bsd/netinet6/in6_src.c

   1 /*      $FreeBSD: src/sys/netinet6/in6_src.c,v 1.1.2.2 2001/07/03 11:01:52 ume Exp $    */
   2 /*      $KAME: in6_src.c,v 1.37 2001/03/29 05:34:31 itojun Exp $        */
   3
   4 /*
   5  * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
   6  * All rights reserved.
   7  *
   8  * Redistribution and use in source and binary forms, with or without
   9  * modification, are permitted provided that the following conditions
  10  * are met:
  11  * 1. Redistributions of source code must retain the above copyright
  12  *    notice, this list of conditions and the following disclaimer.
  13  * 2. Redistributions in binary form must reproduce the above copyright
  14  *    notice, this list of conditions and the following disclaimer in the
  15  *    documentation and/or other materials provided with the distribution.
  16  * 3. Neither the name of the project nor the names of its contributors
  17  *    may be used to endorse or promote products derived from this software
  18  *    without specific prior written permission.
  19  *
  20  * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
  21  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  22  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  23  * ARE DISCLAIMED.  IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
  24  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  25  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  26  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  27  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  28  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  29  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  30  * SUCH DAMAGE.
  31  */
  32
  33 /*
  34  * Copyright (c) 1982, 1986, 1991, 1993
  35  *      The Regents of the University of California.  All rights reserved.
  36  *
  37  * Redistribution and use in source and binary forms, with or without
  38  * modification, are permitted provided that the following conditions
  39  * are met:
  40  * 1. Redistributions of source code must retain the above copyright
  41  *    notice, this list of conditions and the following disclaimer.
  42  * 2. Redistributions in binary form must reproduce the above copyright
  43  *    notice, this list of conditions and the following disclaimer in the
  44  *    documentation and/or other materials provided with the distribution.
  45  * 3. All advertising materials mentioning features or use of this software
  46  *    must display the following acknowledgement:
  47  *      This product includes software developed by the University of
  48  *      California, Berkeley and its contributors.
  49  * 4. Neither the name of the University nor the names of its contributors
  50  *    may be used to endorse or promote products derived from this software
  51  *    without specific prior written permission.
  52  *
  53  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  54  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  55  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  56  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  57  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  58  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  59  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  60  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  61  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  62  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  63  * SUCH DAMAGE.
  64  *
  65  *      @(#)in_pcb.c    8.2 (Berkeley) 1/4/94
  66  */
  67
  68
  69 #include <sys/param.h>
  70 #include <sys/systm.h>
  71 #include <sys/malloc.h>
  72 #include <sys/mbuf.h>
  73 #include <sys/protosw.h>
  74 #include <sys/socket.h>
  75 #include <sys/socketvar.h>
  76 #include <sys/errno.h>
  77 #include <sys/time.h>
  78 #include <sys/proc.h>
  79
  80 #include <net/if.h>
  81 #include <net/route.h>
  82
  83 #include <netinet/in.h>
  84 #include <netinet/in_var.h>
  85 #include <netinet/in_systm.h>
  86 #include <netinet/ip.h>
  87 #include <netinet/in_pcb.h>
  88 #include <netinet6/in6_var.h>
  89 #include <netinet/ip6.h>
  90 #include <netinet6/in6_pcb.h>
  91 #include <netinet6/ip6_var.h>
  92 #include <netinet6/nd6.h>
  93 #if ENABLE_DEFAULT_SCOPE
  94 #include <netinet6/scope6_var.h>
  95 #endif
  96
  97 #include <net/net_osdep.h>
  98
  99 #include "loop.h"
 100
 101 /*
 102  * Return an IPv6 address, which is the most appropriate for a given
 103  * destination and user specified options.
 104  * If necessary, this function lookups the routing table and returns
 105  * an entry to the caller for later use.
 106  */
 107 struct in6_addr *
 108 in6_selectsrc(dstsock, opts, mopts, ro, laddr, errorp)
 109         struct sockaddr_in6 *dstsock;
 110         struct ip6_pktopts *opts;
 111         struct ip6_moptions *mopts;
 112         struct route_in6 *ro;
 113         struct in6_addr *laddr;
 114         int *errorp;
 115 {
 116         struct in6_addr *dst;
 117         struct in6_ifaddr *ia6 = 0;
 118         struct in6_pktinfo *pi = NULL;
 119
 120         dst = &dstsock->sin6_addr;
 121         *errorp = 0;
 122
 123         /*
 124          * If the source address is explicitly specified by the caller,
 125          * use it.
 126          */
 127         if (opts && (pi = opts->ip6po_pktinfo) &&
 128             !IN6_IS_ADDR_UNSPECIFIED(&pi->ipi6_addr))
 129                 return(&pi->ipi6_addr);
 130
 131         /*
 132          * If the source address is not specified but the socket(if any)
 133          * is already bound, use the bound address.
 134          */
 135         if (laddr && !IN6_IS_ADDR_UNSPECIFIED(laddr))
 136                 return(laddr);
 137
 138         /*
 139          * If the caller doesn't specify the source address but
 140          * the outgoing interface, use an address associated with
 141          * the interface.
 142          */
 143         if (pi && pi->ipi6_ifindex) {
 144                 /* XXX boundary check is assumed to be already done. */
 145                 ia6 = in6_ifawithscope(ifindex2ifnet[pi->ipi6_ifindex],
 146                                        dst);
 147                 if (ia6 == 0) {
 148                         *errorp = EADDRNOTAVAIL;
 149                         return(0);
 150                 }
 151                 return(&satosin6(&ia6->ia_addr)->sin6_addr);
 152         }
 153
 154         /*
 155          * If the destination address is a link-local unicast address or
 156          * a multicast address, and if the outgoing interface is specified
 157          * by the sin6_scope_id filed, use an address associated with the
 158          * interface.
 159          * XXX: We're now trying to define more specific semantics of
 160          *      sin6_scope_id field, so this part will be rewritten in
 161          *      the near future.
 162          */
 163         if ((IN6_IS_ADDR_LINKLOCAL(dst) || IN6_IS_ADDR_MULTICAST(dst)) &&
 164             dstsock->sin6_scope_id) {
 165                 /*
 166                  * I'm not sure if boundary check for scope_id is done
 167                  * somewhere...
 168                  */
 169                 if (dstsock->sin6_scope_id < 0 ||
 170                     if_index < dstsock->sin6_scope_id) {
 171                         *errorp = ENXIO; /* XXX: better error? */
 172                         return(0);
 173                 }
 174                 ia6 = in6_ifawithscope(ifindex2ifnet[dstsock->sin6_scope_id],
 175                                        dst);
 176                 if (ia6 == 0) {
 177                         *errorp = EADDRNOTAVAIL;
 178                         return(0);
 179                 }
 180                 return(&satosin6(&ia6->ia_addr)->sin6_addr);
 181         }
 182
 183         /*
 184          * If the destination address is a multicast address and
 185          * the outgoing interface for the address is specified
 186          * by the caller, use an address associated with the interface.
 187          * There is a sanity check here; if the destination has node-local
 188          * scope, the outgoing interfacde should be a loopback address.
 189          * Even if the outgoing interface is not specified, we also
 190          * choose a loopback interface as the outgoing interface.
 191          */
 192         if (IN6_IS_ADDR_MULTICAST(dst)) {
 193                 struct ifnet *ifp = mopts ? mopts->im6o_multicast_ifp : NULL;
 194
 195                 if (ifp == NULL && IN6_IS_ADDR_MC_NODELOCAL(dst)) {
 196                         ifp = &loif[0];
 197                 }
 198
 199                 if (ifp) {
 200                         ia6 = in6_ifawithscope(ifp, dst);
 201                         if (ia6 == 0) {
 202                                 *errorp = EADDRNOTAVAIL;
 203                                 return(0);
 204                         }
 205                         return(&satosin6(&ia6->ia_addr)->sin6_addr);
 206                 }
 207         }
 208
 209         /*
 210          * If the next hop address for the packet is specified
 211          * by caller, use an address associated with the route
 212          * to the next hop.
 213          */
 214         {
 215                 struct sockaddr_in6 *sin6_next;
 216                 struct rtentry *rt;
 217
 218                 if (opts && opts->ip6po_nexthop) {
 219                         sin6_next = satosin6(opts->ip6po_nexthop);
 220                         rt = nd6_lookup(&sin6_next->sin6_addr, 1, NULL);
 221                         if (rt) {
 222                                 ia6 = in6_ifawithscope(rt->rt_ifp, dst);
 223                                 if (ia6 == 0)
 224                                         ia6 = ifatoia6(rt->rt_ifa);
 225                         }
 226                         if (ia6 == 0) {
 227                                 *errorp = EADDRNOTAVAIL;
 228                                 return(0);
 229                         }
 230                         return(&satosin6(&ia6->ia_addr)->sin6_addr);
 231                 }
 232         }
 233
 234         /*
 235          * If route is known or can be allocated now,
 236          * our src addr is taken from the i/f, else punt.
 237          */
 238         if (ro) {
 239                 if (ro->ro_rt &&
 240                     !IN6_ARE_ADDR_EQUAL(&satosin6(&ro->ro_dst)->sin6_addr, dst)) {
 241                         rtfree(ro->ro_rt);
 242                         ro->ro_rt = (struct rtentry *)0;
 243                 }
 244                 if (ro->ro_rt == (struct rtentry *)0 ||
 245                     ro->ro_rt->rt_ifp == (struct ifnet *)0) {
 246                         struct sockaddr_in6 *sa6;
 247
 248                         /* No route yet, so try to acquire one */
 249                         bzero(&ro->ro_dst, sizeof(struct sockaddr_in6));
 250                         sa6 = (struct sockaddr_in6 *)&ro->ro_dst;
 251                         sa6->sin6_family = AF_INET6;
 252                         sa6->sin6_len = sizeof(struct sockaddr_in6);
 253                         sa6->sin6_addr = *dst;
 254                         sa6->sin6_scope_id = dstsock->sin6_scope_id;
 255                         if (IN6_IS_ADDR_MULTICAST(dst)) {
 256                                 ro->ro_rt = rtalloc1(&((struct route *)ro)
 257                                                      ->ro_dst, 0, 0UL);
 258                         } else {
 259                                 rtalloc((struct route *)ro);
 260                         }
 261                 }
 262
 263                 /*
 264                  * in_pcbconnect() checks out IFF_LOOPBACK to skip using
 265                  * the address. But we don't know why it does so.
 266                  * It is necessary to ensure the scope even for lo0
 267                  * so doesn't check out IFF_LOOPBACK.
 268                  */
 269
 270                 if (ro->ro_rt) {
 271                         ia6 = in6_ifawithscope(ro->ro_rt->rt_ifa->ifa_ifp, dst);
 272                         if (ia6 == 0) /* xxx scope error ?*/
 273                                 ia6 = ifatoia6(ro->ro_rt->rt_ifa);
 274                 }
 275 #if 0
 276                 /*
 277                  * xxx The followings are necessary? (kazu)
 278                  * I don't think so.
 279                  * It's for SO_DONTROUTE option in IPv4.(jinmei)
 280                  */
 281                 if (ia6 == 0) {
 282                         struct sockaddr_in6 sin6 = {sizeof(sin6), AF_INET6, 0};
 283
 284                         sin6->sin6_addr = *dst;
 285
 286                         ia6 = ifatoia6(ifa_ifwithdstaddr(sin6tosa(&sin6)));
 287                         if (ia6 == 0)
 288                                 ia6 = ifatoia6(ifa_ifwithnet(sin6tosa(&sin6)));
 289                         if (ia6 == 0)
 290                                 return(0);
 291                         return(&satosin6(&ia6->ia_addr)->sin6_addr);
 292                 }
 293 #endif /* 0 */
 294                 if (ia6 == 0) {
 295                         *errorp = EHOSTUNREACH; /* no route */
 296                         return(0);
 297                 }
 298                 return(&satosin6(&ia6->ia_addr)->sin6_addr);
 299         }
 300
 301         *errorp = EADDRNOTAVAIL;
 302         return(0);
 303 }
 304
 305 /*
 306  * Default hop limit selection. The precedence is as follows:
 307  * 1. Hoplimit value specified via ioctl.
 308  * 2. (If the outgoing interface is detected) the current
 309  *     hop limit of the interface specified by router advertisement.
 310  * 3. The system default hoplimit.
 311 */
 312 int
 313 in6_selecthlim(in6p, ifp)
 314         struct in6pcb *in6p;
 315         struct ifnet *ifp;
 316 {
 317         if (in6p && in6p->in6p_hops >= 0)
 318                 return(in6p->in6p_hops);
 319         else if (ifp)
 320                 return(nd_ifinfo[ifp->if_index].chlim);
 321         else
 322                 return(ip6_defhlim);
 323 }
 324
 325 /*
 326  * XXX: this is borrowed from in6_pcbbind(). If possible, we should
 327  * share this function by all *bsd*...
 328  */
 329 int
 330 in6_pcbsetport(laddr, inp, p)
 331         struct in6_addr *laddr;
 332         struct inpcb *inp;
 333         struct proc *p;
 334 {
 335         struct socket *so = inp->inp_socket;
 336         u_int16_t lport = 0, first, last, *lastport;
 337         int count, error = 0, wild = 0;
 338         struct inpcbinfo *pcbinfo = inp->inp_pcbinfo;
 339
 340         /* XXX: this is redundant when called from in6_pcbbind */
 341         if ((so->so_options & (SO_REUSEADDR|SO_REUSEPORT)) == 0)
 342                 wild = INPLOOKUP_WILDCARD;
 343
 344         inp->inp_flags |= INP_ANONPORT;
 345
 346         if (inp->inp_flags & INP_HIGHPORT) {
 347                 first = ipport_hifirstauto;     /* sysctl */
 348                 last  = ipport_hilastauto;
 349                 lastport = &pcbinfo->lasthi;
 350         } else if (inp->inp_flags & INP_LOWPORT) {
 351 #ifdef __APPLE__
 352                 if (p && (error = suser(p->p_ucred, &p->p_acflag)))
 353 #else
 354                 if (p && (error = suser(p)))
 355 #endif
 356
 357                         return error;
 358                 first = ipport_lowfirstauto;    /* 1023 */
 359                 last  = ipport_lowlastauto;     /* 600 */
 360                 lastport = &pcbinfo->lastlow;
 361         } else {
 362                 first = ipport_firstauto;       /* sysctl */
 363                 last  = ipport_lastauto;
 364                 lastport = &pcbinfo->lastport;
 365         }
 366         /*
 367          * Simple check to ensure all ports are not used up causing
 368          * a deadlock here.
 369          *
 370          * We split the two cases (up and down) so that the direction
 371          * is not being tested on each round of the loop.
 372          */
 373         if (first > last) {
 374                 /*
 375                  * counting down
 376                  */
 377                 count = first - last;
 378
 379                 do {
 380                         if (count-- < 0) {      /* completely used? */
 381                                 /*
 382                                  * Undo any address bind that may have
 383                                  * occurred above.
 384                                  */
 385                                 inp->in6p_laddr = in6addr_any;
 386                                 return (EAGAIN);
 387                         }
 388                         --*lastport;
 389                         if (*lastport > first || *lastport < last)
 390                                 *lastport = first;
 391                         lport = htons(*lastport);
 392                 } while (in6_pcblookup_local(pcbinfo,
 393                                              &inp->in6p_laddr, lport, wild));
 394         } else {
 395                 /*
 396                          * counting up
 397                          */
 398                 count = last - first;
 399
 400                 do {
 401                         if (count-- < 0) {      /* completely used? */
 402                                 /*
 403                                  * Undo any address bind that may have
 404                                  * occurred above.
 405                                  */
 406                                 inp->in6p_laddr = in6addr_any;
 407                                 return (EAGAIN);
 408                         }
 409                         ++*lastport;
 410                         if (*lastport < first || *lastport > last)
 411                                 *lastport = first;
 412                         lport = htons(*lastport);
 413                 } while (in6_pcblookup_local(pcbinfo,
 414                                              &inp->in6p_laddr, lport, wild));
 415         }
 416
 417         inp->inp_lport = lport;
 418         if (in_pcbinshash(inp) != 0) {
 419                 inp->in6p_laddr = in6addr_any;
 420                 inp->inp_lport = 0;
 421                 return (EAGAIN);
 422         }
 423
 424         return(0);
 425 }
 426
 427 /*
 428  * generate kernel-internal form (scopeid embedded into s6_addr16[1]).
 429  * If the address scope of is link-local, embed the interface index in the
 430  * address.  The routine determines our precedence
 431  * between advanced API scope/interface specification and basic API
 432  * specification.
 433  *
 434  * this function should be nuked in the future, when we get rid of
 435  * embedded scopeid thing.
 436  *
 437  * XXX actually, it is over-specification to return ifp against sin6_scope_id.
 438  * there can be multiple interfaces that belong to a particular scope zone
 439  * (in specification, we have 1:N mapping between a scope zone and interfaces).
 440  * we may want to change the function to return something other than ifp.
 441  */
 442 int
 443 in6_embedscope(in6, sin6, in6p, ifpp)
 444         struct in6_addr *in6;
 445         const struct sockaddr_in6 *sin6;
 446 #ifdef HAVE_NRL_INPCB
 447         struct inpcb *in6p;
 448 #define in6p_outputopts inp_outputopts6
 449 #define in6p_moptions   inp_moptions6
 450 #else
 451         struct in6pcb *in6p;
 452 #endif
 453         struct ifnet **ifpp;
 454 {
 455         struct ifnet *ifp = NULL;
 456         u_int32_t scopeid;
 457
 458         *in6 = sin6->sin6_addr;
 459         scopeid = sin6->sin6_scope_id;
 460         if (ifpp)
 461                 *ifpp = NULL;
 462
 463         /*
 464          * don't try to read sin6->sin6_addr beyond here, since the caller may
 465          * ask us to overwrite existing sockaddr_in6
 466          */
 467
 468 #ifdef ENABLE_DEFAULT_SCOPE
 469         if (scopeid == 0)
 470                 scopeid = scope6_addr2default(in6);
 471 #endif
 472
 473         if (IN6_IS_SCOPE_LINKLOCAL(in6)) {
 474                 struct in6_pktinfo *pi;
 475
 476                 /*
 477                  * KAME assumption: link id == interface id
 478                  */
 479
 480                 if (in6p && in6p->in6p_outputopts &&
 481                     (pi = in6p->in6p_outputopts->ip6po_pktinfo) &&
 482                     pi->ipi6_ifindex) {
 483                         ifp = ifindex2ifnet[pi->ipi6_ifindex];
 484                         in6->s6_addr16[1] = htons(pi->ipi6_ifindex);
 485                 } else if (in6p && IN6_IS_ADDR_MULTICAST(in6) &&
 486                            in6p->in6p_moptions &&
 487                            in6p->in6p_moptions->im6o_multicast_ifp) {
 488                         ifp = in6p->in6p_moptions->im6o_multicast_ifp;
 489                         in6->s6_addr16[1] = htons(ifp->if_index);
 490                 } else if (scopeid) {
 491                         /* boundary check */
 492                         if (scopeid < 0 || if_index < scopeid)
 493                                 return ENXIO;  /* XXX EINVAL? */
 494                         ifp = ifindex2ifnet[scopeid];
 495                         /*XXX assignment to 16bit from 32bit variable */
 496                         in6->s6_addr16[1] = htons(scopeid & 0xffff);
 497                 }
 498
 499                 if (ifpp)
 500                         *ifpp = ifp;
 501         }
 502
 503         return 0;
 504 }
 505 #if HAVE_NRL_INPCB
 506 #undef in6p_outputopts
 507 #undef in6p_moptions
 508 #endif
 509
 510 /*
 511  * generate standard sockaddr_in6 from embedded form.
 512  * touches sin6_addr and sin6_scope_id only.
 513  *
 514  * this function should be nuked in the future, when we get rid of
 515  * embedded scopeid thing.
 516  */
 517 int
 518 in6_recoverscope(sin6, in6, ifp)
 519         struct sockaddr_in6 *sin6;
 520         const struct in6_addr *in6;
 521         struct ifnet *ifp;
 522 {
 523         u_int32_t scopeid;
 524
 525         sin6->sin6_addr = *in6;
 526
 527         /*
 528          * don't try to read *in6 beyond here, since the caller may
 529          * ask us to overwrite existing sockaddr_in6
 530          */
 531
 532         sin6->sin6_scope_id = 0;
 533         if (IN6_IS_SCOPE_LINKLOCAL(in6)) {
 534                 /*
 535                  * KAME assumption: link id == interface id
 536                  */
 537                 scopeid = ntohs(sin6->sin6_addr.s6_addr16[1]);
 538                 if (scopeid) {
 539                         /* sanity check */
 540                         if (scopeid < 0 || if_index < scopeid)
 541                                 return ENXIO;
 542                         if (ifp && ifp->if_index != scopeid)
 543                                 return ENXIO;
 544                         sin6->sin6_addr.s6_addr16[1] = 0;
 545                         sin6->sin6_scope_id = scopeid;
 546                 }
 547         }
 548
 549         return 0;
 550 }
 551
 552 /*
 553  * just clear the embedded scope identifer.
 554  * XXX: currently used for bsdi4 only as a supplement function.
 555  */
 556 void
 557 in6_clearscope(addr)
 558         struct in6_addr *addr;
 559 {
 560         if (IN6_IS_SCOPE_LINKLOCAL(addr))
 561                 addr->s6_addr16[1] = 0;
 562 }