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/caps.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 = htons(m->m_pkthdr.len); /* incls header now */ 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 int ip_len; 257 258 if (m->m_pkthdr.len > IP_MAXPACKET) { 259 m_freem(m); 260 return(EMSGSIZE); 261 } 262 if (m->m_len < sizeof(struct ip)) { 263 m = m_pullup(m, sizeof(struct ip)); 264 if (m == NULL) 265 return ENOBUFS; 266 } 267 ip = mtod(m, struct ip *); 268 hlen = IP_VHL_HL(ip->ip_vhl) << 2; 269 ip_len = ntohs(ip->ip_len); /* includes header */ 270 271 /* 272 * Don't allow both user specified and setsockopt options. 273 * Don't allow packet length sizes that will crash. 274 */ 275 if (hlen < sizeof(struct ip) || 276 (hlen != sizeof(struct ip) && inp->inp_options) || 277 ip_len > m->m_pkthdr.len || 278 ip_len < hlen) 279 { 280 m_freem(m); 281 return EINVAL; 282 } 283 284 /* 285 * System supplies ip_id if passed as 0. 286 */ 287 if (ip->ip_id == 0) 288 ip->ip_id = ip_newid(); 289 290 /* Prevent ip_output from overwriting header fields */ 291 flags |= IP_RAWOUTPUT; 292 ipstat.ips_rawout++; 293 } 294 295 return ip_output(m, inp->inp_options, &inp->inp_route, flags, 296 inp->inp_moptions, inp); 297 } 298 299 /* 300 * Raw IP socket option processing. 301 */ 302 void 303 rip_ctloutput(netmsg_t msg) 304 { 305 struct socket *so = msg->base.nm_so; 306 struct sockopt *sopt = msg->ctloutput.nm_sopt; 307 struct inpcb *inp = so->so_pcb; 308 int error, optval; 309 310 ASSERT_NETISR0; 311 312 error = 0; 313 314 /* Get socket's owner cpuid hint */ 315 if (sopt->sopt_level == SOL_SOCKET && 316 sopt->sopt_dir == SOPT_GET && 317 sopt->sopt_name == SO_CPUHINT) { 318 optval = mycpuid; 319 soopt_from_kbuf(sopt, &optval, sizeof(optval)); 320 goto done; 321 } 322 323 if (sopt->sopt_level != IPPROTO_IP) { 324 error = EINVAL; 325 goto done; 326 } 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 soopt_from_kbuf(sopt, &optval, sizeof optval); 334 break; 335 336 case IP_FW_X: 337 error = ip_fw3_sockopt(sopt); 338 break; 339 340 case IP_FW_ADD: /* ADD actually returns the body... */ 341 case IP_FW_GET: 342 case IP_FW_TBL_GET: 343 case IP_FW_TBL_EXPIRE: /* returns # of expired addresses */ 344 error = ip_fw_sockopt(sopt); 345 break; 346 347 case IP_DUMMYNET_GET: 348 error = ip_dn_sockopt(sopt); 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 ip_ctloutput(msg); 369 /* msg invalid now */ 370 return; 371 } 372 break; 373 374 case SOPT_SET: 375 switch (sopt->sopt_name) { 376 case IP_HDRINCL: 377 error = soopt_to_kbuf(sopt, &optval, sizeof optval, 378 sizeof optval); 379 if (error) 380 break; 381 if (optval) 382 inp->inp_flags |= INP_HDRINCL; 383 else 384 inp->inp_flags &= ~INP_HDRINCL; 385 break; 386 387 case IP_FW_X: 388 error = ip_fw3_sockopt(sopt); 389 break; 390 391 case IP_FW_ADD: 392 case IP_FW_DEL: 393 case IP_FW_FLUSH: 394 case IP_FW_ZERO: 395 case IP_FW_RESETLOG: 396 case IP_FW_TBL_CREATE: 397 case IP_FW_TBL_DESTROY: 398 case IP_FW_TBL_ADD: 399 case IP_FW_TBL_DEL: 400 case IP_FW_TBL_FLUSH: 401 case IP_FW_TBL_ZERO: 402 case IP_FW_TBL_EXPIRE: 403 error = ip_fw_sockopt(sopt); 404 break; 405 406 case IP_DUMMYNET_CONFIGURE: 407 case IP_DUMMYNET_DEL: 408 case IP_DUMMYNET_FLUSH: 409 error = ip_dn_sockopt(sopt); 410 break; 411 412 case IP_RSVP_ON: 413 error = ip_rsvp_init(so); 414 break; 415 416 case IP_RSVP_OFF: 417 error = ip_rsvp_done(); 418 break; 419 420 case IP_RSVP_VIF_ON: 421 case IP_RSVP_VIF_OFF: 422 error = ip_rsvp_vif ? 423 ip_rsvp_vif(so, sopt) : EINVAL; 424 break; 425 426 case MRT_INIT: 427 case MRT_DONE: 428 case MRT_ADD_VIF: 429 case MRT_DEL_VIF: 430 case MRT_ADD_MFC: 431 case MRT_DEL_MFC: 432 case MRT_VERSION: 433 case MRT_ASSERT: 434 case MRT_API_SUPPORT: 435 case MRT_API_CONFIG: 436 case MRT_ADD_BW_UPCALL: 437 case MRT_DEL_BW_UPCALL: 438 error = ip_mrouter_set ? ip_mrouter_set(so, sopt) : 439 EOPNOTSUPP; 440 break; 441 442 default: 443 ip_ctloutput(msg); 444 /* msg invalid now */ 445 return; 446 } 447 break; 448 } 449 done: 450 lwkt_replymsg(&msg->lmsg, error); 451 } 452 453 u_long rip_sendspace = RIPSNDQ; 454 u_long rip_recvspace = RIPRCVQ; 455 456 SYSCTL_INT(_net_inet_raw, OID_AUTO, maxdgram, CTLFLAG_RW, 457 &rip_sendspace, 0, "Maximum outgoing raw IP datagram size"); 458 SYSCTL_INT(_net_inet_raw, OID_AUTO, recvspace, CTLFLAG_RW, 459 &rip_recvspace, 0, "Maximum incoming raw IP datagram size"); 460 461 static void 462 rip_attach(netmsg_t msg) 463 { 464 struct socket *so = msg->base.nm_so; 465 int proto = msg->attach.nm_proto; 466 struct pru_attach_info *ai = msg->attach.nm_ai; 467 struct inpcb *inp; 468 int error; 469 470 ASSERT_NETISR0; 471 472 inp = so->so_pcb; 473 if (inp) 474 panic("rip_attach"); 475 error = caps_priv_check(ai->p_ucred, SYSCAP_NONET_RAW | 476 __SYSCAP_NULLCRED); 477 if (error) 478 goto done; 479 480 error = soreserve(so, rip_sendspace, rip_recvspace, ai->sb_rlimit); 481 if (error) 482 goto done; 483 484 error = in_pcballoc(so, &ripcbinfo); 485 if (error == 0) { 486 inp = (struct inpcb *)so->so_pcb; 487 inp->inp_ip_p = proto; 488 inp->inp_ip_ttl = ip_defttl; 489 } 490 done: 491 lwkt_replymsg(&msg->lmsg, error); 492 } 493 494 static void 495 rip_detach(netmsg_t msg) 496 { 497 struct socket *so = msg->base.nm_so; 498 struct inpcb *inp; 499 500 ASSERT_NETISR0; 501 502 inp = so->so_pcb; 503 if (inp == NULL) 504 panic("rip_detach"); 505 if (so == ip_mrouter && ip_mrouter_done) 506 ip_mrouter_done(); 507 if (ip_rsvp_force_done) 508 ip_rsvp_force_done(so); 509 if (so == ip_rsvpd) 510 ip_rsvp_done(); 511 in_pcbdetach(inp); 512 lwkt_replymsg(&msg->lmsg, 0); 513 } 514 515 static void 516 rip_abort(netmsg_t msg) 517 { 518 /* 519 * Raw socket does not support listen(2), 520 * so this should never be called. 521 */ 522 panic("rip_abort is called"); 523 } 524 525 static void 526 rip_disconnect(netmsg_t msg) 527 { 528 struct socket *so = msg->base.nm_so; 529 int error; 530 531 ASSERT_NETISR0; 532 533 if (so->so_state & SS_ISCONNECTED) { 534 soisdisconnected(so); 535 error = 0; 536 } else { 537 error = ENOTCONN; 538 } 539 lwkt_replymsg(&msg->lmsg, error); 540 } 541 542 static void 543 rip_bind(netmsg_t msg) 544 { 545 struct socket *so = msg->base.nm_so; 546 struct sockaddr *nam = msg->bind.nm_nam; 547 struct inpcb *inp = so->so_pcb; 548 struct sockaddr_in *addr = (struct sockaddr_in *)nam; 549 int error; 550 551 ASSERT_NETISR0; 552 553 if (nam->sa_len == sizeof(*addr)) { 554 if (ifnet_array_isempty() || 555 ((addr->sin_family != AF_INET) && 556 (addr->sin_family != AF_IMPLINK)) || 557 (addr->sin_addr.s_addr != INADDR_ANY && 558 ifa_ifwithaddr((struct sockaddr *)addr) == 0)) { 559 error = EADDRNOTAVAIL; 560 } else { 561 inp->inp_laddr = addr->sin_addr; 562 error = 0; 563 } 564 } else { 565 error = EINVAL; 566 } 567 lwkt_replymsg(&msg->lmsg, error); 568 } 569 570 static void 571 rip_connect(netmsg_t msg) 572 { 573 struct socket *so = msg->base.nm_so; 574 struct sockaddr *nam = msg->connect.nm_nam; 575 struct inpcb *inp = so->so_pcb; 576 struct sockaddr_in *addr = (struct sockaddr_in *)nam; 577 int error; 578 579 ASSERT_NETISR0; 580 581 if (nam->sa_len != sizeof(*addr)) { 582 error = EINVAL; 583 } else if (ifnet_array_isempty()) { 584 error = EADDRNOTAVAIL; 585 } else { 586 if ((addr->sin_family != AF_INET) && 587 (addr->sin_family != AF_IMPLINK)) { 588 error = EAFNOSUPPORT; 589 } else { 590 inp->inp_faddr = addr->sin_addr; 591 soisconnected(so); 592 error = 0; 593 } 594 } 595 lwkt_replymsg(&msg->lmsg, error); 596 } 597 598 static void 599 rip_shutdown(netmsg_t msg) 600 { 601 ASSERT_NETISR0; 602 603 socantsendmore(msg->base.nm_so); 604 lwkt_replymsg(&msg->lmsg, 0); 605 } 606 607 static void 608 rip_send(netmsg_t msg) 609 { 610 struct socket *so = msg->base.nm_so; 611 struct mbuf *m = msg->send.nm_m; 612 /*struct mbuf *control = msg->send.nm_control;*/ 613 struct sockaddr *nam = msg->send.nm_addr; 614 /*int flags = msg->send.nm_flags;*/ 615 struct inpcb *inp = so->so_pcb; 616 u_long dst; 617 int error; 618 619 ASSERT_NETISR0; 620 621 if (so->so_state & SS_ISCONNECTED) { 622 if (nam) { 623 m_freem(m); 624 error = EISCONN; 625 } else { 626 dst = inp->inp_faddr.s_addr; 627 error = rip_output(m, so, dst); 628 } 629 } else { 630 if (nam == NULL) { 631 m_freem(m); 632 error = ENOTCONN; 633 } else { 634 dst = ((struct sockaddr_in *)nam)->sin_addr.s_addr; 635 error = rip_output(m, so, dst); 636 } 637 } 638 lwkt_replymsg(&msg->lmsg, error); 639 } 640 641 SYSCTL_PROC(_net_inet_raw, OID_AUTO/*XXX*/, pcblist, CTLFLAG_RD, &ripcbinfo, 1, 642 in_pcblist_range, "S,xinpcb", "List of active raw IP sockets"); 643 644 struct pr_usrreqs rip_usrreqs = { 645 .pru_abort = rip_abort, 646 .pru_accept = pr_generic_notsupp, 647 .pru_attach = rip_attach, 648 .pru_bind = rip_bind, 649 .pru_connect = rip_connect, 650 .pru_connect2 = pr_generic_notsupp, 651 .pru_control = in_control_dispatch, 652 .pru_detach = rip_detach, 653 .pru_disconnect = rip_disconnect, 654 .pru_listen = pr_generic_notsupp, 655 .pru_peeraddr = in_setpeeraddr_dispatch, 656 .pru_rcvd = pr_generic_notsupp, 657 .pru_rcvoob = pr_generic_notsupp, 658 .pru_send = rip_send, 659 .pru_sense = pru_sense_null, 660 .pru_shutdown = rip_shutdown, 661 .pru_sockaddr = in_setsockaddr_dispatch, 662 .pru_sosend = sosend, 663 .pru_soreceive = soreceive 664 }; 665