bsd/netinet/ip_encap.c

   1 /*
   2  * Copyright (c) 2000 Apple Computer, Inc. All rights reserved.
   3  *
   4  * @APPLE_LICENSE_OSREFERENCE_HEADER_START@
   5  *
   6  * This file contains Original Code and/or Modifications of Original Code
   7  * as defined in and that are subject to the Apple Public Source License
   8  * Version 2.0 (the 'License'). You may not use this file except in
   9  * compliance with the License.  The rights granted to you under the
  10  * License may not be used to create, or enable the creation or
  11  * redistribution of, unlawful or unlicensed copies of an Apple operating
  12  * system, or to circumvent, violate, or enable the circumvention or
  13  * violation of, any terms of an Apple operating system software license
  14  * agreement.
  15  *
  16  * Please obtain a copy of the License at
  17  * http://www.opensource.apple.com/apsl/ and read it before using this
  18  * file.
  19  *
  20  * The Original Code and all software distributed under the License are
  21  * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
  22  * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
  23  * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
  24  * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
  25  * Please see the License for the specific language governing rights and
  26  * limitations under the License.
  27  *
  28  * @APPLE_LICENSE_OSREFERENCE_HEADER_END@
  29  */
  30 /*      $FreeBSD: src/sys/netinet/ip_encap.c,v 1.1.2.2 2001/07/03 11:01:46 ume Exp $    */
  31 /*      $KAME: ip_encap.c,v 1.41 2001/03/15 08:35:08 itojun Exp $       */
  32
  33 /*
  34  * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
  35  * 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. Neither the name of the project nor the names of its contributors
  46  *    may be used to endorse or promote products derived from this software
  47  *    without specific prior written permission.
  48  *
  49  * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
  50  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  51  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  52  * ARE DISCLAIMED.  IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
  53  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  54  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  55  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  56  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  57  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  58  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  59  * SUCH DAMAGE.
  60  */
  61 /*
  62  * My grandfather said that there's a devil inside tunnelling technology...
  63  *
  64  * We have surprisingly many protocols that want packets with IP protocol
  65  * #4 or #41.  Here's a list of protocols that want protocol #41:
  66  *      RFC1933 configured tunnel
  67  *      RFC1933 automatic tunnel
  68  *      RFC2401 IPsec tunnel
  69  *      RFC2473 IPv6 generic packet tunnelling
  70  *      RFC2529 6over4 tunnel
  71  *      mobile-ip6 (uses RFC2473)
  72  *      6to4 tunnel
  73  * Here's a list of protocol that want protocol #4:
  74  *      RFC1853 IPv4-in-IPv4 tunnelling
  75  *      RFC2003 IPv4 encapsulation within IPv4
  76  *      RFC2344 reverse tunnelling for mobile-ip4
  77  *      RFC2401 IPsec tunnel
  78  * Well, what can I say.  They impose different en/decapsulation mechanism
  79  * from each other, so they need separate protocol handler.  The only one
  80  * we can easily determine by protocol # is IPsec, which always has
  81  * AH/ESP/IPComp header right after outer IP header.
  82  *
  83  * So, clearly good old protosw does not work for protocol #4 and #41.
  84  * The code will let you match protocol via src/dst address pair.
  85  */
  86 /* XXX is M_NETADDR correct? */
  87
  88 #include <sys/param.h>
  89 #include <sys/systm.h>
  90 #include <sys/socket.h>
  91 #include <sys/sockio.h>
  92 #include <sys/mbuf.h>
  93 #include <sys/errno.h>
  94 #include <sys/protosw.h>
  95 #include <sys/queue.h>
  96
  97 #include <net/if.h>
  98 #include <net/route.h>
  99
 100 #include <netinet/in.h>
 101 #include <netinet/in_systm.h>
 102 #include <netinet/ip.h>
 103 #include <netinet/ip_var.h>
 104 #include <netinet/ip_encap.h>
 105 #if MROUTING
 106 #include <netinet/ip_mroute.h>
 107 #endif /* MROUTING */
 108
 109 #if INET6
 110 #include <netinet/ip6.h>
 111 #include <netinet6/ip6_var.h>
 112 #include <netinet6/ip6protosw.h>
 113 #endif
 114
 115
 116 #include <net/net_osdep.h>
 117
 118 #ifndef __APPLE__
 119 #include <sys/kernel.h>
 120 #include <sys/malloc.h>
 121 MALLOC_DEFINE(M_NETADDR, "Export Host", "Export host address structure");
 122 #endif
 123
 124 static void encap_add(struct encaptab *);
 125 static int mask_match(const struct encaptab *, const struct sockaddr *,
 126                 const struct sockaddr *);
 127 static void encap_fillarg(struct mbuf *, const struct encaptab *);
 128
 129 #ifndef LIST_HEAD_INITIALIZER
 130 /* rely upon BSS initialization */
 131 LIST_HEAD(, encaptab) encaptab;
 132 #else
 133 LIST_HEAD(, encaptab) encaptab = LIST_HEAD_INITIALIZER(&encaptab);
 134 #endif
 135
 136 void
 137 encap_init()
 138 {
 139         static int initialized = 0;
 140
 141         if (initialized)
 142                 return;
 143         initialized++;
 144 #if 0
 145         /*
 146          * we cannot use LIST_INIT() here, since drivers may want to call
 147          * encap_attach(), on driver attach.  encap_init() will be called
 148          * on AF_INET{,6} initialization, which happens after driver
 149          * initialization - using LIST_INIT() here can nuke encap_attach()
 150          * from drivers.
 151          */
 152         LIST_INIT(&encaptab);
 153 #endif
 154 }
 155
 156 #if INET
 157 void
 158 encap4_input(m, off)
 159         struct mbuf *m;
 160         int off;
 161 {
 162         int proto;
 163         struct ip *ip;
 164         struct sockaddr_in s, d;
 165         const struct protosw *psw;
 166         struct encaptab *ep, *match;
 167         int prio, matchprio;
 168
 169 #ifndef __APPLE__
 170         va_start(ap, m);
 171         off = va_arg(ap, int);
 172         proto = va_arg(ap, int);
 173         va_end(ap);
 174 #endif
 175
 176         ip = mtod(m, struct ip *);
 177 #ifdef __APPLE__
 178         proto = ip->ip_p;
 179 #endif
 180
 181         bzero(&s, sizeof(s));
 182         s.sin_family = AF_INET;
 183         s.sin_len = sizeof(struct sockaddr_in);
 184         s.sin_addr = ip->ip_src;
 185         bzero(&d, sizeof(d));
 186         d.sin_family = AF_INET;
 187         d.sin_len = sizeof(struct sockaddr_in);
 188         d.sin_addr = ip->ip_dst;
 189
 190         match = NULL;
 191         matchprio = 0;
 192         for (ep = LIST_FIRST(&encaptab); ep; ep = LIST_NEXT(ep, chain)) {
 193                 if (ep->af != AF_INET)
 194                         continue;
 195                 if (ep->proto >= 0 && ep->proto != proto)
 196                         continue;
 197                 if (ep->func)
 198                         prio = (*ep->func)(m, off, proto, ep->arg);
 199                 else {
 200                         /*
 201                          * it's inbound traffic, we need to match in reverse
 202                          * order
 203                          */
 204                         prio = mask_match(ep, (struct sockaddr *)&d,
 205                             (struct sockaddr *)&s);
 206                 }
 207
 208                 /*
 209                  * We prioritize the matches by using bit length of the
 210                  * matches.  mask_match() and user-supplied matching function
 211                  * should return the bit length of the matches (for example,
 212                  * if both src/dst are matched for IPv4, 64 should be returned).
 213                  * 0 or negative return value means "it did not match".
 214                  *
 215                  * The question is, since we have two "mask" portion, we
 216                  * cannot really define total order between entries.
 217                  * For example, which of these should be preferred?
 218                  * mask_match() returns 48 (32 + 16) for both of them.
 219                  *      src=3ffe::/16, dst=3ffe:501::/32
 220                  *      src=3ffe:501::/32, dst=3ffe::/16
 221                  *
 222                  * We need to loop through all the possible candidates
 223                  * to get the best match - the search takes O(n) for
 224                  * n attachments (i.e. interfaces).
 225                  */
 226                 if (prio <= 0)
 227                         continue;
 228                 if (prio > matchprio) {
 229                         matchprio = prio;
 230                         match = ep;
 231                 }
 232         }
 233
 234         if (match) {
 235                 /* found a match, "match" has the best one */
 236                 psw = (const struct protosw *)match->psw;
 237                 if (psw && psw->pr_input) {
 238                         encap_fillarg(m, match);
 239                         (*psw->pr_input)(m, off);
 240                 } else
 241                         m_freem(m);
 242                 return;
 243         }
 244
 245         /* for backward compatibility */
 246 # if MROUTING
 247 #  define COMPATFUNC    ipip_input
 248 # endif /*MROUTING*/
 249
 250 #if COMPATFUNC
 251         if (proto == IPPROTO_IPV4) {
 252                 COMPATFUNC(m, off);
 253                 return;
 254         }
 255 #endif
 256
 257         /* last resort: inject to raw socket */
 258         rip_input(m, off);
 259 }
 260 #endif
 261
 262 #if INET6
 263 int
 264 encap6_input(mp, offp)
 265         struct mbuf **mp;
 266         int *offp;
 267 {
 268         struct mbuf *m = *mp;
 269         struct ip6_hdr *ip6;
 270         struct sockaddr_in6 s, d;
 271         const struct ip6protosw *psw;
 272         struct encaptab *ep, *match;
 273         int prio, matchprio;
 274         int proto;
 275
 276         ip6 = mtod(m, struct ip6_hdr *);
 277         proto = ip6->ip6_nxt;
 278
 279         bzero(&s, sizeof(s));
 280         s.sin6_family = AF_INET6;
 281         s.sin6_len = sizeof(struct sockaddr_in6);
 282         s.sin6_addr = ip6->ip6_src;
 283         bzero(&d, sizeof(d));
 284         d.sin6_family = AF_INET6;
 285         d.sin6_len = sizeof(struct sockaddr_in6);
 286         d.sin6_addr = ip6->ip6_dst;
 287
 288         match = NULL;
 289         matchprio = 0;
 290         for (ep = LIST_FIRST(&encaptab); ep; ep = LIST_NEXT(ep, chain)) {
 291                 if (ep->af != AF_INET6)
 292                         continue;
 293                 if (ep->proto >= 0 && ep->proto != proto)
 294                         continue;
 295                 if (ep->func)
 296                         prio = (*ep->func)(m, *offp, proto, ep->arg);
 297                 else {
 298                         /*
 299                          * it's inbound traffic, we need to match in reverse
 300                          * order
 301                          */
 302                         prio = mask_match(ep, (struct sockaddr *)&d,
 303                             (struct sockaddr *)&s);
 304                 }
 305
 306                 /* see encap4_input() for issues here */
 307                 if (prio <= 0)
 308                         continue;
 309                 if (prio > matchprio) {
 310                         matchprio = prio;
 311                         match = ep;
 312                 }
 313         }
 314
 315         if (match) {
 316                 /* found a match */
 317                 psw = (const struct ip6protosw *)match->psw;
 318                 if (psw && psw->pr_input) {
 319                         encap_fillarg(m, match);
 320                         return (*psw->pr_input)(mp, offp);
 321                 } else {
 322                         m_freem(m);
 323                         return IPPROTO_DONE;
 324                 }
 325         }
 326
 327         /* last resort: inject to raw socket */
 328         return rip6_input(mp, offp);
 329 }
 330 #endif
 331
 332 static void
 333 encap_add(ep)
 334         struct encaptab *ep;
 335 {
 336
 337         LIST_INSERT_HEAD(&encaptab, ep, chain);
 338 }
 339
 340 /*
 341  * sp (src ptr) is always my side, and dp (dst ptr) is always remote side.
 342  * length of mask (sm and dm) is assumed to be same as sp/dp.
 343  * Return value will be necessary as input (cookie) for encap_detach().
 344  */
 345 const struct encaptab *
 346 encap_attach(af, proto, sp, sm, dp, dm, psw, arg)
 347         int af;
 348         int proto;
 349         const struct sockaddr *sp, *sm;
 350         const struct sockaddr *dp, *dm;
 351         const struct protosw *psw;
 352         void *arg;
 353 {
 354         struct encaptab *ep;
 355         int error;
 356         int s;
 357
 358         s = splnet();
 359         /* sanity check on args */
 360         if (sp->sa_len > sizeof(ep->src) || dp->sa_len > sizeof(ep->dst)) {
 361                 error = EINVAL;
 362                 goto fail;
 363         }
 364         if (sp->sa_len != dp->sa_len) {
 365                 error = EINVAL;
 366                 goto fail;
 367         }
 368         if (af != sp->sa_family || af != dp->sa_family) {
 369                 error = EINVAL;
 370                 goto fail;
 371         }
 372
 373         /* check if anyone have already attached with exactly same config */
 374         for (ep = LIST_FIRST(&encaptab); ep; ep = LIST_NEXT(ep, chain)) {
 375                 if (ep->af != af)
 376                         continue;
 377                 if (ep->proto != proto)
 378                         continue;
 379                 if (ep->src.ss_len != sp->sa_len ||
 380                     bcmp(&ep->src, sp, sp->sa_len) != 0 ||
 381                     bcmp(&ep->srcmask, sm, sp->sa_len) != 0)
 382                         continue;
 383                 if (ep->dst.ss_len != dp->sa_len ||
 384                     bcmp(&ep->dst, dp, dp->sa_len) != 0 ||
 385                     bcmp(&ep->dstmask, dm, dp->sa_len) != 0)
 386                         continue;
 387
 388                 error = EEXIST;
 389                 goto fail;
 390         }
 391
 392         ep = _MALLOC(sizeof(*ep), M_NETADDR, M_WAITOK); /*XXX*/
 393         if (ep == NULL) {
 394                 error = ENOBUFS;
 395                 goto fail;
 396         }
 397         bzero(ep, sizeof(*ep));
 398
 399         ep->af = af;
 400         ep->proto = proto;
 401         bcopy(sp, &ep->src, sp->sa_len);
 402         bcopy(sm, &ep->srcmask, sp->sa_len);
 403         bcopy(dp, &ep->dst, dp->sa_len);
 404         bcopy(dm, &ep->dstmask, dp->sa_len);
 405         ep->psw = psw;
 406         ep->arg = arg;
 407
 408         encap_add(ep);
 409
 410         error = 0;
 411         splx(s);
 412         return ep;
 413
 414 fail:
 415         splx(s);
 416         return NULL;
 417 }
 418
 419 const struct encaptab *
 420 encap_attach_func(af, proto, func, psw, arg)
 421         int af;
 422         int proto;
 423         int (*func)(const struct mbuf *, int, int, void *);
 424         const struct protosw *psw;
 425         void *arg;
 426 {
 427         struct encaptab *ep;
 428         int error;
 429         int s;
 430
 431         s = splnet();
 432         /* sanity check on args */
 433         if (!func) {
 434                 error = EINVAL;
 435                 goto fail;
 436         }
 437
 438         ep = _MALLOC(sizeof(*ep), M_NETADDR, M_WAITOK); /*XXX*/
 439         if (ep == NULL) {
 440                 error = ENOBUFS;
 441                 goto fail;
 442         }
 443         bzero(ep, sizeof(*ep));
 444
 445         ep->af = af;
 446         ep->proto = proto;
 447         ep->func = func;
 448         ep->psw = psw;
 449         ep->arg = arg;
 450
 451         encap_add(ep);
 452
 453         error = 0;
 454         splx(s);
 455         return ep;
 456
 457 fail:
 458         splx(s);
 459         return NULL;
 460 }
 461
 462 int
 463 encap_detach(cookie)
 464         const struct encaptab *cookie;
 465 {
 466         const struct encaptab *ep = cookie;
 467         struct encaptab *p;
 468
 469         for (p = LIST_FIRST(&encaptab); p; p = LIST_NEXT(p, chain)) {
 470                 if (p == ep) {
 471                         LIST_REMOVE(p, chain);
 472                         _FREE(p, M_NETADDR);    /*XXX*/
 473                         return 0;
 474                 }
 475         }
 476
 477         return EINVAL;
 478 }
 479
 480 static int
 481 mask_match(ep, sp, dp)
 482         const struct encaptab *ep;
 483         const struct sockaddr *sp;
 484         const struct sockaddr *dp;
 485 {
 486         struct sockaddr_storage s;
 487         struct sockaddr_storage d;
 488         int i;
 489         const u_int8_t *p, *q;
 490         u_int8_t *r;
 491         int matchlen;
 492
 493         if (sp->sa_len > sizeof(s) || dp->sa_len > sizeof(d))
 494                 return 0;
 495         if (sp->sa_family != ep->af || dp->sa_family != ep->af)
 496                 return 0;
 497         if (sp->sa_len != ep->src.ss_len || dp->sa_len != ep->dst.ss_len)
 498                 return 0;
 499
 500         matchlen = 0;
 501
 502         p = (const u_int8_t *)sp;
 503         q = (const u_int8_t *)&ep->srcmask;
 504         r = (u_int8_t *)&s;
 505         for (i = 0 ; i < sp->sa_len; i++) {
 506                 r[i] = p[i] & q[i];
 507                 /* XXX estimate */
 508                 matchlen += (q[i] ? 8 : 0);
 509         }
 510
 511         p = (const u_int8_t *)dp;
 512         q = (const u_int8_t *)&ep->dstmask;
 513         r = (u_int8_t *)&d;
 514         for (i = 0 ; i < dp->sa_len; i++) {
 515                 r[i] = p[i] & q[i];
 516                 /* XXX rough estimate */
 517                 matchlen += (q[i] ? 8 : 0);
 518         }
 519
 520         /* need to overwrite len/family portion as we don't compare them */
 521         s.ss_len = sp->sa_len;
 522         s.ss_family = sp->sa_family;
 523         d.ss_len = dp->sa_len;
 524         d.ss_family = dp->sa_family;
 525
 526         if (bcmp(&s, &ep->src, ep->src.ss_len) == 0 &&
 527             bcmp(&d, &ep->dst, ep->dst.ss_len) == 0) {
 528                 return matchlen;
 529         } else
 530                 return 0;
 531 }
 532
 533 static void
 534 encap_fillarg(m, ep)
 535         struct mbuf *m;
 536         const struct encaptab *ep;
 537 {
 538 #if 0
 539         m->m_pkthdr.aux = ep->arg;
 540 #else
 541         struct mbuf *n;
 542
 543         n = m_aux_add(m, AF_INET, IPPROTO_IPV4);
 544         if (n) {
 545                 *mtod(n, void **) = ep->arg;
 546                 n->m_len = sizeof(void *);
 547         }
 548 #endif
 549 }
 550
 551 void *
 552 encap_getarg(m)
 553         struct mbuf *m;
 554 {
 555         void *p;
 556 #if 0
 557         p = m->m_pkthdr.aux;
 558         m->m_pkthdr.aux = NULL;
 559         return p;
 560 #else
 561         struct mbuf *n;
 562
 563         p = NULL;
 564         n = m_aux_find(m, AF_INET, IPPROTO_IPV4);
 565         if (n) {
 566                 if (n->m_len == sizeof(void *))
 567                         p = *mtod(n, void **);
 568                 m_aux_delete(m, n);
 569         }
 570         return p;
 571 #endif
 572 }