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? */
60 #include "opt_mrouting.h"
62 #include "opt_inet6.h"
64 #include <sys/param.h>
65 #include <sys/systm.h>
66 #include <sys/socket.h>
67 #include <sys/sockio.h>
69 #include <sys/errno.h>
70 #include <sys/protosw.h>
71 #include <sys/queue.h>
74 #include <net/route.h>
76 #include <netinet/in.h>
77 #include <netinet/in_systm.h>
78 #include <netinet/ip.h>
79 #include <netinet/ip_var.h>
80 #include <netinet/ip_encap.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 static MALLOC_DEFINE(M_NETADDR, "Export Host", "Export host address structure");
96 static void encap_add(struct encaptab *);
97 static int mask_match(const struct encaptab *, const struct sockaddr *,
98 const struct sockaddr *);
99 static void encap_fillarg(struct mbuf *, const struct encaptab *);
102 * All global variables in ip_encap.c are locked using encapmtx.
104 static struct mtx encapmtx;
105 MTX_SYSINIT(encapmtx, &encapmtx, "encapmtx", MTX_DEF);
106 LIST_HEAD(, encaptab) encaptab = LIST_HEAD_INITIALIZER(&encaptab);
109 * We currently keey encap_init() for source code compatibility reasons --
110 * it's referenced by KAME pieces in netinet6.
125 struct sockaddr_in s, d;
126 const struct protosw *psw;
127 struct encaptab *ep, *match;
130 ip = mtod(m, struct ip *);
133 bzero(&s, sizeof(s));
134 s.sin_family = AF_INET;
135 s.sin_len = sizeof(struct sockaddr_in);
136 s.sin_addr = ip->ip_src;
137 bzero(&d, sizeof(d));
138 d.sin_family = AF_INET;
139 d.sin_len = sizeof(struct sockaddr_in);
140 d.sin_addr = ip->ip_dst;
145 LIST_FOREACH(ep, &encaptab, chain) {
146 if (ep->af != AF_INET)
148 if (ep->proto >= 0 && ep->proto != proto)
151 prio = (*ep->func)(m, off, proto, ep->arg);
154 * it's inbound traffic, we need to match in reverse
157 prio = mask_match(ep, (struct sockaddr *)&d,
158 (struct sockaddr *)&s);
162 * We prioritize the matches by using bit length of the
163 * matches. mask_match() and user-supplied matching function
164 * should return the bit length of the matches (for example,
165 * if both src/dst are matched for IPv4, 64 should be returned).
166 * 0 or negative return value means "it did not match".
168 * The question is, since we have two "mask" portion, we
169 * cannot really define total order between entries.
170 * For example, which of these should be preferred?
171 * mask_match() returns 48 (32 + 16) for both of them.
172 * src=3ffe::/16, dst=3ffe:501::/32
173 * src=3ffe:501::/32, dst=3ffe::/16
175 * We need to loop through all the possible candidates
176 * to get the best match - the search takes O(n) for
177 * n attachments (i.e. interfaces).
181 if (prio > matchprio) {
186 mtx_unlock(&encapmtx);
189 /* found a match, "match" has the best one */
191 if (psw && psw->pr_input) {
192 encap_fillarg(m, match);
193 (*psw->pr_input)(m, off);
199 /* last resort: inject to raw socket */
206 encap6_input(mp, offp, proto)
211 struct mbuf *m = *mp;
213 struct sockaddr_in6 s, d;
214 const struct ip6protosw *psw;
215 struct encaptab *ep, *match;
218 ip6 = mtod(m, struct ip6_hdr *);
220 bzero(&s, sizeof(s));
221 s.sin6_family = AF_INET6;
222 s.sin6_len = sizeof(struct sockaddr_in6);
223 s.sin6_addr = ip6->ip6_src;
224 bzero(&d, sizeof(d));
225 d.sin6_family = AF_INET6;
226 d.sin6_len = sizeof(struct sockaddr_in6);
227 d.sin6_addr = ip6->ip6_dst;
232 LIST_FOREACH(ep, &encaptab, chain) {
233 if (ep->af != AF_INET6)
235 if (ep->proto >= 0 && ep->proto != proto)
238 prio = (*ep->func)(m, *offp, proto, ep->arg);
241 * it's inbound traffic, we need to match in reverse
244 prio = mask_match(ep, (struct sockaddr *)&d,
245 (struct sockaddr *)&s);
248 /* see encap4_input() for issues here */
251 if (prio > matchprio) {
256 mtx_unlock(&encapmtx);
260 psw = (const struct ip6protosw *)match->psw;
261 if (psw && psw->pr_input) {
262 encap_fillarg(m, match);
263 return (*psw->pr_input)(mp, offp, proto);
270 /* last resort: inject to raw socket */
271 return rip6_input(mp, offp, proto);
275 /*lint -sem(encap_add, custodial(1)) */
281 mtx_assert(&encapmtx, MA_OWNED);
282 LIST_INSERT_HEAD(&encaptab, ep, chain);
286 * sp (src ptr) is always my side, and dp (dst ptr) is always remote side.
287 * length of mask (sm and dm) is assumed to be same as sp/dp.
288 * Return value will be necessary as input (cookie) for encap_detach().
290 const struct encaptab *
291 encap_attach(af, proto, sp, sm, dp, dm, psw, arg)
294 const struct sockaddr *sp, *sm;
295 const struct sockaddr *dp, *dm;
296 const struct protosw *psw;
301 /* sanity check on args */
302 if (sp->sa_len > sizeof(ep->src) || dp->sa_len > sizeof(ep->dst))
304 if (sp->sa_len != dp->sa_len)
306 if (af != sp->sa_family || af != dp->sa_family)
309 /* check if anyone have already attached with exactly same config */
311 LIST_FOREACH(ep, &encaptab, chain) {
314 if (ep->proto != proto)
316 if (ep->src.ss_len != sp->sa_len ||
317 bcmp(&ep->src, sp, sp->sa_len) != 0 ||
318 bcmp(&ep->srcmask, sm, sp->sa_len) != 0)
320 if (ep->dst.ss_len != dp->sa_len ||
321 bcmp(&ep->dst, dp, dp->sa_len) != 0 ||
322 bcmp(&ep->dstmask, dm, dp->sa_len) != 0)
325 mtx_unlock(&encapmtx);
329 ep = malloc(sizeof(*ep), M_NETADDR, M_NOWAIT); /*XXX*/
331 mtx_unlock(&encapmtx);
334 bzero(ep, sizeof(*ep));
338 bcopy(sp, &ep->src, sp->sa_len);
339 bcopy(sm, &ep->srcmask, sp->sa_len);
340 bcopy(dp, &ep->dst, dp->sa_len);
341 bcopy(dm, &ep->dstmask, dp->sa_len);
346 mtx_unlock(&encapmtx);
350 const struct encaptab *
351 encap_attach_func(af, proto, func, psw, arg)
354 int (*func)(const struct mbuf *, int, int, void *);
355 const struct protosw *psw;
360 /* sanity check on args */
364 ep = malloc(sizeof(*ep), M_NETADDR, M_NOWAIT); /*XXX*/
367 bzero(ep, sizeof(*ep));
377 mtx_unlock(&encapmtx);
383 const struct encaptab *cookie;
385 const struct encaptab *ep = cookie;
389 LIST_FOREACH(p, &encaptab, chain) {
391 LIST_REMOVE(p, chain);
392 mtx_unlock(&encapmtx);
393 free(p, M_NETADDR); /*XXX*/
397 mtx_unlock(&encapmtx);
403 mask_match(ep, sp, dp)
404 const struct encaptab *ep;
405 const struct sockaddr *sp;
406 const struct sockaddr *dp;
408 struct sockaddr_storage s;
409 struct sockaddr_storage d;
411 const u_int8_t *p, *q;
415 if (sp->sa_len > sizeof(s) || dp->sa_len > sizeof(d))
417 if (sp->sa_family != ep->af || dp->sa_family != ep->af)
419 if (sp->sa_len != ep->src.ss_len || dp->sa_len != ep->dst.ss_len)
424 p = (const u_int8_t *)sp;
425 q = (const u_int8_t *)&ep->srcmask;
427 for (i = 0 ; i < sp->sa_len; i++) {
430 matchlen += (q[i] ? 8 : 0);
433 p = (const u_int8_t *)dp;
434 q = (const u_int8_t *)&ep->dstmask;
436 for (i = 0 ; i < dp->sa_len; i++) {
438 /* XXX rough estimate */
439 matchlen += (q[i] ? 8 : 0);
442 /* need to overwrite len/family portion as we don't compare them */
443 s.ss_len = sp->sa_len;
444 s.ss_family = sp->sa_family;
445 d.ss_len = dp->sa_len;
446 d.ss_family = dp->sa_family;
448 if (bcmp(&s, &ep->src, ep->src.ss_len) == 0 &&
449 bcmp(&d, &ep->dst, ep->dst.ss_len) == 0) {
458 const struct encaptab *ep;
462 tag = m_tag_get(PACKET_TAG_ENCAP, sizeof (void*), M_NOWAIT);
464 *(void**)(tag+1) = ep->arg;
465 m_tag_prepend(m, tag);
476 tag = m_tag_find(m, PACKET_TAG_ENCAP, NULL);
478 p = *(void**)(tag+1);
479 m_tag_delete(m, tag);