1 /* 2 * Copyright (c) 1982, 1986, 1988, 1993 3 * The Regents of the University of California. 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. All advertising materials mentioning features or use of this software 14 * must display the following acknowledgement: 15 * This product includes software developed by the University of 16 * California, Berkeley and its contributors. 17 * 4. Neither the name of the University nor the names of its contributors 18 * may be used to endorse or promote products derived from this software 19 * without specific prior written permission. 20 * 21 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 24 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 31 * SUCH DAMAGE. 32 * 33 * @(#)raw_ip.c 8.7 (Berkeley) 5/15/95 34 * $FreeBSD: src/sys/netinet/raw_ip.c,v 1.64.2.16 2003/08/24 08:24:38 hsu Exp $ 35 * $DragonFly: src/sys/netinet/raw_ip.c,v 1.26 2007/04/22 01:13:14 dillon Exp $ 36 */ 37 38 #include "opt_inet6.h" 39 #include "opt_ipsec.h" 40 41 #include <sys/param.h> 42 #include <sys/systm.h> 43 #include <sys/kernel.h> 44 #include <sys/malloc.h> 45 #include <sys/mbuf.h> 46 #include <sys/proc.h> 47 #include <sys/protosw.h> 48 #include <sys/socket.h> 49 #include <sys/socketvar.h> 50 #include <sys/sysctl.h> 51 #include <sys/thread2.h> 52 53 #include <machine/stdarg.h> 54 55 #include <vm/vm_zone.h> 56 57 #include <net/if.h> 58 #include <net/route.h> 59 60 #define _IP_VHL 61 #include <netinet/in.h> 62 #include <netinet/in_systm.h> 63 #include <netinet/ip.h> 64 #include <netinet/in_pcb.h> 65 #include <netinet/in_var.h> 66 #include <netinet/ip_var.h> 67 68 #include <net/ip_mroute/ip_mroute.h> 69 #include <net/ipfw/ip_fw.h> 70 #include <net/dummynet/ip_dummynet.h> 71 72 #ifdef FAST_IPSEC 73 #include <netproto/ipsec/ipsec.h> 74 #endif /*FAST_IPSEC*/ 75 76 #ifdef IPSEC 77 #include <netinet6/ipsec.h> 78 #endif /*IPSEC*/ 79 80 struct inpcbinfo ripcbinfo; 81 82 /* control hooks for ipfw and dummynet */ 83 ip_fw_ctl_t *ip_fw_ctl_ptr; 84 ip_dn_ctl_t *ip_dn_ctl_ptr; 85 86 /* 87 * hooks for multicast routing. They all default to NULL, 88 * so leave them not initialized and rely on BSS being set to 0. 89 */ 90 91 /* The socket used to communicate with the multicast routing daemon. */ 92 struct socket *ip_mrouter; 93 94 /* The various mrouter and rsvp functions */ 95 int (*ip_mrouter_set)(struct socket *, struct sockopt *); 96 int (*ip_mrouter_get)(struct socket *, struct sockopt *); 97 int (*ip_mrouter_done)(void); 98 int (*ip_mforward)(struct ip *, struct ifnet *, struct mbuf *, 99 struct ip_moptions *); 100 int (*mrt_ioctl)(int, caddr_t); 101 int (*legal_vif_num)(int); 102 u_long (*ip_mcast_src)(int); 103 104 void (*rsvp_input_p)(struct mbuf *m, ...); 105 int (*ip_rsvp_vif)(struct socket *, struct sockopt *); 106 void (*ip_rsvp_force_done)(struct socket *); 107 108 /* 109 * Nominal space allocated to a raw ip socket. 110 */ 111 #define RIPSNDQ 8192 112 #define RIPRCVQ 8192 113 114 /* 115 * Raw interface to IP protocol. 116 */ 117 118 /* 119 * Initialize raw connection block queue. 120 */ 121 void 122 rip_init(void) 123 { 124 in_pcbinfo_init(&ripcbinfo); 125 /* 126 * XXX We don't use the hash list for raw IP, but it's easier 127 * to allocate a one entry hash list than it is to check all 128 * over the place for hashbase == NULL. 129 */ 130 ripcbinfo.hashbase = hashinit(1, M_PCB, &ripcbinfo.hashmask); 131 ripcbinfo.porthashbase = hashinit(1, M_PCB, &ripcbinfo.porthashmask); 132 ripcbinfo.wildcardhashbase = hashinit(1, M_PCB, 133 &ripcbinfo.wildcardhashmask); 134 ripcbinfo.ipi_zone = zinit("ripcb", sizeof(struct inpcb), 135 maxsockets, ZONE_INTERRUPT, 0); 136 } 137 138 /* 139 * Setup generic address and protocol structures 140 * for raw_input routine, then pass them along with 141 * mbuf chain. 142 */ 143 void 144 rip_input(struct mbuf *m, ...) 145 { 146 struct sockaddr_in ripsrc = { sizeof ripsrc, AF_INET }; 147 struct ip *ip = mtod(m, struct ip *); 148 struct inpcb *inp; 149 struct inpcb *last = NULL; 150 struct mbuf *opts = NULL; 151 int off, proto; 152 __va_list ap; 153 154 __va_start(ap, m); 155 off = __va_arg(ap, int); 156 proto = __va_arg(ap, int); 157 __va_end(ap); 158 159 ripsrc.sin_addr = ip->ip_src; 160 LIST_FOREACH(inp, &ripcbinfo.pcblisthead, inp_list) { 161 if (inp->inp_flags & INP_PLACEMARKER) 162 continue; 163 #ifdef INET6 164 if ((inp->inp_vflag & INP_IPV4) == 0) 165 continue; 166 #endif 167 if (inp->inp_ip_p && inp->inp_ip_p != proto) 168 continue; 169 if (inp->inp_laddr.s_addr != INADDR_ANY && 170 inp->inp_laddr.s_addr != ip->ip_dst.s_addr) 171 continue; 172 if (inp->inp_faddr.s_addr != INADDR_ANY && 173 inp->inp_faddr.s_addr != ip->ip_src.s_addr) 174 continue; 175 if (last) { 176 struct mbuf *n = m_copypacket(m, MB_DONTWAIT); 177 178 #ifdef IPSEC 179 /* check AH/ESP integrity. */ 180 if (n && ipsec4_in_reject_so(n, last->inp_socket)) { 181 m_freem(n); 182 ipsecstat.in_polvio++; 183 /* do not inject data to pcb */ 184 } else 185 #endif /*IPSEC*/ 186 #ifdef FAST_IPSEC 187 /* check AH/ESP integrity. */ 188 if (ipsec4_in_reject(n, last)) { 189 m_freem(n); 190 /* do not inject data to pcb */ 191 } else 192 #endif /*FAST_IPSEC*/ 193 if (n) { 194 if (last->inp_flags & INP_CONTROLOPTS || 195 last->inp_socket->so_options & SO_TIMESTAMP) 196 ip_savecontrol(last, &opts, ip, n); 197 if (ssb_appendaddr(&last->inp_socket->so_rcv, 198 (struct sockaddr *)&ripsrc, n, 199 opts) == 0) { 200 /* should notify about lost packet */ 201 m_freem(n); 202 if (opts) 203 m_freem(opts); 204 } else 205 sorwakeup(last->inp_socket); 206 opts = 0; 207 } 208 } 209 last = inp; 210 } 211 #ifdef IPSEC 212 /* check AH/ESP integrity. */ 213 if (last && ipsec4_in_reject_so(m, last->inp_socket)) { 214 m_freem(m); 215 ipsecstat.in_polvio++; 216 ipstat.ips_delivered--; 217 /* do not inject data to pcb */ 218 } else 219 #endif /*IPSEC*/ 220 #ifdef FAST_IPSEC 221 /* check AH/ESP integrity. */ 222 if (last && ipsec4_in_reject(m, last)) { 223 m_freem(m); 224 ipstat.ips_delivered--; 225 /* do not inject data to pcb */ 226 } else 227 #endif /*FAST_IPSEC*/ 228 /* Check the minimum TTL for socket. */ 229 if (last && ip->ip_ttl < last->inp_ip_minttl) { 230 m_freem(opts); 231 ipstat.ips_delivered--; 232 } else if (last) { 233 if (last->inp_flags & INP_CONTROLOPTS || 234 last->inp_socket->so_options & SO_TIMESTAMP) 235 ip_savecontrol(last, &opts, ip, m); 236 if (ssb_appendaddr(&last->inp_socket->so_rcv, 237 (struct sockaddr *)&ripsrc, m, opts) == 0) { 238 m_freem(m); 239 if (opts) 240 m_freem(opts); 241 } else 242 sorwakeup(last->inp_socket); 243 } else { 244 m_freem(m); 245 ipstat.ips_noproto++; 246 ipstat.ips_delivered--; 247 } 248 } 249 250 /* 251 * Generate IP header and pass packet to ip_output. 252 * Tack on options user may have setup with control call. 253 */ 254 int 255 rip_output(struct mbuf *m, struct socket *so, ...) 256 { 257 struct ip *ip; 258 struct inpcb *inp = so->so_pcb; 259 __va_list ap; 260 int flags = (so->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST; 261 u_long dst; 262 263 __va_start(ap, so); 264 dst = __va_arg(ap, u_long); 265 __va_end(ap); 266 267 /* 268 * If the user handed us a complete IP packet, use it. 269 * Otherwise, allocate an mbuf for a header and fill it in. 270 */ 271 if ((inp->inp_flags & INP_HDRINCL) == 0) { 272 if (m->m_pkthdr.len + sizeof(struct ip) > IP_MAXPACKET) { 273 m_freem(m); 274 return(EMSGSIZE); 275 } 276 M_PREPEND(m, sizeof(struct ip), MB_WAIT); 277 if (m == NULL) 278 return(ENOBUFS); 279 ip = mtod(m, struct ip *); 280 ip->ip_tos = inp->inp_ip_tos; 281 ip->ip_off = 0; 282 ip->ip_p = inp->inp_ip_p; 283 ip->ip_len = m->m_pkthdr.len; 284 ip->ip_src = inp->inp_laddr; 285 ip->ip_dst.s_addr = dst; 286 ip->ip_ttl = inp->inp_ip_ttl; 287 } else { 288 if (m->m_pkthdr.len > IP_MAXPACKET) { 289 m_freem(m); 290 return(EMSGSIZE); 291 } 292 ip = mtod(m, struct ip *); 293 /* don't allow both user specified and setsockopt options, 294 and don't allow packet length sizes that will crash */ 295 if (((IP_VHL_HL(ip->ip_vhl) != (sizeof (*ip) >> 2)) && 296 inp->inp_options) || 297 (ip->ip_len > m->m_pkthdr.len) || 298 (ip->ip_len < (IP_VHL_HL(ip->ip_vhl) << 2))) { 299 m_freem(m); 300 return EINVAL; 301 } 302 if (ip->ip_id == 0) 303 ip->ip_id = ip_newid(); 304 /* XXX prevent ip_output from overwriting header fields */ 305 flags |= IP_RAWOUTPUT; 306 ipstat.ips_rawout++; 307 } 308 309 return (ip_output(m, inp->inp_options, &inp->inp_route, flags, 310 inp->inp_moptions, inp)); 311 } 312 313 /* 314 * Raw IP socket option processing. 315 */ 316 int 317 rip_ctloutput(struct socket *so, struct sockopt *sopt) 318 { 319 struct inpcb *inp = so->so_pcb; 320 int error, optval; 321 322 if (sopt->sopt_level != IPPROTO_IP) 323 return (EINVAL); 324 325 error = 0; 326 327 switch (sopt->sopt_dir) { 328 case SOPT_GET: 329 switch (sopt->sopt_name) { 330 case IP_HDRINCL: 331 optval = inp->inp_flags & INP_HDRINCL; 332 error = sooptcopyout(sopt, &optval, sizeof optval); 333 break; 334 335 case IP_FW_ADD: /* ADD actually returns the body... */ 336 case IP_FW_GET: 337 if (IPFW_LOADED) 338 error = ip_fw_ctl_ptr(sopt); 339 else 340 error = ENOPROTOOPT; 341 break; 342 343 case IP_DUMMYNET_GET: 344 if (DUMMYNET_LOADED) 345 error = ip_dn_ctl_ptr(sopt); 346 else 347 error = ENOPROTOOPT; 348 break ; 349 350 case MRT_INIT: 351 case MRT_DONE: 352 case MRT_ADD_VIF: 353 case MRT_DEL_VIF: 354 case MRT_ADD_MFC: 355 case MRT_DEL_MFC: 356 case MRT_VERSION: 357 case MRT_ASSERT: 358 case MRT_API_SUPPORT: 359 case MRT_API_CONFIG: 360 case MRT_ADD_BW_UPCALL: 361 case MRT_DEL_BW_UPCALL: 362 error = ip_mrouter_get ? ip_mrouter_get(so, sopt) : 363 EOPNOTSUPP; 364 break; 365 366 default: 367 error = ip_ctloutput(so, sopt); 368 break; 369 } 370 break; 371 372 case SOPT_SET: 373 switch (sopt->sopt_name) { 374 case IP_HDRINCL: 375 error = sooptcopyin(sopt, &optval, sizeof optval, 376 sizeof optval); 377 if (error) 378 break; 379 if (optval) 380 inp->inp_flags |= INP_HDRINCL; 381 else 382 inp->inp_flags &= ~INP_HDRINCL; 383 break; 384 385 case IP_FW_ADD: 386 case IP_FW_DEL: 387 case IP_FW_FLUSH: 388 case IP_FW_ZERO: 389 case IP_FW_RESETLOG: 390 if (IPFW_LOADED) 391 error = ip_fw_ctl_ptr(sopt); 392 else 393 error = ENOPROTOOPT; 394 break; 395 396 case IP_DUMMYNET_CONFIGURE: 397 case IP_DUMMYNET_DEL: 398 case IP_DUMMYNET_FLUSH: 399 if (DUMMYNET_LOADED) 400 error = ip_dn_ctl_ptr(sopt); 401 else 402 error = ENOPROTOOPT ; 403 break ; 404 405 case IP_RSVP_ON: 406 error = ip_rsvp_init(so); 407 break; 408 409 case IP_RSVP_OFF: 410 error = ip_rsvp_done(); 411 break; 412 413 case IP_RSVP_VIF_ON: 414 case IP_RSVP_VIF_OFF: 415 error = ip_rsvp_vif ? 416 ip_rsvp_vif(so, sopt) : EINVAL; 417 break; 418 419 case MRT_INIT: 420 case MRT_DONE: 421 case MRT_ADD_VIF: 422 case MRT_DEL_VIF: 423 case MRT_ADD_MFC: 424 case MRT_DEL_MFC: 425 case MRT_VERSION: 426 case MRT_ASSERT: 427 case MRT_API_SUPPORT: 428 case MRT_API_CONFIG: 429 case MRT_ADD_BW_UPCALL: 430 case MRT_DEL_BW_UPCALL: 431 error = ip_mrouter_set ? ip_mrouter_set(so, sopt) : 432 EOPNOTSUPP; 433 break; 434 435 default: 436 error = ip_ctloutput(so, sopt); 437 break; 438 } 439 break; 440 } 441 442 return (error); 443 } 444 445 /* 446 * This function exists solely to receive the PRC_IFDOWN messages which 447 * are sent by if_down(). It looks for an ifaddr whose ifa_addr is sa, 448 * and calls in_ifadown() to remove all routes corresponding to that address. 449 * It also receives the PRC_IFUP messages from if_up() and reinstalls the 450 * interface routes. 451 */ 452 void 453 rip_ctlinput(int cmd, struct sockaddr *sa, void *vip) 454 { 455 struct in_ifaddr *ia; 456 struct ifnet *ifp; 457 int err; 458 int flags; 459 460 switch (cmd) { 461 case PRC_IFDOWN: 462 TAILQ_FOREACH(ia, &in_ifaddrhead, ia_link) { 463 if (ia->ia_ifa.ifa_addr == sa && 464 (ia->ia_flags & IFA_ROUTE)) { 465 /* 466 * in_ifscrub kills the interface route. 467 */ 468 in_ifscrub(ia->ia_ifp, ia); 469 /* 470 * in_ifadown gets rid of all the rest of 471 * the routes. This is not quite the right 472 * thing to do, but at least if we are running 473 * a routing process they will come back. 474 */ 475 in_ifadown(&ia->ia_ifa, 0); 476 break; 477 } 478 } 479 break; 480 481 case PRC_IFUP: 482 TAILQ_FOREACH(ia, &in_ifaddrhead, ia_link) { 483 if (ia->ia_ifa.ifa_addr == sa) 484 break; 485 } 486 if (ia == 0 || (ia->ia_flags & IFA_ROUTE)) 487 return; 488 flags = RTF_UP; 489 ifp = ia->ia_ifa.ifa_ifp; 490 491 if ((ifp->if_flags & IFF_LOOPBACK) || 492 (ifp->if_flags & IFF_POINTOPOINT)) 493 flags |= RTF_HOST; 494 495 err = rtinit(&ia->ia_ifa, RTM_ADD, flags); 496 if (err == 0) 497 ia->ia_flags |= IFA_ROUTE; 498 break; 499 } 500 } 501 502 u_long rip_sendspace = RIPSNDQ; 503 u_long rip_recvspace = RIPRCVQ; 504 505 SYSCTL_INT(_net_inet_raw, OID_AUTO, maxdgram, CTLFLAG_RW, 506 &rip_sendspace, 0, "Maximum outgoing raw IP datagram size"); 507 SYSCTL_INT(_net_inet_raw, OID_AUTO, recvspace, CTLFLAG_RW, 508 &rip_recvspace, 0, "Maximum incoming raw IP datagram size"); 509 510 static int 511 rip_attach(struct socket *so, int proto, struct pru_attach_info *ai) 512 { 513 struct inpcb *inp; 514 int error; 515 516 inp = so->so_pcb; 517 if (inp) 518 panic("rip_attach"); 519 if ((error = suser_cred(ai->p_ucred, NULL_CRED_OKAY)) != 0) 520 return error; 521 522 error = soreserve(so, rip_sendspace, rip_recvspace, ai->sb_rlimit); 523 if (error) 524 return error; 525 crit_enter(); 526 error = in_pcballoc(so, &ripcbinfo); 527 crit_exit(); 528 if (error) 529 return error; 530 inp = (struct inpcb *)so->so_pcb; 531 inp->inp_vflag |= INP_IPV4; 532 inp->inp_ip_p = proto; 533 inp->inp_ip_ttl = ip_defttl; 534 return 0; 535 } 536 537 static int 538 rip_detach(struct socket *so) 539 { 540 struct inpcb *inp; 541 542 inp = so->so_pcb; 543 if (inp == 0) 544 panic("rip_detach"); 545 if (so == ip_mrouter && ip_mrouter_done) 546 ip_mrouter_done(); 547 if (ip_rsvp_force_done) 548 ip_rsvp_force_done(so); 549 if (so == ip_rsvpd) 550 ip_rsvp_done(); 551 in_pcbdetach(inp); 552 return 0; 553 } 554 555 static int 556 rip_abort(struct socket *so) 557 { 558 soisdisconnected(so); 559 if (so->so_state & SS_NOFDREF) 560 return rip_detach(so); 561 return 0; 562 } 563 564 static int 565 rip_disconnect(struct socket *so) 566 { 567 if ((so->so_state & SS_ISCONNECTED) == 0) 568 return ENOTCONN; 569 return rip_abort(so); 570 } 571 572 static int 573 rip_bind(struct socket *so, struct sockaddr *nam, struct thread *td) 574 { 575 struct inpcb *inp = so->so_pcb; 576 struct sockaddr_in *addr = (struct sockaddr_in *)nam; 577 578 if (nam->sa_len != sizeof(*addr)) 579 return EINVAL; 580 581 if (TAILQ_EMPTY(&ifnet) || ((addr->sin_family != AF_INET) && 582 (addr->sin_family != AF_IMPLINK)) || 583 (addr->sin_addr.s_addr != INADDR_ANY && 584 ifa_ifwithaddr((struct sockaddr *)addr) == 0)) 585 return EADDRNOTAVAIL; 586 inp->inp_laddr = addr->sin_addr; 587 return 0; 588 } 589 590 static int 591 rip_connect(struct socket *so, struct sockaddr *nam, struct thread *td) 592 { 593 struct inpcb *inp = so->so_pcb; 594 struct sockaddr_in *addr = (struct sockaddr_in *)nam; 595 596 if (nam->sa_len != sizeof(*addr)) 597 return EINVAL; 598 if (TAILQ_EMPTY(&ifnet)) 599 return EADDRNOTAVAIL; 600 if ((addr->sin_family != AF_INET) && 601 (addr->sin_family != AF_IMPLINK)) 602 return EAFNOSUPPORT; 603 inp->inp_faddr = addr->sin_addr; 604 soisconnected(so); 605 return 0; 606 } 607 608 static int 609 rip_shutdown(struct socket *so) 610 { 611 socantsendmore(so); 612 return 0; 613 } 614 615 static int 616 rip_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam, 617 struct mbuf *control, struct thread *td) 618 { 619 struct inpcb *inp = so->so_pcb; 620 u_long dst; 621 622 if (so->so_state & SS_ISCONNECTED) { 623 if (nam) { 624 m_freem(m); 625 return EISCONN; 626 } 627 dst = inp->inp_faddr.s_addr; 628 } else { 629 if (nam == NULL) { 630 m_freem(m); 631 return ENOTCONN; 632 } 633 dst = ((struct sockaddr_in *)nam)->sin_addr.s_addr; 634 } 635 return rip_output(m, so, dst); 636 } 637 638 SYSCTL_PROC(_net_inet_raw, OID_AUTO/*XXX*/, pcblist, CTLFLAG_RD, &ripcbinfo, 0, 639 in_pcblist_global, "S,xinpcb", "List of active raw IP sockets"); 640 641 struct pr_usrreqs rip_usrreqs = { 642 .pru_abort = rip_abort, 643 .pru_accept = pru_accept_notsupp, 644 .pru_attach = rip_attach, 645 .pru_bind = rip_bind, 646 .pru_connect = rip_connect, 647 .pru_connect2 = pru_connect2_notsupp, 648 .pru_control = in_control, 649 .pru_detach = rip_detach, 650 .pru_disconnect = rip_disconnect, 651 .pru_listen = pru_listen_notsupp, 652 .pru_peeraddr = in_setpeeraddr, 653 .pru_rcvd = pru_rcvd_notsupp, 654 .pru_rcvoob = pru_rcvoob_notsupp, 655 .pru_send = rip_send, 656 .pru_sense = pru_sense_null, 657 .pru_shutdown = rip_shutdown, 658 .pru_sockaddr = in_setsockaddr, 659 .pru_sosend = sosend, 660 .pru_soreceive = soreceive, 661 .pru_sopoll = sopoll 662 }; 663