ip_input.c revision 1.306
1/* $NetBSD: ip_input.c,v 1.306 2013/06/27 19:38:16 christos 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 1.306 2013/06/27 19:38:16 christos Exp $"); 95 96#include "opt_inet.h" 97#include "opt_compat_netbsd.h" 98#include "opt_gateway.h" 99#include "opt_pfil_hooks.h" 100#include "opt_ipsec.h" 101#include "opt_mrouting.h" 102#include "opt_mbuftrace.h" 103#include "opt_inet_csum.h" 104 105#include <sys/param.h> 106#include <sys/systm.h> 107#include <sys/mbuf.h> 108#include <sys/domain.h> 109#include <sys/protosw.h> 110#include <sys/socket.h> 111#include <sys/socketvar.h> 112#include <sys/errno.h> 113#include <sys/time.h> 114#include <sys/kernel.h> 115#include <sys/pool.h> 116#include <sys/sysctl.h> 117#include <sys/kauth.h> 118 119#include <net/if.h> 120#include <net/if_dl.h> 121#include <net/route.h> 122#include <net/pfil.h> 123 124#include <netinet/in.h> 125#include <netinet/in_systm.h> 126#include <netinet/ip.h> 127#include <netinet/in_pcb.h> 128#include <netinet/in_proto.h> 129#include <netinet/in_var.h> 130#include <netinet/ip_var.h> 131#include <netinet/ip_private.h> 132#include <netinet/ip_icmp.h> 133/* just for gif_ttl */ 134#include <netinet/in_gif.h> 135#include "gif.h" 136#include <net/if_gre.h> 137#include "gre.h" 138 139#ifdef MROUTING 140#include <netinet/ip_mroute.h> 141#endif 142#include <netinet/portalgo.h> 143 144#ifdef IPSEC 145#include <netipsec/ipsec.h> 146#endif 147 148#ifndef IPFORWARDING 149#ifdef GATEWAY 150#define IPFORWARDING 1 /* forward IP packets not for us */ 151#else /* GATEWAY */ 152#define IPFORWARDING 0 /* don't forward IP packets not for us */ 153#endif /* GATEWAY */ 154#endif /* IPFORWARDING */ 155#ifndef IPSENDREDIRECTS 156#define IPSENDREDIRECTS 1 157#endif 158#ifndef IPFORWSRCRT 159#define IPFORWSRCRT 1 /* forward source-routed packets */ 160#endif 161#ifndef IPALLOWSRCRT 162#define IPALLOWSRCRT 1 /* allow source-routed packets */ 163#endif 164#ifndef IPMTUDISC 165#define IPMTUDISC 1 166#endif 167#ifndef IPMTUDISCTIMEOUT 168#define IPMTUDISCTIMEOUT (10 * 60) /* as per RFC 1191 */ 169#endif 170 171#ifdef COMPAT_50 172#include <compat/sys/time.h> 173#include <compat/sys/socket.h> 174#endif 175 176/* 177 * Note: DIRECTED_BROADCAST is handled this way so that previous 178 * configuration using this option will Just Work. 179 */ 180#ifndef IPDIRECTEDBCAST 181#ifdef DIRECTED_BROADCAST 182#define IPDIRECTEDBCAST 1 183#else 184#define IPDIRECTEDBCAST 0 185#endif /* DIRECTED_BROADCAST */ 186#endif /* IPDIRECTEDBCAST */ 187int ipforwarding = IPFORWARDING; 188int ipsendredirects = IPSENDREDIRECTS; 189int ip_defttl = IPDEFTTL; 190int ip_forwsrcrt = IPFORWSRCRT; 191int ip_directedbcast = IPDIRECTEDBCAST; 192int ip_allowsrcrt = IPALLOWSRCRT; 193int ip_mtudisc = IPMTUDISC; 194int ip_mtudisc_timeout = IPMTUDISCTIMEOUT; 195#ifdef DIAGNOSTIC 196int ipprintfs = 0; 197#endif 198 199int ip_do_randomid = 0; 200 201/* 202 * XXX - Setting ip_checkinterface mostly implements the receive side of 203 * the Strong ES model described in RFC 1122, but since the routing table 204 * and transmit implementation do not implement the Strong ES model, 205 * setting this to 1 results in an odd hybrid. 206 * 207 * XXX - ip_checkinterface currently must be disabled if you use ipnat 208 * to translate the destination address to another local interface. 209 * 210 * XXX - ip_checkinterface must be disabled if you add IP aliases 211 * to the loopback interface instead of the interface where the 212 * packets for those addresses are received. 213 */ 214int ip_checkinterface = 0; 215 216 217struct rttimer_queue *ip_mtudisc_timeout_q = NULL; 218 219u_long in_ifaddrhash; /* size of hash table - 1 */ 220int in_ifaddrentries; /* total number of addrs */ 221struct in_ifaddrhead in_ifaddrhead; 222struct in_ifaddrhashhead *in_ifaddrhashtbl; 223u_long in_multihash; /* size of hash table - 1 */ 224int in_multientries; /* total number of addrs */ 225struct in_multihashhead *in_multihashtbl; 226struct ifqueue ipintrq; 227 228ipid_state_t * ip_ids; 229uint16_t ip_id; 230 231percpu_t *ipstat_percpu; 232 233#ifdef PFIL_HOOKS 234struct pfil_head inet_pfil_hook; 235#endif 236 237struct pool inmulti_pool; 238 239#ifdef INET_CSUM_COUNTERS 240#include <sys/device.h> 241 242struct evcnt ip_hwcsum_bad = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 243 NULL, "inet", "hwcsum bad"); 244struct evcnt ip_hwcsum_ok = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 245 NULL, "inet", "hwcsum ok"); 246struct evcnt ip_swcsum = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 247 NULL, "inet", "swcsum"); 248 249#define INET_CSUM_COUNTER_INCR(ev) (ev)->ev_count++ 250 251EVCNT_ATTACH_STATIC(ip_hwcsum_bad); 252EVCNT_ATTACH_STATIC(ip_hwcsum_ok); 253EVCNT_ATTACH_STATIC(ip_swcsum); 254 255#else 256 257#define INET_CSUM_COUNTER_INCR(ev) /* nothing */ 258 259#endif /* INET_CSUM_COUNTERS */ 260 261/* 262 * We need to save the IP options in case a protocol wants to respond 263 * to an incoming packet over the same route if the packet got here 264 * using IP source routing. This allows connection establishment and 265 * maintenance when the remote end is on a network that is not known 266 * to us. 267 */ 268int ip_nhops = 0; 269static struct ip_srcrt { 270 struct in_addr dst; /* final destination */ 271 char nop; /* one NOP to align */ 272 char srcopt[IPOPT_OFFSET + 1]; /* OPTVAL, OLEN and OFFSET */ 273 struct in_addr route[MAX_IPOPTLEN/sizeof(struct in_addr)]; 274} ip_srcrt; 275 276static int ip_drainwanted; 277 278static void save_rte(u_char *, struct in_addr); 279 280#ifdef MBUFTRACE 281struct mowner ip_rx_mowner = MOWNER_INIT("internet", "rx"); 282struct mowner ip_tx_mowner = MOWNER_INIT("internet", "tx"); 283#endif 284 285static void sysctl_net_inet_ip_setup(struct sysctllog **); 286 287/* 288 * IP initialization: fill in IP protocol switch table. 289 * All protocols not implemented in kernel go to raw IP protocol handler. 290 */ 291void 292ip_init(void) 293{ 294 const struct protosw *pr; 295 int i; 296 297 sysctl_net_inet_ip_setup(NULL); 298 299 pool_init(&inmulti_pool, sizeof(struct in_multi), 0, 0, 0, "inmltpl", 300 NULL, IPL_SOFTNET); 301 302 pr = pffindproto(PF_INET, IPPROTO_RAW, SOCK_RAW); 303 if (pr == 0) 304 panic("ip_init"); 305 for (i = 0; i < IPPROTO_MAX; i++) 306 ip_protox[i] = pr - inetsw; 307 for (pr = inetdomain.dom_protosw; 308 pr < inetdomain.dom_protoswNPROTOSW; pr++) 309 if (pr->pr_domain->dom_family == PF_INET && 310 pr->pr_protocol && pr->pr_protocol != IPPROTO_RAW) 311 ip_protox[pr->pr_protocol] = pr - inetsw; 312 313 ip_reass_init(); 314 315 ip_ids = ip_id_init(); 316 ip_id = time_second & 0xfffff; 317 318 ipintrq.ifq_maxlen = IFQ_MAXLEN; 319 320 TAILQ_INIT(&in_ifaddrhead); 321 in_ifaddrhashtbl = hashinit(IN_IFADDR_HASH_SIZE, HASH_LIST, true, 322 &in_ifaddrhash); 323 in_multihashtbl = hashinit(IN_IFADDR_HASH_SIZE, HASH_LIST, true, 324 &in_multihash); 325 ip_mtudisc_timeout_q = rt_timer_queue_create(ip_mtudisc_timeout); 326#ifdef GATEWAY 327 ipflow_init(ip_hashsize); 328#endif 329 330#ifdef PFIL_HOOKS 331 /* Register our Packet Filter hook. */ 332 inet_pfil_hook.ph_type = PFIL_TYPE_AF; 333 inet_pfil_hook.ph_af = AF_INET; 334 i = pfil_head_register(&inet_pfil_hook); 335 if (i != 0) 336 printf("ip_init: WARNING: unable to register pfil hook, " 337 "error %d\n", i); 338#endif /* PFIL_HOOKS */ 339 340#ifdef MBUFTRACE 341 MOWNER_ATTACH(&ip_tx_mowner); 342 MOWNER_ATTACH(&ip_rx_mowner); 343#endif /* MBUFTRACE */ 344 345 ipstat_percpu = percpu_alloc(sizeof(uint64_t) * IP_NSTATS); 346} 347 348struct sockaddr_in ipaddr = { 349 .sin_len = sizeof(ipaddr), 350 .sin_family = AF_INET, 351}; 352struct route ipforward_rt; 353 354/* 355 * IP software interrupt routine 356 */ 357void 358ipintr(void) 359{ 360 int s; 361 struct mbuf *m; 362 struct ifqueue lcl_intrq; 363 364 memset(&lcl_intrq, 0, sizeof(lcl_intrq)); 365 366 mutex_enter(softnet_lock); 367 KERNEL_LOCK(1, NULL); 368 if (!IF_IS_EMPTY(&ipintrq)) { 369 s = splnet(); 370 371 /* Take existing queue onto stack */ 372 lcl_intrq = ipintrq; 373 374 /* Zero out global queue, preserving maxlen and drops */ 375 ipintrq.ifq_head = NULL; 376 ipintrq.ifq_tail = NULL; 377 ipintrq.ifq_len = 0; 378 ipintrq.ifq_maxlen = lcl_intrq.ifq_maxlen; 379 ipintrq.ifq_drops = lcl_intrq.ifq_drops; 380 381 splx(s); 382 } 383 KERNEL_UNLOCK_ONE(NULL); 384 while (!IF_IS_EMPTY(&lcl_intrq)) { 385 IF_DEQUEUE(&lcl_intrq, m); 386 if (m == NULL) 387 break; 388 ip_input(m); 389 } 390 mutex_exit(softnet_lock); 391} 392 393/* 394 * Ip input routine. Checksum and byte swap header. If fragmented 395 * try to reassemble. Process options. Pass to next level. 396 */ 397void 398ip_input(struct mbuf *m) 399{ 400 struct ip *ip = NULL; 401 struct in_ifaddr *ia; 402 struct ifaddr *ifa; 403 int hlen = 0, len; 404 int downmatch; 405 int checkif; 406 int srcrt = 0; 407 408 MCLAIM(m, &ip_rx_mowner); 409 KASSERT((m->m_flags & M_PKTHDR) != 0); 410 411 /* 412 * If no IP addresses have been set yet but the interfaces 413 * are receiving, can't do anything with incoming packets yet. 414 */ 415 if (TAILQ_FIRST(&in_ifaddrhead) == 0) 416 goto bad; 417 IP_STATINC(IP_STAT_TOTAL); 418 /* 419 * If the IP header is not aligned, slurp it up into a new 420 * mbuf with space for link headers, in the event we forward 421 * it. Otherwise, if it is aligned, make sure the entire 422 * base IP header is in the first mbuf of the chain. 423 */ 424 if (IP_HDR_ALIGNED_P(mtod(m, void *)) == 0) { 425 if ((m = m_copyup(m, sizeof(struct ip), 426 (max_linkhdr + 3) & ~3)) == NULL) { 427 /* XXXJRT new stat, please */ 428 IP_STATINC(IP_STAT_TOOSMALL); 429 return; 430 } 431 } else if (__predict_false(m->m_len < sizeof (struct ip))) { 432 if ((m = m_pullup(m, sizeof (struct ip))) == NULL) { 433 IP_STATINC(IP_STAT_TOOSMALL); 434 return; 435 } 436 } 437 ip = mtod(m, struct ip *); 438 if (ip->ip_v != IPVERSION) { 439 IP_STATINC(IP_STAT_BADVERS); 440 goto bad; 441 } 442 hlen = ip->ip_hl << 2; 443 if (hlen < sizeof(struct ip)) { /* minimum header length */ 444 IP_STATINC(IP_STAT_BADHLEN); 445 goto bad; 446 } 447 if (hlen > m->m_len) { 448 if ((m = m_pullup(m, hlen)) == NULL) { 449 IP_STATINC(IP_STAT_BADHLEN); 450 return; 451 } 452 ip = mtod(m, struct ip *); 453 } 454 455 /* 456 * RFC1122: packets with a multicast source address are 457 * not allowed. 458 */ 459 if (IN_MULTICAST(ip->ip_src.s_addr)) { 460 IP_STATINC(IP_STAT_BADADDR); 461 goto bad; 462 } 463 464 /* 127/8 must not appear on wire - RFC1122 */ 465 if ((ntohl(ip->ip_dst.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET || 466 (ntohl(ip->ip_src.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET) { 467 if ((m->m_pkthdr.rcvif->if_flags & IFF_LOOPBACK) == 0) { 468 IP_STATINC(IP_STAT_BADADDR); 469 goto bad; 470 } 471 } 472 473 switch (m->m_pkthdr.csum_flags & 474 ((m->m_pkthdr.rcvif->if_csum_flags_rx & M_CSUM_IPv4) | 475 M_CSUM_IPv4_BAD)) { 476 case M_CSUM_IPv4|M_CSUM_IPv4_BAD: 477 INET_CSUM_COUNTER_INCR(&ip_hwcsum_bad); 478 goto badcsum; 479 480 case M_CSUM_IPv4: 481 /* Checksum was okay. */ 482 INET_CSUM_COUNTER_INCR(&ip_hwcsum_ok); 483 break; 484 485 default: 486 /* 487 * Must compute it ourselves. Maybe skip checksum on 488 * loopback interfaces. 489 */ 490 if (__predict_true(!(m->m_pkthdr.rcvif->if_flags & 491 IFF_LOOPBACK) || ip_do_loopback_cksum)) { 492 INET_CSUM_COUNTER_INCR(&ip_swcsum); 493 if (in_cksum(m, hlen) != 0) 494 goto badcsum; 495 } 496 break; 497 } 498 499 /* Retrieve the packet length. */ 500 len = ntohs(ip->ip_len); 501 502 /* 503 * Check for additional length bogosity 504 */ 505 if (len < hlen) { 506 IP_STATINC(IP_STAT_BADLEN); 507 goto bad; 508 } 509 510 /* 511 * Check that the amount of data in the buffers 512 * is as at least much as the IP header would have us expect. 513 * Trim mbufs if longer than we expect. 514 * Drop packet if shorter than we expect. 515 */ 516 if (m->m_pkthdr.len < len) { 517 IP_STATINC(IP_STAT_TOOSHORT); 518 goto bad; 519 } 520 if (m->m_pkthdr.len > len) { 521 if (m->m_len == m->m_pkthdr.len) { 522 m->m_len = len; 523 m->m_pkthdr.len = len; 524 } else 525 m_adj(m, len - m->m_pkthdr.len); 526 } 527 528 /* 529 * Assume that we can create a fast-forward IP flow entry 530 * based on this packet. 531 */ 532 m->m_flags |= M_CANFASTFWD; 533 534#ifdef PFIL_HOOKS 535 /* 536 * Run through list of hooks for input packets. If there are any 537 * filters which require that additional packets in the flow are 538 * not fast-forwarded, they must clear the M_CANFASTFWD flag. 539 * Note that filters must _never_ set this flag, as another filter 540 * in the list may have previously cleared it. 541 */ 542 /* 543 * let ipfilter look at packet on the wire, 544 * not the decapsulated packet. 545 */ 546#if defined(IPSEC) 547 if (!ipsec_indone(m)) 548#else 549 if (1) 550#endif 551 { 552 struct in_addr odst; 553 554 odst = ip->ip_dst; 555 if (pfil_run_hooks(&inet_pfil_hook, &m, m->m_pkthdr.rcvif, 556 PFIL_IN) != 0) 557 return; 558 if (m == NULL) 559 return; 560 ip = mtod(m, struct ip *); 561 hlen = ip->ip_hl << 2; 562 /* 563 * XXX The setting of "srcrt" here is to prevent ip_forward() 564 * from generating ICMP redirects for packets that have 565 * been redirected by a hook back out on to the same LAN that 566 * they came from and is not an indication that the packet 567 * is being inffluenced by source routing options. This 568 * allows things like 569 * "rdr tlp0 0/0 port 80 -> 1.1.1.200 3128 tcp" 570 * where tlp0 is both on the 1.1.1.0/24 network and is the 571 * default route for hosts on 1.1.1.0/24. Of course this 572 * also requires a "map tlp0 ..." to complete the story. 573 * One might argue whether or not this kind of network config. 574 * should be supported in this manner... 575 */ 576 srcrt = (odst.s_addr != ip->ip_dst.s_addr); 577 } 578#endif /* PFIL_HOOKS */ 579 580#ifdef ALTQ 581 /* XXX Temporary until ALTQ is changed to use a pfil hook */ 582 if (altq_input != NULL && (*altq_input)(m, AF_INET) == 0) { 583 /* packet dropped by traffic conditioner */ 584 return; 585 } 586#endif 587 588 /* 589 * Process options and, if not destined for us, 590 * ship it on. ip_dooptions returns 1 when an 591 * error was detected (causing an icmp message 592 * to be sent and the original packet to be freed). 593 */ 594 ip_nhops = 0; /* for source routed packets */ 595 if (hlen > sizeof (struct ip) && ip_dooptions(m)) 596 return; 597 598 /* 599 * Enable a consistency check between the destination address 600 * and the arrival interface for a unicast packet (the RFC 1122 601 * strong ES model) if IP forwarding is disabled and the packet 602 * is not locally generated. 603 * 604 * XXX - Checking also should be disabled if the destination 605 * address is ipnat'ed to a different interface. 606 * 607 * XXX - Checking is incompatible with IP aliases added 608 * to the loopback interface instead of the interface where 609 * the packets are received. 610 * 611 * XXX - We need to add a per ifaddr flag for this so that 612 * we get finer grain control. 613 */ 614 checkif = ip_checkinterface && (ipforwarding == 0) && 615 (m->m_pkthdr.rcvif != NULL) && 616 ((m->m_pkthdr.rcvif->if_flags & IFF_LOOPBACK) == 0); 617 618 /* 619 * Check our list of addresses, to see if the packet is for us. 620 * 621 * Traditional 4.4BSD did not consult IFF_UP at all. 622 * The behavior here is to treat addresses on !IFF_UP interface 623 * as not mine. 624 */ 625 downmatch = 0; 626 LIST_FOREACH(ia, &IN_IFADDR_HASH(ip->ip_dst.s_addr), ia_hash) { 627 if (in_hosteq(ia->ia_addr.sin_addr, ip->ip_dst)) { 628 if (checkif && ia->ia_ifp != m->m_pkthdr.rcvif) 629 continue; 630 if ((ia->ia_ifp->if_flags & IFF_UP) != 0) 631 break; 632 else 633 downmatch++; 634 } 635 } 636 if (ia != NULL) 637 goto ours; 638 if (m->m_pkthdr.rcvif && m->m_pkthdr.rcvif->if_flags & IFF_BROADCAST) { 639 IFADDR_FOREACH(ifa, m->m_pkthdr.rcvif) { 640 if (ifa->ifa_addr->sa_family != AF_INET) 641 continue; 642 ia = ifatoia(ifa); 643 if (in_hosteq(ip->ip_dst, ia->ia_broadaddr.sin_addr) || 644 in_hosteq(ip->ip_dst, ia->ia_netbroadcast) || 645 /* 646 * Look for all-0's host part (old broadcast addr), 647 * either for subnet or net. 648 */ 649 ip->ip_dst.s_addr == ia->ia_subnet || 650 ip->ip_dst.s_addr == ia->ia_net) 651 goto ours; 652 /* 653 * An interface with IP address zero accepts 654 * all packets that arrive on that interface. 655 */ 656 if (in_nullhost(ia->ia_addr.sin_addr)) 657 goto ours; 658 } 659 } 660 if (IN_MULTICAST(ip->ip_dst.s_addr)) { 661 struct in_multi *inm; 662#ifdef MROUTING 663 extern struct socket *ip_mrouter; 664 665 if (ip_mrouter) { 666 /* 667 * If we are acting as a multicast router, all 668 * incoming multicast packets are passed to the 669 * kernel-level multicast forwarding function. 670 * The packet is returned (relatively) intact; if 671 * ip_mforward() returns a non-zero value, the packet 672 * must be discarded, else it may be accepted below. 673 * 674 * (The IP ident field is put in the same byte order 675 * as expected when ip_mforward() is called from 676 * ip_output().) 677 */ 678 if (ip_mforward(m, m->m_pkthdr.rcvif) != 0) { 679 IP_STATINC(IP_STAT_CANTFORWARD); 680 m_freem(m); 681 return; 682 } 683 684 /* 685 * The process-level routing demon needs to receive 686 * all multicast IGMP packets, whether or not this 687 * host belongs to their destination groups. 688 */ 689 if (ip->ip_p == IPPROTO_IGMP) 690 goto ours; 691 IP_STATINC(IP_STAT_CANTFORWARD); 692 } 693#endif 694 /* 695 * See if we belong to the destination multicast group on the 696 * arrival interface. 697 */ 698 IN_LOOKUP_MULTI(ip->ip_dst, m->m_pkthdr.rcvif, inm); 699 if (inm == NULL) { 700 IP_STATINC(IP_STAT_CANTFORWARD); 701 m_freem(m); 702 return; 703 } 704 goto ours; 705 } 706 if (ip->ip_dst.s_addr == INADDR_BROADCAST || 707 in_nullhost(ip->ip_dst)) 708 goto ours; 709 710 /* 711 * Not for us; forward if possible and desirable. 712 */ 713 if (ipforwarding == 0) { 714 IP_STATINC(IP_STAT_CANTFORWARD); 715 m_freem(m); 716 } else { 717 /* 718 * If ip_dst matched any of my address on !IFF_UP interface, 719 * and there's no IFF_UP interface that matches ip_dst, 720 * send icmp unreach. Forwarding it will result in in-kernel 721 * forwarding loop till TTL goes to 0. 722 */ 723 if (downmatch) { 724 icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_HOST, 0, 0); 725 IP_STATINC(IP_STAT_CANTFORWARD); 726 return; 727 } 728#ifdef IPSEC 729 /* Perform IPsec, if any. */ 730 if (ipsec4_input(m, IP_FORWARDING | (ip_directedbcast ? 731 IP_ALLOWBROADCAST : 0)) != 0) { 732 goto bad; 733 } 734#endif 735 ip_forward(m, srcrt); 736 } 737 return; 738 739ours: 740 /* 741 * If offset or IP_MF are set, must reassemble. 742 */ 743 if (ip->ip_off & ~htons(IP_DF|IP_RF)) { 744 /* 745 * Pass to IP reassembly mechanism. 746 */ 747 if (ip_reass_packet(&m, ip) != 0) { 748 /* Failed; invalid fragment(s) or packet. */ 749 goto bad; 750 } 751 if (m == NULL) { 752 /* More fragments should come; silently return. */ 753 return; 754 } 755 /* 756 * Reassembly is done, we have the final packet. 757 * Updated cached data in local variable(s). 758 */ 759 ip = mtod(m, struct ip *); 760 hlen = ip->ip_hl << 2; 761 } 762 763#ifdef IPSEC 764 /* 765 * Enforce IPsec policy checking if we are seeing last header. 766 * Note that we do not visit this with protocols with PCB layer 767 * code - like UDP/TCP/raw IP. 768 */ 769 if ((inetsw[ip_protox[ip->ip_p]].pr_flags & PR_LASTHDR) != 0) { 770 if (ipsec4_input(m, 0) != 0) { 771 goto bad; 772 } 773 } 774#endif 775 776 /* 777 * Switch out to protocol's input routine. 778 */ 779#if IFA_STATS 780 if (ia && ip) 781 ia->ia_ifa.ifa_data.ifad_inbytes += ntohs(ip->ip_len); 782#endif 783 IP_STATINC(IP_STAT_DELIVERED); 784 { 785 int off = hlen, nh = ip->ip_p; 786 787 (*inetsw[ip_protox[nh]].pr_input)(m, off, nh); 788 return; 789 } 790bad: 791 m_freem(m); 792 return; 793 794badcsum: 795 IP_STATINC(IP_STAT_BADSUM); 796 m_freem(m); 797} 798 799/* 800 * IP timer processing. 801 */ 802void 803ip_slowtimo(void) 804{ 805 806 mutex_enter(softnet_lock); 807 KERNEL_LOCK(1, NULL); 808 809 ip_reass_slowtimo(); 810 811 KERNEL_UNLOCK_ONE(NULL); 812 mutex_exit(softnet_lock); 813} 814 815/* 816 * IP drain processing. 817 */ 818void 819ip_drain(void) 820{ 821 822 KERNEL_LOCK(1, NULL); 823 ip_reass_drain(); 824 KERNEL_UNLOCK_ONE(NULL); 825} 826 827/* 828 * Do option processing on a datagram, 829 * possibly discarding it if bad options are encountered, 830 * or forwarding it if source-routed. 831 * Returns 1 if packet has been forwarded/freed, 832 * 0 if the packet should be processed further. 833 */ 834int 835ip_dooptions(struct mbuf *m) 836{ 837 struct ip *ip = mtod(m, struct ip *); 838 u_char *cp, *cp0; 839 struct ip_timestamp *ipt; 840 struct in_ifaddr *ia; 841 int opt, optlen, cnt, off, code, type = ICMP_PARAMPROB, forward = 0; 842 struct in_addr dst; 843 n_time ntime; 844 845 dst = ip->ip_dst; 846 cp = (u_char *)(ip + 1); 847 cnt = (ip->ip_hl << 2) - sizeof (struct ip); 848 for (; cnt > 0; cnt -= optlen, cp += optlen) { 849 opt = cp[IPOPT_OPTVAL]; 850 if (opt == IPOPT_EOL) 851 break; 852 if (opt == IPOPT_NOP) 853 optlen = 1; 854 else { 855 if (cnt < IPOPT_OLEN + sizeof(*cp)) { 856 code = &cp[IPOPT_OLEN] - (u_char *)ip; 857 goto bad; 858 } 859 optlen = cp[IPOPT_OLEN]; 860 if (optlen < IPOPT_OLEN + sizeof(*cp) || optlen > cnt) { 861 code = &cp[IPOPT_OLEN] - (u_char *)ip; 862 goto bad; 863 } 864 } 865 switch (opt) { 866 867 default: 868 break; 869 870 /* 871 * Source routing with record. 872 * Find interface with current destination address. 873 * If none on this machine then drop if strictly routed, 874 * or do nothing if loosely routed. 875 * Record interface address and bring up next address 876 * component. If strictly routed make sure next 877 * address is on directly accessible net. 878 */ 879 case IPOPT_LSRR: 880 case IPOPT_SSRR: 881 if (ip_allowsrcrt == 0) { 882 type = ICMP_UNREACH; 883 code = ICMP_UNREACH_NET_PROHIB; 884 goto bad; 885 } 886 if (optlen < IPOPT_OFFSET + sizeof(*cp)) { 887 code = &cp[IPOPT_OLEN] - (u_char *)ip; 888 goto bad; 889 } 890 if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) { 891 code = &cp[IPOPT_OFFSET] - (u_char *)ip; 892 goto bad; 893 } 894 ipaddr.sin_addr = ip->ip_dst; 895 ia = ifatoia(ifa_ifwithaddr(sintosa(&ipaddr))); 896 if (ia == 0) { 897 if (opt == IPOPT_SSRR) { 898 type = ICMP_UNREACH; 899 code = ICMP_UNREACH_SRCFAIL; 900 goto bad; 901 } 902 /* 903 * Loose routing, and not at next destination 904 * yet; nothing to do except forward. 905 */ 906 break; 907 } 908 off--; /* 0 origin */ 909 if ((off + sizeof(struct in_addr)) > optlen) { 910 /* 911 * End of source route. Should be for us. 912 */ 913 save_rte(cp, ip->ip_src); 914 break; 915 } 916 /* 917 * locate outgoing interface 918 */ 919 memcpy((void *)&ipaddr.sin_addr, (void *)(cp + off), 920 sizeof(ipaddr.sin_addr)); 921 if (opt == IPOPT_SSRR) 922 ia = ifatoia(ifa_ifwithladdr(sintosa(&ipaddr))); 923 else 924 ia = ip_rtaddr(ipaddr.sin_addr); 925 if (ia == 0) { 926 type = ICMP_UNREACH; 927 code = ICMP_UNREACH_SRCFAIL; 928 goto bad; 929 } 930 ip->ip_dst = ipaddr.sin_addr; 931 bcopy((void *)&ia->ia_addr.sin_addr, 932 (void *)(cp + off), sizeof(struct in_addr)); 933 cp[IPOPT_OFFSET] += sizeof(struct in_addr); 934 /* 935 * Let ip_intr's mcast routing check handle mcast pkts 936 */ 937 forward = !IN_MULTICAST(ip->ip_dst.s_addr); 938 break; 939 940 case IPOPT_RR: 941 if (optlen < IPOPT_OFFSET + sizeof(*cp)) { 942 code = &cp[IPOPT_OLEN] - (u_char *)ip; 943 goto bad; 944 } 945 if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) { 946 code = &cp[IPOPT_OFFSET] - (u_char *)ip; 947 goto bad; 948 } 949 /* 950 * If no space remains, ignore. 951 */ 952 off--; /* 0 origin */ 953 if ((off + sizeof(struct in_addr)) > optlen) 954 break; 955 memcpy((void *)&ipaddr.sin_addr, (void *)(&ip->ip_dst), 956 sizeof(ipaddr.sin_addr)); 957 /* 958 * locate outgoing interface; if we're the destination, 959 * use the incoming interface (should be same). 960 */ 961 if ((ia = ifatoia(ifa_ifwithaddr(sintosa(&ipaddr)))) 962 == NULL && 963 (ia = ip_rtaddr(ipaddr.sin_addr)) == NULL) { 964 type = ICMP_UNREACH; 965 code = ICMP_UNREACH_HOST; 966 goto bad; 967 } 968 bcopy((void *)&ia->ia_addr.sin_addr, 969 (void *)(cp + off), sizeof(struct in_addr)); 970 cp[IPOPT_OFFSET] += sizeof(struct in_addr); 971 break; 972 973 case IPOPT_TS: 974 code = cp - (u_char *)ip; 975 ipt = (struct ip_timestamp *)cp; 976 if (ipt->ipt_len < 4 || ipt->ipt_len > 40) { 977 code = (u_char *)&ipt->ipt_len - (u_char *)ip; 978 goto bad; 979 } 980 if (ipt->ipt_ptr < 5) { 981 code = (u_char *)&ipt->ipt_ptr - (u_char *)ip; 982 goto bad; 983 } 984 if (ipt->ipt_ptr > ipt->ipt_len - sizeof (int32_t)) { 985 if (++ipt->ipt_oflw == 0) { 986 code = (u_char *)&ipt->ipt_ptr - 987 (u_char *)ip; 988 goto bad; 989 } 990 break; 991 } 992 cp0 = (cp + ipt->ipt_ptr - 1); 993 switch (ipt->ipt_flg) { 994 995 case IPOPT_TS_TSONLY: 996 break; 997 998 case IPOPT_TS_TSANDADDR: 999 if (ipt->ipt_ptr - 1 + sizeof(n_time) + 1000 sizeof(struct in_addr) > ipt->ipt_len) { 1001 code = (u_char *)&ipt->ipt_ptr - 1002 (u_char *)ip; 1003 goto bad; 1004 } 1005 ipaddr.sin_addr = dst; 1006 ia = ifatoia(ifaof_ifpforaddr(sintosa(&ipaddr), 1007 m->m_pkthdr.rcvif)); 1008 if (ia == 0) 1009 continue; 1010 bcopy(&ia->ia_addr.sin_addr, 1011 cp0, sizeof(struct in_addr)); 1012 ipt->ipt_ptr += sizeof(struct in_addr); 1013 break; 1014 1015 case IPOPT_TS_PRESPEC: 1016 if (ipt->ipt_ptr - 1 + sizeof(n_time) + 1017 sizeof(struct in_addr) > ipt->ipt_len) { 1018 code = (u_char *)&ipt->ipt_ptr - 1019 (u_char *)ip; 1020 goto bad; 1021 } 1022 memcpy(&ipaddr.sin_addr, cp0, 1023 sizeof(struct in_addr)); 1024 if (ifatoia(ifa_ifwithaddr(sintosa(&ipaddr))) 1025 == NULL) 1026 continue; 1027 ipt->ipt_ptr += sizeof(struct in_addr); 1028 break; 1029 1030 default: 1031 /* XXX can't take &ipt->ipt_flg */ 1032 code = (u_char *)&ipt->ipt_ptr - 1033 (u_char *)ip + 1; 1034 goto bad; 1035 } 1036 ntime = iptime(); 1037 cp0 = (u_char *) &ntime; /* XXX grumble, GCC... */ 1038 memmove((char *)cp + ipt->ipt_ptr - 1, cp0, 1039 sizeof(n_time)); 1040 ipt->ipt_ptr += sizeof(n_time); 1041 } 1042 } 1043 if (forward) { 1044 if (ip_forwsrcrt == 0) { 1045 type = ICMP_UNREACH; 1046 code = ICMP_UNREACH_SRCFAIL; 1047 goto bad; 1048 } 1049 ip_forward(m, 1); 1050 return (1); 1051 } 1052 return (0); 1053bad: 1054 icmp_error(m, type, code, 0, 0); 1055 IP_STATINC(IP_STAT_BADOPTIONS); 1056 return (1); 1057} 1058 1059/* 1060 * Given address of next destination (final or next hop), 1061 * return internet address info of interface to be used to get there. 1062 */ 1063struct in_ifaddr * 1064ip_rtaddr(struct in_addr dst) 1065{ 1066 struct rtentry *rt; 1067 union { 1068 struct sockaddr dst; 1069 struct sockaddr_in dst4; 1070 } u; 1071 1072 sockaddr_in_init(&u.dst4, &dst, 0); 1073 1074 if ((rt = rtcache_lookup(&ipforward_rt, &u.dst)) == NULL) 1075 return NULL; 1076 1077 return ifatoia(rt->rt_ifa); 1078} 1079 1080/* 1081 * Save incoming source route for use in replies, 1082 * to be picked up later by ip_srcroute if the receiver is interested. 1083 */ 1084void 1085save_rte(u_char *option, struct in_addr dst) 1086{ 1087 unsigned olen; 1088 1089 olen = option[IPOPT_OLEN]; 1090#ifdef DIAGNOSTIC 1091 if (ipprintfs) 1092 printf("save_rte: olen %d\n", olen); 1093#endif /* 0 */ 1094 if (olen > sizeof(ip_srcrt) - (1 + sizeof(dst))) 1095 return; 1096 memcpy((void *)ip_srcrt.srcopt, (void *)option, olen); 1097 ip_nhops = (olen - IPOPT_OFFSET - 1) / sizeof(struct in_addr); 1098 ip_srcrt.dst = dst; 1099} 1100 1101/* 1102 * Retrieve incoming source route for use in replies, 1103 * in the same form used by setsockopt. 1104 * The first hop is placed before the options, will be removed later. 1105 */ 1106struct mbuf * 1107ip_srcroute(void) 1108{ 1109 struct in_addr *p, *q; 1110 struct mbuf *m; 1111 1112 if (ip_nhops == 0) 1113 return NULL; 1114 m = m_get(M_DONTWAIT, MT_SOOPTS); 1115 if (m == 0) 1116 return NULL; 1117 1118 MCLAIM(m, &inetdomain.dom_mowner); 1119#define OPTSIZ (sizeof(ip_srcrt.nop) + sizeof(ip_srcrt.srcopt)) 1120 1121 /* length is (nhops+1)*sizeof(addr) + sizeof(nop + srcrt header) */ 1122 m->m_len = ip_nhops * sizeof(struct in_addr) + sizeof(struct in_addr) + 1123 OPTSIZ; 1124#ifdef DIAGNOSTIC 1125 if (ipprintfs) 1126 printf("ip_srcroute: nhops %d mlen %d", ip_nhops, m->m_len); 1127#endif 1128 1129 /* 1130 * First save first hop for return route 1131 */ 1132 p = &ip_srcrt.route[ip_nhops - 1]; 1133 *(mtod(m, struct in_addr *)) = *p--; 1134#ifdef DIAGNOSTIC 1135 if (ipprintfs) 1136 printf(" hops %x", ntohl(mtod(m, struct in_addr *)->s_addr)); 1137#endif 1138 1139 /* 1140 * Copy option fields and padding (nop) to mbuf. 1141 */ 1142 ip_srcrt.nop = IPOPT_NOP; 1143 ip_srcrt.srcopt[IPOPT_OFFSET] = IPOPT_MINOFF; 1144 memmove(mtod(m, char *) + sizeof(struct in_addr), &ip_srcrt.nop, 1145 OPTSIZ); 1146 q = (struct in_addr *)(mtod(m, char *) + 1147 sizeof(struct in_addr) + OPTSIZ); 1148#undef OPTSIZ 1149 /* 1150 * Record return path as an IP source route, 1151 * reversing the path (pointers are now aligned). 1152 */ 1153 while (p >= ip_srcrt.route) { 1154#ifdef DIAGNOSTIC 1155 if (ipprintfs) 1156 printf(" %x", ntohl(q->s_addr)); 1157#endif 1158 *q++ = *p--; 1159 } 1160 /* 1161 * Last hop goes to final destination. 1162 */ 1163 *q = ip_srcrt.dst; 1164#ifdef DIAGNOSTIC 1165 if (ipprintfs) 1166 printf(" %x\n", ntohl(q->s_addr)); 1167#endif 1168 return (m); 1169} 1170 1171const int inetctlerrmap[PRC_NCMDS] = { 1172 [PRC_MSGSIZE] = EMSGSIZE, 1173 [PRC_HOSTDEAD] = EHOSTDOWN, 1174 [PRC_HOSTUNREACH] = EHOSTUNREACH, 1175 [PRC_UNREACH_NET] = EHOSTUNREACH, 1176 [PRC_UNREACH_HOST] = EHOSTUNREACH, 1177 [PRC_UNREACH_PROTOCOL] = ECONNREFUSED, 1178 [PRC_UNREACH_PORT] = ECONNREFUSED, 1179 [PRC_UNREACH_SRCFAIL] = EHOSTUNREACH, 1180 [PRC_PARAMPROB] = ENOPROTOOPT, 1181}; 1182 1183void 1184ip_fasttimo(void) 1185{ 1186 if (ip_drainwanted) { 1187 ip_drain(); 1188 ip_drainwanted = 0; 1189 } 1190} 1191 1192void 1193ip_drainstub(void) 1194{ 1195 ip_drainwanted = 1; 1196} 1197 1198/* 1199 * Forward a packet. If some error occurs return the sender 1200 * an icmp packet. Note we can't always generate a meaningful 1201 * icmp message because icmp doesn't have a large enough repertoire 1202 * of codes and types. 1203 * 1204 * If not forwarding, just drop the packet. This could be confusing 1205 * if ipforwarding was zero but some routing protocol was advancing 1206 * us as a gateway to somewhere. However, we must let the routing 1207 * protocol deal with that. 1208 * 1209 * The srcrt parameter indicates whether the packet is being forwarded 1210 * via a source route. 1211 */ 1212void 1213ip_forward(struct mbuf *m, int srcrt) 1214{ 1215 struct ip *ip = mtod(m, struct ip *); 1216 struct rtentry *rt; 1217 int error, type = 0, code = 0, destmtu = 0; 1218 struct mbuf *mcopy; 1219 n_long dest; 1220 union { 1221 struct sockaddr dst; 1222 struct sockaddr_in dst4; 1223 } u; 1224 1225 /* 1226 * We are now in the output path. 1227 */ 1228 MCLAIM(m, &ip_tx_mowner); 1229 1230 /* 1231 * Clear any in-bound checksum flags for this packet. 1232 */ 1233 m->m_pkthdr.csum_flags = 0; 1234 1235 dest = 0; 1236#ifdef DIAGNOSTIC 1237 if (ipprintfs) { 1238 printf("forward: src %s ", inet_ntoa(ip->ip_src)); 1239 printf("dst %s ttl %x\n", inet_ntoa(ip->ip_dst), ip->ip_ttl); 1240 } 1241#endif 1242 if (m->m_flags & (M_BCAST|M_MCAST) || in_canforward(ip->ip_dst) == 0) { 1243 IP_STATINC(IP_STAT_CANTFORWARD); 1244 m_freem(m); 1245 return; 1246 } 1247 if (ip->ip_ttl <= IPTTLDEC) { 1248 icmp_error(m, ICMP_TIMXCEED, ICMP_TIMXCEED_INTRANS, dest, 0); 1249 return; 1250 } 1251 1252 sockaddr_in_init(&u.dst4, &ip->ip_dst, 0); 1253 if ((rt = rtcache_lookup(&ipforward_rt, &u.dst)) == NULL) { 1254 icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_NET, dest, 0); 1255 return; 1256 } 1257 1258 /* 1259 * Save at most 68 bytes of the packet in case 1260 * we need to generate an ICMP message to the src. 1261 * Pullup to avoid sharing mbuf cluster between m and mcopy. 1262 */ 1263 mcopy = m_copym(m, 0, imin(ntohs(ip->ip_len), 68), M_DONTWAIT); 1264 if (mcopy) 1265 mcopy = m_pullup(mcopy, ip->ip_hl << 2); 1266 1267 ip->ip_ttl -= IPTTLDEC; 1268 1269 /* 1270 * If forwarding packet using same interface that it came in on, 1271 * perhaps should send a redirect to sender to shortcut a hop. 1272 * Only send redirect if source is sending directly to us, 1273 * and if packet was not source routed (or has any options). 1274 * Also, don't send redirect if forwarding using a default route 1275 * or a route modified by a redirect. 1276 */ 1277 if (rt->rt_ifp == m->m_pkthdr.rcvif && 1278 (rt->rt_flags & (RTF_DYNAMIC|RTF_MODIFIED)) == 0 && 1279 !in_nullhost(satocsin(rt_getkey(rt))->sin_addr) && 1280 ipsendredirects && !srcrt) { 1281 if (rt->rt_ifa && 1282 (ip->ip_src.s_addr & ifatoia(rt->rt_ifa)->ia_subnetmask) == 1283 ifatoia(rt->rt_ifa)->ia_subnet) { 1284 if (rt->rt_flags & RTF_GATEWAY) 1285 dest = satosin(rt->rt_gateway)->sin_addr.s_addr; 1286 else 1287 dest = ip->ip_dst.s_addr; 1288 /* 1289 * Router requirements says to only send host 1290 * redirects. 1291 */ 1292 type = ICMP_REDIRECT; 1293 code = ICMP_REDIRECT_HOST; 1294#ifdef DIAGNOSTIC 1295 if (ipprintfs) 1296 printf("redirect (%d) to %x\n", code, 1297 (u_int32_t)dest); 1298#endif 1299 } 1300 } 1301 1302 error = ip_output(m, NULL, &ipforward_rt, 1303 (IP_FORWARDING | (ip_directedbcast ? IP_ALLOWBROADCAST : 0)), 1304 NULL, NULL); 1305 1306 if (error) 1307 IP_STATINC(IP_STAT_CANTFORWARD); 1308 else { 1309 uint64_t *ips = IP_STAT_GETREF(); 1310 ips[IP_STAT_FORWARD]++; 1311 if (type) { 1312 ips[IP_STAT_REDIRECTSENT]++; 1313 IP_STAT_PUTREF(); 1314 } else { 1315 IP_STAT_PUTREF(); 1316 if (mcopy) { 1317#ifdef GATEWAY 1318 if (mcopy->m_flags & M_CANFASTFWD) 1319 ipflow_create(&ipforward_rt, mcopy); 1320#endif 1321 m_freem(mcopy); 1322 } 1323 return; 1324 } 1325 } 1326 if (mcopy == NULL) 1327 return; 1328 1329 switch (error) { 1330 1331 case 0: /* forwarded, but need redirect */ 1332 /* type, code set above */ 1333 break; 1334 1335 case ENETUNREACH: /* shouldn't happen, checked above */ 1336 case EHOSTUNREACH: 1337 case ENETDOWN: 1338 case EHOSTDOWN: 1339 default: 1340 type = ICMP_UNREACH; 1341 code = ICMP_UNREACH_HOST; 1342 break; 1343 1344 case EMSGSIZE: 1345 type = ICMP_UNREACH; 1346 code = ICMP_UNREACH_NEEDFRAG; 1347 1348 if ((rt = rtcache_validate(&ipforward_rt)) != NULL) 1349 destmtu = rt->rt_ifp->if_mtu; 1350#ifdef IPSEC 1351 (void)ipsec4_forward(mcopy, &destmtu); 1352#endif 1353 IP_STATINC(IP_STAT_CANTFRAG); 1354 break; 1355 1356 case ENOBUFS: 1357#if 1 1358 /* 1359 * a router should not generate ICMP_SOURCEQUENCH as 1360 * required in RFC1812 Requirements for IP Version 4 Routers. 1361 * source quench could be a big problem under DoS attacks, 1362 * or if the underlying interface is rate-limited. 1363 */ 1364 if (mcopy) 1365 m_freem(mcopy); 1366 return; 1367#else 1368 type = ICMP_SOURCEQUENCH; 1369 code = 0; 1370 break; 1371#endif 1372 } 1373 icmp_error(mcopy, type, code, dest, destmtu); 1374} 1375 1376void 1377ip_savecontrol(struct inpcb *inp, struct mbuf **mp, struct ip *ip, 1378 struct mbuf *m) 1379{ 1380 1381 if (inp->inp_socket->so_options & SO_TIMESTAMP 1382#ifdef SO_OTIMESTAMP 1383 || inp->inp_socket->so_options & SO_OTIMESTAMP 1384#endif 1385 ) { 1386 struct timeval tv; 1387 1388 microtime(&tv); 1389#ifdef SO_OTIMESTAMP 1390 if (inp->inp_socket->so_options & SO_OTIMESTAMP) { 1391 struct timeval50 tv50; 1392 timeval_to_timeval50(&tv, &tv50); 1393 *mp = sbcreatecontrol((void *) &tv50, sizeof(tv50), 1394 SCM_OTIMESTAMP, SOL_SOCKET); 1395 } else 1396#endif 1397 *mp = sbcreatecontrol((void *) &tv, sizeof(tv), 1398 SCM_TIMESTAMP, SOL_SOCKET); 1399 if (*mp) 1400 mp = &(*mp)->m_next; 1401 } 1402 if (inp->inp_flags & INP_RECVDSTADDR) { 1403 *mp = sbcreatecontrol((void *) &ip->ip_dst, 1404 sizeof(struct in_addr), IP_RECVDSTADDR, IPPROTO_IP); 1405 if (*mp) 1406 mp = &(*mp)->m_next; 1407 } 1408 if (inp->inp_flags & INP_RECVPKTINFO) { 1409 struct in_pktinfo ipi; 1410 ipi.ipi_addr = ip->ip_dst; 1411 ipi.ipi_ifindex = m->m_pkthdr.rcvif->if_index; 1412 *mp = sbcreatecontrol((void *) &ipi, 1413 sizeof(ipi), IP_RECVPKTINFO, IPPROTO_IP); 1414 if (*mp) 1415 mp = &(*mp)->m_next; 1416 } 1417 if (inp->inp_flags & INP_PKTINFO) { 1418 struct in_pktinfo ipi; 1419 ipi.ipi_addr = ip->ip_src; 1420 ipi.ipi_ifindex = m->m_pkthdr.rcvif->if_index; 1421 *mp = sbcreatecontrol((void *) &ipi, 1422 sizeof(ipi), IP_PKTINFO, IPPROTO_IP); 1423 if (*mp) 1424 mp = &(*mp)->m_next; 1425 } 1426#ifdef notyet 1427 /* 1428 * XXX 1429 * Moving these out of udp_input() made them even more broken 1430 * than they already were. 1431 * - fenner@parc.xerox.com 1432 */ 1433 /* options were tossed already */ 1434 if (inp->inp_flags & INP_RECVOPTS) { 1435 *mp = sbcreatecontrol((void *) opts_deleted_above, 1436 sizeof(struct in_addr), IP_RECVOPTS, IPPROTO_IP); 1437 if (*mp) 1438 mp = &(*mp)->m_next; 1439 } 1440 /* ip_srcroute doesn't do what we want here, need to fix */ 1441 if (inp->inp_flags & INP_RECVRETOPTS) { 1442 *mp = sbcreatecontrol((void *) ip_srcroute(), 1443 sizeof(struct in_addr), IP_RECVRETOPTS, IPPROTO_IP); 1444 if (*mp) 1445 mp = &(*mp)->m_next; 1446 } 1447#endif 1448 if (inp->inp_flags & INP_RECVIF) { 1449 struct sockaddr_dl sdl; 1450 1451 sockaddr_dl_init(&sdl, sizeof(sdl), 1452 (m->m_pkthdr.rcvif != NULL) 1453 ? m->m_pkthdr.rcvif->if_index 1454 : 0, 1455 0, NULL, 0, NULL, 0); 1456 *mp = sbcreatecontrol(&sdl, sdl.sdl_len, IP_RECVIF, IPPROTO_IP); 1457 if (*mp) 1458 mp = &(*mp)->m_next; 1459 } 1460 if (inp->inp_flags & INP_RECVTTL) { 1461 *mp = sbcreatecontrol((void *) &ip->ip_ttl, 1462 sizeof(uint8_t), IP_RECVTTL, IPPROTO_IP); 1463 if (*mp) 1464 mp = &(*mp)->m_next; 1465 } 1466} 1467 1468/* 1469 * sysctl helper routine for net.inet.ip.forwsrcrt. 1470 */ 1471static int 1472sysctl_net_inet_ip_forwsrcrt(SYSCTLFN_ARGS) 1473{ 1474 int error, tmp; 1475 struct sysctlnode node; 1476 1477 node = *rnode; 1478 tmp = ip_forwsrcrt; 1479 node.sysctl_data = &tmp; 1480 error = sysctl_lookup(SYSCTLFN_CALL(&node)); 1481 if (error || newp == NULL) 1482 return (error); 1483 1484 error = kauth_authorize_network(l->l_cred, KAUTH_NETWORK_FORWSRCRT, 1485 0, NULL, NULL, NULL); 1486 if (error) 1487 return (error); 1488 1489 ip_forwsrcrt = tmp; 1490 1491 return (0); 1492} 1493 1494/* 1495 * sysctl helper routine for net.inet.ip.mtudisctimeout. checks the 1496 * range of the new value and tweaks timers if it changes. 1497 */ 1498static int 1499sysctl_net_inet_ip_pmtudto(SYSCTLFN_ARGS) 1500{ 1501 int error, tmp; 1502 struct sysctlnode node; 1503 1504 node = *rnode; 1505 tmp = ip_mtudisc_timeout; 1506 node.sysctl_data = &tmp; 1507 error = sysctl_lookup(SYSCTLFN_CALL(&node)); 1508 if (error || newp == NULL) 1509 return (error); 1510 if (tmp < 0) 1511 return (EINVAL); 1512 1513 mutex_enter(softnet_lock); 1514 1515 ip_mtudisc_timeout = tmp; 1516 rt_timer_queue_change(ip_mtudisc_timeout_q, ip_mtudisc_timeout); 1517 1518 mutex_exit(softnet_lock); 1519 1520 return (0); 1521} 1522 1523#ifdef GATEWAY 1524/* 1525 * sysctl helper routine for net.inet.ip.maxflows. 1526 */ 1527static int 1528sysctl_net_inet_ip_maxflows(SYSCTLFN_ARGS) 1529{ 1530 int error; 1531 1532 error = sysctl_lookup(SYSCTLFN_CALL(rnode)); 1533 if (error || newp == NULL) 1534 return (error); 1535 1536 mutex_enter(softnet_lock); 1537 KERNEL_LOCK(1, NULL); 1538 1539 ipflow_prune(); 1540 1541 KERNEL_UNLOCK_ONE(NULL); 1542 mutex_exit(softnet_lock); 1543 1544 return (0); 1545} 1546 1547static int 1548sysctl_net_inet_ip_hashsize(SYSCTLFN_ARGS) 1549{ 1550 int error, tmp; 1551 struct sysctlnode node; 1552 1553 node = *rnode; 1554 tmp = ip_hashsize; 1555 node.sysctl_data = &tmp; 1556 error = sysctl_lookup(SYSCTLFN_CALL(&node)); 1557 if (error || newp == NULL) 1558 return (error); 1559 1560 if ((tmp & (tmp - 1)) == 0 && tmp != 0) { 1561 /* 1562 * Can only fail due to malloc() 1563 */ 1564 mutex_enter(softnet_lock); 1565 KERNEL_LOCK(1, NULL); 1566 1567 error = ipflow_invalidate_all(tmp); 1568 1569 KERNEL_UNLOCK_ONE(NULL); 1570 mutex_exit(softnet_lock); 1571 1572 } else { 1573 /* 1574 * EINVAL if not a power of 2 1575 */ 1576 error = EINVAL; 1577 } 1578 1579 return error; 1580} 1581#endif /* GATEWAY */ 1582 1583static int 1584sysctl_net_inet_ip_stats(SYSCTLFN_ARGS) 1585{ 1586 1587 return (NETSTAT_SYSCTL(ipstat_percpu, IP_NSTATS)); 1588} 1589 1590static void 1591sysctl_net_inet_ip_setup(struct sysctllog **clog) 1592{ 1593 extern int subnetsarelocal, hostzeroisbroadcast; 1594 1595 sysctl_createv(clog, 0, NULL, NULL, 1596 CTLFLAG_PERMANENT, 1597 CTLTYPE_NODE, "net", NULL, 1598 NULL, 0, NULL, 0, 1599 CTL_NET, CTL_EOL); 1600 sysctl_createv(clog, 0, NULL, NULL, 1601 CTLFLAG_PERMANENT, 1602 CTLTYPE_NODE, "inet", 1603 SYSCTL_DESCR("PF_INET related settings"), 1604 NULL, 0, NULL, 0, 1605 CTL_NET, PF_INET, CTL_EOL); 1606 sysctl_createv(clog, 0, NULL, NULL, 1607 CTLFLAG_PERMANENT, 1608 CTLTYPE_NODE, "ip", 1609 SYSCTL_DESCR("IPv4 related settings"), 1610 NULL, 0, NULL, 0, 1611 CTL_NET, PF_INET, IPPROTO_IP, CTL_EOL); 1612 1613 sysctl_createv(clog, 0, NULL, NULL, 1614 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1615 CTLTYPE_INT, "forwarding", 1616 SYSCTL_DESCR("Enable forwarding of INET datagrams"), 1617 NULL, 0, &ipforwarding, 0, 1618 CTL_NET, PF_INET, IPPROTO_IP, 1619 IPCTL_FORWARDING, CTL_EOL); 1620 sysctl_createv(clog, 0, NULL, NULL, 1621 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1622 CTLTYPE_INT, "redirect", 1623 SYSCTL_DESCR("Enable sending of ICMP redirect messages"), 1624 NULL, 0, &ipsendredirects, 0, 1625 CTL_NET, PF_INET, IPPROTO_IP, 1626 IPCTL_SENDREDIRECTS, CTL_EOL); 1627 sysctl_createv(clog, 0, NULL, NULL, 1628 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1629 CTLTYPE_INT, "ttl", 1630 SYSCTL_DESCR("Default TTL for an INET datagram"), 1631 NULL, 0, &ip_defttl, 0, 1632 CTL_NET, PF_INET, IPPROTO_IP, 1633 IPCTL_DEFTTL, CTL_EOL); 1634#ifdef IPCTL_DEFMTU 1635 sysctl_createv(clog, 0, NULL, NULL, 1636 CTLFLAG_PERMANENT /* |CTLFLAG_READWRITE? */, 1637 CTLTYPE_INT, "mtu", 1638 SYSCTL_DESCR("Default MTA for an INET route"), 1639 NULL, 0, &ip_mtu, 0, 1640 CTL_NET, PF_INET, IPPROTO_IP, 1641 IPCTL_DEFMTU, CTL_EOL); 1642#endif /* IPCTL_DEFMTU */ 1643 sysctl_createv(clog, 0, NULL, NULL, 1644 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1645 CTLTYPE_INT, "forwsrcrt", 1646 SYSCTL_DESCR("Enable forwarding of source-routed " 1647 "datagrams"), 1648 sysctl_net_inet_ip_forwsrcrt, 0, &ip_forwsrcrt, 0, 1649 CTL_NET, PF_INET, IPPROTO_IP, 1650 IPCTL_FORWSRCRT, CTL_EOL); 1651 sysctl_createv(clog, 0, NULL, NULL, 1652 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1653 CTLTYPE_INT, "directed-broadcast", 1654 SYSCTL_DESCR("Enable forwarding of broadcast datagrams"), 1655 NULL, 0, &ip_directedbcast, 0, 1656 CTL_NET, PF_INET, IPPROTO_IP, 1657 IPCTL_DIRECTEDBCAST, CTL_EOL); 1658 sysctl_createv(clog, 0, NULL, NULL, 1659 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1660 CTLTYPE_INT, "allowsrcrt", 1661 SYSCTL_DESCR("Accept source-routed datagrams"), 1662 NULL, 0, &ip_allowsrcrt, 0, 1663 CTL_NET, PF_INET, IPPROTO_IP, 1664 IPCTL_ALLOWSRCRT, CTL_EOL); 1665 sysctl_createv(clog, 0, NULL, NULL, 1666 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1667 CTLTYPE_INT, "subnetsarelocal", 1668 SYSCTL_DESCR("Whether logical subnets are considered " 1669 "local"), 1670 NULL, 0, &subnetsarelocal, 0, 1671 CTL_NET, PF_INET, IPPROTO_IP, 1672 IPCTL_SUBNETSARELOCAL, CTL_EOL); 1673 sysctl_createv(clog, 0, NULL, NULL, 1674 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1675 CTLTYPE_INT, "mtudisc", 1676 SYSCTL_DESCR("Use RFC1191 Path MTU Discovery"), 1677 NULL, 0, &ip_mtudisc, 0, 1678 CTL_NET, PF_INET, IPPROTO_IP, 1679 IPCTL_MTUDISC, CTL_EOL); 1680 sysctl_createv(clog, 0, NULL, NULL, 1681 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1682 CTLTYPE_INT, "anonportmin", 1683 SYSCTL_DESCR("Lowest ephemeral port number to assign"), 1684 sysctl_net_inet_ip_ports, 0, &anonportmin, 0, 1685 CTL_NET, PF_INET, IPPROTO_IP, 1686 IPCTL_ANONPORTMIN, CTL_EOL); 1687 sysctl_createv(clog, 0, NULL, NULL, 1688 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1689 CTLTYPE_INT, "anonportmax", 1690 SYSCTL_DESCR("Highest ephemeral port number to assign"), 1691 sysctl_net_inet_ip_ports, 0, &anonportmax, 0, 1692 CTL_NET, PF_INET, IPPROTO_IP, 1693 IPCTL_ANONPORTMAX, CTL_EOL); 1694 sysctl_createv(clog, 0, NULL, NULL, 1695 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1696 CTLTYPE_INT, "mtudisctimeout", 1697 SYSCTL_DESCR("Lifetime of a Path MTU Discovered route"), 1698 sysctl_net_inet_ip_pmtudto, 0, (void *)&ip_mtudisc_timeout, 0, 1699 CTL_NET, PF_INET, IPPROTO_IP, 1700 IPCTL_MTUDISCTIMEOUT, CTL_EOL); 1701#ifdef GATEWAY 1702 sysctl_createv(clog, 0, NULL, NULL, 1703 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1704 CTLTYPE_INT, "maxflows", 1705 SYSCTL_DESCR("Number of flows for fast forwarding"), 1706 sysctl_net_inet_ip_maxflows, 0, &ip_maxflows, 0, 1707 CTL_NET, PF_INET, IPPROTO_IP, 1708 IPCTL_MAXFLOWS, CTL_EOL); 1709 sysctl_createv(clog, 0, NULL, NULL, 1710 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1711 CTLTYPE_INT, "hashsize", 1712 SYSCTL_DESCR("Size of hash table for fast forwarding (IPv4)"), 1713 sysctl_net_inet_ip_hashsize, 0, &ip_hashsize, 0, 1714 CTL_NET, PF_INET, IPPROTO_IP, 1715 CTL_CREATE, CTL_EOL); 1716#endif /* GATEWAY */ 1717 sysctl_createv(clog, 0, NULL, NULL, 1718 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1719 CTLTYPE_INT, "hostzerobroadcast", 1720 SYSCTL_DESCR("All zeroes address is broadcast address"), 1721 NULL, 0, &hostzeroisbroadcast, 0, 1722 CTL_NET, PF_INET, IPPROTO_IP, 1723 IPCTL_HOSTZEROBROADCAST, CTL_EOL); 1724#if NGIF > 0 1725 sysctl_createv(clog, 0, NULL, NULL, 1726 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1727 CTLTYPE_INT, "gifttl", 1728 SYSCTL_DESCR("Default TTL for a gif tunnel datagram"), 1729 NULL, 0, &ip_gif_ttl, 0, 1730 CTL_NET, PF_INET, IPPROTO_IP, 1731 IPCTL_GIF_TTL, CTL_EOL); 1732#endif /* NGIF */ 1733#ifndef IPNOPRIVPORTS 1734 sysctl_createv(clog, 0, NULL, NULL, 1735 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1736 CTLTYPE_INT, "lowportmin", 1737 SYSCTL_DESCR("Lowest privileged ephemeral port number " 1738 "to assign"), 1739 sysctl_net_inet_ip_ports, 0, &lowportmin, 0, 1740 CTL_NET, PF_INET, IPPROTO_IP, 1741 IPCTL_LOWPORTMIN, CTL_EOL); 1742 sysctl_createv(clog, 0, NULL, NULL, 1743 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1744 CTLTYPE_INT, "lowportmax", 1745 SYSCTL_DESCR("Highest privileged ephemeral port number " 1746 "to assign"), 1747 sysctl_net_inet_ip_ports, 0, &lowportmax, 0, 1748 CTL_NET, PF_INET, IPPROTO_IP, 1749 IPCTL_LOWPORTMAX, CTL_EOL); 1750#endif /* IPNOPRIVPORTS */ 1751#if NGRE > 0 1752 sysctl_createv(clog, 0, NULL, NULL, 1753 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1754 CTLTYPE_INT, "grettl", 1755 SYSCTL_DESCR("Default TTL for a gre tunnel datagram"), 1756 NULL, 0, &ip_gre_ttl, 0, 1757 CTL_NET, PF_INET, IPPROTO_IP, 1758 IPCTL_GRE_TTL, CTL_EOL); 1759#endif /* NGRE */ 1760 sysctl_createv(clog, 0, NULL, NULL, 1761 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1762 CTLTYPE_INT, "checkinterface", 1763 SYSCTL_DESCR("Enable receive side of Strong ES model " 1764 "from RFC1122"), 1765 NULL, 0, &ip_checkinterface, 0, 1766 CTL_NET, PF_INET, IPPROTO_IP, 1767 IPCTL_CHECKINTERFACE, CTL_EOL); 1768 sysctl_createv(clog, 0, NULL, NULL, 1769 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1770 CTLTYPE_INT, "random_id", 1771 SYSCTL_DESCR("Assign random ip_id values"), 1772 NULL, 0, &ip_do_randomid, 0, 1773 CTL_NET, PF_INET, IPPROTO_IP, 1774 IPCTL_RANDOMID, CTL_EOL); 1775 sysctl_createv(clog, 0, NULL, NULL, 1776 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1777 CTLTYPE_INT, "do_loopback_cksum", 1778 SYSCTL_DESCR("Perform IP checksum on loopback"), 1779 NULL, 0, &ip_do_loopback_cksum, 0, 1780 CTL_NET, PF_INET, IPPROTO_IP, 1781 IPCTL_LOOPBACKCKSUM, CTL_EOL); 1782 sysctl_createv(clog, 0, NULL, NULL, 1783 CTLFLAG_PERMANENT, 1784 CTLTYPE_STRUCT, "stats", 1785 SYSCTL_DESCR("IP statistics"), 1786 sysctl_net_inet_ip_stats, 0, NULL, 0, 1787 CTL_NET, PF_INET, IPPROTO_IP, IPCTL_STATS, 1788 CTL_EOL); 1789 1790 /* anonportalgo RFC6056 subtree */ 1791 const struct sysctlnode *portalgo_node; 1792 sysctl_createv(clog, 0, NULL, &portalgo_node, 1793 CTLFLAG_PERMANENT, 1794 CTLTYPE_NODE, "anonportalgo", 1795 SYSCTL_DESCR("Anonymous Port Algorithm Selection (RFC 6056)"), 1796 NULL, 0, NULL, 0, 1797 CTL_NET, PF_INET, IPPROTO_IP, CTL_CREATE, CTL_EOL); 1798 sysctl_createv(clog, 0, &portalgo_node, NULL, 1799 CTLFLAG_PERMANENT, 1800 CTLTYPE_STRING, "available", 1801 SYSCTL_DESCR("available algorithms"), 1802 sysctl_portalgo_available, 0, NULL, PORTALGO_MAXLEN, 1803 CTL_CREATE, CTL_EOL); 1804 sysctl_createv(clog, 0, &portalgo_node, NULL, 1805 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1806 CTLTYPE_STRING, "selected", 1807 SYSCTL_DESCR("selected algorithm"), 1808 sysctl_portalgo_selected4, 0, NULL, PORTALGO_MAXLEN, 1809 CTL_CREATE, CTL_EOL); 1810 sysctl_createv(clog, 0, &portalgo_node, NULL, 1811 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 1812 CTLTYPE_STRUCT, "reserve", 1813 SYSCTL_DESCR("bitmap of reserved ports"), 1814 sysctl_portalgo_reserve4, 0, NULL, 0, 1815 CTL_CREATE, CTL_EOL); 1816} 1817 1818void 1819ip_statinc(u_int stat) 1820{ 1821 1822 KASSERT(stat < IP_NSTATS); 1823 IP_STATINC(stat); 1824} 1825