1 /* $KAME: ip_encap.c,v 1.41 2001/03/15 08:35:08 itojun Exp $ */
4 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions
10 * 1. Redistributions of source code must retain the above copyright
11 * notice, this list of conditions and the following disclaimer.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in the
14 * documentation and/or other materials provided with the distribution.
15 * 3. Neither the name of the project nor the names of its contributors
16 * may be used to endorse or promote products derived from this software
17 * without specific prior written permission.
19 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32 * My grandfather said that there's a devil inside tunnelling technology...
34 * We have surprisingly many protocols that want packets with IP protocol
35 * #4 or #41. Here's a list of protocols that want protocol #41:
36 * RFC1933 configured tunnel
37 * RFC1933 automatic tunnel
38 * RFC2401 IPsec tunnel
39 * RFC2473 IPv6 generic packet tunnelling
40 * RFC2529 6over4 tunnel
41 * mobile-ip6 (uses RFC2473)
44 * Here's a list of protocol that want protocol #4:
45 * RFC1853 IPv4-in-IPv4 tunnelling
46 * RFC2003 IPv4 encapsulation within IPv4
47 * RFC2344 reverse tunnelling for mobile-ip4
48 * RFC2401 IPsec tunnel
49 * Well, what can I say. They impose different en/decapsulation mechanism
50 * from each other, so they need separate protocol handler. The only one
51 * we can easily determine by protocol # is IPsec, which always has
52 * AH/ESP/IPComp header right after outer IP header.
54 * So, clearly good old protosw does not work for protocol #4 and #41.
55 * The code will let you match protocol via src/dst address pair.
57 /* XXX is M_NETADDR correct? */
59 #include <sys/cdefs.h>
60 __FBSDID("$FreeBSD$");
62 #include "opt_mrouting.h"
64 #include "opt_inet6.h"
66 #include <sys/param.h>
67 #include <sys/systm.h>
69 #include <sys/mutex.h>
70 #include <sys/socket.h>
71 #include <sys/sockio.h>
73 #include <sys/errno.h>
74 #include <sys/protosw.h>
75 #include <sys/queue.h>
78 #include <net/route.h>
80 #include <netinet/in.h>
81 #include <netinet/in_systm.h>
82 #include <netinet/ip.h>
83 #include <netinet/ip_var.h>
84 #include <netinet/ip_encap.h>
87 #include <netinet/ip6.h>
88 #include <netinet6/ip6_var.h>
91 #include <machine/stdarg.h>
93 #include <sys/kernel.h>
94 #include <sys/malloc.h>
95 static MALLOC_DEFINE(M_NETADDR, "encap_export_host", "Export host address structure");
97 static void encap_add(struct encaptab *);
98 static int mask_match(const struct encaptab *, const struct sockaddr *,
99 const struct sockaddr *);
100 static void encap_fillarg(struct mbuf *, void *);
103 * All global variables in ip_encap.c are locked using encapmtx.
105 static struct mtx encapmtx;
106 MTX_SYSINIT(encapmtx, &encapmtx, "encapmtx", MTX_DEF);
107 static LIST_HEAD(, encaptab) encaptab = LIST_HEAD_INITIALIZER(encaptab);
110 * We currently keey encap_init() for source code compatibility reasons --
111 * it's referenced by KAME pieces in netinet6.
120 encap4_input(struct mbuf **mp, int *offp, int proto)
124 struct sockaddr_in s, d;
125 const struct protosw *psw;
126 struct encaptab *ep, *match;
128 int matchprio, off, prio;
132 ip = mtod(m, struct ip *);
134 bzero(&s, sizeof(s));
135 s.sin_family = AF_INET;
136 s.sin_len = sizeof(struct sockaddr_in);
137 s.sin_addr = ip->ip_src;
138 bzero(&d, sizeof(d));
139 d.sin_family = AF_INET;
140 d.sin_len = sizeof(struct sockaddr_in);
141 d.sin_addr = ip->ip_dst;
148 LIST_FOREACH(ep, &encaptab, chain) {
149 if (ep->af != AF_INET)
151 if (ep->proto >= 0 && ep->proto != proto)
154 prio = (*ep->func)(m, off, proto, ep->arg);
157 * it's inbound traffic, we need to match in reverse
160 prio = mask_match(ep, (struct sockaddr *)&d,
161 (struct sockaddr *)&s);
165 * We prioritize the matches by using bit length of the
166 * matches. mask_match() and user-supplied matching function
167 * should return the bit length of the matches (for example,
168 * if both src/dst are matched for IPv4, 64 should be returned).
169 * 0 or negative return value means "it did not match".
171 * The question is, since we have two "mask" portion, we
172 * cannot really define total order between entries.
173 * For example, which of these should be preferred?
174 * mask_match() returns 48 (32 + 16) for both of them.
175 * src=3ffe::/16, dst=3ffe:501::/32
176 * src=3ffe:501::/32, dst=3ffe::/16
178 * We need to loop through all the possible candidates
179 * to get the best match - the search takes O(n) for
180 * n attachments (i.e. interfaces).
184 if (prio > matchprio) {
193 mtx_unlock(&encapmtx);
196 /* found a match, "match" has the best one */
197 if (psw != NULL && psw->pr_input != NULL) {
198 encap_fillarg(m, arg);
199 (*psw->pr_input)(mp, offp, proto);
202 return (IPPROTO_DONE);
205 /* last resort: inject to raw socket */
206 return (rip_input(mp, offp, proto));
212 encap6_input(struct mbuf **mp, int *offp, int proto)
214 struct mbuf *m = *mp;
216 struct sockaddr_in6 s, d;
217 const struct protosw *psw;
218 struct encaptab *ep, *match;
222 ip6 = mtod(m, struct ip6_hdr *);
224 bzero(&s, sizeof(s));
225 s.sin6_family = AF_INET6;
226 s.sin6_len = sizeof(struct sockaddr_in6);
227 s.sin6_addr = ip6->ip6_src;
228 bzero(&d, sizeof(d));
229 d.sin6_family = AF_INET6;
230 d.sin6_len = sizeof(struct sockaddr_in6);
231 d.sin6_addr = ip6->ip6_dst;
238 LIST_FOREACH(ep, &encaptab, chain) {
239 if (ep->af != AF_INET6)
241 if (ep->proto >= 0 && ep->proto != proto)
244 prio = (*ep->func)(m, *offp, proto, ep->arg);
247 * it's inbound traffic, we need to match in reverse
250 prio = mask_match(ep, (struct sockaddr *)&d,
251 (struct sockaddr *)&s);
254 /* see encap4_input() for issues here */
257 if (prio > matchprio) {
266 mtx_unlock(&encapmtx);
270 if (psw != NULL && psw->pr_input != NULL) {
271 encap_fillarg(m, arg);
272 return (*psw->pr_input)(mp, offp, proto);
275 return (IPPROTO_DONE);
279 /* last resort: inject to raw socket */
280 return rip6_input(mp, offp, proto);
284 /*lint -sem(encap_add, custodial(1)) */
286 encap_add(struct encaptab *ep)
289 mtx_assert(&encapmtx, MA_OWNED);
290 LIST_INSERT_HEAD(&encaptab, ep, chain);
294 * sp (src ptr) is always my side, and dp (dst ptr) is always remote side.
295 * length of mask (sm and dm) is assumed to be same as sp/dp.
296 * Return value will be necessary as input (cookie) for encap_detach().
298 const struct encaptab *
299 encap_attach(int af, int proto, const struct sockaddr *sp,
300 const struct sockaddr *sm, const struct sockaddr *dp,
301 const struct sockaddr *dm, const struct protosw *psw, void *arg)
305 /* sanity check on args */
306 if (sp->sa_len > sizeof(ep->src) || dp->sa_len > sizeof(ep->dst))
308 if (sp->sa_len != dp->sa_len)
310 if (af != sp->sa_family || af != dp->sa_family)
313 /* check if anyone have already attached with exactly same config */
315 LIST_FOREACH(ep, &encaptab, chain) {
318 if (ep->proto != proto)
320 if (ep->src.ss_len != sp->sa_len ||
321 bcmp(&ep->src, sp, sp->sa_len) != 0 ||
322 bcmp(&ep->srcmask, sm, sp->sa_len) != 0)
324 if (ep->dst.ss_len != dp->sa_len ||
325 bcmp(&ep->dst, dp, dp->sa_len) != 0 ||
326 bcmp(&ep->dstmask, dm, dp->sa_len) != 0)
329 mtx_unlock(&encapmtx);
333 ep = malloc(sizeof(*ep), M_NETADDR, M_NOWAIT); /*XXX*/
335 mtx_unlock(&encapmtx);
338 bzero(ep, sizeof(*ep));
342 bcopy(sp, &ep->src, sp->sa_len);
343 bcopy(sm, &ep->srcmask, sp->sa_len);
344 bcopy(dp, &ep->dst, dp->sa_len);
345 bcopy(dm, &ep->dstmask, dp->sa_len);
350 mtx_unlock(&encapmtx);
354 const struct encaptab *
355 encap_attach_func(int af, int proto,
356 int (*func)(const struct mbuf *, int, int, void *),
357 const struct protosw *psw, void *arg)
361 /* sanity check on args */
365 ep = malloc(sizeof(*ep), M_NETADDR, M_NOWAIT); /*XXX*/
368 bzero(ep, sizeof(*ep));
378 mtx_unlock(&encapmtx);
383 encap_detach(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(const struct encaptab *ep, const struct sockaddr *sp,
404 const struct sockaddr *dp)
406 struct sockaddr_storage s;
407 struct sockaddr_storage d;
409 const u_int8_t *p, *q;
413 if (sp->sa_len > sizeof(s) || dp->sa_len > sizeof(d))
415 if (sp->sa_family != ep->af || dp->sa_family != ep->af)
417 if (sp->sa_len != ep->src.ss_len || dp->sa_len != ep->dst.ss_len)
422 p = (const u_int8_t *)sp;
423 q = (const u_int8_t *)&ep->srcmask;
425 for (i = 0 ; i < sp->sa_len; i++) {
428 matchlen += (q[i] ? 8 : 0);
431 p = (const u_int8_t *)dp;
432 q = (const u_int8_t *)&ep->dstmask;
434 for (i = 0 ; i < dp->sa_len; i++) {
436 /* XXX rough estimate */
437 matchlen += (q[i] ? 8 : 0);
440 /* need to overwrite len/family portion as we don't compare them */
441 s.ss_len = sp->sa_len;
442 s.ss_family = sp->sa_family;
443 d.ss_len = dp->sa_len;
444 d.ss_family = dp->sa_family;
446 if (bcmp(&s, &ep->src, ep->src.ss_len) == 0 &&
447 bcmp(&d, &ep->dst, ep->dst.ss_len) == 0) {
454 encap_fillarg(struct mbuf *m, void *arg)
459 tag = m_tag_get(PACKET_TAG_ENCAP, sizeof(void *), M_NOWAIT);
461 *(void**)(tag+1) = arg;
462 m_tag_prepend(m, tag);
468 encap_getarg(struct mbuf *m)
473 tag = m_tag_find(m, PACKET_TAG_ENCAP, NULL);
475 p = *(void**)(tag+1);
476 m_tag_delete(m, tag);