2 * SPDX-License-Identifier: BSD-3-Clause
4 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
5 * Copyright (c) 2018 Andrey V. Elsukov <ae@FreeBSD.org>
8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions
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.
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
32 * $KAME: in6_gif.c,v 1.49 2001/05/14 14:02:17 itojun Exp $
35 #include <sys/cdefs.h>
36 __FBSDID("$FreeBSD$");
39 #include "opt_inet6.h"
41 #include <sys/param.h>
42 #include <sys/systm.h>
44 #include <sys/socket.h>
45 #include <sys/sockio.h>
47 #include <sys/errno.h>
48 #include <sys/kernel.h>
49 #include <sys/syslog.h>
50 #include <sys/sysctl.h>
51 #include <sys/malloc.h>
53 #include <net/ethernet.h>
55 #include <net/if_var.h>
56 #include <net/route.h>
59 #include <netinet/in.h>
60 #include <netinet/in_systm.h>
62 #include <netinet/ip.h>
63 #include <netinet/ip_ecn.h>
65 #include <netinet/ip_encap.h>
66 #include <netinet/ip6.h>
67 #include <netinet6/ip6_var.h>
68 #include <netinet6/in6_var.h>
69 #include <netinet6/scope6_var.h>
70 #include <netinet6/ip6_ecn.h>
71 #include <netinet6/in6_fib.h>
73 #include <net/if_gif.h>
76 static VNET_DEFINE(int, ip6_gif_hlim) = GIF_HLIM;
77 #define V_ip6_gif_hlim VNET(ip6_gif_hlim)
79 SYSCTL_DECL(_net_inet6_ip6);
80 SYSCTL_INT(_net_inet6_ip6, IPV6CTL_GIF_HLIM, gifhlim,
81 CTLFLAG_VNET | CTLFLAG_RW, &VNET_NAME(ip6_gif_hlim), 0,
82 "Default hop limit for encapsulated packets");
85 * We keep interfaces in a hash table using src+dst as key.
86 * Interfaces with GIF_IGNORE_SOURCE flag are linked into plain list.
88 static VNET_DEFINE(struct gif_list *, ipv6_hashtbl) = NULL;
89 static VNET_DEFINE(struct gif_list, ipv6_list) = CK_LIST_HEAD_INITIALIZER();
90 #define V_ipv6_hashtbl VNET(ipv6_hashtbl)
91 #define V_ipv6_list VNET(ipv6_list)
93 #define GIF_HASH(src, dst) (V_ipv6_hashtbl[\
94 in6_gif_hashval((src), (dst)) & (GIF_HASH_SIZE - 1)])
95 #define GIF_HASH_SC(sc) GIF_HASH(&(sc)->gif_ip6hdr->ip6_src,\
96 &(sc)->gif_ip6hdr->ip6_dst)
98 in6_gif_hashval(const struct in6_addr *src, const struct in6_addr *dst)
102 ret = fnv_32_buf(src, sizeof(*src), FNV1_32_INIT);
103 return (fnv_32_buf(dst, sizeof(*dst), ret));
107 in6_gif_checkdup(const struct gif_softc *sc, const struct in6_addr *src,
108 const struct in6_addr *dst)
110 struct gif_softc *tmp;
112 if (sc->gif_family == AF_INET6 &&
113 IN6_ARE_ADDR_EQUAL(&sc->gif_ip6hdr->ip6_src, src) &&
114 IN6_ARE_ADDR_EQUAL(&sc->gif_ip6hdr->ip6_dst, dst))
117 CK_LIST_FOREACH(tmp, &GIF_HASH(src, dst), chain) {
120 if (IN6_ARE_ADDR_EQUAL(&tmp->gif_ip6hdr->ip6_src, src) &&
121 IN6_ARE_ADDR_EQUAL(&tmp->gif_ip6hdr->ip6_dst, dst))
122 return (EADDRNOTAVAIL);
128 in6_gif_attach(struct gif_softc *sc)
131 if (sc->gif_options & GIF_IGNORE_SOURCE)
132 CK_LIST_INSERT_HEAD(&V_ipv6_list, sc, chain);
134 CK_LIST_INSERT_HEAD(&GIF_HASH_SC(sc), sc, chain);
138 in6_gif_setopts(struct gif_softc *sc, u_int options)
141 /* NOTE: we are protected with gif_ioctl_sx lock */
142 MPASS(sc->gif_family == AF_INET6);
143 MPASS(sc->gif_options != options);
145 if ((options & GIF_IGNORE_SOURCE) !=
146 (sc->gif_options & GIF_IGNORE_SOURCE)) {
147 CK_LIST_REMOVE(sc, chain);
148 sc->gif_options = options;
155 in6_gif_ioctl(struct gif_softc *sc, u_long cmd, caddr_t data)
157 struct in6_ifreq *ifr = (struct in6_ifreq *)data;
158 struct sockaddr_in6 *dst, *src;
162 /* NOTE: we are protected with gif_ioctl_sx lock */
165 case SIOCSIFPHYADDR_IN6:
166 src = &((struct in6_aliasreq *)data)->ifra_addr;
167 dst = &((struct in6_aliasreq *)data)->ifra_dstaddr;
170 if (src->sin6_family != dst->sin6_family ||
171 src->sin6_family != AF_INET6 ||
172 src->sin6_len != dst->sin6_len ||
173 src->sin6_len != sizeof(*src))
175 if (IN6_IS_ADDR_UNSPECIFIED(&src->sin6_addr) ||
176 IN6_IS_ADDR_UNSPECIFIED(&dst->sin6_addr)) {
177 error = EADDRNOTAVAIL;
181 * Check validity of the scope zone ID of the
182 * addresses, and convert it into the kernel
183 * internal form if necessary.
185 if ((error = sa6_embedscope(src, 0)) != 0 ||
186 (error = sa6_embedscope(dst, 0)) != 0)
189 if (V_ipv6_hashtbl == NULL)
190 V_ipv6_hashtbl = gif_hashinit();
191 error = in6_gif_checkdup(sc, &src->sin6_addr,
193 if (error == EADDRNOTAVAIL)
195 if (error == EEXIST) {
196 /* Addresses are the same. Just return. */
200 ip6 = malloc(sizeof(*ip6), M_GIF, M_WAITOK | M_ZERO);
201 ip6->ip6_src = src->sin6_addr;
202 ip6->ip6_dst = dst->sin6_addr;
203 if (sc->gif_family != 0) {
204 /* Detach existing tunnel first */
205 CK_LIST_REMOVE(sc, chain);
207 free(sc->gif_hdr, M_GIF);
208 /* XXX: should we notify about link state change? */
210 sc->gif_family = AF_INET6;
211 sc->gif_ip6hdr = ip6;
214 case SIOCGIFPSRCADDR_IN6:
215 case SIOCGIFPDSTADDR_IN6:
216 if (sc->gif_family != AF_INET6) {
217 error = EADDRNOTAVAIL;
220 src = (struct sockaddr_in6 *)&ifr->ifr_addr;
221 memset(src, 0, sizeof(*src));
222 src->sin6_family = AF_INET6;
223 src->sin6_len = sizeof(*src);
224 src->sin6_addr = (cmd == SIOCGIFPSRCADDR_IN6) ?
225 sc->gif_ip6hdr->ip6_src: sc->gif_ip6hdr->ip6_dst;
226 error = prison_if(curthread->td_ucred, (struct sockaddr *)src);
228 error = sa6_recoverscope(src);
230 memset(src, 0, sizeof(*src));
237 in6_gif_output(struct ifnet *ifp, struct mbuf *m, int proto, uint8_t ecn)
239 struct gif_softc *sc = ifp->if_softc;
243 /* prepend new IP header */
245 len = sizeof(struct ip6_hdr);
246 #ifndef __NO_STRICT_ALIGNMENT
247 if (proto == IPPROTO_ETHERIP)
248 len += ETHERIP_ALIGN;
250 M_PREPEND(m, len, M_NOWAIT);
253 #ifndef __NO_STRICT_ALIGNMENT
254 if (proto == IPPROTO_ETHERIP) {
255 len = mtod(m, vm_offset_t) & 3;
256 KASSERT(len == 0 || len == ETHERIP_ALIGN,
257 ("in6_gif_output: unexpected misalignment"));
259 m->m_len -= ETHERIP_ALIGN;
263 ip6 = mtod(m, struct ip6_hdr *);
264 MPASS(sc->gif_family == AF_INET6);
265 bcopy(sc->gif_ip6hdr, ip6, sizeof(struct ip6_hdr));
267 ip6->ip6_flow |= htonl((uint32_t)ecn << 20);
268 ip6->ip6_nxt = proto;
269 ip6->ip6_hlim = V_ip6_gif_hlim;
271 * force fragmentation to minimum MTU, to avoid path MTU discovery.
272 * it is too painful to ask for resend of inner packet, to achieve
273 * path MTU discovery for encapsulated packets.
275 return (ip6_output(m, 0, NULL, IPV6_MINMTU, 0, NULL, NULL));
279 in6_gif_input(struct mbuf *m, int off, int proto, void *arg)
281 struct gif_softc *sc = arg;
289 IP6STAT_INC(ip6s_nogif);
290 return (IPPROTO_DONE);
293 if ((gifp->if_flags & IFF_UP) != 0) {
294 ip6 = mtod(m, struct ip6_hdr *);
295 ecn = (ntohl(ip6->ip6_flow) >> 20) & 0xff;
297 gif_input(m, gifp, proto, ecn);
300 IP6STAT_INC(ip6s_nogif);
302 return (IPPROTO_DONE);
306 in6_gif_lookup(const struct mbuf *m, int off, int proto, void **arg)
308 const struct ip6_hdr *ip6;
309 struct gif_softc *sc;
314 * NOTE: it is safe to iterate without any locking here, because softc
315 * can be reclaimed only when we are not within net_epoch_preempt
316 * section, but ip_encap lookup+input are executed in epoch section.
318 ip6 = mtod(m, const struct ip6_hdr *);
320 CK_LIST_FOREACH(sc, &GIF_HASH(&ip6->ip6_dst, &ip6->ip6_src), chain) {
322 * This is an inbound packet, its ip6_dst is source address
325 if (IN6_ARE_ADDR_EQUAL(&sc->gif_ip6hdr->ip6_src,
327 IN6_ARE_ADDR_EQUAL(&sc->gif_ip6hdr->ip6_dst,
329 ret = ENCAP_DRV_LOOKUP;
335 * Check the list of interfaces with GIF_IGNORE_SOURCE flag.
337 CK_LIST_FOREACH(sc, &V_ipv6_list, chain) {
338 if (IN6_ARE_ADDR_EQUAL(&sc->gif_ip6hdr->ip6_src,
340 ret = 128 + 8; /* src + proto */
346 if ((GIF2IFP(sc)->if_flags & IFF_UP) == 0)
348 /* ingress filters on outer source */
349 if ((GIF2IFP(sc)->if_flags & IFF_LINK2) == 0) {
350 struct nhop6_basic nh6;
352 if (fib6_lookup_nh_basic(sc->gif_fibnum, &ip6->ip6_src,
353 ntohs(in6_getscope(&ip6->ip6_src)), 0, 0, &nh6) != 0)
356 if (nh6.nh_ifp != m->m_pkthdr.rcvif)
364 const struct encap_config encap;
365 const struct encaptab *cookie;
366 } ipv6_encap_cfg[] = {
370 .proto = IPPROTO_IPV4,
371 .min_length = sizeof(struct ip6_hdr) +
373 .exact_match = ENCAP_DRV_LOOKUP,
374 .lookup = in6_gif_lookup,
375 .input = in6_gif_input
381 .proto = IPPROTO_IPV6,
382 .min_length = 2 * sizeof(struct ip6_hdr),
383 .exact_match = ENCAP_DRV_LOOKUP,
384 .lookup = in6_gif_lookup,
385 .input = in6_gif_input
390 .proto = IPPROTO_ETHERIP,
391 .min_length = sizeof(struct ip6_hdr) +
392 sizeof(struct etherip_header) +
393 sizeof(struct ether_header),
394 .exact_match = ENCAP_DRV_LOOKUP,
395 .lookup = in6_gif_lookup,
396 .input = in6_gif_input
406 if (!IS_DEFAULT_VNET(curvnet))
408 for (i = 0; i < nitems(ipv6_encap_cfg); i++)
409 ipv6_encap_cfg[i].cookie = ip6_encap_attach(
410 &ipv6_encap_cfg[i].encap, NULL, M_WAITOK);
418 if (IS_DEFAULT_VNET(curvnet)) {
419 for (i = 0; i < nitems(ipv6_encap_cfg); i++)
420 ip6_encap_detach(ipv6_encap_cfg[i].cookie);
422 if (V_ipv6_hashtbl != NULL)
423 gif_hashdestroy(V_ipv6_hashtbl);