1 /* $FreeBSD: src/sys/netinet6/in6_src.c,v 1.1.2.3 2002/02/26 18:02:06 ume Exp $ */ 2 /* $DragonFly: src/sys/netinet6/in6_src.c,v 1.10 2005/02/01 16:09:37 hrs Exp $ */ 3 /* $KAME: in6_src.c,v 1.37 2001/03/29 05:34:31 itojun Exp $ */ 4 5 /* 6 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. 7 * All rights reserved. 8 * 9 * Redistribution and use in source and binary forms, with or without 10 * modification, are permitted provided that the following conditions 11 * are met: 12 * 1. Redistributions of source code must retain the above copyright 13 * notice, this list of conditions and the following disclaimer. 14 * 2. Redistributions in binary form must reproduce the above copyright 15 * notice, this list of conditions and the following disclaimer in the 16 * documentation and/or other materials provided with the distribution. 17 * 3. Neither the name of the project nor the names of its contributors 18 * may be used to endorse or promote products derived from this software 19 * without specific prior written permission. 20 * 21 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND 22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 24 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE 25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 31 * SUCH DAMAGE. 32 */ 33 34 /* 35 * Copyright (c) 1982, 1986, 1991, 1993 36 * The Regents of the University of California. All rights reserved. 37 * 38 * Redistribution and use in source and binary forms, with or without 39 * modification, are permitted provided that the following conditions 40 * are met: 41 * 1. Redistributions of source code must retain the above copyright 42 * notice, this list of conditions and the following disclaimer. 43 * 2. Redistributions in binary form must reproduce the above copyright 44 * notice, this list of conditions and the following disclaimer in the 45 * documentation and/or other materials provided with the distribution. 46 * 3. All advertising materials mentioning features or use of this software 47 * must display the following acknowledgement: 48 * This product includes software developed by the University of 49 * California, Berkeley and its contributors. 50 * 4. Neither the name of the University nor the names of its contributors 51 * may be used to endorse or promote products derived from this software 52 * without specific prior written permission. 53 * 54 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 55 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 56 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 57 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 58 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 59 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 60 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 61 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 62 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 63 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 64 * SUCH DAMAGE. 65 * 66 * @(#)in_pcb.c 8.2 (Berkeley) 1/4/94 67 */ 68 69 #include "opt_inet.h" 70 #include "opt_inet6.h" 71 72 #include <sys/param.h> 73 #include <sys/systm.h> 74 #include <sys/malloc.h> 75 #include <sys/mbuf.h> 76 #include <sys/protosw.h> 77 #include <sys/socket.h> 78 #include <sys/socketvar.h> 79 #include <sys/errno.h> 80 #include <sys/time.h> 81 #include <sys/proc.h> 82 83 #include <net/if.h> 84 #include <net/route.h> 85 86 #include <netinet/in.h> 87 #include <netinet/in_var.h> 88 #include <netinet/in_systm.h> 89 #include <netinet/ip.h> 90 #include <netinet/in_pcb.h> 91 #include <netinet6/in6_var.h> 92 #include <netinet/ip6.h> 93 #include <netinet6/in6_pcb.h> 94 #include <netinet6/ip6_var.h> 95 #include <netinet6/nd6.h> 96 #ifdef ENABLE_DEFAULT_SCOPE 97 #include <netinet6/scope6_var.h> 98 #endif 99 100 #include <net/net_osdep.h> 101 102 #include "use_loop.h" 103 104 /* 105 * Return an IPv6 address, which is the most appropriate for a given 106 * destination and user specified options. 107 * If necessary, this function lookups the routing table and returns 108 * an entry to the caller for later use. 109 */ 110 struct in6_addr * 111 in6_selectsrc(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts, 112 struct ip6_moptions *mopts, struct route_in6 *ro, 113 struct in6_addr *laddr, int *errorp) 114 { 115 struct in6_addr *dst; 116 struct in6_ifaddr *ia6 = 0; 117 struct in6_pktinfo *pi = NULL; 118 119 dst = &dstsock->sin6_addr; 120 *errorp = 0; 121 122 /* 123 * If the source address is explicitly specified by the caller, 124 * use it. 125 */ 126 if (opts && (pi = opts->ip6po_pktinfo) && 127 !IN6_IS_ADDR_UNSPECIFIED(&pi->ipi6_addr)) 128 return(&pi->ipi6_addr); 129 130 /* 131 * If the source address is not specified but the socket(if any) 132 * is already bound, use the bound address. 133 */ 134 if (laddr && !IN6_IS_ADDR_UNSPECIFIED(laddr)) 135 return(laddr); 136 137 /* 138 * If the caller doesn't specify the source address but 139 * the outgoing interface, use an address associated with 140 * the interface. 141 */ 142 if (pi && pi->ipi6_ifindex) { 143 /* XXX boundary check is assumed to be already done. */ 144 ia6 = in6_ifawithscope(ifindex2ifnet[pi->ipi6_ifindex], 145 dst); 146 if (ia6 == 0) { 147 *errorp = EADDRNOTAVAIL; 148 return(0); 149 } 150 return(&satosin6(&ia6->ia_addr)->sin6_addr); 151 } 152 153 /* 154 * If the destination address is a link-local unicast address or 155 * a multicast address, and if the outgoing interface is specified 156 * by the sin6_scope_id filed, use an address associated with the 157 * interface. 158 * XXX: We're now trying to define more specific semantics of 159 * sin6_scope_id field, so this part will be rewritten in 160 * the near future. 161 */ 162 if ((IN6_IS_ADDR_LINKLOCAL(dst) || IN6_IS_ADDR_MULTICAST(dst)) && 163 dstsock->sin6_scope_id) { 164 /* 165 * I'm not sure if boundary check for scope_id is done 166 * somewhere... 167 */ 168 if (dstsock->sin6_scope_id < 0 || 169 if_index < dstsock->sin6_scope_id) { 170 *errorp = ENXIO; /* XXX: better error? */ 171 return(0); 172 } 173 ia6 = in6_ifawithscope(ifindex2ifnet[dstsock->sin6_scope_id], 174 dst); 175 if (ia6 == 0) { 176 *errorp = EADDRNOTAVAIL; 177 return(0); 178 } 179 return(&satosin6(&ia6->ia_addr)->sin6_addr); 180 } 181 182 /* 183 * If the destination address is a multicast address and 184 * the outgoing interface for the address is specified 185 * by the caller, use an address associated with the interface. 186 * There is a sanity check here; if the destination has node-local 187 * scope, the outgoing interfacde should be a loopback address. 188 * Even if the outgoing interface is not specified, we also 189 * choose a loopback interface as the outgoing interface. 190 */ 191 if (IN6_IS_ADDR_MULTICAST(dst)) { 192 struct ifnet *ifp = mopts ? mopts->im6o_multicast_ifp : NULL; 193 194 if (ifp == NULL && IN6_IS_ADDR_MC_NODELOCAL(dst)) { 195 ifp = &loif[0]; 196 } 197 198 if (ifp) { 199 ia6 = in6_ifawithscope(ifp, dst); 200 if (ia6 == 0) { 201 *errorp = EADDRNOTAVAIL; 202 return(0); 203 } 204 return(&satosin6(&ia6->ia_addr)->sin6_addr); 205 } 206 } 207 208 /* 209 * If the next hop address for the packet is specified 210 * by caller, use an address associated with the route 211 * to the next hop. 212 */ 213 { 214 struct sockaddr_in6 *sin6_next; 215 struct rtentry *rt; 216 217 if (opts && opts->ip6po_nexthop) { 218 sin6_next = satosin6(opts->ip6po_nexthop); 219 rt = nd6_lookup(&sin6_next->sin6_addr, 1, NULL); 220 if (rt) { 221 ia6 = in6_ifawithscope(rt->rt_ifp, dst); 222 if (ia6 == 0) 223 ia6 = ifatoia6(rt->rt_ifa); 224 } 225 if (ia6 == 0) { 226 *errorp = EADDRNOTAVAIL; 227 return(0); 228 } 229 return(&satosin6(&ia6->ia_addr)->sin6_addr); 230 } 231 } 232 233 /* 234 * If route is known or can be allocated now, 235 * our src addr is taken from the i/f, else punt. 236 */ 237 if (ro) { 238 if (ro->ro_rt && 239 (!(ro->ro_rt->rt_flags & RTF_UP) || 240 satosin6(&ro->ro_dst)->sin6_family != AF_INET6 || 241 !IN6_ARE_ADDR_EQUAL(&satosin6(&ro->ro_dst)->sin6_addr, 242 dst))) { 243 RTFREE(ro->ro_rt); 244 ro->ro_rt = NULL; 245 } 246 if (ro->ro_rt == NULL || ro->ro_rt->rt_ifp == NULL) { 247 struct sockaddr_in6 *sa6; 248 249 /* No route yet, so try to acquire one */ 250 bzero(&ro->ro_dst, sizeof(struct sockaddr_in6)); 251 sa6 = &ro->ro_dst; 252 sa6->sin6_family = AF_INET6; 253 sa6->sin6_len = sizeof(struct sockaddr_in6); 254 sa6->sin6_addr = *dst; 255 sa6->sin6_scope_id = dstsock->sin6_scope_id; 256 if (IN6_IS_ADDR_MULTICAST(dst)) { 257 ro->ro_rt = 258 rtpurelookup((struct sockaddr *)&ro->ro_dst); 259 } else { 260 rtalloc((struct route *)ro); 261 } 262 } 263 264 /* 265 * in_pcbconnect() checks out IFF_LOOPBACK to skip using 266 * the address. But we don't know why it does so. 267 * It is necessary to ensure the scope even for lo0 268 * so doesn't check out IFF_LOOPBACK. 269 */ 270 271 if (ro->ro_rt) { 272 ia6 = in6_ifawithscope(ro->ro_rt->rt_ifa->ifa_ifp, dst); 273 if (ia6 == 0) /* xxx scope error ?*/ 274 ia6 = ifatoia6(ro->ro_rt->rt_ifa); 275 } 276 #if 0 277 /* 278 * xxx The followings are necessary? (kazu) 279 * I don't think so. 280 * It's for SO_DONTROUTE option in IPv4.(jinmei) 281 */ 282 if (ia6 == 0) { 283 struct sockaddr_in6 sin6 = {sizeof(sin6), AF_INET6, 0}; 284 285 sin6->sin6_addr = *dst; 286 287 ia6 = ifatoia6(ifa_ifwithdstaddr(sin6tosa(&sin6))); 288 if (ia6 == 0) 289 ia6 = ifatoia6(ifa_ifwithnet(sin6tosa(&sin6))); 290 if (ia6 == 0) 291 return(0); 292 return(&satosin6(&ia6->ia_addr)->sin6_addr); 293 } 294 #endif /* 0 */ 295 if (ia6 == 0) { 296 *errorp = EHOSTUNREACH; /* no route */ 297 return(0); 298 } 299 return(&satosin6(&ia6->ia_addr)->sin6_addr); 300 } 301 302 *errorp = EADDRNOTAVAIL; 303 return(0); 304 } 305 306 /* 307 * Default hop limit selection. The precedence is as follows: 308 * 1. Hoplimit value specified via ioctl. 309 * 2. (If the outgoing interface is detected) the current 310 * hop limit of the interface specified by router advertisement. 311 * 3. The system default hoplimit. 312 */ 313 int 314 in6_selecthlim(struct in6pcb *in6p, struct ifnet *ifp) 315 { 316 if (in6p && in6p->in6p_hops >= 0) 317 return(in6p->in6p_hops); 318 else if (ifp) 319 return(ND_IFINFO(ifp)->chlim); 320 else 321 return(ip6_defhlim); 322 } 323 324 /* 325 * XXX: this is borrowed from in6_pcbbind(). If possible, we should 326 * share this function by all *bsd*... 327 */ 328 int 329 in6_pcbsetport(struct in6_addr *laddr, struct inpcb *inp, struct thread *td) 330 { 331 struct socket *so = inp->inp_socket; 332 u_int16_t lport = 0, first, last, *lastport; 333 int count, error = 0, wild = 0; 334 struct inpcbinfo *pcbinfo = inp->inp_pcbinfo; 335 336 /* XXX: this is redundant when called from in6_pcbbind */ 337 if ((so->so_options & (SO_REUSEADDR|SO_REUSEPORT)) == 0) 338 wild = INPLOOKUP_WILDCARD; 339 340 inp->inp_flags |= INP_ANONPORT; 341 342 if (inp->inp_flags & INP_HIGHPORT) { 343 first = ipport_hifirstauto; /* sysctl */ 344 last = ipport_hilastauto; 345 lastport = &pcbinfo->lasthi; 346 } else if (inp->inp_flags & INP_LOWPORT) { 347 if ((error = suser(td)) != 0) 348 return error; 349 first = ipport_lowfirstauto; /* 1023 */ 350 last = ipport_lowlastauto; /* 600 */ 351 lastport = &pcbinfo->lastlow; 352 } else { 353 first = ipport_firstauto; /* sysctl */ 354 last = ipport_lastauto; 355 lastport = &pcbinfo->lastport; 356 } 357 /* 358 * Simple check to ensure all ports are not used up causing 359 * a deadlock here. 360 * 361 * We split the two cases (up and down) so that the direction 362 * is not being tested on each round of the loop. 363 */ 364 if (first > last) { 365 /* 366 * counting down 367 */ 368 count = first - last; 369 370 do { 371 if (count-- < 0) { /* completely used? */ 372 /* 373 * Undo any address bind that may have 374 * occurred above. 375 */ 376 inp->in6p_laddr = in6addr_any; 377 return (EAGAIN); 378 } 379 --*lastport; 380 if (*lastport > first || *lastport < last) 381 *lastport = first; 382 lport = htons(*lastport); 383 } while (in6_pcblookup_local(pcbinfo, 384 &inp->in6p_laddr, lport, wild)); 385 } else { 386 /* 387 * counting up 388 */ 389 count = last - first; 390 391 do { 392 if (count-- < 0) { /* completely used? */ 393 /* 394 * Undo any address bind that may have 395 * occurred above. 396 */ 397 inp->in6p_laddr = in6addr_any; 398 return (EAGAIN); 399 } 400 ++*lastport; 401 if (*lastport < first || *lastport > last) 402 *lastport = first; 403 lport = htons(*lastport); 404 } while (in6_pcblookup_local(pcbinfo, 405 &inp->in6p_laddr, lport, wild)); 406 } 407 408 inp->inp_lport = lport; 409 if (in_pcbinsporthash(inp) != 0) { 410 inp->in6p_laddr = in6addr_any; 411 inp->inp_lport = 0; 412 return (EAGAIN); 413 } 414 415 return(0); 416 } 417 418 /* 419 * generate kernel-internal form (scopeid embedded into s6_addr16[1]). 420 * If the address scope of is link-local, embed the interface index in the 421 * address. The routine determines our precedence 422 * between advanced API scope/interface specification and basic API 423 * specification. 424 * 425 * this function should be nuked in the future, when we get rid of 426 * embedded scopeid thing. 427 * 428 * XXX actually, it is over-specification to return ifp against sin6_scope_id. 429 * there can be multiple interfaces that belong to a particular scope zone 430 * (in specification, we have 1:N mapping between a scope zone and interfaces). 431 * we may want to change the function to return something other than ifp. 432 */ 433 int 434 in6_embedscope(struct in6_addr *in6, 435 const struct sockaddr_in6 *sin6, 436 #ifdef HAVE_NRL_INPCB 437 struct inpcb *in6p, 438 #define in6p_outputopts inp_outputopts6 439 #define in6p_moptions inp_moptions6 440 #else 441 struct in6pcb *in6p, 442 #endif 443 struct ifnet **ifpp) 444 { 445 struct ifnet *ifp = NULL; 446 u_int32_t scopeid; 447 448 *in6 = sin6->sin6_addr; 449 scopeid = sin6->sin6_scope_id; 450 if (ifpp) 451 *ifpp = NULL; 452 453 /* 454 * don't try to read sin6->sin6_addr beyond here, since the caller may 455 * ask us to overwrite existing sockaddr_in6 456 */ 457 458 #ifdef ENABLE_DEFAULT_SCOPE 459 if (scopeid == 0) 460 scopeid = scope6_addr2default(in6); 461 #endif 462 463 if (IN6_IS_SCOPE_LINKLOCAL(in6)) { 464 struct in6_pktinfo *pi; 465 466 /* 467 * KAME assumption: link id == interface id 468 */ 469 470 if (in6p && in6p->in6p_outputopts && 471 (pi = in6p->in6p_outputopts->ip6po_pktinfo) && 472 pi->ipi6_ifindex) { 473 ifp = ifindex2ifnet[pi->ipi6_ifindex]; 474 in6->s6_addr16[1] = htons(pi->ipi6_ifindex); 475 } else if (in6p && IN6_IS_ADDR_MULTICAST(in6) && 476 in6p->in6p_moptions && 477 in6p->in6p_moptions->im6o_multicast_ifp) { 478 ifp = in6p->in6p_moptions->im6o_multicast_ifp; 479 in6->s6_addr16[1] = htons(ifp->if_index); 480 } else if (scopeid) { 481 /* boundary check */ 482 if (scopeid < 0 || if_index < scopeid) 483 return ENXIO; /* XXX EINVAL? */ 484 ifp = ifindex2ifnet[scopeid]; 485 /*XXX assignment to 16bit from 32bit variable */ 486 in6->s6_addr16[1] = htons(scopeid & 0xffff); 487 } 488 489 if (ifpp) 490 *ifpp = ifp; 491 } 492 493 return 0; 494 } 495 #ifdef HAVE_NRL_INPCB 496 #undef in6p_outputopts 497 #undef in6p_moptions 498 #endif 499 500 /* 501 * generate standard sockaddr_in6 from embedded form. 502 * touches sin6_addr and sin6_scope_id only. 503 * 504 * this function should be nuked in the future, when we get rid of 505 * embedded scopeid thing. 506 */ 507 int 508 in6_recoverscope(struct sockaddr_in6 *sin6, const struct in6_addr *in6, 509 struct ifnet *ifp) 510 { 511 u_int32_t scopeid; 512 513 sin6->sin6_addr = *in6; 514 515 /* 516 * don't try to read *in6 beyond here, since the caller may 517 * ask us to overwrite existing sockaddr_in6 518 */ 519 520 sin6->sin6_scope_id = 0; 521 if (IN6_IS_SCOPE_LINKLOCAL(in6)) { 522 /* 523 * KAME assumption: link id == interface id 524 */ 525 scopeid = ntohs(sin6->sin6_addr.s6_addr16[1]); 526 if (scopeid) { 527 /* sanity check */ 528 if (scopeid < 0 || if_index < scopeid) 529 return ENXIO; 530 if (ifp && ifp->if_index != scopeid) 531 return ENXIO; 532 sin6->sin6_addr.s6_addr16[1] = 0; 533 sin6->sin6_scope_id = scopeid; 534 } 535 } 536 537 return 0; 538 } 539 540 /* 541 * just clear the embedded scope identifer. 542 * XXX: currently used for bsdi4 only as a supplement function. 543 */ 544 void 545 in6_clearscope(struct in6_addr *addr) 546 { 547 if (IN6_IS_SCOPE_LINKLOCAL(addr)) 548 addr->s6_addr16[1] = 0; 549 } 550