1 /*- 2 * SPDX-License-Identifier: BSD-3-Clause 3 * 4 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. 5 * Copyright (c) 2018 Andrey V. Elsukov <ae@FreeBSD.org> 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 * $KAME: in_gif.c,v 1.54 2001/05/14 14:02:16 itojun Exp $ 33 */ 34 35 #include <sys/cdefs.h> 36 __FBSDID("$FreeBSD$"); 37 38 #include "opt_inet.h" 39 #include "opt_inet6.h" 40 41 #include <sys/param.h> 42 #include <sys/systm.h> 43 #include <sys/jail.h> 44 #include <sys/socket.h> 45 #include <sys/sockio.h> 46 #include <sys/mbuf.h> 47 #include <sys/errno.h> 48 #include <sys/kernel.h> 49 #include <sys/sysctl.h> 50 #include <sys/malloc.h> 51 #include <sys/proc.h> 52 53 #include <net/ethernet.h> 54 #include <net/if.h> 55 #include <net/if_var.h> 56 #include <net/route.h> 57 #include <net/vnet.h> 58 59 #include <netinet/in.h> 60 #include <netinet/in_systm.h> 61 #include <netinet/ip.h> 62 #include <netinet/ip_var.h> 63 #include <netinet/in_var.h> 64 #include <netinet/ip_encap.h> 65 #include <netinet/ip_ecn.h> 66 #include <netinet/in_fib.h> 67 68 #ifdef INET6 69 #include <netinet/ip6.h> 70 #endif 71 72 #include <net/if_gif.h> 73 74 #define GIF_TTL 30 75 VNET_DEFINE_STATIC(int, ip_gif_ttl) = GIF_TTL; 76 #define V_ip_gif_ttl VNET(ip_gif_ttl) 77 SYSCTL_INT(_net_inet_ip, IPCTL_GIF_TTL, gifttl, CTLFLAG_VNET | CTLFLAG_RW, 78 &VNET_NAME(ip_gif_ttl), 0, "Default TTL value for encapsulated packets"); 79 80 /* 81 * We keep interfaces in a hash table using src+dst as key. 82 * Interfaces with GIF_IGNORE_SOURCE flag are linked into plain list. 83 */ 84 VNET_DEFINE_STATIC(struct gif_list *, ipv4_hashtbl) = NULL; 85 VNET_DEFINE_STATIC(struct gif_list, ipv4_list) = CK_LIST_HEAD_INITIALIZER(); 86 #define V_ipv4_hashtbl VNET(ipv4_hashtbl) 87 #define V_ipv4_list VNET(ipv4_list) 88 89 #define GIF_HASH(src, dst) (V_ipv4_hashtbl[\ 90 in_gif_hashval((src), (dst)) & (GIF_HASH_SIZE - 1)]) 91 #define GIF_HASH_SC(sc) GIF_HASH((sc)->gif_iphdr->ip_src.s_addr,\ 92 (sc)->gif_iphdr->ip_dst.s_addr) 93 static uint32_t 94 in_gif_hashval(in_addr_t src, in_addr_t dst) 95 { 96 uint32_t ret; 97 98 ret = fnv_32_buf(&src, sizeof(src), FNV1_32_INIT); 99 return (fnv_32_buf(&dst, sizeof(dst), ret)); 100 } 101 102 static int 103 in_gif_checkdup(const struct gif_softc *sc, in_addr_t src, in_addr_t dst) 104 { 105 struct gif_softc *tmp; 106 107 if (sc->gif_family == AF_INET && 108 sc->gif_iphdr->ip_src.s_addr == src && 109 sc->gif_iphdr->ip_dst.s_addr == dst) 110 return (EEXIST); 111 112 CK_LIST_FOREACH(tmp, &GIF_HASH(src, dst), chain) { 113 if (tmp == sc) 114 continue; 115 if (tmp->gif_iphdr->ip_src.s_addr == src && 116 tmp->gif_iphdr->ip_dst.s_addr == dst) 117 return (EADDRNOTAVAIL); 118 } 119 return (0); 120 } 121 122 static void 123 in_gif_attach(struct gif_softc *sc) 124 { 125 126 if (sc->gif_options & GIF_IGNORE_SOURCE) 127 CK_LIST_INSERT_HEAD(&V_ipv4_list, sc, chain); 128 else 129 CK_LIST_INSERT_HEAD(&GIF_HASH_SC(sc), sc, chain); 130 } 131 132 int 133 in_gif_setopts(struct gif_softc *sc, u_int options) 134 { 135 136 /* NOTE: we are protected with gif_ioctl_sx lock */ 137 MPASS(sc->gif_family == AF_INET); 138 MPASS(sc->gif_options != options); 139 140 if ((options & GIF_IGNORE_SOURCE) != 141 (sc->gif_options & GIF_IGNORE_SOURCE)) { 142 CK_LIST_REMOVE(sc, chain); 143 sc->gif_options = options; 144 in_gif_attach(sc); 145 } 146 return (0); 147 } 148 149 int 150 in_gif_ioctl(struct gif_softc *sc, u_long cmd, caddr_t data) 151 { 152 struct ifreq *ifr = (struct ifreq *)data; 153 struct sockaddr_in *dst, *src; 154 struct ip *ip; 155 int error; 156 157 /* NOTE: we are protected with gif_ioctl_sx lock */ 158 error = EINVAL; 159 switch (cmd) { 160 case SIOCSIFPHYADDR: 161 src = &((struct in_aliasreq *)data)->ifra_addr; 162 dst = &((struct in_aliasreq *)data)->ifra_dstaddr; 163 164 /* sanity checks */ 165 if (src->sin_family != dst->sin_family || 166 src->sin_family != AF_INET || 167 src->sin_len != dst->sin_len || 168 src->sin_len != sizeof(*src)) 169 break; 170 if (src->sin_addr.s_addr == INADDR_ANY || 171 dst->sin_addr.s_addr == INADDR_ANY) { 172 error = EADDRNOTAVAIL; 173 break; 174 } 175 if (V_ipv4_hashtbl == NULL) 176 V_ipv4_hashtbl = gif_hashinit(); 177 error = in_gif_checkdup(sc, src->sin_addr.s_addr, 178 dst->sin_addr.s_addr); 179 if (error == EADDRNOTAVAIL) 180 break; 181 if (error == EEXIST) { 182 /* Addresses are the same. Just return. */ 183 error = 0; 184 break; 185 } 186 ip = malloc(sizeof(*ip), M_GIF, M_WAITOK | M_ZERO); 187 ip->ip_src.s_addr = src->sin_addr.s_addr; 188 ip->ip_dst.s_addr = dst->sin_addr.s_addr; 189 if (sc->gif_family != 0) { 190 /* Detach existing tunnel first */ 191 CK_LIST_REMOVE(sc, chain); 192 GIF_WAIT(); 193 free(sc->gif_hdr, M_GIF); 194 /* XXX: should we notify about link state change? */ 195 } 196 sc->gif_family = AF_INET; 197 sc->gif_iphdr = ip; 198 in_gif_attach(sc); 199 break; 200 case SIOCGIFPSRCADDR: 201 case SIOCGIFPDSTADDR: 202 if (sc->gif_family != AF_INET) { 203 error = EADDRNOTAVAIL; 204 break; 205 } 206 src = (struct sockaddr_in *)&ifr->ifr_addr; 207 memset(src, 0, sizeof(*src)); 208 src->sin_family = AF_INET; 209 src->sin_len = sizeof(*src); 210 src->sin_addr = (cmd == SIOCGIFPSRCADDR) ? 211 sc->gif_iphdr->ip_src: sc->gif_iphdr->ip_dst; 212 error = prison_if(curthread->td_ucred, (struct sockaddr *)src); 213 if (error != 0) 214 memset(src, 0, sizeof(*src)); 215 break; 216 } 217 return (error); 218 } 219 220 int 221 in_gif_output(struct ifnet *ifp, struct mbuf *m, int proto, uint8_t ecn) 222 { 223 struct gif_softc *sc = ifp->if_softc; 224 struct ip *ip; 225 int len; 226 227 /* prepend new IP header */ 228 MPASS(in_epoch(net_epoch_preempt)); 229 len = sizeof(struct ip); 230 #ifndef __NO_STRICT_ALIGNMENT 231 if (proto == IPPROTO_ETHERIP) 232 len += ETHERIP_ALIGN; 233 #endif 234 M_PREPEND(m, len, M_NOWAIT); 235 if (m == NULL) 236 return (ENOBUFS); 237 #ifndef __NO_STRICT_ALIGNMENT 238 if (proto == IPPROTO_ETHERIP) { 239 len = mtod(m, vm_offset_t) & 3; 240 KASSERT(len == 0 || len == ETHERIP_ALIGN, 241 ("in_gif_output: unexpected misalignment")); 242 m->m_data += len; 243 m->m_len -= ETHERIP_ALIGN; 244 } 245 #endif 246 ip = mtod(m, struct ip *); 247 248 MPASS(sc->gif_family == AF_INET); 249 bcopy(sc->gif_iphdr, ip, sizeof(struct ip)); 250 ip->ip_p = proto; 251 /* version will be set in ip_output() */ 252 ip->ip_ttl = V_ip_gif_ttl; 253 ip->ip_len = htons(m->m_pkthdr.len); 254 ip->ip_tos = ecn; 255 256 return (ip_output(m, NULL, NULL, 0, NULL, NULL)); 257 } 258 259 static int 260 in_gif_input(struct mbuf *m, int off, int proto, void *arg) 261 { 262 struct gif_softc *sc = arg; 263 struct ifnet *gifp; 264 struct ip *ip; 265 uint8_t ecn; 266 267 MPASS(in_epoch(net_epoch_preempt)); 268 if (sc == NULL) { 269 m_freem(m); 270 KMOD_IPSTAT_INC(ips_nogif); 271 return (IPPROTO_DONE); 272 } 273 gifp = GIF2IFP(sc); 274 if ((gifp->if_flags & IFF_UP) != 0) { 275 ip = mtod(m, struct ip *); 276 ecn = ip->ip_tos; 277 m_adj(m, off); 278 gif_input(m, gifp, proto, ecn); 279 } else { 280 m_freem(m); 281 KMOD_IPSTAT_INC(ips_nogif); 282 } 283 return (IPPROTO_DONE); 284 } 285 286 static int 287 in_gif_lookup(const struct mbuf *m, int off, int proto, void **arg) 288 { 289 const struct ip *ip; 290 struct gif_softc *sc; 291 int ret; 292 293 if (V_ipv4_hashtbl == NULL) 294 return (0); 295 296 MPASS(in_epoch(net_epoch_preempt)); 297 ip = mtod(m, const struct ip *); 298 /* 299 * NOTE: it is safe to iterate without any locking here, because softc 300 * can be reclaimed only when we are not within net_epoch_preempt 301 * section, but ip_encap lookup+input are executed in epoch section. 302 */ 303 ret = 0; 304 CK_LIST_FOREACH(sc, &GIF_HASH(ip->ip_dst.s_addr, 305 ip->ip_src.s_addr), chain) { 306 /* 307 * This is an inbound packet, its ip_dst is source address 308 * in softc. 309 */ 310 if (sc->gif_iphdr->ip_src.s_addr == ip->ip_dst.s_addr && 311 sc->gif_iphdr->ip_dst.s_addr == ip->ip_src.s_addr) { 312 ret = ENCAP_DRV_LOOKUP; 313 goto done; 314 } 315 } 316 /* 317 * No exact match. 318 * Check the list of interfaces with GIF_IGNORE_SOURCE flag. 319 */ 320 CK_LIST_FOREACH(sc, &V_ipv4_list, chain) { 321 if (sc->gif_iphdr->ip_src.s_addr == ip->ip_dst.s_addr) { 322 ret = 32 + 8; /* src + proto */ 323 goto done; 324 } 325 } 326 return (0); 327 done: 328 if ((GIF2IFP(sc)->if_flags & IFF_UP) == 0) 329 return (0); 330 /* ingress filters on outer source */ 331 if ((GIF2IFP(sc)->if_flags & IFF_LINK2) == 0) { 332 struct nhop4_basic nh4; 333 struct in_addr dst; 334 335 dst = ip->ip_src; 336 if (fib4_lookup_nh_basic(sc->gif_fibnum, dst, 0, 0, &nh4) != 0) 337 return (0); 338 if (nh4.nh_ifp != m->m_pkthdr.rcvif) 339 return (0); 340 } 341 *arg = sc; 342 return (ret); 343 } 344 345 static struct { 346 const struct encap_config encap; 347 const struct encaptab *cookie; 348 } ipv4_encap_cfg[] = { 349 { 350 .encap = { 351 .proto = IPPROTO_IPV4, 352 .min_length = 2 * sizeof(struct ip), 353 .exact_match = ENCAP_DRV_LOOKUP, 354 .lookup = in_gif_lookup, 355 .input = in_gif_input 356 }, 357 }, 358 #ifdef INET6 359 { 360 .encap = { 361 .proto = IPPROTO_IPV6, 362 .min_length = sizeof(struct ip) + 363 sizeof(struct ip6_hdr), 364 .exact_match = ENCAP_DRV_LOOKUP, 365 .lookup = in_gif_lookup, 366 .input = in_gif_input 367 }, 368 }, 369 #endif 370 { 371 .encap = { 372 .proto = IPPROTO_ETHERIP, 373 .min_length = sizeof(struct ip) + 374 sizeof(struct etherip_header) + 375 sizeof(struct ether_header), 376 .exact_match = ENCAP_DRV_LOOKUP, 377 .lookup = in_gif_lookup, 378 .input = in_gif_input 379 }, 380 } 381 }; 382 383 void 384 in_gif_init(void) 385 { 386 int i; 387 388 if (!IS_DEFAULT_VNET(curvnet)) 389 return; 390 for (i = 0; i < nitems(ipv4_encap_cfg); i++) 391 ipv4_encap_cfg[i].cookie = ip_encap_attach( 392 &ipv4_encap_cfg[i].encap, NULL, M_WAITOK); 393 } 394 395 void 396 in_gif_uninit(void) 397 { 398 int i; 399 400 if (IS_DEFAULT_VNET(curvnet)) { 401 for (i = 0; i < nitems(ipv4_encap_cfg); i++) 402 ip_encap_detach(ipv4_encap_cfg[i].cookie); 403 } 404 if (V_ipv4_hashtbl != NULL) 405 gif_hashdestroy(V_ipv4_hashtbl); 406 } 407 408