ip_input.c revision 1.347
1/* $NetBSD: ip_input.c,v 1.347 2016/12/12 03:55:57 ozaki-r 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) 1998 The NetBSD Foundation, Inc. 34 * All rights reserved. 35 * 36 * This code is derived from software contributed to The NetBSD Foundation 37 * by Public Access Networks Corporation ("Panix"). It was developed under 38 * contract to Panix by Eric Haszlakiewicz and Thor Lancelot Simon. 39 * 40 * Redistribution and use in source and binary forms, with or without 41 * modification, are permitted provided that the following conditions 42 * are met: 43 * 1. Redistributions of source code must retain the above copyright 44 * notice, this list of conditions and the following disclaimer. 45 * 2. Redistributions in binary form must reproduce the above copyright 46 * notice, this list of conditions and the following disclaimer in the 47 * documentation and/or other materials provided with the distribution. 48 * 49 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 50 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 51 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 52 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 53 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 54 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 55 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 56 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 57 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 58 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 59 * POSSIBILITY OF SUCH DAMAGE. 60 */ 61 62/* 63 * Copyright (c) 1982, 1986, 1988, 1993 64 * The Regents of the University of California. All rights reserved. 65 * 66 * Redistribution and use in source and binary forms, with or without 67 * modification, are permitted provided that the following conditions 68 * are met: 69 * 1. Redistributions of source code must retain the above copyright 70 * notice, this list of conditions and the following disclaimer. 71 * 2. Redistributions in binary form must reproduce the above copyright 72 * notice, this list of conditions and the following disclaimer in the 73 * documentation and/or other materials provided with the distribution. 74 * 3. Neither the name of the University nor the names of its contributors 75 * may be used to endorse or promote products derived from this software 76 * without specific prior written permission. 77 * 78 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 79 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 80 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 81 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 82 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 83 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 84 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 85 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 86 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 87 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 88 * SUCH DAMAGE. 89 * 90 * @(#)ip_input.c 8.2 (Berkeley) 1/4/94 91 */ 92 93#include <sys/cdefs.h> 94__KERNEL_RCSID(0, "$NetBSD: ip_input.c,v 1.347 2016/12/12 03:55:57 ozaki-r Exp $"); 95 96#ifdef _KERNEL_OPT 97#include "opt_inet.h" 98#include "opt_compat_netbsd.h" 99#include "opt_gateway.h" 100#include "opt_ipsec.h" 101#include "opt_mrouting.h" 102#include "opt_mbuftrace.h" 103#include "opt_inet_csum.h" 104#include "opt_net_mpsafe.h" 105#endif 106 107#include "arp.h" 108 109#include <sys/param.h> 110#include <sys/systm.h> 111#include <sys/cpu.h> 112#include <sys/mbuf.h> 113#include <sys/domain.h> 114#include <sys/protosw.h> 115#include <sys/socket.h> 116#include <sys/socketvar.h> 117#include <sys/errno.h> 118#include <sys/time.h> 119#include <sys/kernel.h> 120#include <sys/pool.h> 121#include <sys/sysctl.h> 122#include <sys/kauth.h> 123 124#include <net/if.h> 125#include <net/if_dl.h> 126#include <net/route.h> 127#include <net/pktqueue.h> 128#include <net/pfil.h> 129 130#include <netinet/in.h> 131#include <netinet/in_systm.h> 132#include <netinet/ip.h> 133#include <netinet/in_pcb.h> 134#include <netinet/in_proto.h> 135#include <netinet/in_var.h> 136#include <netinet/ip_var.h> 137#include <netinet/ip_private.h> 138#include <netinet/ip_icmp.h> 139/* just for gif_ttl */ 140#include <netinet/in_gif.h> 141#include "gif.h" 142#include <net/if_gre.h> 143#include "gre.h" 144 145#ifdef MROUTING 146#include <netinet/ip_mroute.h> 147#endif 148#include <netinet/portalgo.h> 149 150#ifdef IPSEC 151#include <netipsec/ipsec.h> 152#endif 153 154#ifndef IPFORWARDING 155#ifdef GATEWAY 156#define IPFORWARDING 1 /* forward IP packets not for us */ 157#else /* GATEWAY */ 158#define IPFORWARDING 0 /* don't forward IP packets not for us */ 159#endif /* GATEWAY */ 160#endif /* IPFORWARDING */ 161#ifndef IPSENDREDIRECTS 162#define IPSENDREDIRECTS 1 163#endif 164#ifndef IPFORWSRCRT 165#define IPFORWSRCRT 1 /* forward source-routed packets */ 166#endif 167#ifndef IPALLOWSRCRT 168#define IPALLOWSRCRT 1 /* allow source-routed packets */ 169#endif 170#ifndef IPMTUDISC 171#define IPMTUDISC 1 172#endif 173#ifndef IPMTUDISCTIMEOUT 174#define IPMTUDISCTIMEOUT (10 * 60) /* as per RFC 1191 */ 175#endif 176 177#ifdef COMPAT_50 178#include <compat/sys/time.h> 179#include <compat/sys/socket.h> 180#endif 181 182/* 183 * Note: DIRECTED_BROADCAST is handled this way so that previous 184 * configuration using this option will Just Work. 185 */ 186#ifndef IPDIRECTEDBCAST 187#ifdef DIRECTED_BROADCAST 188#define IPDIRECTEDBCAST 1 189#else 190#define IPDIRECTEDBCAST 0 191#endif /* DIRECTED_BROADCAST */ 192#endif /* IPDIRECTEDBCAST */ 193int ipforwarding = IPFORWARDING; 194int ipsendredirects = IPSENDREDIRECTS; 195int ip_defttl = IPDEFTTL; 196int ip_forwsrcrt = IPFORWSRCRT; 197int ip_directedbcast = IPDIRECTEDBCAST; 198int ip_allowsrcrt = IPALLOWSRCRT; 199int ip_mtudisc = IPMTUDISC; 200int ip_mtudisc_timeout = IPMTUDISCTIMEOUT; 201#ifdef DIAGNOSTIC 202int ipprintfs = 0; 203#endif 204 205int ip_do_randomid = 0; 206 207/* 208 * XXX - Setting ip_checkinterface mostly implements the receive side of 209 * the Strong ES model described in RFC 1122, but since the routing table 210 * and transmit implementation do not implement the Strong ES model, 211 * setting this to 1 results in an odd hybrid. 212 * 213 * XXX - ip_checkinterface currently must be disabled if you use ipnat 214 * to translate the destination address to another local interface. 215 * 216 * XXX - ip_checkinterface must be disabled if you add IP aliases 217 * to the loopback interface instead of the interface where the 218 * packets for those addresses are received. 219 */ 220static int ip_checkinterface __read_mostly = 0; 221 222struct rttimer_queue *ip_mtudisc_timeout_q = NULL; 223 224pktqueue_t * ip_pktq __read_mostly; 225pfil_head_t * inet_pfil_hook __read_mostly; 226ipid_state_t * ip_ids __read_mostly; 227percpu_t * ipstat_percpu __read_mostly; 228 229static percpu_t *ipforward_rt_percpu __cacheline_aligned; 230 231uint16_t ip_id; 232 233#ifdef INET_CSUM_COUNTERS 234#include <sys/device.h> 235 236struct evcnt ip_hwcsum_bad = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 237 NULL, "inet", "hwcsum bad"); 238struct evcnt ip_hwcsum_ok = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 239 NULL, "inet", "hwcsum ok"); 240struct evcnt ip_swcsum = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 241 NULL, "inet", "swcsum"); 242 243#define INET_CSUM_COUNTER_INCR(ev) (ev)->ev_count++ 244 245EVCNT_ATTACH_STATIC(ip_hwcsum_bad); 246EVCNT_ATTACH_STATIC(ip_hwcsum_ok); 247EVCNT_ATTACH_STATIC(ip_swcsum); 248 249#else 250 251#define INET_CSUM_COUNTER_INCR(ev) /* nothing */ 252 253#endif /* INET_CSUM_COUNTERS */ 254 255/* 256 * We need to save the IP options in case a protocol wants to respond 257 * to an incoming packet over the same route if the packet got here 258 * using IP source routing. This allows connection establishment and 259 * maintenance when the remote end is on a network that is not known 260 * to us. 261 */ 262 263static int ip_nhops = 0; 264 265static struct ip_srcrt { 266 struct in_addr dst; /* final destination */ 267 char nop; /* one NOP to align */ 268 char srcopt[IPOPT_OFFSET + 1]; /* OPTVAL, OLEN and OFFSET */ 269 struct in_addr route[MAX_IPOPTLEN/sizeof(struct in_addr)]; 270} ip_srcrt; 271 272static int ip_drainwanted; 273 274struct sockaddr_in ipaddr = { 275 .sin_len = sizeof(ipaddr), 276 .sin_family = AF_INET, 277}; 278 279static void save_rte(u_char *, struct in_addr); 280 281#ifdef MBUFTRACE 282struct mowner ip_rx_mowner = MOWNER_INIT("internet", "rx"); 283struct mowner ip_tx_mowner = MOWNER_INIT("internet", "tx"); 284#endif 285 286static void ipintr(void *); 287static void ip_input(struct mbuf *); 288static void ip_forward(struct mbuf *, int, struct ifnet *); 289static bool ip_dooptions(struct mbuf *); 290static struct in_ifaddr *ip_rtaddr(struct in_addr, struct psref *); 291static void sysctl_net_inet_ip_setup(struct sysctllog **); 292 293static struct in_ifaddr *ip_match_our_address(struct ifnet *, struct ip *, 294 int *); 295static struct in_ifaddr *ip_match_our_address_broadcast(struct ifnet *, 296 struct ip *); 297 298#ifdef NET_MPSAFE 299#define SOFTNET_LOCK() mutex_enter(softnet_lock) 300#define SOFTNET_UNLOCK() mutex_exit(softnet_lock) 301#else 302#define SOFTNET_LOCK() KASSERT(mutex_owned(softnet_lock)) 303#define SOFTNET_UNLOCK() KASSERT(mutex_owned(softnet_lock)) 304#endif 305 306/* 307 * IP initialization: fill in IP protocol switch table. 308 * All protocols not implemented in kernel go to raw IP protocol handler. 309 */ 310void 311ip_init(void) 312{ 313 const struct protosw *pr; 314 315 in_init(); 316 sysctl_net_inet_ip_setup(NULL); 317 318 pr = pffindproto(PF_INET, IPPROTO_RAW, SOCK_RAW); 319 KASSERT(pr != NULL); 320 321 ip_pktq = pktq_create(IFQ_MAXLEN, ipintr, NULL); 322 KASSERT(ip_pktq != NULL); 323 324 for (u_int i = 0; i < IPPROTO_MAX; i++) { 325 ip_protox[i] = pr - inetsw; 326 } 327 for (pr = inetdomain.dom_protosw; 328 pr < inetdomain.dom_protoswNPROTOSW; pr++) 329 if (pr->pr_domain->dom_family == PF_INET && 330 pr->pr_protocol && pr->pr_protocol != IPPROTO_RAW) 331 ip_protox[pr->pr_protocol] = pr - inetsw; 332 333 ip_reass_init(); 334 335 ip_ids = ip_id_init(); 336 ip_id = time_uptime & 0xfffff; 337 338#ifdef GATEWAY 339 ipflow_init(); 340#endif 341 342 /* Register our Packet Filter hook. */ 343 inet_pfil_hook = pfil_head_create(PFIL_TYPE_AF, (void *)AF_INET); 344 KASSERT(inet_pfil_hook != NULL); 345 346#ifdef MBUFTRACE 347 MOWNER_ATTACH(&ip_tx_mowner); 348 MOWNER_ATTACH(&ip_rx_mowner); 349#endif /* MBUFTRACE */ 350 351 ipstat_percpu = percpu_alloc(sizeof(uint64_t) * IP_NSTATS); 352 353 ipforward_rt_percpu = percpu_alloc(sizeof(struct route)); 354 if (ipforward_rt_percpu == NULL) 355 panic("failed to allocate ipforward_rt_percpu"); 356} 357 358static struct in_ifaddr * 359ip_match_our_address(struct ifnet *ifp, struct ip *ip, int *downmatch) 360{ 361 struct in_ifaddr *ia = NULL; 362 int checkif; 363 364 /* 365 * Enable a consistency check between the destination address 366 * and the arrival interface for a unicast packet (the RFC 1122 367 * strong ES model) if IP forwarding is disabled and the packet 368 * is not locally generated. 369 * 370 * XXX - Checking also should be disabled if the destination 371 * address is ipnat'ed to a different interface. 372 * 373 * XXX - Checking is incompatible with IP aliases added 374 * to the loopback interface instead of the interface where 375 * the packets are received. 376 * 377 * XXX - We need to add a per ifaddr flag for this so that 378 * we get finer grain control. 379 */ 380 checkif = ip_checkinterface && (ipforwarding == 0) && 381 (ifp->if_flags & IFF_LOOPBACK) == 0; 382 383 IN_ADDRHASH_READER_FOREACH(ia, ip->ip_dst.s_addr) { 384 if (in_hosteq(ia->ia_addr.sin_addr, ip->ip_dst)) { 385 if (ia->ia4_flags & IN_IFF_NOTREADY) 386 continue; 387 if (checkif && ia->ia_ifp != ifp) 388 continue; 389 if ((ia->ia_ifp->if_flags & IFF_UP) != 0 && 390 (ia->ia4_flags & IN_IFF_DETACHED) == 0) 391 break; 392 else 393 (*downmatch)++; 394 } 395 } 396 397 return ia; 398} 399 400static struct in_ifaddr * 401ip_match_our_address_broadcast(struct ifnet *ifp, struct ip *ip) 402{ 403 struct in_ifaddr *ia = NULL; 404 struct ifaddr *ifa; 405 406 IFADDR_READER_FOREACH(ifa, ifp) { 407 if (ifa->ifa_addr->sa_family != AF_INET) 408 continue; 409 ia = ifatoia(ifa); 410 if (ia->ia4_flags & (IN_IFF_NOTREADY | IN_IFF_DETACHED)) 411 continue; 412 if (in_hosteq(ip->ip_dst, ia->ia_broadaddr.sin_addr) || 413 in_hosteq(ip->ip_dst, ia->ia_netbroadcast) || 414 /* 415 * Look for all-0's host part (old broadcast addr), 416 * either for subnet or net. 417 */ 418 ip->ip_dst.s_addr == ia->ia_subnet || 419 ip->ip_dst.s_addr == ia->ia_net) 420 goto matched; 421 /* 422 * An interface with IP address zero accepts 423 * all packets that arrive on that interface. 424 */ 425 if (in_nullhost(ia->ia_addr.sin_addr)) 426 goto matched; 427 } 428 ia = NULL; 429 430matched: 431 return ia; 432} 433 434/* 435 * IP software interrupt routine. 436 */ 437static void 438ipintr(void *arg __unused) 439{ 440 struct mbuf *m; 441 442 KASSERT(cpu_softintr_p()); 443 444#ifndef NET_MPSAFE 445 mutex_enter(softnet_lock); 446#endif 447 while ((m = pktq_dequeue(ip_pktq)) != NULL) { 448 ip_input(m); 449 } 450#ifndef NET_MPSAFE 451 mutex_exit(softnet_lock); 452#endif 453} 454 455/* 456 * IP input routine. Checksum and byte swap header. If fragmented 457 * try to reassemble. Process options. Pass to next level. 458 */ 459static void 460ip_input(struct mbuf *m) 461{ 462 struct ip *ip = NULL; 463 struct in_ifaddr *ia = NULL; 464 int hlen = 0, len; 465 int downmatch; 466 int srcrt = 0; 467 ifnet_t *ifp; 468 struct psref psref; 469 int s; 470 471 KASSERTMSG(cpu_softintr_p(), "ip_input: not in the software " 472 "interrupt handler; synchronization assumptions violated"); 473 474 MCLAIM(m, &ip_rx_mowner); 475 KASSERT((m->m_flags & M_PKTHDR) != 0); 476 477 ifp = m_get_rcvif_psref(m, &psref); 478 if (__predict_false(ifp == NULL)) 479 goto out; 480 481 /* 482 * If no IP addresses have been set yet but the interfaces 483 * are receiving, can't do anything with incoming packets yet. 484 * Note: we pre-check without locks held. 485 */ 486 if (IN_ADDRLIST_READER_EMPTY()) 487 goto out; 488 IP_STATINC(IP_STAT_TOTAL); 489 490 /* 491 * If the IP header is not aligned, slurp it up into a new 492 * mbuf with space for link headers, in the event we forward 493 * it. Otherwise, if it is aligned, make sure the entire 494 * base IP header is in the first mbuf of the chain. 495 */ 496 if (IP_HDR_ALIGNED_P(mtod(m, void *)) == 0) { 497 if ((m = m_copyup(m, sizeof(struct ip), 498 (max_linkhdr + 3) & ~3)) == NULL) { 499 /* XXXJRT new stat, please */ 500 IP_STATINC(IP_STAT_TOOSMALL); 501 goto out; 502 } 503 } else if (__predict_false(m->m_len < sizeof (struct ip))) { 504 if ((m = m_pullup(m, sizeof (struct ip))) == NULL) { 505 IP_STATINC(IP_STAT_TOOSMALL); 506 goto out; 507 } 508 } 509 ip = mtod(m, struct ip *); 510 if (ip->ip_v != IPVERSION) { 511 IP_STATINC(IP_STAT_BADVERS); 512 goto out; 513 } 514 hlen = ip->ip_hl << 2; 515 if (hlen < sizeof(struct ip)) { /* minimum header length */ 516 IP_STATINC(IP_STAT_BADHLEN); 517 goto out; 518 } 519 if (hlen > m->m_len) { 520 if ((m = m_pullup(m, hlen)) == NULL) { 521 IP_STATINC(IP_STAT_BADHLEN); 522 goto out; 523 } 524 ip = mtod(m, struct ip *); 525 } 526 527 /* 528 * RFC1122: packets with a multicast source address are 529 * not allowed. 530 */ 531 if (IN_MULTICAST(ip->ip_src.s_addr)) { 532 IP_STATINC(IP_STAT_BADADDR); 533 goto out; 534 } 535 536 /* 127/8 must not appear on wire - RFC1122 */ 537 if ((ntohl(ip->ip_dst.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET || 538 (ntohl(ip->ip_src.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET) { 539 if ((ifp->if_flags & IFF_LOOPBACK) == 0) { 540 IP_STATINC(IP_STAT_BADADDR); 541 goto out; 542 } 543 } 544 545 switch (m->m_pkthdr.csum_flags & 546 ((ifp->if_csum_flags_rx & M_CSUM_IPv4) | 547 M_CSUM_IPv4_BAD)) { 548 case M_CSUM_IPv4|M_CSUM_IPv4_BAD: 549 INET_CSUM_COUNTER_INCR(&ip_hwcsum_bad); 550 IP_STATINC(IP_STAT_BADSUM); 551 goto out; 552 553 case M_CSUM_IPv4: 554 /* Checksum was okay. */ 555 INET_CSUM_COUNTER_INCR(&ip_hwcsum_ok); 556 break; 557 558 default: 559 /* 560 * Must compute it ourselves. Maybe skip checksum on 561 * loopback interfaces. 562 */ 563 if (__predict_true(!(ifp->if_flags & IFF_LOOPBACK) || 564 ip_do_loopback_cksum)) { 565 INET_CSUM_COUNTER_INCR(&ip_swcsum); 566 if (in_cksum(m, hlen) != 0) { 567 IP_STATINC(IP_STAT_BADSUM); 568 goto out; 569 } 570 } 571 break; 572 } 573 574 /* Retrieve the packet length. */ 575 len = ntohs(ip->ip_len); 576 577 /* 578 * Check for additional length bogosity 579 */ 580 if (len < hlen) { 581 IP_STATINC(IP_STAT_BADLEN); 582 goto out; 583 } 584 585 /* 586 * Check that the amount of data in the buffers 587 * is as at least much as the IP header would have us expect. 588 * Trim mbufs if longer than we expect. 589 * Drop packet if shorter than we expect. 590 */ 591 if (m->m_pkthdr.len < len) { 592 IP_STATINC(IP_STAT_TOOSHORT); 593 goto out; 594 } 595 if (m->m_pkthdr.len > len) { 596 if (m->m_len == m->m_pkthdr.len) { 597 m->m_len = len; 598 m->m_pkthdr.len = len; 599 } else 600 m_adj(m, len - m->m_pkthdr.len); 601 } 602 603 /* 604 * Assume that we can create a fast-forward IP flow entry 605 * based on this packet. 606 */ 607 m->m_flags |= M_CANFASTFWD; 608 609 /* 610 * Run through list of hooks for input packets. If there are any 611 * filters which require that additional packets in the flow are 612 * not fast-forwarded, they must clear the M_CANFASTFWD flag. 613 * Note that filters must _never_ set this flag, as another filter 614 * in the list may have previously cleared it. 615 */ 616#if defined(IPSEC) 617 if (!ipsec_used || !ipsec_indone(m)) 618#else 619 if (1) 620#endif 621 { 622 struct in_addr odst = ip->ip_dst; 623 bool freed; 624 625 freed = pfil_run_hooks(inet_pfil_hook, &m, ifp, PFIL_IN) != 0; 626 if (freed || m == NULL) { 627 m = NULL; 628 goto out; 629 } 630 ip = mtod(m, struct ip *); 631 hlen = ip->ip_hl << 2; 632 633 /* 634 * XXX The setting of "srcrt" here is to prevent ip_forward() 635 * from generating ICMP redirects for packets that have 636 * been redirected by a hook back out on to the same LAN that 637 * they came from and is not an indication that the packet 638 * is being inffluenced by source routing options. This 639 * allows things like 640 * "rdr tlp0 0/0 port 80 -> 1.1.1.200 3128 tcp" 641 * where tlp0 is both on the 1.1.1.0/24 network and is the 642 * default route for hosts on 1.1.1.0/24. Of course this 643 * also requires a "map tlp0 ..." to complete the story. 644 * One might argue whether or not this kind of network config. 645 * should be supported in this manner... 646 */ 647 srcrt = (odst.s_addr != ip->ip_dst.s_addr); 648 } 649 650#ifdef ALTQ 651 /* XXX Temporary until ALTQ is changed to use a pfil hook */ 652 if (altq_input) { 653 SOFTNET_LOCK(); 654 if ((*altq_input)(m, AF_INET) == 0) { 655 /* Packet dropped by traffic conditioner. */ 656 SOFTNET_UNLOCK(); 657 m = NULL; 658 goto out; 659 } 660 SOFTNET_UNLOCK(); 661 } 662#endif 663 664 /* 665 * Process options and, if not destined for us, 666 * ship it on. ip_dooptions returns 1 when an 667 * error was detected (causing an icmp message 668 * to be sent and the original packet to be freed). 669 */ 670 ip_nhops = 0; /* for source routed packets */ 671 if (hlen > sizeof (struct ip) && ip_dooptions(m)) { 672 m = NULL; 673 goto out; 674 } 675 676 /* 677 * Check our list of addresses, to see if the packet is for us. 678 * 679 * Traditional 4.4BSD did not consult IFF_UP at all. 680 * The behavior here is to treat addresses on !IFF_UP interface 681 * or IN_IFF_NOTREADY addresses as not mine. 682 */ 683 downmatch = 0; 684 s = pserialize_read_enter(); 685 ia = ip_match_our_address(ifp, ip, &downmatch); 686 if (ia != NULL) { 687 pserialize_read_exit(s); 688 goto ours; 689 } 690 691 if (ifp->if_flags & IFF_BROADCAST) { 692 ia = ip_match_our_address_broadcast(ifp, ip); 693 if (ia != NULL) { 694 pserialize_read_exit(s); 695 goto ours; 696 } 697 } 698 pserialize_read_exit(s); 699 700 if (IN_MULTICAST(ip->ip_dst.s_addr)) { 701#ifdef MROUTING 702 extern struct socket *ip_mrouter; 703 704 if (ip_mrouter) { 705 /* 706 * If we are acting as a multicast router, all 707 * incoming multicast packets are passed to the 708 * kernel-level multicast forwarding function. 709 * The packet is returned (relatively) intact; if 710 * ip_mforward() returns a non-zero value, the packet 711 * must be discarded, else it may be accepted below. 712 * 713 * (The IP ident field is put in the same byte order 714 * as expected when ip_mforward() is called from 715 * ip_output().) 716 */ 717 SOFTNET_LOCK(); 718 if (ip_mforward(m, ifp) != 0) { 719 SOFTNET_UNLOCK(); 720 IP_STATINC(IP_STAT_CANTFORWARD); 721 goto out; 722 } 723 SOFTNET_UNLOCK(); 724 725 /* 726 * The process-level routing demon needs to receive 727 * all multicast IGMP packets, whether or not this 728 * host belongs to their destination groups. 729 */ 730 if (ip->ip_p == IPPROTO_IGMP) { 731 goto ours; 732 } 733 IP_STATINC(IP_STAT_CANTFORWARD); 734 } 735#endif 736 /* 737 * See if we belong to the destination multicast group on the 738 * arrival interface. 739 */ 740 if (!in_multi_group(ip->ip_dst, ifp, 0)) { 741 IP_STATINC(IP_STAT_CANTFORWARD); 742 goto out; 743 } 744 goto ours; 745 } 746 if (ip->ip_dst.s_addr == INADDR_BROADCAST || 747 in_nullhost(ip->ip_dst)) 748 goto ours; 749 750 /* 751 * Not for us; forward if possible and desirable. 752 */ 753 if (ipforwarding == 0) { 754 m_put_rcvif_psref(ifp, &psref); 755 IP_STATINC(IP_STAT_CANTFORWARD); 756 m_freem(m); 757 } else { 758 /* 759 * If ip_dst matched any of my address on !IFF_UP interface, 760 * and there's no IFF_UP interface that matches ip_dst, 761 * send icmp unreach. Forwarding it will result in in-kernel 762 * forwarding loop till TTL goes to 0. 763 */ 764 if (downmatch) { 765 m_put_rcvif_psref(ifp, &psref); 766 icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_HOST, 0, 0); 767 IP_STATINC(IP_STAT_CANTFORWARD); 768 return; 769 } 770#ifdef IPSEC 771 /* Perform IPsec, if any. */ 772 if (ipsec_used) { 773 SOFTNET_LOCK(); 774 if (ipsec4_input(m, IP_FORWARDING | 775 (ip_directedbcast ? IP_ALLOWBROADCAST : 0)) != 0) { 776 SOFTNET_UNLOCK(); 777 goto out; 778 } 779 SOFTNET_UNLOCK(); 780 } 781#endif 782 ip_forward(m, srcrt, ifp); 783 m_put_rcvif_psref(ifp, &psref); 784 } 785 return; 786 787ours: 788 m_put_rcvif_psref(ifp, &psref); 789 ifp = NULL; 790 791 /* 792 * If offset or IP_MF are set, must reassemble. 793 */ 794 if (ip->ip_off & ~htons(IP_DF|IP_RF)) { 795 /* 796 * Pass to IP reassembly mechanism. 797 */ 798 if (ip_reass_packet(&m, ip) != 0) { 799 /* Failed; invalid fragment(s) or packet. */ 800 goto out; 801 } 802 if (m == NULL) { 803 /* More fragments should come; silently return. */ 804 goto out; 805 } 806 /* 807 * Reassembly is done, we have the final packet. 808 * Updated cached data in local variable(s). 809 */ 810 ip = mtod(m, struct ip *); 811 hlen = ip->ip_hl << 2; 812 } 813 814#ifdef IPSEC 815 /* 816 * Enforce IPsec policy checking if we are seeing last header. 817 * Note that we do not visit this with protocols with PCB layer 818 * code - like UDP/TCP/raw IP. 819 */ 820 if (ipsec_used && 821 (inetsw[ip_protox[ip->ip_p]].pr_flags & PR_LASTHDR) != 0) { 822 SOFTNET_LOCK(); 823 if (ipsec4_input(m, 0) != 0) { 824 SOFTNET_UNLOCK(); 825 goto out; 826 } 827 SOFTNET_UNLOCK(); 828 } 829#endif 830 831 /* 832 * Switch out to protocol's input routine. 833 */ 834#if IFA_STATS 835 if (ia && ip) { 836 struct in_ifaddr *_ia; 837 /* 838 * Keep a reference from ip_match_our_address with psref 839 * is expensive, so explore ia here again. 840 */ 841 s = pserialize_read_enter(); 842 _ia = in_get_ia(ip->ip_dst); 843 _ia->ia_ifa.ifa_data.ifad_inbytes += ntohs(ip->ip_len); 844 pserialize_read_exit(s); 845 } 846#endif 847 IP_STATINC(IP_STAT_DELIVERED); 848 849 const int off = hlen, nh = ip->ip_p; 850 851 SOFTNET_LOCK(); 852 (*inetsw[ip_protox[nh]].pr_input)(m, off, nh); 853 SOFTNET_UNLOCK(); 854 return; 855 856out: 857 m_put_rcvif_psref(ifp, &psref); 858 if (m != NULL) 859 m_freem(m); 860} 861 862/* 863 * IP timer processing. 864 */ 865void 866ip_slowtimo(void) 867{ 868 869 mutex_enter(softnet_lock); 870 KERNEL_LOCK(1, NULL); 871 872 ip_reass_slowtimo(); 873 874 KERNEL_UNLOCK_ONE(NULL); 875 mutex_exit(softnet_lock); 876} 877 878/* 879 * IP drain processing. 880 */ 881void 882ip_drain(void) 883{ 884 885 KERNEL_LOCK(1, NULL); 886 ip_reass_drain(); 887 KERNEL_UNLOCK_ONE(NULL); 888} 889 890/* 891 * ip_dooptions: perform option processing on a datagram, possibly discarding 892 * it if bad options are encountered, or forwarding it if source-routed. 893 * 894 * => Returns true if packet has been forwarded/freed. 895 * => Returns false if the packet should be processed further. 896 */ 897static bool 898ip_dooptions(struct mbuf *m) 899{ 900 struct ip *ip = mtod(m, struct ip *); 901 u_char *cp, *cp0; 902 struct ip_timestamp *ipt; 903 struct in_ifaddr *ia; 904 int opt, optlen, cnt, off, code, type = ICMP_PARAMPROB, forward = 0; 905 struct in_addr dst; 906 n_time ntime; 907 struct ifaddr *ifa; 908 int s; 909 910 dst = ip->ip_dst; 911 cp = (u_char *)(ip + 1); 912 cnt = (ip->ip_hl << 2) - sizeof (struct ip); 913 for (; cnt > 0; cnt -= optlen, cp += optlen) { 914 opt = cp[IPOPT_OPTVAL]; 915 if (opt == IPOPT_EOL) 916 break; 917 if (opt == IPOPT_NOP) 918 optlen = 1; 919 else { 920 if (cnt < IPOPT_OLEN + sizeof(*cp)) { 921 code = &cp[IPOPT_OLEN] - (u_char *)ip; 922 goto bad; 923 } 924 optlen = cp[IPOPT_OLEN]; 925 if (optlen < IPOPT_OLEN + sizeof(*cp) || optlen > cnt) { 926 code = &cp[IPOPT_OLEN] - (u_char *)ip; 927 goto bad; 928 } 929 } 930 switch (opt) { 931 932 default: 933 break; 934 935 /* 936 * Source routing with record. 937 * Find interface with current destination address. 938 * If none on this machine then drop if strictly routed, 939 * or do nothing if loosely routed. 940 * Record interface address and bring up next address 941 * component. If strictly routed make sure next 942 * address is on directly accessible net. 943 */ 944 case IPOPT_LSRR: 945 case IPOPT_SSRR: { 946 struct psref psref; 947 if (ip_allowsrcrt == 0) { 948 type = ICMP_UNREACH; 949 code = ICMP_UNREACH_NET_PROHIB; 950 goto bad; 951 } 952 if (optlen < IPOPT_OFFSET + sizeof(*cp)) { 953 code = &cp[IPOPT_OLEN] - (u_char *)ip; 954 goto bad; 955 } 956 if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) { 957 code = &cp[IPOPT_OFFSET] - (u_char *)ip; 958 goto bad; 959 } 960 ipaddr.sin_addr = ip->ip_dst; 961 962 s = pserialize_read_enter(); 963 ifa = ifa_ifwithaddr(sintosa(&ipaddr)); 964 if (ifa == NULL) { 965 pserialize_read_exit(s); 966 if (opt == IPOPT_SSRR) { 967 type = ICMP_UNREACH; 968 code = ICMP_UNREACH_SRCFAIL; 969 goto bad; 970 } 971 /* 972 * Loose routing, and not at next destination 973 * yet; nothing to do except forward. 974 */ 975 break; 976 } 977 pserialize_read_exit(s); 978 979 off--; /* 0 origin */ 980 if ((off + sizeof(struct in_addr)) > optlen) { 981 /* 982 * End of source route. Should be for us. 983 */ 984 save_rte(cp, ip->ip_src); 985 break; 986 } 987 /* 988 * locate outgoing interface 989 */ 990 memcpy((void *)&ipaddr.sin_addr, (void *)(cp + off), 991 sizeof(ipaddr.sin_addr)); 992 if (opt == IPOPT_SSRR) { 993 ifa = ifa_ifwithladdr_psref(sintosa(&ipaddr), 994 &psref); 995 if (ifa != NULL) 996 ia = ifatoia(ifa); 997 else 998 ia = NULL; 999 } else { 1000 ia = ip_rtaddr(ipaddr.sin_addr, &psref); 1001 } 1002 if (ia == NULL) { 1003 type = ICMP_UNREACH; 1004 code = ICMP_UNREACH_SRCFAIL; 1005 goto bad; 1006 } 1007 ip->ip_dst = ipaddr.sin_addr; 1008 bcopy((void *)&ia->ia_addr.sin_addr, 1009 (void *)(cp + off), sizeof(struct in_addr)); 1010 ia4_release(ia, &psref); 1011 cp[IPOPT_OFFSET] += sizeof(struct in_addr); 1012 /* 1013 * Let ip_intr's mcast routing check handle mcast pkts 1014 */ 1015 forward = !IN_MULTICAST(ip->ip_dst.s_addr); 1016 break; 1017 } 1018 1019 case IPOPT_RR: { 1020 struct psref psref; 1021 if (optlen < IPOPT_OFFSET + sizeof(*cp)) { 1022 code = &cp[IPOPT_OLEN] - (u_char *)ip; 1023 goto bad; 1024 } 1025 if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) { 1026 code = &cp[IPOPT_OFFSET] - (u_char *)ip; 1027 goto bad; 1028 } 1029 /* 1030 * If no space remains, ignore. 1031 */ 1032 off--; /* 0 origin */ 1033 if ((off + sizeof(struct in_addr)) > optlen) 1034 break; 1035 memcpy((void *)&ipaddr.sin_addr, (void *)(&ip->ip_dst), 1036 sizeof(ipaddr.sin_addr)); 1037 /* 1038 * locate outgoing interface; if we're the destination, 1039 * use the incoming interface (should be same). 1040 */ 1041 ifa = ifa_ifwithaddr_psref(sintosa(&ipaddr), &psref); 1042 if (ifa == NULL) { 1043 ia = ip_rtaddr(ipaddr.sin_addr, &psref); 1044 if (ia == NULL) { 1045 type = ICMP_UNREACH; 1046 code = ICMP_UNREACH_HOST; 1047 goto bad; 1048 } 1049 } else { 1050 ia = ifatoia(ifa); 1051 } 1052 bcopy((void *)&ia->ia_addr.sin_addr, 1053 (void *)(cp + off), sizeof(struct in_addr)); 1054 ia4_release(ia, &psref); 1055 cp[IPOPT_OFFSET] += sizeof(struct in_addr); 1056 break; 1057 } 1058 1059 case IPOPT_TS: 1060 code = cp - (u_char *)ip; 1061 ipt = (struct ip_timestamp *)cp; 1062 if (ipt->ipt_len < 4 || ipt->ipt_len > 40) { 1063 code = (u_char *)&ipt->ipt_len - (u_char *)ip; 1064 goto bad; 1065 } 1066 if (ipt->ipt_ptr < 5) { 1067 code = (u_char *)&ipt->ipt_ptr - (u_char *)ip; 1068 goto bad; 1069 } 1070 if (ipt->ipt_ptr > ipt->ipt_len - sizeof (int32_t)) { 1071 if (++ipt->ipt_oflw == 0) { 1072 code = (u_char *)&ipt->ipt_ptr - 1073 (u_char *)ip; 1074 goto bad; 1075 } 1076 break; 1077 } 1078 cp0 = (cp + ipt->ipt_ptr - 1); 1079 switch (ipt->ipt_flg) { 1080 1081 case IPOPT_TS_TSONLY: 1082 break; 1083 1084 case IPOPT_TS_TSANDADDR: { 1085 struct ifnet *rcvif; 1086 int _s, _ss; 1087 1088 if (ipt->ipt_ptr - 1 + sizeof(n_time) + 1089 sizeof(struct in_addr) > ipt->ipt_len) { 1090 code = (u_char *)&ipt->ipt_ptr - 1091 (u_char *)ip; 1092 goto bad; 1093 } 1094 ipaddr.sin_addr = dst; 1095 _ss = pserialize_read_enter(); 1096 rcvif = m_get_rcvif(m, &_s); 1097 ifa = ifaof_ifpforaddr(sintosa(&ipaddr), rcvif); 1098 m_put_rcvif(rcvif, &_s); 1099 if (ifa == NULL) { 1100 pserialize_read_exit(_ss); 1101 break; 1102 } 1103 ia = ifatoia(ifa); 1104 bcopy(&ia->ia_addr.sin_addr, 1105 cp0, sizeof(struct in_addr)); 1106 pserialize_read_exit(_ss); 1107 ipt->ipt_ptr += sizeof(struct in_addr); 1108 break; 1109 } 1110 1111 case IPOPT_TS_PRESPEC: 1112 if (ipt->ipt_ptr - 1 + sizeof(n_time) + 1113 sizeof(struct in_addr) > ipt->ipt_len) { 1114 code = (u_char *)&ipt->ipt_ptr - 1115 (u_char *)ip; 1116 goto bad; 1117 } 1118 memcpy(&ipaddr.sin_addr, cp0, 1119 sizeof(struct in_addr)); 1120 s = pserialize_read_enter(); 1121 ifa = ifa_ifwithaddr(sintosa(&ipaddr)); 1122 if (ifa == NULL) { 1123 pserialize_read_exit(s); 1124 continue; 1125 } 1126 pserialize_read_exit(s); 1127 ipt->ipt_ptr += sizeof(struct in_addr); 1128 break; 1129 1130 default: 1131 /* XXX can't take &ipt->ipt_flg */ 1132 code = (u_char *)&ipt->ipt_ptr - 1133 (u_char *)ip + 1; 1134 goto bad; 1135 } 1136 ntime = iptime(); 1137 cp0 = (u_char *) &ntime; /* XXX grumble, GCC... */ 1138 memmove((char *)cp + ipt->ipt_ptr - 1, cp0, 1139 sizeof(n_time)); 1140 ipt->ipt_ptr += sizeof(n_time); 1141 } 1142 } 1143 if (forward) { 1144 struct ifnet *rcvif; 1145 struct psref _psref; 1146 1147 if (ip_forwsrcrt == 0) { 1148 type = ICMP_UNREACH; 1149 code = ICMP_UNREACH_SRCFAIL; 1150 goto bad; 1151 } 1152 1153 rcvif = m_get_rcvif_psref(m, &_psref); 1154 if (__predict_false(rcvif == NULL)) { 1155 type = ICMP_UNREACH; 1156 code = ICMP_UNREACH_HOST; 1157 goto bad; 1158 } 1159 ip_forward(m, 1, rcvif); 1160 m_put_rcvif_psref(rcvif, &_psref); 1161 return true; 1162 } 1163 return false; 1164bad: 1165 icmp_error(m, type, code, 0, 0); 1166 IP_STATINC(IP_STAT_BADOPTIONS); 1167 return true; 1168} 1169 1170/* 1171 * ip_rtaddr: given address of next destination (final or next hop), 1172 * return internet address info of interface to be used to get there. 1173 */ 1174static struct in_ifaddr * 1175ip_rtaddr(struct in_addr dst, struct psref *psref) 1176{ 1177 struct rtentry *rt; 1178 union { 1179 struct sockaddr dst; 1180 struct sockaddr_in dst4; 1181 } u; 1182 struct route *ro; 1183 1184 sockaddr_in_init(&u.dst4, &dst, 0); 1185 1186 ro = percpu_getref(ipforward_rt_percpu); 1187 rt = rtcache_lookup(ro, &u.dst); 1188 if (rt == NULL) { 1189 percpu_putref(ipforward_rt_percpu); 1190 return NULL; 1191 } 1192 1193 ia4_acquire(ifatoia(rt->rt_ifa), psref); 1194 rtcache_unref(rt, ro); 1195 percpu_putref(ipforward_rt_percpu); 1196 1197 return ifatoia(rt->rt_ifa); 1198} 1199 1200/* 1201 * save_rte: save incoming source route for use in replies, to be picked 1202 * up later by ip_srcroute if the receiver is interested. 1203 */ 1204static void 1205save_rte(u_char *option, struct in_addr dst) 1206{ 1207 unsigned olen; 1208 1209 olen = option[IPOPT_OLEN]; 1210 if (olen > sizeof(ip_srcrt) - (1 + sizeof(dst))) 1211 return; 1212 memcpy((void *)ip_srcrt.srcopt, (void *)option, olen); 1213 ip_nhops = (olen - IPOPT_OFFSET - 1) / sizeof(struct in_addr); 1214 ip_srcrt.dst = dst; 1215} 1216 1217/* 1218 * Retrieve incoming source route for use in replies, 1219 * in the same form used by setsockopt. 1220 * The first hop is placed before the options, will be removed later. 1221 */ 1222struct mbuf * 1223ip_srcroute(void) 1224{ 1225 struct in_addr *p, *q; 1226 struct mbuf *m; 1227 1228 if (ip_nhops == 0) 1229 return NULL; 1230 m = m_get(M_DONTWAIT, MT_SOOPTS); 1231 if (m == 0) 1232 return NULL; 1233 1234 MCLAIM(m, &inetdomain.dom_mowner); 1235#define OPTSIZ (sizeof(ip_srcrt.nop) + sizeof(ip_srcrt.srcopt)) 1236 1237 /* length is (nhops+1)*sizeof(addr) + sizeof(nop + srcrt header) */ 1238 m->m_len = ip_nhops * sizeof(struct in_addr) + sizeof(struct in_addr) + 1239 OPTSIZ; 1240 1241 /* 1242 * First save first hop for return route 1243 */ 1244 p = &ip_srcrt.route[ip_nhops - 1]; 1245 *(mtod(m, struct in_addr *)) = *p--; 1246 1247 /* 1248 * Copy option fields and padding (nop) to mbuf. 1249 */ 1250 ip_srcrt.nop = IPOPT_NOP; 1251 ip_srcrt.srcopt[IPOPT_OFFSET] = IPOPT_MINOFF; 1252 memmove(mtod(m, char *) + sizeof(struct in_addr), &ip_srcrt.nop, 1253 OPTSIZ); 1254 q = (struct in_addr *)(mtod(m, char *) + 1255 sizeof(struct in_addr) + OPTSIZ); 1256#undef OPTSIZ 1257 /* 1258 * Record return path as an IP source route, 1259 * reversing the path (pointers are now aligned). 1260 */ 1261 while (p >= ip_srcrt.route) { 1262 *q++ = *p--; 1263 } 1264 /* 1265 * Last hop goes to final destination. 1266 */ 1267 *q = ip_srcrt.dst; 1268 return (m); 1269} 1270 1271const int inetctlerrmap[PRC_NCMDS] = { 1272 [PRC_MSGSIZE] = EMSGSIZE, 1273 [PRC_HOSTDEAD] = EHOSTDOWN, 1274 [PRC_HOSTUNREACH] = EHOSTUNREACH, 1275 [PRC_UNREACH_NET] = EHOSTUNREACH, 1276 [PRC_UNREACH_HOST] = EHOSTUNREACH, 1277 [PRC_UNREACH_PROTOCOL] = ECONNREFUSED, 1278 [PRC_UNREACH_PORT] = ECONNREFUSED, 1279 [PRC_UNREACH_SRCFAIL] = EHOSTUNREACH, 1280 [PRC_PARAMPROB] = ENOPROTOOPT, 1281}; 1282 1283void 1284ip_fasttimo(void) 1285{ 1286 if (ip_drainwanted) { 1287 ip_drain(); 1288 ip_drainwanted = 0; 1289 } 1290} 1291 1292void 1293ip_drainstub(void) 1294{ 1295 ip_drainwanted = 1; 1296} 1297 1298/* 1299 * Forward a packet. If some error occurs return the sender 1300 * an icmp packet. Note we can't always generate a meaningful 1301 * icmp message because icmp doesn't have a large enough repertoire 1302 * of codes and types. 1303 * 1304 * If not forwarding, just drop the packet. This could be confusing 1305 * if ipforwarding was zero but some routing protocol was advancing 1306 * us as a gateway to somewhere. However, we must let the routing 1307 * protocol deal with that. 1308 * 1309 * The srcrt parameter indicates whether the packet is being forwarded 1310 * via a source route. 1311 */ 1312static void 1313ip_forward(struct mbuf *m, int srcrt, struct ifnet *rcvif) 1314{ 1315 struct ip *ip = mtod(m, struct ip *); 1316 struct rtentry *rt; 1317 int error, type = 0, code = 0, destmtu = 0; 1318 struct mbuf *mcopy; 1319 n_long dest; 1320 union { 1321 struct sockaddr dst; 1322 struct sockaddr_in dst4; 1323 } u; 1324 uint64_t *ips; 1325 struct route *ro; 1326 1327 KASSERTMSG(cpu_softintr_p(), "ip_forward: not in the software " 1328 "interrupt handler; synchronization assumptions violated"); 1329 1330 /* 1331 * We are now in the output path. 1332 */ 1333 MCLAIM(m, &ip_tx_mowner); 1334 1335 /* 1336 * Clear any in-bound checksum flags for this packet. 1337 */ 1338 m->m_pkthdr.csum_flags = 0; 1339 1340 dest = 0; 1341 if (m->m_flags & (M_BCAST|M_MCAST) || in_canforward(ip->ip_dst) == 0) { 1342 IP_STATINC(IP_STAT_CANTFORWARD); 1343 m_freem(m); 1344 return; 1345 } 1346 1347 if (ip->ip_ttl <= IPTTLDEC) { 1348 icmp_error(m, ICMP_TIMXCEED, ICMP_TIMXCEED_INTRANS, dest, 0); 1349 return; 1350 } 1351 1352 sockaddr_in_init(&u.dst4, &ip->ip_dst, 0); 1353 1354 ro = percpu_getref(ipforward_rt_percpu); 1355 rt = rtcache_lookup(ro, &u.dst); 1356 if (rt == NULL) { 1357 percpu_putref(ipforward_rt_percpu); 1358 icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_NET, dest, 0); 1359 return; 1360 } 1361 1362 /* 1363 * Save at most 68 bytes of the packet in case 1364 * we need to generate an ICMP message to the src. 1365 * Pullup to avoid sharing mbuf cluster between m and mcopy. 1366 */ 1367 mcopy = m_copym(m, 0, imin(ntohs(ip->ip_len), 68), M_DONTWAIT); 1368 if (mcopy) 1369 mcopy = m_pullup(mcopy, ip->ip_hl << 2); 1370 1371 ip->ip_ttl -= IPTTLDEC; 1372 1373 /* 1374 * If forwarding packet using same interface that it came in on, 1375 * perhaps should send a redirect to sender to shortcut a hop. 1376 * Only send redirect if source is sending directly to us, 1377 * and if packet was not source routed (or has any options). 1378 * Also, don't send redirect if forwarding using a default route 1379 * or a route modified by a redirect. 1380 */ 1381 if (rt->rt_ifp == rcvif && 1382 (rt->rt_flags & (RTF_DYNAMIC|RTF_MODIFIED)) == 0 && 1383 !in_nullhost(satocsin(rt_getkey(rt))->sin_addr) && 1384 ipsendredirects && !srcrt) { 1385 if (rt->rt_ifa && 1386 (ip->ip_src.s_addr & ifatoia(rt->rt_ifa)->ia_subnetmask) == 1387 ifatoia(rt->rt_ifa)->ia_subnet) { 1388 if (rt->rt_flags & RTF_GATEWAY) 1389 dest = satosin(rt->rt_gateway)->sin_addr.s_addr; 1390 else 1391 dest = ip->ip_dst.s_addr; 1392 /* 1393 * Router requirements says to only send host 1394 * redirects. 1395 */ 1396 type = ICMP_REDIRECT; 1397 code = ICMP_REDIRECT_HOST; 1398 } 1399 } 1400 rtcache_unref(rt, ro); 1401 1402 error = ip_output(m, NULL, ro, 1403 (IP_FORWARDING | (ip_directedbcast ? IP_ALLOWBROADCAST : 0)), 1404 NULL, NULL); 1405 1406 if (error) { 1407 IP_STATINC(IP_STAT_CANTFORWARD); 1408 goto error; 1409 } 1410 1411 ips = IP_STAT_GETREF(); 1412 ips[IP_STAT_FORWARD]++; 1413 1414 if (type) { 1415 ips[IP_STAT_REDIRECTSENT]++; 1416 IP_STAT_PUTREF(); 1417 goto redirect; 1418 } 1419 1420 IP_STAT_PUTREF(); 1421 if (mcopy) { 1422#ifdef GATEWAY 1423 if (mcopy->m_flags & M_CANFASTFWD) 1424 ipflow_create(ro, mcopy); 1425#endif 1426 m_freem(mcopy); 1427 } 1428 1429 percpu_putref(ipforward_rt_percpu); 1430 return; 1431 1432redirect: 1433error: 1434 if (mcopy == NULL) { 1435 percpu_putref(ipforward_rt_percpu); 1436 return; 1437 } 1438 1439 switch (error) { 1440 1441 case 0: /* forwarded, but need redirect */ 1442 /* type, code set above */ 1443 break; 1444 1445 case ENETUNREACH: /* shouldn't happen, checked above */ 1446 case EHOSTUNREACH: 1447 case ENETDOWN: 1448 case EHOSTDOWN: 1449 default: 1450 type = ICMP_UNREACH; 1451 code = ICMP_UNREACH_HOST; 1452 break; 1453 1454 case EMSGSIZE: 1455 type = ICMP_UNREACH; 1456 code = ICMP_UNREACH_NEEDFRAG; 1457 1458 if ((rt = rtcache_validate(ro)) != NULL) { 1459 destmtu = rt->rt_ifp->if_mtu; 1460 rtcache_unref(rt, ro); 1461 } 1462#ifdef IPSEC 1463 if (ipsec_used) 1464 (void)ipsec4_forward(mcopy, &destmtu); 1465#endif 1466 IP_STATINC(IP_STAT_CANTFRAG); 1467 break; 1468 1469 case ENOBUFS: 1470 /* 1471 * Do not generate ICMP_SOURCEQUENCH as required in RFC 1812, 1472 * Requirements for IP Version 4 Routers. Source quench can 1473 * big problem under DoS attacks or if the underlying 1474 * interface is rate-limited. 1475 */ 1476 if (mcopy) 1477 m_freem(mcopy); 1478 percpu_putref(ipforward_rt_percpu); 1479 return; 1480 } 1481 icmp_error(mcopy, type, code, dest, destmtu); 1482 percpu_putref(ipforward_rt_percpu); 1483} 1484 1485void 1486ip_savecontrol(struct inpcb *inp, struct mbuf **mp, struct ip *ip, 1487 struct mbuf *m) 1488{ 1489 struct socket *so = inp->inp_socket; 1490 ifnet_t *ifp; 1491 int inpflags = inp->inp_flags; 1492 struct psref psref; 1493 1494 ifp = m_get_rcvif_psref(m, &psref); 1495 if (__predict_false(ifp == NULL)) 1496 return; /* XXX should report error? */ 1497 1498 if (so->so_options & SO_TIMESTAMP 1499#ifdef SO_OTIMESTAMP 1500 || so->so_options & SO_OTIMESTAMP 1501#endif 1502 ) { 1503 struct timeval tv; 1504 1505 microtime(&tv); 1506#ifdef SO_OTIMESTAMP 1507 if (so->so_options & SO_OTIMESTAMP) { 1508 struct timeval50 tv50; 1509 timeval_to_timeval50(&tv, &tv50); 1510 *mp = sbcreatecontrol((void *) &tv50, sizeof(tv50), 1511 SCM_OTIMESTAMP, SOL_SOCKET); 1512 } else 1513#endif 1514 *mp = sbcreatecontrol((void *) &tv, sizeof(tv), 1515 SCM_TIMESTAMP, SOL_SOCKET); 1516 if (*mp) 1517 mp = &(*mp)->m_next; 1518 } 1519 if (inpflags & INP_RECVDSTADDR) { 1520 *mp = sbcreatecontrol((void *) &ip->ip_dst, 1521 sizeof(struct in_addr), IP_RECVDSTADDR, IPPROTO_IP); 1522 if (*mp) 1523 mp = &(*mp)->m_next; 1524 } 1525 if (inpflags & INP_RECVPKTINFO) { 1526 struct in_pktinfo ipi; 1527 ipi.ipi_addr = ip->ip_src; 1528 ipi.ipi_ifindex = ifp->if_index; 1529 *mp = sbcreatecontrol((void *) &ipi, 1530 sizeof(ipi), IP_RECVPKTINFO, IPPROTO_IP); 1531 if (*mp) 1532 mp = &(*mp)->m_next; 1533 } 1534 if (inpflags & INP_PKTINFO) { 1535 struct in_pktinfo ipi; 1536 ipi.ipi_addr = ip->ip_dst; 1537 ipi.ipi_ifindex = ifp->if_index; 1538 *mp = sbcreatecontrol((void *) &ipi, 1539 sizeof(ipi), IP_PKTINFO, IPPROTO_IP); 1540 if (*mp) 1541 mp = &(*mp)->m_next; 1542 } 1543 if (inpflags & INP_RECVIF) { 1544 struct sockaddr_dl sdl; 1545 1546 sockaddr_dl_init(&sdl, sizeof(sdl), ifp->if_index, 0, NULL, 0, 1547 NULL, 0); 1548 *mp = sbcreatecontrol(&sdl, sdl.sdl_len, IP_RECVIF, IPPROTO_IP); 1549 if (*mp) 1550 mp = &(*mp)->m_next; 1551 } 1552 if (inpflags & INP_RECVTTL) { 1553 *mp = sbcreatecontrol((void *) &ip->ip_ttl, 1554 sizeof(uint8_t), IP_RECVTTL, IPPROTO_IP); 1555 if (*mp) 1556 mp = &(*mp)->m_next; 1557 } 1558 m_put_rcvif_psref(ifp, &psref); 1559} 1560 1561/* 1562 * sysctl helper routine for net.inet.ip.forwsrcrt. 1563 */ 1564static int 1565sysctl_net_inet_ip_forwsrcrt(SYSCTLFN_ARGS) 1566{ 1567 int error, tmp; 1568 struct sysctlnode node; 1569 1570 node = *rnode; 1571 tmp = ip_forwsrcrt; 1572 node.sysctl_data = &tmp; 1573 error = sysctl_lookup(SYSCTLFN_CALL(&node)); 1574 if (error || newp == NULL) 1575 return (error); 1576 1577 error = kauth_authorize_network(l->l_cred, KAUTH_NETWORK_FORWSRCRT, 1578 0, NULL, NULL, NULL); 1579 if (error) 1580 return (error); 1581 1582 ip_forwsrcrt = tmp; 1583 1584 return (0); 1585} 1586 1587/* 1588 * sysctl helper routine for net.inet.ip.mtudisctimeout. checks the 1589 * range of the new value and tweaks timers if it changes. 1590 */ 1591static int 1592sysctl_net_inet_ip_pmtudto(SYSCTLFN_ARGS) 1593{ 1594 int error, tmp; 1595 struct sysctlnode node; 1596 1597 node = *rnode; 1598 tmp = ip_mtudisc_timeout; 1599 node.sysctl_data = &tmp; 1600 error = sysctl_lookup(SYSCTLFN_CALL(&node)); 1601 if (error || newp == NULL) 1602 return (error); 1603 if (tmp < 0) 1604 return (EINVAL); 1605 1606 mutex_enter(softnet_lock); 1607 1608 ip_mtudisc_timeout = tmp; 1609 rt_timer_queue_change(ip_mtudisc_timeout_q, ip_mtudisc_timeout); 1610 1611 mutex_exit(softnet_lock); 1612 1613 return (0); 1614} 1615 1616static int 1617sysctl_net_inet_ip_stats(SYSCTLFN_ARGS) 1618{ 1619 1620 return (NETSTAT_SYSCTL(ipstat_percpu, IP_NSTATS)); 1621} 1622 1623static void 1624sysctl_net_inet_ip_setup(struct sysctllog **clog) 1625{ 1626 sysctl_createv(clog, 0, NULL, NULL, 1627 CTLFLAG_PERMANENT, 1628 CTLTYPE_NODE, "inet", 1629 SYSCTL_DESCR("PF_INET related settings"), 1630 NULL, 0, NULL, 0, 1631 CTL_NET, PF_INET, CTL_EOL); 1632 sysctl_createv(clog, 0, NULL, NULL, 1633 CTLFLAG_PERMANENT, 1634 CTLTYPE_NODE, "ip", 1635 SYSCTL_DESCR("IPv4 related settings"), 1636 NULL, 0, NULL, 0, 1637 CTL_NET, PF_INET, IPPROTO_IP, CTL_EOL); 1638 1639 sysctl_createv(clog, 0, NULL, NULL, 1640 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1641 CTLTYPE_INT, "forwarding", 1642 SYSCTL_DESCR("Enable forwarding of INET datagrams"), 1643 NULL, 0, &ipforwarding, 0, 1644 CTL_NET, PF_INET, IPPROTO_IP, 1645 IPCTL_FORWARDING, CTL_EOL); 1646 sysctl_createv(clog, 0, NULL, NULL, 1647 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1648 CTLTYPE_INT, "redirect", 1649 SYSCTL_DESCR("Enable sending of ICMP redirect messages"), 1650 NULL, 0, &ipsendredirects, 0, 1651 CTL_NET, PF_INET, IPPROTO_IP, 1652 IPCTL_SENDREDIRECTS, CTL_EOL); 1653 sysctl_createv(clog, 0, NULL, NULL, 1654 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1655 CTLTYPE_INT, "ttl", 1656 SYSCTL_DESCR("Default TTL for an INET datagram"), 1657 NULL, 0, &ip_defttl, 0, 1658 CTL_NET, PF_INET, IPPROTO_IP, 1659 IPCTL_DEFTTL, CTL_EOL); 1660#ifdef IPCTL_DEFMTU 1661 sysctl_createv(clog, 0, NULL, NULL, 1662 CTLFLAG_PERMANENT /* |CTLFLAG_READWRITE? */, 1663 CTLTYPE_INT, "mtu", 1664 SYSCTL_DESCR("Default MTA for an INET route"), 1665 NULL, 0, &ip_mtu, 0, 1666 CTL_NET, PF_INET, IPPROTO_IP, 1667 IPCTL_DEFMTU, CTL_EOL); 1668#endif /* IPCTL_DEFMTU */ 1669 sysctl_createv(clog, 0, NULL, NULL, 1670 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1671 CTLTYPE_INT, "forwsrcrt", 1672 SYSCTL_DESCR("Enable forwarding of source-routed " 1673 "datagrams"), 1674 sysctl_net_inet_ip_forwsrcrt, 0, &ip_forwsrcrt, 0, 1675 CTL_NET, PF_INET, IPPROTO_IP, 1676 IPCTL_FORWSRCRT, CTL_EOL); 1677 sysctl_createv(clog, 0, NULL, NULL, 1678 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1679 CTLTYPE_INT, "directed-broadcast", 1680 SYSCTL_DESCR("Enable forwarding of broadcast datagrams"), 1681 NULL, 0, &ip_directedbcast, 0, 1682 CTL_NET, PF_INET, IPPROTO_IP, 1683 IPCTL_DIRECTEDBCAST, CTL_EOL); 1684 sysctl_createv(clog, 0, NULL, NULL, 1685 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1686 CTLTYPE_INT, "allowsrcrt", 1687 SYSCTL_DESCR("Accept source-routed datagrams"), 1688 NULL, 0, &ip_allowsrcrt, 0, 1689 CTL_NET, PF_INET, IPPROTO_IP, 1690 IPCTL_ALLOWSRCRT, CTL_EOL); 1691 1692 sysctl_createv(clog, 0, NULL, NULL, 1693 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1694 CTLTYPE_INT, "mtudisc", 1695 SYSCTL_DESCR("Use RFC1191 Path MTU Discovery"), 1696 NULL, 0, &ip_mtudisc, 0, 1697 CTL_NET, PF_INET, IPPROTO_IP, 1698 IPCTL_MTUDISC, CTL_EOL); 1699 sysctl_createv(clog, 0, NULL, NULL, 1700 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1701 CTLTYPE_INT, "anonportmin", 1702 SYSCTL_DESCR("Lowest ephemeral port number to assign"), 1703 sysctl_net_inet_ip_ports, 0, &anonportmin, 0, 1704 CTL_NET, PF_INET, IPPROTO_IP, 1705 IPCTL_ANONPORTMIN, CTL_EOL); 1706 sysctl_createv(clog, 0, NULL, NULL, 1707 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1708 CTLTYPE_INT, "anonportmax", 1709 SYSCTL_DESCR("Highest ephemeral port number to assign"), 1710 sysctl_net_inet_ip_ports, 0, &anonportmax, 0, 1711 CTL_NET, PF_INET, IPPROTO_IP, 1712 IPCTL_ANONPORTMAX, CTL_EOL); 1713 sysctl_createv(clog, 0, NULL, NULL, 1714 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1715 CTLTYPE_INT, "mtudisctimeout", 1716 SYSCTL_DESCR("Lifetime of a Path MTU Discovered route"), 1717 sysctl_net_inet_ip_pmtudto, 0, (void *)&ip_mtudisc_timeout, 0, 1718 CTL_NET, PF_INET, IPPROTO_IP, 1719 IPCTL_MTUDISCTIMEOUT, CTL_EOL); 1720#ifndef IPNOPRIVPORTS 1721 sysctl_createv(clog, 0, NULL, NULL, 1722 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1723 CTLTYPE_INT, "lowportmin", 1724 SYSCTL_DESCR("Lowest privileged ephemeral port number " 1725 "to assign"), 1726 sysctl_net_inet_ip_ports, 0, &lowportmin, 0, 1727 CTL_NET, PF_INET, IPPROTO_IP, 1728 IPCTL_LOWPORTMIN, CTL_EOL); 1729 sysctl_createv(clog, 0, NULL, NULL, 1730 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1731 CTLTYPE_INT, "lowportmax", 1732 SYSCTL_DESCR("Highest privileged ephemeral port number " 1733 "to assign"), 1734 sysctl_net_inet_ip_ports, 0, &lowportmax, 0, 1735 CTL_NET, PF_INET, IPPROTO_IP, 1736 IPCTL_LOWPORTMAX, CTL_EOL); 1737#endif /* IPNOPRIVPORTS */ 1738#if NGRE > 0 1739 sysctl_createv(clog, 0, NULL, NULL, 1740 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1741 CTLTYPE_INT, "grettl", 1742 SYSCTL_DESCR("Default TTL for a gre tunnel datagram"), 1743 NULL, 0, &ip_gre_ttl, 0, 1744 CTL_NET, PF_INET, IPPROTO_IP, 1745 IPCTL_GRE_TTL, CTL_EOL); 1746#endif /* NGRE */ 1747 sysctl_createv(clog, 0, NULL, NULL, 1748 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1749 CTLTYPE_INT, "checkinterface", 1750 SYSCTL_DESCR("Enable receive side of Strong ES model " 1751 "from RFC1122"), 1752 NULL, 0, &ip_checkinterface, 0, 1753 CTL_NET, PF_INET, IPPROTO_IP, 1754 IPCTL_CHECKINTERFACE, CTL_EOL); 1755 sysctl_createv(clog, 0, NULL, NULL, 1756 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1757 CTLTYPE_INT, "random_id", 1758 SYSCTL_DESCR("Assign random ip_id values"), 1759 NULL, 0, &ip_do_randomid, 0, 1760 CTL_NET, PF_INET, IPPROTO_IP, 1761 IPCTL_RANDOMID, CTL_EOL); 1762 sysctl_createv(clog, 0, NULL, NULL, 1763 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1764 CTLTYPE_INT, "do_loopback_cksum", 1765 SYSCTL_DESCR("Perform IP checksum on loopback"), 1766 NULL, 0, &ip_do_loopback_cksum, 0, 1767 CTL_NET, PF_INET, IPPROTO_IP, 1768 IPCTL_LOOPBACKCKSUM, CTL_EOL); 1769 sysctl_createv(clog, 0, NULL, NULL, 1770 CTLFLAG_PERMANENT, 1771 CTLTYPE_STRUCT, "stats", 1772 SYSCTL_DESCR("IP statistics"), 1773 sysctl_net_inet_ip_stats, 0, NULL, 0, 1774 CTL_NET, PF_INET, IPPROTO_IP, IPCTL_STATS, 1775 CTL_EOL); 1776#if NARP 1777 sysctl_createv(clog, 0, NULL, NULL, 1778 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1779 CTLTYPE_INT, "dad_count", 1780 SYSCTL_DESCR("Number of Duplicate Address Detection " 1781 "probes to send"), 1782 NULL, 0, &ip_dad_count, 0, 1783 CTL_NET, PF_INET, IPPROTO_IP, 1784 IPCTL_DAD_COUNT, CTL_EOL); 1785#endif 1786 1787 /* anonportalgo RFC6056 subtree */ 1788 const struct sysctlnode *portalgo_node; 1789 sysctl_createv(clog, 0, NULL, &portalgo_node, 1790 CTLFLAG_PERMANENT, 1791 CTLTYPE_NODE, "anonportalgo", 1792 SYSCTL_DESCR("Anonymous Port Algorithm Selection (RFC 6056)"), 1793 NULL, 0, NULL, 0, 1794 CTL_NET, PF_INET, IPPROTO_IP, CTL_CREATE, CTL_EOL); 1795 sysctl_createv(clog, 0, &portalgo_node, NULL, 1796 CTLFLAG_PERMANENT, 1797 CTLTYPE_STRING, "available", 1798 SYSCTL_DESCR("available algorithms"), 1799 sysctl_portalgo_available, 0, NULL, PORTALGO_MAXLEN, 1800 CTL_CREATE, CTL_EOL); 1801 sysctl_createv(clog, 0, &portalgo_node, NULL, 1802 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1803 CTLTYPE_STRING, "selected", 1804 SYSCTL_DESCR("selected algorithm"), 1805 sysctl_portalgo_selected4, 0, NULL, PORTALGO_MAXLEN, 1806 CTL_CREATE, CTL_EOL); 1807 sysctl_createv(clog, 0, &portalgo_node, NULL, 1808 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1809 CTLTYPE_STRUCT, "reserve", 1810 SYSCTL_DESCR("bitmap of reserved ports"), 1811 sysctl_portalgo_reserve4, 0, NULL, 0, 1812 CTL_CREATE, CTL_EOL); 1813} 1814 1815void 1816ip_statinc(u_int stat) 1817{ 1818 1819 KASSERT(stat < IP_NSTATS); 1820 IP_STATINC(stat); 1821} 1822