Deleted Added
full compact
if_stf.c (103487) if_stf.c (105194)
1/* $FreeBSD: head/sys/net/if_stf.c 103487 2002-09-17 14:25:19Z ume $ */
1/* $FreeBSD: head/sys/net/if_stf.c 105194 2002-10-16 01:54:46Z sam $ */
2/* $KAME: if_stf.c,v 1.73 2001/12/03 11:08:30 keiichi Exp $ */
3
4/*
5 * Copyright (C) 2000 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 * 6to4 interface, based on RFC3056.
35 *
36 * 6to4 interface is NOT capable of link-layer (I mean, IPv4) multicasting.
37 * There is no address mapping defined from IPv6 multicast address to IPv4
38 * address. Therefore, we do not have IFF_MULTICAST on the interface.
39 *
40 * Due to the lack of address mapping for link-local addresses, we cannot
41 * throw packets toward link-local addresses (fe80::x). Also, we cannot throw
42 * packets to link-local multicast addresses (ff02::x).
43 *
44 * Here are interesting symptoms due to the lack of link-local address:
45 *
46 * Unicast routing exchange:
47 * - RIPng: Impossible. Uses link-local multicast packet toward ff02::9,
48 * and link-local addresses as nexthop.
49 * - OSPFv6: Impossible. OSPFv6 assumes that there's link-local address
50 * assigned to the link, and makes use of them. Also, HELLO packets use
51 * link-local multicast addresses (ff02::5 and ff02::6).
52 * - BGP4+: Maybe. You can only use global address as nexthop, and global
53 * address as TCP endpoint address.
54 *
55 * Multicast routing protocols:
56 * - PIM: Hello packet cannot be used to discover adjacent PIM routers.
57 * Adjacent PIM routers must be configured manually (is it really spec-wise
58 * correct thing to do?).
59 *
60 * ICMPv6:
61 * - Redirects cannot be used due to the lack of link-local address.
62 *
63 * stf interface does not have, and will not need, a link-local address.
64 * It seems to have no real benefit and does not help the above symptoms much.
65 * Even if we assign link-locals to interface, we cannot really
66 * use link-local unicast/multicast on top of 6to4 cloud (since there's no
67 * encapsulation defined for link-local address), and the above analysis does
68 * not change. RFC3056 does not mandate the assignment of link-local address
69 * either.
70 *
71 * 6to4 interface has security issues. Refer to
72 * http://playground.iijlab.net/i-d/draft-itojun-ipv6-transition-abuse-00.txt
73 * for details. The code tries to filter out some of malicious packets.
74 * Note that there is no way to be 100% secure.
75 */
76
77#include "opt_inet.h"
78#include "opt_inet6.h"
79
80#include <sys/param.h>
81#include <sys/systm.h>
82#include <sys/socket.h>
83#include <sys/sockio.h>
84#include <sys/mbuf.h>
85#include <sys/errno.h>
86#include <sys/kernel.h>
87#include <sys/protosw.h>
88#include <sys/queue.h>
89#include <machine/cpu.h>
90
91#include <sys/malloc.h>
92
93#include <net/if.h>
94#include <net/route.h>
95#include <net/netisr.h>
96#include <net/if_types.h>
97#include <net/if_stf.h>
98
99#include <netinet/in.h>
100#include <netinet/in_systm.h>
101#include <netinet/ip.h>
102#include <netinet/ip_var.h>
103#include <netinet/in_var.h>
104
105#include <netinet/ip6.h>
106#include <netinet6/ip6_var.h>
107#include <netinet6/in6_var.h>
108#include <netinet/ip_ecn.h>
109
110#include <netinet/ip_encap.h>
111
112#include <machine/stdarg.h>
113
114#include <net/net_osdep.h>
115
116#include <net/bpf.h>
117
118#define STFNAME "stf"
119
120#define IN6_IS_ADDR_6TO4(x) (ntohs((x)->s6_addr16[0]) == 0x2002)
121#define GET_V4(x) ((struct in_addr *)(&(x)->s6_addr16[1]))
122
123struct stf_softc {
124 struct ifnet sc_if; /* common area */
125 union {
126 struct route __sc_ro4;
127 struct route_in6 __sc_ro6; /* just for safety */
128 } __sc_ro46;
129#define sc_ro __sc_ro46.__sc_ro4
130 const struct encaptab *encap_cookie;
131 LIST_ENTRY(stf_softc) sc_list; /* all stf's are linked */
132};
133
134static LIST_HEAD(, stf_softc) stf_softc_list;
135
136static MALLOC_DEFINE(M_STF, STFNAME, "6to4 Tunnel Interface");
137static int ip_stf_ttl = 40;
138
139extern struct domain inetdomain;
140struct protosw in_stf_protosw =
141{ SOCK_RAW, &inetdomain, IPPROTO_IPV6, PR_ATOMIC|PR_ADDR,
142 in_stf_input, (pr_output_t*)rip_output, 0, rip_ctloutput,
143 0,
144 0, 0, 0, 0,
145 &rip_usrreqs
146};
147
148static int stfmodevent(module_t, int, void *);
149static int stf_encapcheck(const struct mbuf *, int, int, void *);
150static struct in6_ifaddr *stf_getsrcifa6(struct ifnet *);
151static int stf_output(struct ifnet *, struct mbuf *, struct sockaddr *,
152 struct rtentry *);
153static int isrfc1918addr(struct in_addr *);
154static int stf_checkaddr4(struct stf_softc *, struct in_addr *,
155 struct ifnet *);
156static int stf_checkaddr6(struct stf_softc *, struct in6_addr *,
157 struct ifnet *);
158static void stf_rtrequest(int, struct rtentry *, struct rt_addrinfo *);
159static int stf_ioctl(struct ifnet *, u_long, caddr_t);
160
161int stf_clone_create(struct if_clone *, int);
162void stf_clone_destroy(struct ifnet *);
163
164/* only one clone is currently allowed */
165struct if_clone stf_cloner =
166 IF_CLONE_INITIALIZER(STFNAME, stf_clone_create, stf_clone_destroy, 0, 0);
167
168int
169stf_clone_create(ifc, unit)
170 struct if_clone *ifc;
171 int unit;
172{
173 struct stf_softc *sc;
174
175 sc = malloc(sizeof(struct stf_softc), M_STF, M_WAITOK | M_ZERO);
176 sc->sc_if.if_name = STFNAME;
177 sc->sc_if.if_unit = unit;
178
179 sc->encap_cookie = encap_attach_func(AF_INET, IPPROTO_IPV6,
180 stf_encapcheck, &in_stf_protosw, sc);
181 if (sc->encap_cookie == NULL) {
182 printf("%s: attach failed\n", if_name(&sc->sc_if));
183 free(sc, M_STF);
184 return (ENOMEM);
185 }
186
187 sc->sc_if.if_mtu = IPV6_MMTU;
188 sc->sc_if.if_ioctl = stf_ioctl;
189 sc->sc_if.if_output = stf_output;
190 sc->sc_if.if_type = IFT_STF;
191 sc->sc_if.if_snd.ifq_maxlen = IFQ_MAXLEN;
192 if_attach(&sc->sc_if);
193 bpfattach(&sc->sc_if, DLT_NULL, sizeof(u_int));
194 LIST_INSERT_HEAD(&stf_softc_list, sc, sc_list);
195 return (0);
196}
197
198void
199stf_clone_destroy(ifp)
200 struct ifnet *ifp;
201{
202 int err;
203 struct stf_softc *sc = (void *) ifp;
204
205 LIST_REMOVE(sc, sc_list);
206 err = encap_detach(sc->encap_cookie);
207 KASSERT(err == 0, ("Unexpected error detaching encap_cookie"));
208 bpfdetach(ifp);
209 if_detach(ifp);
210
211 free(sc, M_STF);
212}
213
214static int
215stfmodevent(mod, type, data)
216 module_t mod;
217 int type;
218 void *data;
219{
220
221 switch (type) {
222 case MOD_LOAD:
223 LIST_INIT(&stf_softc_list);
224 if_clone_attach(&stf_cloner);
225
226 break;
227 case MOD_UNLOAD:
228 if_clone_detach(&stf_cloner);
229
230 while (!LIST_EMPTY(&stf_softc_list))
231 stf_clone_destroy(&LIST_FIRST(&stf_softc_list)->sc_if);
232 break;
233 }
234
235 return (0);
236}
237
238static moduledata_t stf_mod = {
239 "if_stf",
240 stfmodevent,
241 0
242};
243
244DECLARE_MODULE(if_stf, stf_mod, SI_SUB_PSEUDO, SI_ORDER_ANY);
245
246static int
247stf_encapcheck(m, off, proto, arg)
248 const struct mbuf *m;
249 int off;
250 int proto;
251 void *arg;
252{
253 struct ip ip;
254 struct in6_ifaddr *ia6;
255 struct stf_softc *sc;
256 struct in_addr a, b;
257
258 sc = (struct stf_softc *)arg;
259 if (sc == NULL)
260 return 0;
261
262 if ((sc->sc_if.if_flags & IFF_UP) == 0)
263 return 0;
264
265 /* IFF_LINK0 means "no decapsulation" */
266 if ((sc->sc_if.if_flags & IFF_LINK0) != 0)
267 return 0;
268
269 if (proto != IPPROTO_IPV6)
270 return 0;
271
272 /* LINTED const cast */
273 m_copydata((struct mbuf *)(uintptr_t)m, 0, sizeof(ip), (caddr_t)&ip);
274
275 if (ip.ip_v != 4)
276 return 0;
277
278 ia6 = stf_getsrcifa6(&sc->sc_if);
279 if (ia6 == NULL)
280 return 0;
281
282 /*
283 * check if IPv4 dst matches the IPv4 address derived from the
284 * local 6to4 address.
285 * success on: dst = 10.1.1.1, ia6->ia_addr = 2002:0a01:0101:...
286 */
287 if (bcmp(GET_V4(&ia6->ia_addr.sin6_addr), &ip.ip_dst,
288 sizeof(ip.ip_dst)) != 0)
289 return 0;
290
291 /*
292 * check if IPv4 src matches the IPv4 address derived from the
293 * local 6to4 address masked by prefixmask.
294 * success on: src = 10.1.1.1, ia6->ia_addr = 2002:0a00:.../24
295 * fail on: src = 10.1.1.1, ia6->ia_addr = 2002:0b00:.../24
296 */
297 bzero(&a, sizeof(a));
298 a.s_addr = GET_V4(&ia6->ia_addr.sin6_addr)->s_addr;
299 a.s_addr &= GET_V4(&ia6->ia_prefixmask.sin6_addr)->s_addr;
300 b = ip.ip_src;
301 b.s_addr &= GET_V4(&ia6->ia_prefixmask.sin6_addr)->s_addr;
302 if (a.s_addr != b.s_addr)
303 return 0;
304
305 /* stf interface makes single side match only */
306 return 32;
307}
308
309static struct in6_ifaddr *
310stf_getsrcifa6(ifp)
311 struct ifnet *ifp;
312{
313 struct ifaddr *ia;
314 struct in_ifaddr *ia4;
315 struct sockaddr_in6 *sin6;
316 struct in_addr in;
317
318 for (ia = TAILQ_FIRST(&ifp->if_addrlist);
319 ia;
320 ia = TAILQ_NEXT(ia, ifa_list))
321 {
322 if (ia->ifa_addr == NULL)
323 continue;
324 if (ia->ifa_addr->sa_family != AF_INET6)
325 continue;
326 sin6 = (struct sockaddr_in6 *)ia->ifa_addr;
327 if (!IN6_IS_ADDR_6TO4(&sin6->sin6_addr))
328 continue;
329
330 bcopy(GET_V4(&sin6->sin6_addr), &in, sizeof(in));
331 LIST_FOREACH(ia4, INADDR_HASH(in.s_addr), ia_hash)
332 if (ia4->ia_addr.sin_addr.s_addr == in.s_addr)
333 break;
334 if (ia4 == NULL)
335 continue;
336
337 return (struct in6_ifaddr *)ia;
338 }
339
340 return NULL;
341}
342
343static int
344stf_output(ifp, m, dst, rt)
345 struct ifnet *ifp;
346 struct mbuf *m;
347 struct sockaddr *dst;
348 struct rtentry *rt;
349{
350 struct stf_softc *sc;
351 struct sockaddr_in6 *dst6;
352 struct in_addr *in4;
353 struct sockaddr_in *dst4;
354 u_int8_t tos;
355 struct ip *ip;
356 struct ip6_hdr *ip6;
357 struct in6_ifaddr *ia6;
358
359 sc = (struct stf_softc*)ifp;
360 dst6 = (struct sockaddr_in6 *)dst;
361
362 /* just in case */
363 if ((ifp->if_flags & IFF_UP) == 0) {
364 m_freem(m);
365 ifp->if_oerrors++;
366 return ENETDOWN;
367 }
368
369 /*
370 * If we don't have an ip4 address that match my inner ip6 address,
371 * we shouldn't generate output. Without this check, we'll end up
372 * using wrong IPv4 source.
373 */
374 ia6 = stf_getsrcifa6(ifp);
375 if (ia6 == NULL) {
376 m_freem(m);
377 ifp->if_oerrors++;
378 return ENETDOWN;
379 }
380
381 if (m->m_len < sizeof(*ip6)) {
382 m = m_pullup(m, sizeof(*ip6));
383 if (!m) {
384 ifp->if_oerrors++;
385 return ENOBUFS;
386 }
387 }
388 ip6 = mtod(m, struct ip6_hdr *);
389 tos = (ntohl(ip6->ip6_flow) >> 20) & 0xff;
390
391 /*
392 * Pickup the right outer dst addr from the list of candidates.
393 * ip6_dst has priority as it may be able to give us shorter IPv4 hops.
394 */
395 if (IN6_IS_ADDR_6TO4(&ip6->ip6_dst))
396 in4 = GET_V4(&ip6->ip6_dst);
397 else if (IN6_IS_ADDR_6TO4(&dst6->sin6_addr))
398 in4 = GET_V4(&dst6->sin6_addr);
399 else {
400 m_freem(m);
401 ifp->if_oerrors++;
402 return ENETUNREACH;
403 }
404
405#if NBPFILTER > 0
406 if (ifp->if_bpf) {
407 /*
408 * We need to prepend the address family as
409 * a four byte field. Cons up a dummy header
410 * to pacify bpf. This is safe because bpf
411 * will only read from the mbuf (i.e., it won't
412 * try to free it or keep a pointer a to it).
413 */
414 struct mbuf m0;
415 u_int32_t af = AF_INET6;
416
417 m0.m_next = m;
418 m0.m_len = 4;
419 m0.m_data = (char *)&af;
420
421#ifdef HAVE_OLD_BPF
422 bpf_mtap(ifp, &m0);
423#else
424 bpf_mtap(ifp->if_bpf, &m0);
425#endif
426 }
427#endif /*NBPFILTER > 0*/
428
429 M_PREPEND(m, sizeof(struct ip), M_DONTWAIT);
430 if (m && m->m_len < sizeof(struct ip))
431 m = m_pullup(m, sizeof(struct ip));
432 if (m == NULL) {
433 ifp->if_oerrors++;
434 return ENOBUFS;
435 }
436 ip = mtod(m, struct ip *);
437
438 bzero(ip, sizeof(*ip));
439
440 bcopy(GET_V4(&((struct sockaddr_in6 *)&ia6->ia_addr)->sin6_addr),
441 &ip->ip_src, sizeof(ip->ip_src));
442 bcopy(in4, &ip->ip_dst, sizeof(ip->ip_dst));
443 ip->ip_p = IPPROTO_IPV6;
444 ip->ip_ttl = ip_stf_ttl;
445 ip->ip_len = m->m_pkthdr.len; /*host order*/
446 if (ifp->if_flags & IFF_LINK1)
447 ip_ecn_ingress(ECN_ALLOWED, &ip->ip_tos, &tos);
448 else
449 ip_ecn_ingress(ECN_NOCARE, &ip->ip_tos, &tos);
450
451 dst4 = (struct sockaddr_in *)&sc->sc_ro.ro_dst;
452 if (dst4->sin_family != AF_INET ||
453 bcmp(&dst4->sin_addr, &ip->ip_dst, sizeof(ip->ip_dst)) != 0) {
454 /* cache route doesn't match */
455 dst4->sin_family = AF_INET;
456 dst4->sin_len = sizeof(struct sockaddr_in);
457 bcopy(&ip->ip_dst, &dst4->sin_addr, sizeof(dst4->sin_addr));
458 if (sc->sc_ro.ro_rt) {
459 RTFREE(sc->sc_ro.ro_rt);
460 sc->sc_ro.ro_rt = NULL;
461 }
462 }
463
464 if (sc->sc_ro.ro_rt == NULL) {
465 rtalloc(&sc->sc_ro);
466 if (sc->sc_ro.ro_rt == NULL) {
467 m_freem(m);
468 ifp->if_oerrors++;
469 return ENETUNREACH;
470 }
471 }
472
473 ifp->if_opackets++;
2/* $KAME: if_stf.c,v 1.73 2001/12/03 11:08:30 keiichi Exp $ */
3
4/*
5 * Copyright (C) 2000 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 * 6to4 interface, based on RFC3056.
35 *
36 * 6to4 interface is NOT capable of link-layer (I mean, IPv4) multicasting.
37 * There is no address mapping defined from IPv6 multicast address to IPv4
38 * address. Therefore, we do not have IFF_MULTICAST on the interface.
39 *
40 * Due to the lack of address mapping for link-local addresses, we cannot
41 * throw packets toward link-local addresses (fe80::x). Also, we cannot throw
42 * packets to link-local multicast addresses (ff02::x).
43 *
44 * Here are interesting symptoms due to the lack of link-local address:
45 *
46 * Unicast routing exchange:
47 * - RIPng: Impossible. Uses link-local multicast packet toward ff02::9,
48 * and link-local addresses as nexthop.
49 * - OSPFv6: Impossible. OSPFv6 assumes that there's link-local address
50 * assigned to the link, and makes use of them. Also, HELLO packets use
51 * link-local multicast addresses (ff02::5 and ff02::6).
52 * - BGP4+: Maybe. You can only use global address as nexthop, and global
53 * address as TCP endpoint address.
54 *
55 * Multicast routing protocols:
56 * - PIM: Hello packet cannot be used to discover adjacent PIM routers.
57 * Adjacent PIM routers must be configured manually (is it really spec-wise
58 * correct thing to do?).
59 *
60 * ICMPv6:
61 * - Redirects cannot be used due to the lack of link-local address.
62 *
63 * stf interface does not have, and will not need, a link-local address.
64 * It seems to have no real benefit and does not help the above symptoms much.
65 * Even if we assign link-locals to interface, we cannot really
66 * use link-local unicast/multicast on top of 6to4 cloud (since there's no
67 * encapsulation defined for link-local address), and the above analysis does
68 * not change. RFC3056 does not mandate the assignment of link-local address
69 * either.
70 *
71 * 6to4 interface has security issues. Refer to
72 * http://playground.iijlab.net/i-d/draft-itojun-ipv6-transition-abuse-00.txt
73 * for details. The code tries to filter out some of malicious packets.
74 * Note that there is no way to be 100% secure.
75 */
76
77#include "opt_inet.h"
78#include "opt_inet6.h"
79
80#include <sys/param.h>
81#include <sys/systm.h>
82#include <sys/socket.h>
83#include <sys/sockio.h>
84#include <sys/mbuf.h>
85#include <sys/errno.h>
86#include <sys/kernel.h>
87#include <sys/protosw.h>
88#include <sys/queue.h>
89#include <machine/cpu.h>
90
91#include <sys/malloc.h>
92
93#include <net/if.h>
94#include <net/route.h>
95#include <net/netisr.h>
96#include <net/if_types.h>
97#include <net/if_stf.h>
98
99#include <netinet/in.h>
100#include <netinet/in_systm.h>
101#include <netinet/ip.h>
102#include <netinet/ip_var.h>
103#include <netinet/in_var.h>
104
105#include <netinet/ip6.h>
106#include <netinet6/ip6_var.h>
107#include <netinet6/in6_var.h>
108#include <netinet/ip_ecn.h>
109
110#include <netinet/ip_encap.h>
111
112#include <machine/stdarg.h>
113
114#include <net/net_osdep.h>
115
116#include <net/bpf.h>
117
118#define STFNAME "stf"
119
120#define IN6_IS_ADDR_6TO4(x) (ntohs((x)->s6_addr16[0]) == 0x2002)
121#define GET_V4(x) ((struct in_addr *)(&(x)->s6_addr16[1]))
122
123struct stf_softc {
124 struct ifnet sc_if; /* common area */
125 union {
126 struct route __sc_ro4;
127 struct route_in6 __sc_ro6; /* just for safety */
128 } __sc_ro46;
129#define sc_ro __sc_ro46.__sc_ro4
130 const struct encaptab *encap_cookie;
131 LIST_ENTRY(stf_softc) sc_list; /* all stf's are linked */
132};
133
134static LIST_HEAD(, stf_softc) stf_softc_list;
135
136static MALLOC_DEFINE(M_STF, STFNAME, "6to4 Tunnel Interface");
137static int ip_stf_ttl = 40;
138
139extern struct domain inetdomain;
140struct protosw in_stf_protosw =
141{ SOCK_RAW, &inetdomain, IPPROTO_IPV6, PR_ATOMIC|PR_ADDR,
142 in_stf_input, (pr_output_t*)rip_output, 0, rip_ctloutput,
143 0,
144 0, 0, 0, 0,
145 &rip_usrreqs
146};
147
148static int stfmodevent(module_t, int, void *);
149static int stf_encapcheck(const struct mbuf *, int, int, void *);
150static struct in6_ifaddr *stf_getsrcifa6(struct ifnet *);
151static int stf_output(struct ifnet *, struct mbuf *, struct sockaddr *,
152 struct rtentry *);
153static int isrfc1918addr(struct in_addr *);
154static int stf_checkaddr4(struct stf_softc *, struct in_addr *,
155 struct ifnet *);
156static int stf_checkaddr6(struct stf_softc *, struct in6_addr *,
157 struct ifnet *);
158static void stf_rtrequest(int, struct rtentry *, struct rt_addrinfo *);
159static int stf_ioctl(struct ifnet *, u_long, caddr_t);
160
161int stf_clone_create(struct if_clone *, int);
162void stf_clone_destroy(struct ifnet *);
163
164/* only one clone is currently allowed */
165struct if_clone stf_cloner =
166 IF_CLONE_INITIALIZER(STFNAME, stf_clone_create, stf_clone_destroy, 0, 0);
167
168int
169stf_clone_create(ifc, unit)
170 struct if_clone *ifc;
171 int unit;
172{
173 struct stf_softc *sc;
174
175 sc = malloc(sizeof(struct stf_softc), M_STF, M_WAITOK | M_ZERO);
176 sc->sc_if.if_name = STFNAME;
177 sc->sc_if.if_unit = unit;
178
179 sc->encap_cookie = encap_attach_func(AF_INET, IPPROTO_IPV6,
180 stf_encapcheck, &in_stf_protosw, sc);
181 if (sc->encap_cookie == NULL) {
182 printf("%s: attach failed\n", if_name(&sc->sc_if));
183 free(sc, M_STF);
184 return (ENOMEM);
185 }
186
187 sc->sc_if.if_mtu = IPV6_MMTU;
188 sc->sc_if.if_ioctl = stf_ioctl;
189 sc->sc_if.if_output = stf_output;
190 sc->sc_if.if_type = IFT_STF;
191 sc->sc_if.if_snd.ifq_maxlen = IFQ_MAXLEN;
192 if_attach(&sc->sc_if);
193 bpfattach(&sc->sc_if, DLT_NULL, sizeof(u_int));
194 LIST_INSERT_HEAD(&stf_softc_list, sc, sc_list);
195 return (0);
196}
197
198void
199stf_clone_destroy(ifp)
200 struct ifnet *ifp;
201{
202 int err;
203 struct stf_softc *sc = (void *) ifp;
204
205 LIST_REMOVE(sc, sc_list);
206 err = encap_detach(sc->encap_cookie);
207 KASSERT(err == 0, ("Unexpected error detaching encap_cookie"));
208 bpfdetach(ifp);
209 if_detach(ifp);
210
211 free(sc, M_STF);
212}
213
214static int
215stfmodevent(mod, type, data)
216 module_t mod;
217 int type;
218 void *data;
219{
220
221 switch (type) {
222 case MOD_LOAD:
223 LIST_INIT(&stf_softc_list);
224 if_clone_attach(&stf_cloner);
225
226 break;
227 case MOD_UNLOAD:
228 if_clone_detach(&stf_cloner);
229
230 while (!LIST_EMPTY(&stf_softc_list))
231 stf_clone_destroy(&LIST_FIRST(&stf_softc_list)->sc_if);
232 break;
233 }
234
235 return (0);
236}
237
238static moduledata_t stf_mod = {
239 "if_stf",
240 stfmodevent,
241 0
242};
243
244DECLARE_MODULE(if_stf, stf_mod, SI_SUB_PSEUDO, SI_ORDER_ANY);
245
246static int
247stf_encapcheck(m, off, proto, arg)
248 const struct mbuf *m;
249 int off;
250 int proto;
251 void *arg;
252{
253 struct ip ip;
254 struct in6_ifaddr *ia6;
255 struct stf_softc *sc;
256 struct in_addr a, b;
257
258 sc = (struct stf_softc *)arg;
259 if (sc == NULL)
260 return 0;
261
262 if ((sc->sc_if.if_flags & IFF_UP) == 0)
263 return 0;
264
265 /* IFF_LINK0 means "no decapsulation" */
266 if ((sc->sc_if.if_flags & IFF_LINK0) != 0)
267 return 0;
268
269 if (proto != IPPROTO_IPV6)
270 return 0;
271
272 /* LINTED const cast */
273 m_copydata((struct mbuf *)(uintptr_t)m, 0, sizeof(ip), (caddr_t)&ip);
274
275 if (ip.ip_v != 4)
276 return 0;
277
278 ia6 = stf_getsrcifa6(&sc->sc_if);
279 if (ia6 == NULL)
280 return 0;
281
282 /*
283 * check if IPv4 dst matches the IPv4 address derived from the
284 * local 6to4 address.
285 * success on: dst = 10.1.1.1, ia6->ia_addr = 2002:0a01:0101:...
286 */
287 if (bcmp(GET_V4(&ia6->ia_addr.sin6_addr), &ip.ip_dst,
288 sizeof(ip.ip_dst)) != 0)
289 return 0;
290
291 /*
292 * check if IPv4 src matches the IPv4 address derived from the
293 * local 6to4 address masked by prefixmask.
294 * success on: src = 10.1.1.1, ia6->ia_addr = 2002:0a00:.../24
295 * fail on: src = 10.1.1.1, ia6->ia_addr = 2002:0b00:.../24
296 */
297 bzero(&a, sizeof(a));
298 a.s_addr = GET_V4(&ia6->ia_addr.sin6_addr)->s_addr;
299 a.s_addr &= GET_V4(&ia6->ia_prefixmask.sin6_addr)->s_addr;
300 b = ip.ip_src;
301 b.s_addr &= GET_V4(&ia6->ia_prefixmask.sin6_addr)->s_addr;
302 if (a.s_addr != b.s_addr)
303 return 0;
304
305 /* stf interface makes single side match only */
306 return 32;
307}
308
309static struct in6_ifaddr *
310stf_getsrcifa6(ifp)
311 struct ifnet *ifp;
312{
313 struct ifaddr *ia;
314 struct in_ifaddr *ia4;
315 struct sockaddr_in6 *sin6;
316 struct in_addr in;
317
318 for (ia = TAILQ_FIRST(&ifp->if_addrlist);
319 ia;
320 ia = TAILQ_NEXT(ia, ifa_list))
321 {
322 if (ia->ifa_addr == NULL)
323 continue;
324 if (ia->ifa_addr->sa_family != AF_INET6)
325 continue;
326 sin6 = (struct sockaddr_in6 *)ia->ifa_addr;
327 if (!IN6_IS_ADDR_6TO4(&sin6->sin6_addr))
328 continue;
329
330 bcopy(GET_V4(&sin6->sin6_addr), &in, sizeof(in));
331 LIST_FOREACH(ia4, INADDR_HASH(in.s_addr), ia_hash)
332 if (ia4->ia_addr.sin_addr.s_addr == in.s_addr)
333 break;
334 if (ia4 == NULL)
335 continue;
336
337 return (struct in6_ifaddr *)ia;
338 }
339
340 return NULL;
341}
342
343static int
344stf_output(ifp, m, dst, rt)
345 struct ifnet *ifp;
346 struct mbuf *m;
347 struct sockaddr *dst;
348 struct rtentry *rt;
349{
350 struct stf_softc *sc;
351 struct sockaddr_in6 *dst6;
352 struct in_addr *in4;
353 struct sockaddr_in *dst4;
354 u_int8_t tos;
355 struct ip *ip;
356 struct ip6_hdr *ip6;
357 struct in6_ifaddr *ia6;
358
359 sc = (struct stf_softc*)ifp;
360 dst6 = (struct sockaddr_in6 *)dst;
361
362 /* just in case */
363 if ((ifp->if_flags & IFF_UP) == 0) {
364 m_freem(m);
365 ifp->if_oerrors++;
366 return ENETDOWN;
367 }
368
369 /*
370 * If we don't have an ip4 address that match my inner ip6 address,
371 * we shouldn't generate output. Without this check, we'll end up
372 * using wrong IPv4 source.
373 */
374 ia6 = stf_getsrcifa6(ifp);
375 if (ia6 == NULL) {
376 m_freem(m);
377 ifp->if_oerrors++;
378 return ENETDOWN;
379 }
380
381 if (m->m_len < sizeof(*ip6)) {
382 m = m_pullup(m, sizeof(*ip6));
383 if (!m) {
384 ifp->if_oerrors++;
385 return ENOBUFS;
386 }
387 }
388 ip6 = mtod(m, struct ip6_hdr *);
389 tos = (ntohl(ip6->ip6_flow) >> 20) & 0xff;
390
391 /*
392 * Pickup the right outer dst addr from the list of candidates.
393 * ip6_dst has priority as it may be able to give us shorter IPv4 hops.
394 */
395 if (IN6_IS_ADDR_6TO4(&ip6->ip6_dst))
396 in4 = GET_V4(&ip6->ip6_dst);
397 else if (IN6_IS_ADDR_6TO4(&dst6->sin6_addr))
398 in4 = GET_V4(&dst6->sin6_addr);
399 else {
400 m_freem(m);
401 ifp->if_oerrors++;
402 return ENETUNREACH;
403 }
404
405#if NBPFILTER > 0
406 if (ifp->if_bpf) {
407 /*
408 * We need to prepend the address family as
409 * a four byte field. Cons up a dummy header
410 * to pacify bpf. This is safe because bpf
411 * will only read from the mbuf (i.e., it won't
412 * try to free it or keep a pointer a to it).
413 */
414 struct mbuf m0;
415 u_int32_t af = AF_INET6;
416
417 m0.m_next = m;
418 m0.m_len = 4;
419 m0.m_data = (char *)&af;
420
421#ifdef HAVE_OLD_BPF
422 bpf_mtap(ifp, &m0);
423#else
424 bpf_mtap(ifp->if_bpf, &m0);
425#endif
426 }
427#endif /*NBPFILTER > 0*/
428
429 M_PREPEND(m, sizeof(struct ip), M_DONTWAIT);
430 if (m && m->m_len < sizeof(struct ip))
431 m = m_pullup(m, sizeof(struct ip));
432 if (m == NULL) {
433 ifp->if_oerrors++;
434 return ENOBUFS;
435 }
436 ip = mtod(m, struct ip *);
437
438 bzero(ip, sizeof(*ip));
439
440 bcopy(GET_V4(&((struct sockaddr_in6 *)&ia6->ia_addr)->sin6_addr),
441 &ip->ip_src, sizeof(ip->ip_src));
442 bcopy(in4, &ip->ip_dst, sizeof(ip->ip_dst));
443 ip->ip_p = IPPROTO_IPV6;
444 ip->ip_ttl = ip_stf_ttl;
445 ip->ip_len = m->m_pkthdr.len; /*host order*/
446 if (ifp->if_flags & IFF_LINK1)
447 ip_ecn_ingress(ECN_ALLOWED, &ip->ip_tos, &tos);
448 else
449 ip_ecn_ingress(ECN_NOCARE, &ip->ip_tos, &tos);
450
451 dst4 = (struct sockaddr_in *)&sc->sc_ro.ro_dst;
452 if (dst4->sin_family != AF_INET ||
453 bcmp(&dst4->sin_addr, &ip->ip_dst, sizeof(ip->ip_dst)) != 0) {
454 /* cache route doesn't match */
455 dst4->sin_family = AF_INET;
456 dst4->sin_len = sizeof(struct sockaddr_in);
457 bcopy(&ip->ip_dst, &dst4->sin_addr, sizeof(dst4->sin_addr));
458 if (sc->sc_ro.ro_rt) {
459 RTFREE(sc->sc_ro.ro_rt);
460 sc->sc_ro.ro_rt = NULL;
461 }
462 }
463
464 if (sc->sc_ro.ro_rt == NULL) {
465 rtalloc(&sc->sc_ro);
466 if (sc->sc_ro.ro_rt == NULL) {
467 m_freem(m);
468 ifp->if_oerrors++;
469 return ENETUNREACH;
470 }
471 }
472
473 ifp->if_opackets++;
474 return ip_output(m, NULL, &sc->sc_ro, 0, NULL);
474 return ip_output(m, NULL, &sc->sc_ro, 0, NULL, NULL);
475}
476
477static int
478isrfc1918addr(in)
479 struct in_addr *in;
480{
481 /*
482 * returns 1 if private address range:
483 * 10.0.0.0/8 172.16.0.0/12 192.168.0.0/16
484 */
485 if ((ntohl(in->s_addr) & 0xff000000) >> 24 == 10 ||
486 (ntohl(in->s_addr) & 0xfff00000) >> 16 == 172 * 256 + 16 ||
487 (ntohl(in->s_addr) & 0xffff0000) >> 16 == 192 * 256 + 168)
488 return 1;
489
490 return 0;
491}
492
493static int
494stf_checkaddr4(sc, in, inifp)
495 struct stf_softc *sc;
496 struct in_addr *in;
497 struct ifnet *inifp; /* incoming interface */
498{
499 struct in_ifaddr *ia4;
500
501 /*
502 * reject packets with the following address:
503 * 224.0.0.0/4 0.0.0.0/8 127.0.0.0/8 255.0.0.0/8
504 */
505 if (IN_MULTICAST(ntohl(in->s_addr)))
506 return -1;
507 switch ((ntohl(in->s_addr) & 0xff000000) >> 24) {
508 case 0: case 127: case 255:
509 return -1;
510 }
511
512 /*
513 * reject packets with private address range.
514 * (requirement from RFC3056 section 2 1st paragraph)
515 */
516 if (isrfc1918addr(in))
517 return -1;
518
519 /*
520 * reject packets with broadcast
521 */
522 for (ia4 = TAILQ_FIRST(&in_ifaddrhead);
523 ia4;
524 ia4 = TAILQ_NEXT(ia4, ia_link))
525 {
526 if ((ia4->ia_ifa.ifa_ifp->if_flags & IFF_BROADCAST) == 0)
527 continue;
528 if (in->s_addr == ia4->ia_broadaddr.sin_addr.s_addr)
529 return -1;
530 }
531
532 /*
533 * perform ingress filter
534 */
535 if (sc && (sc->sc_if.if_flags & IFF_LINK2) == 0 && inifp) {
536 struct sockaddr_in sin;
537 struct rtentry *rt;
538
539 bzero(&sin, sizeof(sin));
540 sin.sin_family = AF_INET;
541 sin.sin_len = sizeof(struct sockaddr_in);
542 sin.sin_addr = *in;
543 rt = rtalloc1((struct sockaddr *)&sin, 0, 0UL);
544 if (!rt || rt->rt_ifp != inifp) {
545#if 0
546 log(LOG_WARNING, "%s: packet from 0x%x dropped "
547 "due to ingress filter\n", if_name(&sc->sc_if),
548 (u_int32_t)ntohl(sin.sin_addr.s_addr));
549#endif
550 if (rt)
551 rtfree(rt);
552 return -1;
553 }
554 rtfree(rt);
555 }
556
557 return 0;
558}
559
560static int
561stf_checkaddr6(sc, in6, inifp)
562 struct stf_softc *sc;
563 struct in6_addr *in6;
564 struct ifnet *inifp; /* incoming interface */
565{
566 /*
567 * check 6to4 addresses
568 */
569 if (IN6_IS_ADDR_6TO4(in6))
570 return stf_checkaddr4(sc, GET_V4(in6), inifp);
571
572 /*
573 * reject anything that look suspicious. the test is implemented
574 * in ip6_input too, but we check here as well to
575 * (1) reject bad packets earlier, and
576 * (2) to be safe against future ip6_input change.
577 */
578 if (IN6_IS_ADDR_V4COMPAT(in6) || IN6_IS_ADDR_V4MAPPED(in6))
579 return -1;
580
581 return 0;
582}
583
584void
585in_stf_input(m, off)
586 struct mbuf *m;
587 int off;
588{
589 int proto;
590 struct stf_softc *sc;
591 struct ip *ip;
592 struct ip6_hdr *ip6;
593 u_int8_t otos, itos;
594 int len, isr;
595 struct ifqueue *ifq = NULL;
596 struct ifnet *ifp;
597
598 proto = mtod(m, struct ip *)->ip_p;
599
600 if (proto != IPPROTO_IPV6) {
601 m_freem(m);
602 return;
603 }
604
605 ip = mtod(m, struct ip *);
606
607 sc = (struct stf_softc *)encap_getarg(m);
608
609 if (sc == NULL || (sc->sc_if.if_flags & IFF_UP) == 0) {
610 m_freem(m);
611 return;
612 }
613
614 ifp = &sc->sc_if;
615
616 /*
617 * perform sanity check against outer src/dst.
618 * for source, perform ingress filter as well.
619 */
620 if (stf_checkaddr4(sc, &ip->ip_dst, NULL) < 0 ||
621 stf_checkaddr4(sc, &ip->ip_src, m->m_pkthdr.rcvif) < 0) {
622 m_freem(m);
623 return;
624 }
625
626 otos = ip->ip_tos;
627 m_adj(m, off);
628
629 if (m->m_len < sizeof(*ip6)) {
630 m = m_pullup(m, sizeof(*ip6));
631 if (!m)
632 return;
633 }
634 ip6 = mtod(m, struct ip6_hdr *);
635
636 /*
637 * perform sanity check against inner src/dst.
638 * for source, perform ingress filter as well.
639 */
640 if (stf_checkaddr6(sc, &ip6->ip6_dst, NULL) < 0 ||
641 stf_checkaddr6(sc, &ip6->ip6_src, m->m_pkthdr.rcvif) < 0) {
642 m_freem(m);
643 return;
644 }
645
646 itos = (ntohl(ip6->ip6_flow) >> 20) & 0xff;
647 if ((ifp->if_flags & IFF_LINK1) != 0)
648 ip_ecn_egress(ECN_ALLOWED, &otos, &itos);
649 else
650 ip_ecn_egress(ECN_NOCARE, &otos, &itos);
651 ip6->ip6_flow &= ~htonl(0xff << 20);
652 ip6->ip6_flow |= htonl((u_int32_t)itos << 20);
653
654 m->m_pkthdr.rcvif = ifp;
655
656 if (ifp->if_bpf) {
657 /*
658 * We need to prepend the address family as
659 * a four byte field. Cons up a dummy header
660 * to pacify bpf. This is safe because bpf
661 * will only read from the mbuf (i.e., it won't
662 * try to free it or keep a pointer a to it).
663 */
664 struct mbuf m0;
665 u_int32_t af = AF_INET6;
666
667 m0.m_next = m;
668 m0.m_len = 4;
669 m0.m_data = (char *)&af;
670
671#ifdef HAVE_OLD_BPF
672 bpf_mtap(ifp, &m0);
673#else
674 bpf_mtap(ifp->if_bpf, &m0);
675#endif
676 }
677
678 /*
679 * Put the packet to the network layer input queue according to the
680 * specified address family.
681 * See net/if_gif.c for possible issues with packet processing
682 * reorder due to extra queueing.
683 */
684 ifq = &ip6intrq;
685 isr = NETISR_IPV6;
686
687 len = m->m_pkthdr.len;
688 if (! IF_HANDOFF(ifq, m, NULL))
689 return;
690 schednetisr(isr);
691 ifp->if_ipackets++;
692 ifp->if_ibytes += len;
693}
694
695/* ARGSUSED */
696static void
697stf_rtrequest(cmd, rt, info)
698 int cmd;
699 struct rtentry *rt;
700 struct rt_addrinfo *info;
701{
702
703 if (rt)
704 rt->rt_rmx.rmx_mtu = IPV6_MMTU;
705}
706
707static int
708stf_ioctl(ifp, cmd, data)
709 struct ifnet *ifp;
710 u_long cmd;
711 caddr_t data;
712{
713 struct ifaddr *ifa;
714 struct ifreq *ifr;
715 struct sockaddr_in6 *sin6;
716 int error;
717
718 error = 0;
719 switch (cmd) {
720 case SIOCSIFADDR:
721 ifa = (struct ifaddr *)data;
722 if (ifa == NULL || ifa->ifa_addr->sa_family != AF_INET6) {
723 error = EAFNOSUPPORT;
724 break;
725 }
726 sin6 = (struct sockaddr_in6 *)ifa->ifa_addr;
727 if (IN6_IS_ADDR_6TO4(&sin6->sin6_addr) &&
728 !isrfc1918addr(GET_V4(&sin6->sin6_addr))) {
729 ifa->ifa_rtrequest = stf_rtrequest;
730 ifp->if_flags |= IFF_UP;
731 } else
732 error = EINVAL;
733 break;
734
735 case SIOCADDMULTI:
736 case SIOCDELMULTI:
737 ifr = (struct ifreq *)data;
738 if (ifr && ifr->ifr_addr.sa_family == AF_INET6)
739 ;
740 else
741 error = EAFNOSUPPORT;
742 break;
743
744 default:
745 error = EINVAL;
746 break;
747 }
748
749 return error;
750}
475}
476
477static int
478isrfc1918addr(in)
479 struct in_addr *in;
480{
481 /*
482 * returns 1 if private address range:
483 * 10.0.0.0/8 172.16.0.0/12 192.168.0.0/16
484 */
485 if ((ntohl(in->s_addr) & 0xff000000) >> 24 == 10 ||
486 (ntohl(in->s_addr) & 0xfff00000) >> 16 == 172 * 256 + 16 ||
487 (ntohl(in->s_addr) & 0xffff0000) >> 16 == 192 * 256 + 168)
488 return 1;
489
490 return 0;
491}
492
493static int
494stf_checkaddr4(sc, in, inifp)
495 struct stf_softc *sc;
496 struct in_addr *in;
497 struct ifnet *inifp; /* incoming interface */
498{
499 struct in_ifaddr *ia4;
500
501 /*
502 * reject packets with the following address:
503 * 224.0.0.0/4 0.0.0.0/8 127.0.0.0/8 255.0.0.0/8
504 */
505 if (IN_MULTICAST(ntohl(in->s_addr)))
506 return -1;
507 switch ((ntohl(in->s_addr) & 0xff000000) >> 24) {
508 case 0: case 127: case 255:
509 return -1;
510 }
511
512 /*
513 * reject packets with private address range.
514 * (requirement from RFC3056 section 2 1st paragraph)
515 */
516 if (isrfc1918addr(in))
517 return -1;
518
519 /*
520 * reject packets with broadcast
521 */
522 for (ia4 = TAILQ_FIRST(&in_ifaddrhead);
523 ia4;
524 ia4 = TAILQ_NEXT(ia4, ia_link))
525 {
526 if ((ia4->ia_ifa.ifa_ifp->if_flags & IFF_BROADCAST) == 0)
527 continue;
528 if (in->s_addr == ia4->ia_broadaddr.sin_addr.s_addr)
529 return -1;
530 }
531
532 /*
533 * perform ingress filter
534 */
535 if (sc && (sc->sc_if.if_flags & IFF_LINK2) == 0 && inifp) {
536 struct sockaddr_in sin;
537 struct rtentry *rt;
538
539 bzero(&sin, sizeof(sin));
540 sin.sin_family = AF_INET;
541 sin.sin_len = sizeof(struct sockaddr_in);
542 sin.sin_addr = *in;
543 rt = rtalloc1((struct sockaddr *)&sin, 0, 0UL);
544 if (!rt || rt->rt_ifp != inifp) {
545#if 0
546 log(LOG_WARNING, "%s: packet from 0x%x dropped "
547 "due to ingress filter\n", if_name(&sc->sc_if),
548 (u_int32_t)ntohl(sin.sin_addr.s_addr));
549#endif
550 if (rt)
551 rtfree(rt);
552 return -1;
553 }
554 rtfree(rt);
555 }
556
557 return 0;
558}
559
560static int
561stf_checkaddr6(sc, in6, inifp)
562 struct stf_softc *sc;
563 struct in6_addr *in6;
564 struct ifnet *inifp; /* incoming interface */
565{
566 /*
567 * check 6to4 addresses
568 */
569 if (IN6_IS_ADDR_6TO4(in6))
570 return stf_checkaddr4(sc, GET_V4(in6), inifp);
571
572 /*
573 * reject anything that look suspicious. the test is implemented
574 * in ip6_input too, but we check here as well to
575 * (1) reject bad packets earlier, and
576 * (2) to be safe against future ip6_input change.
577 */
578 if (IN6_IS_ADDR_V4COMPAT(in6) || IN6_IS_ADDR_V4MAPPED(in6))
579 return -1;
580
581 return 0;
582}
583
584void
585in_stf_input(m, off)
586 struct mbuf *m;
587 int off;
588{
589 int proto;
590 struct stf_softc *sc;
591 struct ip *ip;
592 struct ip6_hdr *ip6;
593 u_int8_t otos, itos;
594 int len, isr;
595 struct ifqueue *ifq = NULL;
596 struct ifnet *ifp;
597
598 proto = mtod(m, struct ip *)->ip_p;
599
600 if (proto != IPPROTO_IPV6) {
601 m_freem(m);
602 return;
603 }
604
605 ip = mtod(m, struct ip *);
606
607 sc = (struct stf_softc *)encap_getarg(m);
608
609 if (sc == NULL || (sc->sc_if.if_flags & IFF_UP) == 0) {
610 m_freem(m);
611 return;
612 }
613
614 ifp = &sc->sc_if;
615
616 /*
617 * perform sanity check against outer src/dst.
618 * for source, perform ingress filter as well.
619 */
620 if (stf_checkaddr4(sc, &ip->ip_dst, NULL) < 0 ||
621 stf_checkaddr4(sc, &ip->ip_src, m->m_pkthdr.rcvif) < 0) {
622 m_freem(m);
623 return;
624 }
625
626 otos = ip->ip_tos;
627 m_adj(m, off);
628
629 if (m->m_len < sizeof(*ip6)) {
630 m = m_pullup(m, sizeof(*ip6));
631 if (!m)
632 return;
633 }
634 ip6 = mtod(m, struct ip6_hdr *);
635
636 /*
637 * perform sanity check against inner src/dst.
638 * for source, perform ingress filter as well.
639 */
640 if (stf_checkaddr6(sc, &ip6->ip6_dst, NULL) < 0 ||
641 stf_checkaddr6(sc, &ip6->ip6_src, m->m_pkthdr.rcvif) < 0) {
642 m_freem(m);
643 return;
644 }
645
646 itos = (ntohl(ip6->ip6_flow) >> 20) & 0xff;
647 if ((ifp->if_flags & IFF_LINK1) != 0)
648 ip_ecn_egress(ECN_ALLOWED, &otos, &itos);
649 else
650 ip_ecn_egress(ECN_NOCARE, &otos, &itos);
651 ip6->ip6_flow &= ~htonl(0xff << 20);
652 ip6->ip6_flow |= htonl((u_int32_t)itos << 20);
653
654 m->m_pkthdr.rcvif = ifp;
655
656 if (ifp->if_bpf) {
657 /*
658 * We need to prepend the address family as
659 * a four byte field. Cons up a dummy header
660 * to pacify bpf. This is safe because bpf
661 * will only read from the mbuf (i.e., it won't
662 * try to free it or keep a pointer a to it).
663 */
664 struct mbuf m0;
665 u_int32_t af = AF_INET6;
666
667 m0.m_next = m;
668 m0.m_len = 4;
669 m0.m_data = (char *)&af;
670
671#ifdef HAVE_OLD_BPF
672 bpf_mtap(ifp, &m0);
673#else
674 bpf_mtap(ifp->if_bpf, &m0);
675#endif
676 }
677
678 /*
679 * Put the packet to the network layer input queue according to the
680 * specified address family.
681 * See net/if_gif.c for possible issues with packet processing
682 * reorder due to extra queueing.
683 */
684 ifq = &ip6intrq;
685 isr = NETISR_IPV6;
686
687 len = m->m_pkthdr.len;
688 if (! IF_HANDOFF(ifq, m, NULL))
689 return;
690 schednetisr(isr);
691 ifp->if_ipackets++;
692 ifp->if_ibytes += len;
693}
694
695/* ARGSUSED */
696static void
697stf_rtrequest(cmd, rt, info)
698 int cmd;
699 struct rtentry *rt;
700 struct rt_addrinfo *info;
701{
702
703 if (rt)
704 rt->rt_rmx.rmx_mtu = IPV6_MMTU;
705}
706
707static int
708stf_ioctl(ifp, cmd, data)
709 struct ifnet *ifp;
710 u_long cmd;
711 caddr_t data;
712{
713 struct ifaddr *ifa;
714 struct ifreq *ifr;
715 struct sockaddr_in6 *sin6;
716 int error;
717
718 error = 0;
719 switch (cmd) {
720 case SIOCSIFADDR:
721 ifa = (struct ifaddr *)data;
722 if (ifa == NULL || ifa->ifa_addr->sa_family != AF_INET6) {
723 error = EAFNOSUPPORT;
724 break;
725 }
726 sin6 = (struct sockaddr_in6 *)ifa->ifa_addr;
727 if (IN6_IS_ADDR_6TO4(&sin6->sin6_addr) &&
728 !isrfc1918addr(GET_V4(&sin6->sin6_addr))) {
729 ifa->ifa_rtrequest = stf_rtrequest;
730 ifp->if_flags |= IFF_UP;
731 } else
732 error = EINVAL;
733 break;
734
735 case SIOCADDMULTI:
736 case SIOCDELMULTI:
737 ifr = (struct ifreq *)data;
738 if (ifr && ifr->ifr_addr.sa_family == AF_INET6)
739 ;
740 else
741 error = EAFNOSUPPORT;
742 break;
743
744 default:
745 error = EINVAL;
746 break;
747 }
748
749 return error;
750}