in6_src.c revision 121353
1/* $FreeBSD: head/sys/netinet6/in6_src.c 121353 2003-10-22 18:52:57Z ume $ */ 2/* $KAME: in6_src.c,v 1.37 2001/03/29 05:34:31 itojun Exp $ */ 3 4/* 5 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. 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 33/* 34 * Copyright (c) 1982, 1986, 1991, 1993 35 * The Regents of the University of California. All rights reserved. 36 * 37 * Redistribution and use in source and binary forms, with or without 38 * modification, are permitted provided that the following conditions 39 * are met: 40 * 1. Redistributions of source code must retain the above copyright 41 * notice, this list of conditions and the following disclaimer. 42 * 2. Redistributions in binary form must reproduce the above copyright 43 * notice, this list of conditions and the following disclaimer in the 44 * documentation and/or other materials provided with the distribution. 45 * 3. All advertising materials mentioning features or use of this software 46 * must display the following acknowledgement: 47 * This product includes software developed by the University of 48 * California, Berkeley and its contributors. 49 * 4. Neither the name of the University nor the names of its contributors 50 * may be used to endorse or promote products derived from this software 51 * without specific prior written permission. 52 * 53 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 54 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 55 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 56 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 57 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 58 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 59 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 60 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 61 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 62 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 63 * SUCH DAMAGE. 64 * 65 * @(#)in_pcb.c 8.2 (Berkeley) 1/4/94 66 */ 67 68#include "opt_inet.h" 69#include "opt_inet6.h" 70 71#include <sys/param.h> 72#include <sys/systm.h> 73#include <sys/malloc.h> 74#include <sys/mbuf.h> 75#include <sys/protosw.h> 76#include <sys/socket.h> 77#include <sys/socketvar.h> 78#include <sys/errno.h> 79#include <sys/time.h> 80 81#include <net/if.h> 82#include <net/route.h> 83 84#include <netinet/in.h> 85#include <netinet/in_var.h> 86#include <netinet/in_systm.h> 87#include <netinet/ip.h> 88#include <netinet/in_pcb.h> 89#include <netinet6/in6_var.h> 90#include <netinet/ip6.h> 91#include <netinet6/in6_pcb.h> 92#include <netinet6/ip6_var.h> 93#include <netinet6/nd6.h> 94#ifdef ENABLE_DEFAULT_SCOPE 95#include <netinet6/scope6_var.h> 96#endif 97 98#include <net/net_osdep.h> 99 100/* 101 * Return an IPv6 address, which is the most appropriate for a given 102 * destination and user specified options. 103 * If necessary, this function lookups the routing table and returns 104 * an entry to the caller for later use. 105 */ 106struct in6_addr * 107in6_selectsrc(dstsock, opts, mopts, ro, laddr, errorp) 108 struct sockaddr_in6 *dstsock; 109 struct ip6_pktopts *opts; 110 struct ip6_moptions *mopts; 111 struct route_in6 *ro; 112 struct in6_addr *laddr; 113 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(ifnet_byindex(pi->ipi6_ifindex), dst); 145 if (ia6 == 0) { 146 *errorp = EADDRNOTAVAIL; 147 return (0); 148 } 149 return (&satosin6(&ia6->ia_addr)->sin6_addr); 150 } 151 152 /* 153 * If the destination address is a link-local unicast address or 154 * a multicast address, and if the outgoing interface is specified 155 * by the sin6_scope_id filed, use an address associated with the 156 * interface. 157 * XXX: We're now trying to define more specific semantics of 158 * sin6_scope_id field, so this part will be rewritten in 159 * the near future. 160 */ 161 if ((IN6_IS_ADDR_LINKLOCAL(dst) || IN6_IS_ADDR_MULTICAST(dst)) && 162 dstsock->sin6_scope_id) { 163 /* 164 * I'm not sure if boundary check for scope_id is done 165 * somewhere... 166 */ 167 if (dstsock->sin6_scope_id < 0 || 168 if_index < dstsock->sin6_scope_id) { 169 *errorp = ENXIO; /* XXX: better error? */ 170 return (0); 171 } 172 ia6 = in6_ifawithscope(ifnet_byindex(dstsock->sin6_scope_id), 173 dst); 174 if (ia6 == 0) { 175 *errorp = EADDRNOTAVAIL; 176 return (0); 177 } 178 return (&satosin6(&ia6->ia_addr)->sin6_addr); 179 } 180 181 /* 182 * If the destination address is a multicast address and 183 * the outgoing interface for the address is specified 184 * by the caller, use an address associated with the interface. 185 * There is a sanity check here; if the destination has node-local 186 * scope, the outgoing interfacde should be a loopback address. 187 * Even if the outgoing interface is not specified, we also 188 * choose a loopback interface as the outgoing interface. 189 */ 190 if (IN6_IS_ADDR_MULTICAST(dst)) { 191 struct ifnet *ifp = mopts ? mopts->im6o_multicast_ifp : NULL; 192 193 if (ifp == NULL && IN6_IS_ADDR_MC_INTFACELOCAL(dst)) { 194 ifp = &loif[0]; 195 } 196 197 if (ifp) { 198 ia6 = in6_ifawithscope(ifp, dst); 199 if (ia6 == 0) { 200 *errorp = EADDRNOTAVAIL; 201 return (0); 202 } 203 return (&satosin6(&ia6->ia_addr)->sin6_addr); 204 } 205 } 206 207 /* 208 * If the next hop address for the packet is specified 209 * by caller, use an address associated with the route 210 * to the next hop. 211 */ 212 { 213 struct sockaddr_in6 *sin6_next; 214 struct rtentry *rt; 215 216 if (opts && opts->ip6po_nexthop) { 217 sin6_next = satosin6(opts->ip6po_nexthop); 218 rt = nd6_lookup(&sin6_next->sin6_addr, 1, NULL); 219 if (rt) { 220 ia6 = in6_ifawithscope(rt->rt_ifp, dst); 221 if (ia6 == 0) 222 ia6 = ifatoia6(rt->rt_ifa); 223 } 224 if (ia6 == 0) { 225 *errorp = EADDRNOTAVAIL; 226 return (0); 227 } 228 return (&satosin6(&ia6->ia_addr)->sin6_addr); 229 } 230 } 231 232 /* 233 * If route is known or can be allocated now, 234 * our src addr is taken from the i/f, else punt. 235 */ 236 if (ro) { 237 if (ro->ro_rt && 238 (!(ro->ro_rt->rt_flags & RTF_UP) || 239 satosin6(&ro->ro_dst)->sin6_family != AF_INET6 || 240 !IN6_ARE_ADDR_EQUAL(&satosin6(&ro->ro_dst)->sin6_addr, 241 dst))) { 242 RTFREE(ro->ro_rt); 243 ro->ro_rt = (struct rtentry *)0; 244 } 245 if (ro->ro_rt == (struct rtentry *)0 || 246 ro->ro_rt->rt_ifp == (struct ifnet *)0) { 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 = (struct sockaddr_in6 *)&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 = rtalloc1(&((struct route *)ro) 258 ->ro_dst, 0, 0UL); 259 RT_UNLOCK(ro->ro_rt); 260 } else { 261 rtalloc((struct route *)ro); 262 } 263 } 264 265 /* 266 * in_pcbconnect() checks out IFF_LOOPBACK to skip using 267 * the address. But we don't know why it does so. 268 * It is necessary to ensure the scope even for lo0 269 * so doesn't check out IFF_LOOPBACK. 270 */ 271 272 if (ro->ro_rt) { 273 ia6 = in6_ifawithscope(ro->ro_rt->rt_ifa->ifa_ifp, dst); 274 if (ia6 == 0) /* xxx scope error ?*/ 275 ia6 = ifatoia6(ro->ro_rt->rt_ifa); 276 } 277 if (ia6 == 0) { 278 *errorp = EHOSTUNREACH; /* no route */ 279 return (0); 280 } 281 return (&satosin6(&ia6->ia_addr)->sin6_addr); 282 } 283 284 *errorp = EADDRNOTAVAIL; 285 return (0); 286} 287 288/* 289 * Default hop limit selection. The precedence is as follows: 290 * 1. Hoplimit value specified via ioctl. 291 * 2. (If the outgoing interface is detected) the current 292 * hop limit of the interface specified by router advertisement. 293 * 3. The system default hoplimit. 294*/ 295int 296in6_selecthlim(in6p, ifp) 297 struct in6pcb *in6p; 298 struct ifnet *ifp; 299{ 300 if (in6p && in6p->in6p_hops >= 0) 301 return (in6p->in6p_hops); 302 else if (ifp) 303 return (ND_IFINFO(ifp)->chlim); 304 else 305 return (ip6_defhlim); 306} 307 308/* 309 * XXX: this is borrowed from in6_pcbbind(). If possible, we should 310 * share this function by all *bsd*... 311 */ 312int 313in6_pcbsetport(laddr, inp, td) 314 struct in6_addr *laddr; 315 struct inpcb *inp; 316 struct thread *td; 317{ 318 struct socket *so = inp->inp_socket; 319 u_int16_t lport = 0, first, last, *lastport; 320 int count, error = 0, wild = 0; 321 struct inpcbinfo *pcbinfo = inp->inp_pcbinfo; 322 323 /* XXX: this is redundant when called from in6_pcbbind */ 324 if ((so->so_options & (SO_REUSEADDR|SO_REUSEPORT)) == 0) 325 wild = INPLOOKUP_WILDCARD; 326 327 inp->inp_flags |= INP_ANONPORT; 328 329 if (inp->inp_flags & INP_HIGHPORT) { 330 first = ipport_hifirstauto; /* sysctl */ 331 last = ipport_hilastauto; 332 lastport = &pcbinfo->lasthi; 333 } else if (inp->inp_flags & INP_LOWPORT) { 334 if (td && (error = suser(td))) 335 return error; 336 first = ipport_lowfirstauto; /* 1023 */ 337 last = ipport_lowlastauto; /* 600 */ 338 lastport = &pcbinfo->lastlow; 339 } else { 340 first = ipport_firstauto; /* sysctl */ 341 last = ipport_lastauto; 342 lastport = &pcbinfo->lastport; 343 } 344 /* 345 * Simple check to ensure all ports are not used up causing 346 * a deadlock here. 347 * 348 * We split the two cases (up and down) so that the direction 349 * is not being tested on each round of the loop. 350 */ 351 if (first > last) { 352 /* 353 * counting down 354 */ 355 count = first - last; 356 357 do { 358 if (count-- < 0) { /* completely used? */ 359 /* 360 * Undo any address bind that may have 361 * occurred above. 362 */ 363 inp->in6p_laddr = in6addr_any; 364 return (EAGAIN); 365 } 366 --*lastport; 367 if (*lastport > first || *lastport < last) 368 *lastport = first; 369 lport = htons(*lastport); 370 } while (in6_pcblookup_local(pcbinfo, 371 &inp->in6p_laddr, lport, wild)); 372 } else { 373 /* 374 * counting up 375 */ 376 count = last - first; 377 378 do { 379 if (count-- < 0) { /* completely used? */ 380 /* 381 * Undo any address bind that may have 382 * occurred above. 383 */ 384 inp->in6p_laddr = in6addr_any; 385 return (EAGAIN); 386 } 387 ++*lastport; 388 if (*lastport < first || *lastport > last) 389 *lastport = first; 390 lport = htons(*lastport); 391 } while (in6_pcblookup_local(pcbinfo, 392 &inp->in6p_laddr, lport, wild)); 393 } 394 395 inp->inp_lport = lport; 396 if (in_pcbinshash(inp) != 0) { 397 inp->in6p_laddr = in6addr_any; 398 inp->inp_lport = 0; 399 return (EAGAIN); 400 } 401 402 return (0); 403} 404 405/* 406 * Generate kernel-internal form (scopeid embedded into s6_addr16[1]). 407 * If the address scope of is link-local, embed the interface index in the 408 * address. The routine determines our precedence 409 * between advanced API scope/interface specification and basic API 410 * specification. 411 * 412 * This function should be nuked in the future, when we get rid of embedded 413 * scopeid thing. 414 * 415 * XXX actually, it is over-specification to return ifp against sin6_scope_id. 416 * there can be multiple interfaces that belong to a particular scope zone 417 * (in specification, we have 1:N mapping between a scope zone and interfaces). 418 * we may want to change the function to return something other than ifp. 419 */ 420int 421in6_embedscope(in6, sin6, in6p, ifpp) 422 struct in6_addr *in6; 423 const struct sockaddr_in6 *sin6; 424 struct in6pcb *in6p; 425 struct ifnet **ifpp; 426{ 427 struct ifnet *ifp = NULL; 428 u_int32_t scopeid; 429 430 *in6 = sin6->sin6_addr; 431 scopeid = sin6->sin6_scope_id; 432 if (ifpp) 433 *ifpp = NULL; 434 435 /* 436 * don't try to read sin6->sin6_addr beyond here, since the caller may 437 * ask us to overwrite existing sockaddr_in6 438 */ 439 440#ifdef ENABLE_DEFAULT_SCOPE 441 if (scopeid == 0) 442 scopeid = scope6_addr2default(in6); 443#endif 444 445 if (IN6_IS_SCOPE_LINKLOCAL(in6)) { 446 struct in6_pktinfo *pi; 447 448 /* 449 * KAME assumption: link id == interface id 450 */ 451 452 if (in6p && in6p->in6p_outputopts && 453 (pi = in6p->in6p_outputopts->ip6po_pktinfo) && 454 pi->ipi6_ifindex) { 455 ifp = ifnet_byindex(pi->ipi6_ifindex); 456 in6->s6_addr16[1] = htons(pi->ipi6_ifindex); 457 } else if (in6p && IN6_IS_ADDR_MULTICAST(in6) && 458 in6p->in6p_moptions && 459 in6p->in6p_moptions->im6o_multicast_ifp) { 460 ifp = in6p->in6p_moptions->im6o_multicast_ifp; 461 in6->s6_addr16[1] = htons(ifp->if_index); 462 } else if (scopeid) { 463 /* boundary check */ 464 if (scopeid < 0 || if_index < scopeid) 465 return ENXIO; /* XXX EINVAL? */ 466 ifp = ifnet_byindex(scopeid); 467 /* XXX assignment to 16bit from 32bit variable */ 468 in6->s6_addr16[1] = htons(scopeid & 0xffff); 469 } 470 471 if (ifpp) 472 *ifpp = ifp; 473 } 474 475 return 0; 476} 477 478/* 479 * generate standard sockaddr_in6 from embedded form. 480 * touches sin6_addr and sin6_scope_id only. 481 * 482 * this function should be nuked in the future, when we get rid of 483 * embedded scopeid thing. 484 */ 485int 486in6_recoverscope(sin6, in6, ifp) 487 struct sockaddr_in6 *sin6; 488 const struct in6_addr *in6; 489 struct ifnet *ifp; 490{ 491 u_int32_t zoneid; 492 493 sin6->sin6_addr = *in6; 494 495 /* 496 * don't try to read *in6 beyond here, since the caller may 497 * ask us to overwrite existing sockaddr_in6 498 */ 499 500 sin6->sin6_scope_id = 0; 501 if (IN6_IS_SCOPE_LINKLOCAL(in6) || IN6_IS_ADDR_MC_INTFACELOCAL(in6)) { 502 /* 503 * KAME assumption: link id == interface id 504 */ 505 zoneid = ntohs(sin6->sin6_addr.s6_addr16[1]); 506 if (zoneid) { 507 /* sanity check */ 508 if (zoneid < 0 || if_index < zoneid) 509 return ENXIO; 510 if (ifp && ifp->if_index != zoneid) 511 return ENXIO; 512 sin6->sin6_addr.s6_addr16[1] = 0; 513 sin6->sin6_scope_id = zoneid; 514 } 515 } 516 517 return 0; 518} 519 520/* 521 * just clear the embedded scope identifier. 522 */ 523void 524in6_clearscope(addr) 525 struct in6_addr *addr; 526{ 527 if (IN6_IS_SCOPE_LINKLOCAL(addr) || IN6_IS_ADDR_MC_INTFACELOCAL(addr)) 528 addr->s6_addr16[1] = 0; 529} 530