1 /* $FreeBSD: src/sys/netinet/ip_encap.c,v 1.1.2.5 2003/01/23 21:06:45 sam Exp $ */
2 /* $KAME: ip_encap.c,v 1.41 2001/03/15 08:35:08 itojun Exp $ */
5 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions
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.
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
33 * My grandfather said that there's a devil inside tunnelling technology...
35 * We have surprisingly many protocols that want packets with IP protocol
36 * #4 or #41. Here's a list of protocols that want protocol #41:
37 * RFC1933 configured tunnel
38 * RFC1933 automatic tunnel
39 * RFC2401 IPsec tunnel
40 * RFC2473 IPv6 generic packet tunnelling
41 * RFC2529 6over4 tunnel
42 * mobile-ip6 (uses RFC2473)
45 * Here's a list of protocol that want protocol #4:
46 * RFC1853 IPv4-in-IPv4 tunnelling
47 * RFC2003 IPv4 encapsulation within IPv4
48 * RFC2344 reverse tunnelling for mobile-ip4
49 * RFC2401 IPsec tunnel
50 * Well, what can I say. They impose different en/decapsulation mechanism
51 * from each other, so they need separate protocol handler. The only one
52 * we can easily determine by protocol # is IPsec, which always has
53 * AH/ESP/IPComp header right after outer IP header.
55 * So, clearly good old protosw does not work for protocol #4 and #41.
56 * The code will let you match protocol via src/dst address pair.
58 /* XXX is M_NETADDR correct? */
61 #include "opt_inet6.h"
63 #include <sys/param.h>
64 #include <sys/systm.h>
65 #include <sys/socket.h>
66 #include <sys/sockio.h>
68 #include <sys/errno.h>
69 #include <sys/protosw.h>
70 #include <sys/queue.h>
73 #include <net/route.h>
75 #include <netinet/in.h>
76 #include <netinet/in_systm.h>
77 #include <netinet/ip.h>
78 #include <netinet/ip_var.h>
79 #include <netinet/ip_encap.h>
80 #include <netinet/ipprotosw.h>
83 #include <netinet/ip6.h>
84 #include <netinet6/ip6_var.h>
85 #include <netinet6/ip6protosw.h>
88 #include <machine/stdarg.h>
90 #include <net/net_osdep.h>
92 #include <sys/kernel.h>
93 #include <sys/malloc.h>
94 MALLOC_DEFINE(M_NETADDR, "Export Host", "Export host address structure");
96 static void encap_add __P((struct encaptab *));
97 static int mask_match __P((const struct encaptab *, const struct sockaddr *,
98 const struct sockaddr *));
99 static void encap_fillarg __P((struct mbuf *, const struct encaptab *));
101 #ifndef LIST_HEAD_INITIALIZER
102 /* rely upon BSS initialization */
103 LIST_HEAD(, encaptab) encaptab;
105 LIST_HEAD(, encaptab) encaptab = LIST_HEAD_INITIALIZER(&encaptab);
108 void (*ipip_input)(struct mbuf *, int, int); /* hook for mrouting */
113 static int initialized = 0;
120 * we cannot use LIST_INIT() here, since drivers may want to call
121 * encap_attach(), on driver attach. encap_init() will be called
122 * on AF_INET{,6} initialization, which happens after driver
123 * initialization - using LIST_INIT() here can nuke encap_attach()
126 LIST_INIT(&encaptab);
132 encap4_input(struct mbuf *m, int off, int proto)
135 struct sockaddr_in s, d;
136 const struct ipprotosw *psw;
137 struct encaptab *ep, *match;
140 ip = mtod(m, struct ip *);
142 bzero(&s, sizeof(s));
143 s.sin_family = AF_INET;
144 s.sin_len = sizeof(struct sockaddr_in);
145 s.sin_addr = ip->ip_src;
146 bzero(&d, sizeof(d));
147 d.sin_family = AF_INET;
148 d.sin_len = sizeof(struct sockaddr_in);
149 d.sin_addr = ip->ip_dst;
153 for (ep = LIST_FIRST(&encaptab); ep; ep = LIST_NEXT(ep, chain)) {
154 if (ep->af != AF_INET)
156 if (ep->proto >= 0 && ep->proto != proto)
159 prio = (*ep->func)(m, off, proto, ep->arg);
162 * it's inbound traffic, we need to match in reverse
165 prio = mask_match(ep, (struct sockaddr *)&d,
166 (struct sockaddr *)&s);
170 * We prioritize the matches by using bit length of the
171 * matches. mask_match() and user-supplied matching function
172 * should return the bit length of the matches (for example,
173 * if both src/dst are matched for IPv4, 64 should be returned).
174 * 0 or negative return value means "it did not match".
176 * The question is, since we have two "mask" portion, we
177 * cannot really define total order between entries.
178 * For example, which of these should be preferred?
179 * mask_match() returns 48 (32 + 16) for both of them.
180 * src=3ffe::/16, dst=3ffe:501::/32
181 * src=3ffe:501::/32, dst=3ffe::/16
183 * We need to loop through all the possible candidates
184 * to get the best match - the search takes O(n) for
185 * n attachments (i.e. interfaces).
189 if (prio > matchprio) {
196 /* found a match, "match" has the best one */
197 psw = (const struct ipprotosw *)match->psw;
198 if (psw && psw->pr_input) {
199 encap_fillarg(m, match);
200 (*psw->pr_input)(m, off, proto);
206 /* for backward compatibility */
207 if (proto == IPPROTO_IPV4 && ipip_input) {
208 ipip_input(m, off, proto);
212 /* last resort: inject to raw socket */
213 rip_input(m, off, proto);
219 encap6_input(mp, offp, proto)
224 struct mbuf *m = *mp;
226 struct sockaddr_in6 s, d;
227 const struct ip6protosw *psw;
228 struct encaptab *ep, *match;
231 ip6 = mtod(m, struct ip6_hdr *);
233 bzero(&s, sizeof(s));
234 s.sin6_family = AF_INET6;
235 s.sin6_len = sizeof(struct sockaddr_in6);
236 s.sin6_addr = ip6->ip6_src;
237 bzero(&d, sizeof(d));
238 d.sin6_family = AF_INET6;
239 d.sin6_len = sizeof(struct sockaddr_in6);
240 d.sin6_addr = ip6->ip6_dst;
244 for (ep = LIST_FIRST(&encaptab); ep; ep = LIST_NEXT(ep, chain)) {
245 if (ep->af != AF_INET6)
247 if (ep->proto >= 0 && ep->proto != proto)
250 prio = (*ep->func)(m, *offp, proto, ep->arg);
253 * it's inbound traffic, we need to match in reverse
256 prio = mask_match(ep, (struct sockaddr *)&d,
257 (struct sockaddr *)&s);
260 /* see encap4_input() for issues here */
263 if (prio > matchprio) {
271 psw = (const struct ip6protosw *)match->psw;
272 if (psw && psw->pr_input) {
273 encap_fillarg(m, match);
274 return (*psw->pr_input)(mp, offp, proto);
281 /* last resort: inject to raw socket */
282 return rip6_input(mp, offp, proto);
291 LIST_INSERT_HEAD(&encaptab, ep, chain);
295 * sp (src ptr) is always my side, and dp (dst ptr) is always remote side.
296 * length of mask (sm and dm) is assumed to be same as sp/dp.
297 * Return value will be necessary as input (cookie) for encap_detach().
299 const struct encaptab *
300 encap_attach(af, proto, sp, sm, dp, dm, psw, arg)
303 const struct sockaddr *sp, *sm;
304 const struct sockaddr *dp, *dm;
305 const struct protosw *psw;
313 /* sanity check on args */
314 if (sp->sa_len > sizeof(ep->src) || dp->sa_len > sizeof(ep->dst)) {
318 if (sp->sa_len != dp->sa_len) {
322 if (af != sp->sa_family || af != dp->sa_family) {
327 /* check if anyone have already attached with exactly same config */
328 for (ep = LIST_FIRST(&encaptab); ep; ep = LIST_NEXT(ep, chain)) {
331 if (ep->proto != proto)
333 if (ep->src.ss_len != sp->sa_len ||
334 bcmp(&ep->src, sp, sp->sa_len) != 0 ||
335 bcmp(&ep->srcmask, sm, sp->sa_len) != 0)
337 if (ep->dst.ss_len != dp->sa_len ||
338 bcmp(&ep->dst, dp, dp->sa_len) != 0 ||
339 bcmp(&ep->dstmask, dm, dp->sa_len) != 0)
346 ep = malloc(sizeof(*ep), M_NETADDR, M_NOWAIT); /*XXX*/
351 bzero(ep, sizeof(*ep));
355 bcopy(sp, &ep->src, sp->sa_len);
356 bcopy(sm, &ep->srcmask, sp->sa_len);
357 bcopy(dp, &ep->dst, dp->sa_len);
358 bcopy(dm, &ep->dstmask, dp->sa_len);
373 const struct encaptab *
374 encap_attach_func(af, proto, func, psw, arg)
377 int (*func) __P((const struct mbuf *, int, int, void *));
378 const struct protosw *psw;
386 /* sanity check on args */
392 ep = malloc(sizeof(*ep), M_NETADDR, M_NOWAIT); /*XXX*/
397 bzero(ep, sizeof(*ep));
418 const struct encaptab *cookie;
420 const struct encaptab *ep = cookie;
423 for (p = LIST_FIRST(&encaptab); p; p = LIST_NEXT(p, chain)) {
425 LIST_REMOVE(p, chain);
426 free(p, M_NETADDR); /*XXX*/
435 mask_match(ep, sp, dp)
436 const struct encaptab *ep;
437 const struct sockaddr *sp;
438 const struct sockaddr *dp;
440 struct sockaddr_storage s;
441 struct sockaddr_storage d;
443 const u_int8_t *p, *q;
447 if (sp->sa_len > sizeof(s) || dp->sa_len > sizeof(d))
449 if (sp->sa_family != ep->af || dp->sa_family != ep->af)
451 if (sp->sa_len != ep->src.ss_len || dp->sa_len != ep->dst.ss_len)
456 p = (const u_int8_t *)sp;
457 q = (const u_int8_t *)&ep->srcmask;
459 for (i = 0 ; i < sp->sa_len; i++) {
462 matchlen += (q[i] ? 8 : 0);
465 p = (const u_int8_t *)dp;
466 q = (const u_int8_t *)&ep->dstmask;
468 for (i = 0 ; i < dp->sa_len; i++) {
470 /* XXX rough estimate */
471 matchlen += (q[i] ? 8 : 0);
474 /* need to overwrite len/family portion as we don't compare them */
475 s.ss_len = sp->sa_len;
476 s.ss_family = sp->sa_family;
477 d.ss_len = dp->sa_len;
478 d.ss_family = dp->sa_family;
480 if (bcmp(&s, &ep->src, ep->src.ss_len) == 0 &&
481 bcmp(&d, &ep->dst, ep->dst.ss_len) == 0) {
490 const struct encaptab *ep;
494 tag = m_tag_get(PACKET_TAG_ENCAP, sizeof (void*), M_NOWAIT);
496 *(void**)(tag+1) = ep->arg;
497 m_tag_prepend(m, tag);
508 tag = m_tag_find(m, PACKET_TAG_ENCAP, NULL);
510 p = *(void**)(tag+1);
511 m_tag_delete(m, tag);