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 30 /*- 31 * Copyright (c) 1982, 1986, 1988, 1993 32 * The Regents of the University of California. 33 * All rights reserved. 34 * 35 * Redistribution and use in source and binary forms, with or without 36 * modification, are permitted provided that the following conditions 37 * are met: 38 * 1. Redistributions of source code must retain the above copyright 39 * notice, this list of conditions and the following disclaimer. 40 * 2. Redistributions in binary form must reproduce the above copyright 41 * notice, this list of conditions and the following disclaimer in the 42 * documentation and/or other materials provided with the distribution. 43 * 4. Neither the name of the University nor the names of its contributors 44 * may be used to endorse or promote products derived from this software 45 * without specific prior written permission. 46 * 47 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 48 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 49 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 50 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 51 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 52 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 53 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 54 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 55 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 56 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 57 * SUCH DAMAGE. 58 * 59 * @(#)raw_ip.c 8.2 (Berkeley) 1/4/94 60 */ 61 62 #include <sys/cdefs.h> 63 __FBSDID("$FreeBSD$"); 64 65 #include "opt_ipsec.h" 66 #include "opt_inet6.h" 67 68 #include <sys/param.h> 69 #include <sys/errno.h> 70 #include <sys/jail.h> 71 #include <sys/kernel.h> 72 #include <sys/lock.h> 73 #include <sys/malloc.h> 74 #include <sys/mbuf.h> 75 #include <sys/priv.h> 76 #include <sys/proc.h> 77 #include <sys/protosw.h> 78 #include <sys/signalvar.h> 79 #include <sys/socket.h> 80 #include <sys/socketvar.h> 81 #include <sys/sx.h> 82 #include <sys/syslog.h> 83 84 #include <net/if.h> 85 #include <net/if_var.h> 86 #include <net/if_types.h> 87 #include <net/route.h> 88 #include <net/vnet.h> 89 90 #include <netinet/in.h> 91 #include <netinet/in_var.h> 92 #include <netinet/in_systm.h> 93 #include <netinet/in_pcb.h> 94 95 #include <netinet/icmp6.h> 96 #include <netinet/ip6.h> 97 #include <netinet/ip_var.h> 98 #include <netinet6/ip6protosw.h> 99 #include <netinet6/ip6_mroute.h> 100 #include <netinet6/in6_pcb.h> 101 #include <netinet6/ip6_var.h> 102 #include <netinet6/nd6.h> 103 #include <netinet6/raw_ip6.h> 104 #include <netinet6/scope6_var.h> 105 #include <netinet6/send.h> 106 107 #ifdef IPSEC 108 #include <netipsec/ipsec.h> 109 #include <netipsec/ipsec6.h> 110 #endif /* IPSEC */ 111 112 #include <machine/stdarg.h> 113 114 #define satosin6(sa) ((struct sockaddr_in6 *)(sa)) 115 #define ifatoia6(ifa) ((struct in6_ifaddr *)(ifa)) 116 117 /* 118 * Raw interface to IP6 protocol. 119 */ 120 121 VNET_DECLARE(struct inpcbhead, ripcb); 122 VNET_DECLARE(struct inpcbinfo, ripcbinfo); 123 #define V_ripcb VNET(ripcb) 124 #define V_ripcbinfo VNET(ripcbinfo) 125 126 extern u_long rip_sendspace; 127 extern u_long rip_recvspace; 128 129 VNET_PCPUSTAT_DEFINE(struct rip6stat, rip6stat); 130 VNET_PCPUSTAT_SYSINIT(rip6stat); 131 132 #ifdef VIMAGE 133 VNET_PCPUSTAT_SYSUNINIT(rip6stat); 134 #endif /* VIMAGE */ 135 136 /* 137 * Hooks for multicast routing. They all default to NULL, so leave them not 138 * initialized and rely on BSS being set to 0. 139 */ 140 141 /* 142 * The socket used to communicate with the multicast routing daemon. 143 */ 144 VNET_DEFINE(struct socket *, ip6_mrouter); 145 146 /* 147 * The various mrouter functions. 148 */ 149 int (*ip6_mrouter_set)(struct socket *, struct sockopt *); 150 int (*ip6_mrouter_get)(struct socket *, struct sockopt *); 151 int (*ip6_mrouter_done)(void); 152 int (*ip6_mforward)(struct ip6_hdr *, struct ifnet *, struct mbuf *); 153 int (*mrt6_ioctl)(u_long, caddr_t); 154 155 /* 156 * Setup generic address and protocol structures for raw_input routine, then 157 * pass them along with mbuf chain. 158 */ 159 int 160 rip6_input(struct mbuf **mp, int *offp, int proto) 161 { 162 struct ifnet *ifp; 163 struct mbuf *m = *mp; 164 register struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *); 165 register struct inpcb *in6p; 166 struct inpcb *last = 0; 167 struct mbuf *opts = NULL; 168 struct sockaddr_in6 fromsa; 169 170 RIP6STAT_INC(rip6s_ipackets); 171 172 if (faithprefix_p != NULL && (*faithprefix_p)(&ip6->ip6_dst)) { 173 /* XXX Send icmp6 host/port unreach? */ 174 m_freem(m); 175 return (IPPROTO_DONE); 176 } 177 178 init_sin6(&fromsa, m); /* general init */ 179 180 ifp = m->m_pkthdr.rcvif; 181 182 INP_INFO_RLOCK(&V_ripcbinfo); 183 LIST_FOREACH(in6p, &V_ripcb, inp_list) { 184 /* XXX inp locking */ 185 if ((in6p->inp_vflag & INP_IPV6) == 0) 186 continue; 187 if (in6p->inp_ip_p && 188 in6p->inp_ip_p != proto) 189 continue; 190 if (!IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_laddr) && 191 !IN6_ARE_ADDR_EQUAL(&in6p->in6p_laddr, &ip6->ip6_dst)) 192 continue; 193 if (!IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_faddr) && 194 !IN6_ARE_ADDR_EQUAL(&in6p->in6p_faddr, &ip6->ip6_src)) 195 continue; 196 if (jailed_without_vnet(in6p->inp_cred)) { 197 /* 198 * Allow raw socket in jail to receive multicast; 199 * assume process had PRIV_NETINET_RAW at attach, 200 * and fall through into normal filter path if so. 201 */ 202 if (!IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) && 203 prison_check_ip6(in6p->inp_cred, 204 &ip6->ip6_dst) != 0) 205 continue; 206 } 207 INP_RLOCK(in6p); 208 if (in6p->in6p_cksum != -1) { 209 RIP6STAT_INC(rip6s_isum); 210 if (in6_cksum(m, proto, *offp, 211 m->m_pkthdr.len - *offp)) { 212 INP_RUNLOCK(in6p); 213 RIP6STAT_INC(rip6s_badsum); 214 continue; 215 } 216 } 217 /* 218 * If this raw socket has multicast state, and we 219 * have received a multicast, check if this socket 220 * should receive it, as multicast filtering is now 221 * the responsibility of the transport layer. 222 */ 223 if (in6p->in6p_moptions && 224 IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) { 225 /* 226 * If the incoming datagram is for MLD, allow it 227 * through unconditionally to the raw socket. 228 * 229 * Use the M_RTALERT_MLD flag to check for MLD 230 * traffic without having to inspect the mbuf chain 231 * more deeply, as all MLDv1/v2 host messages MUST 232 * contain the Router Alert option. 233 * 234 * In the case of MLDv1, we may not have explicitly 235 * joined the group, and may have set IFF_ALLMULTI 236 * on the interface. im6o_mc_filter() may discard 237 * control traffic we actually need to see. 238 * 239 * Userland multicast routing daemons should continue 240 * filter the control traffic appropriately. 241 */ 242 int blocked; 243 244 blocked = MCAST_PASS; 245 if ((m->m_flags & M_RTALERT_MLD) == 0) { 246 struct sockaddr_in6 mcaddr; 247 248 bzero(&mcaddr, sizeof(struct sockaddr_in6)); 249 mcaddr.sin6_len = sizeof(struct sockaddr_in6); 250 mcaddr.sin6_family = AF_INET6; 251 mcaddr.sin6_addr = ip6->ip6_dst; 252 253 blocked = im6o_mc_filter(in6p->in6p_moptions, 254 ifp, 255 (struct sockaddr *)&mcaddr, 256 (struct sockaddr *)&fromsa); 257 } 258 if (blocked != MCAST_PASS) { 259 IP6STAT_INC(ip6s_notmember); 260 INP_RUNLOCK(in6p); 261 continue; 262 } 263 } 264 if (last != NULL) { 265 struct mbuf *n = m_copy(m, 0, (int)M_COPYALL); 266 267 #ifdef IPSEC 268 /* 269 * Check AH/ESP integrity. 270 */ 271 if (n && ipsec6_in_reject(n, last)) { 272 m_freem(n); 273 IPSEC6STAT_INC(ips_in_polvio); 274 /* Do not inject data into pcb. */ 275 } else 276 #endif /* IPSEC */ 277 if (n) { 278 if (last->inp_flags & INP_CONTROLOPTS || 279 last->inp_socket->so_options & SO_TIMESTAMP) 280 ip6_savecontrol(last, n, &opts); 281 /* strip intermediate headers */ 282 m_adj(n, *offp); 283 if (sbappendaddr(&last->inp_socket->so_rcv, 284 (struct sockaddr *)&fromsa, 285 n, opts) == 0) { 286 m_freem(n); 287 if (opts) 288 m_freem(opts); 289 RIP6STAT_INC(rip6s_fullsock); 290 } else 291 sorwakeup(last->inp_socket); 292 opts = NULL; 293 } 294 INP_RUNLOCK(last); 295 } 296 last = in6p; 297 } 298 INP_INFO_RUNLOCK(&V_ripcbinfo); 299 #ifdef IPSEC 300 /* 301 * Check AH/ESP integrity. 302 */ 303 if ((last != NULL) && ipsec6_in_reject(m, last)) { 304 m_freem(m); 305 IPSEC6STAT_INC(ips_in_polvio); 306 IP6STAT_DEC(ip6s_delivered); 307 /* Do not inject data into pcb. */ 308 INP_RUNLOCK(last); 309 } else 310 #endif /* IPSEC */ 311 if (last != NULL) { 312 if (last->inp_flags & INP_CONTROLOPTS || 313 last->inp_socket->so_options & SO_TIMESTAMP) 314 ip6_savecontrol(last, m, &opts); 315 /* Strip intermediate headers. */ 316 m_adj(m, *offp); 317 if (sbappendaddr(&last->inp_socket->so_rcv, 318 (struct sockaddr *)&fromsa, m, opts) == 0) { 319 m_freem(m); 320 if (opts) 321 m_freem(opts); 322 RIP6STAT_INC(rip6s_fullsock); 323 } else 324 sorwakeup(last->inp_socket); 325 INP_RUNLOCK(last); 326 } else { 327 RIP6STAT_INC(rip6s_nosock); 328 if (m->m_flags & M_MCAST) 329 RIP6STAT_INC(rip6s_nosockmcast); 330 if (proto == IPPROTO_NONE) 331 m_freem(m); 332 else { 333 char *prvnxtp = ip6_get_prevhdr(m, *offp); /* XXX */ 334 icmp6_error(m, ICMP6_PARAM_PROB, 335 ICMP6_PARAMPROB_NEXTHEADER, 336 prvnxtp - mtod(m, char *)); 337 } 338 IP6STAT_DEC(ip6s_delivered); 339 } 340 return (IPPROTO_DONE); 341 } 342 343 void 344 rip6_ctlinput(int cmd, struct sockaddr *sa, void *d) 345 { 346 struct ip6_hdr *ip6; 347 struct mbuf *m; 348 int off = 0; 349 struct ip6ctlparam *ip6cp = NULL; 350 const struct sockaddr_in6 *sa6_src = NULL; 351 void *cmdarg; 352 struct inpcb *(*notify)(struct inpcb *, int) = in6_rtchange; 353 354 if (sa->sa_family != AF_INET6 || 355 sa->sa_len != sizeof(struct sockaddr_in6)) 356 return; 357 358 if ((unsigned)cmd >= PRC_NCMDS) 359 return; 360 if (PRC_IS_REDIRECT(cmd)) 361 notify = in6_rtchange, d = NULL; 362 else if (cmd == PRC_HOSTDEAD) 363 d = NULL; 364 else if (inet6ctlerrmap[cmd] == 0) 365 return; 366 367 /* 368 * If the parameter is from icmp6, decode it. 369 */ 370 if (d != NULL) { 371 ip6cp = (struct ip6ctlparam *)d; 372 m = ip6cp->ip6c_m; 373 ip6 = ip6cp->ip6c_ip6; 374 off = ip6cp->ip6c_off; 375 cmdarg = ip6cp->ip6c_cmdarg; 376 sa6_src = ip6cp->ip6c_src; 377 } else { 378 m = NULL; 379 ip6 = NULL; 380 cmdarg = NULL; 381 sa6_src = &sa6_any; 382 } 383 384 (void) in6_pcbnotify(&V_ripcbinfo, sa, 0, 385 (const struct sockaddr *)sa6_src, 0, cmd, cmdarg, notify); 386 } 387 388 /* 389 * Generate IPv6 header and pass packet to ip6_output. Tack on options user 390 * may have setup with control call. 391 */ 392 int 393 #if __STDC__ 394 rip6_output(struct mbuf *m, ...) 395 #else 396 rip6_output(m, va_alist) 397 struct mbuf *m; 398 va_dcl 399 #endif 400 { 401 struct mbuf *control; 402 struct m_tag *mtag; 403 struct socket *so; 404 struct sockaddr_in6 *dstsock; 405 struct in6_addr *dst; 406 struct ip6_hdr *ip6; 407 struct inpcb *in6p; 408 u_int plen = m->m_pkthdr.len; 409 int error = 0; 410 struct ip6_pktopts opt, *optp; 411 struct ifnet *oifp = NULL; 412 int type = 0, code = 0; /* for ICMPv6 output statistics only */ 413 int scope_ambiguous = 0; 414 int use_defzone = 0; 415 struct in6_addr in6a; 416 va_list ap; 417 418 va_start(ap, m); 419 so = va_arg(ap, struct socket *); 420 dstsock = va_arg(ap, struct sockaddr_in6 *); 421 control = va_arg(ap, struct mbuf *); 422 va_end(ap); 423 424 in6p = sotoinpcb(so); 425 INP_WLOCK(in6p); 426 427 dst = &dstsock->sin6_addr; 428 if (control != NULL) { 429 if ((error = ip6_setpktopts(control, &opt, 430 in6p->in6p_outputopts, so->so_cred, 431 so->so_proto->pr_protocol)) != 0) { 432 goto bad; 433 } 434 optp = &opt; 435 } else 436 optp = in6p->in6p_outputopts; 437 438 /* 439 * Check and convert scope zone ID into internal form. 440 * 441 * XXX: we may still need to determine the zone later. 442 */ 443 if (!(so->so_state & SS_ISCONNECTED)) { 444 if (!optp || !optp->ip6po_pktinfo || 445 !optp->ip6po_pktinfo->ipi6_ifindex) 446 use_defzone = V_ip6_use_defzone; 447 if (dstsock->sin6_scope_id == 0 && !use_defzone) 448 scope_ambiguous = 1; 449 if ((error = sa6_embedscope(dstsock, use_defzone)) != 0) 450 goto bad; 451 } 452 453 /* 454 * For an ICMPv6 packet, we should know its type and code to update 455 * statistics. 456 */ 457 if (so->so_proto->pr_protocol == IPPROTO_ICMPV6) { 458 struct icmp6_hdr *icmp6; 459 if (m->m_len < sizeof(struct icmp6_hdr) && 460 (m = m_pullup(m, sizeof(struct icmp6_hdr))) == NULL) { 461 error = ENOBUFS; 462 goto bad; 463 } 464 icmp6 = mtod(m, struct icmp6_hdr *); 465 type = icmp6->icmp6_type; 466 code = icmp6->icmp6_code; 467 } 468 469 M_PREPEND(m, sizeof(*ip6), M_NOWAIT); 470 if (m == NULL) { 471 error = ENOBUFS; 472 goto bad; 473 } 474 ip6 = mtod(m, struct ip6_hdr *); 475 476 /* 477 * Source address selection. 478 */ 479 error = in6_selectsrc(dstsock, optp, in6p, NULL, so->so_cred, 480 &oifp, &in6a); 481 if (error) 482 goto bad; 483 error = prison_check_ip6(in6p->inp_cred, &in6a); 484 if (error != 0) 485 goto bad; 486 ip6->ip6_src = in6a; 487 488 if (oifp && scope_ambiguous) { 489 /* 490 * Application should provide a proper zone ID or the use of 491 * default zone IDs should be enabled. Unfortunately, some 492 * applications do not behave as it should, so we need a 493 * workaround. Even if an appropriate ID is not determined 494 * (when it's required), if we can determine the outgoing 495 * interface. determine the zone ID based on the interface. 496 */ 497 error = in6_setscope(&dstsock->sin6_addr, oifp, NULL); 498 if (error != 0) 499 goto bad; 500 } 501 ip6->ip6_dst = dstsock->sin6_addr; 502 503 /* 504 * Fill in the rest of the IPv6 header fields. 505 */ 506 ip6->ip6_flow = (ip6->ip6_flow & ~IPV6_FLOWINFO_MASK) | 507 (in6p->inp_flow & IPV6_FLOWINFO_MASK); 508 ip6->ip6_vfc = (ip6->ip6_vfc & ~IPV6_VERSION_MASK) | 509 (IPV6_VERSION & IPV6_VERSION_MASK); 510 511 /* 512 * ip6_plen will be filled in ip6_output, so not fill it here. 513 */ 514 ip6->ip6_nxt = in6p->inp_ip_p; 515 ip6->ip6_hlim = in6_selecthlim(in6p, oifp); 516 517 if (so->so_proto->pr_protocol == IPPROTO_ICMPV6 || 518 in6p->in6p_cksum != -1) { 519 struct mbuf *n; 520 int off; 521 u_int16_t *p; 522 523 /* Compute checksum. */ 524 if (so->so_proto->pr_protocol == IPPROTO_ICMPV6) 525 off = offsetof(struct icmp6_hdr, icmp6_cksum); 526 else 527 off = in6p->in6p_cksum; 528 if (plen < off + 1) { 529 error = EINVAL; 530 goto bad; 531 } 532 off += sizeof(struct ip6_hdr); 533 534 n = m; 535 while (n && n->m_len <= off) { 536 off -= n->m_len; 537 n = n->m_next; 538 } 539 if (!n) 540 goto bad; 541 p = (u_int16_t *)(mtod(n, caddr_t) + off); 542 *p = 0; 543 *p = in6_cksum(m, ip6->ip6_nxt, sizeof(*ip6), plen); 544 } 545 546 /* 547 * Send RA/RS messages to user land for protection, before sending 548 * them to rtadvd/rtsol. 549 */ 550 if ((send_sendso_input_hook != NULL) && 551 so->so_proto->pr_protocol == IPPROTO_ICMPV6) { 552 switch (type) { 553 case ND_ROUTER_ADVERT: 554 case ND_ROUTER_SOLICIT: 555 mtag = m_tag_get(PACKET_TAG_ND_OUTGOING, 556 sizeof(unsigned short), M_NOWAIT); 557 if (mtag == NULL) 558 goto bad; 559 m_tag_prepend(m, mtag); 560 } 561 } 562 563 error = ip6_output(m, optp, NULL, 0, in6p->in6p_moptions, &oifp, in6p); 564 if (so->so_proto->pr_protocol == IPPROTO_ICMPV6) { 565 if (oifp) 566 icmp6_ifoutstat_inc(oifp, type, code); 567 ICMP6STAT_INC(icp6s_outhist[type]); 568 } else 569 RIP6STAT_INC(rip6s_opackets); 570 571 goto freectl; 572 573 bad: 574 if (m) 575 m_freem(m); 576 577 freectl: 578 if (control != NULL) { 579 ip6_clearpktopts(&opt, -1); 580 m_freem(control); 581 } 582 INP_WUNLOCK(in6p); 583 return (error); 584 } 585 586 /* 587 * Raw IPv6 socket option processing. 588 */ 589 int 590 rip6_ctloutput(struct socket *so, struct sockopt *sopt) 591 { 592 struct inpcb *inp; 593 int error; 594 595 if (sopt->sopt_level == IPPROTO_ICMPV6) 596 /* 597 * XXX: is it better to call icmp6_ctloutput() directly 598 * from protosw? 599 */ 600 return (icmp6_ctloutput(so, sopt)); 601 else if (sopt->sopt_level != IPPROTO_IPV6) { 602 if (sopt->sopt_level == SOL_SOCKET && 603 sopt->sopt_name == SO_SETFIB) { 604 inp = sotoinpcb(so); 605 INP_WLOCK(inp); 606 inp->inp_inc.inc_fibnum = so->so_fibnum; 607 INP_WUNLOCK(inp); 608 return (0); 609 } 610 return (EINVAL); 611 } 612 613 error = 0; 614 615 switch (sopt->sopt_dir) { 616 case SOPT_GET: 617 switch (sopt->sopt_name) { 618 case MRT6_INIT: 619 case MRT6_DONE: 620 case MRT6_ADD_MIF: 621 case MRT6_DEL_MIF: 622 case MRT6_ADD_MFC: 623 case MRT6_DEL_MFC: 624 case MRT6_PIM: 625 error = ip6_mrouter_get ? ip6_mrouter_get(so, sopt) : 626 EOPNOTSUPP; 627 break; 628 case IPV6_CHECKSUM: 629 error = ip6_raw_ctloutput(so, sopt); 630 break; 631 default: 632 error = ip6_ctloutput(so, sopt); 633 break; 634 } 635 break; 636 637 case SOPT_SET: 638 switch (sopt->sopt_name) { 639 case MRT6_INIT: 640 case MRT6_DONE: 641 case MRT6_ADD_MIF: 642 case MRT6_DEL_MIF: 643 case MRT6_ADD_MFC: 644 case MRT6_DEL_MFC: 645 case MRT6_PIM: 646 error = ip6_mrouter_set ? ip6_mrouter_set(so, sopt) : 647 EOPNOTSUPP; 648 break; 649 case IPV6_CHECKSUM: 650 error = ip6_raw_ctloutput(so, sopt); 651 break; 652 default: 653 error = ip6_ctloutput(so, sopt); 654 break; 655 } 656 break; 657 } 658 659 return (error); 660 } 661 662 static int 663 rip6_attach(struct socket *so, int proto, struct thread *td) 664 { 665 struct inpcb *inp; 666 struct icmp6_filter *filter; 667 int error; 668 669 inp = sotoinpcb(so); 670 KASSERT(inp == NULL, ("rip6_attach: inp != NULL")); 671 672 error = priv_check(td, PRIV_NETINET_RAW); 673 if (error) 674 return (error); 675 error = soreserve(so, rip_sendspace, rip_recvspace); 676 if (error) 677 return (error); 678 filter = malloc(sizeof(struct icmp6_filter), M_PCB, M_NOWAIT); 679 if (filter == NULL) 680 return (ENOMEM); 681 INP_INFO_WLOCK(&V_ripcbinfo); 682 error = in_pcballoc(so, &V_ripcbinfo); 683 if (error) { 684 INP_INFO_WUNLOCK(&V_ripcbinfo); 685 free(filter, M_PCB); 686 return (error); 687 } 688 inp = (struct inpcb *)so->so_pcb; 689 INP_INFO_WUNLOCK(&V_ripcbinfo); 690 inp->inp_vflag |= INP_IPV6; 691 inp->inp_ip_p = (long)proto; 692 inp->in6p_hops = -1; /* use kernel default */ 693 inp->in6p_cksum = -1; 694 inp->in6p_icmp6filt = filter; 695 ICMP6_FILTER_SETPASSALL(inp->in6p_icmp6filt); 696 INP_WUNLOCK(inp); 697 return (0); 698 } 699 700 static void 701 rip6_detach(struct socket *so) 702 { 703 struct inpcb *inp; 704 705 inp = sotoinpcb(so); 706 KASSERT(inp != NULL, ("rip6_detach: inp == NULL")); 707 708 if (so == V_ip6_mrouter && ip6_mrouter_done) 709 ip6_mrouter_done(); 710 /* xxx: RSVP */ 711 INP_INFO_WLOCK(&V_ripcbinfo); 712 INP_WLOCK(inp); 713 free(inp->in6p_icmp6filt, M_PCB); 714 in_pcbdetach(inp); 715 in_pcbfree(inp); 716 INP_INFO_WUNLOCK(&V_ripcbinfo); 717 } 718 719 /* XXXRW: This can't ever be called. */ 720 static void 721 rip6_abort(struct socket *so) 722 { 723 struct inpcb *inp; 724 725 inp = sotoinpcb(so); 726 KASSERT(inp != NULL, ("rip6_abort: inp == NULL")); 727 728 soisdisconnected(so); 729 } 730 731 static void 732 rip6_close(struct socket *so) 733 { 734 struct inpcb *inp; 735 736 inp = sotoinpcb(so); 737 KASSERT(inp != NULL, ("rip6_close: inp == NULL")); 738 739 soisdisconnected(so); 740 } 741 742 static int 743 rip6_disconnect(struct socket *so) 744 { 745 struct inpcb *inp; 746 747 inp = sotoinpcb(so); 748 KASSERT(inp != NULL, ("rip6_disconnect: inp == NULL")); 749 750 if ((so->so_state & SS_ISCONNECTED) == 0) 751 return (ENOTCONN); 752 inp->in6p_faddr = in6addr_any; 753 rip6_abort(so); 754 return (0); 755 } 756 757 static int 758 rip6_bind(struct socket *so, struct sockaddr *nam, struct thread *td) 759 { 760 struct inpcb *inp; 761 struct sockaddr_in6 *addr = (struct sockaddr_in6 *)nam; 762 struct ifaddr *ifa = NULL; 763 int error = 0; 764 765 inp = sotoinpcb(so); 766 KASSERT(inp != NULL, ("rip6_bind: inp == NULL")); 767 768 if (nam->sa_len != sizeof(*addr)) 769 return (EINVAL); 770 if ((error = prison_check_ip6(td->td_ucred, &addr->sin6_addr)) != 0) 771 return (error); 772 if (TAILQ_EMPTY(&V_ifnet) || addr->sin6_family != AF_INET6) 773 return (EADDRNOTAVAIL); 774 if ((error = sa6_embedscope(addr, V_ip6_use_defzone)) != 0) 775 return (error); 776 777 if (!IN6_IS_ADDR_UNSPECIFIED(&addr->sin6_addr) && 778 (ifa = ifa_ifwithaddr((struct sockaddr *)addr)) == NULL) 779 return (EADDRNOTAVAIL); 780 if (ifa != NULL && 781 ((struct in6_ifaddr *)ifa)->ia6_flags & 782 (IN6_IFF_ANYCAST|IN6_IFF_NOTREADY| 783 IN6_IFF_DETACHED|IN6_IFF_DEPRECATED)) { 784 ifa_free(ifa); 785 return (EADDRNOTAVAIL); 786 } 787 if (ifa != NULL) 788 ifa_free(ifa); 789 INP_INFO_WLOCK(&V_ripcbinfo); 790 INP_WLOCK(inp); 791 inp->in6p_laddr = addr->sin6_addr; 792 INP_WUNLOCK(inp); 793 INP_INFO_WUNLOCK(&V_ripcbinfo); 794 return (0); 795 } 796 797 static int 798 rip6_connect(struct socket *so, struct sockaddr *nam, struct thread *td) 799 { 800 struct inpcb *inp; 801 struct sockaddr_in6 *addr = (struct sockaddr_in6 *)nam; 802 struct in6_addr in6a; 803 struct ifnet *ifp = NULL; 804 int error = 0, scope_ambiguous = 0; 805 806 inp = sotoinpcb(so); 807 KASSERT(inp != NULL, ("rip6_connect: inp == NULL")); 808 809 if (nam->sa_len != sizeof(*addr)) 810 return (EINVAL); 811 if (TAILQ_EMPTY(&V_ifnet)) 812 return (EADDRNOTAVAIL); 813 if (addr->sin6_family != AF_INET6) 814 return (EAFNOSUPPORT); 815 816 /* 817 * Application should provide a proper zone ID or the use of default 818 * zone IDs should be enabled. Unfortunately, some applications do 819 * not behave as it should, so we need a workaround. Even if an 820 * appropriate ID is not determined, we'll see if we can determine 821 * the outgoing interface. If we can, determine the zone ID based on 822 * the interface below. 823 */ 824 if (addr->sin6_scope_id == 0 && !V_ip6_use_defzone) 825 scope_ambiguous = 1; 826 if ((error = sa6_embedscope(addr, V_ip6_use_defzone)) != 0) 827 return (error); 828 829 INP_INFO_WLOCK(&V_ripcbinfo); 830 INP_WLOCK(inp); 831 /* Source address selection. XXX: need pcblookup? */ 832 error = in6_selectsrc(addr, inp->in6p_outputopts, 833 inp, NULL, so->so_cred, &ifp, &in6a); 834 if (error) { 835 INP_WUNLOCK(inp); 836 INP_INFO_WUNLOCK(&V_ripcbinfo); 837 return (error); 838 } 839 840 /* XXX: see above */ 841 if (ifp && scope_ambiguous && 842 (error = in6_setscope(&addr->sin6_addr, ifp, NULL)) != 0) { 843 INP_WUNLOCK(inp); 844 INP_INFO_WUNLOCK(&V_ripcbinfo); 845 return (error); 846 } 847 inp->in6p_faddr = addr->sin6_addr; 848 inp->in6p_laddr = in6a; 849 soisconnected(so); 850 INP_WUNLOCK(inp); 851 INP_INFO_WUNLOCK(&V_ripcbinfo); 852 return (0); 853 } 854 855 static int 856 rip6_shutdown(struct socket *so) 857 { 858 struct inpcb *inp; 859 860 inp = sotoinpcb(so); 861 KASSERT(inp != NULL, ("rip6_shutdown: inp == NULL")); 862 863 INP_WLOCK(inp); 864 socantsendmore(so); 865 INP_WUNLOCK(inp); 866 return (0); 867 } 868 869 static int 870 rip6_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam, 871 struct mbuf *control, struct thread *td) 872 { 873 struct inpcb *inp; 874 struct sockaddr_in6 tmp; 875 struct sockaddr_in6 *dst; 876 int ret; 877 878 inp = sotoinpcb(so); 879 KASSERT(inp != NULL, ("rip6_send: inp == NULL")); 880 881 /* Always copy sockaddr to avoid overwrites. */ 882 /* Unlocked read. */ 883 if (so->so_state & SS_ISCONNECTED) { 884 if (nam) { 885 m_freem(m); 886 return (EISCONN); 887 } 888 /* XXX */ 889 bzero(&tmp, sizeof(tmp)); 890 tmp.sin6_family = AF_INET6; 891 tmp.sin6_len = sizeof(struct sockaddr_in6); 892 INP_RLOCK(inp); 893 bcopy(&inp->in6p_faddr, &tmp.sin6_addr, 894 sizeof(struct in6_addr)); 895 INP_RUNLOCK(inp); 896 dst = &tmp; 897 } else { 898 if (nam == NULL) { 899 m_freem(m); 900 return (ENOTCONN); 901 } 902 if (nam->sa_len != sizeof(struct sockaddr_in6)) { 903 m_freem(m); 904 return (EINVAL); 905 } 906 tmp = *(struct sockaddr_in6 *)nam; 907 dst = &tmp; 908 909 if (dst->sin6_family == AF_UNSPEC) { 910 /* 911 * XXX: we allow this case for backward 912 * compatibility to buggy applications that 913 * rely on old (and wrong) kernel behavior. 914 */ 915 log(LOG_INFO, "rip6 SEND: address family is " 916 "unspec. Assume AF_INET6\n"); 917 dst->sin6_family = AF_INET6; 918 } else if (dst->sin6_family != AF_INET6) { 919 m_freem(m); 920 return(EAFNOSUPPORT); 921 } 922 } 923 ret = rip6_output(m, so, dst, control); 924 return (ret); 925 } 926 927 struct pr_usrreqs rip6_usrreqs = { 928 .pru_abort = rip6_abort, 929 .pru_attach = rip6_attach, 930 .pru_bind = rip6_bind, 931 .pru_connect = rip6_connect, 932 .pru_control = in6_control, 933 .pru_detach = rip6_detach, 934 .pru_disconnect = rip6_disconnect, 935 .pru_peeraddr = in6_getpeeraddr, 936 .pru_send = rip6_send, 937 .pru_shutdown = rip6_shutdown, 938 .pru_sockaddr = in6_getsockaddr, 939 .pru_close = rip6_close, 940 }; 941