1 /* $KAME: ip_encap.c,v 1.41 2001/03/15 08:35:08 itojun Exp $ */
4 * SPDX-License-Identifier: BSD-3-Clause
6 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
9 * Redistribution and use in source and binary forms, with or without
10 * modification, are permitted provided that the following conditions
12 * 1. Redistributions of source code must retain the above copyright
13 * notice, this list of conditions and the following disclaimer.
14 * 2. Redistributions in binary form must reproduce the above copyright
15 * notice, this list of conditions and the following disclaimer in the
16 * documentation and/or other materials provided with the distribution.
17 * 3. Neither the name of the project nor the names of its contributors
18 * may be used to endorse or promote products derived from this software
19 * without specific prior written permission.
21 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34 * My grandfather said that there's a devil inside tunnelling technology...
36 * We have surprisingly many protocols that want packets with IP protocol
37 * #4 or #41. Here's a list of protocols that want protocol #41:
38 * RFC1933 configured tunnel
39 * RFC1933 automatic tunnel
40 * RFC2401 IPsec tunnel
41 * RFC2473 IPv6 generic packet tunnelling
42 * RFC2529 6over4 tunnel
43 * mobile-ip6 (uses RFC2473)
46 * Here's a list of protocol that want protocol #4:
47 * RFC1853 IPv4-in-IPv4 tunnelling
48 * RFC2003 IPv4 encapsulation within IPv4
49 * RFC2344 reverse tunnelling for mobile-ip4
50 * RFC2401 IPsec tunnel
51 * Well, what can I say. They impose different en/decapsulation mechanism
52 * from each other, so they need separate protocol handler. The only one
53 * we can easily determine by protocol # is IPsec, which always has
54 * AH/ESP/IPComp header right after outer IP header.
56 * So, clearly good old protosw does not work for protocol #4 and #41.
57 * The code will let you match protocol via src/dst address pair.
59 /* XXX is M_NETADDR correct? */
61 #include <sys/cdefs.h>
62 __FBSDID("$FreeBSD$");
64 #include "opt_mrouting.h"
66 #include "opt_inet6.h"
68 #include <sys/param.h>
69 #include <sys/systm.h>
71 #include <sys/mutex.h>
72 #include <sys/socket.h>
73 #include <sys/sockio.h>
75 #include <sys/errno.h>
76 #include <sys/protosw.h>
77 #include <sys/queue.h>
80 #include <net/route.h>
82 #include <netinet/in.h>
83 #include <netinet/in_systm.h>
84 #include <netinet/ip.h>
85 #include <netinet/ip_var.h>
86 #include <netinet/ip_encap.h>
89 #include <netinet/ip6.h>
90 #include <netinet6/ip6_var.h>
93 #include <machine/stdarg.h>
95 #include <sys/kernel.h>
96 #include <sys/malloc.h>
97 static MALLOC_DEFINE(M_NETADDR, "encap_export_host", "Export host address structure");
99 static void encap_add(struct encaptab *);
100 static int mask_match(const struct encaptab *, const struct sockaddr *,
101 const struct sockaddr *);
102 static void encap_fillarg(struct mbuf *, void *);
105 * All global variables in ip_encap.c are locked using encapmtx.
107 static struct mtx encapmtx;
108 MTX_SYSINIT(encapmtx, &encapmtx, "encapmtx", MTX_DEF);
109 static LIST_HEAD(, encaptab) encaptab = LIST_HEAD_INITIALIZER(encaptab);
112 * We currently keey encap_init() for source code compatibility reasons --
113 * it's referenced by KAME pieces in netinet6.
122 encap4_input(struct mbuf **mp, int *offp, int proto)
126 struct sockaddr_in s, d;
127 const struct protosw *psw;
128 struct encaptab *ep, *match;
130 int matchprio, off, prio;
134 ip = mtod(m, struct ip *);
136 bzero(&s, sizeof(s));
137 s.sin_family = AF_INET;
138 s.sin_len = sizeof(struct sockaddr_in);
139 s.sin_addr = ip->ip_src;
140 bzero(&d, sizeof(d));
141 d.sin_family = AF_INET;
142 d.sin_len = sizeof(struct sockaddr_in);
143 d.sin_addr = ip->ip_dst;
150 LIST_FOREACH(ep, &encaptab, chain) {
151 if (ep->af != AF_INET)
153 if (ep->proto >= 0 && ep->proto != proto)
156 prio = (*ep->func)(m, off, proto, ep->arg);
159 * it's inbound traffic, we need to match in reverse
162 prio = mask_match(ep, (struct sockaddr *)&d,
163 (struct sockaddr *)&s);
167 * We prioritize the matches by using bit length of the
168 * matches. mask_match() and user-supplied matching function
169 * should return the bit length of the matches (for example,
170 * if both src/dst are matched for IPv4, 64 should be returned).
171 * 0 or negative return value means "it did not match".
173 * The question is, since we have two "mask" portion, we
174 * cannot really define total order between entries.
175 * For example, which of these should be preferred?
176 * mask_match() returns 48 (32 + 16) for both of them.
177 * src=3ffe::/16, dst=3ffe:501::/32
178 * src=3ffe:501::/32, dst=3ffe::/16
180 * We need to loop through all the possible candidates
181 * to get the best match - the search takes O(n) for
182 * n attachments (i.e. interfaces).
186 if (prio > matchprio) {
195 mtx_unlock(&encapmtx);
198 /* found a match, "match" has the best one */
199 if (psw != NULL && psw->pr_input != NULL) {
200 encap_fillarg(m, arg);
201 (*psw->pr_input)(mp, offp, proto);
204 return (IPPROTO_DONE);
207 /* last resort: inject to raw socket */
208 return (rip_input(mp, offp, proto));
214 encap6_input(struct mbuf **mp, int *offp, int proto)
216 struct mbuf *m = *mp;
218 struct sockaddr_in6 s, d;
219 const struct protosw *psw;
220 struct encaptab *ep, *match;
224 ip6 = mtod(m, struct ip6_hdr *);
226 bzero(&s, sizeof(s));
227 s.sin6_family = AF_INET6;
228 s.sin6_len = sizeof(struct sockaddr_in6);
229 s.sin6_addr = ip6->ip6_src;
230 bzero(&d, sizeof(d));
231 d.sin6_family = AF_INET6;
232 d.sin6_len = sizeof(struct sockaddr_in6);
233 d.sin6_addr = ip6->ip6_dst;
240 LIST_FOREACH(ep, &encaptab, chain) {
241 if (ep->af != AF_INET6)
243 if (ep->proto >= 0 && ep->proto != proto)
246 prio = (*ep->func)(m, *offp, proto, ep->arg);
249 * it's inbound traffic, we need to match in reverse
252 prio = mask_match(ep, (struct sockaddr *)&d,
253 (struct sockaddr *)&s);
256 /* see encap4_input() for issues here */
259 if (prio > matchprio) {
268 mtx_unlock(&encapmtx);
272 if (psw != NULL && psw->pr_input != NULL) {
273 encap_fillarg(m, arg);
274 return (*psw->pr_input)(mp, offp, proto);
277 return (IPPROTO_DONE);
281 /* last resort: inject to raw socket */
282 return rip6_input(mp, offp, proto);
286 /*lint -sem(encap_add, custodial(1)) */
288 encap_add(struct encaptab *ep)
291 mtx_assert(&encapmtx, MA_OWNED);
292 LIST_INSERT_HEAD(&encaptab, ep, chain);
296 * sp (src ptr) is always my side, and dp (dst ptr) is always remote side.
297 * length of mask (sm and dm) is assumed to be same as sp/dp.
298 * Return value will be necessary as input (cookie) for encap_detach().
300 const struct encaptab *
301 encap_attach(int af, int proto, const struct sockaddr *sp,
302 const struct sockaddr *sm, const struct sockaddr *dp,
303 const struct sockaddr *dm, const struct protosw *psw, void *arg)
307 /* sanity check on args */
308 if (sp->sa_len > sizeof(ep->src) || dp->sa_len > sizeof(ep->dst))
310 if (sp->sa_len != dp->sa_len)
312 if (af != sp->sa_family || af != dp->sa_family)
315 /* check if anyone have already attached with exactly same config */
317 LIST_FOREACH(ep, &encaptab, chain) {
320 if (ep->proto != proto)
322 if (ep->src.ss_len != sp->sa_len ||
323 bcmp(&ep->src, sp, sp->sa_len) != 0 ||
324 bcmp(&ep->srcmask, sm, sp->sa_len) != 0)
326 if (ep->dst.ss_len != dp->sa_len ||
327 bcmp(&ep->dst, dp, dp->sa_len) != 0 ||
328 bcmp(&ep->dstmask, dm, dp->sa_len) != 0)
331 mtx_unlock(&encapmtx);
335 ep = malloc(sizeof(*ep), M_NETADDR, M_NOWAIT); /*XXX*/
337 mtx_unlock(&encapmtx);
340 bzero(ep, sizeof(*ep));
344 bcopy(sp, &ep->src, sp->sa_len);
345 bcopy(sm, &ep->srcmask, sp->sa_len);
346 bcopy(dp, &ep->dst, dp->sa_len);
347 bcopy(dm, &ep->dstmask, dp->sa_len);
352 mtx_unlock(&encapmtx);
356 const struct encaptab *
357 encap_attach_func(int af, int proto,
358 int (*func)(const struct mbuf *, int, int, void *),
359 const struct protosw *psw, void *arg)
363 /* sanity check on args */
367 ep = malloc(sizeof(*ep), M_NETADDR, M_NOWAIT); /*XXX*/
370 bzero(ep, sizeof(*ep));
380 mtx_unlock(&encapmtx);
385 encap_detach(const struct encaptab *cookie)
387 const struct encaptab *ep = cookie;
391 LIST_FOREACH(p, &encaptab, chain) {
393 LIST_REMOVE(p, chain);
394 mtx_unlock(&encapmtx);
395 free(p, M_NETADDR); /*XXX*/
399 mtx_unlock(&encapmtx);
405 mask_match(const struct encaptab *ep, 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) {
456 encap_fillarg(struct mbuf *m, void *arg)
461 tag = m_tag_get(PACKET_TAG_ENCAP, sizeof(void *), M_NOWAIT);
463 *(void**)(tag+1) = arg;
464 m_tag_prepend(m, tag);
470 encap_getarg(struct mbuf *m)
475 tag = m_tag_find(m, PACKET_TAG_ENCAP, NULL);
477 p = *(void**)(tag+1);
478 m_tag_delete(m, tag);