PF_INET6, IPPROTO_RAW, SOCK_RAW);
169 if (pr == 0)
170 panic("ip6_init");
171 for (i = 0; i < IPPROTO_MAX; i++)
172 ip6_protox[i] = pr - inet6sw;
173 for (pr = (const struct ip6protosw *)inet6domain.dom_protosw;
174 pr < (const struct ip6protosw *)inet6domain.dom_protoswNPROTOSW; pr++)
175 if (pr->pr_domain->dom_family == PF_INET6 &&
176 pr->pr_protocol && pr->pr_protocol != IPPROTO_RAW)
177 ip6_protox[pr->pr_protocol] = pr - inet6sw;
178
179 ip6_pktq = pktq_create(IFQ_MAXLEN, ip6intr, NULL);
180 KASSERT(ip6_pktq != NULL);
181
182 scope6_init();
183 addrsel_policy_init();
184 nd6_init();
185 frag6_init();
186 ip6_desync_factor = cprng_fast32() % MAX_TEMP_DESYNC_FACTOR;
187
188 ip6_init2();
189 #ifdef GATEWAY
190 ip6flow_init(ip6_hashsize);
191 #endif
192 /* Register our Packet Filter hook. */
193 inet6_pfil_hook = pfil_head_create(PFIL_TYPE_AF, (void *)AF_INET6);
194 KASSERT(inet6_pfil_hook != NULL);
195
196 ip6stat_percpu = percpu_alloc(sizeof(uint64_t) * IP6_NSTATS);
197 }
198
199 static void
ip6_init2(void)200 ip6_init2(void)
201 {
202
203 /* timer for regeneranation of temporary addresses randomize ID */
204 callout_init(&in6_tmpaddrtimer_ch, CALLOUT_MPSAFE);
205 callout_reset(&in6_tmpaddrtimer_ch,
206 (ip6_temp_preferred_lifetime - ip6_desync_factor -
207 ip6_temp_regen_advance) * hz,
208 in6_tmpaddrtimer, NULL);
209 }
210
211 /*
212 * IP6 input interrupt handling. Just pass the packet to ip6_input.
213 */
214 static void
ip6intr(void * arg __unused)215 ip6intr(void *arg __unused)
216 {
217 struct mbuf *m;
218
219 mutex_enter(softnet_lock);
220 while ((m = pktq_dequeue(ip6_pktq)) != NULL) {
221 struct psref psref;
222 struct ifnet *rcvif = m_get_rcvif_psref(m, &psref);
223
224 if (rcvif == NULL) {
225 m_freem(m);
226 continue;
227 }
228 /*
229 * Drop the packet if IPv6 is disabled on the interface.
230 */
231 if ((ND_IFINFO(rcvif)->flags & ND6_IFF_IFDISABLED)) {
232 m_put_rcvif_psref(rcvif, &psref);
233 m_freem(m);
234 continue;
235 }
236 ip6_input(m, rcvif);
237 m_put_rcvif_psref(rcvif, &psref);
238 }
239 mutex_exit(softnet_lock);
240 }
241
242 extern struct route ip6_forward_rt;
243
244 void
ip6_input(struct mbuf * m,struct ifnet * rcvif)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 const struct rtentry *rt;
255 union {
256 struct sockaddr dst;
257 struct sockaddr_in6 dst6;
258 } u;
259
260 /*
261 * make sure we don't have onion peering information into m_tag.
262 */
263 ip6_delaux(m);
264
265 /*
266 * mbuf statistics
267 */
268 if (m->m_flags & M_EXT) {
269 if (m->m_next)
270 IP6_STATINC(IP6_STAT_MEXT2M);
271 else
272 IP6_STATINC(IP6_STAT_MEXT1);
273 } else {
274 #define M2MMAX 32
275 if (m->m_next) {
276 if (m->m_flags & M_LOOP)
277 /*XXX*/ IP6_STATINC(IP6_STAT_M2M + lo0ifp->if_index);
278 else if (rcvif->if_index < M2MMAX)
279 IP6_STATINC(IP6_STAT_M2M + rcvif->if_index);
280 else
281 IP6_STATINC(IP6_STAT_M2M);
282 } else
283 IP6_STATINC(IP6_STAT_M1);
284 #undef M2MMAX
285 }
286
287 in6_ifstat_inc(rcvif, ifs6_in_receive);
288 IP6_STATINC(IP6_STAT_TOTAL);
289
290 /*
291 * If the IPv6 header is not aligned, slurp it up into a new
292 * mbuf with space for link headers, in the event we forward
293 * it. Otherwise, if it is aligned, make sure the entire base
294 * IPv6 header is in the first mbuf of the chain.
295 */
296 if (IP6_HDR_ALIGNED_P(mtod(m, void *)) == 0) {
297 if ((m = m_copyup(m, sizeof(struct ip6_hdr),
298 (max_linkhdr + 3) & ~3)) == NULL) {
299 /* XXXJRT new stat, please */
300 IP6_STATINC(IP6_STAT_TOOSMALL);
301 in6_ifstat_inc(rcvif, ifs6_in_hdrerr);
302 return;
303 }
304 } else if (__predict_false(m->m_len < sizeof(struct ip6_hdr))) {
305 if ((m = m_pullup(m, sizeof(struct ip6_hdr))) == NULL) {
306 IP6_STATINC(IP6_STAT_TOOSMALL);
307 in6_ifstat_inc(rcvif, ifs6_in_hdrerr);
308 return;
309 }
310 }
311
312 ip6 = mtod(m, struct ip6_hdr *);
313
314 if ((ip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION) {
315 IP6_STATINC(IP6_STAT_BADVERS);
316 in6_ifstat_inc(rcvif, ifs6_in_hdrerr);
317 goto bad;
318 }
319
320 /*
321 * Assume that we can create a fast-forward IP flow entry
322 * based on this packet.
323 */
324 m->m_flags |= M_CANFASTFWD;
325
326 /*
327 * Run through list of hooks for input packets. If there are any
328 * filters which require that additional packets in the flow are
329 * not fast-forwarded, they must clear the M_CANFASTFWD flag.
330 * Note that filters must _never_ set this flag, as another filter
331 * in the list may have previously cleared it.
332 */
333 /*
334 * let ipfilter look at packet on the wire,
335 * not the decapsulated packet.
336 */
337 #if defined(IPSEC)
338 if (!ipsec_used || !ipsec_indone(m))
339 #else
340 if (1)
341 #endif
342 {
343 struct in6_addr odst;
344
345 odst = ip6->ip6_dst;
346 if (pfil_run_hooks(inet6_pfil_hook, &m, rcvif, PFIL_IN) != 0)
347 return;
348 if (m == NULL)
349 return;
350 ip6 = mtod(m, struct ip6_hdr *);
351 srcrt = !IN6_ARE_ADDR_EQUAL(&odst, &ip6->ip6_dst);
352 }
353
354 IP6_STATINC(IP6_STAT_NXTHIST + ip6->ip6_nxt);
355
356 #ifdef ALTQ
357 if (altq_input != NULL && (*altq_input)(m, AF_INET6) == 0) {
358 /* packet is dropped by traffic conditioner */
359 return;
360 }
361 #endif
362
363 /*
364 * Check against address spoofing/corruption.
365 */
366 if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_src) ||
367 IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_dst)) {
368 /*
369 * XXX: "badscope" is not very suitable for a multicast source.
370 */
371 IP6_STATINC(IP6_STAT_BADSCOPE);
372 in6_ifstat_inc(rcvif, ifs6_in_addrerr);
373 goto bad;
374 }
375 /*
376 * The following check is not documented in specs. A malicious
377 * party may be able to use IPv4 mapped addr to confuse tcp/udp stack
378 * and bypass security checks (act as if it was from 127.0.0.1 by using
379 * IPv6 src ::ffff:127.0.0.1). Be cautious.
380 *
381 * This check chokes if we are in an SIIT cloud. As none of BSDs
382 * support IPv4-less kernel compilation, we cannot support SIIT
383 * environment at all. So, it makes more sense for us to reject any
384 * malicious packets for non-SIIT environment, than try to do a
385 * partial support for SIIT environment.
386 */
387 if (IN6_IS_ADDR_V4MAPPED(&ip6->ip6_src) ||
388 IN6_IS_ADDR_V4MAPPED(&ip6->ip6_dst)) {
389 IP6_STATINC(IP6_STAT_BADSCOPE);
390 in6_ifstat_inc(rcvif, ifs6_in_addrerr);
391 goto bad;
392 }
393 #if 0
394 /*
395 * Reject packets with IPv4 compatible addresses (auto tunnel).
396 *
397 * The code forbids auto tunnel relay case in RFC1933 (the check is
398 * stronger than RFC1933). We may want to re-enable it if mech-xx
399 * is revised to forbid relaying case.
400 */
401 if (IN6_IS_ADDR_V4COMPAT(&ip6->ip6_src) ||
402 IN6_IS_ADDR_V4COMPAT(&ip6->ip6_dst)) {
403 IP6_STATINC(IP6_STAT_BADSCOPE);
404 in6_ifstat_inc(rcvif, ifs6_in_addrerr);
405 goto bad;
406 }
407 #endif
408
409 /*
410 * Disambiguate address scope zones (if there is ambiguity).
411 * We first make sure that the original source or destination address
412 * is not in our internal form for scoped addresses. Such addresses
413 * are not necessarily invalid spec-wise, but we cannot accept them due
414 * to the usage conflict.
415 * in6_setscope() then also checks and rejects the cases where src or
416 * dst are the loopback address and the receiving interface
417 * is not loopback.
418 */
419 if (__predict_false(
420 m_makewritable(&m, 0, sizeof(struct ip6_hdr), M_DONTWAIT)))
421 goto bad;
422 ip6 = mtod(m, struct ip6_hdr *);
423 if (in6_clearscope(&ip6->ip6_src) || in6_clearscope(&ip6->ip6_dst)) {
424 IP6_STATINC(IP6_STAT_BADSCOPE); /* XXX */
425 goto bad;
426 }
427 if (in6_setscope(&ip6->ip6_src, rcvif, NULL) ||
428 in6_setscope(&ip6->ip6_dst, rcvif, NULL)) {
429 IP6_STATINC(IP6_STAT_BADSCOPE);
430 goto bad;
431 }
432
433 /*
434 * Multicast check
435 */
436 if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) {
437 struct in6_multi *in6m = 0;
438
439 in6_ifstat_inc(rcvif, ifs6_in_mcast);
440 /*
441 * See if we belong to the destination multicast group on the
442 * arrival interface.
443 */
444 IN6_LOOKUP_MULTI(ip6->ip6_dst, rcvif, in6m);
445 if (in6m)
446 ours = 1;
447 else if (!ip6_mrouter) {
448 uint64_t *ip6s = IP6_STAT_GETREF();
449 ip6s[IP6_STAT_NOTMEMBER]++;
450 ip6s[IP6_STAT_CANTFORWARD]++;
451 IP6_STAT_PUTREF();
452 in6_ifstat_inc(rcvif, ifs6_in_discard);
453 goto bad;
454 }
455 deliverifp = rcvif;
456 goto hbhcheck;
457 }
458
459 sockaddr_in6_init(&u.dst6, &ip6->ip6_dst, 0, 0, 0);
460
461 /*
462 * Unicast check
463 */
464 rt = rtcache_lookup2(&ip6_forward_rt, &u.dst, 1, &hit);
465 if (hit)
466 IP6_STATINC(IP6_STAT_FORWARD_CACHEHIT);
467 else
468 IP6_STATINC(IP6_STAT_FORWARD_CACHEMISS);
469
470 #define rt6_getkey(__rt) satocsin6(rt_getkey(__rt))
471
472 /*
473 * Accept the packet if the forwarding interface to the destination
474 * according to the routing table is the loopback interface,
475 * unless the associated route has a gateway.
476 * Note that this approach causes to accept a packet if there is a
477 * route to the loopback interface for the destination of the packet.
478 * But we think it's even useful in some situations, e.g. when using
479 * a special daemon which wants to intercept the packet.
480 */
481 if (rt != NULL &&
482 (rt->rt_flags & (RTF_HOST|RTF_GATEWAY)) == RTF_HOST &&
483 #if 0
484 /*
485 * The check below is redundant since the comparison of
486 * the destination and the key of the rtentry has
487 * already done through looking up the routing table.
488 */
489 IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst, &rt6_getkey(rt)->sin6_addr) &&
490 #endif
491 rt->rt_ifp->if_type == IFT_LOOP) {
492 struct in6_ifaddr *ia6 = (struct in6_ifaddr *)rt->rt_ifa;
493 if (ia6->ia6_flags & IN6_IFF_ANYCAST)
494 m->m_flags |= M_ANYCAST6;
495 /*
496 * packets to a tentative, duplicated, or somehow invalid
497 * address must not be accepted.
498 */
499 if (!(ia6->ia6_flags & IN6_IFF_NOTREADY)) {
500 /* this address is ready */
501 ours = 1;
502 deliverifp = ia6->ia_ifp; /* correct? */
503 goto hbhcheck;
504 } else {
505 /* address is not ready, so discard the packet. */
506 nd6log(LOG_INFO, "packet to an unready address %s->%s\n",
507 ip6_sprintf(&ip6->ip6_src),
508 ip6_sprintf(&ip6->ip6_dst));
509
510 goto bad;
511 }
512 }
513
514 /*
515 * FAITH (Firewall Aided Internet Translator)
516 */
517 #if defined(NFAITH) && 0 < NFAITH
518 if (ip6_keepfaith) {
519 if (rt != NULL && rt->rt_ifp != NULL &&
520 rt->rt_ifp->if_type == IFT_FAITH) {
521 /* XXX do we need more sanity checks? */
522 ours = 1;
523 deliverifp = rt->rt_ifp; /* faith */
524 goto hbhcheck;
525 }
526 }
527 #endif
528
529 #if 0
530 {
531 /*
532 * Last resort: check in6_ifaddr for incoming interface.
533 * The code is here until I update the "goto ours hack" code above
534 * working right.
535 */
536 struct ifaddr *ifa;
537 IFADDR_READER_FOREACH(ifa, rcvif) {
538 if (ifa->ifa_addr->sa_family != AF_INET6)
539 continue;
540 if (IN6_ARE_ADDR_EQUAL(IFA_IN6(ifa), &ip6->ip6_dst)) {
541 ours = 1;
542 deliverifp = ifa->ifa_ifp;
543 goto hbhcheck;
544 }
545 }
546 }
547 #endif
548
549 /*
550 * Now there is no reason to process the packet if it's not our own
551 * and we're not a router.
552 */
553 if (!ip6_forwarding) {
554 IP6_STATINC(IP6_STAT_CANTFORWARD);
555 in6_ifstat_inc(rcvif, ifs6_in_discard);
556 goto bad;
557 }
558
559 hbhcheck:
560 /*
561 * record address information into m_tag, if we don't have one yet.
562 * note that we are unable to record it, if the address is not listed
563 * as our interface address (e.g. multicast addresses, addresses
564 * within FAITH prefixes and such).
565 */
566 if (deliverifp && ip6_getdstifaddr(m) == NULL) {
567 struct in6_ifaddr *ia6;
568
569 ia6 = in6_ifawithifp(deliverifp, &ip6->ip6_dst);
570 if (ia6 != NULL && ip6_setdstifaddr(m, ia6) == NULL) {
571 /*
572 * XXX maybe we should drop the packet here,
573 * as we could not provide enough information
574 * to the upper layers.
575 */
576 }
577 }
578
579 /*
580 * Process Hop-by-Hop options header if it's contained.
581 * m may be modified in ip6_hopopts_input().
582 * If a JumboPayload option is included, plen will also be modified.
583 */
584 plen = (u_int32_t)ntohs(ip6->ip6_plen);
585 if (ip6->ip6_nxt == IPPROTO_HOPOPTS) {
586 struct ip6_hbh *hbh;
587
588 if (ip6_hopopts_input(&plen, &rtalert, &m, &off)) {
589 #if 0 /*touches NULL pointer*/
590 in6_ifstat_inc(rcvif, ifs6_in_discard);
591 #endif
592 return; /* m have already been freed */
593 }
594
595 /* adjust pointer */
596 ip6 = mtod(m, struct ip6_hdr *);
597
598 /*
599 * if the payload length field is 0 and the next header field
600 * indicates Hop-by-Hop Options header, then a Jumbo Payload
601 * option MUST be included.
602 */
603 if (ip6->ip6_plen == 0 && plen == 0) {
604 /*
605 * Note that if a valid jumbo payload option is
606 * contained, ip6_hopopts_input() must set a valid
607 * (non-zero) payload length to the variable plen.
608 */
609 IP6_STATINC(IP6_STAT_BADOPTIONS);
610 in6_ifstat_inc(rcvif, ifs6_in_discard);
611 in6_ifstat_inc(rcvif, ifs6_in_hdrerr);
612 icmp6_error(m, ICMP6_PARAM_PROB,
613 ICMP6_PARAMPROB_HEADER,
614 (char *)&ip6->ip6_plen - (char *)ip6);
615 return;
616 }
617 IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m, sizeof(struct ip6_hdr),
618 sizeof(struct ip6_hbh));
619 if (hbh == NULL) {
620 IP6_STATINC(IP6_STAT_TOOSHORT);
621 return;
622 }
623 KASSERT(IP6_HDR_ALIGNED_P(hbh));
624 nxt = hbh->ip6h_nxt;
625
626 /*
627 * accept the packet if a router alert option is included
628 * and we act as an IPv6 router.
629 */
630 if (rtalert != ~0 && ip6_forwarding)
631 ours = 1;
632 } else
633 nxt = ip6->ip6_nxt;
634
635 /*
636 * Check that the amount of data in the buffers
637 * is as at least much as the IPv6 header would have us expect.
638 * Trim mbufs if longer than we expect.
639 * Drop packet if shorter than we expect.
640 */
641 if (m->m_pkthdr.len - sizeof(struct ip6_hdr) < plen) {
642 IP6_STATINC(IP6_STAT_TOOSHORT);
643 in6_ifstat_inc(rcvif, ifs6_in_truncated);
644 goto bad;
645 }
646 if (m->m_pkthdr.len > sizeof(struct ip6_hdr) + plen) {
647 if (m->m_len == m->m_pkthdr.len) {
648 m->m_len = sizeof(struct ip6_hdr) + plen;
649 m->m_pkthdr.len = sizeof(struct ip6_hdr) + plen;
650 } else
651 m_adj(m, sizeof(struct ip6_hdr) + plen - m->m_pkthdr.len);
652 }
653
654 /*
655 * Forward if desirable.
656 */
657 if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) {
658 /*
659 * If we are acting as a multicast router, all
660 * incoming multicast packets are passed to the
661 * kernel-level multicast forwarding function.
662 * The packet is returned (relatively) intact; if
663 * ip6_mforward() returns a non-zero value, the packet
664 * must be discarded, else it may be accepted below.
665 */
666 if (ip6_mrouter && ip6_mforward(ip6, rcvif, m)) {
667 IP6_STATINC(IP6_STAT_CANTFORWARD);
668 m_freem(m);
669 return;
670 }
671 if (!ours) {
672 m_freem(m);
673 return;
674 }
675 } else if (!ours) {
676 ip6_forward(m, srcrt);
677 return;
678 }
679
680 ip6 = mtod(m, struct ip6_hdr *);
681
682 /*
683 * Malicious party may be able to use IPv4 mapped addr to confuse
684 * tcp/udp stack and bypass security checks (act as if it was from
685 * 127.0.0.1 by using IPv6 src ::ffff:127.0.0.1). Be cautious.
686 *
687 * For SIIT end node behavior, you may want to disable the check.
688 * However, you will become vulnerable to attacks using IPv4 mapped
689 * source.
690 */
691 if (IN6_IS_ADDR_V4MAPPED(&ip6->ip6_src) ||
692 IN6_IS_ADDR_V4MAPPED(&ip6->ip6_dst)) {
693 IP6_STATINC(IP6_STAT_BADSCOPE);
694 in6_ifstat_inc(rcvif, ifs6_in_addrerr);
695 goto bad;
696 }
697
698 /*
699 * Tell launch routine the next header
700 */
701 #ifdef IFA_STATS
702 if (deliverifp != NULL) {
703 struct in6_ifaddr *ia6;
704 ia6 = in6_ifawithifp(deliverifp, &ip6->ip6_dst);
705 if (ia6)
706 ia6->ia_ifa.ifa_data.ifad_inbytes += m->m_pkthdr.len;
707 }
708 #endif
709 IP6_STATINC(IP6_STAT_DELIVERED);
710 in6_ifstat_inc(deliverifp, ifs6_in_deliver);
711 nest = 0;
712
713 rh_present = 0;
714 while (nxt != IPPROTO_DONE) {
715 if (ip6_hdrnestlimit && (++nest > ip6_hdrnestlimit)) {
716 IP6_STATINC(IP6_STAT_TOOMANYHDR);
717 in6_ifstat_inc(rcvif, ifs6_in_hdrerr);
718 goto bad;
719 }
720
721 /*
722 * protection against faulty packet - there should be
723 * more sanity checks in header chain processing.
724 */
725 if (m->m_pkthdr.len < off) {
726 IP6_STATINC(IP6_STAT_TOOSHORT);
727 in6_ifstat_inc(rcvif, ifs6_in_truncated);
728 goto bad;
729 }
730
731 if (nxt == IPPROTO_ROUTING) {
732 if (rh_present++) {
733 in6_ifstat_inc(rcvif, ifs6_in_hdrerr);
734 IP6_STATINC(IP6_STAT_BADOPTIONS);
735 goto bad;
736 }
737 }
738
739 #ifdef IPSEC
740 if (ipsec_used) {
741 /*
742 * enforce IPsec policy checking if we are seeing last
743 * header. note that we do not visit this with
744 * protocols with pcb layer code - like udp/tcp/raw ip.
745 */
746 if ((inet6sw[ip_protox[nxt]].pr_flags
747 & PR_LASTHDR) != 0) {
748 int error = ipsec6_input(m);
749 if (error)
750 goto bad;
751 }
752 }
753 #endif /* IPSEC */
754
755 nxt = (*inet6sw[ip6_protox[nxt]].pr_input)(&m, &off, nxt);
756 }
757 return;
758 bad:
759 m_freem(m);
760 }
761
762 /*
763 * set/grab in6_ifaddr correspond to IPv6 destination address.
764 */
765 static struct m_tag *
ip6_setdstifaddr(struct mbuf * m,const struct in6_ifaddr * ia)766 ip6_setdstifaddr(struct mbuf *m, const struct in6_ifaddr *ia)
767 {
768 struct m_tag *mtag;
769 struct ip6aux *ip6a;
770
771 mtag = ip6_addaux(m);
772 if (mtag == NULL)
773 return NULL;
774
775 ip6a = (struct ip6aux *)(mtag + 1);
776 if (in6_setscope(&ip6a->ip6a_src, ia->ia_ifp, &ip6a->ip6a_scope_id)) {
777 IP6_STATINC(IP6_STAT_BADSCOPE);
778 return NULL;
779 }
780
781 ip6a->ip6a_src = ia->ia_addr.sin6_addr;
782 ip6a->ip6a_flags = ia->ia6_flags;
783 return mtag;
784 }
785
786 const struct ip6aux *
ip6_getdstifaddr(struct mbuf * m)787 ip6_getdstifaddr(struct mbuf *m)
788 {
789 struct m_tag *mtag;
790
791 mtag = ip6_findaux(m);
792 if (mtag != NULL)
793 return (struct ip6aux *)(mtag + 1);
794 else
795 return NULL;
796 }
797
798 /*
799 * Hop-by-Hop options header processing. If a valid jumbo payload option is
800 * included, the real payload length will be stored in plenp.
801 *
802 * rtalertp - XXX: should be stored more smart way
803 */
804 int
ip6_hopopts_input(u_int32_t * plenp,u_int32_t * rtalertp,struct mbuf ** mp,int * offp)805 ip6_hopopts_input(u_int32_t *plenp, u_int32_t *rtalertp,
806 struct mbuf **mp, int *offp)
807 {
808 struct mbuf *m = *mp;
809 int off = *offp, hbhlen;
810 struct ip6_hbh *hbh;
811
812 /* validation of the length of the header */
813 IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m,
814 sizeof(struct ip6_hdr), sizeof(struct ip6_hbh));
815 if (hbh == NULL) {
816 IP6_STATINC(IP6_STAT_TOOSHORT);
817 return -1;
818 }
819 hbhlen = (hbh->ip6h_len + 1) << 3;
820 IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m, sizeof(struct ip6_hdr),
821 hbhlen);
822 if (hbh == NULL) {
823 IP6_STATINC(IP6_STAT_TOOSHORT);
824 return -1;
825 }
826 KASSERT(IP6_HDR_ALIGNED_P(hbh));
827 off += hbhlen;
828 hbhlen -= sizeof(struct ip6_hbh);
829
830 if (ip6_process_hopopts(m, (u_int8_t *)hbh + sizeof(struct ip6_hbh),
831 hbhlen, rtalertp, plenp) < 0)
832 return (-1);
833
834 *offp = off;
835 *mp = m;
836 return (0);
837 }
838
839 /*
840 * Search header for all Hop-by-hop options and process each option.
841 * This function is separate from ip6_hopopts_input() in order to
842 * handle a case where the sending node itself process its hop-by-hop
843 * options header. In such a case, the function is called from ip6_output().
844 *
845 * The function assumes that hbh header is located right after the IPv6 header
846 * (RFC2460 p7), opthead is pointer into data content in m, and opthead to
847 * opthead + hbhlen is located in continuous memory region.
848 */
849 static int
ip6_process_hopopts(struct mbuf * m,u_int8_t * opthead,int hbhlen,u_int32_t * rtalertp,u_int32_t * plenp)850 ip6_process_hopopts(struct mbuf *m, u_int8_t *opthead, int hbhlen,
851 u_int32_t *rtalertp, u_int32_t *plenp)
852 {
853 struct ip6_hdr *ip6;
854 int optlen = 0;
855 u_int8_t *opt = opthead;
856 u_int16_t rtalert_val;
857 u_int32_t jumboplen;
858 const int erroff = sizeof(struct ip6_hdr) + sizeof(struct ip6_hbh);
859
860 for (; hbhlen > 0; hbhlen -= optlen, opt += optlen) {
861 switch (*opt) {
862 case IP6OPT_PAD1:
863 optlen = 1;
864 break;
865 case IP6OPT_PADN:
866 if (hbhlen < IP6OPT_MINLEN) {
867 IP6_STATINC(IP6_STAT_TOOSMALL);
868 goto bad;
869 }
870 optlen = *(opt + 1) + 2;
871 break;
872 case IP6OPT_RTALERT:
873 /* XXX may need check for alignment */
874 if (hbhlen < IP6OPT_RTALERT_LEN) {
875 IP6_STATINC(IP6_STAT_TOOSMALL);
876 goto bad;
877 }
878 if (*(opt + 1) != IP6OPT_RTALERT_LEN - 2) {
879 /* XXX stat */
880 icmp6_error(m, ICMP6_PARAM_PROB,
881 ICMP6_PARAMPROB_HEADER,
882 erroff + opt + 1 - opthead);
883 return (-1);
884 }
885 optlen = IP6OPT_RTALERT_LEN;
886 memcpy((void *)&rtalert_val, (void *)(opt + 2), 2);
887 *rtalertp = ntohs(rtalert_val);
888 break;
889 case IP6OPT_JUMBO:
890 /* XXX may need check for alignment */
891 if (hbhlen < IP6OPT_JUMBO_LEN) {
892 IP6_STATINC(IP6_STAT_TOOSMALL);
893 goto bad;
894 }
895 if (*(opt + 1) != IP6OPT_JUMBO_LEN - 2) {
896 /* XXX stat */
897 icmp6_error(m, ICMP6_PARAM_PROB,
898 ICMP6_PARAMPROB_HEADER,
899 erroff + opt + 1 - opthead);
900 return (-1);
901 }
902 optlen = IP6OPT_JUMBO_LEN;
903
904 /*
905 * IPv6 packets that have non 0 payload length
906 * must not contain a jumbo payload option.
907 */
908 ip6 = mtod(m, struct ip6_hdr *);
909 if (ip6->ip6_plen) {
910 IP6_STATINC(IP6_STAT_BADOPTIONS);
911 icmp6_error(m, ICMP6_PARAM_PROB,
912 ICMP6_PARAMPROB_HEADER,
913 erroff + opt - opthead);
914 return (-1);
915 }
916
917 /*
918 * We may see jumbolen in unaligned location, so
919 * we'd need to perform bcopy().
920 */
921 memcpy(&jumboplen, opt + 2, sizeof(jumboplen));
922 jumboplen = (u_int32_t)htonl(jumboplen);
923
924 #if 1
925 /*
926 * if there are multiple jumbo payload options,
927 * *plenp will be non-zero and the packet will be
928 * rejected.
929 * the behavior may need some debate in ipngwg -
930 * multiple options does not make sense, however,
931 * there's no explicit mention in specification.
932 */
933 if (*plenp != 0) {
934 IP6_STATINC(IP6_STAT_BADOPTIONS);
935 icmp6_error(m, ICMP6_PARAM_PROB,
936 ICMP6_PARAMPROB_HEADER,
937 erroff + opt + 2 - opthead);
938 return (-1);
939 }
940 #endif
941
942 /*
943 * jumbo payload length must be larger than 65535.
944 */
945 if (jumboplen <= IPV6_MAXPACKET) {
946 IP6_STATINC(IP6_STAT_BADOPTIONS);
947 icmp6_error(m, ICMP6_PARAM_PROB,
948 ICMP6_PARAMPROB_HEADER,
949 erroff + opt + 2 - opthead);
950 return (-1);
951 }
952 *plenp = jumboplen;
953
954 break;
955 default: /* unknown option */
956 if (hbhlen < IP6OPT_MINLEN) {
957 IP6_STATINC(IP6_STAT_TOOSMALL);
958 goto bad;
959 }
960 optlen = ip6_unknown_opt(opt, m,
961 erroff + opt - opthead);
962 if (optlen == -1)
963 return (-1);
964 optlen += 2;
965 break;
966 }
967 }
968
969 return (0);
970
971 bad:
972 m_freem(m);
973 return (-1);
974 }
975
976 /*
977 * Unknown option processing.
978 * The third argument `off' is the offset from the IPv6 header to the option,
979 * which is necessary if the IPv6 header the and option header and IPv6 header
980 * is not continuous in order to return an ICMPv6 error.
981 */
982 int
ip6_unknown_opt(u_int8_t * optp,struct mbuf * m,int off)983 ip6_unknown_opt(u_int8_t *optp, struct mbuf *m, int off)
984 {
985 struct ip6_hdr *ip6;
986
987 switch (IP6OPT_TYPE(*optp)) {
988 case IP6OPT_TYPE_SKIP: /* ignore the option */
989 return ((int)*(optp + 1));
990 case IP6OPT_TYPE_DISCARD: /* silently discard */
991 m_freem(m);
992 return (-1);
993 case IP6OPT_TYPE_FORCEICMP: /* send ICMP even if multicasted */
994 IP6_STATINC(IP6_STAT_BADOPTIONS);
995 icmp6_error(m, ICMP6_PARAM_PROB, ICMP6_PARAMPROB_OPTION, off);
996 return (-1);
997 case IP6OPT_TYPE_ICMP: /* send ICMP if not multicasted */
998 IP6_STATINC(IP6_STAT_BADOPTIONS);
999 ip6 = mtod(m, struct ip6_hdr *);
1000 if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) ||
1001 (m->m_flags & (M_BCAST|M_MCAST)))
1002 m_freem(m);
1003 else
1004 icmp6_error(m, ICMP6_PARAM_PROB,
1005 ICMP6_PARAMPROB_OPTION, off);
1006 return (-1);
1007 }
1008
1009 m_freem(m); /* XXX: NOTREACHED */
1010 return (-1);
1011 }
1012
1013 /*
1014 * Create the "control" list for this pcb.
1015 *
1016 * The routine will be called from upper layer handlers like tcp6_input().
1017 * Thus the routine assumes that the caller (tcp6_input) have already
1018 * called IP6_EXTHDR_CHECK() and all the extension headers are located in the
1019 * very first mbuf on the mbuf chain.
1020 * We may want to add some infinite loop prevention or sanity checks for safety.
1021 * (This applies only when you are using KAME mbuf chain restriction, i.e.
1022 * you are using IP6_EXTHDR_CHECK() not m_pulldown())
1023 */
1024 void
ip6_savecontrol(struct in6pcb * in6p,struct mbuf ** mp,struct ip6_hdr * ip6,struct mbuf * m)1025 ip6_savecontrol(struct in6pcb *in6p, struct mbuf **mp,
1026 struct ip6_hdr *ip6, struct mbuf *m)
1027 {
1028 #ifdef RFC2292
1029 #define IS2292(x, y) ((in6p->in6p_flags & IN6P_RFC2292) ? (x) : (y))
1030 #else
1031 #define IS2292(x, y) (y)
1032 #endif
1033
1034 if (in6p->in6p_socket->so_options & SO_TIMESTAMP
1035 #ifdef SO_OTIMESTAMP
1036 || in6p->in6p_socket->so_options & SO_OTIMESTAMP
1037 #endif
1038 ) {
1039 struct timeval tv;
1040
1041 microtime(&tv);
1042 #ifdef SO_OTIMESTAMP
1043 if (in6p->in6p_socket->so_options & SO_OTIMESTAMP) {
1044 struct timeval50 tv50;
1045 timeval_to_timeval50(&tv, &tv50);
1046 *mp = sbcreatecontrol((void *) &tv50, sizeof(tv50),
1047 SCM_OTIMESTAMP, SOL_SOCKET);
1048 } else
1049 #endif
1050 *mp = sbcreatecontrol((void *) &tv, sizeof(tv),
1051 SCM_TIMESTAMP, SOL_SOCKET);
1052 if (*mp)
1053 mp = &(*mp)->m_next;
1054 }
1055
1056 /* some OSes call this logic with IPv4 packet, for SO_TIMESTAMP */
1057 if ((ip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION)
1058 return;
1059
1060 /* RFC 2292 sec. 5 */
1061 if ((in6p->in6p_flags & IN6P_PKTINFO) != 0) {
1062 struct in6_pktinfo pi6;
1063
1064 memcpy(&pi6.ipi6_addr, &ip6->ip6_dst, sizeof(struct in6_addr));
1065 in6_clearscope(&pi6.ipi6_addr); /* XXX */
1066 pi6.ipi6_ifindex = m->m_pkthdr.rcvif_index;
1067 *mp = sbcreatecontrol((void *) &pi6,
1068 sizeof(struct in6_pktinfo),
1069 IS2292(IPV6_2292PKTINFO, IPV6_PKTINFO), IPPROTO_IPV6);
1070 if (*mp)
1071 mp = &(*mp)->m_next;
1072 }
1073
1074 if (in6p->in6p_flags & IN6P_HOPLIMIT) {
1075 int hlim = ip6->ip6_hlim & 0xff;
1076
1077 *mp = sbcreatecontrol((void *) &hlim, sizeof(int),
1078 IS2292(IPV6_2292HOPLIMIT, IPV6_HOPLIMIT), IPPROTO_IPV6);
1079 if (*mp)
1080 mp = &(*mp)->m_next;
1081 }
1082
1083 if ((in6p->in6p_flags & IN6P_TCLASS) != 0) {
1084 u_int32_t flowinfo;
1085 int tclass;
1086
1087 flowinfo = (u_int32_t)ntohl(ip6->ip6_flow & IPV6_FLOWINFO_MASK);
1088 flowinfo >>= 20;
1089
1090 tclass = flowinfo & 0xff;
1091 *mp = sbcreatecontrol((void *)&tclass, sizeof(tclass),
1092 IPV6_TCLASS, IPPROTO_IPV6);
1093
1094 if (*mp)
1095 mp = &(*mp)->m_next;
1096 }
1097
1098 /*
1099 * IPV6_HOPOPTS socket option. Recall that we required super-user
1100 * privilege for the option (see ip6_ctloutput), but it might be too
1101 * strict, since there might be some hop-by-hop options which can be
1102 * returned to normal user.
1103 * See also RFC3542 section 8 (or RFC2292 section 6).
1104 */
1105 if ((in6p->in6p_flags & IN6P_HOPOPTS) != 0) {
1106 /*
1107 * Check if a hop-by-hop options header is contatined in the
1108 * received packet, and if so, store the options as ancillary
1109 * data. Note that a hop-by-hop options header must be
1110 * just after the IPv6 header, which fact is assured through
1111 * the IPv6 input processing.
1112 */
1113 struct ip6_hdr *xip6 = mtod(m, struct ip6_hdr *);
1114 if (xip6->ip6_nxt == IPPROTO_HOPOPTS) {
1115 struct ip6_hbh *hbh;
1116 int hbhlen;
1117 struct mbuf *ext;
1118
1119 ext = ip6_pullexthdr(m, sizeof(struct ip6_hdr),
1120 xip6->ip6_nxt);
1121 if (ext == NULL) {
1122 IP6_STATINC(IP6_STAT_TOOSHORT);
1123 return;
1124 }
1125 hbh = mtod(ext, struct ip6_hbh *);
1126 hbhlen = (hbh->ip6h_len + 1) << 3;
1127 if (hbhlen != ext->m_len) {
1128 m_freem(ext);
1129 IP6_STATINC(IP6_STAT_TOOSHORT);
1130 return;
1131 }
1132
1133 /*
1134 * XXX: We copy whole the header even if a jumbo
1135 * payload option is included, which option is to
1136 * be removed before returning in the RFC 2292.
1137 * Note: this constraint is removed in RFC3542.
1138 */
1139 *mp = sbcreatecontrol((void *)hbh, hbhlen,
1140 IS2292(IPV6_2292HOPOPTS, IPV6_HOPOPTS),
1141 IPPROTO_IPV6);
1142 if (*mp)
1143 mp = &(*mp)->m_next;
1144 m_freem(ext);
1145 }
1146 }
1147
1148 /* IPV6_DSTOPTS and IPV6_RTHDR socket options */
1149 if (in6p->in6p_flags & (IN6P_DSTOPTS | IN6P_RTHDR)) {
1150 struct ip6_hdr *xip6 = mtod(m, struct ip6_hdr *);
1151 int nxt = xip6->ip6_nxt, off = sizeof(struct ip6_hdr);
1152
1153 /*
1154 * Search for destination options headers or routing
1155 * header(s) through the header chain, and stores each
1156 * header as ancillary data.
1157 * Note that the order of the headers remains in
1158 * the chain of ancillary data.
1159 */
1160 for (;;) { /* is explicit loop prevention necessary? */
1161 struct ip6_ext *ip6e = NULL;
1162 int elen;
1163 struct mbuf *ext = NULL;
1164
1165 /*
1166 * if it is not an extension header, don't try to
1167 * pull it from the chain.
1168 */
1169 switch (nxt) {
1170 case IPPROTO_DSTOPTS:
1171 case IPPROTO_ROUTING:
1172 case IPPROTO_HOPOPTS:
1173 case IPPROTO_AH: /* is it possible? */
1174 break;
1175 default:
1176 goto loopend;
1177 }
1178
1179 ext = ip6_pullexthdr(m, off, nxt);
1180 if (ext == NULL) {
1181 IP6_STATINC(IP6_STAT_TOOSHORT);
1182 return;
1183 }
1184 ip6e = mtod(ext, struct ip6_ext *);
1185 if (nxt == IPPROTO_AH)
1186 elen = (ip6e->ip6e_len + 2) << 2;
1187 else
1188 elen = (ip6e->ip6e_len + 1) << 3;
1189 if (elen != ext->m_len) {
1190 m_freem(ext);
1191 IP6_STATINC(IP6_STAT_TOOSHORT);
1192 return;
1193 }
1194 KASSERT(IP6_HDR_ALIGNED_P(ip6e));
1195
1196 switch (nxt) {
1197 case IPPROTO_DSTOPTS:
1198 if (!(in6p->in6p_flags & IN6P_DSTOPTS))
1199 break;
1200
1201 *mp = sbcreatecontrol((void *)ip6e, elen,
1202 IS2292(IPV6_2292DSTOPTS, IPV6_DSTOPTS),
1203 IPPROTO_IPV6);
1204 if (*mp)
1205 mp = &(*mp)->m_next;
1206 break;
1207
1208 case IPPROTO_ROUTING:
1209 if (!(in6p->in6p_flags & IN6P_RTHDR))
1210 break;
1211
1212 *mp = sbcreatecontrol((void *)ip6e, elen,
1213 IS2292(IPV6_2292RTHDR, IPV6_RTHDR),
1214 IPPROTO_IPV6);
1215 if (*mp)
1216 mp = &(*mp)->m_next;
1217 break;
1218
1219 case IPPROTO_HOPOPTS:
1220 case IPPROTO_AH: /* is it possible? */
1221 break;
1222
1223 default:
1224 /*
1225 * other cases have been filtered in the above.
1226 * none will visit this case. here we supply
1227 * the code just in case (nxt overwritten or
1228 * other cases).
1229 */
1230 m_freem(ext);
1231 goto loopend;
1232
1233 }
1234
1235 /* proceed with the next header. */
1236 off += elen;
1237 nxt = ip6e->ip6e_nxt;
1238 ip6e = NULL;
1239 m_freem(ext);
1240 ext = NULL;
1241 }
1242 loopend:
1243 ;
1244 }
1245 }
1246 #undef IS2292
1247
1248
1249 void
ip6_notify_pmtu(struct in6pcb * in6p,const struct sockaddr_in6 * dst,uint32_t * mtu)1250 ip6_notify_pmtu(struct in6pcb *in6p, const struct sockaddr_in6 *dst,
1251 uint32_t *mtu)
1252 {
1253 struct socket *so;
1254 struct mbuf *m_mtu;
1255 struct ip6_mtuinfo mtuctl;
1256
1257 so = in6p->in6p_socket;
1258
1259 if (mtu == NULL)
1260 return;
1261
1262 #ifdef DIAGNOSTIC
1263 if (so == NULL) /* I believe this is impossible */
1264 panic("ip6_notify_pmtu: socket is NULL");
1265 #endif
1266
1267 memset(&mtuctl, 0, sizeof(mtuctl)); /* zero-clear for safety */
1268 mtuctl.ip6m_mtu = *mtu;
1269 mtuctl.ip6m_addr = *dst;
1270 if (sa6_recoverscope(&mtuctl.ip6m_addr))
1271 return;
1272
1273 if ((m_mtu = sbcreatecontrol((void *)&mtuctl, sizeof(mtuctl),
1274 IPV6_PATHMTU, IPPROTO_IPV6)) == NULL)
1275 return;
1276
1277 if (sbappendaddr(&so->so_rcv, (const struct sockaddr *)dst, NULL, m_mtu)
1278 == 0) {
1279 m_freem(m_mtu);
1280 /* XXX: should count statistics */
1281 } else
1282 sorwakeup(so);
1283
1284 return;
1285 }
1286
1287 /*
1288 * pull single extension header from mbuf chain. returns single mbuf that
1289 * contains the result, or NULL on error.
1290 */
1291 static struct mbuf *
ip6_pullexthdr(struct mbuf * m,size_t off,int nxt)1292 ip6_pullexthdr(struct mbuf *m, size_t off, int nxt)
1293 {
1294 struct ip6_ext ip6e;
1295 size_t elen;
1296 struct mbuf *n;
1297
1298 #ifdef DIAGNOSTIC
1299 switch (nxt) {
1300 case IPPROTO_DSTOPTS:
1301 case IPPROTO_ROUTING:
1302 case IPPROTO_HOPOPTS:
1303 case IPPROTO_AH: /* is it possible? */
1304 break;
1305 default:
1306 printf("ip6_pullexthdr: invalid nxt=%d\n", nxt);
1307 }
1308 #endif
1309
1310 m_copydata(m, off, sizeof(ip6e), (void *)&ip6e);
1311 if (nxt == IPPROTO_AH)
1312 elen = (ip6e.ip6e_len + 2) << 2;
1313 else
1314 elen = (ip6e.ip6e_len + 1) << 3;
1315
1316 MGET(n, M_DONTWAIT, MT_DATA);
1317 if (n && elen >= MLEN) {
1318 MCLGET(n, M_DONTWAIT);
1319 if ((n->m_flags & M_EXT) == 0) {
1320 m_free(n);
1321 n = NULL;
1322 }
1323 }
1324 if (!n)
1325 return NULL;
1326
1327 n->m_len = 0;
1328 if (elen >= M_TRAILINGSPACE(n)) {
1329 m_free(n);
1330 return NULL;
1331 }
1332
1333 m_copydata(m, off, elen, mtod(n, void *));
1334 n->m_len = elen;
1335 return n;
1336 }
1337
1338 /*
1339 * Get pointer to the previous header followed by the header
1340 * currently processed.
1341 * XXX: This function supposes that
1342 * M includes all headers,
1343 * the next header field and the header length field of each header
1344 * are valid, and
1345 * the sum of each header length equals to OFF.
1346 * Because of these assumptions, this function must be called very
1347 * carefully. Moreover, it will not be used in the near future when
1348 * we develop `neater' mechanism to process extension headers.
1349 */
1350 u_int8_t *
ip6_get_prevhdr(struct mbuf * m,int off)1351 ip6_get_prevhdr(struct mbuf *m, int off)
1352 {
1353 struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *);
1354
1355 if (off == sizeof(struct ip6_hdr))
1356 return (&ip6->ip6_nxt);
1357 else {
1358 int len, nxt;
1359 struct ip6_ext *ip6e = NULL;
1360
1361 nxt = ip6->ip6_nxt;
1362 len = sizeof(struct ip6_hdr);
1363 while (len < off) {
1364 ip6e = (struct ip6_ext *)(mtod(m, char *) + len);
1365
1366 switch (nxt) {
1367 case IPPROTO_FRAGMENT:
1368 len += sizeof(struct ip6_frag);
1369 break;
1370 case IPPROTO_AH:
1371 len += (ip6e->ip6e_len + 2) << 2;
1372 break;
1373 default:
1374 len += (ip6e->ip6e_len + 1) << 3;
1375 break;
1376 }
1377 nxt = ip6e->ip6e_nxt;
1378 }
1379 if (ip6e)
1380 return (&ip6e->ip6e_nxt);
1381 else
1382 return NULL;
1383 }
1384 }
1385
1386 /*
1387 * get next header offset. m will be retained.
1388 */
1389 int
ip6_nexthdr(struct mbuf * m,int off,int proto,int * nxtp)1390 ip6_nexthdr(struct mbuf *m, int off, int proto, int *nxtp)
1391 {
1392 struct ip6_hdr ip6;
1393 struct ip6_ext ip6e;
1394 struct ip6_frag fh;
1395
1396 /* just in case */
1397 if (m == NULL)
1398 panic("ip6_nexthdr: m == NULL");
1399 if ((m->m_flags & M_PKTHDR) == 0 || m->m_pkthdr.len < off)
1400 return -1;
1401
1402 switch (proto) {
1403 case IPPROTO_IPV6:
1404 /* do not chase beyond intermediate IPv6 headers */
1405 if (off != 0)
1406 return -1;
1407 if (m->m_pkthdr.len < off + sizeof(ip6))
1408 return -1;
1409 m_copydata(m, off, sizeof(ip6), (void *)&ip6);
1410 if (nxtp)
1411 *nxtp = ip6.ip6_nxt;
1412 off += sizeof(ip6);
1413 return off;
1414
1415 case IPPROTO_FRAGMENT:
1416 /*
1417 * terminate parsing if it is not the first fragment,
1418 * it does not make sense to parse through it.
1419 */
1420 if (m->m_pkthdr.len < off + sizeof(fh))
1421 return -1;
1422 m_copydata(m, off, sizeof(fh), (void *)&fh);
1423 if ((fh.ip6f_offlg & IP6F_OFF_MASK) != 0)
1424 return -1;
1425 if (nxtp)
1426 *nxtp = fh.ip6f_nxt;
1427 off += sizeof(struct ip6_frag);
1428 return off;
1429
1430 case IPPROTO_AH:
1431 if (m->m_pkthdr.len < off + sizeof(ip6e))
1432 return -1;
1433 m_copydata(m, off, sizeof(ip6e), (void *)&ip6e);
1434 if (nxtp)
1435 *nxtp = ip6e.ip6e_nxt;
1436 off += (ip6e.ip6e_len + 2) << 2;
1437 if (m->m_pkthdr.len < off)
1438 return -1;
1439 return off;
1440
1441 case IPPROTO_HOPOPTS:
1442 case IPPROTO_ROUTING:
1443 case IPPROTO_DSTOPTS:
1444 if (m->m_pkthdr.len < off + sizeof(ip6e))
1445 return -1;
1446 m_copydata(m, off, sizeof(ip6e), (void *)&ip6e);
1447 if (nxtp)
1448 *nxtp = ip6e.ip6e_nxt;
1449 off += (ip6e.ip6e_len + 1) << 3;
1450 if (m->m_pkthdr.len < off)
1451 return -1;
1452 return off;
1453
1454 case IPPROTO_NONE:
1455 case IPPROTO_ESP:
1456 case IPPROTO_IPCOMP:
1457 /* give up */
1458 return -1;
1459
1460 default:
1461 return -1;
1462 }
1463 }
1464
1465 /*
1466 * get offset for the last header in the chain. m will be kept untainted.
1467 */
1468 int
ip6_lasthdr(struct mbuf * m,int off,int proto,int * nxtp)1469 ip6_lasthdr(struct mbuf *m, int off, int proto, int *nxtp)
1470 {
1471 int newoff;
1472 int nxt;
1473
1474 if (!nxtp) {
1475 nxt = -1;
1476 nxtp = &nxt;
1477 }
1478 for (;;) {
1479 newoff = ip6_nexthdr(m, off, proto, nxtp);
1480 if (newoff < 0)
1481 return off;
1482 else if (newoff < off)
1483 return -1; /* invalid */
1484 else if (newoff == off)
1485 return newoff;
1486
1487 off = newoff;
1488 proto = *nxtp;
1489 }
1490 }
1491
1492 struct m_tag *
ip6_addaux(struct mbuf * m)1493 ip6_addaux(struct mbuf *m)
1494 {
1495 struct m_tag *mtag;
1496
1497 mtag = m_tag_find(m, PACKET_TAG_INET6, NULL);
1498 if (!mtag) {
1499 mtag = m_tag_get(PACKET_TAG_INET6, sizeof(struct ip6aux),
1500 M_NOWAIT);
1501 if (mtag) {
1502 m_tag_prepend(m, mtag);
1503 memset(mtag + 1, 0, sizeof(struct ip6aux));
1504 }
1505 }
1506 return mtag;
1507 }
1508
1509 struct m_tag *
ip6_findaux(struct mbuf * m)1510 ip6_findaux(struct mbuf *m)
1511 {
1512 struct m_tag *mtag;
1513
1514 mtag = m_tag_find(m, PACKET_TAG_INET6, NULL);
1515 return mtag;
1516 }
1517
1518 void
ip6_delaux(struct mbuf * m)1519 ip6_delaux(struct mbuf *m)
1520 {
1521 struct m_tag *mtag;
1522
1523 mtag = m_tag_find(m, PACKET_TAG_INET6, NULL);
1524 if (mtag)
1525 m_tag_delete(m, mtag);
1526 }
1527
1528 #ifdef GATEWAY
1529 /*
1530 * sysctl helper routine for net.inet.ip6.maxflows. Since
1531 * we could reduce this value, call ip6flow_reap();
1532 */
1533 static int
sysctl_net_inet6_ip6_maxflows(SYSCTLFN_ARGS)1534 sysctl_net_inet6_ip6_maxflows(SYSCTLFN_ARGS)
1535 {
1536 int error;
1537
1538 error = sysctl_lookup(SYSCTLFN_CALL(rnode));
1539 if (error || newp == NULL)
1540 return (error);
1541
1542 mutex_enter(softnet_lock);
1543 KERNEL_LOCK(1, NULL);
1544
1545 ip6flow_reap(0);
1546
1547 KERNEL_UNLOCK_ONE(NULL);
1548 mutex_exit(softnet_lock);
1549
1550 return (0);
1551 }
1552
1553 static int
sysctl_net_inet6_ip6_hashsize(SYSCTLFN_ARGS)1554 sysctl_net_inet6_ip6_hashsize(SYSCTLFN_ARGS)
1555 {
1556 int error, tmp;
1557 struct sysctlnode node;
1558
1559 node = *rnode;
1560 tmp = ip6_hashsize;
1561 node.sysctl_data = &tmp;
1562 error = sysctl_lookup(SYSCTLFN_CALL(&node));
1563 if (error || newp == NULL)
1564 return (error);
1565
1566 if ((tmp & (tmp - 1)) == 0 && tmp != 0) {
1567 /*
1568 * Can only fail due to malloc()
1569 */
1570 mutex_enter(softnet_lock);
1571 KERNEL_LOCK(1, NULL);
1572
1573 error = ip6flow_invalidate_all(tmp);
1574
1575 KERNEL_UNLOCK_ONE(NULL);
1576 mutex_exit(softnet_lock);
1577 } else {
1578 /*
1579 * EINVAL if not a power of 2
1580 */
1581 error = EINVAL;
1582 }
1583
1584 return error;
1585 }
1586 #endif /* GATEWAY */
1587
1588 /*
1589 * System control for IP6
1590 */
1591
1592 const u_char inet6ctlerrmap[PRC_NCMDS] = {
1593 0, 0, 0, 0,
1594 0, EMSGSIZE, EHOSTDOWN, EHOSTUNREACH,
1595 EHOSTUNREACH, EHOSTUNREACH, ECONNREFUSED, ECONNREFUSED,
1596 EMSGSIZE, EHOSTUNREACH, 0, 0,
1597 0, 0, 0, 0,
1598 ENOPROTOOPT
1599 };
1600
1601 extern int sysctl_net_inet6_addrctlpolicy(SYSCTLFN_ARGS);
1602
1603 static int
sysctl_net_inet6_ip6_stats(SYSCTLFN_ARGS)1604 sysctl_net_inet6_ip6_stats(SYSCTLFN_ARGS)
1605 {
1606
1607 return (NETSTAT_SYSCTL(ip6stat_percpu, IP6_NSTATS));
1608 }
1609
1610 static void
sysctl_net_inet6_ip6_setup(struct sysctllog ** clog)1611 sysctl_net_inet6_ip6_setup(struct sysctllog **clog)
1612 {
1613 #ifdef RFC2292
1614 #define IS2292(x, y) ((in6p->in6p_flags & IN6P_RFC2292) ? (x) : (y))
1615 #else
1616 #define IS2292(x, y) (y)
1617 #endif
1618
1619 sysctl_createv(clog, 0, NULL, NULL,
1620 CTLFLAG_PERMANENT,
1621 CTLTYPE_NODE, "inet6",
1622 SYSCTL_DESCR("PF_INET6 related settings"),
1623 NULL, 0, NULL, 0,
1624 CTL_NET, PF_INET6, CTL_EOL);
1625 sysctl_createv(clog, 0, NULL, NULL,
1626 CTLFLAG_PERMANENT,
1627 CTLTYPE_NODE, "ip6",
1628 SYSCTL_DESCR("IPv6 related settings"),
1629 NULL, 0, NULL, 0,
1630 CTL_NET, PF_INET6, IPPROTO_IPV6, CTL_EOL);
1631
1632 sysctl_createv(clog, 0, NULL, NULL,
1633 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1634 CTLTYPE_INT, "forwarding",
1635 SYSCTL_DESCR("Enable forwarding of INET6 datagrams"),
1636 NULL, 0, &ip6_forwarding, 0,
1637 CTL_NET, PF_INET6, IPPROTO_IPV6,
1638 IPV6CTL_FORWARDING, CTL_EOL);
1639 sysctl_createv(clog, 0, NULL, NULL,
1640 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1641 CTLTYPE_INT, "redirect",
1642 SYSCTL_DESCR("Enable sending of ICMPv6 redirect messages"),
1643 NULL, 0, &ip6_sendredirects, 0,
1644 CTL_NET, PF_INET6, IPPROTO_IPV6,
1645 IPV6CTL_SENDREDIRECTS, CTL_EOL);
1646 sysctl_createv(clog, 0, NULL, NULL,
1647 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1648 CTLTYPE_INT, "hlim",
1649 SYSCTL_DESCR("Hop limit for an INET6 datagram"),
1650 NULL, 0, &ip6_defhlim, 0,
1651 CTL_NET, PF_INET6, IPPROTO_IPV6,
1652 IPV6CTL_DEFHLIM, CTL_EOL);
1653 #ifdef notyet
1654 sysctl_createv(clog, 0, NULL, NULL,
1655 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1656 CTLTYPE_INT, "mtu", NULL,
1657 NULL, 0, &, 0,
1658 CTL_NET, PF_INET6, IPPROTO_IPV6,
1659 IPV6CTL_DEFMTU, CTL_EOL);
1660 #endif
1661 #ifdef __no_idea__
1662 sysctl_createv(clog, 0, NULL, NULL,
1663 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1664 CTLTYPE_INT, "forwsrcrt", NULL,
1665 NULL, 0, &?, 0,
1666 CTL_NET, PF_INET6, IPPROTO_IPV6,
1667 IPV6CTL_FORWSRCRT, CTL_EOL);
1668 sysctl_createv(clog, 0, NULL, NULL,
1669 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1670 CTLTYPE_STRUCT, "mrtstats", NULL,
1671 NULL, 0, &?, sizeof(?),
1672 CTL_NET, PF_INET6, IPPROTO_IPV6,
1673 IPV6CTL_MRTSTATS, CTL_EOL);
1674 sysctl_createv(clog, 0, NULL, NULL,
1675 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1676 CTLTYPE_?, "mrtproto", NULL,
1677 NULL, 0, &?, sizeof(?),
1678 CTL_NET, PF_INET6, IPPROTO_IPV6,
1679 IPV6CTL_MRTPROTO, CTL_EOL);
1680 #endif
1681 sysctl_createv(clog, 0, NULL, NULL,
1682 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1683 CTLTYPE_INT, "maxfragpackets",
1684 SYSCTL_DESCR("Maximum number of fragments to buffer "
1685 "for reassembly"),
1686 NULL, 0, &ip6_maxfragpackets, 0,
1687 CTL_NET, PF_INET6, IPPROTO_IPV6,
1688 IPV6CTL_MAXFRAGPACKETS, CTL_EOL);
1689 #ifdef __no_idea__
1690 sysctl_createv(clog, 0, NULL, NULL,
1691 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1692 CTLTYPE_INT, "sourcecheck", NULL,
1693 NULL, 0, &?, 0,
1694 CTL_NET, PF_INET6, IPPROTO_IPV6,
1695 IPV6CTL_SOURCECHECK, CTL_EOL);
1696 sysctl_createv(clog, 0, NULL, NULL,
1697 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1698 CTLTYPE_INT, "sourcecheck_logint", NULL,
1699 NULL, 0, &?, 0,
1700 CTL_NET, PF_INET6, IPPROTO_IPV6,
1701 IPV6CTL_SOURCECHECK_LOGINT, CTL_EOL);
1702 #endif
1703 sysctl_createv(clog, 0, NULL, NULL,
1704 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1705 CTLTYPE_INT, "accept_rtadv",
1706 SYSCTL_DESCR("Accept router advertisements"),
1707 NULL, 0, &ip6_accept_rtadv, 0,
1708 CTL_NET, PF_INET6, IPPROTO_IPV6,
1709 IPV6CTL_ACCEPT_RTADV, CTL_EOL);
1710 sysctl_createv(clog, 0, NULL, NULL,
1711 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1712 CTLTYPE_INT, "rtadv_maxroutes",
1713 SYSCTL_DESCR("Maximum number of routes accepted via router advertisements"),
1714 NULL, 0, &ip6_rtadv_maxroutes, 0,
1715 CTL_NET, PF_INET6, IPPROTO_IPV6,
1716 IPV6CTL_RTADV_MAXROUTES, CTL_EOL);
1717 sysctl_createv(clog, 0, NULL, NULL,
1718 CTLFLAG_PERMANENT,
1719 CTLTYPE_INT, "rtadv_numroutes",
1720 SYSCTL_DESCR("Current number of routes accepted via router advertisements"),
1721 NULL, 0, &nd6_numroutes, 0,
1722 CTL_NET, PF_INET6, IPPROTO_IPV6,
1723 IPV6CTL_RTADV_NUMROUTES, CTL_EOL);
1724 sysctl_createv(clog, 0, NULL, NULL,
1725 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1726 CTLTYPE_INT, "keepfaith",
1727 SYSCTL_DESCR("Activate faith interface"),
1728 NULL, 0, &ip6_keepfaith, 0,
1729 CTL_NET, PF_INET6, IPPROTO_IPV6,
1730 IPV6CTL_KEEPFAITH, CTL_EOL);
1731 sysctl_createv(clog, 0, NULL, NULL,
1732 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1733 CTLTYPE_INT, "log_interval",
1734 SYSCTL_DESCR("Minumum interval between logging "
1735 "unroutable packets"),
1736 NULL, 0, &ip6_log_interval, 0,
1737 CTL_NET, PF_INET6, IPPROTO_IPV6,
1738 IPV6CTL_LOG_INTERVAL, CTL_EOL);
1739 sysctl_createv(clog, 0, NULL, NULL,
1740 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1741 CTLTYPE_INT, "hdrnestlimit",
1742 SYSCTL_DESCR("Maximum number of nested IPv6 headers"),
1743 NULL, 0, &ip6_hdrnestlimit, 0,
1744 CTL_NET, PF_INET6, IPPROTO_IPV6,
1745 IPV6CTL_HDRNESTLIMIT, CTL_EOL);
1746 sysctl_createv(clog, 0, NULL, NULL,
1747 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1748 CTLTYPE_INT, "dad_count",
1749 SYSCTL_DESCR("Number of Duplicate Address Detection "
1750 "probes to send"),
1751 NULL, 0, &ip6_dad_count, 0,
1752 CTL_NET, PF_INET6, IPPROTO_IPV6,
1753 IPV6CTL_DAD_COUNT, CTL_EOL);
1754 sysctl_createv(clog, 0, NULL, NULL,
1755 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1756 CTLTYPE_INT, "auto_flowlabel",
1757 SYSCTL_DESCR("Assign random IPv6 flow labels"),
1758 NULL, 0, &ip6_auto_flowlabel, 0,
1759 CTL_NET, PF_INET6, IPPROTO_IPV6,
1760 IPV6CTL_AUTO_FLOWLABEL, CTL_EOL);
1761 sysctl_createv(clog, 0, NULL, NULL,
1762 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1763 CTLTYPE_INT, "defmcasthlim",
1764 SYSCTL_DESCR("Default multicast hop limit"),
1765 NULL, 0, &ip6_defmcasthlim, 0,
1766 CTL_NET, PF_INET6, IPPROTO_IPV6,
1767 IPV6CTL_DEFMCASTHLIM, CTL_EOL);
1768 sysctl_createv(clog, 0, NULL, NULL,
1769 CTLFLAG_PERMANENT,
1770 CTLTYPE_STRING, "kame_version",
1771 SYSCTL_DESCR("KAME Version"),
1772 NULL, 0, __UNCONST(__KAME_VERSION), 0,
1773 CTL_NET, PF_INET6, IPPROTO_IPV6,
1774 IPV6CTL_KAME_VERSION, CTL_EOL);
1775 sysctl_createv(clog, 0, NULL, NULL,
1776 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1777 CTLTYPE_INT, "use_deprecated",
1778 SYSCTL_DESCR("Allow use of deprecated addresses as "
1779 "source addresses"),
1780 NULL, 0, &ip6_use_deprecated, 0,
1781 CTL_NET, PF_INET6, IPPROTO_IPV6,
1782 IPV6CTL_USE_DEPRECATED, CTL_EOL);
1783 sysctl_createv(clog, 0, NULL, NULL,
1784 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1785 CTLTYPE_INT, "rr_prune", NULL,
1786 NULL, 0, &ip6_rr_prune, 0,
1787 CTL_NET, PF_INET6, IPPROTO_IPV6,
1788 IPV6CTL_RR_PRUNE, CTL_EOL);
1789 sysctl_createv(clog, 0, NULL, NULL,
1790 CTLFLAG_PERMANENT
1791 #ifndef INET6_BINDV6ONLY
1792 |CTLFLAG_READWRITE,
1793 #endif
1794 CTLTYPE_INT, "v6only",
1795 SYSCTL_DESCR("Disallow PF_INET6 sockets from connecting "
1796 "to PF_INET sockets"),
1797 NULL, 0, &ip6_v6only, 0,
1798 CTL_NET, PF_INET6, IPPROTO_IPV6,
1799 IPV6CTL_V6ONLY, CTL_EOL);
1800 sysctl_createv(clog, 0, NULL, NULL,
1801 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1802 CTLTYPE_INT, "anonportmin",
1803 SYSCTL_DESCR("Lowest ephemeral port number to assign"),
1804 sysctl_net_inet_ip_ports, 0, &ip6_anonportmin, 0,
1805 CTL_NET, PF_INET6, IPPROTO_IPV6,
1806 IPV6CTL_ANONPORTMIN, CTL_EOL);
1807 sysctl_createv(clog, 0, NULL, NULL,
1808 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1809 CTLTYPE_INT, "anonportmax",
1810 SYSCTL_DESCR("Highest ephemeral port number to assign"),
1811 sysctl_net_inet_ip_ports, 0, &ip6_anonportmax, 0,
1812 CTL_NET, PF_INET6, IPPROTO_IPV6,
1813 IPV6CTL_ANONPORTMAX, CTL_EOL);
1814 #ifndef IPNOPRIVPORTS
1815 sysctl_createv(clog, 0, NULL, NULL,
1816 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1817 CTLTYPE_INT, "lowportmin",
1818 SYSCTL_DESCR("Lowest privileged ephemeral port number "
1819 "to assign"),
1820 sysctl_net_inet_ip_ports, 0, &ip6_lowportmin, 0,
1821 CTL_NET, PF_INET6, IPPROTO_IPV6,
1822 IPV6CTL_LOWPORTMIN, CTL_EOL);
1823 sysctl_createv(clog, 0, NULL, NULL,
1824 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1825 CTLTYPE_INT, "lowportmax",
1826 SYSCTL_DESCR("Highest privileged ephemeral port number "
1827 "to assign"),
1828 sysctl_net_inet_ip_ports, 0, &ip6_lowportmax, 0,
1829 CTL_NET, PF_INET6, IPPROTO_IPV6,
1830 IPV6CTL_LOWPORTMAX, CTL_EOL);
1831 #endif /* IPNOPRIVPORTS */
1832 sysctl_createv(clog, 0, NULL, NULL,
1833 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1834 CTLTYPE_INT, "auto_linklocal",
1835 SYSCTL_DESCR("Default value of per-interface flag for "
1836 "adding an IPv6 link-local address to "
1837 "interfaces when attached"),
1838 NULL, 0, &ip6_auto_linklocal, 0,
1839 CTL_NET, PF_INET6, IPPROTO_IPV6,
1840 IPV6CTL_AUTO_LINKLOCAL, CTL_EOL);
1841 sysctl_createv(clog, 0, NULL, NULL,
1842 CTLFLAG_PERMANENT|CTLFLAG_READONLY,
1843 CTLTYPE_STRUCT, "addctlpolicy",
1844 SYSCTL_DESCR("Return the current address control"
1845 " policy"),
1846 sysctl_net_inet6_addrctlpolicy, 0, NULL, 0,
1847 CTL_NET, PF_INET6, IPPROTO_IPV6,
1848 IPV6CTL_ADDRCTLPOLICY, CTL_EOL);
1849 sysctl_createv(clog, 0, NULL, NULL,
1850 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1851 CTLTYPE_INT, "use_tempaddr",
1852 SYSCTL_DESCR("Use temporary address"),
1853 NULL, 0, &ip6_use_tempaddr, 0,
1854 CTL_NET, PF_INET6, IPPROTO_IPV6,
1855 CTL_CREATE, CTL_EOL);
1856 sysctl_createv(clog, 0, NULL, NULL,
1857 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1858 CTLTYPE_INT, "prefer_tempaddr",
1859 SYSCTL_DESCR("Prefer temporary address as source "
1860 "address"),
1861 NULL, 0, &ip6_prefer_tempaddr, 0,
1862 CTL_NET, PF_INET6, IPPROTO_IPV6,
1863 CTL_CREATE, CTL_EOL);
1864 sysctl_createv(clog, 0, NULL, NULL,
1865 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1866 CTLTYPE_INT, "temppltime",
1867 SYSCTL_DESCR("preferred lifetime of a temporary address"),
1868 NULL, 0, &ip6_temp_preferred_lifetime, 0,
1869 CTL_NET, PF_INET6, IPPROTO_IPV6,
1870 CTL_CREATE, CTL_EOL);
1871 sysctl_createv(clog, 0, NULL, NULL,
1872 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1873 CTLTYPE_INT, "tempvltime",
1874 SYSCTL_DESCR("valid lifetime of a temporary address"),
1875 NULL, 0, &ip6_temp_valid_lifetime, 0,
1876 CTL_NET, PF_INET6, IPPROTO_IPV6,
1877 CTL_CREATE, CTL_EOL);
1878 sysctl_createv(clog, 0, NULL, NULL,
1879 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1880 CTLTYPE_INT, "maxfrags",
1881 SYSCTL_DESCR("Maximum fragments in reassembly queue"),
1882 NULL, 0, &ip6_maxfrags, 0,
1883 CTL_NET, PF_INET6, IPPROTO_IPV6,
1884 IPV6CTL_MAXFRAGS, CTL_EOL);
1885 sysctl_createv(clog, 0, NULL, NULL,
1886 CTLFLAG_PERMANENT,
1887 CTLTYPE_STRUCT, "stats",
1888 SYSCTL_DESCR("IPv6 statistics"),
1889 sysctl_net_inet6_ip6_stats, 0, NULL, 0,
1890 CTL_NET, PF_INET6, IPPROTO_IPV6,
1891 IPV6CTL_STATS, CTL_EOL);
1892 sysctl_createv(clog, 0, NULL, NULL,
1893 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1894 CTLTYPE_INT, "use_defaultzone",
1895 SYSCTL_DESCR("Whether to use the default scope zones"),
1896 NULL, 0, &ip6_use_defzone, 0,
1897 CTL_NET, PF_INET6, IPPROTO_IPV6,
1898 IPV6CTL_USE_DEFAULTZONE, CTL_EOL);
1899 sysctl_createv(clog, 0, NULL, NULL,
1900 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1901 CTLTYPE_INT, "mcast_pmtu",
1902 SYSCTL_DESCR("Enable pMTU discovery for multicast packet"),
1903 NULL, 0, &ip6_mcast_pmtu, 0,
1904 CTL_NET, PF_INET6, IPPROTO_IPV6,
1905 CTL_CREATE, CTL_EOL);
1906 #ifdef GATEWAY
1907 sysctl_createv(clog, 0, NULL, NULL,
1908 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1909 CTLTYPE_INT, "maxflows",
1910 SYSCTL_DESCR("Number of flows for fast forwarding (IPv6)"),
1911 sysctl_net_inet6_ip6_maxflows, 0, &ip6_maxflows, 0,
1912 CTL_NET, PF_INET6, IPPROTO_IPV6,
1913 CTL_CREATE, CTL_EOL);
1914 sysctl_createv(clog, 0, NULL, NULL,
1915 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1916 CTLTYPE_INT, "hashsize",
1917 SYSCTL_DESCR("Size of hash table for fast forwarding (IPv6)"),
1918 sysctl_net_inet6_ip6_hashsize, 0, &ip6_hashsize, 0,
1919 CTL_NET, PF_INET6, IPPROTO_IPV6,
1920 CTL_CREATE, CTL_EOL);
1921 #endif
1922 /* anonportalgo RFC6056 subtree */
1923 const struct sysctlnode *portalgo_node;
1924 sysctl_createv(clog, 0, NULL, &portalgo_node,
1925 CTLFLAG_PERMANENT,
1926 CTLTYPE_NODE, "anonportalgo",
1927 SYSCTL_DESCR("Anonymous port algorithm selection (RFC 6056)"),
1928 NULL, 0, NULL, 0,
1929 CTL_NET, PF_INET6, IPPROTO_IPV6, CTL_CREATE, CTL_EOL);
1930 sysctl_createv(clog, 0, &portalgo_node, NULL,
1931 CTLFLAG_PERMANENT,
1932 CTLTYPE_STRING, "available",
1933 SYSCTL_DESCR("available algorithms"),
1934 sysctl_portalgo_available, 0, NULL, PORTALGO_MAXLEN,
1935 CTL_CREATE, CTL_EOL);
1936 sysctl_createv(clog, 0, &portalgo_node, NULL,
1937 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1938 CTLTYPE_STRING, "selected",
1939 SYSCTL_DESCR("selected algorithm"),
1940 sysctl_portalgo_selected6, 0, NULL, PORTALGO_MAXLEN,
1941 CTL_CREATE, CTL_EOL);
1942 sysctl_createv(clog, 0, &portalgo_node, NULL,
1943 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1944 CTLTYPE_STRUCT, "reserve",
1945 SYSCTL_DESCR("bitmap of reserved ports"),
1946 sysctl_portalgo_reserve6, 0, NULL, 0,
1947 CTL_CREATE, CTL_EOL);
1948 sysctl_createv(clog, 0, NULL, NULL,
1949 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1950 CTLTYPE_INT, "neighborgcthresh",
1951 SYSCTL_DESCR("Maximum number of entries in neighbor"
1952 " cache"),
1953 NULL, 1, &ip6_neighborgcthresh, 0,
1954 CTL_NET, PF_INET6, IPPROTO_IPV6,
1955 CTL_CREATE, CTL_EOL);
1956 sysctl_createv(clog, 0, NULL, NULL,
1957 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1958 CTLTYPE_INT, "maxifprefixes",
1959 SYSCTL_DESCR("Maximum number of prefixes created by"
1960 " route advertisement per interface"),
1961 NULL, 1, &ip6_maxifprefixes, 0,
1962 CTL_NET, PF_INET6, IPPROTO_IPV6,
1963 CTL_CREATE, CTL_EOL);
1964 sysctl_createv(clog, 0, NULL, NULL,
1965 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1966 CTLTYPE_INT, "maxifdefrouters",
1967 SYSCTL_DESCR("Maximum number of default routers created"
1968 " by route advertisement per interface"),
1969 NULL, 1, &ip6_maxifdefrouters, 0,
1970 CTL_NET, PF_INET6, IPPROTO_IPV6,
1971 CTL_CREATE, CTL_EOL);
1972 sysctl_createv(clog, 0, NULL, NULL,
1973 CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1974 CTLTYPE_INT, "maxdynroutes",
1975 SYSCTL_DESCR("Maximum number of routes created via"
1976 " redirect"),
1977 NULL, 1, &ip6_maxdynroutes, 0,
1978 CTL_NET, PF_INET6, IPPROTO_IPV6,
1979 CTL_CREATE, CTL_EOL);
1980 }
1981
1982 void
ip6_statinc(u_int stat)1983 ip6_statinc(u_int stat)
1984 {
1985
1986 KASSERT(stat < IP6_NSTATS);
1987 IP6_STATINC(stat);
1988 }
1989