1 /* $KAME: in_gif.c,v 1.54 2001/05/14 14:02:16 itojun Exp $ */ 2 3 /*- 4 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. 5 * All rights reserved. 6 * 7 * Redistribution and use in source and binary forms, with or without 8 * modification, are permitted provided that the following conditions 9 * are met: 10 * 1. Redistributions of source code must retain the above copyright 11 * notice, this list of conditions and the following disclaimer. 12 * 2. Redistributions in binary form must reproduce the above copyright 13 * notice, this list of conditions and the following disclaimer in the 14 * documentation and/or other materials provided with the distribution. 15 * 3. Neither the name of the project nor the names of its contributors 16 * may be used to endorse or promote products derived from this software 17 * without specific prior written permission. 18 * 19 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND 20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 22 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE 23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 29 * SUCH DAMAGE. 30 */ 31 32 #include <sys/cdefs.h> 33 __FBSDID("$FreeBSD$"); 34 35 #include "opt_mrouting.h" 36 #include "opt_inet.h" 37 #include "opt_inet6.h" 38 39 #include <sys/param.h> 40 #include <sys/systm.h> 41 #include <sys/socket.h> 42 #include <sys/sockio.h> 43 #include <sys/mbuf.h> 44 #include <sys/errno.h> 45 #include <sys/kernel.h> 46 #include <sys/sysctl.h> 47 #include <sys/protosw.h> 48 #include <sys/malloc.h> 49 #include <sys/vimage.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 #include <netinet/ip.h> 57 #include <netinet/ip_var.h> 58 #include <netinet/in_gif.h> 59 #include <netinet/in_var.h> 60 #include <netinet/ip_encap.h> 61 #include <netinet/ip_ecn.h> 62 #include <netinet/vinet.h> 63 64 #ifdef INET6 65 #include <netinet/ip6.h> 66 #endif 67 68 #ifdef MROUTING 69 #include <netinet/ip_mroute.h> 70 #endif /* MROUTING */ 71 72 #include <net/if_gif.h> 73 74 static int gif_validate4(const struct ip *, struct gif_softc *, 75 struct ifnet *); 76 77 extern struct domain inetdomain; 78 struct protosw in_gif_protosw = { 79 .pr_type = SOCK_RAW, 80 .pr_domain = &inetdomain, 81 .pr_protocol = 0/* IPPROTO_IPV[46] */, 82 .pr_flags = PR_ATOMIC|PR_ADDR, 83 .pr_input = in_gif_input, 84 .pr_output = (pr_output_t*)rip_output, 85 .pr_ctloutput = rip_ctloutput, 86 .pr_usrreqs = &rip_usrreqs 87 }; 88 89 #ifdef VIMAGE_GLOBALS 90 extern int ip_gif_ttl; 91 #endif 92 SYSCTL_V_INT(V_NET, vnet_gif, _net_inet_ip, IPCTL_GIF_TTL, gifttl, 93 CTLFLAG_RW, ip_gif_ttl, 0, ""); 94 95 int 96 in_gif_output(struct ifnet *ifp, int family, struct mbuf *m) 97 { 98 INIT_VNET_GIF(ifp->if_vnet); 99 struct gif_softc *sc = ifp->if_softc; 100 struct sockaddr_in *dst = (struct sockaddr_in *)&sc->gif_ro.ro_dst; 101 struct sockaddr_in *sin_src = (struct sockaddr_in *)sc->gif_psrc; 102 struct sockaddr_in *sin_dst = (struct sockaddr_in *)sc->gif_pdst; 103 struct ip iphdr; /* capsule IP header, host byte ordered */ 104 struct etherip_header eiphdr; 105 int error, len, proto; 106 u_int8_t tos; 107 108 GIF_LOCK_ASSERT(sc); 109 110 if (sin_src == NULL || sin_dst == NULL || 111 sin_src->sin_family != AF_INET || 112 sin_dst->sin_family != AF_INET) { 113 m_freem(m); 114 return EAFNOSUPPORT; 115 } 116 117 switch (family) { 118 #ifdef INET 119 case AF_INET: 120 { 121 struct ip *ip; 122 123 proto = IPPROTO_IPV4; 124 if (m->m_len < sizeof(*ip)) { 125 m = m_pullup(m, sizeof(*ip)); 126 if (!m) 127 return ENOBUFS; 128 } 129 ip = mtod(m, struct ip *); 130 tos = ip->ip_tos; 131 break; 132 } 133 #endif /* INET */ 134 #ifdef INET6 135 case AF_INET6: 136 { 137 struct ip6_hdr *ip6; 138 proto = IPPROTO_IPV6; 139 if (m->m_len < sizeof(*ip6)) { 140 m = m_pullup(m, sizeof(*ip6)); 141 if (!m) 142 return ENOBUFS; 143 } 144 ip6 = mtod(m, struct ip6_hdr *); 145 tos = (ntohl(ip6->ip6_flow) >> 20) & 0xff; 146 break; 147 } 148 #endif /* INET6 */ 149 case AF_LINK: 150 proto = IPPROTO_ETHERIP; 151 eiphdr.eip_ver = ETHERIP_VERSION & ETHERIP_VER_VERS_MASK; 152 eiphdr.eip_pad = 0; 153 /* prepend Ethernet-in-IP header */ 154 M_PREPEND(m, sizeof(struct etherip_header), M_DONTWAIT); 155 if (m && m->m_len < sizeof(struct etherip_header)) 156 m = m_pullup(m, sizeof(struct etherip_header)); 157 if (m == NULL) 158 return ENOBUFS; 159 bcopy(&eiphdr, mtod(m, struct etherip_header *), 160 sizeof(struct etherip_header)); 161 break; 162 163 default: 164 #ifdef DEBUG 165 printf("in_gif_output: warning: unknown family %d passed\n", 166 family); 167 #endif 168 m_freem(m); 169 return EAFNOSUPPORT; 170 } 171 172 bzero(&iphdr, sizeof(iphdr)); 173 iphdr.ip_src = sin_src->sin_addr; 174 /* bidirectional configured tunnel mode */ 175 if (sin_dst->sin_addr.s_addr != INADDR_ANY) 176 iphdr.ip_dst = sin_dst->sin_addr; 177 else { 178 m_freem(m); 179 return ENETUNREACH; 180 } 181 iphdr.ip_p = proto; 182 /* version will be set in ip_output() */ 183 iphdr.ip_ttl = V_ip_gif_ttl; 184 iphdr.ip_len = m->m_pkthdr.len + sizeof(struct ip); 185 ip_ecn_ingress((ifp->if_flags & IFF_LINK1) ? ECN_ALLOWED : ECN_NOCARE, 186 &iphdr.ip_tos, &tos); 187 188 /* prepend new IP header */ 189 len = sizeof(struct ip); 190 #ifndef __NO_STRICT_ALIGNMENT 191 if (family == AF_LINK) 192 len += ETHERIP_ALIGN; 193 #endif 194 M_PREPEND(m, len, M_DONTWAIT); 195 if (m != NULL && m->m_len < len) 196 m = m_pullup(m, len); 197 if (m == NULL) { 198 printf("ENOBUFS in in_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 ("in_gif_output: unexpected misalignment")); 206 m->m_data += len; 207 m->m_len -= ETHERIP_ALIGN; 208 } 209 #endif 210 bcopy(&iphdr, mtod(m, struct ip *), sizeof(struct ip)); 211 212 M_SETFIB(m, sc->gif_fibnum); 213 214 if (dst->sin_family != sin_dst->sin_family || 215 dst->sin_addr.s_addr != sin_dst->sin_addr.s_addr) { 216 /* cache route doesn't match */ 217 bzero(dst, sizeof(*dst)); 218 dst->sin_family = sin_dst->sin_family; 219 dst->sin_len = sizeof(struct sockaddr_in); 220 dst->sin_addr = sin_dst->sin_addr; 221 if (sc->gif_ro.ro_rt) { 222 RTFREE(sc->gif_ro.ro_rt); 223 sc->gif_ro.ro_rt = NULL; 224 } 225 #if 0 226 GIF2IFP(sc)->if_mtu = GIF_MTU; 227 #endif 228 } 229 230 if (sc->gif_ro.ro_rt == NULL) { 231 in_rtalloc_ign(&sc->gif_ro, 0, sc->gif_fibnum); 232 if (sc->gif_ro.ro_rt == NULL) { 233 m_freem(m); 234 return ENETUNREACH; 235 } 236 237 /* if it constitutes infinite encapsulation, punt. */ 238 if (sc->gif_ro.ro_rt->rt_ifp == ifp) { 239 m_freem(m); 240 return ENETUNREACH; /* XXX */ 241 } 242 #if 0 243 ifp->if_mtu = sc->gif_ro.ro_rt->rt_ifp->if_mtu 244 - sizeof(struct ip); 245 #endif 246 } 247 248 error = ip_output(m, NULL, &sc->gif_ro, 0, NULL, NULL); 249 250 if (!(GIF2IFP(sc)->if_flags & IFF_LINK0) && 251 sc->gif_ro.ro_rt != NULL) { 252 RTFREE(sc->gif_ro.ro_rt); 253 sc->gif_ro.ro_rt = NULL; 254 } 255 256 return (error); 257 } 258 259 void 260 in_gif_input(struct mbuf *m, int off) 261 { 262 INIT_VNET_INET(curvnet); 263 struct ifnet *gifp = NULL; 264 struct gif_softc *sc; 265 struct ip *ip; 266 int af; 267 u_int8_t otos; 268 int proto; 269 270 ip = mtod(m, struct ip *); 271 proto = ip->ip_p; 272 273 sc = (struct gif_softc *)encap_getarg(m); 274 if (sc == NULL) { 275 m_freem(m); 276 IPSTAT_INC(ips_nogif); 277 return; 278 } 279 280 gifp = GIF2IFP(sc); 281 if (gifp == NULL || (gifp->if_flags & IFF_UP) == 0) { 282 m_freem(m); 283 IPSTAT_INC(ips_nogif); 284 return; 285 } 286 287 otos = ip->ip_tos; 288 m_adj(m, off); 289 290 switch (proto) { 291 #ifdef INET 292 case IPPROTO_IPV4: 293 { 294 struct ip *ip; 295 af = AF_INET; 296 if (m->m_len < sizeof(*ip)) { 297 m = m_pullup(m, sizeof(*ip)); 298 if (!m) 299 return; 300 } 301 ip = mtod(m, struct ip *); 302 if (ip_ecn_egress((gifp->if_flags & IFF_LINK1) ? 303 ECN_ALLOWED : ECN_NOCARE, 304 &otos, &ip->ip_tos) == 0) { 305 m_freem(m); 306 return; 307 } 308 break; 309 } 310 #endif 311 #ifdef INET6 312 case IPPROTO_IPV6: 313 { 314 struct ip6_hdr *ip6; 315 u_int8_t itos, oitos; 316 317 af = AF_INET6; 318 if (m->m_len < sizeof(*ip6)) { 319 m = m_pullup(m, sizeof(*ip6)); 320 if (!m) 321 return; 322 } 323 ip6 = mtod(m, struct ip6_hdr *); 324 itos = oitos = (ntohl(ip6->ip6_flow) >> 20) & 0xff; 325 if (ip_ecn_egress((gifp->if_flags & IFF_LINK1) ? 326 ECN_ALLOWED : ECN_NOCARE, 327 &otos, &itos) == 0) { 328 m_freem(m); 329 return; 330 } 331 if (itos != oitos) { 332 ip6->ip6_flow &= ~htonl(0xff << 20); 333 ip6->ip6_flow |= htonl((u_int32_t)itos << 20); 334 } 335 break; 336 } 337 #endif /* INET6 */ 338 case IPPROTO_ETHERIP: 339 af = AF_LINK; 340 break; 341 342 default: 343 IPSTAT_INC(ips_nogif); 344 m_freem(m); 345 return; 346 } 347 gif_input(m, af, gifp); 348 return; 349 } 350 351 /* 352 * validate outer address. 353 */ 354 static int 355 gif_validate4(const struct ip *ip, struct gif_softc *sc, struct ifnet *ifp) 356 { 357 INIT_VNET_INET(curvnet); 358 struct sockaddr_in *src, *dst; 359 struct in_ifaddr *ia4; 360 361 src = (struct sockaddr_in *)sc->gif_psrc; 362 dst = (struct sockaddr_in *)sc->gif_pdst; 363 364 /* check for address match */ 365 if (src->sin_addr.s_addr != ip->ip_dst.s_addr || 366 dst->sin_addr.s_addr != ip->ip_src.s_addr) 367 return 0; 368 369 /* martian filters on outer source - NOT done in ip_input! */ 370 if (IN_MULTICAST(ntohl(ip->ip_src.s_addr))) 371 return 0; 372 switch ((ntohl(ip->ip_src.s_addr) & 0xff000000) >> 24) { 373 case 0: case 127: case 255: 374 return 0; 375 } 376 /* reject packets with broadcast on source */ 377 TAILQ_FOREACH(ia4, &V_in_ifaddrhead, ia_link) { 378 if ((ia4->ia_ifa.ifa_ifp->if_flags & IFF_BROADCAST) == 0) 379 continue; 380 if (ip->ip_src.s_addr == ia4->ia_broadaddr.sin_addr.s_addr) 381 return 0; 382 } 383 384 /* ingress filters on outer source */ 385 if ((GIF2IFP(sc)->if_flags & IFF_LINK2) == 0 && ifp) { 386 struct sockaddr_in sin; 387 struct rtentry *rt; 388 389 bzero(&sin, sizeof(sin)); 390 sin.sin_family = AF_INET; 391 sin.sin_len = sizeof(struct sockaddr_in); 392 sin.sin_addr = ip->ip_src; 393 /* XXX MRT check for the interface we would use on output */ 394 rt = in_rtalloc1((struct sockaddr *)&sin, 0, 395 0UL, sc->gif_fibnum); 396 if (!rt || rt->rt_ifp != ifp) { 397 #if 0 398 log(LOG_WARNING, "%s: packet from 0x%x dropped " 399 "due to ingress filter\n", if_name(GIF2IFP(sc)), 400 (u_int32_t)ntohl(sin.sin_addr.s_addr)); 401 #endif 402 if (rt) 403 RTFREE_LOCKED(rt); 404 return 0; 405 } 406 RTFREE_LOCKED(rt); 407 } 408 409 return 32 * 2; 410 } 411 412 /* 413 * we know that we are in IFF_UP, outer address available, and outer family 414 * matched the physical addr family. see gif_encapcheck(). 415 */ 416 int 417 gif_encapcheck4(const struct mbuf *m, int off, int proto, void *arg) 418 { 419 struct ip ip; 420 struct gif_softc *sc; 421 struct ifnet *ifp; 422 423 /* sanity check done in caller */ 424 sc = (struct gif_softc *)arg; 425 426 /* LINTED const cast */ 427 m_copydata(m, 0, sizeof(ip), (caddr_t)&ip); 428 ifp = ((m->m_flags & M_PKTHDR) != 0) ? m->m_pkthdr.rcvif : NULL; 429 430 return gif_validate4(&ip, sc, ifp); 431 } 432 433 int 434 in_gif_attach(struct gif_softc *sc) 435 { 436 sc->encap_cookie4 = encap_attach_func(AF_INET, -1, gif_encapcheck, 437 &in_gif_protosw, sc); 438 if (sc->encap_cookie4 == NULL) 439 return EEXIST; 440 return 0; 441 } 442 443 int 444 in_gif_detach(struct gif_softc *sc) 445 { 446 int error; 447 448 error = encap_detach(sc->encap_cookie4); 449 if (error == 0) 450 sc->encap_cookie4 = NULL; 451 return error; 452 } 453