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 ip = mtod(m, struct ip *); 245 ip->ip_tos = inp->inp_ip_tos; 246 ip->ip_off = 0; 247 ip->ip_p = inp->inp_ip_p; 248 ip->ip_len = htons(m->m_pkthdr.len); /* incls header now */ 249 ip->ip_src = inp->inp_laddr; 250 ip->ip_dst.s_addr = dst; 251 ip->ip_ttl = inp->inp_ip_ttl; 252 } else { 253 int hlen; 254 int ip_len; 255 256 if (m->m_pkthdr.len > IP_MAXPACKET) { 257 m_freem(m); 258 return(EMSGSIZE); 259 } 260 if (m->m_len < sizeof(struct ip)) { 261 m = m_pullup(m, sizeof(struct ip)); 262 if (m == NULL) 263 return ENOBUFS; 264 } 265 ip = mtod(m, struct ip *); 266 hlen = IP_VHL_HL(ip->ip_vhl) << 2; 267 ip_len = ntohs(ip->ip_len); /* includes header */ 268 269 /* 270 * Don't allow both user specified and setsockopt options. 271 * Don't allow packet length sizes that will crash. 272 */ 273 if (hlen < sizeof(struct ip) || 274 (hlen != sizeof(struct ip) && inp->inp_options) || 275 ip_len > m->m_pkthdr.len || 276 ip_len < hlen) 277 { 278 m_freem(m); 279 return EINVAL; 280 } 281 282 /* 283 * System supplies ip_id if passed as 0. 284 */ 285 if (ip->ip_id == 0) 286 ip->ip_id = ip_newid(); 287 288 /* Prevent ip_output from overwriting header fields */ 289 flags |= IP_RAWOUTPUT; 290 ipstat.ips_rawout++; 291 } 292 293 return ip_output(m, inp->inp_options, &inp->inp_route, flags, 294 inp->inp_moptions, inp); 295 } 296 297 /* 298 * Raw IP socket option processing. 299 */ 300 void 301 rip_ctloutput(netmsg_t msg) 302 { 303 struct socket *so = msg->base.nm_so; 304 struct sockopt *sopt = msg->ctloutput.nm_sopt; 305 struct inpcb *inp = so->so_pcb; 306 int error, optval; 307 308 ASSERT_NETISR0; 309 310 error = 0; 311 312 /* Get socket's owner cpuid hint */ 313 if (sopt->sopt_level == SOL_SOCKET && 314 sopt->sopt_dir == SOPT_GET && 315 sopt->sopt_name == SO_CPUHINT) { 316 optval = mycpuid; 317 soopt_from_kbuf(sopt, &optval, sizeof(optval)); 318 goto done; 319 } 320 321 if (sopt->sopt_level != IPPROTO_IP) { 322 error = EINVAL; 323 goto done; 324 } 325 326 switch (sopt->sopt_dir) { 327 case SOPT_GET: 328 switch (sopt->sopt_name) { 329 case IP_HDRINCL: 330 optval = inp->inp_flags & INP_HDRINCL; 331 soopt_from_kbuf(sopt, &optval, sizeof optval); 332 break; 333 334 case IP_FW_X: 335 error = ip_fw3_sockopt(sopt); 336 break; 337 338 case IP_FW_ADD: /* ADD actually returns the body... */ 339 case IP_FW_GET: 340 case IP_FW_TBL_GET: 341 case IP_FW_TBL_EXPIRE: /* returns # of expired addresses */ 342 error = ip_fw_sockopt(sopt); 343 break; 344 345 case IP_DUMMYNET_GET: 346 error = ip_dn_sockopt(sopt); 347 break ; 348 349 case MRT_INIT: 350 case MRT_DONE: 351 case MRT_ADD_VIF: 352 case MRT_DEL_VIF: 353 case MRT_ADD_MFC: 354 case MRT_DEL_MFC: 355 case MRT_VERSION: 356 case MRT_ASSERT: 357 case MRT_API_SUPPORT: 358 case MRT_API_CONFIG: 359 case MRT_ADD_BW_UPCALL: 360 case MRT_DEL_BW_UPCALL: 361 error = ip_mrouter_get ? ip_mrouter_get(so, sopt) : 362 EOPNOTSUPP; 363 break; 364 365 default: 366 ip_ctloutput(msg); 367 /* msg invalid now */ 368 return; 369 } 370 break; 371 372 case SOPT_SET: 373 switch (sopt->sopt_name) { 374 case IP_HDRINCL: 375 error = soopt_to_kbuf(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_X: 386 error = ip_fw3_sockopt(sopt); 387 break; 388 389 case IP_FW_ADD: 390 case IP_FW_DEL: 391 case IP_FW_FLUSH: 392 case IP_FW_ZERO: 393 case IP_FW_RESETLOG: 394 case IP_FW_TBL_CREATE: 395 case IP_FW_TBL_DESTROY: 396 case IP_FW_TBL_ADD: 397 case IP_FW_TBL_DEL: 398 case IP_FW_TBL_FLUSH: 399 case IP_FW_TBL_ZERO: 400 case IP_FW_TBL_EXPIRE: 401 error = ip_fw_sockopt(sopt); 402 break; 403 404 case IP_DUMMYNET_CONFIGURE: 405 case IP_DUMMYNET_DEL: 406 case IP_DUMMYNET_FLUSH: 407 error = ip_dn_sockopt(sopt); 408 break; 409 410 case IP_RSVP_ON: 411 error = ip_rsvp_init(so); 412 break; 413 414 case IP_RSVP_OFF: 415 error = ip_rsvp_done(); 416 break; 417 418 case IP_RSVP_VIF_ON: 419 case IP_RSVP_VIF_OFF: 420 error = ip_rsvp_vif ? 421 ip_rsvp_vif(so, sopt) : EINVAL; 422 break; 423 424 case MRT_INIT: 425 case MRT_DONE: 426 case MRT_ADD_VIF: 427 case MRT_DEL_VIF: 428 case MRT_ADD_MFC: 429 case MRT_DEL_MFC: 430 case MRT_VERSION: 431 case MRT_ASSERT: 432 case MRT_API_SUPPORT: 433 case MRT_API_CONFIG: 434 case MRT_ADD_BW_UPCALL: 435 case MRT_DEL_BW_UPCALL: 436 error = ip_mrouter_set ? ip_mrouter_set(so, sopt) : 437 EOPNOTSUPP; 438 break; 439 440 default: 441 ip_ctloutput(msg); 442 /* msg invalid now */ 443 return; 444 } 445 break; 446 } 447 done: 448 lwkt_replymsg(&msg->lmsg, error); 449 } 450 451 u_long rip_sendspace = RIPSNDQ; 452 u_long rip_recvspace = RIPRCVQ; 453 454 SYSCTL_INT(_net_inet_raw, OID_AUTO, maxdgram, CTLFLAG_RW, 455 &rip_sendspace, 0, "Maximum outgoing raw IP datagram size"); 456 SYSCTL_INT(_net_inet_raw, OID_AUTO, recvspace, CTLFLAG_RW, 457 &rip_recvspace, 0, "Maximum incoming raw IP datagram size"); 458 459 static void 460 rip_attach(netmsg_t msg) 461 { 462 struct socket *so = msg->base.nm_so; 463 int proto = msg->attach.nm_proto; 464 struct pru_attach_info *ai = msg->attach.nm_ai; 465 struct inpcb *inp; 466 int error; 467 468 ASSERT_NETISR0; 469 470 inp = so->so_pcb; 471 if (inp) 472 panic("rip_attach"); 473 error = caps_priv_check(ai->p_ucred, SYSCAP_NONET_RAW | 474 __SYSCAP_NULLCRED); 475 if (error) 476 goto done; 477 478 error = soreserve(so, rip_sendspace, rip_recvspace, ai->sb_rlimit); 479 if (error) 480 goto done; 481 482 error = in_pcballoc(so, &ripcbinfo); 483 if (error == 0) { 484 inp = (struct inpcb *)so->so_pcb; 485 inp->inp_ip_p = proto; 486 inp->inp_ip_ttl = ip_defttl; 487 } 488 done: 489 lwkt_replymsg(&msg->lmsg, error); 490 } 491 492 static void 493 rip_detach(netmsg_t msg) 494 { 495 struct socket *so = msg->base.nm_so; 496 struct inpcb *inp; 497 498 ASSERT_NETISR0; 499 500 inp = so->so_pcb; 501 if (inp == NULL) 502 panic("rip_detach"); 503 if (so == ip_mrouter && ip_mrouter_done) 504 ip_mrouter_done(); 505 if (ip_rsvp_force_done) 506 ip_rsvp_force_done(so); 507 if (so == ip_rsvpd) 508 ip_rsvp_done(); 509 in_pcbdetach(inp); 510 lwkt_replymsg(&msg->lmsg, 0); 511 } 512 513 static void 514 rip_abort(netmsg_t msg) 515 { 516 /* 517 * Raw socket does not support listen(2), 518 * so this should never be called. 519 */ 520 panic("rip_abort is called"); 521 } 522 523 static void 524 rip_disconnect(netmsg_t msg) 525 { 526 struct socket *so = msg->base.nm_so; 527 int error; 528 529 ASSERT_NETISR0; 530 531 if (so->so_state & SS_ISCONNECTED) { 532 soisdisconnected(so); 533 error = 0; 534 } else { 535 error = ENOTCONN; 536 } 537 lwkt_replymsg(&msg->lmsg, error); 538 } 539 540 static void 541 rip_bind(netmsg_t msg) 542 { 543 struct socket *so = msg->base.nm_so; 544 struct sockaddr *nam = msg->bind.nm_nam; 545 struct inpcb *inp = so->so_pcb; 546 struct sockaddr_in *addr = (struct sockaddr_in *)nam; 547 int error; 548 549 ASSERT_NETISR0; 550 551 if (nam->sa_len == sizeof(*addr)) { 552 if (ifnet_array_isempty() || 553 ((addr->sin_family != AF_INET) && 554 (addr->sin_family != AF_IMPLINK)) || 555 (addr->sin_addr.s_addr != INADDR_ANY && 556 ifa_ifwithaddr((struct sockaddr *)addr) == 0)) { 557 error = EADDRNOTAVAIL; 558 } else { 559 inp->inp_laddr = addr->sin_addr; 560 error = 0; 561 } 562 } else { 563 error = EINVAL; 564 } 565 lwkt_replymsg(&msg->lmsg, error); 566 } 567 568 static void 569 rip_connect(netmsg_t msg) 570 { 571 struct socket *so = msg->base.nm_so; 572 struct sockaddr *nam = msg->connect.nm_nam; 573 struct inpcb *inp = so->so_pcb; 574 struct sockaddr_in *addr = (struct sockaddr_in *)nam; 575 int error; 576 577 ASSERT_NETISR0; 578 579 if (nam->sa_len != sizeof(*addr)) { 580 error = EINVAL; 581 } else if (ifnet_array_isempty()) { 582 error = EADDRNOTAVAIL; 583 } else { 584 if ((addr->sin_family != AF_INET) && 585 (addr->sin_family != AF_IMPLINK)) { 586 error = EAFNOSUPPORT; 587 } else { 588 inp->inp_faddr = addr->sin_addr; 589 soisconnected(so); 590 error = 0; 591 } 592 } 593 lwkt_replymsg(&msg->lmsg, error); 594 } 595 596 static void 597 rip_shutdown(netmsg_t msg) 598 { 599 ASSERT_NETISR0; 600 601 socantsendmore(msg->base.nm_so); 602 lwkt_replymsg(&msg->lmsg, 0); 603 } 604 605 static void 606 rip_send(netmsg_t msg) 607 { 608 struct socket *so = msg->base.nm_so; 609 struct mbuf *m = msg->send.nm_m; 610 /*struct mbuf *control = msg->send.nm_control;*/ 611 struct sockaddr *nam = msg->send.nm_addr; 612 /*int flags = msg->send.nm_flags;*/ 613 struct inpcb *inp = so->so_pcb; 614 u_long dst; 615 int error; 616 617 ASSERT_NETISR0; 618 619 if (so->so_state & SS_ISCONNECTED) { 620 if (nam) { 621 m_freem(m); 622 error = EISCONN; 623 } else { 624 dst = inp->inp_faddr.s_addr; 625 error = rip_output(m, so, dst); 626 } 627 } else { 628 if (nam == NULL) { 629 m_freem(m); 630 error = ENOTCONN; 631 } else { 632 dst = ((struct sockaddr_in *)nam)->sin_addr.s_addr; 633 error = rip_output(m, so, dst); 634 } 635 } 636 lwkt_replymsg(&msg->lmsg, error); 637 } 638 639 SYSCTL_PROC(_net_inet_raw, OID_AUTO/*XXX*/, pcblist, CTLFLAG_RD, &ripcbinfo, 1, 640 in_pcblist_range, "S,xinpcb", "List of active raw IP sockets"); 641 642 struct pr_usrreqs rip_usrreqs = { 643 .pru_abort = rip_abort, 644 .pru_accept = pr_generic_notsupp, 645 .pru_attach = rip_attach, 646 .pru_bind = rip_bind, 647 .pru_connect = rip_connect, 648 .pru_connect2 = pr_generic_notsupp, 649 .pru_control = in_control_dispatch, 650 .pru_detach = rip_detach, 651 .pru_disconnect = rip_disconnect, 652 .pru_listen = pr_generic_notsupp, 653 .pru_peeraddr = in_setpeeraddr_dispatch, 654 .pru_rcvd = pr_generic_notsupp, 655 .pru_rcvoob = pr_generic_notsupp, 656 .pru_send = rip_send, 657 .pru_sense = pru_sense_null, 658 .pru_shutdown = rip_shutdown, 659 .pru_sockaddr = in_setsockaddr_dispatch, 660 .pru_sosend = sosend, 661 .pru_soreceive = soreceive 662 }; 663