in6_pcb.c revision 120727
1/* $FreeBSD: head/sys/netinet6/in6_pcb.c 120727 2003-10-04 03:44:50Z sam $ */ 2/* $KAME: in6_pcb.c,v 1.31 2001/05/21 05:45:10 jinmei 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/* 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#include "opt_ipsec.h" 72#include "opt_random_ip_id.h" 73 74#include <sys/param.h> 75#include <sys/systm.h> 76#include <sys/malloc.h> 77#include <sys/mbuf.h> 78#include <sys/domain.h> 79#include <sys/protosw.h> 80#include <sys/socket.h> 81#include <sys/socketvar.h> 82#include <sys/sockio.h> 83#include <sys/errno.h> 84#include <sys/time.h> 85#include <sys/proc.h> 86#include <sys/jail.h> 87 88#include <vm/uma.h> 89 90#include <net/if.h> 91#include <net/if_types.h> 92#include <net/route.h> 93 94#include <netinet/in.h> 95#include <netinet/in_var.h> 96#include <netinet/in_systm.h> 97#include <netinet/tcp_var.h> 98#include <netinet/ip6.h> 99#include <netinet/ip_var.h> 100#include <netinet6/ip6_var.h> 101#include <netinet6/nd6.h> 102#include <netinet/in_pcb.h> 103#include <netinet6/in6_pcb.h> 104 105#ifdef IPSEC 106#include <netinet6/ipsec.h> 107#ifdef INET6 108#include <netinet6/ipsec6.h> 109#endif 110#include <netinet6/ah.h> 111#ifdef INET6 112#include <netinet6/ah6.h> 113#endif 114#include <netkey/key.h> 115#endif /* IPSEC */ 116 117#ifdef FAST_IPSEC 118#include <netipsec/ipsec.h> 119#include <netipsec/ipsec6.h> 120#include <netipsec/key.h> 121#define IPSEC 122#endif /* FAST_IPSEC */ 123 124struct in6_addr zeroin6_addr; 125 126int 127in6_pcbbind(inp, nam, td) 128 register struct inpcb *inp; 129 struct sockaddr *nam; 130 struct thread *td; 131{ 132 struct socket *so = inp->inp_socket; 133 struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)NULL; 134 struct inpcbinfo *pcbinfo = inp->inp_pcbinfo; 135 u_short lport = 0; 136 int wild = 0, reuseport = (so->so_options & SO_REUSEPORT); 137 138 if (!in6_ifaddr) /* XXX broken! */ 139 return (EADDRNOTAVAIL); 140 if (inp->inp_lport || !IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr)) 141 return(EINVAL); 142 if ((so->so_options & (SO_REUSEADDR|SO_REUSEPORT)) == 0) 143 wild = 1; 144 if (nam) { 145 sin6 = (struct sockaddr_in6 *)nam; 146 if (nam->sa_len != sizeof(*sin6)) 147 return(EINVAL); 148 /* 149 * family check. 150 */ 151 if (nam->sa_family != AF_INET6) 152 return(EAFNOSUPPORT); 153 154 /* KAME hack: embed scopeid */ 155 if (in6_embedscope(&sin6->sin6_addr, sin6, inp, NULL) != 0) 156 return EINVAL; 157 /* this must be cleared for ifa_ifwithaddr() */ 158 sin6->sin6_scope_id = 0; 159 160 lport = sin6->sin6_port; 161 if (IN6_IS_ADDR_MULTICAST(&sin6->sin6_addr)) { 162 /* 163 * Treat SO_REUSEADDR as SO_REUSEPORT for multicast; 164 * allow compepte duplication of binding if 165 * SO_REUSEPORT is set, or if SO_REUSEADDR is set 166 * and a multicast address is bound on both 167 * new and duplicated sockets. 168 */ 169 if (so->so_options & SO_REUSEADDR) 170 reuseport = SO_REUSEADDR|SO_REUSEPORT; 171 } else if (!IN6_IS_ADDR_UNSPECIFIED(&sin6->sin6_addr)) { 172 struct ifaddr *ia = NULL; 173 174 sin6->sin6_port = 0; /* yech... */ 175 if ((ia = ifa_ifwithaddr((struct sockaddr *)sin6)) == 0) 176 return(EADDRNOTAVAIL); 177 178 /* 179 * XXX: bind to an anycast address might accidentally 180 * cause sending a packet with anycast source address. 181 * We should allow to bind to a deprecated address, since 182 * the application dare to use it. 183 */ 184 if (ia && 185 ((struct in6_ifaddr *)ia)->ia6_flags & 186 (IN6_IFF_ANYCAST|IN6_IFF_NOTREADY|IN6_IFF_DETACHED)) { 187 return(EADDRNOTAVAIL); 188 } 189 } 190 if (lport) { 191 struct inpcb *t; 192 193 /* GROSS */ 194 if (ntohs(lport) < IPV6PORT_RESERVED && td && 195 suser_cred(td->td_ucred, PRISON_ROOT)) 196 return(EACCES); 197 if (so->so_cred->cr_uid != 0 && 198 !IN6_IS_ADDR_MULTICAST(&sin6->sin6_addr)) { 199 t = in6_pcblookup_local(pcbinfo, 200 &sin6->sin6_addr, lport, 201 INPLOOKUP_WILDCARD); 202 if (t && (t->inp_vflag & INP_TIMEWAIT)) { 203 if ((!IN6_IS_ADDR_UNSPECIFIED(&sin6->sin6_addr) || 204 !IN6_IS_ADDR_UNSPECIFIED(&t->in6p_laddr) || 205 !(intotw(t)->tw_so_options & SO_REUSEPORT)) 206 && so->so_cred->cr_uid != 207 intotw(t)->tw_cred->cr_uid) 208 return (EADDRINUSE); 209 } else if (t && 210 (!IN6_IS_ADDR_UNSPECIFIED(&sin6->sin6_addr) || 211 !IN6_IS_ADDR_UNSPECIFIED(&t->in6p_laddr) || 212 (t->inp_socket->so_options & SO_REUSEPORT) 213 == 0) && (so->so_cred->cr_uid != 214 t->inp_socket->so_cred->cr_uid)) 215 return (EADDRINUSE); 216 if ((inp->inp_flags & IN6P_IPV6_V6ONLY) == 0 && 217 IN6_IS_ADDR_UNSPECIFIED(&sin6->sin6_addr)) { 218 struct sockaddr_in sin; 219 220 in6_sin6_2_sin(&sin, sin6); 221 t = in_pcblookup_local(pcbinfo, 222 sin.sin_addr, lport, 223 INPLOOKUP_WILDCARD); 224 if (t && (t->inp_vflag & INP_TIMEWAIT)) { 225 if (so->so_cred->cr_uid != 226 intotw(t)->tw_cred->cr_uid && 227 (ntohl(t->inp_laddr.s_addr) != 228 INADDR_ANY || 229 ((inp->inp_vflag & 230 INP_IPV6PROTO) == 231 (t->inp_vflag & 232 INP_IPV6PROTO)))) 233 return (EADDRINUSE); 234 } else if (t && 235 (so->so_cred->cr_uid != 236 t->inp_socket->so_cred->cr_uid) && 237 (ntohl(t->inp_laddr.s_addr) != 238 INADDR_ANY || 239 INP_SOCKAF(so) == 240 INP_SOCKAF(t->inp_socket))) 241 return (EADDRINUSE); 242 } 243 } 244 t = in6_pcblookup_local(pcbinfo, &sin6->sin6_addr, 245 lport, wild); 246 if (t && (reuseport & ((t->inp_vflag & INP_TIMEWAIT) ? 247 intotw(t)->tw_so_options : 248 t->inp_socket->so_options)) == 0) 249 return(EADDRINUSE); 250 if ((inp->inp_flags & IN6P_IPV6_V6ONLY) == 0 && 251 IN6_IS_ADDR_UNSPECIFIED(&sin6->sin6_addr)) { 252 struct sockaddr_in sin; 253 254 in6_sin6_2_sin(&sin, sin6); 255 t = in_pcblookup_local(pcbinfo, sin.sin_addr, 256 lport, wild); 257 if (t && t->inp_vflag & INP_TIMEWAIT) { 258 if ((reuseport & 259 intotw(t)->tw_so_options) == 0 && 260 (ntohl(t->inp_laddr.s_addr) != 261 INADDR_ANY || ((inp->inp_vflag & 262 INP_IPV6PROTO) == 263 (t->inp_vflag & INP_IPV6PROTO)))) 264 return (EADDRINUSE); 265 } 266 else if (t && 267 (reuseport & t->inp_socket->so_options) 268 == 0 && (ntohl(t->inp_laddr.s_addr) != 269 INADDR_ANY || INP_SOCKAF(so) == 270 INP_SOCKAF(t->inp_socket))) 271 return (EADDRINUSE); 272 } 273 } 274 inp->in6p_laddr = sin6->sin6_addr; 275 } 276 if (lport == 0) { 277 int e; 278 if ((e = in6_pcbsetport(&inp->in6p_laddr, inp, td)) != 0) 279 return(e); 280 } 281 else { 282 inp->inp_lport = lport; 283 if (in_pcbinshash(inp) != 0) { 284 inp->in6p_laddr = in6addr_any; 285 inp->inp_lport = 0; 286 return (EAGAIN); 287 } 288 } 289 return(0); 290} 291 292/* 293 * Transform old in6_pcbconnect() into an inner subroutine for new 294 * in6_pcbconnect(): Do some validity-checking on the remote 295 * address (in mbuf 'nam') and then determine local host address 296 * (i.e., which interface) to use to access that remote host. 297 * 298 * This preserves definition of in6_pcbconnect(), while supporting a 299 * slightly different version for T/TCP. (This is more than 300 * a bit of a kludge, but cleaning up the internal interfaces would 301 * have forced minor changes in every protocol). 302 */ 303 304int 305in6_pcbladdr(inp, nam, plocal_addr6) 306 register struct inpcb *inp; 307 struct sockaddr *nam; 308 struct in6_addr **plocal_addr6; 309{ 310 register struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)nam; 311 struct ifnet *ifp = NULL; 312 int error = 0; 313 314 if (nam->sa_len != sizeof (*sin6)) 315 return (EINVAL); 316 if (sin6->sin6_family != AF_INET6) 317 return (EAFNOSUPPORT); 318 if (sin6->sin6_port == 0) 319 return (EADDRNOTAVAIL); 320 321 /* KAME hack: embed scopeid */ 322 if (in6_embedscope(&sin6->sin6_addr, sin6, inp, &ifp) != 0) 323 return EINVAL; 324 325 if (in6_ifaddr) { 326 /* 327 * If the destination address is UNSPECIFIED addr, 328 * use the loopback addr, e.g ::1. 329 */ 330 if (IN6_IS_ADDR_UNSPECIFIED(&sin6->sin6_addr)) 331 sin6->sin6_addr = in6addr_loopback; 332 } 333 { 334 /* 335 * XXX: in6_selectsrc might replace the bound local address 336 * with the address specified by setsockopt(IPV6_PKTINFO). 337 * Is it the intended behavior? 338 */ 339 *plocal_addr6 = in6_selectsrc(sin6, inp->in6p_outputopts, 340 inp->in6p_moptions, 341 &inp->in6p_route, 342 &inp->in6p_laddr, &error); 343 if (*plocal_addr6 == 0) { 344 if (error == 0) 345 error = EADDRNOTAVAIL; 346 return(error); 347 } 348 /* 349 * Don't do pcblookup call here; return interface in 350 * plocal_addr6 351 * and exit to caller, that will do the lookup. 352 */ 353 } 354 355 if (inp->in6p_route.ro_rt) 356 ifp = inp->in6p_route.ro_rt->rt_ifp; 357 358 return(0); 359} 360 361/* 362 * Outer subroutine: 363 * Connect from a socket to a specified address. 364 * Both address and port must be specified in argument sin. 365 * If don't have a local address for this socket yet, 366 * then pick one. 367 */ 368int 369in6_pcbconnect(inp, nam, td) 370 register struct inpcb *inp; 371 struct sockaddr *nam; 372 struct thread *td; 373{ 374 struct in6_addr *addr6; 375 register struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)nam; 376 int error; 377 378 /* 379 * Call inner routine, to assign local interface address. 380 * in6_pcbladdr() may automatically fill in sin6_scope_id. 381 */ 382 if ((error = in6_pcbladdr(inp, nam, &addr6)) != 0) 383 return(error); 384 385 if (in6_pcblookup_hash(inp->inp_pcbinfo, &sin6->sin6_addr, 386 sin6->sin6_port, 387 IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr) 388 ? addr6 : &inp->in6p_laddr, 389 inp->inp_lport, 0, NULL) != NULL) { 390 return (EADDRINUSE); 391 } 392 if (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr)) { 393 if (inp->inp_lport == 0) { 394 error = in6_pcbbind(inp, (struct sockaddr *)0, td); 395 if (error) 396 return (error); 397 } 398 inp->in6p_laddr = *addr6; 399 } 400 inp->in6p_faddr = sin6->sin6_addr; 401 inp->inp_fport = sin6->sin6_port; 402 /* update flowinfo - draft-itojun-ipv6-flowlabel-api-00 */ 403 inp->in6p_flowinfo &= ~IPV6_FLOWLABEL_MASK; 404 if (inp->in6p_flags & IN6P_AUTOFLOWLABEL) 405 inp->in6p_flowinfo |= 406#ifdef RANDOM_IP_ID 407 (htonl(ip6_randomflowlabel()) & IPV6_FLOWLABEL_MASK); 408#else 409 (htonl(ip6_flow_seq++) & IPV6_FLOWLABEL_MASK); 410#endif 411 412 in_pcbrehash(inp); 413 return (0); 414} 415 416#if 0 417/* 418 * Return an IPv6 address, which is the most appropriate for given 419 * destination and user specified options. 420 * If necessary, this function lookups the routing table and return 421 * an entry to the caller for later use. 422 */ 423struct in6_addr * 424in6_selectsrc(dstsock, opts, mopts, ro, laddr, errorp) 425 struct sockaddr_in6 *dstsock; 426 struct ip6_pktopts *opts; 427 struct ip6_moptions *mopts; 428 struct route_in6 *ro; 429 struct in6_addr *laddr; 430 int *errorp; 431{ 432 struct in6_addr *dst; 433 struct in6_ifaddr *ia6 = 0; 434 struct in6_pktinfo *pi = NULL; 435 436 dst = &dstsock->sin6_addr; 437 *errorp = 0; 438 439 /* 440 * If the source address is explicitly specified by the caller, 441 * use it. 442 */ 443 if (opts && (pi = opts->ip6po_pktinfo) && 444 !IN6_IS_ADDR_UNSPECIFIED(&pi->ipi6_addr)) 445 return(&pi->ipi6_addr); 446 447 /* 448 * If the source address is not specified but the socket(if any) 449 * is already bound, use the bound address. 450 */ 451 if (laddr && !IN6_IS_ADDR_UNSPECIFIED(laddr)) 452 return(laddr); 453 454 /* 455 * If the caller doesn't specify the source address but 456 * the outgoing interface, use an address associated with 457 * the interface. 458 */ 459 if (pi && pi->ipi6_ifindex) { 460 /* XXX boundary check is assumed to be already done. */ 461 ia6 = in6_ifawithscope(ifnet_byindex(pi->ipi6_ifindex), dst); 462 if (ia6 == 0) { 463 *errorp = EADDRNOTAVAIL; 464 return(0); 465 } 466 return(&satosin6(&ia6->ia_addr)->sin6_addr); 467 } 468 469 /* 470 * If the destination address is a link-local unicast address or 471 * a multicast address, and if the outgoing interface is specified 472 * by the sin6_scope_id filed, use an address associated with the 473 * interface. 474 * XXX: We're now trying to define more specific semantics of 475 * sin6_scope_id field, so this part will be rewritten in 476 * the near future. 477 */ 478 if ((IN6_IS_ADDR_LINKLOCAL(dst) || IN6_IS_ADDR_MULTICAST(dst)) && 479 dstsock->sin6_scope_id) { 480 /* 481 * I'm not sure if boundary check for scope_id is done 482 * somewhere... 483 */ 484 if (dstsock->sin6_scope_id < 0 || 485 if_index < dstsock->sin6_scope_id) { 486 *errorp = ENXIO; /* XXX: better error? */ 487 return(0); 488 } 489 ia6 = in6_ifawithscope(ifnet_byindex(dstsock->sin6_scope_id), 490 dst); 491 if (ia6 == 0) { 492 *errorp = EADDRNOTAVAIL; 493 return(0); 494 } 495 return(&satosin6(&ia6->ia_addr)->sin6_addr); 496 } 497 498 /* 499 * If the destination address is a multicast address and 500 * the outgoing interface for the address is specified 501 * by the caller, use an address associated with the interface. 502 * There is a sanity check here; if the destination has node-local 503 * scope, the outgoing interfacde should be a loopback address. 504 * Even if the outgoing interface is not specified, we also 505 * choose a loopback interface as the outgoing interface. 506 */ 507 if (IN6_IS_ADDR_MULTICAST(dst)) { 508 struct ifnet *ifp = mopts ? mopts->im6o_multicast_ifp : NULL; 509 510 if (ifp == NULL && IN6_IS_ADDR_MC_NODELOCAL(dst)) { 511 ifp = &loif[0]; 512 } 513 514 if (ifp) { 515 ia6 = in6_ifawithscope(ifp, dst); 516 if (ia6 == 0) { 517 *errorp = EADDRNOTAVAIL; 518 return(0); 519 } 520 return(&ia6->ia_addr.sin6_addr); 521 } 522 } 523 524 /* 525 * If the next hop address for the packet is specified 526 * by caller, use an address associated with the route 527 * to the next hop. 528 */ 529 { 530 struct sockaddr_in6 *sin6_next; 531 struct rtentry *rt; 532 533 if (opts && opts->ip6po_nexthop) { 534 sin6_next = satosin6(opts->ip6po_nexthop); 535 rt = nd6_lookup(&sin6_next->sin6_addr, 1, NULL); 536 if (rt) { 537 ia6 = in6_ifawithscope(rt->rt_ifp, dst); 538 if (ia6 == 0) 539 ia6 = ifatoia6(rt->rt_ifa); 540 } 541 if (ia6 == 0) { 542 *errorp = EADDRNOTAVAIL; 543 return(0); 544 } 545 return(&satosin6(&ia6->ia_addr)->sin6_addr); 546 } 547 } 548 549 /* 550 * If route is known or can be allocated now, 551 * our src addr is taken from the i/f, else punt. 552 */ 553 if (ro) { 554 if (ro->ro_rt && 555 !IN6_ARE_ADDR_EQUAL(&satosin6(&ro->ro_dst)->sin6_addr, dst)) { 556 RTFREE(ro->ro_rt); 557 ro->ro_rt = (struct rtentry *)0; 558 } 559 if (ro->ro_rt == (struct rtentry *)0 || 560 ro->ro_rt->rt_ifp == (struct ifnet *)0) { 561 struct sockaddr_in6 *dst6; 562 563 /* No route yet, so try to acquire one */ 564 bzero(&ro->ro_dst, sizeof(struct sockaddr_in6)); 565 dst6 = (struct sockaddr_in6 *)&ro->ro_dst; 566 dst6->sin6_family = AF_INET6; 567 dst6->sin6_len = sizeof(struct sockaddr_in6); 568 dst6->sin6_addr = *dst; 569 if (IN6_IS_ADDR_MULTICAST(dst)) { 570 ro->ro_rt = rtalloc1(&((struct route *)ro) 571 ->ro_dst, 0, 0UL); 572 RT_UNLOCK(ro->ro_rt); 573 } else { 574 rtalloc((struct route *)ro); 575 } 576 } 577 578 /* 579 * in_pcbconnect() checks out IFF_LOOPBACK to skip using 580 * the address. But we don't know why it does so. 581 * It is necessary to ensure the scope even for lo0 582 * so doesn't check out IFF_LOOPBACK. 583 */ 584 585 if (ro->ro_rt) { 586 ia6 = in6_ifawithscope(ro->ro_rt->rt_ifa->ifa_ifp, dst); 587 if (ia6 == 0) /* xxx scope error ?*/ 588 ia6 = ifatoia6(ro->ro_rt->rt_ifa); 589 } 590 if (ia6 == 0) { 591 *errorp = EHOSTUNREACH; /* no route */ 592 return(0); 593 } 594 return(&satosin6(&ia6->ia_addr)->sin6_addr); 595 } 596 597 *errorp = EADDRNOTAVAIL; 598 return(0); 599} 600 601/* 602 * Default hop limit selection. The precedence is as follows: 603 * 1. Hoplimit valued specified via ioctl. 604 * 2. (If the outgoing interface is detected) the current 605 * hop limit of the interface specified by router advertisement. 606 * 3. The system default hoplimit. 607*/ 608int 609in6_selecthlim(in6p, ifp) 610 struct in6pcb *in6p; 611 struct ifnet *ifp; 612{ 613 if (in6p && in6p->in6p_hops >= 0) 614 return(in6p->in6p_hops); 615 else if (ifp) 616 return(nd_ifinfo[ifp->if_index].chlim); 617 else 618 return(ip6_defhlim); 619} 620#endif 621 622void 623in6_pcbdisconnect(inp) 624 struct inpcb *inp; 625{ 626 bzero((caddr_t)&inp->in6p_faddr, sizeof(inp->in6p_faddr)); 627 inp->inp_fport = 0; 628 /* clear flowinfo - draft-itojun-ipv6-flowlabel-api-00 */ 629 inp->in6p_flowinfo &= ~IPV6_FLOWLABEL_MASK; 630 in_pcbrehash(inp); 631 if (inp->inp_socket->so_state & SS_NOFDREF) 632 in6_pcbdetach(inp); 633} 634 635void 636in6_pcbdetach(inp) 637 struct inpcb *inp; 638{ 639 struct socket *so = inp->inp_socket; 640 struct inpcbinfo *ipi = inp->inp_pcbinfo; 641 642#ifdef IPSEC 643 if (inp->in6p_sp != NULL) 644 ipsec6_delete_pcbpolicy(inp); 645#endif /* IPSEC */ 646 inp->inp_gencnt = ++ipi->ipi_gencnt; 647 in_pcbremlists(inp); 648 if (so) { 649 so->so_pcb = NULL; 650 sotryfree(so); 651 } 652 if (inp->in6p_options) 653 m_freem(inp->in6p_options); 654 ip6_freepcbopts(inp->in6p_outputopts); 655 ip6_freemoptions(inp->in6p_moptions); 656 if (inp->in6p_route.ro_rt) 657 RTFREE(inp->in6p_route.ro_rt); 658 /* Check and free IPv4 related resources in case of mapped addr */ 659 if (inp->inp_options) 660 (void)m_free(inp->inp_options); 661 ip_freemoptions(inp->inp_moptions); 662 inp->inp_vflag = 0; 663 INP_LOCK_DESTROY(inp); 664 uma_zfree(ipi->ipi_zone, inp); 665} 666 667struct sockaddr * 668in6_sockaddr(port, addr_p) 669 in_port_t port; 670 struct in6_addr *addr_p; 671{ 672 struct sockaddr_in6 *sin6; 673 674 MALLOC(sin6, struct sockaddr_in6 *, sizeof *sin6, M_SONAME, M_WAITOK); 675 bzero(sin6, sizeof *sin6); 676 sin6->sin6_family = AF_INET6; 677 sin6->sin6_len = sizeof(*sin6); 678 sin6->sin6_port = port; 679 sin6->sin6_addr = *addr_p; 680 if (IN6_IS_SCOPE_LINKLOCAL(&sin6->sin6_addr)) 681 sin6->sin6_scope_id = ntohs(sin6->sin6_addr.s6_addr16[1]); 682 else 683 sin6->sin6_scope_id = 0; /*XXX*/ 684 if (IN6_IS_SCOPE_LINKLOCAL(&sin6->sin6_addr)) 685 sin6->sin6_addr.s6_addr16[1] = 0; 686 687 return (struct sockaddr *)sin6; 688} 689 690struct sockaddr * 691in6_v4mapsin6_sockaddr(port, addr_p) 692 in_port_t port; 693 struct in_addr *addr_p; 694{ 695 struct sockaddr_in sin; 696 struct sockaddr_in6 *sin6_p; 697 698 bzero(&sin, sizeof sin); 699 sin.sin_family = AF_INET; 700 sin.sin_len = sizeof(sin); 701 sin.sin_port = port; 702 sin.sin_addr = *addr_p; 703 704 MALLOC(sin6_p, struct sockaddr_in6 *, sizeof *sin6_p, M_SONAME, 705 M_WAITOK); 706 in6_sin_2_v4mapsin6(&sin, sin6_p); 707 708 return (struct sockaddr *)sin6_p; 709} 710 711/* 712 * The calling convention of in6_setsockaddr() and in6_setpeeraddr() was 713 * modified to match the pru_sockaddr() and pru_peeraddr() entry points 714 * in struct pr_usrreqs, so that protocols can just reference then directly 715 * without the need for a wrapper function. The socket must have a valid 716 * (i.e., non-nil) PCB, but it should be impossible to get an invalid one 717 * except through a kernel programming error, so it is acceptable to panic 718 * (or in this case trap) if the PCB is invalid. (Actually, we don't trap 719 * because there actually /is/ a programming error somewhere... XXX) 720 */ 721int 722in6_setsockaddr(so, nam) 723 struct socket *so; 724 struct sockaddr **nam; 725{ 726 int s; 727 register struct inpcb *inp; 728 struct in6_addr addr; 729 in_port_t port; 730 731 s = splnet(); 732 inp = sotoinpcb(so); 733 if (!inp) { 734 splx(s); 735 return EINVAL; 736 } 737 port = inp->inp_lport; 738 addr = inp->in6p_laddr; 739 splx(s); 740 741 *nam = in6_sockaddr(port, &addr); 742 return 0; 743} 744 745int 746in6_setpeeraddr(so, nam) 747 struct socket *so; 748 struct sockaddr **nam; 749{ 750 int s; 751 struct inpcb *inp; 752 struct in6_addr addr; 753 in_port_t port; 754 755 s = splnet(); 756 inp = sotoinpcb(so); 757 if (!inp) { 758 splx(s); 759 return EINVAL; 760 } 761 port = inp->inp_fport; 762 addr = inp->in6p_faddr; 763 splx(s); 764 765 *nam = in6_sockaddr(port, &addr); 766 return 0; 767} 768 769int 770in6_mapped_sockaddr(struct socket *so, struct sockaddr **nam) 771{ 772 struct inpcb *inp = sotoinpcb(so); 773 int error; 774 775 if (inp == NULL) 776 return EINVAL; 777 if (inp->inp_vflag & INP_IPV4) { 778 error = in_setsockaddr(so, nam, &tcbinfo); 779 if (error == 0) 780 in6_sin_2_v4mapsin6_in_sock(nam); 781 } else 782 /* scope issues will be handled in in6_setsockaddr(). */ 783 error = in6_setsockaddr(so, nam); 784 785 return error; 786} 787 788int 789in6_mapped_peeraddr(struct socket *so, struct sockaddr **nam) 790{ 791 struct inpcb *inp = sotoinpcb(so); 792 int error; 793 794 if (inp == NULL) 795 return EINVAL; 796 if (inp->inp_vflag & INP_IPV4) { 797 error = in_setpeeraddr(so, nam, &tcbinfo); 798 if (error == 0) 799 in6_sin_2_v4mapsin6_in_sock(nam); 800 } else 801 /* scope issues will be handled in in6_setpeeraddr(). */ 802 error = in6_setpeeraddr(so, nam); 803 804 return error; 805} 806 807/* 808 * Pass some notification to all connections of a protocol 809 * associated with address dst. The local address and/or port numbers 810 * may be specified to limit the search. The "usual action" will be 811 * taken, depending on the ctlinput cmd. The caller must filter any 812 * cmds that are uninteresting (e.g., no error in the map). 813 * Call the protocol specific routine (if any) to report 814 * any errors for each matching socket. 815 * 816 * Must be called at splnet. 817 */ 818void 819in6_pcbnotify(head, dst, fport_arg, src, lport_arg, cmd, notify) 820 struct inpcbhead *head; 821 struct sockaddr *dst; 822 const struct sockaddr *src; 823 u_int fport_arg, lport_arg; 824 int cmd; 825 struct inpcb *(*notify) __P((struct inpcb *, int)); 826{ 827 struct inpcb *inp, *ninp; 828 struct sockaddr_in6 sa6_src, *sa6_dst; 829 u_short fport = fport_arg, lport = lport_arg; 830 u_int32_t flowinfo; 831 int errno, s; 832 833 if ((unsigned)cmd >= PRC_NCMDS || dst->sa_family != AF_INET6) 834 return; 835 836 sa6_dst = (struct sockaddr_in6 *)dst; 837 if (IN6_IS_ADDR_UNSPECIFIED(&sa6_dst->sin6_addr)) 838 return; 839 840 /* 841 * note that src can be NULL when we get notify by local fragmentation. 842 */ 843 sa6_src = (src == NULL) ? sa6_any : *(const struct sockaddr_in6 *)src; 844 flowinfo = sa6_src.sin6_flowinfo; 845 846 /* 847 * Redirects go to all references to the destination, 848 * and use in6_rtchange to invalidate the route cache. 849 * Dead host indications: also use in6_rtchange to invalidate 850 * the cache, and deliver the error to all the sockets. 851 * Otherwise, if we have knowledge of the local port and address, 852 * deliver only to that socket. 853 */ 854 if (PRC_IS_REDIRECT(cmd) || cmd == PRC_HOSTDEAD) { 855 fport = 0; 856 lport = 0; 857 bzero((caddr_t)&sa6_src.sin6_addr, sizeof(sa6_src.sin6_addr)); 858 859 if (cmd != PRC_HOSTDEAD) 860 notify = in6_rtchange; 861 } 862 errno = inet6ctlerrmap[cmd]; 863 s = splnet(); 864 for (inp = LIST_FIRST(head); inp != NULL; inp = ninp) { 865 ninp = LIST_NEXT(inp, inp_list); 866 867 if ((inp->inp_vflag & INP_IPV6) == 0) 868 continue; 869 870 /* 871 * Detect if we should notify the error. If no source and 872 * destination ports are specifed, but non-zero flowinfo and 873 * local address match, notify the error. This is the case 874 * when the error is delivered with an encrypted buffer 875 * by ESP. Otherwise, just compare addresses and ports 876 * as usual. 877 */ 878 if (lport == 0 && fport == 0 && flowinfo && 879 inp->inp_socket != NULL && 880 flowinfo == (inp->in6p_flowinfo & IPV6_FLOWLABEL_MASK) && 881 IN6_ARE_ADDR_EQUAL(&inp->in6p_laddr, &sa6_src.sin6_addr)) 882 goto do_notify; 883 else if (!IN6_ARE_ADDR_EQUAL(&inp->in6p_faddr, 884 &sa6_dst->sin6_addr) || 885 inp->inp_socket == 0 || 886 (lport && inp->inp_lport != lport) || 887 (!IN6_IS_ADDR_UNSPECIFIED(&sa6_src.sin6_addr) && 888 !IN6_ARE_ADDR_EQUAL(&inp->in6p_laddr, 889 &sa6_src.sin6_addr)) || 890 (fport && inp->inp_fport != fport)) 891 continue; 892 893 do_notify: 894 if (notify) 895 (*notify)(inp, errno); 896 } 897 splx(s); 898} 899 900/* 901 * Lookup a PCB based on the local address and port. 902 */ 903struct inpcb * 904in6_pcblookup_local(pcbinfo, laddr, lport_arg, wild_okay) 905 struct inpcbinfo *pcbinfo; 906 struct in6_addr *laddr; 907 u_int lport_arg; 908 int wild_okay; 909{ 910 register struct inpcb *inp; 911 int matchwild = 3, wildcard; 912 u_short lport = lport_arg; 913 914 if (!wild_okay) { 915 struct inpcbhead *head; 916 /* 917 * Look for an unconnected (wildcard foreign addr) PCB that 918 * matches the local address and port we're looking for. 919 */ 920 head = &pcbinfo->hashbase[INP_PCBHASH(INADDR_ANY, lport, 0, 921 pcbinfo->hashmask)]; 922 LIST_FOREACH(inp, head, inp_hash) { 923 if ((inp->inp_vflag & INP_IPV6) == 0) 924 continue; 925 if (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_faddr) && 926 IN6_ARE_ADDR_EQUAL(&inp->in6p_laddr, laddr) && 927 inp->inp_lport == lport) { 928 /* 929 * Found. 930 */ 931 return (inp); 932 } 933 } 934 /* 935 * Not found. 936 */ 937 return (NULL); 938 } else { 939 struct inpcbporthead *porthash; 940 struct inpcbport *phd; 941 struct inpcb *match = NULL; 942 /* 943 * Best fit PCB lookup. 944 * 945 * First see if this local port is in use by looking on the 946 * port hash list. 947 */ 948 porthash = &pcbinfo->porthashbase[INP_PCBPORTHASH(lport, 949 pcbinfo->porthashmask)]; 950 LIST_FOREACH(phd, porthash, phd_hash) { 951 if (phd->phd_port == lport) 952 break; 953 } 954 if (phd != NULL) { 955 /* 956 * Port is in use by one or more PCBs. Look for best 957 * fit. 958 */ 959 LIST_FOREACH(inp, &phd->phd_pcblist, inp_portlist) { 960 wildcard = 0; 961 if ((inp->inp_vflag & INP_IPV6) == 0) 962 continue; 963 if (!IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_faddr)) 964 wildcard++; 965 if (!IN6_IS_ADDR_UNSPECIFIED( 966 &inp->in6p_laddr)) { 967 if (IN6_IS_ADDR_UNSPECIFIED(laddr)) 968 wildcard++; 969 else if (!IN6_ARE_ADDR_EQUAL( 970 &inp->in6p_laddr, laddr)) 971 continue; 972 } else { 973 if (!IN6_IS_ADDR_UNSPECIFIED(laddr)) 974 wildcard++; 975 } 976 if (wildcard < matchwild) { 977 match = inp; 978 matchwild = wildcard; 979 if (matchwild == 0) { 980 break; 981 } 982 } 983 } 984 } 985 return (match); 986 } 987} 988 989void 990in6_pcbpurgeif0(head, ifp) 991 struct in6pcb *head; 992 struct ifnet *ifp; 993{ 994 struct in6pcb *in6p; 995 struct ip6_moptions *im6o; 996 struct in6_multi_mship *imm, *nimm; 997 998 for (in6p = head; in6p != NULL; in6p = LIST_NEXT(in6p, inp_list)) { 999 im6o = in6p->in6p_moptions; 1000 if ((in6p->inp_vflag & INP_IPV6) && 1001 im6o) { 1002 /* 1003 * Unselect the outgoing interface if it is being 1004 * detached. 1005 */ 1006 if (im6o->im6o_multicast_ifp == ifp) 1007 im6o->im6o_multicast_ifp = NULL; 1008 1009 /* 1010 * Drop multicast group membership if we joined 1011 * through the interface being detached. 1012 * XXX controversial - is it really legal for kernel 1013 * to force this? 1014 */ 1015 for (imm = im6o->im6o_memberships.lh_first; 1016 imm != NULL; imm = nimm) { 1017 nimm = imm->i6mm_chain.le_next; 1018 if (imm->i6mm_maddr->in6m_ifp == ifp) { 1019 LIST_REMOVE(imm, i6mm_chain); 1020 in6_delmulti(imm->i6mm_maddr); 1021 free(imm, M_IPMADDR); 1022 } 1023 } 1024 } 1025 } 1026} 1027 1028/* 1029 * Check for alternatives when higher level complains 1030 * about service problems. For now, invalidate cached 1031 * routing information. If the route was created dynamically 1032 * (by a redirect), time to try a default gateway again. 1033 */ 1034void 1035in6_losing(in6p) 1036 struct inpcb *in6p; 1037{ 1038 struct rtentry *rt; 1039 struct rt_addrinfo info; 1040 1041 if ((rt = in6p->in6p_route.ro_rt) != NULL) { 1042 RT_LOCK(rt); 1043 in6p->in6p_route.ro_rt = NULL; 1044 bzero((caddr_t)&info, sizeof(info)); 1045 info.rti_flags = rt->rt_flags; 1046 info.rti_info[RTAX_DST] = rt_key(rt); 1047 info.rti_info[RTAX_GATEWAY] = rt->rt_gateway; 1048 info.rti_info[RTAX_NETMASK] = rt_mask(rt); 1049 rt_missmsg(RTM_LOSING, &info, rt->rt_flags, 0); 1050 if (rt->rt_flags & RTF_DYNAMIC) { 1051 RT_UNLOCK(rt); /* XXX refcnt? */ 1052 (void)rtrequest1(RTM_DELETE, &info, NULL); 1053 } else 1054 rtfree(rt); 1055 /* 1056 * A new route can be allocated 1057 * the next time output is attempted. 1058 */ 1059 } 1060} 1061 1062/* 1063 * After a routing change, flush old routing 1064 * and allocate a (hopefully) better one. 1065 */ 1066struct inpcb * 1067in6_rtchange(inp, errno) 1068 struct inpcb *inp; 1069 int errno; 1070{ 1071 if (inp->in6p_route.ro_rt) { 1072 RTFREE(inp->in6p_route.ro_rt); 1073 inp->in6p_route.ro_rt = 0; 1074 /* 1075 * A new route can be allocated the next time 1076 * output is attempted. 1077 */ 1078 } 1079 return inp; 1080} 1081 1082/* 1083 * Lookup PCB in hash list. 1084 */ 1085struct inpcb * 1086in6_pcblookup_hash(pcbinfo, faddr, fport_arg, laddr, lport_arg, wildcard, ifp) 1087 struct inpcbinfo *pcbinfo; 1088 struct in6_addr *faddr, *laddr; 1089 u_int fport_arg, lport_arg; 1090 int wildcard; 1091 struct ifnet *ifp; 1092{ 1093 struct inpcbhead *head; 1094 register struct inpcb *inp; 1095 u_short fport = fport_arg, lport = lport_arg; 1096 int faith; 1097 1098 if (faithprefix_p != NULL) 1099 faith = (*faithprefix_p)(laddr); 1100 else 1101 faith = 0; 1102 1103 /* 1104 * First look for an exact match. 1105 */ 1106 head = &pcbinfo->hashbase[INP_PCBHASH(faddr->s6_addr32[3] /* XXX */, 1107 lport, fport, 1108 pcbinfo->hashmask)]; 1109 LIST_FOREACH(inp, head, inp_hash) { 1110 if ((inp->inp_vflag & INP_IPV6) == 0) 1111 continue; 1112 if (IN6_ARE_ADDR_EQUAL(&inp->in6p_faddr, faddr) && 1113 IN6_ARE_ADDR_EQUAL(&inp->in6p_laddr, laddr) && 1114 inp->inp_fport == fport && 1115 inp->inp_lport == lport) { 1116 /* 1117 * Found. 1118 */ 1119 return (inp); 1120 } 1121 } 1122 if (wildcard) { 1123 struct inpcb *local_wild = NULL; 1124 1125 head = &pcbinfo->hashbase[INP_PCBHASH(INADDR_ANY, lport, 0, 1126 pcbinfo->hashmask)]; 1127 LIST_FOREACH(inp, head, inp_hash) { 1128 if ((inp->inp_vflag & INP_IPV6) == 0) 1129 continue; 1130 if (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_faddr) && 1131 inp->inp_lport == lport) { 1132 if (faith && (inp->inp_flags & INP_FAITH) == 0) 1133 continue; 1134 if (IN6_ARE_ADDR_EQUAL(&inp->in6p_laddr, 1135 laddr)) 1136 return (inp); 1137 else if (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr)) 1138 local_wild = inp; 1139 } 1140 } 1141 return (local_wild); 1142 } 1143 1144 /* 1145 * Not found. 1146 */ 1147 return (NULL); 1148} 1149 1150void 1151init_sin6(struct sockaddr_in6 *sin6, struct mbuf *m) 1152{ 1153 struct ip6_hdr *ip; 1154 1155 ip = mtod(m, struct ip6_hdr *); 1156 bzero(sin6, sizeof(*sin6)); 1157 sin6->sin6_len = sizeof(*sin6); 1158 sin6->sin6_family = AF_INET6; 1159 sin6->sin6_addr = ip->ip6_src; 1160 if (IN6_IS_SCOPE_LINKLOCAL(&sin6->sin6_addr)) 1161 sin6->sin6_addr.s6_addr16[1] = 0; 1162 sin6->sin6_scope_id = 1163 (m->m_pkthdr.rcvif && IN6_IS_SCOPE_LINKLOCAL(&sin6->sin6_addr)) 1164 ? m->m_pkthdr.rcvif->if_index : 0; 1165 1166 return; 1167} 1168