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 1.22 2005/06/02 23:52:42 dillon Exp $ 36 */ 37 38 #include "opt_inet6.h" 39 #include "opt_ipsec.h" 40 #include "opt_random_ip_id.h" 41 42 #include <sys/param.h> 43 #include <sys/systm.h> 44 #include <sys/kernel.h> 45 #include <sys/malloc.h> 46 #include <sys/mbuf.h> 47 #include <sys/proc.h> 48 #include <sys/protosw.h> 49 #include <sys/socket.h> 50 #include <sys/socketvar.h> 51 #include <sys/sysctl.h> 52 #include <sys/thread2.h> 53 54 #include <machine/stdarg.h> 55 56 #include <vm/vm_zone.h> 57 58 #include <net/if.h> 59 #include <net/route.h> 60 61 #define _IP_VHL 62 #include <netinet/in.h> 63 #include <netinet/in_systm.h> 64 #include <netinet/ip.h> 65 #include <netinet/in_pcb.h> 66 #include <netinet/in_var.h> 67 #include <netinet/ip_var.h> 68 69 #include <net/ip_mroute/ip_mroute.h> 70 #include <net/ipfw/ip_fw.h> 71 #include <net/dummynet/ip_dummynet.h> 72 73 #ifdef FAST_IPSEC 74 #include <netproto/ipsec/ipsec.h> 75 #endif /*FAST_IPSEC*/ 76 77 #ifdef IPSEC 78 #include <netinet6/ipsec.h> 79 #endif /*IPSEC*/ 80 81 struct inpcbinfo ripcbinfo; 82 83 /* control hooks for ipfw and dummynet */ 84 ip_fw_ctl_t *ip_fw_ctl_ptr; 85 ip_dn_ctl_t *ip_dn_ctl_ptr; 86 87 /* 88 * hooks for multicast routing. They all default to NULL, 89 * so leave them not initialized and rely on BSS being set to 0. 90 */ 91 92 /* The socket used to communicate with the multicast routing daemon. */ 93 struct socket *ip_mrouter; 94 95 /* The various mrouter and rsvp functions */ 96 int (*ip_mrouter_set)(struct socket *, struct sockopt *); 97 int (*ip_mrouter_get)(struct socket *, struct sockopt *); 98 int (*ip_mrouter_done)(void); 99 int (*ip_mforward)(struct ip *, struct ifnet *, struct mbuf *, 100 struct ip_moptions *); 101 int (*mrt_ioctl)(int, caddr_t); 102 int (*legal_vif_num)(int); 103 u_long (*ip_mcast_src)(int); 104 105 void (*rsvp_input_p)(struct mbuf *m, ...); 106 int (*ip_rsvp_vif)(struct socket *, struct sockopt *); 107 void (*ip_rsvp_force_done)(struct socket *); 108 109 /* 110 * Nominal space allocated to a raw ip socket. 111 */ 112 #define RIPSNDQ 8192 113 #define RIPRCVQ 8192 114 115 /* 116 * Raw interface to IP protocol. 117 */ 118 119 /* 120 * Initialize raw connection block queue. 121 */ 122 void 123 rip_init(void) 124 { 125 in_pcbinfo_init(&ripcbinfo); 126 /* 127 * XXX We don't use the hash list for raw IP, but it's easier 128 * to allocate a one entry hash list than it is to check all 129 * over the place for hashbase == NULL. 130 */ 131 ripcbinfo.hashbase = hashinit(1, M_PCB, &ripcbinfo.hashmask); 132 ripcbinfo.porthashbase = hashinit(1, M_PCB, &ripcbinfo.porthashmask); 133 ripcbinfo.wildcardhashbase = hashinit(1, M_PCB, 134 &ripcbinfo.wildcardhashmask); 135 ripcbinfo.ipi_zone = zinit("ripcb", sizeof(struct inpcb), 136 maxsockets, ZONE_INTERRUPT, 0); 137 } 138 139 /* 140 * Setup generic address and protocol structures 141 * for raw_input routine, then pass them along with 142 * mbuf chain. 143 */ 144 void 145 rip_input(struct mbuf *m, ...) 146 { 147 struct sockaddr_in ripsrc = { sizeof ripsrc, AF_INET }; 148 struct ip *ip = mtod(m, struct ip *); 149 struct inpcb *inp; 150 struct inpcb *last = NULL; 151 struct mbuf *opts = NULL; 152 int off, proto; 153 __va_list ap; 154 155 __va_start(ap, m); 156 off = __va_arg(ap, int); 157 proto = __va_arg(ap, int); 158 __va_end(ap); 159 160 ripsrc.sin_addr = ip->ip_src; 161 LIST_FOREACH(inp, &ripcbinfo.pcblisthead, inp_list) { 162 if (inp->inp_flags & INP_PLACEMARKER) 163 continue; 164 #ifdef INET6 165 if ((inp->inp_vflag & INP_IPV4) == 0) 166 continue; 167 #endif 168 if (inp->inp_ip_p && inp->inp_ip_p != proto) 169 continue; 170 if (inp->inp_laddr.s_addr != INADDR_ANY && 171 inp->inp_laddr.s_addr != ip->ip_dst.s_addr) 172 continue; 173 if (inp->inp_faddr.s_addr != INADDR_ANY && 174 inp->inp_faddr.s_addr != ip->ip_src.s_addr) 175 continue; 176 if (last) { 177 struct mbuf *n = m_copypacket(m, MB_DONTWAIT); 178 179 #ifdef IPSEC 180 /* check AH/ESP integrity. */ 181 if (n && ipsec4_in_reject_so(n, last->inp_socket)) { 182 m_freem(n); 183 ipsecstat.in_polvio++; 184 /* do not inject data to pcb */ 185 } else 186 #endif /*IPSEC*/ 187 #ifdef FAST_IPSEC 188 /* check AH/ESP integrity. */ 189 if (ipsec4_in_reject(n, last)) { 190 m_freem(n); 191 /* do not inject data to pcb */ 192 } else 193 #endif /*FAST_IPSEC*/ 194 if (n) { 195 if (last->inp_flags & INP_CONTROLOPTS || 196 last->inp_socket->so_options & SO_TIMESTAMP) 197 ip_savecontrol(last, &opts, ip, n); 198 if (sbappendaddr(&last->inp_socket->so_rcv, 199 (struct sockaddr *)&ripsrc, n, 200 opts) == 0) { 201 /* should notify about lost packet */ 202 m_freem(n); 203 if (opts) 204 m_freem(opts); 205 } else 206 sorwakeup(last->inp_socket); 207 opts = 0; 208 } 209 } 210 last = inp; 211 } 212 #ifdef IPSEC 213 /* check AH/ESP integrity. */ 214 if (last && ipsec4_in_reject_so(m, last->inp_socket)) { 215 m_freem(m); 216 ipsecstat.in_polvio++; 217 ipstat.ips_delivered--; 218 /* do not inject data to pcb */ 219 } else 220 #endif /*IPSEC*/ 221 #ifdef FAST_IPSEC 222 /* check AH/ESP integrity. */ 223 if (last && ipsec4_in_reject(m, last)) { 224 m_freem(m); 225 ipstat.ips_delivered--; 226 /* do not inject data to pcb */ 227 } else 228 #endif /*FAST_IPSEC*/ 229 if (last) { 230 if (last->inp_flags & INP_CONTROLOPTS || 231 last->inp_socket->so_options & SO_TIMESTAMP) 232 ip_savecontrol(last, &opts, ip, m); 233 if (sbappendaddr(&last->inp_socket->so_rcv, 234 (struct sockaddr *)&ripsrc, m, opts) == 0) { 235 m_freem(m); 236 if (opts) 237 m_freem(opts); 238 } else 239 sorwakeup(last->inp_socket); 240 } else { 241 m_freem(m); 242 ipstat.ips_noproto++; 243 ipstat.ips_delivered--; 244 } 245 } 246 247 /* 248 * Generate IP header and pass packet to ip_output. 249 * Tack on options user may have setup with control call. 250 */ 251 int 252 rip_output(struct mbuf *m, struct socket *so, ...) 253 { 254 struct ip *ip; 255 struct inpcb *inp = so->so_pcb; 256 __va_list ap; 257 int flags = (so->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST; 258 u_long dst; 259 260 __va_start(ap, so); 261 dst = __va_arg(ap, u_long); 262 __va_end(ap); 263 264 /* 265 * If the user handed us a complete IP packet, use it. 266 * Otherwise, allocate an mbuf for a header and fill it in. 267 */ 268 if ((inp->inp_flags & INP_HDRINCL) == 0) { 269 if (m->m_pkthdr.len + sizeof(struct ip) > IP_MAXPACKET) { 270 m_freem(m); 271 return(EMSGSIZE); 272 } 273 M_PREPEND(m, sizeof(struct ip), MB_WAIT); 274 if (m == NULL) 275 return(ENOBUFS); 276 ip = mtod(m, struct ip *); 277 ip->ip_tos = inp->inp_ip_tos; 278 ip->ip_off = 0; 279 ip->ip_p = inp->inp_ip_p; 280 ip->ip_len = m->m_pkthdr.len; 281 ip->ip_src = inp->inp_laddr; 282 ip->ip_dst.s_addr = dst; 283 ip->ip_ttl = inp->inp_ip_ttl; 284 } else { 285 if (m->m_pkthdr.len > IP_MAXPACKET) { 286 m_freem(m); 287 return(EMSGSIZE); 288 } 289 ip = mtod(m, struct ip *); 290 /* don't allow both user specified and setsockopt options, 291 and don't allow packet length sizes that will crash */ 292 if (((IP_VHL_HL(ip->ip_vhl) != (sizeof (*ip) >> 2)) && 293 inp->inp_options) || 294 (ip->ip_len > m->m_pkthdr.len) || 295 (ip->ip_len < (IP_VHL_HL(ip->ip_vhl) << 2))) { 296 m_freem(m); 297 return EINVAL; 298 } 299 if (ip->ip_id == 0) 300 #ifdef RANDOM_IP_ID 301 ip->ip_id = ip_randomid(); 302 #else 303 ip->ip_id = htons(ip_id++); 304 #endif 305 /* XXX prevent ip_output from overwriting header fields */ 306 flags |= IP_RAWOUTPUT; 307 ipstat.ips_rawout++; 308 } 309 310 return (ip_output(m, inp->inp_options, &inp->inp_route, flags, 311 inp->inp_moptions, inp)); 312 } 313 314 /* 315 * Raw IP socket option processing. 316 */ 317 int 318 rip_ctloutput(struct socket *so, struct sockopt *sopt) 319 { 320 struct inpcb *inp = so->so_pcb; 321 int error, optval; 322 323 if (sopt->sopt_level != IPPROTO_IP) 324 return (EINVAL); 325 326 error = 0; 327 328 switch (sopt->sopt_dir) { 329 case SOPT_GET: 330 switch (sopt->sopt_name) { 331 case IP_HDRINCL: 332 optval = inp->inp_flags & INP_HDRINCL; 333 error = sooptcopyout(sopt, &optval, sizeof optval); 334 break; 335 336 case IP_FW_ADD: /* ADD actually returns the body... */ 337 case IP_FW_GET: 338 if (IPFW_LOADED) 339 error = ip_fw_ctl_ptr(sopt); 340 else 341 error = ENOPROTOOPT; 342 break; 343 344 case IP_DUMMYNET_GET: 345 if (DUMMYNET_LOADED) 346 error = ip_dn_ctl_ptr(sopt); 347 else 348 error = ENOPROTOOPT; 349 break ; 350 351 case MRT_INIT: 352 case MRT_DONE: 353 case MRT_ADD_VIF: 354 case MRT_DEL_VIF: 355 case MRT_ADD_MFC: 356 case MRT_DEL_MFC: 357 case MRT_VERSION: 358 case MRT_ASSERT: 359 case MRT_API_SUPPORT: 360 case MRT_API_CONFIG: 361 case MRT_ADD_BW_UPCALL: 362 case MRT_DEL_BW_UPCALL: 363 error = ip_mrouter_get ? ip_mrouter_get(so, sopt) : 364 EOPNOTSUPP; 365 break; 366 367 default: 368 error = ip_ctloutput(so, sopt); 369 break; 370 } 371 break; 372 373 case SOPT_SET: 374 switch (sopt->sopt_name) { 375 case IP_HDRINCL: 376 error = sooptcopyin(sopt, &optval, sizeof optval, 377 sizeof optval); 378 if (error) 379 break; 380 if (optval) 381 inp->inp_flags |= INP_HDRINCL; 382 else 383 inp->inp_flags &= ~INP_HDRINCL; 384 break; 385 386 case IP_FW_ADD: 387 case IP_FW_DEL: 388 case IP_FW_FLUSH: 389 case IP_FW_ZERO: 390 case IP_FW_RESETLOG: 391 if (IPFW_LOADED) 392 error = ip_fw_ctl_ptr(sopt); 393 else 394 error = ENOPROTOOPT; 395 break; 396 397 case IP_DUMMYNET_CONFIGURE: 398 case IP_DUMMYNET_DEL: 399 case IP_DUMMYNET_FLUSH: 400 if (DUMMYNET_LOADED) 401 error = ip_dn_ctl_ptr(sopt); 402 else 403 error = ENOPROTOOPT ; 404 break ; 405 406 case IP_RSVP_ON: 407 error = ip_rsvp_init(so); 408 break; 409 410 case IP_RSVP_OFF: 411 error = ip_rsvp_done(); 412 break; 413 414 case IP_RSVP_VIF_ON: 415 case IP_RSVP_VIF_OFF: 416 error = ip_rsvp_vif ? 417 ip_rsvp_vif(so, sopt) : EINVAL; 418 break; 419 420 case MRT_INIT: 421 case MRT_DONE: 422 case MRT_ADD_VIF: 423 case MRT_DEL_VIF: 424 case MRT_ADD_MFC: 425 case MRT_DEL_MFC: 426 case MRT_VERSION: 427 case MRT_ASSERT: 428 case MRT_API_SUPPORT: 429 case MRT_API_CONFIG: 430 case MRT_ADD_BW_UPCALL: 431 case MRT_DEL_BW_UPCALL: 432 error = ip_mrouter_set ? ip_mrouter_set(so, sopt) : 433 EOPNOTSUPP; 434 break; 435 436 default: 437 error = ip_ctloutput(so, sopt); 438 break; 439 } 440 break; 441 } 442 443 return (error); 444 } 445 446 /* 447 * This function exists solely to receive the PRC_IFDOWN messages which 448 * are sent by if_down(). It looks for an ifaddr whose ifa_addr is sa, 449 * and calls in_ifadown() to remove all routes corresponding to that address. 450 * It also receives the PRC_IFUP messages from if_up() and reinstalls the 451 * interface routes. 452 */ 453 void 454 rip_ctlinput(int cmd, struct sockaddr *sa, void *vip) 455 { 456 struct in_ifaddr *ia; 457 struct ifnet *ifp; 458 int err; 459 int flags; 460 461 switch (cmd) { 462 case PRC_IFDOWN: 463 TAILQ_FOREACH(ia, &in_ifaddrhead, ia_link) { 464 if (ia->ia_ifa.ifa_addr == sa && 465 (ia->ia_flags & IFA_ROUTE)) { 466 /* 467 * in_ifscrub kills the interface route. 468 */ 469 in_ifscrub(ia->ia_ifp, ia); 470 /* 471 * in_ifadown gets rid of all the rest of 472 * the routes. This is not quite the right 473 * thing to do, but at least if we are running 474 * a routing process they will come back. 475 */ 476 in_ifadown(&ia->ia_ifa, 0); 477 break; 478 } 479 } 480 break; 481 482 case PRC_IFUP: 483 TAILQ_FOREACH(ia, &in_ifaddrhead, ia_link) { 484 if (ia->ia_ifa.ifa_addr == sa) 485 break; 486 } 487 if (ia == 0 || (ia->ia_flags & IFA_ROUTE)) 488 return; 489 flags = RTF_UP; 490 ifp = ia->ia_ifa.ifa_ifp; 491 492 if ((ifp->if_flags & IFF_LOOPBACK) || 493 (ifp->if_flags & IFF_POINTOPOINT)) 494 flags |= RTF_HOST; 495 496 err = rtinit(&ia->ia_ifa, RTM_ADD, flags); 497 if (err == 0) 498 ia->ia_flags |= IFA_ROUTE; 499 break; 500 } 501 } 502 503 u_long rip_sendspace = RIPSNDQ; 504 u_long rip_recvspace = RIPRCVQ; 505 506 SYSCTL_INT(_net_inet_raw, OID_AUTO, maxdgram, CTLFLAG_RW, 507 &rip_sendspace, 0, "Maximum outgoing raw IP datagram size"); 508 SYSCTL_INT(_net_inet_raw, OID_AUTO, recvspace, CTLFLAG_RW, 509 &rip_recvspace, 0, "Maximum incoming raw IP datagram size"); 510 511 static int 512 rip_attach(struct socket *so, int proto, struct pru_attach_info *ai) 513 { 514 struct inpcb *inp; 515 int error; 516 517 inp = so->so_pcb; 518 if (inp) 519 panic("rip_attach"); 520 if ((error = suser_cred(ai->p_ucred, NULL_CRED_OKAY)) != 0) 521 return error; 522 523 error = soreserve(so, rip_sendspace, rip_recvspace, ai->sb_rlimit); 524 if (error) 525 return error; 526 crit_enter(); 527 error = in_pcballoc(so, &ripcbinfo); 528 crit_exit(); 529 if (error) 530 return error; 531 inp = (struct inpcb *)so->so_pcb; 532 inp->inp_vflag |= INP_IPV4; 533 inp->inp_ip_p = proto; 534 inp->inp_ip_ttl = ip_defttl; 535 return 0; 536 } 537 538 static int 539 rip_detach(struct socket *so) 540 { 541 struct inpcb *inp; 542 543 inp = so->so_pcb; 544 if (inp == 0) 545 panic("rip_detach"); 546 if (so == ip_mrouter && ip_mrouter_done) 547 ip_mrouter_done(); 548 if (ip_rsvp_force_done) 549 ip_rsvp_force_done(so); 550 if (so == ip_rsvpd) 551 ip_rsvp_done(); 552 in_pcbdetach(inp); 553 return 0; 554 } 555 556 static int 557 rip_abort(struct socket *so) 558 { 559 soisdisconnected(so); 560 if (so->so_state & SS_NOFDREF) 561 return rip_detach(so); 562 return 0; 563 } 564 565 static int 566 rip_disconnect(struct socket *so) 567 { 568 if ((so->so_state & SS_ISCONNECTED) == 0) 569 return ENOTCONN; 570 return rip_abort(so); 571 } 572 573 static int 574 rip_bind(struct socket *so, struct sockaddr *nam, struct thread *td) 575 { 576 struct inpcb *inp = so->so_pcb; 577 struct sockaddr_in *addr = (struct sockaddr_in *)nam; 578 579 if (nam->sa_len != sizeof(*addr)) 580 return EINVAL; 581 582 if (TAILQ_EMPTY(&ifnet) || ((addr->sin_family != AF_INET) && 583 (addr->sin_family != AF_IMPLINK)) || 584 (addr->sin_addr.s_addr != INADDR_ANY && 585 ifa_ifwithaddr((struct sockaddr *)addr) == 0)) 586 return EADDRNOTAVAIL; 587 inp->inp_laddr = addr->sin_addr; 588 return 0; 589 } 590 591 static int 592 rip_connect(struct socket *so, struct sockaddr *nam, struct thread *td) 593 { 594 struct inpcb *inp = so->so_pcb; 595 struct sockaddr_in *addr = (struct sockaddr_in *)nam; 596 597 if (nam->sa_len != sizeof(*addr)) 598 return EINVAL; 599 if (TAILQ_EMPTY(&ifnet)) 600 return EADDRNOTAVAIL; 601 if ((addr->sin_family != AF_INET) && 602 (addr->sin_family != AF_IMPLINK)) 603 return EAFNOSUPPORT; 604 inp->inp_faddr = addr->sin_addr; 605 soisconnected(so); 606 return 0; 607 } 608 609 static int 610 rip_shutdown(struct socket *so) 611 { 612 socantsendmore(so); 613 return 0; 614 } 615 616 static int 617 rip_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam, 618 struct mbuf *control, struct thread *td) 619 { 620 struct inpcb *inp = so->so_pcb; 621 u_long dst; 622 623 if (so->so_state & SS_ISCONNECTED) { 624 if (nam) { 625 m_freem(m); 626 return EISCONN; 627 } 628 dst = inp->inp_faddr.s_addr; 629 } else { 630 if (nam == NULL) { 631 m_freem(m); 632 return ENOTCONN; 633 } 634 dst = ((struct sockaddr_in *)nam)->sin_addr.s_addr; 635 } 636 return rip_output(m, so, dst); 637 } 638 639 SYSCTL_PROC(_net_inet_raw, OID_AUTO/*XXX*/, pcblist, CTLFLAG_RD, &ripcbinfo, 0, 640 in_pcblist_global, "S,xinpcb", "List of active raw IP sockets"); 641 642 struct pr_usrreqs rip_usrreqs = { 643 rip_abort, pru_accept_notsupp, rip_attach, rip_bind, rip_connect, 644 pru_connect2_notsupp, in_control, rip_detach, rip_disconnect, 645 pru_listen_notsupp, in_setpeeraddr, pru_rcvd_notsupp, 646 pru_rcvoob_notsupp, rip_send, pru_sense_null, rip_shutdown, 647 in_setsockaddr, sosend, soreceive, sopoll 648 }; 649