in6_gif.c revision 8e20fa5ae93243e19700ca06c01524b90fe3b784
1/*-
2 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
3 * All rights reserved.
4 *
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
7 * are met:
8 * 1. Redistributions of source code must retain the above copyright
9 *    notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 *    notice, this list of conditions and the following disclaimer in the
12 *    documentation and/or other materials provided with the distribution.
13 * 3. Neither the name of the project nor the names of its contributors
14 *    may be used to endorse or promote products derived from this software
15 *    without specific prior written permission.
16 *
17 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20 * ARE DISCLAIMED.  IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
27 * SUCH DAMAGE.
28 *
29 *	$KAME: in6_gif.c,v 1.49 2001/05/14 14:02:17 itojun Exp $
30 */
31
32#include <sys/cdefs.h>
33__FBSDID("$FreeBSD$");
34
35#include "opt_inet.h"
36#include "opt_inet6.h"
37
38#include <sys/param.h>
39#include <sys/systm.h>
40#include <sys/socket.h>
41#include <sys/sockio.h>
42#include <sys/mbuf.h>
43#include <sys/errno.h>
44#include <sys/kernel.h>
45#include <sys/queue.h>
46#include <sys/syslog.h>
47#include <sys/sysctl.h>
48#include <sys/protosw.h>
49#include <sys/malloc.h>
50
51#include <net/if.h>
52#include <net/route.h>
53
54#include <netinet/in.h>
55#include <netinet/in_systm.h>
56#ifdef INET
57#include <netinet/ip.h>
58#endif
59#include <netinet/ip_encap.h>
60#ifdef INET6
61#include <netinet/ip6.h>
62#include <netinet6/ip6_var.h>
63#include <netinet6/in6_gif.h>
64#include <netinet6/in6_var.h>
65#endif
66#include <netinet6/ip6protosw.h>
67#include <netinet/ip_ecn.h>
68#ifdef INET6
69#include <netinet6/ip6_ecn.h>
70#endif
71
72#include <net/if_gif.h>
73
74VNET_DEFINE(int, ip6_gif_hlim) = GIF_HLIM;
75#define	V_ip6_gif_hlim			VNET(ip6_gif_hlim)
76
77SYSCTL_DECL(_net_inet6_ip6);
78SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_GIF_HLIM, gifhlim, CTLFLAG_RW,
79    &VNET_NAME(ip6_gif_hlim), 0, "");
80
81static int gif_validate6(const struct ip6_hdr *, struct gif_softc *,
82			 struct ifnet *);
83
84extern  struct domain inet6domain;
85struct ip6protosw in6_gif_protosw = {
86	.pr_type =	SOCK_RAW,
87	.pr_domain =	&inet6domain,
88	.pr_protocol =	0,			/* IPPROTO_IPV[46] */
89	.pr_flags =	PR_ATOMIC|PR_ADDR,
90	.pr_input =	in6_gif_input,
91	.pr_output =	rip6_output,
92	.pr_ctloutput =	rip6_ctloutput,
93	.pr_usrreqs =	&rip6_usrreqs
94};
95
96int
97in6_gif_output(struct ifnet *ifp,
98    int family,			/* family of the packet to be encapsulate */
99    struct mbuf *m)
100{
101	struct gif_softc *sc = ifp->if_softc;
102	struct sockaddr_in6 *dst = (struct sockaddr_in6 *)&sc->gif_ro6.ro_dst;
103	struct sockaddr_in6 *sin6_src = (struct sockaddr_in6 *)sc->gif_psrc;
104	struct sockaddr_in6 *sin6_dst = (struct sockaddr_in6 *)sc->gif_pdst;
105	struct ip6_hdr *ip6;
106	struct etherip_header eiphdr;
107	int error, len, proto;
108	u_int8_t itos, otos;
109
110	GIF_LOCK_ASSERT(sc);
111
112	if (sin6_src == NULL || sin6_dst == NULL ||
113	    sin6_src->sin6_family != AF_INET6 ||
114	    sin6_dst->sin6_family != AF_INET6) {
115		m_freem(m);
116		return EAFNOSUPPORT;
117	}
118
119	switch (family) {
120#ifdef INET
121	case AF_INET:
122	    {
123		struct ip *ip;
124
125		proto = IPPROTO_IPV4;
126		if (m->m_len < sizeof(*ip)) {
127			m = m_pullup(m, sizeof(*ip));
128			if (!m)
129				return ENOBUFS;
130		}
131		ip = mtod(m, struct ip *);
132		itos = ip->ip_tos;
133		break;
134	    }
135#endif
136#ifdef INET6
137	case AF_INET6:
138	    {
139		struct ip6_hdr *ip6;
140		proto = IPPROTO_IPV6;
141		if (m->m_len < sizeof(*ip6)) {
142			m = m_pullup(m, sizeof(*ip6));
143			if (!m)
144				return ENOBUFS;
145		}
146		ip6 = mtod(m, struct ip6_hdr *);
147		itos = (ntohl(ip6->ip6_flow) >> 20) & 0xff;
148		break;
149	    }
150#endif
151	case AF_LINK:
152		proto = IPPROTO_ETHERIP;
153
154		/*
155		 * GIF_SEND_REVETHIP (disabled by default) intentionally
156		 * sends an EtherIP packet with revered version field in
157		 * the header.  This is a knob for backward compatibility
158		 * with FreeBSD 7.2R or prior.
159		 */
160		if ((sc->gif_options & GIF_SEND_REVETHIP)) {
161			eiphdr.eip_ver = 0;
162			eiphdr.eip_resvl = ETHERIP_VERSION;
163			eiphdr.eip_resvh = 0;
164		} else {
165			eiphdr.eip_ver = ETHERIP_VERSION;
166			eiphdr.eip_resvl = 0;
167			eiphdr.eip_resvh = 0;
168		}
169		/* prepend Ethernet-in-IP header */
170		M_PREPEND(m, sizeof(struct etherip_header), M_NOWAIT);
171		if (m && m->m_len < sizeof(struct etherip_header))
172			m = m_pullup(m, sizeof(struct etherip_header));
173		if (m == NULL)
174			return ENOBUFS;
175		bcopy(&eiphdr, mtod(m, struct etherip_header *),
176		    sizeof(struct etherip_header));
177		break;
178
179	default:
180#ifdef DEBUG
181		printf("in6_gif_output: warning: unknown family %d passed\n",
182			family);
183#endif
184		m_freem(m);
185		return EAFNOSUPPORT;
186	}
187
188	/* prepend new IP header */
189	len = sizeof(struct ip6_hdr);
190#ifndef __NO_STRICT_ALIGNMENT
191	if (family == AF_LINK)
192		len += ETHERIP_ALIGN;
193#endif
194	M_PREPEND(m, len, M_NOWAIT);
195	if (m != NULL && m->m_len < len)
196		m = m_pullup(m, len);
197	if (m == NULL) {
198		printf("ENOBUFS in in6_gif_output %d\n", __LINE__);
199		return ENOBUFS;
200	}
201#ifndef __NO_STRICT_ALIGNMENT
202	if (family == AF_LINK) {
203		len = mtod(m, vm_offset_t) & 3;
204		KASSERT(len == 0 || len == ETHERIP_ALIGN,
205		    ("in6_gif_output: unexpected misalignment"));
206		m->m_data += len;
207		m->m_len -= ETHERIP_ALIGN;
208	}
209#endif
210
211	ip6 = mtod(m, struct ip6_hdr *);
212	ip6->ip6_flow	= 0;
213	ip6->ip6_vfc	&= ~IPV6_VERSION_MASK;
214	ip6->ip6_vfc	|= IPV6_VERSION;
215	ip6->ip6_plen	= htons((u_short)m->m_pkthdr.len);
216	ip6->ip6_nxt	= proto;
217	ip6->ip6_hlim	= V_ip6_gif_hlim;
218	ip6->ip6_src	= sin6_src->sin6_addr;
219	/* bidirectional configured tunnel mode */
220	if (!IN6_IS_ADDR_UNSPECIFIED(&sin6_dst->sin6_addr))
221		ip6->ip6_dst = sin6_dst->sin6_addr;
222	else  {
223		m_freem(m);
224		return ENETUNREACH;
225	}
226	ip_ecn_ingress((ifp->if_flags & IFF_LINK1) ? ECN_ALLOWED : ECN_NOCARE,
227		       &otos, &itos);
228	ip6->ip6_flow &= ~htonl(0xff << 20);
229	ip6->ip6_flow |= htonl((u_int32_t)otos << 20);
230
231	M_SETFIB(m, sc->gif_fibnum);
232
233	if (dst->sin6_family != sin6_dst->sin6_family ||
234	     !IN6_ARE_ADDR_EQUAL(&dst->sin6_addr, &sin6_dst->sin6_addr)) {
235		/* cache route doesn't match */
236		bzero(dst, sizeof(*dst));
237		dst->sin6_family = sin6_dst->sin6_family;
238		dst->sin6_len = sizeof(struct sockaddr_in6);
239		dst->sin6_addr = sin6_dst->sin6_addr;
240		if (sc->gif_ro6.ro_rt) {
241			RTFREE(sc->gif_ro6.ro_rt);
242			sc->gif_ro6.ro_rt = NULL;
243		}
244#if 0
245		GIF2IFP(sc)->if_mtu = GIF_MTU;
246#endif
247	}
248
249	if (sc->gif_ro6.ro_rt == NULL) {
250		in6_rtalloc(&sc->gif_ro6, sc->gif_fibnum);
251		if (sc->gif_ro6.ro_rt == NULL) {
252			m_freem(m);
253			return ENETUNREACH;
254		}
255
256		/* if it constitutes infinite encapsulation, punt. */
257		if (sc->gif_ro.ro_rt->rt_ifp == ifp) {
258			m_freem(m);
259			return ENETUNREACH;	/*XXX*/
260		}
261#if 0
262		ifp->if_mtu = sc->gif_ro6.ro_rt->rt_ifp->if_mtu
263			- sizeof(struct ip6_hdr);
264#endif
265	}
266
267#ifdef IPV6_MINMTU
268	/*
269	 * force fragmentation to minimum MTU, to avoid path MTU discovery.
270	 * it is too painful to ask for resend of inner packet, to achieve
271	 * path MTU discovery for encapsulated packets.
272	 */
273	error = ip6_output(m, 0, &sc->gif_ro6, IPV6_MINMTU, 0, NULL, NULL);
274#else
275	error = ip6_output(m, 0, &sc->gif_ro6, 0, 0, NULL, NULL);
276#endif
277
278	if (!(GIF2IFP(sc)->if_flags & IFF_LINK0) &&
279	    sc->gif_ro6.ro_rt != NULL) {
280		RTFREE(sc->gif_ro6.ro_rt);
281		sc->gif_ro6.ro_rt = NULL;
282	}
283
284	return (error);
285}
286
287int
288in6_gif_input(struct mbuf **mp, int *offp, int proto)
289{
290	struct mbuf *m = *mp;
291	struct ifnet *gifp = NULL;
292	struct gif_softc *sc;
293	struct ip6_hdr *ip6;
294	int af = 0;
295	u_int32_t otos;
296
297	ip6 = mtod(m, struct ip6_hdr *);
298
299	sc = (struct gif_softc *)encap_getarg(m);
300	if (sc == NULL) {
301		m_freem(m);
302		V_ip6stat.ip6s_nogif++;
303		return IPPROTO_DONE;
304	}
305
306	gifp = GIF2IFP(sc);
307	if (gifp == NULL || (gifp->if_flags & IFF_UP) == 0) {
308		m_freem(m);
309		V_ip6stat.ip6s_nogif++;
310		return IPPROTO_DONE;
311	}
312
313	otos = ip6->ip6_flow;
314	m_adj(m, *offp);
315
316	switch (proto) {
317#ifdef INET
318	case IPPROTO_IPV4:
319	    {
320		struct ip *ip;
321		u_int8_t otos8;
322		af = AF_INET;
323		otos8 = (ntohl(otos) >> 20) & 0xff;
324		if (m->m_len < sizeof(*ip)) {
325			m = m_pullup(m, sizeof(*ip));
326			if (!m)
327				return IPPROTO_DONE;
328		}
329		ip = mtod(m, struct ip *);
330		if (ip_ecn_egress((gifp->if_flags & IFF_LINK1) ?
331				  ECN_ALLOWED : ECN_NOCARE,
332				  &otos8, &ip->ip_tos) == 0) {
333			m_freem(m);
334			return IPPROTO_DONE;
335		}
336		break;
337	    }
338#endif /* INET */
339#ifdef INET6
340	case IPPROTO_IPV6:
341	    {
342		struct ip6_hdr *ip6;
343		af = AF_INET6;
344		if (m->m_len < sizeof(*ip6)) {
345			m = m_pullup(m, sizeof(*ip6));
346			if (!m)
347				return IPPROTO_DONE;
348		}
349		ip6 = mtod(m, struct ip6_hdr *);
350		if (ip6_ecn_egress((gifp->if_flags & IFF_LINK1) ?
351				   ECN_ALLOWED : ECN_NOCARE,
352				   &otos, &ip6->ip6_flow) == 0) {
353			m_freem(m);
354			return IPPROTO_DONE;
355		}
356		break;
357	    }
358#endif
359	case IPPROTO_ETHERIP:
360		af = AF_LINK;
361		break;
362
363	default:
364		V_ip6stat.ip6s_nogif++;
365		m_freem(m);
366		return IPPROTO_DONE;
367	}
368
369	gif_input(m, af, gifp);
370	return IPPROTO_DONE;
371}
372
373/*
374 * validate outer address.
375 */
376static int
377gif_validate6(const struct ip6_hdr *ip6, struct gif_softc *sc,
378    struct ifnet *ifp)
379{
380	struct sockaddr_in6 *src, *dst;
381
382	src = (struct sockaddr_in6 *)sc->gif_psrc;
383	dst = (struct sockaddr_in6 *)sc->gif_pdst;
384
385	/*
386	 * Check for address match.  Note that the check is for an incoming
387	 * packet.  We should compare the *source* address in our configuration
388	 * and the *destination* address of the packet, and vice versa.
389	 */
390	if (!IN6_ARE_ADDR_EQUAL(&src->sin6_addr, &ip6->ip6_dst) ||
391	    !IN6_ARE_ADDR_EQUAL(&dst->sin6_addr, &ip6->ip6_src))
392		return 0;
393
394	/* martian filters on outer source - done in ip6_input */
395
396	/* ingress filters on outer source */
397	if ((GIF2IFP(sc)->if_flags & IFF_LINK2) == 0 && ifp) {
398		struct sockaddr_in6 sin6;
399		struct rtentry *rt;
400
401		bzero(&sin6, sizeof(sin6));
402		sin6.sin6_family = AF_INET6;
403		sin6.sin6_len = sizeof(struct sockaddr_in6);
404		sin6.sin6_addr = ip6->ip6_src;
405		sin6.sin6_scope_id = 0; /* XXX */
406
407		rt = in6_rtalloc1((struct sockaddr *)&sin6, 0, 0UL,
408		    sc->gif_fibnum);
409		if (!rt || rt->rt_ifp != ifp) {
410#if 0
411			char ip6buf[INET6_ADDRSTRLEN];
412			log(LOG_WARNING, "%s: packet from %s dropped "
413			    "due to ingress filter\n", if_name(GIF2IFP(sc)),
414			    ip6_sprintf(ip6buf, &sin6.sin6_addr));
415#endif
416			if (rt)
417				RTFREE_LOCKED(rt);
418			return 0;
419		}
420		RTFREE_LOCKED(rt);
421	}
422
423	return 128 * 2;
424}
425
426/*
427 * we know that we are in IFF_UP, outer address available, and outer family
428 * matched the physical addr family.  see gif_encapcheck().
429 * sanity check for arg should have been done in the caller.
430 */
431int
432gif_encapcheck6(const struct mbuf *m, int off, int proto, void *arg)
433{
434	struct ip6_hdr ip6;
435	struct gif_softc *sc;
436	struct ifnet *ifp;
437
438	/* sanity check done in caller */
439	sc = (struct gif_softc *)arg;
440
441	/* LINTED const cast */
442	m_copydata(m, 0, sizeof(ip6), (caddr_t)&ip6);
443	ifp = ((m->m_flags & M_PKTHDR) != 0) ? m->m_pkthdr.rcvif : NULL;
444
445	return gif_validate6(&ip6, sc, ifp);
446}
447
448int
449in6_gif_attach(struct gif_softc *sc)
450{
451	sc->encap_cookie6 = encap_attach_func(AF_INET6, -1, gif_encapcheck,
452	    (void *)&in6_gif_protosw, sc);
453	if (sc->encap_cookie6 == NULL)
454		return EEXIST;
455	return 0;
456}
457
458int
459in6_gif_detach(struct gif_softc *sc)
460{
461	int error;
462
463	error = encap_detach(sc->encap_cookie6);
464	if (error == 0)
465		sc->encap_cookie6 = NULL;
466	return error;
467}
468