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. Neither the name of the University 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 REGENTS 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 REGENTS 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 * @(#)raw_ip.c 8.7 (Berkeley) 5/15/95 30 * $FreeBSD: src/sys/netinet/raw_ip.c,v 1.64.2.16 2003/08/24 08:24:38 hsu Exp $ 31 */ 32 33 #include "opt_inet6.h" 34 #include "opt_carp.h" 35 36 #include <sys/param.h> 37 #include <sys/systm.h> 38 #include <sys/kernel.h> 39 #include <sys/jail.h> 40 #include <sys/malloc.h> 41 #include <sys/mbuf.h> 42 #include <sys/proc.h> 43 #include <sys/priv.h> 44 #include <sys/protosw.h> 45 #include <sys/socket.h> 46 #include <sys/socketvar.h> 47 #include <sys/sysctl.h> 48 49 #include <sys/socketvar2.h> 50 #include <sys/msgport2.h> 51 52 #include <machine/stdarg.h> 53 54 #include <net/if.h> 55 #ifdef CARP 56 #include <net/if_types.h> 57 #endif 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/ipfw3/ip_fw.h> 71 #include <net/dummynet/ip_dummynet.h> 72 #include <net/dummynet3/ip_dummynet3.h> 73 74 struct inpcbinfo ripcbinfo; 75 struct inpcbportinfo ripcbportinfo; 76 77 /* 78 * hooks for multicast routing. They all default to NULL, 79 * so leave them not initialized and rely on BSS being set to 0. 80 */ 81 82 /* The socket used to communicate with the multicast routing daemon. */ 83 struct socket *ip_mrouter; 84 85 /* The various mrouter and rsvp functions */ 86 int (*ip_mrouter_set)(struct socket *, struct sockopt *); 87 int (*ip_mrouter_get)(struct socket *, struct sockopt *); 88 int (*ip_mrouter_done)(void); 89 int (*ip_mforward)(struct ip *, struct ifnet *, struct mbuf *, 90 struct ip_moptions *); 91 int (*mrt_ioctl)(u_long, caddr_t); 92 int (*legal_vif_num)(int); 93 u_long (*ip_mcast_src)(int); 94 95 int (*rsvp_input_p)(struct mbuf **, int *, int); 96 int (*ip_rsvp_vif)(struct socket *, struct sockopt *); 97 void (*ip_rsvp_force_done)(struct socket *); 98 99 /* 100 * Nominal space allocated to a raw ip socket. 101 */ 102 #define RIPSNDQ 8192 103 #define RIPRCVQ 8192 104 105 /* 106 * Raw interface to IP protocol. 107 */ 108 109 /* 110 * Initialize raw connection block queue. 111 */ 112 void 113 rip_init(void) 114 { 115 in_pcbinfo_init(&ripcbinfo, 0, FALSE); 116 in_pcbportinfo_init(&ripcbportinfo, 1, 0); 117 /* 118 * XXX We don't use the hash list for raw IP, but it's easier 119 * to allocate a one entry hash list than it is to check all 120 * over the place for hashbase == NULL. 121 */ 122 ripcbinfo.hashbase = hashinit(1, M_PCB, &ripcbinfo.hashmask); 123 in_pcbportinfo_set(&ripcbinfo, &ripcbportinfo, 1); 124 ripcbinfo.wildcardhashbase = hashinit(1, M_PCB, 125 &ripcbinfo.wildcardhashmask); 126 ripcbinfo.ipi_size = sizeof(struct inpcb); 127 } 128 129 /* 130 * Setup generic address and protocol structures 131 * for raw_input routine, then pass them along with 132 * mbuf chain. 133 */ 134 int 135 rip_input(struct mbuf **mp, int *offp, int proto) 136 { 137 struct sockaddr_in ripsrc = { sizeof ripsrc, AF_INET }; 138 struct mbuf *m = *mp; 139 struct ip *ip = mtod(m, struct ip *); 140 struct inpcb *inp; 141 struct inpcb *last = NULL; 142 struct mbuf *opts = NULL; 143 144 ASSERT_NETISR0; 145 146 *mp = NULL; 147 148 ripsrc.sin_addr = ip->ip_src; 149 LIST_FOREACH(inp, &ripcbinfo.pcblisthead, inp_list) { 150 if (inp->inp_flags & INP_PLACEMARKER) 151 continue; 152 #ifdef INET6 153 if (!INP_ISIPV4(inp)) 154 continue; 155 #endif 156 if (inp->inp_ip_p && inp->inp_ip_p != proto) 157 continue; 158 if (inp->inp_laddr.s_addr != INADDR_ANY && 159 inp->inp_laddr.s_addr != ip->ip_dst.s_addr) 160 continue; 161 if (inp->inp_faddr.s_addr != INADDR_ANY && 162 inp->inp_faddr.s_addr != ip->ip_src.s_addr) 163 continue; 164 if (last) { 165 struct mbuf *n = m_copypacket(m, M_NOWAIT); 166 167 if (n) { 168 lwkt_gettoken(&last->inp_socket->so_rcv.ssb_token); 169 if (last->inp_flags & INP_CONTROLOPTS || 170 last->inp_socket->so_options & SO_TIMESTAMP) 171 ip_savecontrol(last, &opts, ip, n); 172 if (ssb_appendaddr(&last->inp_socket->so_rcv, 173 (struct sockaddr *)&ripsrc, n, 174 opts) == 0) { 175 m_freem(n); 176 if (opts) 177 m_freem(opts); 178 soroverflow(last->inp_socket); 179 } else { 180 sorwakeup(last->inp_socket); 181 } 182 lwkt_reltoken(&last->inp_socket->so_rcv.ssb_token); 183 opts = NULL; 184 } 185 } 186 last = inp; 187 } 188 /* Check the minimum TTL for socket. */ 189 if (last && ip->ip_ttl < last->inp_ip_minttl) { 190 m_freem(opts); 191 ipstat.ips_delivered--; 192 } else if (last) { 193 if (last->inp_flags & INP_CONTROLOPTS || 194 last->inp_socket->so_options & SO_TIMESTAMP) 195 ip_savecontrol(last, &opts, ip, m); 196 lwkt_gettoken(&last->inp_socket->so_rcv.ssb_token); 197 if (ssb_appendaddr(&last->inp_socket->so_rcv, 198 (struct sockaddr *)&ripsrc, m, opts) == 0) { 199 m_freem(m); 200 if (opts) 201 m_freem(opts); 202 soroverflow(last->inp_socket); 203 } else { 204 sorwakeup(last->inp_socket); 205 } 206 lwkt_reltoken(&last->inp_socket->so_rcv.ssb_token); 207 } else { 208 m_freem(m); 209 ipstat.ips_noproto++; 210 ipstat.ips_delivered--; 211 } 212 return(IPPROTO_DONE); 213 } 214 215 /* 216 * Generate IP header and pass packet to ip_output. 217 * Tack on options user may have setup with control call. 218 */ 219 int 220 rip_output(struct mbuf *m, struct socket *so, ...) 221 { 222 struct ip *ip; 223 struct inpcb *inp = so->so_pcb; 224 __va_list ap; 225 int flags = (so->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST; 226 u_long dst; 227 228 ASSERT_NETISR0; 229 230 __va_start(ap, so); 231 dst = __va_arg(ap, u_long); 232 __va_end(ap); 233 234 /* 235 * If the user handed us a complete IP packet, use it. 236 * Otherwise, allocate an mbuf for a header and fill it in. 237 */ 238 if ((inp->inp_flags & INP_HDRINCL) == 0) { 239 if (m->m_pkthdr.len + sizeof(struct ip) > IP_MAXPACKET) { 240 m_freem(m); 241 return(EMSGSIZE); 242 } 243 M_PREPEND(m, sizeof(struct ip), M_WAITOK); 244 if (m == NULL) 245 return(ENOBUFS); 246 ip = mtod(m, struct ip *); 247 ip->ip_tos = inp->inp_ip_tos; 248 ip->ip_off = 0; 249 ip->ip_p = inp->inp_ip_p; 250 ip->ip_len = m->m_pkthdr.len; 251 ip->ip_src = inp->inp_laddr; 252 ip->ip_dst.s_addr = dst; 253 ip->ip_ttl = inp->inp_ip_ttl; 254 } else { 255 int hlen; 256 257 if (m->m_pkthdr.len > IP_MAXPACKET) { 258 m_freem(m); 259 return(EMSGSIZE); 260 } 261 if (m->m_len < sizeof(struct ip)) { 262 m = m_pullup(m, sizeof(struct ip)); 263 if (m == NULL) 264 return ENOBUFS; 265 } 266 ip = mtod(m, struct ip *); 267 hlen = IP_VHL_HL(ip->ip_vhl) << 2; 268 269 /* Don't allow header length less than the minimum. */ 270 if (hlen < sizeof(struct ip)) { 271 m_freem(m); 272 return EINVAL; 273 } 274 275 /* 276 * Don't allow both user specified and setsockopt options. 277 * Don't allow packet length sizes that will crash. 278 */ 279 if ((hlen != sizeof(struct ip) && inp->inp_options) || 280 ip->ip_len > m->m_pkthdr.len || ip->ip_len < hlen) { 281 m_freem(m); 282 return EINVAL; 283 } 284 if (ip->ip_id == 0) 285 ip->ip_id = ip_newid(); 286 287 /* Prevent ip_output from overwriting header fields */ 288 flags |= IP_RAWOUTPUT; 289 ipstat.ips_rawout++; 290 } 291 292 return ip_output(m, inp->inp_options, &inp->inp_route, flags, 293 inp->inp_moptions, inp); 294 } 295 296 /* 297 * Raw IP socket option processing. 298 */ 299 void 300 rip_ctloutput(netmsg_t msg) 301 { 302 struct socket *so = msg->base.nm_so; 303 struct sockopt *sopt = msg->ctloutput.nm_sopt; 304 struct inpcb *inp = so->so_pcb; 305 int error, optval; 306 307 ASSERT_NETISR0; 308 309 error = 0; 310 311 /* Get socket's owner cpuid hint */ 312 if (sopt->sopt_level == SOL_SOCKET && 313 sopt->sopt_dir == SOPT_GET && 314 sopt->sopt_name == SO_CPUHINT) { 315 optval = mycpuid; 316 soopt_from_kbuf(sopt, &optval, sizeof(optval)); 317 goto done; 318 } 319 320 if (sopt->sopt_level != IPPROTO_IP) { 321 error = EINVAL; 322 goto done; 323 } 324 325 switch (sopt->sopt_dir) { 326 case SOPT_GET: 327 switch (sopt->sopt_name) { 328 case IP_HDRINCL: 329 optval = inp->inp_flags & INP_HDRINCL; 330 soopt_from_kbuf(sopt, &optval, sizeof optval); 331 break; 332 333 case IP_FW_X: 334 error = ip_fw3_sockopt(sopt); 335 break; 336 337 case IP_FW_ADD: /* ADD actually returns the body... */ 338 case IP_FW_GET: 339 case IP_FW_TBL_GET: 340 case IP_FW_TBL_EXPIRE: /* returns # of expired addresses */ 341 error = ip_fw_sockopt(sopt); 342 break; 343 344 case IP_DUMMYNET_GET: 345 error = ip_dn_sockopt(sopt); 346 break ; 347 348 case MRT_INIT: 349 case MRT_DONE: 350 case MRT_ADD_VIF: 351 case MRT_DEL_VIF: 352 case MRT_ADD_MFC: 353 case MRT_DEL_MFC: 354 case MRT_VERSION: 355 case MRT_ASSERT: 356 case MRT_API_SUPPORT: 357 case MRT_API_CONFIG: 358 case MRT_ADD_BW_UPCALL: 359 case MRT_DEL_BW_UPCALL: 360 error = ip_mrouter_get ? ip_mrouter_get(so, sopt) : 361 EOPNOTSUPP; 362 break; 363 364 default: 365 ip_ctloutput(msg); 366 /* msg invalid now */ 367 return; 368 } 369 break; 370 371 case SOPT_SET: 372 switch (sopt->sopt_name) { 373 case IP_HDRINCL: 374 error = soopt_to_kbuf(sopt, &optval, sizeof optval, 375 sizeof optval); 376 if (error) 377 break; 378 if (optval) 379 inp->inp_flags |= INP_HDRINCL; 380 else 381 inp->inp_flags &= ~INP_HDRINCL; 382 break; 383 384 case IP_FW_X: 385 error = ip_fw3_sockopt(sopt); 386 break; 387 388 case IP_FW_ADD: 389 case IP_FW_DEL: 390 case IP_FW_FLUSH: 391 case IP_FW_ZERO: 392 case IP_FW_RESETLOG: 393 case IP_FW_TBL_CREATE: 394 case IP_FW_TBL_DESTROY: 395 case IP_FW_TBL_ADD: 396 case IP_FW_TBL_DEL: 397 case IP_FW_TBL_FLUSH: 398 case IP_FW_TBL_ZERO: 399 case IP_FW_TBL_EXPIRE: 400 error = ip_fw_sockopt(sopt); 401 break; 402 403 case IP_DUMMYNET_CONFIGURE: 404 case IP_DUMMYNET_DEL: 405 case IP_DUMMYNET_FLUSH: 406 error = ip_dn_sockopt(sopt); 407 break; 408 409 case IP_RSVP_ON: 410 error = ip_rsvp_init(so); 411 break; 412 413 case IP_RSVP_OFF: 414 error = ip_rsvp_done(); 415 break; 416 417 case IP_RSVP_VIF_ON: 418 case IP_RSVP_VIF_OFF: 419 error = ip_rsvp_vif ? 420 ip_rsvp_vif(so, sopt) : EINVAL; 421 break; 422 423 case MRT_INIT: 424 case MRT_DONE: 425 case MRT_ADD_VIF: 426 case MRT_DEL_VIF: 427 case MRT_ADD_MFC: 428 case MRT_DEL_MFC: 429 case MRT_VERSION: 430 case MRT_ASSERT: 431 case MRT_API_SUPPORT: 432 case MRT_API_CONFIG: 433 case MRT_ADD_BW_UPCALL: 434 case MRT_DEL_BW_UPCALL: 435 error = ip_mrouter_set ? ip_mrouter_set(so, sopt) : 436 EOPNOTSUPP; 437 break; 438 439 default: 440 ip_ctloutput(msg); 441 /* msg invalid now */ 442 return; 443 } 444 break; 445 } 446 done: 447 lwkt_replymsg(&msg->lmsg, error); 448 } 449 450 u_long rip_sendspace = RIPSNDQ; 451 u_long rip_recvspace = RIPRCVQ; 452 453 SYSCTL_INT(_net_inet_raw, OID_AUTO, maxdgram, CTLFLAG_RW, 454 &rip_sendspace, 0, "Maximum outgoing raw IP datagram size"); 455 SYSCTL_INT(_net_inet_raw, OID_AUTO, recvspace, CTLFLAG_RW, 456 &rip_recvspace, 0, "Maximum incoming raw IP datagram size"); 457 458 static void 459 rip_attach(netmsg_t msg) 460 { 461 struct socket *so = msg->base.nm_so; 462 int proto = msg->attach.nm_proto; 463 struct pru_attach_info *ai = msg->attach.nm_ai; 464 struct inpcb *inp; 465 int error; 466 467 ASSERT_NETISR0; 468 469 inp = so->so_pcb; 470 if (inp) 471 panic("rip_attach"); 472 error = priv_check_cred(ai->p_ucred, PRIV_NETINET_RAW, NULL_CRED_OKAY); 473 if (error) 474 goto done; 475 476 error = soreserve(so, rip_sendspace, rip_recvspace, ai->sb_rlimit); 477 if (error) 478 goto done; 479 480 error = in_pcballoc(so, &ripcbinfo); 481 if (error == 0) { 482 inp = (struct inpcb *)so->so_pcb; 483 inp->inp_ip_p = proto; 484 inp->inp_ip_ttl = ip_defttl; 485 } 486 done: 487 lwkt_replymsg(&msg->lmsg, error); 488 } 489 490 static void 491 rip_detach(netmsg_t msg) 492 { 493 struct socket *so = msg->base.nm_so; 494 struct inpcb *inp; 495 496 ASSERT_NETISR0; 497 498 inp = so->so_pcb; 499 if (inp == NULL) 500 panic("rip_detach"); 501 if (so == ip_mrouter && ip_mrouter_done) 502 ip_mrouter_done(); 503 if (ip_rsvp_force_done) 504 ip_rsvp_force_done(so); 505 if (so == ip_rsvpd) 506 ip_rsvp_done(); 507 in_pcbdetach(inp); 508 lwkt_replymsg(&msg->lmsg, 0); 509 } 510 511 static void 512 rip_abort(netmsg_t msg) 513 { 514 /* 515 * Raw socket does not support listen(2), 516 * so this should never be called. 517 */ 518 panic("rip_abort is called"); 519 } 520 521 static void 522 rip_disconnect(netmsg_t msg) 523 { 524 struct socket *so = msg->base.nm_so; 525 int error; 526 527 ASSERT_NETISR0; 528 529 if (so->so_state & SS_ISCONNECTED) { 530 soisdisconnected(so); 531 error = 0; 532 } else { 533 error = ENOTCONN; 534 } 535 lwkt_replymsg(&msg->lmsg, error); 536 } 537 538 static void 539 rip_bind(netmsg_t msg) 540 { 541 struct socket *so = msg->base.nm_so; 542 struct sockaddr *nam = msg->bind.nm_nam; 543 struct inpcb *inp = so->so_pcb; 544 struct sockaddr_in *addr = (struct sockaddr_in *)nam; 545 int error; 546 547 ASSERT_NETISR0; 548 549 if (nam->sa_len == sizeof(*addr)) { 550 if (ifnet_array_isempty() || 551 ((addr->sin_family != AF_INET) && 552 (addr->sin_family != AF_IMPLINK)) || 553 (addr->sin_addr.s_addr != INADDR_ANY && 554 ifa_ifwithaddr((struct sockaddr *)addr) == 0)) { 555 error = EADDRNOTAVAIL; 556 } else { 557 inp->inp_laddr = addr->sin_addr; 558 error = 0; 559 } 560 } else { 561 error = EINVAL; 562 } 563 lwkt_replymsg(&msg->lmsg, error); 564 } 565 566 static void 567 rip_connect(netmsg_t msg) 568 { 569 struct socket *so = msg->base.nm_so; 570 struct sockaddr *nam = msg->connect.nm_nam; 571 struct inpcb *inp = so->so_pcb; 572 struct sockaddr_in *addr = (struct sockaddr_in *)nam; 573 int error; 574 575 ASSERT_NETISR0; 576 577 if (nam->sa_len != sizeof(*addr)) { 578 error = EINVAL; 579 } else if (ifnet_array_isempty()) { 580 error = EADDRNOTAVAIL; 581 } else { 582 if ((addr->sin_family != AF_INET) && 583 (addr->sin_family != AF_IMPLINK)) { 584 error = EAFNOSUPPORT; 585 } else { 586 inp->inp_faddr = addr->sin_addr; 587 soisconnected(so); 588 error = 0; 589 } 590 } 591 lwkt_replymsg(&msg->lmsg, error); 592 } 593 594 static void 595 rip_shutdown(netmsg_t msg) 596 { 597 ASSERT_NETISR0; 598 599 socantsendmore(msg->base.nm_so); 600 lwkt_replymsg(&msg->lmsg, 0); 601 } 602 603 static void 604 rip_send(netmsg_t msg) 605 { 606 struct socket *so = msg->base.nm_so; 607 struct mbuf *m = msg->send.nm_m; 608 /*struct mbuf *control = msg->send.nm_control;*/ 609 struct sockaddr *nam = msg->send.nm_addr; 610 /*int flags = msg->send.nm_flags;*/ 611 struct inpcb *inp = so->so_pcb; 612 u_long dst; 613 int error; 614 615 ASSERT_NETISR0; 616 617 if (so->so_state & SS_ISCONNECTED) { 618 if (nam) { 619 m_freem(m); 620 error = EISCONN; 621 } else { 622 dst = inp->inp_faddr.s_addr; 623 error = rip_output(m, so, dst); 624 } 625 } else { 626 if (nam == NULL) { 627 m_freem(m); 628 error = ENOTCONN; 629 } else { 630 dst = ((struct sockaddr_in *)nam)->sin_addr.s_addr; 631 error = rip_output(m, so, dst); 632 } 633 } 634 lwkt_replymsg(&msg->lmsg, error); 635 } 636 637 SYSCTL_PROC(_net_inet_raw, OID_AUTO/*XXX*/, pcblist, CTLFLAG_RD, &ripcbinfo, 1, 638 in_pcblist_range, "S,xinpcb", "List of active raw IP sockets"); 639 640 struct pr_usrreqs rip_usrreqs = { 641 .pru_abort = rip_abort, 642 .pru_accept = pr_generic_notsupp, 643 .pru_attach = rip_attach, 644 .pru_bind = rip_bind, 645 .pru_connect = rip_connect, 646 .pru_connect2 = pr_generic_notsupp, 647 .pru_control = in_control_dispatch, 648 .pru_detach = rip_detach, 649 .pru_disconnect = rip_disconnect, 650 .pru_listen = pr_generic_notsupp, 651 .pru_peeraddr = in_setpeeraddr_dispatch, 652 .pru_rcvd = pr_generic_notsupp, 653 .pru_rcvoob = pr_generic_notsupp, 654 .pru_send = rip_send, 655 .pru_sense = pru_sense_null, 656 .pru_shutdown = rip_shutdown, 657 .pru_sockaddr = in_setsockaddr_dispatch, 658 .pru_sosend = sosend, 659 .pru_soreceive = soreceive 660 }; 661