2 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
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.
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
29 * $KAME: ip6_forward.c,v 1.69 2001/05/17 03:48:30 itojun Exp $
32 #include <sys/cdefs.h>
33 __FBSDID("$FreeBSD$");
36 #include "opt_inet6.h"
37 #include "opt_ipsec.h"
38 #include "opt_ipstealth.h"
40 #include <sys/param.h>
41 #include <sys/systm.h>
42 #include <sys/malloc.h>
44 #include <sys/domain.h>
45 #include <sys/protosw.h>
46 #include <sys/socket.h>
47 #include <sys/errno.h>
49 #include <sys/kernel.h>
50 #include <sys/syslog.h>
53 #include <net/if_var.h>
54 #include <net/netisr.h>
55 #include <net/route.h>
58 #include <netinet/in.h>
59 #include <netinet/in_var.h>
60 #include <netinet/in_systm.h>
61 #include <netinet/ip.h>
62 #include <netinet/ip_var.h>
63 #include <netinet6/in6_var.h>
64 #include <netinet/ip6.h>
65 #include <netinet6/ip6_var.h>
66 #include <netinet6/scope6_var.h>
67 #include <netinet/icmp6.h>
68 #include <netinet6/nd6.h>
70 #include <netinet/in_pcb.h>
72 #include <netipsec/ipsec_support.h>
75 * Forward a packet. If some error occurs return the sender
76 * an icmp packet. Note we can't always generate a meaningful
77 * icmp message because icmp doesn't have a large enough repertoire
80 * If not forwarding, just drop the packet. This could be confusing
81 * if ipforwarding was zero but some routing protocol was advancing
82 * us as a gateway to somewhere. However, we must let the routing
83 * protocol deal with that.
87 ip6_forward(struct mbuf *m, int srcrt)
89 struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *);
90 struct sockaddr_in6 *dst = NULL;
91 struct rtentry *rt = NULL;
92 struct route_in6 rin6;
93 int error, type = 0, code = 0;
94 struct mbuf *mcopy = NULL;
95 struct ifnet *origifp; /* maybe unnecessary */
96 u_int32_t inzone, outzone;
97 struct in6_addr src_in6, dst_in6, odst;
98 struct m_tag *fwd_tag;
99 char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN];
102 * Do not forward packets to multicast destination (should be handled
104 * Do not forward packets with unspecified source. It was discussed
105 * in July 2000, on the ipngwg mailing list.
107 if ((m->m_flags & (M_BCAST|M_MCAST)) != 0 ||
108 IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) ||
109 IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_src)) {
110 IP6STAT_INC(ip6s_cantforward);
111 /* XXX in6_ifstat_inc(rt->rt_ifp, ifs6_in_discard) */
112 if (V_ip6_log_time + V_ip6_log_interval < time_uptime) {
113 V_ip6_log_time = time_uptime;
116 "from %s to %s nxt %d received on %s\n",
117 ip6_sprintf(ip6bufs, &ip6->ip6_src),
118 ip6_sprintf(ip6bufd, &ip6->ip6_dst),
120 if_name(m->m_pkthdr.rcvif));
130 ip6->ip6_hlim <= IPV6_HLIMDEC) {
131 /* XXX in6_ifstat_inc(rt->rt_ifp, ifs6_in_discard) */
132 icmp6_error(m, ICMP6_TIME_EXCEEDED,
133 ICMP6_TIME_EXCEED_TRANSIT, 0);
138 * Save at most ICMPV6_PLD_MAXLEN (= the min IPv6 MTU -
139 * size of IPv6 + ICMPv6 headers) bytes of the packet in case
140 * we need to generate an ICMP6 message to the src.
141 * Thanks to M_EXT, in most cases copy will not occur.
143 * It is important to save it before IPsec processing as IPsec
144 * processing may modify the mbuf.
146 mcopy = m_copym(m, 0, imin(m->m_pkthdr.len, ICMPV6_PLD_MAXLEN),
149 if (V_ip6stealth == 0)
151 ip6->ip6_hlim -= IPV6_HLIMDEC;
153 #if defined(IPSEC) || defined(IPSEC_SUPPORT)
154 if (IPSEC_ENABLED(ipv6)) {
155 if ((error = IPSEC_FORWARD(ipv6, m)) != 0) {
156 /* mbuf consumed by IPsec */
158 if (error != EINPROGRESS)
159 IP6STAT_INC(ip6s_cantforward);
162 /* No IPsec processing required */
166 bzero(&rin6, sizeof(struct route_in6));
167 dst = (struct sockaddr_in6 *)&rin6.ro_dst;
168 dst->sin6_len = sizeof(struct sockaddr_in6);
169 dst->sin6_family = AF_INET6;
170 dst->sin6_addr = ip6->ip6_dst;
172 rin6.ro_rt = in6_rtalloc1((struct sockaddr *)dst, 0, 0, M_GETFIB(m));
174 if (rin6.ro_rt != NULL)
175 RT_UNLOCK(rin6.ro_rt);
177 IP6STAT_INC(ip6s_noroute);
178 in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_noroute);
180 icmp6_error(mcopy, ICMP6_DST_UNREACH,
181 ICMP6_DST_UNREACH_NOROUTE, 0);
187 * Source scope check: if a packet can't be delivered to its
188 * destination for the reason that the destination is beyond the scope
189 * of the source address, discard the packet and return an icmp6
190 * destination unreachable error with Code 2 (beyond scope of source
191 * address). We use a local copy of ip6_src, since in6_setscope()
192 * will possibly modify its first argument.
193 * [draft-ietf-ipngwg-icmp-v3-04.txt, Section 3.1]
195 src_in6 = ip6->ip6_src;
196 if (in6_setscope(&src_in6, rt->rt_ifp, &outzone)) {
197 /* XXX: this should not happen */
198 IP6STAT_INC(ip6s_cantforward);
199 IP6STAT_INC(ip6s_badscope);
202 if (in6_setscope(&src_in6, m->m_pkthdr.rcvif, &inzone)) {
203 IP6STAT_INC(ip6s_cantforward);
204 IP6STAT_INC(ip6s_badscope);
207 if (inzone != outzone) {
208 IP6STAT_INC(ip6s_cantforward);
209 IP6STAT_INC(ip6s_badscope);
210 in6_ifstat_inc(rt->rt_ifp, ifs6_in_discard);
212 if (V_ip6_log_time + V_ip6_log_interval < time_uptime) {
213 V_ip6_log_time = time_uptime;
216 "src %s, dst %s, nxt %d, rcvif %s, outif %s\n",
217 ip6_sprintf(ip6bufs, &ip6->ip6_src),
218 ip6_sprintf(ip6bufd, &ip6->ip6_dst),
220 if_name(m->m_pkthdr.rcvif), if_name(rt->rt_ifp));
223 icmp6_error(mcopy, ICMP6_DST_UNREACH,
224 ICMP6_DST_UNREACH_BEYONDSCOPE, 0);
229 * Destination scope check: if a packet is going to break the scope
230 * zone of packet's destination address, discard it. This case should
231 * usually be prevented by appropriately-configured routing table, but
232 * we need an explicit check because we may mistakenly forward the
233 * packet to a different zone by (e.g.) a default route.
235 dst_in6 = ip6->ip6_dst;
236 if (in6_setscope(&dst_in6, m->m_pkthdr.rcvif, &inzone) != 0 ||
237 in6_setscope(&dst_in6, rt->rt_ifp, &outzone) != 0 ||
239 IP6STAT_INC(ip6s_cantforward);
240 IP6STAT_INC(ip6s_badscope);
244 if (rt->rt_flags & RTF_GATEWAY)
245 dst = (struct sockaddr_in6 *)rt->rt_gateway;
248 * If we are to forward the packet using the same interface
249 * as one we got the packet from, perhaps we should send a redirect
250 * to sender to shortcut a hop.
251 * Only send redirect if source is sending directly to us,
252 * and if packet was not source routed (or has any options).
253 * Also, don't send redirect if forwarding using a route
254 * modified by a redirect.
256 if (V_ip6_sendredirects && rt->rt_ifp == m->m_pkthdr.rcvif && !srcrt &&
257 (rt->rt_flags & (RTF_DYNAMIC|RTF_MODIFIED)) == 0) {
258 if ((rt->rt_ifp->if_flags & IFF_POINTOPOINT) != 0) {
260 * If the incoming interface is equal to the outgoing
261 * one, and the link attached to the interface is
262 * point-to-point, then it will be highly probable
263 * that a routing loop occurs. Thus, we immediately
264 * drop the packet and send an ICMPv6 error message.
266 * type/code is based on suggestion by Rich Draves.
267 * not sure if it is the best pick.
269 icmp6_error(mcopy, ICMP6_DST_UNREACH,
270 ICMP6_DST_UNREACH_ADDR, 0);
277 * Fake scoped addresses. Note that even link-local source or
278 * destinaion can appear, if the originating node just sends the
279 * packet to us (without address resolution for the destination).
280 * Since both icmp6_error and icmp6_redirect_output fill the embedded
281 * link identifiers, we can do this stuff after making a copy for
282 * returning an error.
284 if ((rt->rt_ifp->if_flags & IFF_LOOPBACK) != 0) {
286 * See corresponding comments in ip6_output.
287 * XXX: but is it possible that ip6_forward() sends a packet
288 * to a loopback interface? I don't think so, and thus
289 * I bark here. (jinmei@kame.net)
290 * XXX: it is common to route invalid packets to loopback.
291 * also, the codepath will be visited on use of ::1 in
297 if ((rt->rt_flags & (RTF_BLACKHOLE|RTF_REJECT)) == 0)
300 printf("ip6_forward: outgoing interface is loopback. "
301 "src %s, dst %s, nxt %d, rcvif %s, outif %s\n",
302 ip6_sprintf(ip6bufs, &ip6->ip6_src),
303 ip6_sprintf(ip6bufd, &ip6->ip6_dst),
304 ip6->ip6_nxt, if_name(m->m_pkthdr.rcvif),
305 if_name(rt->rt_ifp));
308 /* we can just use rcvif in forwarding. */
309 origifp = m->m_pkthdr.rcvif;
312 origifp = rt->rt_ifp;
314 * clear embedded scope identifiers if necessary.
315 * in6_clearscope will touch the addresses only when necessary.
317 in6_clearscope(&ip6->ip6_src);
318 in6_clearscope(&ip6->ip6_dst);
320 /* Jump over all PFIL processing if hooks are not active. */
321 if (!PFIL_HOOKED(&V_inet6_pfil_hook))
325 /* Run through list of hooks for output packets. */
326 error = pfil_run_hooks(&V_inet6_pfil_hook, &m, rt->rt_ifp, PFIL_OUT, NULL);
327 if (error != 0 || m == NULL)
328 goto freecopy; /* consumed by filter */
329 ip6 = mtod(m, struct ip6_hdr *);
331 /* See if destination IP address was changed by packet filter. */
332 if (!IN6_ARE_ADDR_EQUAL(&odst, &ip6->ip6_dst)) {
333 m->m_flags |= M_SKIP_FIREWALL;
334 /* If destination is now ourself drop to ip6_input(). */
335 if (in6_localip(&ip6->ip6_dst))
336 m->m_flags |= M_FASTFWD_OURS;
339 goto again; /* Redo the routing table lookup. */
343 /* See if local, if yes, send it to netisr. */
344 if (m->m_flags & M_FASTFWD_OURS) {
345 if (m->m_pkthdr.rcvif == NULL)
346 m->m_pkthdr.rcvif = V_loif;
347 if (m->m_pkthdr.csum_flags & CSUM_DELAY_DATA_IPV6) {
348 m->m_pkthdr.csum_flags |=
349 CSUM_DATA_VALID_IPV6 | CSUM_PSEUDO_HDR;
350 m->m_pkthdr.csum_data = 0xffff;
353 if (m->m_pkthdr.csum_flags & CSUM_SCTP_IPV6)
354 m->m_pkthdr.csum_flags |= CSUM_SCTP_VALID;
356 error = netisr_queue(NETISR_IPV6, m);
359 /* Or forward to some other address? */
360 if ((m->m_flags & M_IP6_NEXTHOP) &&
361 (fwd_tag = m_tag_find(m, PACKET_TAG_IPFORWARD, NULL)) != NULL) {
362 dst = (struct sockaddr_in6 *)&rin6.ro_dst;
363 bcopy((fwd_tag+1), dst, sizeof(struct sockaddr_in6));
364 m->m_flags |= M_SKIP_FIREWALL;
365 m->m_flags &= ~M_IP6_NEXTHOP;
366 m_tag_delete(m, fwd_tag);
372 /* See if the size was changed by the packet filter. */
373 if (m->m_pkthdr.len > IN6_LINKMTU(rt->rt_ifp)) {
374 in6_ifstat_inc(rt->rt_ifp, ifs6_in_toobig);
376 icmp6_error(mcopy, ICMP6_PACKET_TOO_BIG, 0,
377 IN6_LINKMTU(rt->rt_ifp));
381 error = nd6_output_ifp(rt->rt_ifp, origifp, m, dst, NULL);
383 in6_ifstat_inc(rt->rt_ifp, ifs6_out_discard);
384 IP6STAT_INC(ip6s_cantforward);
386 IP6STAT_INC(ip6s_forward);
387 in6_ifstat_inc(rt->rt_ifp, ifs6_out_forward);
389 IP6STAT_INC(ip6s_redirectsent);
400 if (type == ND_REDIRECT) {
401 icmp6_redirect_output(mcopy, rt);
407 /* xxx MTU is constant in PPP? */
411 /* Tell source to slow down like source quench in IP? */
414 case ENETUNREACH: /* shouldn't happen, checked above */
419 type = ICMP6_DST_UNREACH;
420 code = ICMP6_DST_UNREACH_ADDR;
423 icmp6_error(mcopy, type, code, 0);