1 /*- 2 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. 3 * All rights reserved. 4 * 5 * Redistribution and use in source and binary forms, with or without 6 * modification, are permitted provided that the following conditions 7 * are met: 8 * 1. Redistributions of source code must retain the above copyright 9 * notice, this list of conditions and the following disclaimer. 10 * 2. Redistributions in binary form must reproduce the above copyright 11 * notice, this list of conditions and the following disclaimer in the 12 * documentation and/or other materials provided with the distribution. 13 * 3. Neither the name of the project nor the names of its contributors 14 * may be used to endorse or promote products derived from this software 15 * without specific prior written permission. 16 * 17 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND 18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 20 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE 21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 27 * SUCH DAMAGE. 28 * 29 * $KAME: nd6_rtr.c,v 1.111 2001/04/27 01:37:15 jinmei Exp $ 30 */ 31 32 #include <sys/cdefs.h> 33 __FBSDID("$FreeBSD$"); 34 35 #include "opt_inet.h" 36 #include "opt_inet6.h" 37 38 #include <sys/param.h> 39 #include <sys/systm.h> 40 #include <sys/malloc.h> 41 #include <sys/mbuf.h> 42 #include <sys/socket.h> 43 #include <sys/sockio.h> 44 #include <sys/time.h> 45 #include <sys/kernel.h> 46 #include <sys/lock.h> 47 #include <sys/errno.h> 48 #include <sys/rwlock.h> 49 #include <sys/syslog.h> 50 #include <sys/queue.h> 51 52 #include <net/if.h> 53 #include <net/if_types.h> 54 #include <net/if_dl.h> 55 #include <net/route.h> 56 #include <net/radix.h> 57 #include <net/vnet.h> 58 59 #include <netinet/in.h> 60 #include <net/if_llatbl.h> 61 #include <netinet6/in6_var.h> 62 #include <netinet6/in6_ifattach.h> 63 #include <netinet/ip6.h> 64 #include <netinet6/ip6_var.h> 65 #include <netinet6/nd6.h> 66 #include <netinet/icmp6.h> 67 #include <netinet6/scope6_var.h> 68 69 static int rtpref(struct nd_defrouter *); 70 static struct nd_defrouter *defrtrlist_update(struct nd_defrouter *); 71 static int prelist_update __P((struct nd_prefixctl *, struct nd_defrouter *, 72 struct mbuf *, int)); 73 static struct in6_ifaddr *in6_ifadd(struct nd_prefixctl *, int); 74 static struct nd_pfxrouter *pfxrtr_lookup __P((struct nd_prefix *, 75 struct nd_defrouter *)); 76 static void pfxrtr_add(struct nd_prefix *, struct nd_defrouter *); 77 static void pfxrtr_del(struct nd_pfxrouter *); 78 static struct nd_pfxrouter *find_pfxlist_reachable_router 79 (struct nd_prefix *); 80 static void defrouter_delreq(struct nd_defrouter *); 81 static void nd6_rtmsg(int, struct rtentry *); 82 83 static int in6_init_prefix_ltimes(struct nd_prefix *); 84 static void in6_init_address_ltimes __P((struct nd_prefix *, 85 struct in6_addrlifetime *)); 86 87 static int rt6_deleteroute(struct radix_node *, void *); 88 89 VNET_DECLARE(int, nd6_recalc_reachtm_interval); 90 #define V_nd6_recalc_reachtm_interval VNET(nd6_recalc_reachtm_interval) 91 92 static VNET_DEFINE(struct ifnet *, nd6_defifp); 93 #define V_nd6_defifp VNET(nd6_defifp) 94 95 VNET_DEFINE(int, nd6_defifindex); 96 VNET_DEFINE(int, ip6_use_tempaddr); 97 VNET_DEFINE(int, ip6_desync_factor); 98 VNET_DEFINE(u_int32_t, ip6_temp_preferred_lifetime); 99 VNET_DEFINE(u_int32_t, ip6_temp_valid_lifetime); 100 VNET_DEFINE(int, ip6_temp_regen_advance); 101 102 /* RTPREF_MEDIUM has to be 0! */ 103 #define RTPREF_HIGH 1 104 #define RTPREF_MEDIUM 0 105 #define RTPREF_LOW (-1) 106 #define RTPREF_RESERVED (-2) 107 #define RTPREF_INVALID (-3) /* internal */ 108 109 /* 110 * Receive Router Solicitation Message - just for routers. 111 * Router solicitation/advertisement is mostly managed by userland program 112 * (rtadvd) so here we have no function like nd6_ra_output(). 113 * 114 * Based on RFC 2461 115 */ 116 void 117 nd6_rs_input(struct mbuf *m, int off, int icmp6len) 118 { 119 struct ifnet *ifp = m->m_pkthdr.rcvif; 120 struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *); 121 struct nd_router_solicit *nd_rs; 122 struct in6_addr saddr6 = ip6->ip6_src; 123 char *lladdr = NULL; 124 int lladdrlen = 0; 125 union nd_opts ndopts; 126 char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN]; 127 128 /* If I'm not a router, ignore it. */ 129 if (!V_ip6_forwarding) 130 goto freeit; 131 132 /* Sanity checks */ 133 if (ip6->ip6_hlim != 255) { 134 nd6log((LOG_ERR, 135 "nd6_rs_input: invalid hlim (%d) from %s to %s on %s\n", 136 ip6->ip6_hlim, ip6_sprintf(ip6bufs, &ip6->ip6_src), 137 ip6_sprintf(ip6bufd, &ip6->ip6_dst), if_name(ifp))); 138 goto bad; 139 } 140 141 /* 142 * Don't update the neighbor cache, if src = ::. 143 * This indicates that the src has no IP address assigned yet. 144 */ 145 if (IN6_IS_ADDR_UNSPECIFIED(&saddr6)) 146 goto freeit; 147 148 #ifndef PULLDOWN_TEST 149 IP6_EXTHDR_CHECK(m, off, icmp6len,); 150 nd_rs = (struct nd_router_solicit *)((caddr_t)ip6 + off); 151 #else 152 IP6_EXTHDR_GET(nd_rs, struct nd_router_solicit *, m, off, icmp6len); 153 if (nd_rs == NULL) { 154 ICMP6STAT_INC(icp6s_tooshort); 155 return; 156 } 157 #endif 158 159 icmp6len -= sizeof(*nd_rs); 160 nd6_option_init(nd_rs + 1, icmp6len, &ndopts); 161 if (nd6_options(&ndopts) < 0) { 162 nd6log((LOG_INFO, 163 "nd6_rs_input: invalid ND option, ignored\n")); 164 /* nd6_options have incremented stats */ 165 goto freeit; 166 } 167 168 if (ndopts.nd_opts_src_lladdr) { 169 lladdr = (char *)(ndopts.nd_opts_src_lladdr + 1); 170 lladdrlen = ndopts.nd_opts_src_lladdr->nd_opt_len << 3; 171 } 172 173 if (lladdr && ((ifp->if_addrlen + 2 + 7) & ~7) != lladdrlen) { 174 nd6log((LOG_INFO, 175 "nd6_rs_input: lladdrlen mismatch for %s " 176 "(if %d, RS packet %d)\n", 177 ip6_sprintf(ip6bufs, &saddr6), 178 ifp->if_addrlen, lladdrlen - 2)); 179 goto bad; 180 } 181 182 nd6_cache_lladdr(ifp, &saddr6, lladdr, lladdrlen, ND_ROUTER_SOLICIT, 0); 183 184 freeit: 185 m_freem(m); 186 return; 187 188 bad: 189 ICMP6STAT_INC(icp6s_badrs); 190 m_freem(m); 191 } 192 193 /* 194 * Receive Router Advertisement Message. 195 * 196 * Based on RFC 2461 197 * TODO: on-link bit on prefix information 198 * TODO: ND_RA_FLAG_{OTHER,MANAGED} processing 199 */ 200 void 201 nd6_ra_input(struct mbuf *m, int off, int icmp6len) 202 { 203 struct ifnet *ifp = m->m_pkthdr.rcvif; 204 struct nd_ifinfo *ndi = ND_IFINFO(ifp); 205 struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *); 206 struct nd_router_advert *nd_ra; 207 struct in6_addr saddr6 = ip6->ip6_src; 208 int mcast = 0; 209 union nd_opts ndopts; 210 struct nd_defrouter *dr; 211 char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN]; 212 213 /* 214 * We only accept RAs only when 215 * the node is not a router and 216 * per-interface variable allows RAs on the receiving interface. 217 */ 218 if (V_ip6_forwarding || !(ndi->flags & ND6_IFF_ACCEPT_RTADV)) 219 goto freeit; 220 221 if (ip6->ip6_hlim != 255) { 222 nd6log((LOG_ERR, 223 "nd6_ra_input: invalid hlim (%d) from %s to %s on %s\n", 224 ip6->ip6_hlim, ip6_sprintf(ip6bufs, &ip6->ip6_src), 225 ip6_sprintf(ip6bufd, &ip6->ip6_dst), if_name(ifp))); 226 goto bad; 227 } 228 229 if (!IN6_IS_ADDR_LINKLOCAL(&saddr6)) { 230 nd6log((LOG_ERR, 231 "nd6_ra_input: src %s is not link-local\n", 232 ip6_sprintf(ip6bufs, &saddr6))); 233 goto bad; 234 } 235 236 #ifndef PULLDOWN_TEST 237 IP6_EXTHDR_CHECK(m, off, icmp6len,); 238 nd_ra = (struct nd_router_advert *)((caddr_t)ip6 + off); 239 #else 240 IP6_EXTHDR_GET(nd_ra, struct nd_router_advert *, m, off, icmp6len); 241 if (nd_ra == NULL) { 242 ICMP6STAT_INC(icp6s_tooshort); 243 return; 244 } 245 #endif 246 247 icmp6len -= sizeof(*nd_ra); 248 nd6_option_init(nd_ra + 1, icmp6len, &ndopts); 249 if (nd6_options(&ndopts) < 0) { 250 nd6log((LOG_INFO, 251 "nd6_ra_input: invalid ND option, ignored\n")); 252 /* nd6_options have incremented stats */ 253 goto freeit; 254 } 255 256 { 257 struct nd_defrouter dr0; 258 u_int32_t advreachable = nd_ra->nd_ra_reachable; 259 260 /* remember if this is a multicasted advertisement */ 261 if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) 262 mcast = 1; 263 264 bzero(&dr0, sizeof(dr0)); 265 dr0.rtaddr = saddr6; 266 dr0.flags = nd_ra->nd_ra_flags_reserved; 267 dr0.rtlifetime = ntohs(nd_ra->nd_ra_router_lifetime); 268 dr0.expire = time_second + dr0.rtlifetime; 269 dr0.ifp = ifp; 270 /* unspecified or not? (RFC 2461 6.3.4) */ 271 if (advreachable) { 272 advreachable = ntohl(advreachable); 273 if (advreachable <= MAX_REACHABLE_TIME && 274 ndi->basereachable != advreachable) { 275 ndi->basereachable = advreachable; 276 ndi->reachable = ND_COMPUTE_RTIME(ndi->basereachable); 277 ndi->recalctm = V_nd6_recalc_reachtm_interval; /* reset */ 278 } 279 } 280 if (nd_ra->nd_ra_retransmit) 281 ndi->retrans = ntohl(nd_ra->nd_ra_retransmit); 282 if (nd_ra->nd_ra_curhoplimit) 283 ndi->chlim = nd_ra->nd_ra_curhoplimit; 284 dr = defrtrlist_update(&dr0); 285 } 286 287 /* 288 * prefix 289 */ 290 if (ndopts.nd_opts_pi) { 291 struct nd_opt_hdr *pt; 292 struct nd_opt_prefix_info *pi = NULL; 293 struct nd_prefixctl pr; 294 295 for (pt = (struct nd_opt_hdr *)ndopts.nd_opts_pi; 296 pt <= (struct nd_opt_hdr *)ndopts.nd_opts_pi_end; 297 pt = (struct nd_opt_hdr *)((caddr_t)pt + 298 (pt->nd_opt_len << 3))) { 299 if (pt->nd_opt_type != ND_OPT_PREFIX_INFORMATION) 300 continue; 301 pi = (struct nd_opt_prefix_info *)pt; 302 303 if (pi->nd_opt_pi_len != 4) { 304 nd6log((LOG_INFO, 305 "nd6_ra_input: invalid option " 306 "len %d for prefix information option, " 307 "ignored\n", pi->nd_opt_pi_len)); 308 continue; 309 } 310 311 if (128 < pi->nd_opt_pi_prefix_len) { 312 nd6log((LOG_INFO, 313 "nd6_ra_input: invalid prefix " 314 "len %d for prefix information option, " 315 "ignored\n", pi->nd_opt_pi_prefix_len)); 316 continue; 317 } 318 319 if (IN6_IS_ADDR_MULTICAST(&pi->nd_opt_pi_prefix) 320 || IN6_IS_ADDR_LINKLOCAL(&pi->nd_opt_pi_prefix)) { 321 nd6log((LOG_INFO, 322 "nd6_ra_input: invalid prefix " 323 "%s, ignored\n", 324 ip6_sprintf(ip6bufs, 325 &pi->nd_opt_pi_prefix))); 326 continue; 327 } 328 329 bzero(&pr, sizeof(pr)); 330 pr.ndpr_prefix.sin6_family = AF_INET6; 331 pr.ndpr_prefix.sin6_len = sizeof(pr.ndpr_prefix); 332 pr.ndpr_prefix.sin6_addr = pi->nd_opt_pi_prefix; 333 pr.ndpr_ifp = (struct ifnet *)m->m_pkthdr.rcvif; 334 335 pr.ndpr_raf_onlink = (pi->nd_opt_pi_flags_reserved & 336 ND_OPT_PI_FLAG_ONLINK) ? 1 : 0; 337 pr.ndpr_raf_auto = (pi->nd_opt_pi_flags_reserved & 338 ND_OPT_PI_FLAG_AUTO) ? 1 : 0; 339 pr.ndpr_plen = pi->nd_opt_pi_prefix_len; 340 pr.ndpr_vltime = ntohl(pi->nd_opt_pi_valid_time); 341 pr.ndpr_pltime = ntohl(pi->nd_opt_pi_preferred_time); 342 (void)prelist_update(&pr, dr, m, mcast); 343 } 344 } 345 346 /* 347 * MTU 348 */ 349 if (ndopts.nd_opts_mtu && ndopts.nd_opts_mtu->nd_opt_mtu_len == 1) { 350 u_long mtu; 351 u_long maxmtu; 352 353 mtu = (u_long)ntohl(ndopts.nd_opts_mtu->nd_opt_mtu_mtu); 354 355 /* lower bound */ 356 if (mtu < IPV6_MMTU) { 357 nd6log((LOG_INFO, "nd6_ra_input: bogus mtu option " 358 "mtu=%lu sent from %s, ignoring\n", 359 mtu, ip6_sprintf(ip6bufs, &ip6->ip6_src))); 360 goto skip; 361 } 362 363 /* upper bound */ 364 maxmtu = (ndi->maxmtu && ndi->maxmtu < ifp->if_mtu) 365 ? ndi->maxmtu : ifp->if_mtu; 366 if (mtu <= maxmtu) { 367 int change = (ndi->linkmtu != mtu); 368 369 ndi->linkmtu = mtu; 370 if (change) /* in6_maxmtu may change */ 371 in6_setmaxmtu(); 372 } else { 373 nd6log((LOG_INFO, "nd6_ra_input: bogus mtu " 374 "mtu=%lu sent from %s; " 375 "exceeds maxmtu %lu, ignoring\n", 376 mtu, ip6_sprintf(ip6bufs, &ip6->ip6_src), maxmtu)); 377 } 378 } 379 380 skip: 381 382 /* 383 * Source link layer address 384 */ 385 { 386 char *lladdr = NULL; 387 int lladdrlen = 0; 388 389 if (ndopts.nd_opts_src_lladdr) { 390 lladdr = (char *)(ndopts.nd_opts_src_lladdr + 1); 391 lladdrlen = ndopts.nd_opts_src_lladdr->nd_opt_len << 3; 392 } 393 394 if (lladdr && ((ifp->if_addrlen + 2 + 7) & ~7) != lladdrlen) { 395 nd6log((LOG_INFO, 396 "nd6_ra_input: lladdrlen mismatch for %s " 397 "(if %d, RA packet %d)\n", ip6_sprintf(ip6bufs, &saddr6), 398 ifp->if_addrlen, lladdrlen - 2)); 399 goto bad; 400 } 401 402 nd6_cache_lladdr(ifp, &saddr6, lladdr, 403 lladdrlen, ND_ROUTER_ADVERT, 0); 404 405 /* 406 * Installing a link-layer address might change the state of the 407 * router's neighbor cache, which might also affect our on-link 408 * detection of adveritsed prefixes. 409 */ 410 pfxlist_onlink_check(); 411 } 412 413 freeit: 414 m_freem(m); 415 return; 416 417 bad: 418 ICMP6STAT_INC(icp6s_badra); 419 m_freem(m); 420 } 421 422 /* 423 * default router list proccessing sub routines 424 */ 425 426 /* tell the change to user processes watching the routing socket. */ 427 static void 428 nd6_rtmsg(int cmd, struct rtentry *rt) 429 { 430 struct rt_addrinfo info; 431 struct ifnet *ifp; 432 struct ifaddr *ifa; 433 434 bzero((caddr_t)&info, sizeof(info)); 435 info.rti_info[RTAX_DST] = rt_key(rt); 436 info.rti_info[RTAX_GATEWAY] = rt->rt_gateway; 437 info.rti_info[RTAX_NETMASK] = rt_mask(rt); 438 ifp = rt->rt_ifp; 439 if (ifp != NULL) { 440 IF_ADDR_LOCK(ifp); 441 ifa = TAILQ_FIRST(&ifp->if_addrhead); 442 info.rti_info[RTAX_IFP] = ifa->ifa_addr; 443 ifa_ref(ifa); 444 IF_ADDR_UNLOCK(ifp); 445 info.rti_info[RTAX_IFA] = rt->rt_ifa->ifa_addr; 446 } else 447 ifa = NULL; 448 449 rt_missmsg(cmd, &info, rt->rt_flags, 0); 450 if (ifa != NULL) 451 ifa_free(ifa); 452 } 453 454 void 455 defrouter_addreq(struct nd_defrouter *new) 456 { 457 struct sockaddr_in6 def, mask, gate; 458 struct rtentry *newrt = NULL; 459 int s; 460 int error; 461 462 bzero(&def, sizeof(def)); 463 bzero(&mask, sizeof(mask)); 464 bzero(&gate, sizeof(gate)); 465 466 def.sin6_len = mask.sin6_len = gate.sin6_len = 467 sizeof(struct sockaddr_in6); 468 def.sin6_family = gate.sin6_family = AF_INET6; 469 gate.sin6_addr = new->rtaddr; 470 471 s = splnet(); 472 error = rtrequest(RTM_ADD, (struct sockaddr *)&def, 473 (struct sockaddr *)&gate, (struct sockaddr *)&mask, 474 RTF_GATEWAY, &newrt); 475 if (newrt) { 476 nd6_rtmsg(RTM_ADD, newrt); /* tell user process */ 477 RTFREE(newrt); 478 } 479 if (error == 0) 480 new->installed = 1; 481 splx(s); 482 return; 483 } 484 485 struct nd_defrouter * 486 defrouter_lookup(struct in6_addr *addr, struct ifnet *ifp) 487 { 488 struct nd_defrouter *dr; 489 490 for (dr = TAILQ_FIRST(&V_nd_defrouter); dr; 491 dr = TAILQ_NEXT(dr, dr_entry)) { 492 if (dr->ifp == ifp && IN6_ARE_ADDR_EQUAL(addr, &dr->rtaddr)) 493 return (dr); 494 } 495 496 return (NULL); /* search failed */ 497 } 498 499 /* 500 * Remove the default route for a given router. 501 * This is just a subroutine function for defrouter_select(), and should 502 * not be called from anywhere else. 503 */ 504 static void 505 defrouter_delreq(struct nd_defrouter *dr) 506 { 507 struct sockaddr_in6 def, mask, gate; 508 struct rtentry *oldrt = NULL; 509 510 bzero(&def, sizeof(def)); 511 bzero(&mask, sizeof(mask)); 512 bzero(&gate, sizeof(gate)); 513 514 def.sin6_len = mask.sin6_len = gate.sin6_len = 515 sizeof(struct sockaddr_in6); 516 def.sin6_family = gate.sin6_family = AF_INET6; 517 gate.sin6_addr = dr->rtaddr; 518 519 rtrequest(RTM_DELETE, (struct sockaddr *)&def, 520 (struct sockaddr *)&gate, 521 (struct sockaddr *)&mask, RTF_GATEWAY, &oldrt); 522 if (oldrt) { 523 nd6_rtmsg(RTM_DELETE, oldrt); 524 RTFREE(oldrt); 525 } 526 527 dr->installed = 0; 528 } 529 530 /* 531 * remove all default routes from default router list 532 */ 533 void 534 defrouter_reset(void) 535 { 536 struct nd_defrouter *dr; 537 538 for (dr = TAILQ_FIRST(&V_nd_defrouter); dr; 539 dr = TAILQ_NEXT(dr, dr_entry)) 540 defrouter_delreq(dr); 541 542 /* 543 * XXX should we also nuke any default routers in the kernel, by 544 * going through them by rtalloc1()? 545 */ 546 } 547 548 void 549 defrtrlist_del(struct nd_defrouter *dr) 550 { 551 struct nd_defrouter *deldr = NULL; 552 struct nd_prefix *pr; 553 554 /* 555 * Flush all the routing table entries that use the router 556 * as a next hop. 557 */ 558 if (!V_ip6_forwarding) 559 rt6_flush(&dr->rtaddr, dr->ifp); 560 561 if (dr->installed) { 562 deldr = dr; 563 defrouter_delreq(dr); 564 } 565 TAILQ_REMOVE(&V_nd_defrouter, dr, dr_entry); 566 567 /* 568 * Also delete all the pointers to the router in each prefix lists. 569 */ 570 for (pr = V_nd_prefix.lh_first; pr; pr = pr->ndpr_next) { 571 struct nd_pfxrouter *pfxrtr; 572 if ((pfxrtr = pfxrtr_lookup(pr, dr)) != NULL) 573 pfxrtr_del(pfxrtr); 574 } 575 pfxlist_onlink_check(); 576 577 /* 578 * If the router is the primary one, choose a new one. 579 * Note that defrouter_select() will remove the current gateway 580 * from the routing table. 581 */ 582 if (deldr) 583 defrouter_select(); 584 585 free(dr, M_IP6NDP); 586 } 587 588 /* 589 * Default Router Selection according to Section 6.3.6 of RFC 2461 and 590 * draft-ietf-ipngwg-router-selection: 591 * 1) Routers that are reachable or probably reachable should be preferred. 592 * If we have more than one (probably) reachable router, prefer ones 593 * with the highest router preference. 594 * 2) When no routers on the list are known to be reachable or 595 * probably reachable, routers SHOULD be selected in a round-robin 596 * fashion, regardless of router preference values. 597 * 3) If the Default Router List is empty, assume that all 598 * destinations are on-link. 599 * 600 * We assume nd_defrouter is sorted by router preference value. 601 * Since the code below covers both with and without router preference cases, 602 * we do not need to classify the cases by ifdef. 603 * 604 * At this moment, we do not try to install more than one default router, 605 * even when the multipath routing is available, because we're not sure about 606 * the benefits for stub hosts comparing to the risk of making the code 607 * complicated and the possibility of introducing bugs. 608 */ 609 void 610 defrouter_select(void) 611 { 612 int s = splnet(); 613 struct nd_defrouter *dr, *selected_dr = NULL, *installed_dr = NULL; 614 struct llentry *ln = NULL; 615 616 /* 617 * This function should be called only when acting as an autoconfigured 618 * host. Although the remaining part of this function is not effective 619 * if the node is not an autoconfigured host, we explicitly exclude 620 * such cases here for safety. 621 */ 622 if (V_ip6_forwarding) { 623 nd6log((LOG_WARNING, 624 "defrouter_select: called unexpectedly (forwarding=%d)\n", 625 V_ip6_forwarding)); 626 splx(s); 627 return; 628 } 629 630 /* 631 * Let's handle easy case (3) first: 632 * If default router list is empty, there's nothing to be done. 633 */ 634 if (!TAILQ_FIRST(&V_nd_defrouter)) { 635 splx(s); 636 return; 637 } 638 639 /* 640 * Search for a (probably) reachable router from the list. 641 * We just pick up the first reachable one (if any), assuming that 642 * the ordering rule of the list described in defrtrlist_update(). 643 */ 644 for (dr = TAILQ_FIRST(&V_nd_defrouter); dr; 645 dr = TAILQ_NEXT(dr, dr_entry)) { 646 IF_AFDATA_LOCK(dr->ifp); 647 if (selected_dr == NULL && 648 (ln = nd6_lookup(&dr->rtaddr, 0, dr->ifp)) && 649 ND6_IS_LLINFO_PROBREACH(ln)) { 650 selected_dr = dr; 651 } 652 IF_AFDATA_UNLOCK(dr->ifp); 653 if (ln != NULL) { 654 LLE_RUNLOCK(ln); 655 ln = NULL; 656 } 657 658 if (dr->installed && installed_dr == NULL) 659 installed_dr = dr; 660 else if (dr->installed && installed_dr) { 661 /* this should not happen. warn for diagnosis. */ 662 log(LOG_ERR, "defrouter_select: more than one router" 663 " is installed\n"); 664 } 665 } 666 /* 667 * If none of the default routers was found to be reachable, 668 * round-robin the list regardless of preference. 669 * Otherwise, if we have an installed router, check if the selected 670 * (reachable) router should really be preferred to the installed one. 671 * We only prefer the new router when the old one is not reachable 672 * or when the new one has a really higher preference value. 673 */ 674 if (selected_dr == NULL) { 675 if (installed_dr == NULL || !TAILQ_NEXT(installed_dr, dr_entry)) 676 selected_dr = TAILQ_FIRST(&V_nd_defrouter); 677 else 678 selected_dr = TAILQ_NEXT(installed_dr, dr_entry); 679 } else if (installed_dr) { 680 IF_AFDATA_LOCK(installed_dr->ifp); 681 if ((ln = nd6_lookup(&installed_dr->rtaddr, 0, installed_dr->ifp)) && 682 ND6_IS_LLINFO_PROBREACH(ln) && 683 rtpref(selected_dr) <= rtpref(installed_dr)) { 684 selected_dr = installed_dr; 685 } 686 IF_AFDATA_UNLOCK(installed_dr->ifp); 687 if (ln != NULL) 688 LLE_RUNLOCK(ln); 689 } 690 691 /* 692 * If the selected router is different than the installed one, 693 * remove the installed router and install the selected one. 694 * Note that the selected router is never NULL here. 695 */ 696 if (installed_dr != selected_dr) { 697 if (installed_dr) 698 defrouter_delreq(installed_dr); 699 defrouter_addreq(selected_dr); 700 } 701 702 splx(s); 703 return; 704 } 705 706 /* 707 * for default router selection 708 * regards router-preference field as a 2-bit signed integer 709 */ 710 static int 711 rtpref(struct nd_defrouter *dr) 712 { 713 switch (dr->flags & ND_RA_FLAG_RTPREF_MASK) { 714 case ND_RA_FLAG_RTPREF_HIGH: 715 return (RTPREF_HIGH); 716 case ND_RA_FLAG_RTPREF_MEDIUM: 717 case ND_RA_FLAG_RTPREF_RSV: 718 return (RTPREF_MEDIUM); 719 case ND_RA_FLAG_RTPREF_LOW: 720 return (RTPREF_LOW); 721 default: 722 /* 723 * This case should never happen. If it did, it would mean a 724 * serious bug of kernel internal. We thus always bark here. 725 * Or, can we even panic? 726 */ 727 log(LOG_ERR, "rtpref: impossible RA flag %x\n", dr->flags); 728 return (RTPREF_INVALID); 729 } 730 /* NOTREACHED */ 731 } 732 733 static struct nd_defrouter * 734 defrtrlist_update(struct nd_defrouter *new) 735 { 736 struct nd_defrouter *dr, *n; 737 int s = splnet(); 738 739 if ((dr = defrouter_lookup(&new->rtaddr, new->ifp)) != NULL) { 740 /* entry exists */ 741 if (new->rtlifetime == 0) { 742 defrtrlist_del(dr); 743 dr = NULL; 744 } else { 745 int oldpref = rtpref(dr); 746 747 /* override */ 748 dr->flags = new->flags; /* xxx flag check */ 749 dr->rtlifetime = new->rtlifetime; 750 dr->expire = new->expire; 751 752 /* 753 * If the preference does not change, there's no need 754 * to sort the entries. 755 */ 756 if (rtpref(new) == oldpref) { 757 splx(s); 758 return (dr); 759 } 760 761 /* 762 * preferred router may be changed, so relocate 763 * this router. 764 * XXX: calling TAILQ_REMOVE directly is a bad manner. 765 * However, since defrtrlist_del() has many side 766 * effects, we intentionally do so here. 767 * defrouter_select() below will handle routing 768 * changes later. 769 */ 770 TAILQ_REMOVE(&V_nd_defrouter, dr, dr_entry); 771 n = dr; 772 goto insert; 773 } 774 splx(s); 775 return (dr); 776 } 777 778 /* entry does not exist */ 779 if (new->rtlifetime == 0) { 780 splx(s); 781 return (NULL); 782 } 783 784 n = (struct nd_defrouter *)malloc(sizeof(*n), M_IP6NDP, M_NOWAIT); 785 if (n == NULL) { 786 splx(s); 787 return (NULL); 788 } 789 bzero(n, sizeof(*n)); 790 *n = *new; 791 792 insert: 793 /* 794 * Insert the new router in the Default Router List; 795 * The Default Router List should be in the descending order 796 * of router-preferece. Routers with the same preference are 797 * sorted in the arriving time order. 798 */ 799 800 /* insert at the end of the group */ 801 for (dr = TAILQ_FIRST(&V_nd_defrouter); dr; 802 dr = TAILQ_NEXT(dr, dr_entry)) { 803 if (rtpref(n) > rtpref(dr)) 804 break; 805 } 806 if (dr) 807 TAILQ_INSERT_BEFORE(dr, n, dr_entry); 808 else 809 TAILQ_INSERT_TAIL(&V_nd_defrouter, n, dr_entry); 810 811 defrouter_select(); 812 813 splx(s); 814 815 return (n); 816 } 817 818 static struct nd_pfxrouter * 819 pfxrtr_lookup(struct nd_prefix *pr, struct nd_defrouter *dr) 820 { 821 struct nd_pfxrouter *search; 822 823 for (search = pr->ndpr_advrtrs.lh_first; search; search = search->pfr_next) { 824 if (search->router == dr) 825 break; 826 } 827 828 return (search); 829 } 830 831 static void 832 pfxrtr_add(struct nd_prefix *pr, struct nd_defrouter *dr) 833 { 834 struct nd_pfxrouter *new; 835 836 new = (struct nd_pfxrouter *)malloc(sizeof(*new), M_IP6NDP, M_NOWAIT); 837 if (new == NULL) 838 return; 839 bzero(new, sizeof(*new)); 840 new->router = dr; 841 842 LIST_INSERT_HEAD(&pr->ndpr_advrtrs, new, pfr_entry); 843 844 pfxlist_onlink_check(); 845 } 846 847 static void 848 pfxrtr_del(struct nd_pfxrouter *pfr) 849 { 850 LIST_REMOVE(pfr, pfr_entry); 851 free(pfr, M_IP6NDP); 852 } 853 854 struct nd_prefix * 855 nd6_prefix_lookup(struct nd_prefixctl *key) 856 { 857 struct nd_prefix *search; 858 859 for (search = V_nd_prefix.lh_first; 860 search; search = search->ndpr_next) { 861 if (key->ndpr_ifp == search->ndpr_ifp && 862 key->ndpr_plen == search->ndpr_plen && 863 in6_are_prefix_equal(&key->ndpr_prefix.sin6_addr, 864 &search->ndpr_prefix.sin6_addr, key->ndpr_plen)) { 865 break; 866 } 867 } 868 869 return (search); 870 } 871 872 int 873 nd6_prelist_add(struct nd_prefixctl *pr, struct nd_defrouter *dr, 874 struct nd_prefix **newp) 875 { 876 struct nd_prefix *new = NULL; 877 int error = 0; 878 int i, s; 879 char ip6buf[INET6_ADDRSTRLEN]; 880 881 new = (struct nd_prefix *)malloc(sizeof(*new), M_IP6NDP, M_NOWAIT); 882 if (new == NULL) 883 return(ENOMEM); 884 bzero(new, sizeof(*new)); 885 new->ndpr_ifp = pr->ndpr_ifp; 886 new->ndpr_prefix = pr->ndpr_prefix; 887 new->ndpr_plen = pr->ndpr_plen; 888 new->ndpr_vltime = pr->ndpr_vltime; 889 new->ndpr_pltime = pr->ndpr_pltime; 890 new->ndpr_flags = pr->ndpr_flags; 891 if ((error = in6_init_prefix_ltimes(new)) != 0) { 892 free(new, M_IP6NDP); 893 return(error); 894 } 895 new->ndpr_lastupdate = time_second; 896 if (newp != NULL) 897 *newp = new; 898 899 /* initialization */ 900 LIST_INIT(&new->ndpr_advrtrs); 901 in6_prefixlen2mask(&new->ndpr_mask, new->ndpr_plen); 902 /* make prefix in the canonical form */ 903 for (i = 0; i < 4; i++) 904 new->ndpr_prefix.sin6_addr.s6_addr32[i] &= 905 new->ndpr_mask.s6_addr32[i]; 906 907 s = splnet(); 908 /* link ndpr_entry to nd_prefix list */ 909 LIST_INSERT_HEAD(&V_nd_prefix, new, ndpr_entry); 910 splx(s); 911 912 /* ND_OPT_PI_FLAG_ONLINK processing */ 913 if (new->ndpr_raf_onlink) { 914 int e; 915 916 if ((e = nd6_prefix_onlink(new)) != 0) { 917 nd6log((LOG_ERR, "nd6_prelist_add: failed to make " 918 "the prefix %s/%d on-link on %s (errno=%d)\n", 919 ip6_sprintf(ip6buf, &pr->ndpr_prefix.sin6_addr), 920 pr->ndpr_plen, if_name(pr->ndpr_ifp), e)); 921 /* proceed anyway. XXX: is it correct? */ 922 } 923 } 924 925 if (dr) 926 pfxrtr_add(new, dr); 927 928 return 0; 929 } 930 931 void 932 prelist_remove(struct nd_prefix *pr) 933 { 934 struct nd_pfxrouter *pfr, *next; 935 int e, s; 936 char ip6buf[INET6_ADDRSTRLEN]; 937 938 /* make sure to invalidate the prefix until it is really freed. */ 939 pr->ndpr_vltime = 0; 940 pr->ndpr_pltime = 0; 941 942 /* 943 * Though these flags are now meaningless, we'd rather keep the value 944 * of pr->ndpr_raf_onlink and pr->ndpr_raf_auto not to confuse users 945 * when executing "ndp -p". 946 */ 947 948 if ((pr->ndpr_stateflags & NDPRF_ONLINK) != 0 && 949 (e = nd6_prefix_offlink(pr)) != 0) { 950 nd6log((LOG_ERR, "prelist_remove: failed to make %s/%d offlink " 951 "on %s, errno=%d\n", 952 ip6_sprintf(ip6buf, &pr->ndpr_prefix.sin6_addr), 953 pr->ndpr_plen, if_name(pr->ndpr_ifp), e)); 954 /* what should we do? */ 955 } 956 957 if (pr->ndpr_refcnt > 0) 958 return; /* notice here? */ 959 960 s = splnet(); 961 962 /* unlink ndpr_entry from nd_prefix list */ 963 LIST_REMOVE(pr, ndpr_entry); 964 965 /* free list of routers that adversed the prefix */ 966 for (pfr = pr->ndpr_advrtrs.lh_first; pfr; pfr = next) { 967 next = pfr->pfr_next; 968 969 free(pfr, M_IP6NDP); 970 } 971 splx(s); 972 973 free(pr, M_IP6NDP); 974 975 pfxlist_onlink_check(); 976 } 977 978 /* 979 * dr - may be NULL 980 */ 981 982 static int 983 prelist_update(struct nd_prefixctl *new, struct nd_defrouter *dr, 984 struct mbuf *m, int mcast) 985 { 986 struct in6_ifaddr *ia6 = NULL, *ia6_match = NULL; 987 struct ifaddr *ifa; 988 struct ifnet *ifp = new->ndpr_ifp; 989 struct nd_prefix *pr; 990 int s = splnet(); 991 int error = 0; 992 int newprefix = 0; 993 int auth; 994 struct in6_addrlifetime lt6_tmp; 995 char ip6buf[INET6_ADDRSTRLEN]; 996 997 auth = 0; 998 if (m) { 999 /* 1000 * Authenticity for NA consists authentication for 1001 * both IP header and IP datagrams, doesn't it ? 1002 */ 1003 #if defined(M_AUTHIPHDR) && defined(M_AUTHIPDGM) 1004 auth = ((m->m_flags & M_AUTHIPHDR) && 1005 (m->m_flags & M_AUTHIPDGM)); 1006 #endif 1007 } 1008 1009 if ((pr = nd6_prefix_lookup(new)) != NULL) { 1010 /* 1011 * nd6_prefix_lookup() ensures that pr and new have the same 1012 * prefix on a same interface. 1013 */ 1014 1015 /* 1016 * Update prefix information. Note that the on-link (L) bit 1017 * and the autonomous (A) bit should NOT be changed from 1 1018 * to 0. 1019 */ 1020 if (new->ndpr_raf_onlink == 1) 1021 pr->ndpr_raf_onlink = 1; 1022 if (new->ndpr_raf_auto == 1) 1023 pr->ndpr_raf_auto = 1; 1024 if (new->ndpr_raf_onlink) { 1025 pr->ndpr_vltime = new->ndpr_vltime; 1026 pr->ndpr_pltime = new->ndpr_pltime; 1027 (void)in6_init_prefix_ltimes(pr); /* XXX error case? */ 1028 pr->ndpr_lastupdate = time_second; 1029 } 1030 1031 if (new->ndpr_raf_onlink && 1032 (pr->ndpr_stateflags & NDPRF_ONLINK) == 0) { 1033 int e; 1034 1035 if ((e = nd6_prefix_onlink(pr)) != 0) { 1036 nd6log((LOG_ERR, 1037 "prelist_update: failed to make " 1038 "the prefix %s/%d on-link on %s " 1039 "(errno=%d)\n", 1040 ip6_sprintf(ip6buf, 1041 &pr->ndpr_prefix.sin6_addr), 1042 pr->ndpr_plen, if_name(pr->ndpr_ifp), e)); 1043 /* proceed anyway. XXX: is it correct? */ 1044 } 1045 } 1046 1047 if (dr && pfxrtr_lookup(pr, dr) == NULL) 1048 pfxrtr_add(pr, dr); 1049 } else { 1050 struct nd_prefix *newpr = NULL; 1051 1052 newprefix = 1; 1053 1054 if (new->ndpr_vltime == 0) 1055 goto end; 1056 if (new->ndpr_raf_onlink == 0 && new->ndpr_raf_auto == 0) 1057 goto end; 1058 1059 error = nd6_prelist_add(new, dr, &newpr); 1060 if (error != 0 || newpr == NULL) { 1061 nd6log((LOG_NOTICE, "prelist_update: " 1062 "nd6_prelist_add failed for %s/%d on %s " 1063 "errno=%d, returnpr=%p\n", 1064 ip6_sprintf(ip6buf, &new->ndpr_prefix.sin6_addr), 1065 new->ndpr_plen, if_name(new->ndpr_ifp), 1066 error, newpr)); 1067 goto end; /* we should just give up in this case. */ 1068 } 1069 1070 /* 1071 * XXX: from the ND point of view, we can ignore a prefix 1072 * with the on-link bit being zero. However, we need a 1073 * prefix structure for references from autoconfigured 1074 * addresses. Thus, we explicitly make sure that the prefix 1075 * itself expires now. 1076 */ 1077 if (newpr->ndpr_raf_onlink == 0) { 1078 newpr->ndpr_vltime = 0; 1079 newpr->ndpr_pltime = 0; 1080 in6_init_prefix_ltimes(newpr); 1081 } 1082 1083 pr = newpr; 1084 } 1085 1086 /* 1087 * Address autoconfiguration based on Section 5.5.3 of RFC 2462. 1088 * Note that pr must be non NULL at this point. 1089 */ 1090 1091 /* 5.5.3 (a). Ignore the prefix without the A bit set. */ 1092 if (!new->ndpr_raf_auto) 1093 goto end; 1094 1095 /* 1096 * 5.5.3 (b). the link-local prefix should have been ignored in 1097 * nd6_ra_input. 1098 */ 1099 1100 /* 5.5.3 (c). Consistency check on lifetimes: pltime <= vltime. */ 1101 if (new->ndpr_pltime > new->ndpr_vltime) { 1102 error = EINVAL; /* XXX: won't be used */ 1103 goto end; 1104 } 1105 1106 /* 1107 * 5.5.3 (d). If the prefix advertised is not equal to the prefix of 1108 * an address configured by stateless autoconfiguration already in the 1109 * list of addresses associated with the interface, and the Valid 1110 * Lifetime is not 0, form an address. We first check if we have 1111 * a matching prefix. 1112 * Note: we apply a clarification in rfc2462bis-02 here. We only 1113 * consider autoconfigured addresses while RFC2462 simply said 1114 * "address". 1115 */ 1116 IF_ADDR_LOCK(ifp); 1117 TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { 1118 struct in6_ifaddr *ifa6; 1119 u_int32_t remaininglifetime; 1120 1121 if (ifa->ifa_addr->sa_family != AF_INET6) 1122 continue; 1123 1124 ifa6 = (struct in6_ifaddr *)ifa; 1125 1126 /* 1127 * We only consider autoconfigured addresses as per rfc2462bis. 1128 */ 1129 if (!(ifa6->ia6_flags & IN6_IFF_AUTOCONF)) 1130 continue; 1131 1132 /* 1133 * Spec is not clear here, but I believe we should concentrate 1134 * on unicast (i.e. not anycast) addresses. 1135 * XXX: other ia6_flags? detached or duplicated? 1136 */ 1137 if ((ifa6->ia6_flags & IN6_IFF_ANYCAST) != 0) 1138 continue; 1139 1140 /* 1141 * Ignore the address if it is not associated with a prefix 1142 * or is associated with a prefix that is different from this 1143 * one. (pr is never NULL here) 1144 */ 1145 if (ifa6->ia6_ndpr != pr) 1146 continue; 1147 1148 if (ia6_match == NULL) /* remember the first one */ 1149 ia6_match = ifa6; 1150 1151 /* 1152 * An already autoconfigured address matched. Now that we 1153 * are sure there is at least one matched address, we can 1154 * proceed to 5.5.3. (e): update the lifetimes according to the 1155 * "two hours" rule and the privacy extension. 1156 * We apply some clarifications in rfc2462bis: 1157 * - use remaininglifetime instead of storedlifetime as a 1158 * variable name 1159 * - remove the dead code in the "two-hour" rule 1160 */ 1161 #define TWOHOUR (120*60) 1162 lt6_tmp = ifa6->ia6_lifetime; 1163 1164 if (lt6_tmp.ia6t_vltime == ND6_INFINITE_LIFETIME) 1165 remaininglifetime = ND6_INFINITE_LIFETIME; 1166 else if (time_second - ifa6->ia6_updatetime > 1167 lt6_tmp.ia6t_vltime) { 1168 /* 1169 * The case of "invalid" address. We should usually 1170 * not see this case. 1171 */ 1172 remaininglifetime = 0; 1173 } else 1174 remaininglifetime = lt6_tmp.ia6t_vltime - 1175 (time_second - ifa6->ia6_updatetime); 1176 1177 /* when not updating, keep the current stored lifetime. */ 1178 lt6_tmp.ia6t_vltime = remaininglifetime; 1179 1180 if (TWOHOUR < new->ndpr_vltime || 1181 remaininglifetime < new->ndpr_vltime) { 1182 lt6_tmp.ia6t_vltime = new->ndpr_vltime; 1183 } else if (remaininglifetime <= TWOHOUR) { 1184 if (auth) { 1185 lt6_tmp.ia6t_vltime = new->ndpr_vltime; 1186 } 1187 } else { 1188 /* 1189 * new->ndpr_vltime <= TWOHOUR && 1190 * TWOHOUR < remaininglifetime 1191 */ 1192 lt6_tmp.ia6t_vltime = TWOHOUR; 1193 } 1194 1195 /* The 2 hour rule is not imposed for preferred lifetime. */ 1196 lt6_tmp.ia6t_pltime = new->ndpr_pltime; 1197 1198 in6_init_address_ltimes(pr, <6_tmp); 1199 1200 /* 1201 * We need to treat lifetimes for temporary addresses 1202 * differently, according to 1203 * draft-ietf-ipv6-privacy-addrs-v2-01.txt 3.3 (1); 1204 * we only update the lifetimes when they are in the maximum 1205 * intervals. 1206 */ 1207 if ((ifa6->ia6_flags & IN6_IFF_TEMPORARY) != 0) { 1208 u_int32_t maxvltime, maxpltime; 1209 1210 if (V_ip6_temp_valid_lifetime > 1211 (u_int32_t)((time_second - ifa6->ia6_createtime) + 1212 V_ip6_desync_factor)) { 1213 maxvltime = V_ip6_temp_valid_lifetime - 1214 (time_second - ifa6->ia6_createtime) - 1215 V_ip6_desync_factor; 1216 } else 1217 maxvltime = 0; 1218 if (V_ip6_temp_preferred_lifetime > 1219 (u_int32_t)((time_second - ifa6->ia6_createtime) + 1220 V_ip6_desync_factor)) { 1221 maxpltime = V_ip6_temp_preferred_lifetime - 1222 (time_second - ifa6->ia6_createtime) - 1223 V_ip6_desync_factor; 1224 } else 1225 maxpltime = 0; 1226 1227 if (lt6_tmp.ia6t_vltime == ND6_INFINITE_LIFETIME || 1228 lt6_tmp.ia6t_vltime > maxvltime) { 1229 lt6_tmp.ia6t_vltime = maxvltime; 1230 } 1231 if (lt6_tmp.ia6t_pltime == ND6_INFINITE_LIFETIME || 1232 lt6_tmp.ia6t_pltime > maxpltime) { 1233 lt6_tmp.ia6t_pltime = maxpltime; 1234 } 1235 } 1236 ifa6->ia6_lifetime = lt6_tmp; 1237 ifa6->ia6_updatetime = time_second; 1238 } 1239 IF_ADDR_UNLOCK(ifp); 1240 if (ia6_match == NULL && new->ndpr_vltime) { 1241 int ifidlen; 1242 1243 /* 1244 * 5.5.3 (d) (continued) 1245 * No address matched and the valid lifetime is non-zero. 1246 * Create a new address. 1247 */ 1248 1249 /* 1250 * Prefix Length check: 1251 * If the sum of the prefix length and interface identifier 1252 * length does not equal 128 bits, the Prefix Information 1253 * option MUST be ignored. The length of the interface 1254 * identifier is defined in a separate link-type specific 1255 * document. 1256 */ 1257 ifidlen = in6_if2idlen(ifp); 1258 if (ifidlen < 0) { 1259 /* this should not happen, so we always log it. */ 1260 log(LOG_ERR, "prelist_update: IFID undefined (%s)\n", 1261 if_name(ifp)); 1262 goto end; 1263 } 1264 if (ifidlen + pr->ndpr_plen != 128) { 1265 nd6log((LOG_INFO, 1266 "prelist_update: invalid prefixlen " 1267 "%d for %s, ignored\n", 1268 pr->ndpr_plen, if_name(ifp))); 1269 goto end; 1270 } 1271 1272 if ((ia6 = in6_ifadd(new, mcast)) != NULL) { 1273 /* 1274 * note that we should use pr (not new) for reference. 1275 */ 1276 pr->ndpr_refcnt++; 1277 ia6->ia6_ndpr = pr; 1278 1279 /* 1280 * RFC 3041 3.3 (2). 1281 * When a new public address is created as described 1282 * in RFC2462, also create a new temporary address. 1283 * 1284 * RFC 3041 3.5. 1285 * When an interface connects to a new link, a new 1286 * randomized interface identifier should be generated 1287 * immediately together with a new set of temporary 1288 * addresses. Thus, we specifiy 1 as the 2nd arg of 1289 * in6_tmpifadd(). 1290 */ 1291 if (V_ip6_use_tempaddr) { 1292 int e; 1293 if ((e = in6_tmpifadd(ia6, 1, 1)) != 0) { 1294 nd6log((LOG_NOTICE, "prelist_update: " 1295 "failed to create a temporary " 1296 "address, errno=%d\n", 1297 e)); 1298 } 1299 } 1300 ifa_free(&ia6->ia_ifa); 1301 1302 /* 1303 * A newly added address might affect the status 1304 * of other addresses, so we check and update it. 1305 * XXX: what if address duplication happens? 1306 */ 1307 pfxlist_onlink_check(); 1308 } else { 1309 /* just set an error. do not bark here. */ 1310 error = EADDRNOTAVAIL; /* XXX: might be unused. */ 1311 } 1312 } 1313 1314 end: 1315 splx(s); 1316 return error; 1317 } 1318 1319 /* 1320 * A supplement function used in the on-link detection below; 1321 * detect if a given prefix has a (probably) reachable advertising router. 1322 * XXX: lengthy function name... 1323 */ 1324 static struct nd_pfxrouter * 1325 find_pfxlist_reachable_router(struct nd_prefix *pr) 1326 { 1327 struct nd_pfxrouter *pfxrtr; 1328 struct llentry *ln; 1329 int canreach; 1330 1331 for (pfxrtr = LIST_FIRST(&pr->ndpr_advrtrs); pfxrtr != NULL; 1332 pfxrtr = LIST_NEXT(pfxrtr, pfr_entry)) { 1333 IF_AFDATA_LOCK(pfxrtr->router->ifp); 1334 ln = nd6_lookup(&pfxrtr->router->rtaddr, 0, pfxrtr->router->ifp); 1335 IF_AFDATA_UNLOCK(pfxrtr->router->ifp); 1336 if (ln == NULL) 1337 continue; 1338 canreach = ND6_IS_LLINFO_PROBREACH(ln); 1339 LLE_RUNLOCK(ln); 1340 if (canreach) 1341 break; 1342 } 1343 return (pfxrtr); 1344 } 1345 1346 /* 1347 * Check if each prefix in the prefix list has at least one available router 1348 * that advertised the prefix (a router is "available" if its neighbor cache 1349 * entry is reachable or probably reachable). 1350 * If the check fails, the prefix may be off-link, because, for example, 1351 * we have moved from the network but the lifetime of the prefix has not 1352 * expired yet. So we should not use the prefix if there is another prefix 1353 * that has an available router. 1354 * But, if there is no prefix that has an available router, we still regards 1355 * all the prefixes as on-link. This is because we can't tell if all the 1356 * routers are simply dead or if we really moved from the network and there 1357 * is no router around us. 1358 */ 1359 void 1360 pfxlist_onlink_check() 1361 { 1362 struct nd_prefix *pr; 1363 struct in6_ifaddr *ifa; 1364 struct nd_defrouter *dr; 1365 struct nd_pfxrouter *pfxrtr = NULL; 1366 1367 /* 1368 * Check if there is a prefix that has a reachable advertising 1369 * router. 1370 */ 1371 for (pr = V_nd_prefix.lh_first; pr; pr = pr->ndpr_next) { 1372 if (pr->ndpr_raf_onlink && find_pfxlist_reachable_router(pr)) 1373 break; 1374 } 1375 1376 /* 1377 * If we have no such prefix, check whether we still have a router 1378 * that does not advertise any prefixes. 1379 */ 1380 if (pr == NULL) { 1381 for (dr = TAILQ_FIRST(&V_nd_defrouter); dr; 1382 dr = TAILQ_NEXT(dr, dr_entry)) { 1383 struct nd_prefix *pr0; 1384 1385 for (pr0 = V_nd_prefix.lh_first; pr0; 1386 pr0 = pr0->ndpr_next) { 1387 if ((pfxrtr = pfxrtr_lookup(pr0, dr)) != NULL) 1388 break; 1389 } 1390 if (pfxrtr != NULL) 1391 break; 1392 } 1393 } 1394 if (pr != NULL || (TAILQ_FIRST(&V_nd_defrouter) && pfxrtr == NULL)) { 1395 /* 1396 * There is at least one prefix that has a reachable router, 1397 * or at least a router which probably does not advertise 1398 * any prefixes. The latter would be the case when we move 1399 * to a new link where we have a router that does not provide 1400 * prefixes and we configure an address by hand. 1401 * Detach prefixes which have no reachable advertising 1402 * router, and attach other prefixes. 1403 */ 1404 for (pr = V_nd_prefix.lh_first; pr; pr = pr->ndpr_next) { 1405 /* XXX: a link-local prefix should never be detached */ 1406 if (IN6_IS_ADDR_LINKLOCAL(&pr->ndpr_prefix.sin6_addr)) 1407 continue; 1408 1409 /* 1410 * we aren't interested in prefixes without the L bit 1411 * set. 1412 */ 1413 if (pr->ndpr_raf_onlink == 0) 1414 continue; 1415 1416 if (pr->ndpr_raf_auto == 0) 1417 continue; 1418 1419 if ((pr->ndpr_stateflags & NDPRF_DETACHED) == 0 && 1420 find_pfxlist_reachable_router(pr) == NULL) 1421 pr->ndpr_stateflags |= NDPRF_DETACHED; 1422 if ((pr->ndpr_stateflags & NDPRF_DETACHED) != 0 && 1423 find_pfxlist_reachable_router(pr) != 0) 1424 pr->ndpr_stateflags &= ~NDPRF_DETACHED; 1425 } 1426 } else { 1427 /* there is no prefix that has a reachable router */ 1428 for (pr = V_nd_prefix.lh_first; pr; pr = pr->ndpr_next) { 1429 if (IN6_IS_ADDR_LINKLOCAL(&pr->ndpr_prefix.sin6_addr)) 1430 continue; 1431 1432 if (pr->ndpr_raf_onlink == 0) 1433 continue; 1434 1435 if (pr->ndpr_raf_auto == 0) 1436 continue; 1437 1438 if ((pr->ndpr_stateflags & NDPRF_DETACHED) != 0) 1439 pr->ndpr_stateflags &= ~NDPRF_DETACHED; 1440 } 1441 } 1442 1443 /* 1444 * Remove each interface route associated with a (just) detached 1445 * prefix, and reinstall the interface route for a (just) attached 1446 * prefix. Note that all attempt of reinstallation does not 1447 * necessarily success, when a same prefix is shared among multiple 1448 * interfaces. Such cases will be handled in nd6_prefix_onlink, 1449 * so we don't have to care about them. 1450 */ 1451 for (pr = V_nd_prefix.lh_first; pr; pr = pr->ndpr_next) { 1452 int e; 1453 char ip6buf[INET6_ADDRSTRLEN]; 1454 1455 if (IN6_IS_ADDR_LINKLOCAL(&pr->ndpr_prefix.sin6_addr)) 1456 continue; 1457 1458 if (pr->ndpr_raf_onlink == 0) 1459 continue; 1460 1461 if (pr->ndpr_raf_auto == 0) 1462 continue; 1463 1464 if ((pr->ndpr_stateflags & NDPRF_DETACHED) != 0 && 1465 (pr->ndpr_stateflags & NDPRF_ONLINK) != 0) { 1466 if ((e = nd6_prefix_offlink(pr)) != 0) { 1467 nd6log((LOG_ERR, 1468 "pfxlist_onlink_check: failed to " 1469 "make %s/%d offlink, errno=%d\n", 1470 ip6_sprintf(ip6buf, 1471 &pr->ndpr_prefix.sin6_addr), 1472 pr->ndpr_plen, e)); 1473 } 1474 } 1475 if ((pr->ndpr_stateflags & NDPRF_DETACHED) == 0 && 1476 (pr->ndpr_stateflags & NDPRF_ONLINK) == 0 && 1477 pr->ndpr_raf_onlink) { 1478 if ((e = nd6_prefix_onlink(pr)) != 0) { 1479 nd6log((LOG_ERR, 1480 "pfxlist_onlink_check: failed to " 1481 "make %s/%d onlink, errno=%d\n", 1482 ip6_sprintf(ip6buf, 1483 &pr->ndpr_prefix.sin6_addr), 1484 pr->ndpr_plen, e)); 1485 } 1486 } 1487 } 1488 1489 /* 1490 * Changes on the prefix status might affect address status as well. 1491 * Make sure that all addresses derived from an attached prefix are 1492 * attached, and that all addresses derived from a detached prefix are 1493 * detached. Note, however, that a manually configured address should 1494 * always be attached. 1495 * The precise detection logic is same as the one for prefixes. 1496 * 1497 * XXXRW: in6_ifaddrhead locking. 1498 */ 1499 TAILQ_FOREACH(ifa, &V_in6_ifaddrhead, ia_link) { 1500 if (!(ifa->ia6_flags & IN6_IFF_AUTOCONF)) 1501 continue; 1502 1503 if (ifa->ia6_ndpr == NULL) { 1504 /* 1505 * This can happen when we first configure the address 1506 * (i.e. the address exists, but the prefix does not). 1507 * XXX: complicated relationships... 1508 */ 1509 continue; 1510 } 1511 1512 if (find_pfxlist_reachable_router(ifa->ia6_ndpr)) 1513 break; 1514 } 1515 if (ifa) { 1516 TAILQ_FOREACH(ifa, &V_in6_ifaddrhead, ia_link) { 1517 if ((ifa->ia6_flags & IN6_IFF_AUTOCONF) == 0) 1518 continue; 1519 1520 if (ifa->ia6_ndpr == NULL) /* XXX: see above. */ 1521 continue; 1522 1523 if (find_pfxlist_reachable_router(ifa->ia6_ndpr)) { 1524 if (ifa->ia6_flags & IN6_IFF_DETACHED) { 1525 ifa->ia6_flags &= ~IN6_IFF_DETACHED; 1526 ifa->ia6_flags |= IN6_IFF_TENTATIVE; 1527 nd6_dad_start((struct ifaddr *)ifa, 0); 1528 } 1529 } else { 1530 ifa->ia6_flags |= IN6_IFF_DETACHED; 1531 } 1532 } 1533 } 1534 else { 1535 TAILQ_FOREACH(ifa, &V_in6_ifaddrhead, ia_link) { 1536 if ((ifa->ia6_flags & IN6_IFF_AUTOCONF) == 0) 1537 continue; 1538 1539 if (ifa->ia6_flags & IN6_IFF_DETACHED) { 1540 ifa->ia6_flags &= ~IN6_IFF_DETACHED; 1541 ifa->ia6_flags |= IN6_IFF_TENTATIVE; 1542 /* Do we need a delay in this case? */ 1543 nd6_dad_start((struct ifaddr *)ifa, 0); 1544 } 1545 } 1546 } 1547 } 1548 1549 int 1550 nd6_prefix_onlink(struct nd_prefix *pr) 1551 { 1552 struct ifaddr *ifa; 1553 struct ifnet *ifp = pr->ndpr_ifp; 1554 struct sockaddr_in6 mask6; 1555 struct nd_prefix *opr; 1556 u_long rtflags; 1557 int error = 0; 1558 struct radix_node_head *rnh; 1559 struct rtentry *rt = NULL; 1560 char ip6buf[INET6_ADDRSTRLEN]; 1561 struct sockaddr_dl null_sdl = {sizeof(null_sdl), AF_LINK}; 1562 1563 /* sanity check */ 1564 if ((pr->ndpr_stateflags & NDPRF_ONLINK) != 0) { 1565 nd6log((LOG_ERR, 1566 "nd6_prefix_onlink: %s/%d is already on-link\n", 1567 ip6_sprintf(ip6buf, &pr->ndpr_prefix.sin6_addr), 1568 pr->ndpr_plen)); 1569 return (EEXIST); 1570 } 1571 1572 /* 1573 * Add the interface route associated with the prefix. Before 1574 * installing the route, check if there's the same prefix on another 1575 * interface, and the prefix has already installed the interface route. 1576 * Although such a configuration is expected to be rare, we explicitly 1577 * allow it. 1578 */ 1579 for (opr = V_nd_prefix.lh_first; opr; opr = opr->ndpr_next) { 1580 if (opr == pr) 1581 continue; 1582 1583 if ((opr->ndpr_stateflags & NDPRF_ONLINK) == 0) 1584 continue; 1585 1586 if (opr->ndpr_plen == pr->ndpr_plen && 1587 in6_are_prefix_equal(&pr->ndpr_prefix.sin6_addr, 1588 &opr->ndpr_prefix.sin6_addr, pr->ndpr_plen)) 1589 return (0); 1590 } 1591 1592 /* 1593 * We prefer link-local addresses as the associated interface address. 1594 */ 1595 /* search for a link-local addr */ 1596 ifa = (struct ifaddr *)in6ifa_ifpforlinklocal(ifp, 1597 IN6_IFF_NOTREADY | IN6_IFF_ANYCAST); 1598 if (ifa == NULL) { 1599 /* XXX: freebsd does not have ifa_ifwithaf */ 1600 IF_ADDR_LOCK(ifp); 1601 TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { 1602 if (ifa->ifa_addr->sa_family == AF_INET6) 1603 break; 1604 } 1605 if (ifa != NULL) 1606 ifa_ref(ifa); 1607 IF_ADDR_UNLOCK(ifp); 1608 /* should we care about ia6_flags? */ 1609 } 1610 if (ifa == NULL) { 1611 /* 1612 * This can still happen, when, for example, we receive an RA 1613 * containing a prefix with the L bit set and the A bit clear, 1614 * after removing all IPv6 addresses on the receiving 1615 * interface. This should, of course, be rare though. 1616 */ 1617 nd6log((LOG_NOTICE, 1618 "nd6_prefix_onlink: failed to find any ifaddr" 1619 " to add route for a prefix(%s/%d) on %s\n", 1620 ip6_sprintf(ip6buf, &pr->ndpr_prefix.sin6_addr), 1621 pr->ndpr_plen, if_name(ifp))); 1622 return (0); 1623 } 1624 1625 /* 1626 * in6_ifinit() sets nd6_rtrequest to ifa_rtrequest for all ifaddrs. 1627 * ifa->ifa_rtrequest = nd6_rtrequest; 1628 */ 1629 bzero(&mask6, sizeof(mask6)); 1630 mask6.sin6_len = sizeof(mask6); 1631 mask6.sin6_addr = pr->ndpr_mask; 1632 rtflags = (ifa->ifa_flags & ~IFA_RTSELF) | RTF_UP; 1633 error = rtrequest(RTM_ADD, (struct sockaddr *)&pr->ndpr_prefix, 1634 ifa->ifa_addr, (struct sockaddr *)&mask6, rtflags, &rt); 1635 if (error == 0) { 1636 if (rt != NULL) /* this should be non NULL, though */ { 1637 rnh = rt_tables_get_rnh(rt->rt_fibnum, AF_INET6); 1638 /* XXX what if rhn == NULL? */ 1639 RADIX_NODE_HEAD_LOCK(rnh); 1640 RT_LOCK(rt); 1641 if (!rt_setgate(rt, rt_key(rt), (struct sockaddr *)&null_sdl)) { 1642 ((struct sockaddr_dl *)rt->rt_gateway)->sdl_type = 1643 rt->rt_ifp->if_type; 1644 ((struct sockaddr_dl *)rt->rt_gateway)->sdl_index = 1645 rt->rt_ifp->if_index; 1646 } 1647 RADIX_NODE_HEAD_UNLOCK(rnh); 1648 nd6_rtmsg(RTM_ADD, rt); 1649 RT_UNLOCK(rt); 1650 } 1651 pr->ndpr_stateflags |= NDPRF_ONLINK; 1652 } else { 1653 char ip6bufg[INET6_ADDRSTRLEN], ip6bufm[INET6_ADDRSTRLEN]; 1654 nd6log((LOG_ERR, "nd6_prefix_onlink: failed to add route for a" 1655 " prefix (%s/%d) on %s, gw=%s, mask=%s, flags=%lx " 1656 "errno = %d\n", 1657 ip6_sprintf(ip6buf, &pr->ndpr_prefix.sin6_addr), 1658 pr->ndpr_plen, if_name(ifp), 1659 ip6_sprintf(ip6bufg, &((struct sockaddr_in6 *)ifa->ifa_addr)->sin6_addr), 1660 ip6_sprintf(ip6bufm, &mask6.sin6_addr), rtflags, error)); 1661 } 1662 1663 if (rt != NULL) { 1664 RT_LOCK(rt); 1665 RT_REMREF(rt); 1666 RT_UNLOCK(rt); 1667 } 1668 if (ifa != NULL) 1669 ifa_free(ifa); 1670 1671 return (error); 1672 } 1673 1674 int 1675 nd6_prefix_offlink(struct nd_prefix *pr) 1676 { 1677 int error = 0; 1678 struct ifnet *ifp = pr->ndpr_ifp; 1679 struct nd_prefix *opr; 1680 struct sockaddr_in6 sa6, mask6; 1681 struct rtentry *rt = NULL; 1682 char ip6buf[INET6_ADDRSTRLEN]; 1683 1684 /* sanity check */ 1685 if ((pr->ndpr_stateflags & NDPRF_ONLINK) == 0) { 1686 nd6log((LOG_ERR, 1687 "nd6_prefix_offlink: %s/%d is already off-link\n", 1688 ip6_sprintf(ip6buf, &pr->ndpr_prefix.sin6_addr), 1689 pr->ndpr_plen)); 1690 return (EEXIST); 1691 } 1692 1693 bzero(&sa6, sizeof(sa6)); 1694 sa6.sin6_family = AF_INET6; 1695 sa6.sin6_len = sizeof(sa6); 1696 bcopy(&pr->ndpr_prefix.sin6_addr, &sa6.sin6_addr, 1697 sizeof(struct in6_addr)); 1698 bzero(&mask6, sizeof(mask6)); 1699 mask6.sin6_family = AF_INET6; 1700 mask6.sin6_len = sizeof(sa6); 1701 bcopy(&pr->ndpr_mask, &mask6.sin6_addr, sizeof(struct in6_addr)); 1702 error = rtrequest(RTM_DELETE, (struct sockaddr *)&sa6, NULL, 1703 (struct sockaddr *)&mask6, 0, &rt); 1704 if (error == 0) { 1705 pr->ndpr_stateflags &= ~NDPRF_ONLINK; 1706 1707 /* report the route deletion to the routing socket. */ 1708 if (rt != NULL) 1709 nd6_rtmsg(RTM_DELETE, rt); 1710 1711 /* 1712 * There might be the same prefix on another interface, 1713 * the prefix which could not be on-link just because we have 1714 * the interface route (see comments in nd6_prefix_onlink). 1715 * If there's one, try to make the prefix on-link on the 1716 * interface. 1717 */ 1718 for (opr = V_nd_prefix.lh_first; opr; opr = opr->ndpr_next) { 1719 if (opr == pr) 1720 continue; 1721 1722 if ((opr->ndpr_stateflags & NDPRF_ONLINK) != 0) 1723 continue; 1724 1725 /* 1726 * KAME specific: detached prefixes should not be 1727 * on-link. 1728 */ 1729 if ((opr->ndpr_stateflags & NDPRF_DETACHED) != 0) 1730 continue; 1731 1732 if (opr->ndpr_plen == pr->ndpr_plen && 1733 in6_are_prefix_equal(&pr->ndpr_prefix.sin6_addr, 1734 &opr->ndpr_prefix.sin6_addr, pr->ndpr_plen)) { 1735 int e; 1736 1737 if ((e = nd6_prefix_onlink(opr)) != 0) { 1738 nd6log((LOG_ERR, 1739 "nd6_prefix_offlink: failed to " 1740 "recover a prefix %s/%d from %s " 1741 "to %s (errno = %d)\n", 1742 ip6_sprintf(ip6buf, 1743 &opr->ndpr_prefix.sin6_addr), 1744 opr->ndpr_plen, if_name(ifp), 1745 if_name(opr->ndpr_ifp), e)); 1746 } 1747 } 1748 } 1749 } else { 1750 /* XXX: can we still set the NDPRF_ONLINK flag? */ 1751 nd6log((LOG_ERR, 1752 "nd6_prefix_offlink: failed to delete route: " 1753 "%s/%d on %s (errno = %d)\n", 1754 ip6_sprintf(ip6buf, &sa6.sin6_addr), pr->ndpr_plen, 1755 if_name(ifp), error)); 1756 } 1757 1758 if (rt != NULL) { 1759 RTFREE(rt); 1760 } 1761 1762 return (error); 1763 } 1764 1765 static struct in6_ifaddr * 1766 in6_ifadd(struct nd_prefixctl *pr, int mcast) 1767 { 1768 struct ifnet *ifp = pr->ndpr_ifp; 1769 struct ifaddr *ifa; 1770 struct in6_aliasreq ifra; 1771 struct in6_ifaddr *ia, *ib; 1772 int error, plen0; 1773 struct in6_addr mask; 1774 int prefixlen = pr->ndpr_plen; 1775 int updateflags; 1776 char ip6buf[INET6_ADDRSTRLEN]; 1777 1778 in6_prefixlen2mask(&mask, prefixlen); 1779 1780 /* 1781 * find a link-local address (will be interface ID). 1782 * Is it really mandatory? Theoretically, a global or a site-local 1783 * address can be configured without a link-local address, if we 1784 * have a unique interface identifier... 1785 * 1786 * it is not mandatory to have a link-local address, we can generate 1787 * interface identifier on the fly. we do this because: 1788 * (1) it should be the easiest way to find interface identifier. 1789 * (2) RFC2462 5.4 suggesting the use of the same interface identifier 1790 * for multiple addresses on a single interface, and possible shortcut 1791 * of DAD. we omitted DAD for this reason in the past. 1792 * (3) a user can prevent autoconfiguration of global address 1793 * by removing link-local address by hand (this is partly because we 1794 * don't have other way to control the use of IPv6 on an interface. 1795 * this has been our design choice - cf. NRL's "ifconfig auto"). 1796 * (4) it is easier to manage when an interface has addresses 1797 * with the same interface identifier, than to have multiple addresses 1798 * with different interface identifiers. 1799 */ 1800 ifa = (struct ifaddr *)in6ifa_ifpforlinklocal(ifp, 0); /* 0 is OK? */ 1801 if (ifa) 1802 ib = (struct in6_ifaddr *)ifa; 1803 else 1804 return NULL; 1805 1806 /* prefixlen + ifidlen must be equal to 128 */ 1807 plen0 = in6_mask2len(&ib->ia_prefixmask.sin6_addr, NULL); 1808 if (prefixlen != plen0) { 1809 ifa_free(ifa); 1810 nd6log((LOG_INFO, "in6_ifadd: wrong prefixlen for %s " 1811 "(prefix=%d ifid=%d)\n", 1812 if_name(ifp), prefixlen, 128 - plen0)); 1813 return NULL; 1814 } 1815 1816 /* make ifaddr */ 1817 1818 bzero(&ifra, sizeof(ifra)); 1819 /* 1820 * in6_update_ifa() does not use ifra_name, but we accurately set it 1821 * for safety. 1822 */ 1823 strncpy(ifra.ifra_name, if_name(ifp), sizeof(ifra.ifra_name)); 1824 ifra.ifra_addr.sin6_family = AF_INET6; 1825 ifra.ifra_addr.sin6_len = sizeof(struct sockaddr_in6); 1826 /* prefix */ 1827 ifra.ifra_addr.sin6_addr = pr->ndpr_prefix.sin6_addr; 1828 ifra.ifra_addr.sin6_addr.s6_addr32[0] &= mask.s6_addr32[0]; 1829 ifra.ifra_addr.sin6_addr.s6_addr32[1] &= mask.s6_addr32[1]; 1830 ifra.ifra_addr.sin6_addr.s6_addr32[2] &= mask.s6_addr32[2]; 1831 ifra.ifra_addr.sin6_addr.s6_addr32[3] &= mask.s6_addr32[3]; 1832 1833 /* interface ID */ 1834 ifra.ifra_addr.sin6_addr.s6_addr32[0] |= 1835 (ib->ia_addr.sin6_addr.s6_addr32[0] & ~mask.s6_addr32[0]); 1836 ifra.ifra_addr.sin6_addr.s6_addr32[1] |= 1837 (ib->ia_addr.sin6_addr.s6_addr32[1] & ~mask.s6_addr32[1]); 1838 ifra.ifra_addr.sin6_addr.s6_addr32[2] |= 1839 (ib->ia_addr.sin6_addr.s6_addr32[2] & ~mask.s6_addr32[2]); 1840 ifra.ifra_addr.sin6_addr.s6_addr32[3] |= 1841 (ib->ia_addr.sin6_addr.s6_addr32[3] & ~mask.s6_addr32[3]); 1842 ifa_free(ifa); 1843 1844 /* new prefix mask. */ 1845 ifra.ifra_prefixmask.sin6_len = sizeof(struct sockaddr_in6); 1846 ifra.ifra_prefixmask.sin6_family = AF_INET6; 1847 bcopy(&mask, &ifra.ifra_prefixmask.sin6_addr, 1848 sizeof(ifra.ifra_prefixmask.sin6_addr)); 1849 1850 /* lifetimes. */ 1851 ifra.ifra_lifetime.ia6t_vltime = pr->ndpr_vltime; 1852 ifra.ifra_lifetime.ia6t_pltime = pr->ndpr_pltime; 1853 1854 /* XXX: scope zone ID? */ 1855 1856 ifra.ifra_flags |= IN6_IFF_AUTOCONF; /* obey autoconf */ 1857 1858 /* 1859 * Make sure that we do not have this address already. This should 1860 * usually not happen, but we can still see this case, e.g., if we 1861 * have manually configured the exact address to be configured. 1862 */ 1863 ifa = (struct ifaddr *)in6ifa_ifpwithaddr(ifp, 1864 &ifra.ifra_addr.sin6_addr); 1865 if (ifa != NULL) { 1866 ifa_free(ifa); 1867 /* this should be rare enough to make an explicit log */ 1868 log(LOG_INFO, "in6_ifadd: %s is already configured\n", 1869 ip6_sprintf(ip6buf, &ifra.ifra_addr.sin6_addr)); 1870 return (NULL); 1871 } 1872 1873 /* 1874 * Allocate ifaddr structure, link into chain, etc. 1875 * If we are going to create a new address upon receiving a multicasted 1876 * RA, we need to impose a random delay before starting DAD. 1877 * [draft-ietf-ipv6-rfc2462bis-02.txt, Section 5.4.2] 1878 */ 1879 updateflags = 0; 1880 if (mcast) 1881 updateflags |= IN6_IFAUPDATE_DADDELAY; 1882 if ((error = in6_update_ifa(ifp, &ifra, NULL, updateflags)) != 0) { 1883 nd6log((LOG_ERR, 1884 "in6_ifadd: failed to make ifaddr %s on %s (errno=%d)\n", 1885 ip6_sprintf(ip6buf, &ifra.ifra_addr.sin6_addr), 1886 if_name(ifp), error)); 1887 return (NULL); /* ifaddr must not have been allocated. */ 1888 } 1889 1890 ia = in6ifa_ifpwithaddr(ifp, &ifra.ifra_addr.sin6_addr); 1891 /* 1892 * XXXRW: Assumption of non-NULLness here might not be true with 1893 * fine-grained locking -- should we validate it? Or just return 1894 * earlier ifa rather than looking it up again? 1895 */ 1896 return (ia); /* this is always non-NULL and referenced. */ 1897 } 1898 1899 /* 1900 * ia0 - corresponding public address 1901 */ 1902 int 1903 in6_tmpifadd(const struct in6_ifaddr *ia0, int forcegen, int delay) 1904 { 1905 struct ifnet *ifp = ia0->ia_ifa.ifa_ifp; 1906 struct in6_ifaddr *newia, *ia; 1907 struct in6_aliasreq ifra; 1908 int i, error; 1909 int trylimit = 3; /* XXX: adhoc value */ 1910 int updateflags; 1911 u_int32_t randid[2]; 1912 time_t vltime0, pltime0; 1913 1914 bzero(&ifra, sizeof(ifra)); 1915 strncpy(ifra.ifra_name, if_name(ifp), sizeof(ifra.ifra_name)); 1916 ifra.ifra_addr = ia0->ia_addr; 1917 /* copy prefix mask */ 1918 ifra.ifra_prefixmask = ia0->ia_prefixmask; 1919 /* clear the old IFID */ 1920 for (i = 0; i < 4; i++) { 1921 ifra.ifra_addr.sin6_addr.s6_addr32[i] &= 1922 ifra.ifra_prefixmask.sin6_addr.s6_addr32[i]; 1923 } 1924 1925 again: 1926 if (in6_get_tmpifid(ifp, (u_int8_t *)randid, 1927 (const u_int8_t *)&ia0->ia_addr.sin6_addr.s6_addr[8], forcegen)) { 1928 nd6log((LOG_NOTICE, "in6_tmpifadd: failed to find a good " 1929 "random IFID\n")); 1930 return (EINVAL); 1931 } 1932 ifra.ifra_addr.sin6_addr.s6_addr32[2] |= 1933 (randid[0] & ~(ifra.ifra_prefixmask.sin6_addr.s6_addr32[2])); 1934 ifra.ifra_addr.sin6_addr.s6_addr32[3] |= 1935 (randid[1] & ~(ifra.ifra_prefixmask.sin6_addr.s6_addr32[3])); 1936 1937 /* 1938 * in6_get_tmpifid() quite likely provided a unique interface ID. 1939 * However, we may still have a chance to see collision, because 1940 * there may be a time lag between generation of the ID and generation 1941 * of the address. So, we'll do one more sanity check. 1942 */ 1943 IN6_IFADDR_RLOCK(); 1944 TAILQ_FOREACH(ia, &V_in6_ifaddrhead, ia_link) { 1945 if (IN6_ARE_ADDR_EQUAL(&ia->ia_addr.sin6_addr, 1946 &ifra.ifra_addr.sin6_addr)) { 1947 if (trylimit-- == 0) { 1948 IN6_IFADDR_RUNLOCK(); 1949 /* 1950 * Give up. Something strange should have 1951 * happened. 1952 */ 1953 nd6log((LOG_NOTICE, "in6_tmpifadd: failed to " 1954 "find a unique random IFID\n")); 1955 return (EEXIST); 1956 } 1957 IN6_IFADDR_RUNLOCK(); 1958 forcegen = 1; 1959 goto again; 1960 } 1961 } 1962 IN6_IFADDR_RUNLOCK(); 1963 1964 /* 1965 * The Valid Lifetime is the lower of the Valid Lifetime of the 1966 * public address or TEMP_VALID_LIFETIME. 1967 * The Preferred Lifetime is the lower of the Preferred Lifetime 1968 * of the public address or TEMP_PREFERRED_LIFETIME - 1969 * DESYNC_FACTOR. 1970 */ 1971 if (ia0->ia6_lifetime.ia6t_vltime != ND6_INFINITE_LIFETIME) { 1972 vltime0 = IFA6_IS_INVALID(ia0) ? 0 : 1973 (ia0->ia6_lifetime.ia6t_vltime - 1974 (time_second - ia0->ia6_updatetime)); 1975 if (vltime0 > V_ip6_temp_valid_lifetime) 1976 vltime0 = V_ip6_temp_valid_lifetime; 1977 } else 1978 vltime0 = V_ip6_temp_valid_lifetime; 1979 if (ia0->ia6_lifetime.ia6t_pltime != ND6_INFINITE_LIFETIME) { 1980 pltime0 = IFA6_IS_DEPRECATED(ia0) ? 0 : 1981 (ia0->ia6_lifetime.ia6t_pltime - 1982 (time_second - ia0->ia6_updatetime)); 1983 if (pltime0 > V_ip6_temp_preferred_lifetime - V_ip6_desync_factor){ 1984 pltime0 = V_ip6_temp_preferred_lifetime - 1985 V_ip6_desync_factor; 1986 } 1987 } else 1988 pltime0 = V_ip6_temp_preferred_lifetime - V_ip6_desync_factor; 1989 ifra.ifra_lifetime.ia6t_vltime = vltime0; 1990 ifra.ifra_lifetime.ia6t_pltime = pltime0; 1991 1992 /* 1993 * A temporary address is created only if this calculated Preferred 1994 * Lifetime is greater than REGEN_ADVANCE time units. 1995 */ 1996 if (ifra.ifra_lifetime.ia6t_pltime <= V_ip6_temp_regen_advance) 1997 return (0); 1998 1999 /* XXX: scope zone ID? */ 2000 2001 ifra.ifra_flags |= (IN6_IFF_AUTOCONF|IN6_IFF_TEMPORARY); 2002 2003 /* allocate ifaddr structure, link into chain, etc. */ 2004 updateflags = 0; 2005 if (delay) 2006 updateflags |= IN6_IFAUPDATE_DADDELAY; 2007 if ((error = in6_update_ifa(ifp, &ifra, NULL, updateflags)) != 0) 2008 return (error); 2009 2010 newia = in6ifa_ifpwithaddr(ifp, &ifra.ifra_addr.sin6_addr); 2011 if (newia == NULL) { /* XXX: can it happen? */ 2012 nd6log((LOG_ERR, 2013 "in6_tmpifadd: ifa update succeeded, but we got " 2014 "no ifaddr\n")); 2015 return (EINVAL); /* XXX */ 2016 } 2017 newia->ia6_ndpr = ia0->ia6_ndpr; 2018 newia->ia6_ndpr->ndpr_refcnt++; 2019 ifa_free(&newia->ia_ifa); 2020 2021 /* 2022 * A newly added address might affect the status of other addresses. 2023 * XXX: when the temporary address is generated with a new public 2024 * address, the onlink check is redundant. However, it would be safe 2025 * to do the check explicitly everywhere a new address is generated, 2026 * and, in fact, we surely need the check when we create a new 2027 * temporary address due to deprecation of an old temporary address. 2028 */ 2029 pfxlist_onlink_check(); 2030 2031 return (0); 2032 } 2033 2034 static int 2035 in6_init_prefix_ltimes(struct nd_prefix *ndpr) 2036 { 2037 if (ndpr->ndpr_pltime == ND6_INFINITE_LIFETIME) 2038 ndpr->ndpr_preferred = 0; 2039 else 2040 ndpr->ndpr_preferred = time_second + ndpr->ndpr_pltime; 2041 if (ndpr->ndpr_vltime == ND6_INFINITE_LIFETIME) 2042 ndpr->ndpr_expire = 0; 2043 else 2044 ndpr->ndpr_expire = time_second + ndpr->ndpr_vltime; 2045 2046 return 0; 2047 } 2048 2049 static void 2050 in6_init_address_ltimes(struct nd_prefix *new, struct in6_addrlifetime *lt6) 2051 { 2052 /* init ia6t_expire */ 2053 if (lt6->ia6t_vltime == ND6_INFINITE_LIFETIME) 2054 lt6->ia6t_expire = 0; 2055 else { 2056 lt6->ia6t_expire = time_second; 2057 lt6->ia6t_expire += lt6->ia6t_vltime; 2058 } 2059 2060 /* init ia6t_preferred */ 2061 if (lt6->ia6t_pltime == ND6_INFINITE_LIFETIME) 2062 lt6->ia6t_preferred = 0; 2063 else { 2064 lt6->ia6t_preferred = time_second; 2065 lt6->ia6t_preferred += lt6->ia6t_pltime; 2066 } 2067 } 2068 2069 /* 2070 * Delete all the routing table entries that use the specified gateway. 2071 * XXX: this function causes search through all entries of routing table, so 2072 * it shouldn't be called when acting as a router. 2073 */ 2074 void 2075 rt6_flush(struct in6_addr *gateway, struct ifnet *ifp) 2076 { 2077 struct radix_node_head *rnh; 2078 int s = splnet(); 2079 2080 /* We'll care only link-local addresses */ 2081 if (!IN6_IS_ADDR_LINKLOCAL(gateway)) { 2082 splx(s); 2083 return; 2084 } 2085 2086 rnh = rt_tables_get_rnh(0, AF_INET6); 2087 if (rnh == NULL) 2088 return; 2089 2090 RADIX_NODE_HEAD_LOCK(rnh); 2091 rnh->rnh_walktree(rnh, rt6_deleteroute, (void *)gateway); 2092 RADIX_NODE_HEAD_UNLOCK(rnh); 2093 splx(s); 2094 } 2095 2096 static int 2097 rt6_deleteroute(struct radix_node *rn, void *arg) 2098 { 2099 #define SIN6(s) ((struct sockaddr_in6 *)s) 2100 struct rtentry *rt = (struct rtentry *)rn; 2101 struct in6_addr *gate = (struct in6_addr *)arg; 2102 2103 if (rt->rt_gateway == NULL || rt->rt_gateway->sa_family != AF_INET6) 2104 return (0); 2105 2106 if (!IN6_ARE_ADDR_EQUAL(gate, &SIN6(rt->rt_gateway)->sin6_addr)) { 2107 return (0); 2108 } 2109 2110 /* 2111 * Do not delete a static route. 2112 * XXX: this seems to be a bit ad-hoc. Should we consider the 2113 * 'cloned' bit instead? 2114 */ 2115 if ((rt->rt_flags & RTF_STATIC) != 0) 2116 return (0); 2117 2118 /* 2119 * We delete only host route. This means, in particular, we don't 2120 * delete default route. 2121 */ 2122 if ((rt->rt_flags & RTF_HOST) == 0) 2123 return (0); 2124 2125 return (rtrequest(RTM_DELETE, rt_key(rt), rt->rt_gateway, 2126 rt_mask(rt), rt->rt_flags, 0)); 2127 #undef SIN6 2128 } 2129 2130 int 2131 nd6_setdefaultiface(int ifindex) 2132 { 2133 int error = 0; 2134 2135 if (ifindex < 0 || V_if_index < ifindex) 2136 return (EINVAL); 2137 if (ifindex != 0 && !ifnet_byindex(ifindex)) 2138 return (EINVAL); 2139 2140 if (V_nd6_defifindex != ifindex) { 2141 V_nd6_defifindex = ifindex; 2142 if (V_nd6_defifindex > 0) 2143 V_nd6_defifp = ifnet_byindex(V_nd6_defifindex); 2144 else 2145 V_nd6_defifp = NULL; 2146 2147 /* 2148 * Our current implementation assumes one-to-one maping between 2149 * interfaces and links, so it would be natural to use the 2150 * default interface as the default link. 2151 */ 2152 scope6_setdefault(V_nd6_defifp); 2153 } 2154 2155 return (error); 2156 } 2157