Annotation of src/sys/netinet/ip_input.c, Revision 1.294
1.294 ! dyoung 1: /* $NetBSD: ip_input.c,v 1.293 2010/12/13 14:18:50 matt Exp $ */
1.89 itojun 2:
3: /*
4: * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
5: * All rights reserved.
1.152 itojun 6: *
1.89 itojun 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.
1.152 itojun 18: *
1.89 itojun 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: */
1.76 thorpej 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: */
1.14 cgd 61:
1.1 cgd 62: /*
1.13 mycroft 63: * Copyright (c) 1982, 1986, 1988, 1993
64: * The Regents of the University of California. All rights reserved.
1.1 cgd 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.
1.172 agc 74: * 3. Neither the name of the University nor the names of its contributors
1.1 cgd 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: *
1.14 cgd 90: * @(#)ip_input.c 8.2 (Berkeley) 1/4/94
1.1 cgd 91: */
1.141 lukem 92:
93: #include <sys/cdefs.h>
1.294 ! dyoung 94: __KERNEL_RCSID(0, "$NetBSD: ip_input.c,v 1.293 2010/12/13 14:18:50 matt Exp $");
1.55 scottr 95:
1.184 jonathan 96: #include "opt_inet.h"
1.278 christos 97: #include "opt_compat_netbsd.h"
1.62 matt 98: #include "opt_gateway.h"
1.69 mrg 99: #include "opt_pfil_hooks.h"
1.91 thorpej 100: #include "opt_ipsec.h"
1.55 scottr 101: #include "opt_mrouting.h"
1.167 martin 102: #include "opt_mbuftrace.h"
1.135 thorpej 103: #include "opt_inet_csum.h"
1.1 cgd 104:
1.5 mycroft 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>
1.44 thorpej 111: #include <sys/socketvar.h>
1.5 mycroft 112: #include <sys/errno.h>
113: #include <sys/time.h>
114: #include <sys/kernel.h>
1.72 thorpej 115: #include <sys/pool.h>
1.28 christos 116: #include <sys/sysctl.h>
1.230 elad 117: #include <sys/kauth.h>
1.1 cgd 118:
1.5 mycroft 119: #include <net/if.h>
1.44 thorpej 120: #include <net/if_dl.h>
1.5 mycroft 121: #include <net/route.h>
1.45 mrg 122: #include <net/pfil.h>
1.1 cgd 123:
1.5 mycroft 124: #include <netinet/in.h>
125: #include <netinet/in_systm.h>
126: #include <netinet/ip.h>
127: #include <netinet/in_pcb.h>
1.215 yamt 128: #include <netinet/in_proto.h>
1.5 mycroft 129: #include <netinet/in_var.h>
130: #include <netinet/ip_var.h>
1.266 thorpej 131: #include <netinet/ip_private.h>
1.5 mycroft 132: #include <netinet/ip_icmp.h>
1.89 itojun 133: /* just for gif_ttl */
134: #include <netinet/in_gif.h>
135: #include "gif.h"
1.144 martin 136: #include <net/if_gre.h>
137: #include "gre.h"
1.111 jdolecek 138:
139: #ifdef MROUTING
140: #include <netinet/ip_mroute.h>
141: #endif
1.89 itojun 142:
143: #ifdef IPSEC
144: #include <netinet6/ipsec.h>
1.267 thorpej 145: #include <netinet6/ipsec_private.h>
1.89 itojun 146: #include <netkey/key.h>
147: #endif
1.173 jonathan 148: #ifdef FAST_IPSEC
149: #include <netipsec/ipsec.h>
150: #include <netipsec/key.h>
151: #endif /* FAST_IPSEC*/
1.44 thorpej 152:
1.1 cgd 153: #ifndef IPFORWARDING
154: #ifdef GATEWAY
155: #define IPFORWARDING 1 /* forward IP packets not for us */
156: #else /* GATEWAY */
157: #define IPFORWARDING 0 /* don't forward IP packets not for us */
158: #endif /* GATEWAY */
159: #endif /* IPFORWARDING */
160: #ifndef IPSENDREDIRECTS
161: #define IPSENDREDIRECTS 1
162: #endif
1.26 thorpej 163: #ifndef IPFORWSRCRT
1.47 cjs 164: #define IPFORWSRCRT 1 /* forward source-routed packets */
165: #endif
166: #ifndef IPALLOWSRCRT
1.48 mrg 167: #define IPALLOWSRCRT 1 /* allow source-routed packets */
1.26 thorpej 168: #endif
1.53 kml 169: #ifndef IPMTUDISC
1.153 itojun 170: #define IPMTUDISC 1
1.53 kml 171: #endif
1.60 kml 172: #ifndef IPMTUDISCTIMEOUT
1.61 kml 173: #define IPMTUDISCTIMEOUT (10 * 60) /* as per RFC 1191 */
1.60 kml 174: #endif
1.53 kml 175:
1.278 christos 176: #ifdef COMPAT_50
177: #include <compat/sys/time.h>
178: #include <compat/sys/socket.h>
179: #endif
180:
1.27 thorpej 181: /*
182: * Note: DIRECTED_BROADCAST is handled this way so that previous
183: * configuration using this option will Just Work.
184: */
185: #ifndef IPDIRECTEDBCAST
186: #ifdef DIRECTED_BROADCAST
187: #define IPDIRECTEDBCAST 1
188: #else
189: #define IPDIRECTEDBCAST 0
190: #endif /* DIRECTED_BROADCAST */
191: #endif /* IPDIRECTEDBCAST */
1.1 cgd 192: int ipforwarding = IPFORWARDING;
193: int ipsendredirects = IPSENDREDIRECTS;
1.13 mycroft 194: int ip_defttl = IPDEFTTL;
1.26 thorpej 195: int ip_forwsrcrt = IPFORWSRCRT;
1.27 thorpej 196: int ip_directedbcast = IPDIRECTEDBCAST;
1.47 cjs 197: int ip_allowsrcrt = IPALLOWSRCRT;
1.53 kml 198: int ip_mtudisc = IPMTUDISC;
1.156 itojun 199: int ip_mtudisc_timeout = IPMTUDISCTIMEOUT;
1.1 cgd 200: #ifdef DIAGNOSTIC
201: int ipprintfs = 0;
202: #endif
1.184 jonathan 203:
204: int ip_do_randomid = 0;
205:
1.165 christos 206: /*
207: * XXX - Setting ip_checkinterface mostly implements the receive side of
208: * the Strong ES model described in RFC 1122, but since the routing table
209: * and transmit implementation do not implement the Strong ES model,
210: * setting this to 1 results in an odd hybrid.
211: *
212: * XXX - ip_checkinterface currently must be disabled if you use ipnat
213: * to translate the destination address to another local interface.
214: *
215: * XXX - ip_checkinterface must be disabled if you add IP aliases
216: * to the loopback interface instead of the interface where the
217: * packets for those addresses are received.
218: */
219: int ip_checkinterface = 0;
220:
1.1 cgd 221:
1.60 kml 222: struct rttimer_queue *ip_mtudisc_timeout_q = NULL;
223:
1.150 matt 224: u_long in_ifaddrhash; /* size of hash table - 1 */
225: int in_ifaddrentries; /* total number of addrs */
1.212 perry 226: struct in_ifaddrhead in_ifaddrhead;
1.57 tls 227: struct in_ifaddrhashhead *in_ifaddrhashtbl;
1.166 matt 228: u_long in_multihash; /* size of hash table - 1 */
229: int in_multientries; /* total number of addrs */
230: struct in_multihashhead *in_multihashtbl;
1.13 mycroft 231: struct ifqueue ipintrq;
1.286 tls 232:
1.291 rmind 233: ipid_state_t * ip_ids;
1.183 jonathan 234: uint16_t ip_id;
1.75 thorpej 235:
1.266 thorpej 236: percpu_t *ipstat_percpu;
237:
1.121 thorpej 238: #ifdef PFIL_HOOKS
239: struct pfil_head inet_pfil_hook;
240: #endif
241:
1.275 pooka 242: struct pool inmulti_pool;
1.72 thorpej 243:
1.135 thorpej 244: #ifdef INET_CSUM_COUNTERS
245: #include <sys/device.h>
246:
247: struct evcnt ip_hwcsum_bad = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
248: NULL, "inet", "hwcsum bad");
249: struct evcnt ip_hwcsum_ok = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
250: NULL, "inet", "hwcsum ok");
251: struct evcnt ip_swcsum = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
252: NULL, "inet", "swcsum");
253:
254: #define INET_CSUM_COUNTER_INCR(ev) (ev)->ev_count++
255:
1.201 matt 256: EVCNT_ATTACH_STATIC(ip_hwcsum_bad);
257: EVCNT_ATTACH_STATIC(ip_hwcsum_ok);
258: EVCNT_ATTACH_STATIC(ip_swcsum);
259:
1.135 thorpej 260: #else
261:
262: #define INET_CSUM_COUNTER_INCR(ev) /* nothing */
263:
264: #endif /* INET_CSUM_COUNTERS */
265:
1.1 cgd 266: /*
267: * We need to save the IP options in case a protocol wants to respond
268: * to an incoming packet over the same route if the packet got here
269: * using IP source routing. This allows connection establishment and
270: * maintenance when the remote end is on a network that is not known
271: * to us.
272: */
273: int ip_nhops = 0;
274: static struct ip_srcrt {
275: struct in_addr dst; /* final destination */
276: char nop; /* one NOP to align */
277: char srcopt[IPOPT_OFFSET + 1]; /* OPTVAL, OLEN and OFFSET */
278: struct in_addr route[MAX_IPOPTLEN/sizeof(struct in_addr)];
279: } ip_srcrt;
280:
1.210 perry 281: static void save_rte(u_char *, struct in_addr);
1.35 mycroft 282:
1.164 matt 283: #ifdef MBUFTRACE
1.234 dogcow 284: struct mowner ip_rx_mowner = MOWNER_INIT("internet", "rx");
285: struct mowner ip_tx_mowner = MOWNER_INIT("internet", "tx");
1.164 matt 286: #endif
287:
1.284 pooka 288: static void sysctl_net_inet_ip_setup(struct sysctllog **);
289:
1.1 cgd 290: /*
291: * IP initialization: fill in IP protocol switch table.
292: * All protocols not implemented in kernel go to raw IP protocol handler.
293: */
1.8 mycroft 294: void
1.211 perry 295: ip_init(void)
1.1 cgd 296: {
1.199 matt 297: const struct protosw *pr;
1.109 augustss 298: int i;
1.1 cgd 299:
1.284 pooka 300: sysctl_net_inet_ip_setup(NULL);
301:
1.275 pooka 302: pool_init(&inmulti_pool, sizeof(struct in_multi), 0, 0, 0, "inmltpl",
303: NULL, IPL_SOFTNET);
304:
1.1 cgd 305: pr = pffindproto(PF_INET, IPPROTO_RAW, SOCK_RAW);
306: if (pr == 0)
307: panic("ip_init");
308: for (i = 0; i < IPPROTO_MAX; i++)
309: ip_protox[i] = pr - inetsw;
310: for (pr = inetdomain.dom_protosw;
311: pr < inetdomain.dom_protoswNPROTOSW; pr++)
312: if (pr->pr_domain->dom_family == PF_INET &&
313: pr->pr_protocol && pr->pr_protocol != IPPROTO_RAW)
314: ip_protox[pr->pr_protocol] = pr - inetsw;
1.192 jonathan 315:
1.288 rmind 316: ip_reass_init();
1.190 jonathan 317:
1.291 rmind 318: ip_ids = ip_id_init();
1.227 kardel 319: ip_id = time_second & 0xfffff;
1.194 jonathan 320:
1.294 ! dyoung 321: ipintrq.ifq_maxlen = IFQ_MAXLEN;
1.194 jonathan 322:
1.181 jonathan 323: TAILQ_INIT(&in_ifaddrhead);
1.272 ad 324: in_ifaddrhashtbl = hashinit(IN_IFADDR_HASH_SIZE, HASH_LIST, true,
325: &in_ifaddrhash);
326: in_multihashtbl = hashinit(IN_IFADDR_HASH_SIZE, HASH_LIST, true,
327: &in_multihash);
1.160 itojun 328: ip_mtudisc_timeout_q = rt_timer_queue_create(ip_mtudisc_timeout);
1.73 thorpej 329: #ifdef GATEWAY
1.248 liamjfoy 330: ipflow_init(ip_hashsize);
1.73 thorpej 331: #endif
1.121 thorpej 332:
333: #ifdef PFIL_HOOKS
334: /* Register our Packet Filter hook. */
1.126 thorpej 335: inet_pfil_hook.ph_type = PFIL_TYPE_AF;
336: inet_pfil_hook.ph_af = AF_INET;
1.121 thorpej 337: i = pfil_head_register(&inet_pfil_hook);
338: if (i != 0)
339: printf("ip_init: WARNING: unable to register pfil hook, "
340: "error %d\n", i);
341: #endif /* PFIL_HOOKS */
1.135 thorpej 342:
1.164 matt 343: #ifdef MBUFTRACE
344: MOWNER_ATTACH(&ip_tx_mowner);
345: MOWNER_ATTACH(&ip_rx_mowner);
346: #endif /* MBUFTRACE */
1.266 thorpej 347:
348: ipstat_percpu = percpu_alloc(sizeof(uint64_t) * IP_NSTATS);
1.1 cgd 349: }
350:
1.229 christos 351: struct sockaddr_in ipaddr = {
352: .sin_len = sizeof(ipaddr),
353: .sin_family = AF_INET,
354: };
1.1 cgd 355: struct route ipforward_rt;
356:
357: /*
1.89 itojun 358: * IP software interrupt routine
359: */
360: void
1.211 perry 361: ipintr(void)
1.89 itojun 362: {
363: int s;
364: struct mbuf *m;
1.286 tls 365: struct ifqueue lcl_intrq;
366:
367: memset(&lcl_intrq, 0, sizeof(lcl_intrq));
1.89 itojun 368:
1.268 ad 369: mutex_enter(softnet_lock);
370: KERNEL_LOCK(1, NULL);
1.286 tls 371: if (!IF_IS_EMPTY(&ipintrq)) {
1.132 thorpej 372: s = splnet();
1.286 tls 373:
374: /* Take existing queue onto stack */
375: lcl_intrq = ipintrq;
376:
377: /* Zero out global queue, preserving maxlen and drops */
378: ipintrq.ifq_head = NULL;
379: ipintrq.ifq_tail = NULL;
380: ipintrq.ifq_len = 0;
381: ipintrq.ifq_maxlen = lcl_intrq.ifq_maxlen;
382: ipintrq.ifq_drops = lcl_intrq.ifq_drops;
383:
1.89 itojun 384: splx(s);
1.286 tls 385: }
386: KERNEL_UNLOCK_ONE(NULL);
387: while (!IF_IS_EMPTY(&lcl_intrq)) {
388: IF_DEQUEUE(&lcl_intrq, m);
1.268 ad 389: if (m == NULL)
390: break;
1.89 itojun 391: ip_input(m);
392: }
1.268 ad 393: mutex_exit(softnet_lock);
1.89 itojun 394: }
395:
396: /*
1.1 cgd 397: * Ip input routine. Checksum and byte swap header. If fragmented
398: * try to reassemble. Process options. Pass to next level.
399: */
1.8 mycroft 400: void
1.89 itojun 401: ip_input(struct mbuf *m)
1.1 cgd 402: {
1.109 augustss 403: struct ip *ip = NULL;
404: struct in_ifaddr *ia;
405: struct ifaddr *ifa;
1.288 rmind 406: int hlen = 0, len;
1.100 itojun 407: int downmatch;
1.165 christos 408: int checkif;
1.169 itojun 409: int srcrt = 0;
1.173 jonathan 410: #ifdef FAST_IPSEC
411: struct m_tag *mtag;
412: struct tdb_ident *tdbi;
413: struct secpolicy *sp;
1.289 rmind 414: int error, s;
1.173 jonathan 415: #endif /* FAST_IPSEC */
1.1 cgd 416:
1.164 matt 417: MCLAIM(m, &ip_rx_mowner);
1.289 rmind 418: KASSERT((m->m_flags & M_PKTHDR) != 0);
1.164 matt 419:
1.1 cgd 420: /*
421: * If no IP addresses have been set yet but the interfaces
422: * are receiving, can't do anything with incoming packets yet.
423: */
1.181 jonathan 424: if (TAILQ_FIRST(&in_ifaddrhead) == 0)
1.1 cgd 425: goto bad;
1.266 thorpej 426: IP_STATINC(IP_STAT_TOTAL);
1.154 thorpej 427: /*
428: * If the IP header is not aligned, slurp it up into a new
429: * mbuf with space for link headers, in the event we forward
430: * it. Otherwise, if it is aligned, make sure the entire
431: * base IP header is in the first mbuf of the chain.
432: */
1.244 christos 433: if (IP_HDR_ALIGNED_P(mtod(m, void *)) == 0) {
1.154 thorpej 434: if ((m = m_copyup(m, sizeof(struct ip),
435: (max_linkhdr + 3) & ~3)) == NULL) {
436: /* XXXJRT new stat, please */
1.266 thorpej 437: IP_STATINC(IP_STAT_TOOSMALL);
1.154 thorpej 438: return;
439: }
440: } else if (__predict_false(m->m_len < sizeof (struct ip))) {
441: if ((m = m_pullup(m, sizeof (struct ip))) == NULL) {
1.266 thorpej 442: IP_STATINC(IP_STAT_TOOSMALL);
1.154 thorpej 443: return;
444: }
1.1 cgd 445: }
446: ip = mtod(m, struct ip *);
1.13 mycroft 447: if (ip->ip_v != IPVERSION) {
1.266 thorpej 448: IP_STATINC(IP_STAT_BADVERS);
1.13 mycroft 449: goto bad;
450: }
1.1 cgd 451: hlen = ip->ip_hl << 2;
452: if (hlen < sizeof(struct ip)) { /* minimum header length */
1.266 thorpej 453: IP_STATINC(IP_STAT_BADHLEN);
1.1 cgd 454: goto bad;
455: }
456: if (hlen > m->m_len) {
457: if ((m = m_pullup(m, hlen)) == 0) {
1.266 thorpej 458: IP_STATINC(IP_STAT_BADHLEN);
1.89 itojun 459: return;
1.1 cgd 460: }
461: ip = mtod(m, struct ip *);
462: }
1.98 thorpej 463:
1.85 hwr 464: /*
1.99 thorpej 465: * RFC1122: packets with a multicast source address are
1.98 thorpej 466: * not allowed.
1.85 hwr 467: */
468: if (IN_MULTICAST(ip->ip_src.s_addr)) {
1.266 thorpej 469: IP_STATINC(IP_STAT_BADADDR);
1.85 hwr 470: goto bad;
1.129 itojun 471: }
472:
473: /* 127/8 must not appear on wire - RFC1122 */
474: if ((ntohl(ip->ip_dst.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET ||
475: (ntohl(ip->ip_src.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET) {
1.130 itojun 476: if ((m->m_pkthdr.rcvif->if_flags & IFF_LOOPBACK) == 0) {
1.266 thorpej 477: IP_STATINC(IP_STAT_BADADDR);
1.129 itojun 478: goto bad;
1.130 itojun 479: }
1.85 hwr 480: }
481:
1.135 thorpej 482: switch (m->m_pkthdr.csum_flags &
1.137 thorpej 483: ((m->m_pkthdr.rcvif->if_csum_flags_rx & M_CSUM_IPv4) |
1.135 thorpej 484: M_CSUM_IPv4_BAD)) {
485: case M_CSUM_IPv4|M_CSUM_IPv4_BAD:
486: INET_CSUM_COUNTER_INCR(&ip_hwcsum_bad);
487: goto badcsum;
488:
489: case M_CSUM_IPv4:
490: /* Checksum was okay. */
491: INET_CSUM_COUNTER_INCR(&ip_hwcsum_ok);
492: break;
493:
494: default:
1.206 thorpej 495: /*
496: * Must compute it ourselves. Maybe skip checksum on
497: * loopback interfaces.
498: */
499: if (__predict_true(!(m->m_pkthdr.rcvif->if_flags &
500: IFF_LOOPBACK) || ip_do_loopback_cksum)) {
501: INET_CSUM_COUNTER_INCR(&ip_swcsum);
502: if (in_cksum(m, hlen) != 0)
503: goto badcsum;
504: }
1.135 thorpej 505: break;
1.1 cgd 506: }
507:
1.121 thorpej 508: /* Retrieve the packet length. */
509: len = ntohs(ip->ip_len);
1.81 proff 510:
511: /*
512: * Check for additional length bogosity
513: */
1.84 proff 514: if (len < hlen) {
1.266 thorpej 515: IP_STATINC(IP_STAT_BADLEN);
1.81 proff 516: goto bad;
517: }
1.1 cgd 518:
519: /*
520: * Check that the amount of data in the buffers
521: * is as at least much as the IP header would have us expect.
522: * Trim mbufs if longer than we expect.
523: * Drop packet if shorter than we expect.
524: */
1.35 mycroft 525: if (m->m_pkthdr.len < len) {
1.266 thorpej 526: IP_STATINC(IP_STAT_TOOSHORT);
1.1 cgd 527: goto bad;
528: }
1.35 mycroft 529: if (m->m_pkthdr.len > len) {
1.1 cgd 530: if (m->m_len == m->m_pkthdr.len) {
1.35 mycroft 531: m->m_len = len;
532: m->m_pkthdr.len = len;
1.1 cgd 533: } else
1.35 mycroft 534: m_adj(m, len - m->m_pkthdr.len);
1.1 cgd 535: }
536:
1.193 scw 537: #if defined(IPSEC)
1.149 wiz 538: /* ipflow (IP fast forwarding) is not compatible with IPsec. */
1.94 itojun 539: m->m_flags &= ~M_CANFASTFWD;
540: #else
1.64 thorpej 541: /*
542: * Assume that we can create a fast-forward IP flow entry
543: * based on this packet.
544: */
545: m->m_flags |= M_CANFASTFWD;
1.94 itojun 546: #endif
1.64 thorpej 547:
1.36 mrg 548: #ifdef PFIL_HOOKS
1.33 mrg 549: /*
1.64 thorpej 550: * Run through list of hooks for input packets. If there are any
551: * filters which require that additional packets in the flow are
552: * not fast-forwarded, they must clear the M_CANFASTFWD flag.
553: * Note that filters must _never_ set this flag, as another filter
554: * in the list may have previously cleared it.
1.33 mrg 555: */
1.127 itojun 556: /*
557: * let ipfilter look at packet on the wire,
558: * not the decapsulated packet.
559: */
560: #ifdef IPSEC
1.136 itojun 561: if (!ipsec_getnhist(m))
1.186 scw 562: #elif defined(FAST_IPSEC)
563: if (!ipsec_indone(m))
1.127 itojun 564: #else
565: if (1)
566: #endif
567: {
1.169 itojun 568: struct in_addr odst;
569:
570: odst = ip->ip_dst;
1.127 itojun 571: if (pfil_run_hooks(&inet_pfil_hook, &m, m->m_pkthdr.rcvif,
1.168 itojun 572: PFIL_IN) != 0)
573: return;
1.127 itojun 574: if (m == NULL)
575: return;
576: ip = mtod(m, struct ip *);
1.142 darrenr 577: hlen = ip->ip_hl << 2;
1.205 darrenr 578: /*
579: * XXX The setting of "srcrt" here is to prevent ip_forward()
580: * from generating ICMP redirects for packets that have
581: * been redirected by a hook back out on to the same LAN that
582: * they came from and is not an indication that the packet
583: * is being inffluenced by source routing options. This
584: * allows things like
585: * "rdr tlp0 0/0 port 80 -> 1.1.1.200 3128 tcp"
586: * where tlp0 is both on the 1.1.1.0/24 network and is the
587: * default route for hosts on 1.1.1.0/24. Of course this
588: * also requires a "map tlp0 ..." to complete the story.
589: * One might argue whether or not this kind of network config.
1.212 perry 590: * should be supported in this manner...
1.205 darrenr 591: */
1.169 itojun 592: srcrt = (odst.s_addr != ip->ip_dst.s_addr);
1.127 itojun 593: }
1.36 mrg 594: #endif /* PFIL_HOOKS */
1.123 thorpej 595:
596: #ifdef ALTQ
597: /* XXX Temporary until ALTQ is changed to use a pfil hook */
598: if (altq_input != NULL && (*altq_input)(m, AF_INET) == 0) {
599: /* packet dropped by traffic conditioner */
600: return;
601: }
602: #endif
1.121 thorpej 603:
604: /*
1.1 cgd 605: * Process options and, if not destined for us,
606: * ship it on. ip_dooptions returns 1 when an
607: * error was detected (causing an icmp message
608: * to be sent and the original packet to be freed).
609: */
610: ip_nhops = 0; /* for source routed packets */
611: if (hlen > sizeof (struct ip) && ip_dooptions(m))
1.89 itojun 612: return;
1.1 cgd 613:
614: /*
1.165 christos 615: * Enable a consistency check between the destination address
616: * and the arrival interface for a unicast packet (the RFC 1122
617: * strong ES model) if IP forwarding is disabled and the packet
618: * is not locally generated.
619: *
620: * XXX - Checking also should be disabled if the destination
621: * address is ipnat'ed to a different interface.
622: *
623: * XXX - Checking is incompatible with IP aliases added
624: * to the loopback interface instead of the interface where
625: * the packets are received.
626: *
627: * XXX - We need to add a per ifaddr flag for this so that
628: * we get finer grain control.
629: */
630: checkif = ip_checkinterface && (ipforwarding == 0) &&
631: (m->m_pkthdr.rcvif != NULL) &&
632: ((m->m_pkthdr.rcvif->if_flags & IFF_LOOPBACK) == 0);
633:
634: /*
1.1 cgd 635: * Check our list of addresses, to see if the packet is for us.
1.100 itojun 636: *
637: * Traditional 4.4BSD did not consult IFF_UP at all.
638: * The behavior here is to treat addresses on !IFF_UP interface
639: * as not mine.
1.1 cgd 640: */
1.100 itojun 641: downmatch = 0;
1.140 matt 642: LIST_FOREACH(ia, &IN_IFADDR_HASH(ip->ip_dst.s_addr), ia_hash) {
1.97 itojun 643: if (in_hosteq(ia->ia_addr.sin_addr, ip->ip_dst)) {
1.165 christos 644: if (checkif && ia->ia_ifp != m->m_pkthdr.rcvif)
645: continue;
1.97 itojun 646: if ((ia->ia_ifp->if_flags & IFF_UP) != 0)
647: break;
1.100 itojun 648: else
649: downmatch++;
1.97 itojun 650: }
651: }
1.86 thorpej 652: if (ia != NULL)
653: goto ours;
1.225 christos 654: if (m->m_pkthdr.rcvif && m->m_pkthdr.rcvif->if_flags & IFF_BROADCAST) {
1.209 matt 655: IFADDR_FOREACH(ifa, m->m_pkthdr.rcvif) {
1.140 matt 656: if (ifa->ifa_addr->sa_family != AF_INET)
657: continue;
1.57 tls 658: ia = ifatoia(ifa);
1.35 mycroft 659: if (in_hosteq(ip->ip_dst, ia->ia_broadaddr.sin_addr) ||
660: in_hosteq(ip->ip_dst, ia->ia_netbroadcast) ||
1.20 mycroft 661: /*
662: * Look for all-0's host part (old broadcast addr),
663: * either for subnet or net.
664: */
665: ip->ip_dst.s_addr == ia->ia_subnet ||
1.18 mycroft 666: ip->ip_dst.s_addr == ia->ia_net)
1.1 cgd 667: goto ours;
1.57 tls 668: /*
669: * An interface with IP address zero accepts
670: * all packets that arrive on that interface.
671: */
672: if (in_nullhost(ia->ia_addr.sin_addr))
673: goto ours;
1.1 cgd 674: }
675: }
1.18 mycroft 676: if (IN_MULTICAST(ip->ip_dst.s_addr)) {
1.4 hpeyerl 677: struct in_multi *inm;
678: #ifdef MROUTING
679: extern struct socket *ip_mrouter;
1.10 brezak 680:
1.4 hpeyerl 681: if (ip_mrouter) {
682: /*
683: * If we are acting as a multicast router, all
684: * incoming multicast packets are passed to the
685: * kernel-level multicast forwarding function.
686: * The packet is returned (relatively) intact; if
687: * ip_mforward() returns a non-zero value, the packet
688: * must be discarded, else it may be accepted below.
689: *
690: * (The IP ident field is put in the same byte order
691: * as expected when ip_mforward() is called from
692: * ip_output().)
693: */
1.13 mycroft 694: if (ip_mforward(m, m->m_pkthdr.rcvif) != 0) {
1.266 thorpej 695: IP_STATINC(IP_STAT_CANTFORWARD);
1.4 hpeyerl 696: m_freem(m);
1.89 itojun 697: return;
1.4 hpeyerl 698: }
699:
700: /*
701: * The process-level routing demon needs to receive
702: * all multicast IGMP packets, whether or not this
703: * host belongs to their destination groups.
704: */
705: if (ip->ip_p == IPPROTO_IGMP)
706: goto ours;
1.266 thorpej 707: IP_STATINC(IP_STAT_CANTFORWARD);
1.4 hpeyerl 708: }
709: #endif
710: /*
711: * See if we belong to the destination multicast group on the
712: * arrival interface.
713: */
714: IN_LOOKUP_MULTI(ip->ip_dst, m->m_pkthdr.rcvif, inm);
715: if (inm == NULL) {
1.266 thorpej 716: IP_STATINC(IP_STAT_CANTFORWARD);
1.4 hpeyerl 717: m_freem(m);
1.89 itojun 718: return;
1.4 hpeyerl 719: }
720: goto ours;
721: }
1.19 mycroft 722: if (ip->ip_dst.s_addr == INADDR_BROADCAST ||
1.35 mycroft 723: in_nullhost(ip->ip_dst))
1.1 cgd 724: goto ours;
725:
726: /*
727: * Not for us; forward if possible and desirable.
728: */
729: if (ipforwarding == 0) {
1.266 thorpej 730: IP_STATINC(IP_STAT_CANTFORWARD);
1.1 cgd 731: m_freem(m);
1.100 itojun 732: } else {
733: /*
734: * If ip_dst matched any of my address on !IFF_UP interface,
735: * and there's no IFF_UP interface that matches ip_dst,
736: * send icmp unreach. Forwarding it will result in in-kernel
737: * forwarding loop till TTL goes to 0.
738: */
739: if (downmatch) {
740: icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_HOST, 0, 0);
1.266 thorpej 741: IP_STATINC(IP_STAT_CANTFORWARD);
1.100 itojun 742: return;
743: }
1.145 itojun 744: #ifdef IPSEC
745: if (ipsec4_in_reject(m, NULL)) {
1.267 thorpej 746: IPSEC_STATINC(IPSEC_STAT_IN_POLVIO);
1.145 itojun 747: goto bad;
748: }
749: #endif
1.173 jonathan 750: #ifdef FAST_IPSEC
751: mtag = m_tag_find(m, PACKET_TAG_IPSEC_IN_DONE, NULL);
752: s = splsoftnet();
753: if (mtag != NULL) {
754: tdbi = (struct tdb_ident *)(mtag + 1);
755: sp = ipsec_getpolicy(tdbi, IPSEC_DIR_INBOUND);
756: } else {
757: sp = ipsec_getpolicybyaddr(m, IPSEC_DIR_INBOUND,
1.212 perry 758: IP_FORWARDING, &error);
1.173 jonathan 759: }
760: if (sp == NULL) { /* NB: can happen if error */
761: splx(s);
762: /*XXX error stat???*/
763: DPRINTF(("ip_input: no SP for forwarding\n")); /*XXX*/
764: goto bad;
765: }
766:
767: /*
768: * Check security policy against packet attributes.
769: */
770: error = ipsec_in_reject(sp, m);
771: KEY_FREESP(&sp);
772: splx(s);
773: if (error) {
1.266 thorpej 774: IP_STATINC(IP_STAT_CANTFORWARD);
1.173 jonathan 775: goto bad;
1.193 scw 776: }
777:
778: /*
779: * Peek at the outbound SP for this packet to determine if
780: * it's a Fast Forward candidate.
781: */
782: mtag = m_tag_find(m, PACKET_TAG_IPSEC_PENDING_TDB, NULL);
783: if (mtag != NULL)
784: m->m_flags &= ~M_CANFASTFWD;
785: else {
786: s = splsoftnet();
787: sp = ipsec4_checkpolicy(m, IPSEC_DIR_OUTBOUND,
788: (IP_FORWARDING |
789: (ip_directedbcast ? IP_ALLOWBROADCAST : 0)),
790: &error, NULL);
791: if (sp != NULL) {
792: m->m_flags &= ~M_CANFASTFWD;
793: KEY_FREESP(&sp);
794: }
795: splx(s);
1.173 jonathan 796: }
797: #endif /* FAST_IPSEC */
1.145 itojun 798:
1.169 itojun 799: ip_forward(m, srcrt);
1.100 itojun 800: }
1.89 itojun 801: return;
1.1 cgd 802:
803: ours:
804: /*
805: * If offset or IP_MF are set, must reassemble.
806: */
1.155 itojun 807: if (ip->ip_off & ~htons(IP_DF|IP_RF)) {
1.289 rmind 808: /*
809: * Pass to IP reassembly mechanism.
810: */
1.290 rmind 811: if (ip_reass_packet(&m, ip) != 0) {
1.289 rmind 812: /* Failed; invalid fragment(s) or packet. */
1.288 rmind 813: goto bad;
1.16 cgd 814: }
1.290 rmind 815: if (m == NULL) {
1.289 rmind 816: /* More fragments should come; silently return. */
817: return;
818: }
1.290 rmind 819: /*
820: * Reassembly is done, we have the final packet.
821: * Updated cached data in local variable(s).
822: */
1.289 rmind 823: ip = mtod(m, struct ip *);
824: hlen = ip->ip_hl << 2;
1.79 mycroft 825: }
1.128 itojun 826:
1.173 jonathan 827: #if defined(IPSEC)
1.128 itojun 828: /*
829: * enforce IPsec policy checking if we are seeing last header.
830: * note that we do not visit this with protocols with pcb layer
831: * code - like udp/tcp/raw ip.
832: */
833: if ((inetsw[ip_protox[ip->ip_p]].pr_flags & PR_LASTHDR) != 0 &&
834: ipsec4_in_reject(m, NULL)) {
1.267 thorpej 835: IPSEC_STATINC(IPSEC_STAT_IN_POLVIO);
1.128 itojun 836: goto bad;
837: }
838: #endif
1.226 liamjfoy 839: #ifdef FAST_IPSEC
1.173 jonathan 840: /*
841: * enforce IPsec policy checking if we are seeing last header.
842: * note that we do not visit this with protocols with pcb layer
843: * code - like udp/tcp/raw ip.
844: */
845: if ((inetsw[ip_protox[ip->ip_p]].pr_flags & PR_LASTHDR) != 0) {
846: /*
847: * Check if the packet has already had IPsec processing
848: * done. If so, then just pass it along. This tag gets
849: * set during AH, ESP, etc. input handling, before the
850: * packet is returned to the ip input queue for delivery.
1.212 perry 851: */
1.173 jonathan 852: mtag = m_tag_find(m, PACKET_TAG_IPSEC_IN_DONE, NULL);
853: s = splsoftnet();
854: if (mtag != NULL) {
855: tdbi = (struct tdb_ident *)(mtag + 1);
856: sp = ipsec_getpolicy(tdbi, IPSEC_DIR_INBOUND);
857: } else {
858: sp = ipsec_getpolicybyaddr(m, IPSEC_DIR_INBOUND,
1.212 perry 859: IP_FORWARDING, &error);
1.173 jonathan 860: }
861: if (sp != NULL) {
862: /*
863: * Check security policy against packet attributes.
864: */
865: error = ipsec_in_reject(sp, m);
866: KEY_FREESP(&sp);
867: } else {
868: /* XXX error stat??? */
869: error = EINVAL;
870: DPRINTF(("ip_input: no SP, packet discarded\n"));/*XXX*/
871: }
872: splx(s);
873: if (error)
874: goto bad;
875: }
876: #endif /* FAST_IPSEC */
1.1 cgd 877:
878: /*
879: * Switch out to protocol's input routine.
880: */
1.82 aidan 881: #if IFA_STATS
1.122 itojun 882: if (ia && ip)
1.155 itojun 883: ia->ia_ifa.ifa_data.ifad_inbytes += ntohs(ip->ip_len);
1.82 aidan 884: #endif
1.266 thorpej 885: IP_STATINC(IP_STAT_DELIVERED);
1.89 itojun 886: {
887: int off = hlen, nh = ip->ip_p;
888:
889: (*inetsw[ip_protox[nh]].pr_input)(m, off, nh);
890: return;
891: }
1.1 cgd 892: bad:
893: m_freem(m);
1.135 thorpej 894: return;
895:
896: badcsum:
1.266 thorpej 897: IP_STATINC(IP_STAT_BADSUM);
1.135 thorpej 898: m_freem(m);
1.1 cgd 899: }
900:
901: /*
1.288 rmind 902: * IP timer processing.
1.1 cgd 903: */
1.8 mycroft 904: void
1.211 perry 905: ip_slowtimo(void)
1.1 cgd 906: {
1.268 ad 907:
908: mutex_enter(softnet_lock);
909: KERNEL_LOCK(1, NULL);
1.1 cgd 910:
1.288 rmind 911: ip_reass_slowtimo();
1.268 ad 912:
913: KERNEL_UNLOCK_ONE(NULL);
914: mutex_exit(softnet_lock);
1.1 cgd 915: }
916:
917: /*
1.288 rmind 918: * IP drain processing.
1.1 cgd 919: */
1.8 mycroft 920: void
1.211 perry 921: ip_drain(void)
1.1 cgd 922: {
923:
1.268 ad 924: KERNEL_LOCK(1, NULL);
1.288 rmind 925: ip_reass_drain();
1.268 ad 926: KERNEL_UNLOCK_ONE(NULL);
1.1 cgd 927: }
928:
929: /*
930: * Do option processing on a datagram,
931: * possibly discarding it if bad options are encountered,
932: * or forwarding it if source-routed.
933: * Returns 1 if packet has been forwarded/freed,
934: * 0 if the packet should be processed further.
935: */
1.8 mycroft 936: int
1.211 perry 937: ip_dooptions(struct mbuf *m)
1.1 cgd 938: {
1.109 augustss 939: struct ip *ip = mtod(m, struct ip *);
940: u_char *cp, *cp0;
941: struct ip_timestamp *ipt;
942: struct in_ifaddr *ia;
1.1 cgd 943: int opt, optlen, cnt, off, code, type = ICMP_PARAMPROB, forward = 0;
1.104 thorpej 944: struct in_addr dst;
1.1 cgd 945: n_time ntime;
946:
1.13 mycroft 947: dst = ip->ip_dst;
1.1 cgd 948: cp = (u_char *)(ip + 1);
949: cnt = (ip->ip_hl << 2) - sizeof (struct ip);
950: for (; cnt > 0; cnt -= optlen, cp += optlen) {
951: opt = cp[IPOPT_OPTVAL];
952: if (opt == IPOPT_EOL)
953: break;
954: if (opt == IPOPT_NOP)
955: optlen = 1;
956: else {
1.113 itojun 957: if (cnt < IPOPT_OLEN + sizeof(*cp)) {
958: code = &cp[IPOPT_OLEN] - (u_char *)ip;
959: goto bad;
960: }
1.1 cgd 961: optlen = cp[IPOPT_OLEN];
1.114 itojun 962: if (optlen < IPOPT_OLEN + sizeof(*cp) || optlen > cnt) {
1.1 cgd 963: code = &cp[IPOPT_OLEN] - (u_char *)ip;
964: goto bad;
965: }
966: }
967: switch (opt) {
968:
969: default:
970: break;
971:
972: /*
973: * Source routing with record.
974: * Find interface with current destination address.
975: * If none on this machine then drop if strictly routed,
976: * or do nothing if loosely routed.
977: * Record interface address and bring up next address
978: * component. If strictly routed make sure next
979: * address is on directly accessible net.
980: */
981: case IPOPT_LSRR:
982: case IPOPT_SSRR:
1.47 cjs 983: if (ip_allowsrcrt == 0) {
984: type = ICMP_UNREACH;
985: code = ICMP_UNREACH_NET_PROHIB;
986: goto bad;
987: }
1.114 itojun 988: if (optlen < IPOPT_OFFSET + sizeof(*cp)) {
989: code = &cp[IPOPT_OLEN] - (u_char *)ip;
990: goto bad;
991: }
1.1 cgd 992: if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) {
993: code = &cp[IPOPT_OFFSET] - (u_char *)ip;
994: goto bad;
995: }
996: ipaddr.sin_addr = ip->ip_dst;
1.19 mycroft 997: ia = ifatoia(ifa_ifwithaddr(sintosa(&ipaddr)));
1.1 cgd 998: if (ia == 0) {
999: if (opt == IPOPT_SSRR) {
1000: type = ICMP_UNREACH;
1001: code = ICMP_UNREACH_SRCFAIL;
1002: goto bad;
1003: }
1004: /*
1005: * Loose routing, and not at next destination
1006: * yet; nothing to do except forward.
1007: */
1008: break;
1009: }
1010: off--; /* 0 origin */
1.112 sommerfe 1011: if ((off + sizeof(struct in_addr)) > optlen) {
1.1 cgd 1012: /*
1013: * End of source route. Should be for us.
1014: */
1015: save_rte(cp, ip->ip_src);
1016: break;
1017: }
1018: /*
1019: * locate outgoing interface
1020: */
1.281 tsutsui 1021: memcpy((void *)&ipaddr.sin_addr, (void *)(cp + off),
1.1 cgd 1022: sizeof(ipaddr.sin_addr));
1.96 thorpej 1023: if (opt == IPOPT_SSRR)
1.196 itojun 1024: ia = ifatoia(ifa_ifwithladdr(sintosa(&ipaddr)));
1.96 thorpej 1025: else
1.1 cgd 1026: ia = ip_rtaddr(ipaddr.sin_addr);
1027: if (ia == 0) {
1028: type = ICMP_UNREACH;
1029: code = ICMP_UNREACH_SRCFAIL;
1030: goto bad;
1031: }
1032: ip->ip_dst = ipaddr.sin_addr;
1.244 christos 1033: bcopy((void *)&ia->ia_addr.sin_addr,
1034: (void *)(cp + off), sizeof(struct in_addr));
1.1 cgd 1035: cp[IPOPT_OFFSET] += sizeof(struct in_addr);
1.13 mycroft 1036: /*
1037: * Let ip_intr's mcast routing check handle mcast pkts
1038: */
1.18 mycroft 1039: forward = !IN_MULTICAST(ip->ip_dst.s_addr);
1.1 cgd 1040: break;
1041:
1042: case IPOPT_RR:
1.114 itojun 1043: if (optlen < IPOPT_OFFSET + sizeof(*cp)) {
1044: code = &cp[IPOPT_OLEN] - (u_char *)ip;
1045: goto bad;
1046: }
1.1 cgd 1047: if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) {
1048: code = &cp[IPOPT_OFFSET] - (u_char *)ip;
1049: goto bad;
1050: }
1051: /*
1052: * If no space remains, ignore.
1053: */
1054: off--; /* 0 origin */
1.112 sommerfe 1055: if ((off + sizeof(struct in_addr)) > optlen)
1.1 cgd 1056: break;
1.281 tsutsui 1057: memcpy((void *)&ipaddr.sin_addr, (void *)(&ip->ip_dst),
1.1 cgd 1058: sizeof(ipaddr.sin_addr));
1059: /*
1060: * locate outgoing interface; if we're the destination,
1061: * use the incoming interface (should be same).
1062: */
1.96 thorpej 1063: if ((ia = ifatoia(ifa_ifwithaddr(sintosa(&ipaddr))))
1064: == NULL &&
1065: (ia = ip_rtaddr(ipaddr.sin_addr)) == NULL) {
1.1 cgd 1066: type = ICMP_UNREACH;
1067: code = ICMP_UNREACH_HOST;
1068: goto bad;
1069: }
1.244 christos 1070: bcopy((void *)&ia->ia_addr.sin_addr,
1071: (void *)(cp + off), sizeof(struct in_addr));
1.1 cgd 1072: cp[IPOPT_OFFSET] += sizeof(struct in_addr);
1073: break;
1074:
1075: case IPOPT_TS:
1076: code = cp - (u_char *)ip;
1077: ipt = (struct ip_timestamp *)cp;
1.114 itojun 1078: if (ipt->ipt_len < 4 || ipt->ipt_len > 40) {
1079: code = (u_char *)&ipt->ipt_len - (u_char *)ip;
1.1 cgd 1080: goto bad;
1.114 itojun 1081: }
1082: if (ipt->ipt_ptr < 5) {
1083: code = (u_char *)&ipt->ipt_ptr - (u_char *)ip;
1084: goto bad;
1085: }
1.15 cgd 1086: if (ipt->ipt_ptr > ipt->ipt_len - sizeof (int32_t)) {
1.114 itojun 1087: if (++ipt->ipt_oflw == 0) {
1088: code = (u_char *)&ipt->ipt_ptr -
1089: (u_char *)ip;
1.1 cgd 1090: goto bad;
1.114 itojun 1091: }
1.1 cgd 1092: break;
1093: }
1.104 thorpej 1094: cp0 = (cp + ipt->ipt_ptr - 1);
1.1 cgd 1095: switch (ipt->ipt_flg) {
1096:
1097: case IPOPT_TS_TSONLY:
1098: break;
1099:
1100: case IPOPT_TS_TSANDADDR:
1.66 thorpej 1101: if (ipt->ipt_ptr - 1 + sizeof(n_time) +
1.114 itojun 1102: sizeof(struct in_addr) > ipt->ipt_len) {
1103: code = (u_char *)&ipt->ipt_ptr -
1104: (u_char *)ip;
1.1 cgd 1105: goto bad;
1.114 itojun 1106: }
1.13 mycroft 1107: ipaddr.sin_addr = dst;
1.96 thorpej 1108: ia = ifatoia(ifaof_ifpforaddr(sintosa(&ipaddr),
1109: m->m_pkthdr.rcvif));
1.13 mycroft 1110: if (ia == 0)
1111: continue;
1.104 thorpej 1112: bcopy(&ia->ia_addr.sin_addr,
1113: cp0, sizeof(struct in_addr));
1.1 cgd 1114: ipt->ipt_ptr += sizeof(struct in_addr);
1115: break;
1116:
1117: case IPOPT_TS_PRESPEC:
1.66 thorpej 1118: if (ipt->ipt_ptr - 1 + sizeof(n_time) +
1.114 itojun 1119: sizeof(struct in_addr) > ipt->ipt_len) {
1120: code = (u_char *)&ipt->ipt_ptr -
1121: (u_char *)ip;
1.1 cgd 1122: goto bad;
1.114 itojun 1123: }
1.281 tsutsui 1124: memcpy(&ipaddr.sin_addr, cp0,
1.1 cgd 1125: sizeof(struct in_addr));
1.96 thorpej 1126: if (ifatoia(ifa_ifwithaddr(sintosa(&ipaddr)))
1127: == NULL)
1.1 cgd 1128: continue;
1129: ipt->ipt_ptr += sizeof(struct in_addr);
1130: break;
1131:
1132: default:
1.114 itojun 1133: /* XXX can't take &ipt->ipt_flg */
1134: code = (u_char *)&ipt->ipt_ptr -
1135: (u_char *)ip + 1;
1.1 cgd 1136: goto bad;
1137: }
1138: ntime = iptime();
1.107 thorpej 1139: cp0 = (u_char *) &ntime; /* XXX grumble, GCC... */
1.244 christos 1140: memmove((char *)cp + ipt->ipt_ptr - 1, cp0,
1.1 cgd 1141: sizeof(n_time));
1142: ipt->ipt_ptr += sizeof(n_time);
1143: }
1144: }
1145: if (forward) {
1.26 thorpej 1146: if (ip_forwsrcrt == 0) {
1147: type = ICMP_UNREACH;
1148: code = ICMP_UNREACH_SRCFAIL;
1149: goto bad;
1150: }
1.1 cgd 1151: ip_forward(m, 1);
1152: return (1);
1.13 mycroft 1153: }
1154: return (0);
1.1 cgd 1155: bad:
1.13 mycroft 1156: icmp_error(m, type, code, 0, 0);
1.266 thorpej 1157: IP_STATINC(IP_STAT_BADOPTIONS);
1.1 cgd 1158: return (1);
1159: }
1160:
1161: /*
1162: * Given address of next destination (final or next hop),
1163: * return internet address info of interface to be used to get there.
1164: */
1165: struct in_ifaddr *
1.211 perry 1166: ip_rtaddr(struct in_addr dst)
1.1 cgd 1167: {
1.249 dyoung 1168: struct rtentry *rt;
1169: union {
1170: struct sockaddr dst;
1171: struct sockaddr_in dst4;
1172: } u;
1173:
1174: sockaddr_in_init(&u.dst4, &dst, 0);
1175:
1176: if ((rt = rtcache_lookup(&ipforward_rt, &u.dst)) == NULL)
1177: return NULL;
1178:
1179: return ifatoia(rt->rt_ifa);
1.1 cgd 1180: }
1181:
1182: /*
1183: * Save incoming source route for use in replies,
1184: * to be picked up later by ip_srcroute if the receiver is interested.
1185: */
1.13 mycroft 1186: void
1.211 perry 1187: save_rte(u_char *option, struct in_addr dst)
1.1 cgd 1188: {
1189: unsigned olen;
1190:
1191: olen = option[IPOPT_OLEN];
1192: #ifdef DIAGNOSTIC
1193: if (ipprintfs)
1.39 christos 1194: printf("save_rte: olen %d\n", olen);
1.89 itojun 1195: #endif /* 0 */
1.1 cgd 1196: if (olen > sizeof(ip_srcrt) - (1 + sizeof(dst)))
1197: return;
1.281 tsutsui 1198: memcpy((void *)ip_srcrt.srcopt, (void *)option, olen);
1.1 cgd 1199: ip_nhops = (olen - IPOPT_OFFSET - 1) / sizeof(struct in_addr);
1200: ip_srcrt.dst = dst;
1201: }
1202:
1203: /*
1204: * Retrieve incoming source route for use in replies,
1205: * in the same form used by setsockopt.
1206: * The first hop is placed before the options, will be removed later.
1207: */
1208: struct mbuf *
1.211 perry 1209: ip_srcroute(void)
1.1 cgd 1210: {
1.109 augustss 1211: struct in_addr *p, *q;
1212: struct mbuf *m;
1.1 cgd 1213:
1214: if (ip_nhops == 0)
1.237 dyoung 1215: return NULL;
1.1 cgd 1216: m = m_get(M_DONTWAIT, MT_SOOPTS);
1217: if (m == 0)
1.237 dyoung 1218: return NULL;
1.1 cgd 1219:
1.164 matt 1220: MCLAIM(m, &inetdomain.dom_mowner);
1.13 mycroft 1221: #define OPTSIZ (sizeof(ip_srcrt.nop) + sizeof(ip_srcrt.srcopt))
1.1 cgd 1222:
1223: /* length is (nhops+1)*sizeof(addr) + sizeof(nop + srcrt header) */
1224: m->m_len = ip_nhops * sizeof(struct in_addr) + sizeof(struct in_addr) +
1225: OPTSIZ;
1226: #ifdef DIAGNOSTIC
1227: if (ipprintfs)
1.39 christos 1228: printf("ip_srcroute: nhops %d mlen %d", ip_nhops, m->m_len);
1.1 cgd 1229: #endif
1230:
1231: /*
1232: * First save first hop for return route
1233: */
1234: p = &ip_srcrt.route[ip_nhops - 1];
1235: *(mtod(m, struct in_addr *)) = *p--;
1236: #ifdef DIAGNOSTIC
1237: if (ipprintfs)
1.39 christos 1238: printf(" hops %x", ntohl(mtod(m, struct in_addr *)->s_addr));
1.1 cgd 1239: #endif
1240:
1241: /*
1242: * Copy option fields and padding (nop) to mbuf.
1243: */
1244: ip_srcrt.nop = IPOPT_NOP;
1245: ip_srcrt.srcopt[IPOPT_OFFSET] = IPOPT_MINOFF;
1.244 christos 1246: memmove(mtod(m, char *) + sizeof(struct in_addr), &ip_srcrt.nop,
1247: OPTSIZ);
1248: q = (struct in_addr *)(mtod(m, char *) +
1.1 cgd 1249: sizeof(struct in_addr) + OPTSIZ);
1250: #undef OPTSIZ
1251: /*
1252: * Record return path as an IP source route,
1253: * reversing the path (pointers are now aligned).
1254: */
1255: while (p >= ip_srcrt.route) {
1256: #ifdef DIAGNOSTIC
1257: if (ipprintfs)
1.39 christos 1258: printf(" %x", ntohl(q->s_addr));
1.1 cgd 1259: #endif
1260: *q++ = *p--;
1261: }
1262: /*
1263: * Last hop goes to final destination.
1264: */
1265: *q = ip_srcrt.dst;
1266: #ifdef DIAGNOSTIC
1267: if (ipprintfs)
1.39 christos 1268: printf(" %x\n", ntohl(q->s_addr));
1.1 cgd 1269: #endif
1270: return (m);
1271: }
1272:
1.139 matt 1273: const int inetctlerrmap[PRC_NCMDS] = {
1.256 yamt 1274: [PRC_MSGSIZE] = EMSGSIZE,
1275: [PRC_HOSTDEAD] = EHOSTDOWN,
1276: [PRC_HOSTUNREACH] = EHOSTUNREACH,
1277: [PRC_UNREACH_NET] = EHOSTUNREACH,
1278: [PRC_UNREACH_HOST] = EHOSTUNREACH,
1279: [PRC_UNREACH_PROTOCOL] = ECONNREFUSED,
1280: [PRC_UNREACH_PORT] = ECONNREFUSED,
1281: [PRC_UNREACH_SRCFAIL] = EHOSTUNREACH,
1282: [PRC_PARAMPROB] = ENOPROTOOPT,
1.1 cgd 1283: };
1284:
1285: /*
1286: * Forward a packet. If some error occurs return the sender
1287: * an icmp packet. Note we can't always generate a meaningful
1288: * icmp message because icmp doesn't have a large enough repertoire
1289: * of codes and types.
1290: *
1291: * If not forwarding, just drop the packet. This could be confusing
1292: * if ipforwarding was zero but some routing protocol was advancing
1293: * us as a gateway to somewhere. However, we must let the routing
1294: * protocol deal with that.
1295: *
1296: * The srcrt parameter indicates whether the packet is being forwarded
1297: * via a source route.
1298: */
1.13 mycroft 1299: void
1.211 perry 1300: ip_forward(struct mbuf *m, int srcrt)
1.1 cgd 1301: {
1.109 augustss 1302: struct ip *ip = mtod(m, struct ip *);
1303: struct rtentry *rt;
1.220 christos 1304: int error, type = 0, code = 0, destmtu = 0;
1.1 cgd 1305: struct mbuf *mcopy;
1.13 mycroft 1306: n_long dest;
1.249 dyoung 1307: union {
1308: struct sockaddr dst;
1309: struct sockaddr_in dst4;
1310: } u;
1.164 matt 1311:
1312: /*
1313: * We are now in the output path.
1314: */
1315: MCLAIM(m, &ip_tx_mowner);
1.135 thorpej 1316:
1317: /*
1318: * Clear any in-bound checksum flags for this packet.
1319: */
1320: m->m_pkthdr.csum_flags = 0;
1.1 cgd 1321:
1.13 mycroft 1322: dest = 0;
1.1 cgd 1323: #ifdef DIAGNOSTIC
1.224 joerg 1324: if (ipprintfs) {
1325: printf("forward: src %s ", inet_ntoa(ip->ip_src));
1326: printf("dst %s ttl %x\n", inet_ntoa(ip->ip_dst), ip->ip_ttl);
1327: }
1.1 cgd 1328: #endif
1.93 sommerfe 1329: if (m->m_flags & (M_BCAST|M_MCAST) || in_canforward(ip->ip_dst) == 0) {
1.266 thorpej 1330: IP_STATINC(IP_STAT_CANTFORWARD);
1.1 cgd 1331: m_freem(m);
1332: return;
1333: }
1334: if (ip->ip_ttl <= IPTTLDEC) {
1.13 mycroft 1335: icmp_error(m, ICMP_TIMXCEED, ICMP_TIMXCEED_INTRANS, dest, 0);
1.1 cgd 1336: return;
1337: }
1338:
1.249 dyoung 1339: sockaddr_in_init(&u.dst4, &ip->ip_dst, 0);
1340: if ((rt = rtcache_lookup(&ipforward_rt, &u.dst)) == NULL) {
1341: icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_NET, dest, 0);
1342: return;
1.1 cgd 1343: }
1344:
1345: /*
1.34 mycroft 1346: * Save at most 68 bytes of the packet in case
1.1 cgd 1347: * we need to generate an ICMP message to the src.
1.119 itojun 1348: * Pullup to avoid sharing mbuf cluster between m and mcopy.
1.1 cgd 1349: */
1.155 itojun 1350: mcopy = m_copym(m, 0, imin(ntohs(ip->ip_len), 68), M_DONTWAIT);
1.119 itojun 1351: if (mcopy)
1352: mcopy = m_pullup(mcopy, ip->ip_hl << 2);
1.1 cgd 1353:
1.221 christos 1354: ip->ip_ttl -= IPTTLDEC;
1355:
1.1 cgd 1356: /*
1357: * If forwarding packet using same interface that it came in on,
1358: * perhaps should send a redirect to sender to shortcut a hop.
1359: * Only send redirect if source is sending directly to us,
1360: * and if packet was not source routed (or has any options).
1361: * Also, don't send redirect if forwarding using a default route
1362: * or a route modified by a redirect.
1363: */
1364: if (rt->rt_ifp == m->m_pkthdr.rcvif &&
1365: (rt->rt_flags & (RTF_DYNAMIC|RTF_MODIFIED)) == 0 &&
1.250 dyoung 1366: !in_nullhost(satocsin(rt_getkey(rt))->sin_addr) &&
1.1 cgd 1367: ipsendredirects && !srcrt) {
1.19 mycroft 1368: if (rt->rt_ifa &&
1369: (ip->ip_src.s_addr & ifatoia(rt->rt_ifa)->ia_subnetmask) ==
1370: ifatoia(rt->rt_ifa)->ia_subnet) {
1.77 thorpej 1371: if (rt->rt_flags & RTF_GATEWAY)
1372: dest = satosin(rt->rt_gateway)->sin_addr.s_addr;
1373: else
1374: dest = ip->ip_dst.s_addr;
1375: /*
1376: * Router requirements says to only send host
1377: * redirects.
1378: */
1379: type = ICMP_REDIRECT;
1380: code = ICMP_REDIRECT_HOST;
1.1 cgd 1381: #ifdef DIAGNOSTIC
1.77 thorpej 1382: if (ipprintfs)
1383: printf("redirect (%d) to %x\n", code,
1384: (u_int32_t)dest);
1.1 cgd 1385: #endif
1386: }
1387: }
1388:
1.238 dyoung 1389: error = ip_output(m, NULL, &ipforward_rt,
1.173 jonathan 1390: (IP_FORWARDING | (ip_directedbcast ? IP_ALLOWBROADCAST : 0)),
1.174 itojun 1391: (struct ip_moptions *)NULL, (struct socket *)NULL);
1.173 jonathan 1392:
1.1 cgd 1393: if (error)
1.266 thorpej 1394: IP_STATINC(IP_STAT_CANTFORWARD);
1.1 cgd 1395: else {
1.266 thorpej 1396: uint64_t *ips = IP_STAT_GETREF();
1397: ips[IP_STAT_FORWARD]++;
1398: if (type) {
1399: ips[IP_STAT_REDIRECTSENT]++;
1400: IP_STAT_PUTREF();
1401: } else {
1402: IP_STAT_PUTREF();
1.63 matt 1403: if (mcopy) {
1404: #ifdef GATEWAY
1.64 thorpej 1405: if (mcopy->m_flags & M_CANFASTFWD)
1406: ipflow_create(&ipforward_rt, mcopy);
1.63 matt 1407: #endif
1.1 cgd 1408: m_freem(mcopy);
1.63 matt 1409: }
1.1 cgd 1410: return;
1411: }
1412: }
1413: if (mcopy == NULL)
1414: return;
1.13 mycroft 1415:
1.1 cgd 1416: switch (error) {
1417:
1418: case 0: /* forwarded, but need redirect */
1419: /* type, code set above */
1420: break;
1421:
1422: case ENETUNREACH: /* shouldn't happen, checked above */
1423: case EHOSTUNREACH:
1424: case ENETDOWN:
1425: case EHOSTDOWN:
1426: default:
1427: type = ICMP_UNREACH;
1428: code = ICMP_UNREACH_HOST;
1429: break;
1430:
1431: case EMSGSIZE:
1432: type = ICMP_UNREACH;
1433: code = ICMP_UNREACH_NEEDFRAG;
1.263 cube 1434:
1.274 seanb 1435: if ((rt = rtcache_validate(&ipforward_rt)) != NULL)
1436: destmtu = rt->rt_ifp->if_mtu;
1.263 cube 1437:
1438: #if defined(IPSEC) || defined(FAST_IPSEC)
1.274 seanb 1439: {
1.263 cube 1440: /*
1441: * If the packet is routed over IPsec tunnel, tell the
1442: * originator the tunnel MTU.
1443: * tunnel MTU = if MTU - sizeof(IP) - ESP/AH hdrsiz
1444: * XXX quickhack!!!
1445: */
1446:
1.89 itojun 1447: struct secpolicy *sp;
1448: int ipsecerror;
1.95 itojun 1449: size_t ipsechdr;
1.89 itojun 1450: struct route *ro;
1451:
1452: sp = ipsec4_getpolicybyaddr(mcopy,
1.170 itojun 1453: IPSEC_DIR_OUTBOUND, IP_FORWARDING,
1454: &ipsecerror);
1.89 itojun 1455:
1.263 cube 1456: if (sp != NULL) {
1.89 itojun 1457: /* count IPsec header size */
1.95 itojun 1458: ipsechdr = ipsec4_hdrsiz(mcopy,
1.170 itojun 1459: IPSEC_DIR_OUTBOUND, NULL);
1.89 itojun 1460:
1461: /*
1462: * find the correct route for outer IPv4
1463: * header, compute tunnel MTU.
1464: */
1.220 christos 1465:
1.89 itojun 1466: if (sp->req != NULL
1.95 itojun 1467: && sp->req->sav != NULL
1468: && sp->req->sav->sah != NULL) {
1469: ro = &sp->req->sav->sah->sa_route;
1.274 seanb 1470: rt = rtcache_validate(ro);
1.257 dyoung 1471: if (rt && rt->rt_ifp) {
1.220 christos 1472: destmtu =
1.257 dyoung 1473: rt->rt_rmx.rmx_mtu ?
1474: rt->rt_rmx.rmx_mtu :
1475: rt->rt_ifp->if_mtu;
1.220 christos 1476: destmtu -= ipsechdr;
1.89 itojun 1477: }
1478: }
1479:
1.173 jonathan 1480: #ifdef IPSEC
1.89 itojun 1481: key_freesp(sp);
1.173 jonathan 1482: #else
1483: KEY_FREESP(&sp);
1484: #endif
1.89 itojun 1485: }
1.274 seanb 1486: }
1.263 cube 1487: #endif /*defined(IPSEC) || defined(FAST_IPSEC)*/
1.266 thorpej 1488: IP_STATINC(IP_STAT_CANTFRAG);
1.1 cgd 1489: break;
1490:
1491: case ENOBUFS:
1.143 itojun 1492: #if 1
1493: /*
1494: * a router should not generate ICMP_SOURCEQUENCH as
1495: * required in RFC1812 Requirements for IP Version 4 Routers.
1496: * source quench could be a big problem under DoS attacks,
1.149 wiz 1497: * or if the underlying interface is rate-limited.
1.143 itojun 1498: */
1499: if (mcopy)
1500: m_freem(mcopy);
1501: return;
1502: #else
1.1 cgd 1503: type = ICMP_SOURCEQUENCH;
1504: code = 0;
1505: break;
1.143 itojun 1506: #endif
1.1 cgd 1507: }
1.220 christos 1508: icmp_error(mcopy, type, code, dest, destmtu);
1.44 thorpej 1509: }
1510:
1511: void
1.211 perry 1512: ip_savecontrol(struct inpcb *inp, struct mbuf **mp, struct ip *ip,
1513: struct mbuf *m)
1.44 thorpej 1514: {
1515:
1.283 minskim 1516: if (inp->inp_socket->so_options & SO_TIMESTAMP
1.278 christos 1517: #ifdef SO_OTIMESTAMP
1.283 minskim 1518: || inp->inp_socket->so_options & SO_OTIMESTAMP
1.278 christos 1519: #endif
1520: ) {
1.44 thorpej 1521: struct timeval tv;
1522:
1523: microtime(&tv);
1.278 christos 1524: #ifdef SO_OTIMESTAMP
1525: if (inp->inp_socket->so_options & SO_OTIMESTAMP) {
1526: struct timeval50 tv50;
1527: timeval_to_timeval50(&tv, &tv50);
1528: *mp = sbcreatecontrol((void *) &tv50, sizeof(tv50),
1529: SCM_OTIMESTAMP, SOL_SOCKET);
1530: } else
1531: #endif
1.244 christos 1532: *mp = sbcreatecontrol((void *) &tv, sizeof(tv),
1.44 thorpej 1533: SCM_TIMESTAMP, SOL_SOCKET);
1534: if (*mp)
1535: mp = &(*mp)->m_next;
1536: }
1537: if (inp->inp_flags & INP_RECVDSTADDR) {
1.244 christos 1538: *mp = sbcreatecontrol((void *) &ip->ip_dst,
1.44 thorpej 1539: sizeof(struct in_addr), IP_RECVDSTADDR, IPPROTO_IP);
1540: if (*mp)
1541: mp = &(*mp)->m_next;
1542: }
1543: #ifdef notyet
1544: /*
1545: * XXX
1546: * Moving these out of udp_input() made them even more broken
1547: * than they already were.
1548: * - fenner@parc.xerox.com
1549: */
1550: /* options were tossed already */
1551: if (inp->inp_flags & INP_RECVOPTS) {
1.244 christos 1552: *mp = sbcreatecontrol((void *) opts_deleted_above,
1.44 thorpej 1553: sizeof(struct in_addr), IP_RECVOPTS, IPPROTO_IP);
1554: if (*mp)
1555: mp = &(*mp)->m_next;
1556: }
1557: /* ip_srcroute doesn't do what we want here, need to fix */
1558: if (inp->inp_flags & INP_RECVRETOPTS) {
1.244 christos 1559: *mp = sbcreatecontrol((void *) ip_srcroute(),
1.44 thorpej 1560: sizeof(struct in_addr), IP_RECVRETOPTS, IPPROTO_IP);
1561: if (*mp)
1562: mp = &(*mp)->m_next;
1563: }
1564: #endif
1565: if (inp->inp_flags & INP_RECVIF) {
1566: struct sockaddr_dl sdl;
1567:
1.252 dyoung 1568: sockaddr_dl_init(&sdl, sizeof(sdl),
1569: (m->m_pkthdr.rcvif != NULL)
1570: ? m->m_pkthdr.rcvif->if_index
1571: : 0,
1572: 0, NULL, 0, NULL, 0);
1.251 dyoung 1573: *mp = sbcreatecontrol(&sdl, sdl.sdl_len, IP_RECVIF, IPPROTO_IP);
1.44 thorpej 1574: if (*mp)
1575: mp = &(*mp)->m_next;
1576: }
1.282 minskim 1577: if (inp->inp_flags & INP_RECVTTL) {
1578: *mp = sbcreatecontrol((void *) &ip->ip_ttl,
1579: sizeof(uint8_t), IP_RECVTTL, IPPROTO_IP);
1580: if (*mp)
1581: mp = &(*mp)->m_next;
1582: }
1.13 mycroft 1583: }
1584:
1.189 atatat 1585: /*
1.228 elad 1586: * sysctl helper routine for net.inet.ip.forwsrcrt.
1587: */
1588: static int
1589: sysctl_net_inet_ip_forwsrcrt(SYSCTLFN_ARGS)
1590: {
1591: int error, tmp;
1592: struct sysctlnode node;
1593:
1594: node = *rnode;
1595: tmp = ip_forwsrcrt;
1596: node.sysctl_data = &tmp;
1597: error = sysctl_lookup(SYSCTLFN_CALL(&node));
1598: if (error || newp == NULL)
1599: return (error);
1600:
1.280 elad 1601: error = kauth_authorize_network(l->l_cred, KAUTH_NETWORK_FORWSRCRT,
1602: 0, NULL, NULL, NULL);
1603: if (error)
1604: return (error);
1.228 elad 1605:
1606: ip_forwsrcrt = tmp;
1607:
1608: return (0);
1609: }
1610:
1611: /*
1.189 atatat 1612: * sysctl helper routine for net.inet.ip.mtudisctimeout. checks the
1613: * range of the new value and tweaks timers if it changes.
1614: */
1615: static int
1616: sysctl_net_inet_ip_pmtudto(SYSCTLFN_ARGS)
1.13 mycroft 1617: {
1.189 atatat 1618: int error, tmp;
1619: struct sysctlnode node;
1620:
1621: node = *rnode;
1622: tmp = ip_mtudisc_timeout;
1623: node.sysctl_data = &tmp;
1624: error = sysctl_lookup(SYSCTLFN_CALL(&node));
1625: if (error || newp == NULL)
1626: return (error);
1627: if (tmp < 0)
1628: return (EINVAL);
1.52 thorpej 1629:
1.273 matt 1630: mutex_enter(softnet_lock);
1631:
1.189 atatat 1632: ip_mtudisc_timeout = tmp;
1633: rt_timer_queue_change(ip_mtudisc_timeout_q, ip_mtudisc_timeout);
1634:
1.273 matt 1635: mutex_exit(softnet_lock);
1636:
1.189 atatat 1637: return (0);
1638: }
1.54 lukem 1639:
1.65 matt 1640: #ifdef GATEWAY
1.189 atatat 1641: /*
1.247 liamjfoy 1642: * sysctl helper routine for net.inet.ip.maxflows.
1.189 atatat 1643: */
1644: static int
1645: sysctl_net_inet_ip_maxflows(SYSCTLFN_ARGS)
1646: {
1.273 matt 1647: int error;
1.67 thorpej 1648:
1.273 matt 1649: error = sysctl_lookup(SYSCTLFN_CALL(rnode));
1650: if (error || newp == NULL)
1651: return (error);
1652:
1653: mutex_enter(softnet_lock);
1654: KERNEL_LOCK(1, NULL);
1.212 perry 1655:
1.265 thorpej 1656: ipflow_prune();
1.273 matt 1657:
1658: KERNEL_UNLOCK_ONE(NULL);
1659: mutex_exit(softnet_lock);
1.144 martin 1660:
1.189 atatat 1661: return (0);
1662: }
1.248 liamjfoy 1663:
1664: static int
1665: sysctl_net_inet_ip_hashsize(SYSCTLFN_ARGS)
1.283 minskim 1666: {
1.248 liamjfoy 1667: int error, tmp;
1668: struct sysctlnode node;
1669:
1670: node = *rnode;
1671: tmp = ip_hashsize;
1672: node.sysctl_data = &tmp;
1673: error = sysctl_lookup(SYSCTLFN_CALL(&node));
1674: if (error || newp == NULL)
1675: return (error);
1676:
1677: if ((tmp & (tmp - 1)) == 0 && tmp != 0) {
1678: /*
1679: * Can only fail due to malloc()
1680: */
1.273 matt 1681: mutex_enter(softnet_lock);
1682: KERNEL_LOCK(1, NULL);
1683:
1684: error = ipflow_invalidate_all(tmp);
1685:
1686: KERNEL_UNLOCK_ONE(NULL);
1687: mutex_exit(softnet_lock);
1688:
1.248 liamjfoy 1689: } else {
1690: /*
1691: * EINVAL if not a power of 2
1692: */
1.273 matt 1693: error = EINVAL;
1.283 minskim 1694: }
1.248 liamjfoy 1695:
1.273 matt 1696: return error;
1.248 liamjfoy 1697: }
1.189 atatat 1698: #endif /* GATEWAY */
1.117 tron 1699:
1.266 thorpej 1700: static int
1701: sysctl_net_inet_ip_stats(SYSCTLFN_ARGS)
1702: {
1703:
1.271 thorpej 1704: return (NETSTAT_SYSCTL(ipstat_percpu, IP_NSTATS));
1.266 thorpej 1705: }
1.131 itojun 1706:
1.284 pooka 1707: static void
1708: sysctl_net_inet_ip_setup(struct sysctllog **clog)
1.189 atatat 1709: {
1710: extern int subnetsarelocal, hostzeroisbroadcast;
1.180 jonathan 1711:
1.197 atatat 1712: sysctl_createv(clog, 0, NULL, NULL,
1713: CTLFLAG_PERMANENT,
1.189 atatat 1714: CTLTYPE_NODE, "net", NULL,
1715: NULL, 0, NULL, 0,
1716: CTL_NET, CTL_EOL);
1.197 atatat 1717: sysctl_createv(clog, 0, NULL, NULL,
1718: CTLFLAG_PERMANENT,
1.203 atatat 1719: CTLTYPE_NODE, "inet",
1720: SYSCTL_DESCR("PF_INET related settings"),
1.189 atatat 1721: NULL, 0, NULL, 0,
1722: CTL_NET, PF_INET, CTL_EOL);
1.197 atatat 1723: sysctl_createv(clog, 0, NULL, NULL,
1724: CTLFLAG_PERMANENT,
1.203 atatat 1725: CTLTYPE_NODE, "ip",
1726: SYSCTL_DESCR("IPv4 related settings"),
1.189 atatat 1727: NULL, 0, NULL, 0,
1728: CTL_NET, PF_INET, IPPROTO_IP, CTL_EOL);
1.212 perry 1729:
1.197 atatat 1730: sysctl_createv(clog, 0, NULL, NULL,
1731: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1732: CTLTYPE_INT, "forwarding",
1733: SYSCTL_DESCR("Enable forwarding of INET datagrams"),
1.189 atatat 1734: NULL, 0, &ipforwarding, 0,
1735: CTL_NET, PF_INET, IPPROTO_IP,
1736: IPCTL_FORWARDING, CTL_EOL);
1.197 atatat 1737: sysctl_createv(clog, 0, NULL, NULL,
1738: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1739: CTLTYPE_INT, "redirect",
1740: SYSCTL_DESCR("Enable sending of ICMP redirect messages"),
1.189 atatat 1741: NULL, 0, &ipsendredirects, 0,
1742: CTL_NET, PF_INET, IPPROTO_IP,
1743: IPCTL_SENDREDIRECTS, CTL_EOL);
1.197 atatat 1744: sysctl_createv(clog, 0, NULL, NULL,
1745: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1746: CTLTYPE_INT, "ttl",
1747: SYSCTL_DESCR("Default TTL for an INET datagram"),
1.189 atatat 1748: NULL, 0, &ip_defttl, 0,
1749: CTL_NET, PF_INET, IPPROTO_IP,
1750: IPCTL_DEFTTL, CTL_EOL);
1751: #ifdef IPCTL_DEFMTU
1.197 atatat 1752: sysctl_createv(clog, 0, NULL, NULL,
1753: CTLFLAG_PERMANENT /* |CTLFLAG_READWRITE? */,
1.203 atatat 1754: CTLTYPE_INT, "mtu",
1755: SYSCTL_DESCR("Default MTA for an INET route"),
1.189 atatat 1756: NULL, 0, &ip_mtu, 0,
1757: CTL_NET, PF_INET, IPPROTO_IP,
1758: IPCTL_DEFMTU, CTL_EOL);
1759: #endif /* IPCTL_DEFMTU */
1.197 atatat 1760: sysctl_createv(clog, 0, NULL, NULL,
1.228 elad 1761: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1762: CTLTYPE_INT, "forwsrcrt",
1763: SYSCTL_DESCR("Enable forwarding of source-routed "
1764: "datagrams"),
1.228 elad 1765: sysctl_net_inet_ip_forwsrcrt, 0, &ip_forwsrcrt, 0,
1.189 atatat 1766: CTL_NET, PF_INET, IPPROTO_IP,
1767: IPCTL_FORWSRCRT, CTL_EOL);
1.197 atatat 1768: sysctl_createv(clog, 0, NULL, NULL,
1769: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1770: CTLTYPE_INT, "directed-broadcast",
1771: SYSCTL_DESCR("Enable forwarding of broadcast datagrams"),
1.189 atatat 1772: NULL, 0, &ip_directedbcast, 0,
1773: CTL_NET, PF_INET, IPPROTO_IP,
1774: IPCTL_DIRECTEDBCAST, CTL_EOL);
1.197 atatat 1775: sysctl_createv(clog, 0, NULL, NULL,
1776: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1777: CTLTYPE_INT, "allowsrcrt",
1778: SYSCTL_DESCR("Accept source-routed datagrams"),
1.189 atatat 1779: NULL, 0, &ip_allowsrcrt, 0,
1780: CTL_NET, PF_INET, IPPROTO_IP,
1781: IPCTL_ALLOWSRCRT, CTL_EOL);
1.197 atatat 1782: sysctl_createv(clog, 0, NULL, NULL,
1783: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1784: CTLTYPE_INT, "subnetsarelocal",
1785: SYSCTL_DESCR("Whether logical subnets are considered "
1786: "local"),
1.189 atatat 1787: NULL, 0, &subnetsarelocal, 0,
1788: CTL_NET, PF_INET, IPPROTO_IP,
1789: IPCTL_SUBNETSARELOCAL, CTL_EOL);
1.197 atatat 1790: sysctl_createv(clog, 0, NULL, NULL,
1791: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1792: CTLTYPE_INT, "mtudisc",
1793: SYSCTL_DESCR("Use RFC1191 Path MTU Discovery"),
1.189 atatat 1794: NULL, 0, &ip_mtudisc, 0,
1795: CTL_NET, PF_INET, IPPROTO_IP,
1796: IPCTL_MTUDISC, CTL_EOL);
1.197 atatat 1797: sysctl_createv(clog, 0, NULL, NULL,
1798: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1799: CTLTYPE_INT, "anonportmin",
1800: SYSCTL_DESCR("Lowest ephemeral port number to assign"),
1.189 atatat 1801: sysctl_net_inet_ip_ports, 0, &anonportmin, 0,
1802: CTL_NET, PF_INET, IPPROTO_IP,
1803: IPCTL_ANONPORTMIN, CTL_EOL);
1.197 atatat 1804: sysctl_createv(clog, 0, NULL, NULL,
1805: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1806: CTLTYPE_INT, "anonportmax",
1807: SYSCTL_DESCR("Highest ephemeral port number to assign"),
1.189 atatat 1808: sysctl_net_inet_ip_ports, 0, &anonportmax, 0,
1809: CTL_NET, PF_INET, IPPROTO_IP,
1810: IPCTL_ANONPORTMAX, CTL_EOL);
1.197 atatat 1811: sysctl_createv(clog, 0, NULL, NULL,
1812: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1813: CTLTYPE_INT, "mtudisctimeout",
1814: SYSCTL_DESCR("Lifetime of a Path MTU Discovered route"),
1.189 atatat 1815: sysctl_net_inet_ip_pmtudto, 0, &ip_mtudisc_timeout, 0,
1816: CTL_NET, PF_INET, IPPROTO_IP,
1817: IPCTL_MTUDISCTIMEOUT, CTL_EOL);
1818: #ifdef GATEWAY
1.197 atatat 1819: sysctl_createv(clog, 0, NULL, NULL,
1820: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1821: CTLTYPE_INT, "maxflows",
1822: SYSCTL_DESCR("Number of flows for fast forwarding"),
1.189 atatat 1823: sysctl_net_inet_ip_maxflows, 0, &ip_maxflows, 0,
1824: CTL_NET, PF_INET, IPPROTO_IP,
1825: IPCTL_MAXFLOWS, CTL_EOL);
1.248 liamjfoy 1826: sysctl_createv(clog, 0, NULL, NULL,
1827: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1828: CTLTYPE_INT, "hashsize",
1829: SYSCTL_DESCR("Size of hash table for fast forwarding (IPv4)"),
1830: sysctl_net_inet_ip_hashsize, 0, &ip_hashsize, 0,
1831: CTL_NET, PF_INET, IPPROTO_IP,
1832: CTL_CREATE, CTL_EOL);
1.189 atatat 1833: #endif /* GATEWAY */
1.197 atatat 1834: sysctl_createv(clog, 0, NULL, NULL,
1835: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1836: CTLTYPE_INT, "hostzerobroadcast",
1837: SYSCTL_DESCR("All zeroes address is broadcast address"),
1.189 atatat 1838: NULL, 0, &hostzeroisbroadcast, 0,
1839: CTL_NET, PF_INET, IPPROTO_IP,
1840: IPCTL_HOSTZEROBROADCAST, CTL_EOL);
1841: #if NGIF > 0
1.197 atatat 1842: sysctl_createv(clog, 0, NULL, NULL,
1843: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1844: CTLTYPE_INT, "gifttl",
1845: SYSCTL_DESCR("Default TTL for a gif tunnel datagram"),
1.189 atatat 1846: NULL, 0, &ip_gif_ttl, 0,
1847: CTL_NET, PF_INET, IPPROTO_IP,
1848: IPCTL_GIF_TTL, CTL_EOL);
1849: #endif /* NGIF */
1850: #ifndef IPNOPRIVPORTS
1.197 atatat 1851: sysctl_createv(clog, 0, NULL, NULL,
1852: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1853: CTLTYPE_INT, "lowportmin",
1854: SYSCTL_DESCR("Lowest privileged ephemeral port number "
1855: "to assign"),
1.189 atatat 1856: sysctl_net_inet_ip_ports, 0, &lowportmin, 0,
1857: CTL_NET, PF_INET, IPPROTO_IP,
1858: IPCTL_LOWPORTMIN, CTL_EOL);
1.197 atatat 1859: sysctl_createv(clog, 0, NULL, NULL,
1860: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1861: CTLTYPE_INT, "lowportmax",
1862: SYSCTL_DESCR("Highest privileged ephemeral port number "
1863: "to assign"),
1.189 atatat 1864: sysctl_net_inet_ip_ports, 0, &lowportmax, 0,
1865: CTL_NET, PF_INET, IPPROTO_IP,
1866: IPCTL_LOWPORTMAX, CTL_EOL);
1867: #endif /* IPNOPRIVPORTS */
1868: #if NGRE > 0
1.197 atatat 1869: sysctl_createv(clog, 0, NULL, NULL,
1870: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1871: CTLTYPE_INT, "grettl",
1872: SYSCTL_DESCR("Default TTL for a gre tunnel datagram"),
1.189 atatat 1873: NULL, 0, &ip_gre_ttl, 0,
1874: CTL_NET, PF_INET, IPPROTO_IP,
1875: IPCTL_GRE_TTL, CTL_EOL);
1876: #endif /* NGRE */
1.197 atatat 1877: sysctl_createv(clog, 0, NULL, NULL,
1878: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1879: CTLTYPE_INT, "checkinterface",
1880: SYSCTL_DESCR("Enable receive side of Strong ES model "
1881: "from RFC1122"),
1.189 atatat 1882: NULL, 0, &ip_checkinterface, 0,
1883: CTL_NET, PF_INET, IPPROTO_IP,
1884: IPCTL_CHECKINTERFACE, CTL_EOL);
1.197 atatat 1885: sysctl_createv(clog, 0, NULL, NULL,
1886: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1887: CTLTYPE_INT, "random_id",
1888: SYSCTL_DESCR("Assign random ip_id values"),
1.189 atatat 1889: NULL, 0, &ip_do_randomid, 0,
1890: CTL_NET, PF_INET, IPPROTO_IP,
1891: IPCTL_RANDOMID, CTL_EOL);
1.206 thorpej 1892: sysctl_createv(clog, 0, NULL, NULL,
1893: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1894: CTLTYPE_INT, "do_loopback_cksum",
1895: SYSCTL_DESCR("Perform IP checksum on loopback"),
1896: NULL, 0, &ip_do_loopback_cksum, 0,
1897: CTL_NET, PF_INET, IPPROTO_IP,
1898: IPCTL_LOOPBACKCKSUM, CTL_EOL);
1.219 elad 1899: sysctl_createv(clog, 0, NULL, NULL,
1900: CTLFLAG_PERMANENT,
1901: CTLTYPE_STRUCT, "stats",
1902: SYSCTL_DESCR("IP statistics"),
1.266 thorpej 1903: sysctl_net_inet_ip_stats, 0, NULL, 0,
1.219 elad 1904: CTL_NET, PF_INET, IPPROTO_IP, IPCTL_STATS,
1905: CTL_EOL);
1.1 cgd 1906: }
1.266 thorpej 1907:
1908: void
1909: ip_statinc(u_int stat)
1910: {
1911:
1912: KASSERT(stat < IP_NSTATS);
1913: IP_STATINC(stat);
1914: }
CVSweb <webmaster@jp.NetBSD.org>