1 /* $NetBSD: raw_ip.c,v 1.83 2005/02/03 23:16:40 perry Exp $ */ 2 3 /* 4 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. 5 * All rights reserved. 6 * 7 * Redistribution and use in source and binary forms, with or without 8 * modification, are permitted provided that the following conditions 9 * are met: 10 * 1. Redistributions of source code must retain the above copyright 11 * notice, this list of conditions and the following disclaimer. 12 * 2. Redistributions in binary form must reproduce the above copyright 13 * notice, this list of conditions and the following disclaimer in the 14 * documentation and/or other materials provided with the distribution. 15 * 3. Neither the name of the project nor the names of its contributors 16 * may be used to endorse or promote products derived from this software 17 * without specific prior written permission. 18 * 19 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND 20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 22 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE 23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 29 * SUCH DAMAGE. 30 */ 31 32 /* 33 * Copyright (c) 1982, 1986, 1988, 1993 34 * The Regents of the University of California. All rights reserved. 35 * 36 * Redistribution and use in source and binary forms, with or without 37 * modification, are permitted provided that the following conditions 38 * are met: 39 * 1. Redistributions of source code must retain the above copyright 40 * notice, this list of conditions and the following disclaimer. 41 * 2. Redistributions in binary form must reproduce the above copyright 42 * notice, this list of conditions and the following disclaimer in the 43 * documentation and/or other materials provided with the distribution. 44 * 3. Neither the name of the University nor the names of its contributors 45 * may be used to endorse or promote products derived from this software 46 * without specific prior written permission. 47 * 48 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 49 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 50 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 51 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 52 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 53 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 54 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 55 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 56 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 57 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 58 * SUCH DAMAGE. 59 * 60 * @(#)raw_ip.c 8.7 (Berkeley) 5/15/95 61 */ 62 63 #include <sys/cdefs.h> 64 __KERNEL_RCSID(0, "$NetBSD: raw_ip.c,v 1.83 2005/02/03 23:16:40 perry Exp $"); 65 66 #include "opt_inet.h" 67 #include "opt_ipsec.h" 68 #include "opt_mrouting.h" 69 70 #include <sys/param.h> 71 #include <sys/malloc.h> 72 #include <sys/mbuf.h> 73 #include <sys/socket.h> 74 #include <sys/protosw.h> 75 #include <sys/socketvar.h> 76 #include <sys/errno.h> 77 #include <sys/systm.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_systm.h> 85 #include <netinet/ip.h> 86 #include <netinet/ip_var.h> 87 #include <netinet/ip_mroute.h> 88 #include <netinet/ip_icmp.h> 89 #include <netinet/in_pcb.h> 90 #include <netinet/in_var.h> 91 92 #include <machine/stdarg.h> 93 94 #ifdef IPSEC 95 #include <netinet6/ipsec.h> 96 #endif /*IPSEC*/ 97 98 #ifdef FAST_IPSEC 99 #include <netipsec/ipsec.h> 100 #include <netipsec/ipsec_var.h> /* XXX ipsecstat namespace */ 101 #endif /* FAST_IPSEC*/ 102 103 struct inpcbtable rawcbtable; 104 105 int rip_pcbnotify(struct inpcbtable *, struct in_addr, 106 struct in_addr, int, int, void (*)(struct inpcb *, int)); 107 int rip_bind(struct inpcb *, struct mbuf *); 108 int rip_connect(struct inpcb *, struct mbuf *); 109 void rip_disconnect(struct inpcb *); 110 111 /* 112 * Nominal space allocated to a raw ip socket. 113 */ 114 #define RIPSNDQ 8192 115 #define RIPRCVQ 8192 116 117 /* 118 * Raw interface to IP protocol. 119 */ 120 121 /* 122 * Initialize raw connection block q. 123 */ 124 void 125 rip_init(void) 126 { 127 128 in_pcbinit(&rawcbtable, 1, 1); 129 } 130 131 /* 132 * Setup generic address and protocol structures 133 * for raw_input routine, then pass them along with 134 * mbuf chain. 135 */ 136 void 137 rip_input(struct mbuf *m, ...) 138 { 139 int proto; 140 struct ip *ip = mtod(m, struct ip *); 141 struct inpcb_hdr *inph; 142 struct inpcb *inp; 143 struct inpcb *last = 0; 144 struct mbuf *opts = 0; 145 struct sockaddr_in ripsrc; 146 va_list ap; 147 148 va_start(ap, m); 149 (void)va_arg(ap, int); /* ignore value, advance ap */ 150 proto = va_arg(ap, int); 151 va_end(ap); 152 153 ripsrc.sin_family = AF_INET; 154 ripsrc.sin_len = sizeof(struct sockaddr_in); 155 ripsrc.sin_addr = ip->ip_src; 156 ripsrc.sin_port = 0; 157 bzero((caddr_t)ripsrc.sin_zero, sizeof(ripsrc.sin_zero)); 158 159 /* 160 * XXX Compatibility: programs using raw IP expect ip_len 161 * XXX to have the header length subtracted, and in host order. 162 * XXX ip_off is also expected to be host order. 163 */ 164 ip->ip_len = ntohs(ip->ip_len) - (ip->ip_hl << 2); 165 NTOHS(ip->ip_off); 166 167 CIRCLEQ_FOREACH(inph, &rawcbtable.inpt_queue, inph_queue) { 168 inp = (struct inpcb *)inph; 169 if (inp->inp_af != AF_INET) 170 continue; 171 if (inp->inp_ip.ip_p && inp->inp_ip.ip_p != proto) 172 continue; 173 if (!in_nullhost(inp->inp_laddr) && 174 !in_hosteq(inp->inp_laddr, ip->ip_dst)) 175 continue; 176 if (!in_nullhost(inp->inp_faddr) && 177 !in_hosteq(inp->inp_faddr, ip->ip_src)) 178 continue; 179 if (last) { 180 struct mbuf *n; 181 182 #if defined(IPSEC) || defined(FAST_IPSEC) 183 /* check AH/ESP integrity. */ 184 if (ipsec4_in_reject_so(m, last->inp_socket)) { 185 ipsecstat.in_polvio++; 186 /* do not inject data to pcb */ 187 } else 188 #endif /*IPSEC*/ 189 if ((n = m_copy(m, 0, (int)M_COPYALL)) != NULL) { 190 if (last->inp_flags & INP_CONTROLOPTS || 191 last->inp_socket->so_options & SO_TIMESTAMP) 192 ip_savecontrol(last, &opts, ip, n); 193 if (sbappendaddr(&last->inp_socket->so_rcv, 194 sintosa(&ripsrc), n, opts) == 0) { 195 /* should notify about lost packet */ 196 m_freem(n); 197 if (opts) 198 m_freem(opts); 199 } else 200 sorwakeup(last->inp_socket); 201 opts = NULL; 202 } 203 } 204 last = inp; 205 } 206 #if defined(IPSEC) || defined(FAST_IPSEC) 207 /* check AH/ESP integrity. */ 208 if (last && ipsec4_in_reject_so(m, last->inp_socket)) { 209 m_freem(m); 210 ipsecstat.in_polvio++; 211 ipstat.ips_delivered--; 212 /* do not inject data to pcb */ 213 } else 214 #endif /*IPSEC*/ 215 if (last) { 216 if (last->inp_flags & INP_CONTROLOPTS || 217 last->inp_socket->so_options & SO_TIMESTAMP) 218 ip_savecontrol(last, &opts, ip, m); 219 if (sbappendaddr(&last->inp_socket->so_rcv, 220 sintosa(&ripsrc), m, opts) == 0) { 221 m_freem(m); 222 if (opts) 223 m_freem(opts); 224 } else 225 sorwakeup(last->inp_socket); 226 } else { 227 if (inetsw[ip_protox[ip->ip_p]].pr_input == rip_input) { 228 icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_PROTOCOL, 229 0, 0); 230 ipstat.ips_noproto++; 231 ipstat.ips_delivered--; 232 } else 233 m_freem(m); 234 } 235 return; 236 } 237 238 int 239 rip_pcbnotify(struct inpcbtable *table, 240 struct in_addr faddr, struct in_addr laddr, int proto, int errno, 241 void (*notify)(struct inpcb *, int)) 242 { 243 struct inpcb *inp, *ninp; 244 int nmatch; 245 246 nmatch = 0; 247 for (inp = (struct inpcb *)CIRCLEQ_FIRST(&table->inpt_queue); 248 inp != (struct inpcb *)&table->inpt_queue; 249 inp = ninp) { 250 ninp = (struct inpcb *)inp->inp_queue.cqe_next; 251 if (inp->inp_af != AF_INET) 252 continue; 253 if (inp->inp_ip.ip_p && inp->inp_ip.ip_p != proto) 254 continue; 255 if (in_hosteq(inp->inp_faddr, faddr) && 256 in_hosteq(inp->inp_laddr, laddr)) { 257 (*notify)(inp, errno); 258 nmatch++; 259 } 260 } 261 262 return nmatch; 263 } 264 265 void * 266 rip_ctlinput(int cmd, struct sockaddr *sa, void *v) 267 { 268 struct ip *ip = v; 269 void (*notify)(struct inpcb *, int) = in_rtchange; 270 int errno; 271 272 if (sa->sa_family != AF_INET || 273 sa->sa_len != sizeof(struct sockaddr_in)) 274 return NULL; 275 if ((unsigned)cmd >= PRC_NCMDS) 276 return NULL; 277 errno = inetctlerrmap[cmd]; 278 if (PRC_IS_REDIRECT(cmd)) 279 notify = in_rtchange, ip = 0; 280 else if (cmd == PRC_HOSTDEAD) 281 ip = 0; 282 else if (errno == 0) 283 return NULL; 284 if (ip) { 285 rip_pcbnotify(&rawcbtable, satosin(sa)->sin_addr, 286 ip->ip_src, ip->ip_p, errno, notify); 287 288 /* XXX mapped address case */ 289 } else 290 in_pcbnotifyall(&rawcbtable, satosin(sa)->sin_addr, errno, 291 notify); 292 return NULL; 293 } 294 295 /* 296 * Generate IP header and pass packet to ip_output. 297 * Tack on options user may have setup with control call. 298 */ 299 int 300 rip_output(struct mbuf *m, ...) 301 { 302 struct inpcb *inp; 303 struct ip *ip; 304 struct mbuf *opts; 305 int flags; 306 va_list ap; 307 308 va_start(ap, m); 309 inp = va_arg(ap, struct inpcb *); 310 va_end(ap); 311 312 flags = 313 (inp->inp_socket->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST 314 | IP_RETURNMTU; 315 316 /* 317 * If the user handed us a complete IP packet, use it. 318 * Otherwise, allocate an mbuf for a header and fill it in. 319 */ 320 if ((inp->inp_flags & INP_HDRINCL) == 0) { 321 if ((m->m_pkthdr.len + sizeof(struct ip)) > IP_MAXPACKET) { 322 m_freem(m); 323 return (EMSGSIZE); 324 } 325 M_PREPEND(m, sizeof(struct ip), M_DONTWAIT); 326 if (!m) 327 return (ENOBUFS); 328 ip = mtod(m, struct ip *); 329 ip->ip_tos = 0; 330 ip->ip_off = htons(0); 331 ip->ip_p = inp->inp_ip.ip_p; 332 ip->ip_len = htons(m->m_pkthdr.len); 333 ip->ip_src = inp->inp_laddr; 334 ip->ip_dst = inp->inp_faddr; 335 ip->ip_ttl = MAXTTL; 336 opts = inp->inp_options; 337 } else { 338 if (m->m_pkthdr.len > IP_MAXPACKET) { 339 m_freem(m); 340 return (EMSGSIZE); 341 } 342 ip = mtod(m, struct ip *); 343 344 /* 345 * If the mbuf is read-only, we need to allocate 346 * a new mbuf for the header, since we need to 347 * modify the header. 348 */ 349 if (M_READONLY(m)) { 350 int hlen = ip->ip_hl << 2; 351 352 m = m_copyup(m, hlen, (max_linkhdr + 3) & ~3); 353 if (m == NULL) 354 return (ENOMEM); /* XXX */ 355 ip = mtod(m, struct ip *); 356 } 357 358 /* XXX userland passes ip_len and ip_off in host order */ 359 if (m->m_pkthdr.len != ip->ip_len) { 360 m_freem(m); 361 return (EINVAL); 362 } 363 HTONS(ip->ip_len); 364 HTONS(ip->ip_off); 365 if (ip->ip_id == 0) 366 ip->ip_id = ip_newid(); 367 opts = NULL; 368 /* XXX prevent ip_output from overwriting header fields */ 369 flags |= IP_RAWOUTPUT; 370 ipstat.ips_rawout++; 371 } 372 return (ip_output(m, opts, &inp->inp_route, flags, inp->inp_moptions, 373 inp->inp_socket, &inp->inp_errormtu)); 374 } 375 376 /* 377 * Raw IP socket option processing. 378 */ 379 int 380 rip_ctloutput(int op, struct socket *so, int level, int optname, 381 struct mbuf **m) 382 { 383 struct inpcb *inp = sotoinpcb(so); 384 int error = 0; 385 386 if (level != IPPROTO_IP) { 387 error = ENOPROTOOPT; 388 if (op == PRCO_SETOPT && *m != 0) 389 (void) m_free(*m); 390 } else switch (op) { 391 392 case PRCO_SETOPT: 393 switch (optname) { 394 case IP_HDRINCL: 395 if (*m == 0 || (*m)->m_len < sizeof (int)) 396 error = EINVAL; 397 else { 398 if (*mtod(*m, int *)) 399 inp->inp_flags |= INP_HDRINCL; 400 else 401 inp->inp_flags &= ~INP_HDRINCL; 402 } 403 if (*m != 0) 404 (void) m_free(*m); 405 break; 406 407 #ifdef MROUTING 408 case MRT_INIT: 409 case MRT_DONE: 410 case MRT_ADD_VIF: 411 case MRT_DEL_VIF: 412 case MRT_ADD_MFC: 413 case MRT_DEL_MFC: 414 case MRT_ASSERT: 415 case MRT_API_CONFIG: 416 case MRT_ADD_BW_UPCALL: 417 case MRT_DEL_BW_UPCALL: 418 error = ip_mrouter_set(so, optname, m); 419 break; 420 #endif 421 422 default: 423 error = ip_ctloutput(op, so, level, optname, m); 424 break; 425 } 426 break; 427 428 case PRCO_GETOPT: 429 switch (optname) { 430 case IP_HDRINCL: 431 *m = m_get(M_WAIT, MT_SOOPTS); 432 MCLAIM((*m), so->so_mowner); 433 (*m)->m_len = sizeof (int); 434 *mtod(*m, int *) = inp->inp_flags & INP_HDRINCL ? 1 : 0; 435 break; 436 437 #ifdef MROUTING 438 case MRT_VERSION: 439 case MRT_ASSERT: 440 case MRT_API_SUPPORT: 441 case MRT_API_CONFIG: 442 error = ip_mrouter_get(so, optname, m); 443 break; 444 #endif 445 446 default: 447 error = ip_ctloutput(op, so, level, optname, m); 448 break; 449 } 450 break; 451 } 452 return (error); 453 } 454 455 int 456 rip_bind(struct inpcb *inp, struct mbuf *nam) 457 { 458 struct sockaddr_in *addr = mtod(nam, struct sockaddr_in *); 459 460 if (nam->m_len != sizeof(*addr)) 461 return (EINVAL); 462 if (TAILQ_FIRST(&ifnet) == 0) 463 return (EADDRNOTAVAIL); 464 if (addr->sin_family != AF_INET && 465 addr->sin_family != AF_IMPLINK) 466 return (EAFNOSUPPORT); 467 if (!in_nullhost(addr->sin_addr) && 468 ifa_ifwithaddr(sintosa(addr)) == 0) 469 return (EADDRNOTAVAIL); 470 inp->inp_laddr = addr->sin_addr; 471 return (0); 472 } 473 474 int 475 rip_connect(struct inpcb *inp, struct mbuf *nam) 476 { 477 struct sockaddr_in *addr = mtod(nam, struct sockaddr_in *); 478 479 if (nam->m_len != sizeof(*addr)) 480 return (EINVAL); 481 if (TAILQ_FIRST(&ifnet) == 0) 482 return (EADDRNOTAVAIL); 483 if (addr->sin_family != AF_INET && 484 addr->sin_family != AF_IMPLINK) 485 return (EAFNOSUPPORT); 486 inp->inp_faddr = addr->sin_addr; 487 return (0); 488 } 489 490 void 491 rip_disconnect(struct inpcb *inp) 492 { 493 494 inp->inp_faddr = zeroin_addr; 495 } 496 497 u_long rip_sendspace = RIPSNDQ; 498 u_long rip_recvspace = RIPRCVQ; 499 500 /*ARGSUSED*/ 501 int 502 rip_usrreq(struct socket *so, int req, 503 struct mbuf *m, struct mbuf *nam, struct mbuf *control, struct proc *p) 504 { 505 struct inpcb *inp; 506 int s; 507 int error = 0; 508 #ifdef MROUTING 509 extern struct socket *ip_mrouter; 510 #endif 511 512 if (req == PRU_CONTROL) 513 return (in_control(so, (long)m, (caddr_t)nam, 514 (struct ifnet *)control, p)); 515 516 if (req == PRU_PURGEIF) { 517 in_pcbpurgeif0(&rawcbtable, (struct ifnet *)control); 518 in_purgeif((struct ifnet *)control); 519 in_pcbpurgeif(&rawcbtable, (struct ifnet *)control); 520 return (0); 521 } 522 523 s = splsoftnet(); 524 inp = sotoinpcb(so); 525 #ifdef DIAGNOSTIC 526 if (req != PRU_SEND && req != PRU_SENDOOB && control) 527 panic("rip_usrreq: unexpected control mbuf"); 528 #endif 529 if (inp == 0 && req != PRU_ATTACH) { 530 error = EINVAL; 531 goto release; 532 } 533 534 switch (req) { 535 536 case PRU_ATTACH: 537 if (inp != 0) { 538 error = EISCONN; 539 break; 540 } 541 if (p == 0 || (error = suser(p->p_ucred, &p->p_acflag))) { 542 error = EACCES; 543 break; 544 } 545 if (so->so_snd.sb_hiwat == 0 || so->so_rcv.sb_hiwat == 0) { 546 error = soreserve(so, rip_sendspace, rip_recvspace); 547 if (error) 548 break; 549 } 550 error = in_pcballoc(so, &rawcbtable); 551 if (error) 552 break; 553 inp = sotoinpcb(so); 554 inp->inp_ip.ip_p = (long)nam; 555 break; 556 557 case PRU_DETACH: 558 #ifdef MROUTING 559 if (so == ip_mrouter) 560 ip_mrouter_done(); 561 #endif 562 in_pcbdetach(inp); 563 break; 564 565 case PRU_BIND: 566 error = rip_bind(inp, nam); 567 break; 568 569 case PRU_LISTEN: 570 error = EOPNOTSUPP; 571 break; 572 573 case PRU_CONNECT: 574 error = rip_connect(inp, nam); 575 if (error) 576 break; 577 soisconnected(so); 578 break; 579 580 case PRU_CONNECT2: 581 error = EOPNOTSUPP; 582 break; 583 584 case PRU_DISCONNECT: 585 soisdisconnected(so); 586 rip_disconnect(inp); 587 break; 588 589 /* 590 * Mark the connection as being incapable of further input. 591 */ 592 case PRU_SHUTDOWN: 593 socantsendmore(so); 594 break; 595 596 case PRU_RCVD: 597 error = EOPNOTSUPP; 598 break; 599 600 /* 601 * Ship a packet out. The appropriate raw output 602 * routine handles any massaging necessary. 603 */ 604 case PRU_SEND: 605 if (control && control->m_len) { 606 m_freem(control); 607 m_freem(m); 608 error = EINVAL; 609 break; 610 } 611 { 612 if (nam) { 613 if ((so->so_state & SS_ISCONNECTED) != 0) { 614 error = EISCONN; 615 goto die; 616 } 617 error = rip_connect(inp, nam); 618 if (error) { 619 die: 620 m_freem(m); 621 break; 622 } 623 } else { 624 if ((so->so_state & SS_ISCONNECTED) == 0) { 625 error = ENOTCONN; 626 goto die; 627 } 628 } 629 error = rip_output(m, inp); 630 if (nam) 631 rip_disconnect(inp); 632 } 633 break; 634 635 case PRU_SENSE: 636 /* 637 * stat: don't bother with a blocksize. 638 */ 639 splx(s); 640 return (0); 641 642 case PRU_RCVOOB: 643 error = EOPNOTSUPP; 644 break; 645 646 case PRU_SENDOOB: 647 m_freem(control); 648 m_freem(m); 649 error = EOPNOTSUPP; 650 break; 651 652 case PRU_SOCKADDR: 653 in_setsockaddr(inp, nam); 654 break; 655 656 case PRU_PEERADDR: 657 in_setpeeraddr(inp, nam); 658 break; 659 660 default: 661 panic("rip_usrreq"); 662 } 663 664 release: 665 splx(s); 666 return (error); 667 } 668