Annotation of src/sys/netinet/ip_input.c, Revision 1.299
1.299 ! drochner 1: /* $NetBSD: ip_input.c,v 1.298 2012/01/09 14:31:22 liamjfoy 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>
1.299 ! drochner 94: __KERNEL_RCSID(0, "$NetBSD: ip_input.c,v 1.298 2012/01/09 14:31:22 liamjfoy 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:
1.173 jonathan 143: #ifdef FAST_IPSEC
144: #include <netipsec/ipsec.h>
145: #include <netipsec/key.h>
146: #endif /* FAST_IPSEC*/
1.44 thorpej 147:
1.1 cgd 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
1.26 thorpej 158: #ifndef IPFORWSRCRT
1.47 cjs 159: #define IPFORWSRCRT 1 /* forward source-routed packets */
160: #endif
161: #ifndef IPALLOWSRCRT
1.48 mrg 162: #define IPALLOWSRCRT 1 /* allow source-routed packets */
1.26 thorpej 163: #endif
1.53 kml 164: #ifndef IPMTUDISC
1.153 itojun 165: #define IPMTUDISC 1
1.53 kml 166: #endif
1.60 kml 167: #ifndef IPMTUDISCTIMEOUT
1.61 kml 168: #define IPMTUDISCTIMEOUT (10 * 60) /* as per RFC 1191 */
1.60 kml 169: #endif
1.53 kml 170:
1.278 christos 171: #ifdef COMPAT_50
172: #include <compat/sys/time.h>
173: #include <compat/sys/socket.h>
174: #endif
175:
1.27 thorpej 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 */
1.1 cgd 187: int ipforwarding = IPFORWARDING;
188: int ipsendredirects = IPSENDREDIRECTS;
1.13 mycroft 189: int ip_defttl = IPDEFTTL;
1.26 thorpej 190: int ip_forwsrcrt = IPFORWSRCRT;
1.27 thorpej 191: int ip_directedbcast = IPDIRECTEDBCAST;
1.47 cjs 192: int ip_allowsrcrt = IPALLOWSRCRT;
1.53 kml 193: int ip_mtudisc = IPMTUDISC;
1.156 itojun 194: int ip_mtudisc_timeout = IPMTUDISCTIMEOUT;
1.1 cgd 195: #ifdef DIAGNOSTIC
196: int ipprintfs = 0;
197: #endif
1.184 jonathan 198:
199: int ip_do_randomid = 0;
200:
1.165 christos 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: */
214: int ip_checkinterface = 0;
215:
1.1 cgd 216:
1.60 kml 217: struct rttimer_queue *ip_mtudisc_timeout_q = NULL;
218:
1.150 matt 219: u_long in_ifaddrhash; /* size of hash table - 1 */
220: int in_ifaddrentries; /* total number of addrs */
1.212 perry 221: struct in_ifaddrhead in_ifaddrhead;
1.57 tls 222: struct in_ifaddrhashhead *in_ifaddrhashtbl;
1.166 matt 223: u_long in_multihash; /* size of hash table - 1 */
224: int in_multientries; /* total number of addrs */
225: struct in_multihashhead *in_multihashtbl;
1.13 mycroft 226: struct ifqueue ipintrq;
1.286 tls 227:
1.291 rmind 228: ipid_state_t * ip_ids;
1.183 jonathan 229: uint16_t ip_id;
1.75 thorpej 230:
1.266 thorpej 231: percpu_t *ipstat_percpu;
232:
1.121 thorpej 233: #ifdef PFIL_HOOKS
234: struct pfil_head inet_pfil_hook;
235: #endif
236:
1.275 pooka 237: struct pool inmulti_pool;
1.72 thorpej 238:
1.135 thorpej 239: #ifdef INET_CSUM_COUNTERS
240: #include <sys/device.h>
241:
242: struct evcnt ip_hwcsum_bad = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
243: NULL, "inet", "hwcsum bad");
244: struct evcnt ip_hwcsum_ok = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
245: NULL, "inet", "hwcsum ok");
246: struct evcnt ip_swcsum = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
247: NULL, "inet", "swcsum");
248:
249: #define INET_CSUM_COUNTER_INCR(ev) (ev)->ev_count++
250:
1.201 matt 251: EVCNT_ATTACH_STATIC(ip_hwcsum_bad);
252: EVCNT_ATTACH_STATIC(ip_hwcsum_ok);
253: EVCNT_ATTACH_STATIC(ip_swcsum);
254:
1.135 thorpej 255: #else
256:
257: #define INET_CSUM_COUNTER_INCR(ev) /* nothing */
258:
259: #endif /* INET_CSUM_COUNTERS */
260:
1.1 cgd 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: */
268: int ip_nhops = 0;
269: static 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:
1.295 dyoung 276: static int ip_drainwanted;
277:
1.210 perry 278: static void save_rte(u_char *, struct in_addr);
1.35 mycroft 279:
1.164 matt 280: #ifdef MBUFTRACE
1.234 dogcow 281: struct mowner ip_rx_mowner = MOWNER_INIT("internet", "rx");
282: struct mowner ip_tx_mowner = MOWNER_INIT("internet", "tx");
1.164 matt 283: #endif
284:
1.284 pooka 285: static void sysctl_net_inet_ip_setup(struct sysctllog **);
286:
1.1 cgd 287: /*
288: * IP initialization: fill in IP protocol switch table.
289: * All protocols not implemented in kernel go to raw IP protocol handler.
290: */
1.8 mycroft 291: void
1.211 perry 292: ip_init(void)
1.1 cgd 293: {
1.199 matt 294: const struct protosw *pr;
1.109 augustss 295: int i;
1.1 cgd 296:
1.284 pooka 297: sysctl_net_inet_ip_setup(NULL);
298:
1.275 pooka 299: pool_init(&inmulti_pool, sizeof(struct in_multi), 0, 0, 0, "inmltpl",
300: NULL, IPL_SOFTNET);
301:
1.1 cgd 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;
1.192 jonathan 312:
1.288 rmind 313: ip_reass_init();
1.190 jonathan 314:
1.291 rmind 315: ip_ids = ip_id_init();
1.227 kardel 316: ip_id = time_second & 0xfffff;
1.194 jonathan 317:
1.294 dyoung 318: ipintrq.ifq_maxlen = IFQ_MAXLEN;
1.194 jonathan 319:
1.181 jonathan 320: TAILQ_INIT(&in_ifaddrhead);
1.272 ad 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);
1.160 itojun 325: ip_mtudisc_timeout_q = rt_timer_queue_create(ip_mtudisc_timeout);
1.73 thorpej 326: #ifdef GATEWAY
1.248 liamjfoy 327: ipflow_init(ip_hashsize);
1.73 thorpej 328: #endif
1.121 thorpej 329:
330: #ifdef PFIL_HOOKS
331: /* Register our Packet Filter hook. */
1.126 thorpej 332: inet_pfil_hook.ph_type = PFIL_TYPE_AF;
333: inet_pfil_hook.ph_af = AF_INET;
1.121 thorpej 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 */
1.135 thorpej 339:
1.164 matt 340: #ifdef MBUFTRACE
341: MOWNER_ATTACH(&ip_tx_mowner);
342: MOWNER_ATTACH(&ip_rx_mowner);
343: #endif /* MBUFTRACE */
1.266 thorpej 344:
345: ipstat_percpu = percpu_alloc(sizeof(uint64_t) * IP_NSTATS);
1.1 cgd 346: }
347:
1.229 christos 348: struct sockaddr_in ipaddr = {
349: .sin_len = sizeof(ipaddr),
350: .sin_family = AF_INET,
351: };
1.1 cgd 352: struct route ipforward_rt;
353:
354: /*
1.89 itojun 355: * IP software interrupt routine
356: */
357: void
1.211 perry 358: ipintr(void)
1.89 itojun 359: {
360: int s;
361: struct mbuf *m;
1.286 tls 362: struct ifqueue lcl_intrq;
363:
364: memset(&lcl_intrq, 0, sizeof(lcl_intrq));
1.89 itojun 365:
1.268 ad 366: mutex_enter(softnet_lock);
367: KERNEL_LOCK(1, NULL);
1.286 tls 368: if (!IF_IS_EMPTY(&ipintrq)) {
1.132 thorpej 369: s = splnet();
1.286 tls 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:
1.89 itojun 381: splx(s);
1.286 tls 382: }
383: KERNEL_UNLOCK_ONE(NULL);
384: while (!IF_IS_EMPTY(&lcl_intrq)) {
385: IF_DEQUEUE(&lcl_intrq, m);
1.268 ad 386: if (m == NULL)
387: break;
1.89 itojun 388: ip_input(m);
389: }
1.268 ad 390: mutex_exit(softnet_lock);
1.89 itojun 391: }
392:
393: /*
1.1 cgd 394: * Ip input routine. Checksum and byte swap header. If fragmented
395: * try to reassemble. Process options. Pass to next level.
396: */
1.8 mycroft 397: void
1.89 itojun 398: ip_input(struct mbuf *m)
1.1 cgd 399: {
1.109 augustss 400: struct ip *ip = NULL;
401: struct in_ifaddr *ia;
402: struct ifaddr *ifa;
1.288 rmind 403: int hlen = 0, len;
1.100 itojun 404: int downmatch;
1.165 christos 405: int checkif;
1.169 itojun 406: int srcrt = 0;
1.173 jonathan 407: #ifdef FAST_IPSEC
408: struct m_tag *mtag;
409: struct tdb_ident *tdbi;
410: struct secpolicy *sp;
1.289 rmind 411: int error, s;
1.173 jonathan 412: #endif /* FAST_IPSEC */
1.1 cgd 413:
1.164 matt 414: MCLAIM(m, &ip_rx_mowner);
1.289 rmind 415: KASSERT((m->m_flags & M_PKTHDR) != 0);
1.164 matt 416:
1.1 cgd 417: /*
418: * If no IP addresses have been set yet but the interfaces
419: * are receiving, can't do anything with incoming packets yet.
420: */
1.181 jonathan 421: if (TAILQ_FIRST(&in_ifaddrhead) == 0)
1.1 cgd 422: goto bad;
1.266 thorpej 423: IP_STATINC(IP_STAT_TOTAL);
1.154 thorpej 424: /*
425: * If the IP header is not aligned, slurp it up into a new
426: * mbuf with space for link headers, in the event we forward
427: * it. Otherwise, if it is aligned, make sure the entire
428: * base IP header is in the first mbuf of the chain.
429: */
1.244 christos 430: if (IP_HDR_ALIGNED_P(mtod(m, void *)) == 0) {
1.154 thorpej 431: if ((m = m_copyup(m, sizeof(struct ip),
432: (max_linkhdr + 3) & ~3)) == NULL) {
433: /* XXXJRT new stat, please */
1.266 thorpej 434: IP_STATINC(IP_STAT_TOOSMALL);
1.154 thorpej 435: return;
436: }
437: } else if (__predict_false(m->m_len < sizeof (struct ip))) {
438: if ((m = m_pullup(m, sizeof (struct ip))) == NULL) {
1.266 thorpej 439: IP_STATINC(IP_STAT_TOOSMALL);
1.154 thorpej 440: return;
441: }
1.1 cgd 442: }
443: ip = mtod(m, struct ip *);
1.13 mycroft 444: if (ip->ip_v != IPVERSION) {
1.266 thorpej 445: IP_STATINC(IP_STAT_BADVERS);
1.13 mycroft 446: goto bad;
447: }
1.1 cgd 448: hlen = ip->ip_hl << 2;
449: if (hlen < sizeof(struct ip)) { /* minimum header length */
1.266 thorpej 450: IP_STATINC(IP_STAT_BADHLEN);
1.1 cgd 451: goto bad;
452: }
453: if (hlen > m->m_len) {
1.298 liamjfoy 454: if ((m = m_pullup(m, hlen)) == NULL) {
1.266 thorpej 455: IP_STATINC(IP_STAT_BADHLEN);
1.89 itojun 456: return;
1.1 cgd 457: }
458: ip = mtod(m, struct ip *);
459: }
1.98 thorpej 460:
1.85 hwr 461: /*
1.99 thorpej 462: * RFC1122: packets with a multicast source address are
1.98 thorpej 463: * not allowed.
1.85 hwr 464: */
465: if (IN_MULTICAST(ip->ip_src.s_addr)) {
1.266 thorpej 466: IP_STATINC(IP_STAT_BADADDR);
1.85 hwr 467: goto bad;
1.129 itojun 468: }
469:
470: /* 127/8 must not appear on wire - RFC1122 */
471: if ((ntohl(ip->ip_dst.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET ||
472: (ntohl(ip->ip_src.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET) {
1.130 itojun 473: if ((m->m_pkthdr.rcvif->if_flags & IFF_LOOPBACK) == 0) {
1.266 thorpej 474: IP_STATINC(IP_STAT_BADADDR);
1.129 itojun 475: goto bad;
1.130 itojun 476: }
1.85 hwr 477: }
478:
1.135 thorpej 479: switch (m->m_pkthdr.csum_flags &
1.137 thorpej 480: ((m->m_pkthdr.rcvif->if_csum_flags_rx & M_CSUM_IPv4) |
1.135 thorpej 481: M_CSUM_IPv4_BAD)) {
482: case M_CSUM_IPv4|M_CSUM_IPv4_BAD:
483: INET_CSUM_COUNTER_INCR(&ip_hwcsum_bad);
484: goto badcsum;
485:
486: case M_CSUM_IPv4:
487: /* Checksum was okay. */
488: INET_CSUM_COUNTER_INCR(&ip_hwcsum_ok);
489: break;
490:
491: default:
1.206 thorpej 492: /*
493: * Must compute it ourselves. Maybe skip checksum on
494: * loopback interfaces.
495: */
496: if (__predict_true(!(m->m_pkthdr.rcvif->if_flags &
497: IFF_LOOPBACK) || ip_do_loopback_cksum)) {
498: INET_CSUM_COUNTER_INCR(&ip_swcsum);
499: if (in_cksum(m, hlen) != 0)
500: goto badcsum;
501: }
1.135 thorpej 502: break;
1.1 cgd 503: }
504:
1.121 thorpej 505: /* Retrieve the packet length. */
506: len = ntohs(ip->ip_len);
1.81 proff 507:
508: /*
509: * Check for additional length bogosity
510: */
1.84 proff 511: if (len < hlen) {
1.266 thorpej 512: IP_STATINC(IP_STAT_BADLEN);
1.81 proff 513: goto bad;
514: }
1.1 cgd 515:
516: /*
517: * Check that the amount of data in the buffers
518: * is as at least much as the IP header would have us expect.
519: * Trim mbufs if longer than we expect.
520: * Drop packet if shorter than we expect.
521: */
1.35 mycroft 522: if (m->m_pkthdr.len < len) {
1.266 thorpej 523: IP_STATINC(IP_STAT_TOOSHORT);
1.1 cgd 524: goto bad;
525: }
1.35 mycroft 526: if (m->m_pkthdr.len > len) {
1.1 cgd 527: if (m->m_len == m->m_pkthdr.len) {
1.35 mycroft 528: m->m_len = len;
529: m->m_pkthdr.len = len;
1.1 cgd 530: } else
1.35 mycroft 531: m_adj(m, len - m->m_pkthdr.len);
1.1 cgd 532: }
533:
1.64 thorpej 534: /*
535: * Assume that we can create a fast-forward IP flow entry
536: * based on this packet.
537: */
538: m->m_flags |= M_CANFASTFWD;
539:
1.36 mrg 540: #ifdef PFIL_HOOKS
1.33 mrg 541: /*
1.64 thorpej 542: * Run through list of hooks for input packets. If there are any
543: * filters which require that additional packets in the flow are
544: * not fast-forwarded, they must clear the M_CANFASTFWD flag.
545: * Note that filters must _never_ set this flag, as another filter
546: * in the list may have previously cleared it.
1.33 mrg 547: */
1.127 itojun 548: /*
549: * let ipfilter look at packet on the wire,
550: * not the decapsulated packet.
551: */
1.299 ! drochner 552: #if defined(FAST_IPSEC)
1.186 scw 553: if (!ipsec_indone(m))
1.127 itojun 554: #else
555: if (1)
556: #endif
557: {
1.169 itojun 558: struct in_addr odst;
559:
560: odst = ip->ip_dst;
1.127 itojun 561: if (pfil_run_hooks(&inet_pfil_hook, &m, m->m_pkthdr.rcvif,
1.168 itojun 562: PFIL_IN) != 0)
563: return;
1.127 itojun 564: if (m == NULL)
565: return;
566: ip = mtod(m, struct ip *);
1.142 darrenr 567: hlen = ip->ip_hl << 2;
1.205 darrenr 568: /*
569: * XXX The setting of "srcrt" here is to prevent ip_forward()
570: * from generating ICMP redirects for packets that have
571: * been redirected by a hook back out on to the same LAN that
572: * they came from and is not an indication that the packet
573: * is being inffluenced by source routing options. This
574: * allows things like
575: * "rdr tlp0 0/0 port 80 -> 1.1.1.200 3128 tcp"
576: * where tlp0 is both on the 1.1.1.0/24 network and is the
577: * default route for hosts on 1.1.1.0/24. Of course this
578: * also requires a "map tlp0 ..." to complete the story.
579: * One might argue whether or not this kind of network config.
1.212 perry 580: * should be supported in this manner...
1.205 darrenr 581: */
1.169 itojun 582: srcrt = (odst.s_addr != ip->ip_dst.s_addr);
1.127 itojun 583: }
1.36 mrg 584: #endif /* PFIL_HOOKS */
1.123 thorpej 585:
586: #ifdef ALTQ
587: /* XXX Temporary until ALTQ is changed to use a pfil hook */
588: if (altq_input != NULL && (*altq_input)(m, AF_INET) == 0) {
589: /* packet dropped by traffic conditioner */
590: return;
591: }
592: #endif
1.121 thorpej 593:
594: /*
1.1 cgd 595: * Process options and, if not destined for us,
596: * ship it on. ip_dooptions returns 1 when an
597: * error was detected (causing an icmp message
598: * to be sent and the original packet to be freed).
599: */
600: ip_nhops = 0; /* for source routed packets */
601: if (hlen > sizeof (struct ip) && ip_dooptions(m))
1.89 itojun 602: return;
1.1 cgd 603:
604: /*
1.165 christos 605: * Enable a consistency check between the destination address
606: * and the arrival interface for a unicast packet (the RFC 1122
607: * strong ES model) if IP forwarding is disabled and the packet
608: * is not locally generated.
609: *
610: * XXX - Checking also should be disabled if the destination
611: * address is ipnat'ed to a different interface.
612: *
613: * XXX - Checking is incompatible with IP aliases added
614: * to the loopback interface instead of the interface where
615: * the packets are received.
616: *
617: * XXX - We need to add a per ifaddr flag for this so that
618: * we get finer grain control.
619: */
620: checkif = ip_checkinterface && (ipforwarding == 0) &&
621: (m->m_pkthdr.rcvif != NULL) &&
622: ((m->m_pkthdr.rcvif->if_flags & IFF_LOOPBACK) == 0);
623:
624: /*
1.1 cgd 625: * Check our list of addresses, to see if the packet is for us.
1.100 itojun 626: *
627: * Traditional 4.4BSD did not consult IFF_UP at all.
628: * The behavior here is to treat addresses on !IFF_UP interface
629: * as not mine.
1.1 cgd 630: */
1.100 itojun 631: downmatch = 0;
1.140 matt 632: LIST_FOREACH(ia, &IN_IFADDR_HASH(ip->ip_dst.s_addr), ia_hash) {
1.97 itojun 633: if (in_hosteq(ia->ia_addr.sin_addr, ip->ip_dst)) {
1.165 christos 634: if (checkif && ia->ia_ifp != m->m_pkthdr.rcvif)
635: continue;
1.97 itojun 636: if ((ia->ia_ifp->if_flags & IFF_UP) != 0)
637: break;
1.100 itojun 638: else
639: downmatch++;
1.97 itojun 640: }
641: }
1.86 thorpej 642: if (ia != NULL)
643: goto ours;
1.225 christos 644: if (m->m_pkthdr.rcvif && m->m_pkthdr.rcvif->if_flags & IFF_BROADCAST) {
1.209 matt 645: IFADDR_FOREACH(ifa, m->m_pkthdr.rcvif) {
1.140 matt 646: if (ifa->ifa_addr->sa_family != AF_INET)
647: continue;
1.57 tls 648: ia = ifatoia(ifa);
1.35 mycroft 649: if (in_hosteq(ip->ip_dst, ia->ia_broadaddr.sin_addr) ||
650: in_hosteq(ip->ip_dst, ia->ia_netbroadcast) ||
1.20 mycroft 651: /*
652: * Look for all-0's host part (old broadcast addr),
653: * either for subnet or net.
654: */
655: ip->ip_dst.s_addr == ia->ia_subnet ||
1.18 mycroft 656: ip->ip_dst.s_addr == ia->ia_net)
1.1 cgd 657: goto ours;
1.57 tls 658: /*
659: * An interface with IP address zero accepts
660: * all packets that arrive on that interface.
661: */
662: if (in_nullhost(ia->ia_addr.sin_addr))
663: goto ours;
1.1 cgd 664: }
665: }
1.18 mycroft 666: if (IN_MULTICAST(ip->ip_dst.s_addr)) {
1.4 hpeyerl 667: struct in_multi *inm;
668: #ifdef MROUTING
669: extern struct socket *ip_mrouter;
1.10 brezak 670:
1.4 hpeyerl 671: if (ip_mrouter) {
672: /*
673: * If we are acting as a multicast router, all
674: * incoming multicast packets are passed to the
675: * kernel-level multicast forwarding function.
676: * The packet is returned (relatively) intact; if
677: * ip_mforward() returns a non-zero value, the packet
678: * must be discarded, else it may be accepted below.
679: *
680: * (The IP ident field is put in the same byte order
681: * as expected when ip_mforward() is called from
682: * ip_output().)
683: */
1.13 mycroft 684: if (ip_mforward(m, m->m_pkthdr.rcvif) != 0) {
1.266 thorpej 685: IP_STATINC(IP_STAT_CANTFORWARD);
1.4 hpeyerl 686: m_freem(m);
1.89 itojun 687: return;
1.4 hpeyerl 688: }
689:
690: /*
691: * The process-level routing demon needs to receive
692: * all multicast IGMP packets, whether or not this
693: * host belongs to their destination groups.
694: */
695: if (ip->ip_p == IPPROTO_IGMP)
696: goto ours;
1.266 thorpej 697: IP_STATINC(IP_STAT_CANTFORWARD);
1.4 hpeyerl 698: }
699: #endif
700: /*
701: * See if we belong to the destination multicast group on the
702: * arrival interface.
703: */
704: IN_LOOKUP_MULTI(ip->ip_dst, m->m_pkthdr.rcvif, inm);
705: if (inm == NULL) {
1.266 thorpej 706: IP_STATINC(IP_STAT_CANTFORWARD);
1.4 hpeyerl 707: m_freem(m);
1.89 itojun 708: return;
1.4 hpeyerl 709: }
710: goto ours;
711: }
1.19 mycroft 712: if (ip->ip_dst.s_addr == INADDR_BROADCAST ||
1.35 mycroft 713: in_nullhost(ip->ip_dst))
1.1 cgd 714: goto ours;
715:
716: /*
717: * Not for us; forward if possible and desirable.
718: */
719: if (ipforwarding == 0) {
1.266 thorpej 720: IP_STATINC(IP_STAT_CANTFORWARD);
1.1 cgd 721: m_freem(m);
1.100 itojun 722: } else {
723: /*
724: * If ip_dst matched any of my address on !IFF_UP interface,
725: * and there's no IFF_UP interface that matches ip_dst,
726: * send icmp unreach. Forwarding it will result in in-kernel
727: * forwarding loop till TTL goes to 0.
728: */
729: if (downmatch) {
730: icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_HOST, 0, 0);
1.266 thorpej 731: IP_STATINC(IP_STAT_CANTFORWARD);
1.100 itojun 732: return;
733: }
1.173 jonathan 734: #ifdef FAST_IPSEC
735: mtag = m_tag_find(m, PACKET_TAG_IPSEC_IN_DONE, NULL);
736: s = splsoftnet();
737: if (mtag != NULL) {
738: tdbi = (struct tdb_ident *)(mtag + 1);
739: sp = ipsec_getpolicy(tdbi, IPSEC_DIR_INBOUND);
740: } else {
741: sp = ipsec_getpolicybyaddr(m, IPSEC_DIR_INBOUND,
1.212 perry 742: IP_FORWARDING, &error);
1.173 jonathan 743: }
744: if (sp == NULL) { /* NB: can happen if error */
745: splx(s);
746: /*XXX error stat???*/
747: DPRINTF(("ip_input: no SP for forwarding\n")); /*XXX*/
748: goto bad;
749: }
750:
751: /*
752: * Check security policy against packet attributes.
753: */
754: error = ipsec_in_reject(sp, m);
755: KEY_FREESP(&sp);
756: splx(s);
757: if (error) {
1.266 thorpej 758: IP_STATINC(IP_STAT_CANTFORWARD);
1.173 jonathan 759: goto bad;
1.193 scw 760: }
761:
762: /*
763: * Peek at the outbound SP for this packet to determine if
764: * it's a Fast Forward candidate.
765: */
766: mtag = m_tag_find(m, PACKET_TAG_IPSEC_PENDING_TDB, NULL);
767: if (mtag != NULL)
768: m->m_flags &= ~M_CANFASTFWD;
769: else {
770: s = splsoftnet();
771: sp = ipsec4_checkpolicy(m, IPSEC_DIR_OUTBOUND,
772: (IP_FORWARDING |
773: (ip_directedbcast ? IP_ALLOWBROADCAST : 0)),
774: &error, NULL);
775: if (sp != NULL) {
776: m->m_flags &= ~M_CANFASTFWD;
777: KEY_FREESP(&sp);
778: }
779: splx(s);
1.173 jonathan 780: }
781: #endif /* FAST_IPSEC */
1.145 itojun 782:
1.169 itojun 783: ip_forward(m, srcrt);
1.100 itojun 784: }
1.89 itojun 785: return;
1.1 cgd 786:
787: ours:
788: /*
789: * If offset or IP_MF are set, must reassemble.
790: */
1.155 itojun 791: if (ip->ip_off & ~htons(IP_DF|IP_RF)) {
1.289 rmind 792: /*
793: * Pass to IP reassembly mechanism.
794: */
1.290 rmind 795: if (ip_reass_packet(&m, ip) != 0) {
1.289 rmind 796: /* Failed; invalid fragment(s) or packet. */
1.288 rmind 797: goto bad;
1.16 cgd 798: }
1.290 rmind 799: if (m == NULL) {
1.289 rmind 800: /* More fragments should come; silently return. */
801: return;
802: }
1.290 rmind 803: /*
804: * Reassembly is done, we have the final packet.
805: * Updated cached data in local variable(s).
806: */
1.289 rmind 807: ip = mtod(m, struct ip *);
808: hlen = ip->ip_hl << 2;
1.79 mycroft 809: }
1.128 itojun 810:
1.226 liamjfoy 811: #ifdef FAST_IPSEC
1.173 jonathan 812: /*
813: * enforce IPsec policy checking if we are seeing last header.
814: * note that we do not visit this with protocols with pcb layer
815: * code - like udp/tcp/raw ip.
816: */
817: if ((inetsw[ip_protox[ip->ip_p]].pr_flags & PR_LASTHDR) != 0) {
818: /*
819: * Check if the packet has already had IPsec processing
820: * done. If so, then just pass it along. This tag gets
821: * set during AH, ESP, etc. input handling, before the
822: * packet is returned to the ip input queue for delivery.
1.212 perry 823: */
1.173 jonathan 824: mtag = m_tag_find(m, PACKET_TAG_IPSEC_IN_DONE, NULL);
825: s = splsoftnet();
826: if (mtag != NULL) {
827: tdbi = (struct tdb_ident *)(mtag + 1);
828: sp = ipsec_getpolicy(tdbi, IPSEC_DIR_INBOUND);
829: } else {
830: sp = ipsec_getpolicybyaddr(m, IPSEC_DIR_INBOUND,
1.212 perry 831: IP_FORWARDING, &error);
1.173 jonathan 832: }
833: if (sp != NULL) {
834: /*
835: * Check security policy against packet attributes.
836: */
837: error = ipsec_in_reject(sp, m);
838: KEY_FREESP(&sp);
839: } else {
840: /* XXX error stat??? */
841: error = EINVAL;
842: DPRINTF(("ip_input: no SP, packet discarded\n"));/*XXX*/
843: }
844: splx(s);
845: if (error)
846: goto bad;
847: }
848: #endif /* FAST_IPSEC */
1.1 cgd 849:
850: /*
851: * Switch out to protocol's input routine.
852: */
1.82 aidan 853: #if IFA_STATS
1.122 itojun 854: if (ia && ip)
1.155 itojun 855: ia->ia_ifa.ifa_data.ifad_inbytes += ntohs(ip->ip_len);
1.82 aidan 856: #endif
1.266 thorpej 857: IP_STATINC(IP_STAT_DELIVERED);
1.89 itojun 858: {
859: int off = hlen, nh = ip->ip_p;
860:
861: (*inetsw[ip_protox[nh]].pr_input)(m, off, nh);
862: return;
863: }
1.1 cgd 864: bad:
865: m_freem(m);
1.135 thorpej 866: return;
867:
868: badcsum:
1.266 thorpej 869: IP_STATINC(IP_STAT_BADSUM);
1.135 thorpej 870: m_freem(m);
1.1 cgd 871: }
872:
873: /*
1.288 rmind 874: * IP timer processing.
1.1 cgd 875: */
1.8 mycroft 876: void
1.211 perry 877: ip_slowtimo(void)
1.1 cgd 878: {
1.268 ad 879:
880: mutex_enter(softnet_lock);
881: KERNEL_LOCK(1, NULL);
1.1 cgd 882:
1.288 rmind 883: ip_reass_slowtimo();
1.268 ad 884:
885: KERNEL_UNLOCK_ONE(NULL);
886: mutex_exit(softnet_lock);
1.1 cgd 887: }
888:
889: /*
1.288 rmind 890: * IP drain processing.
1.1 cgd 891: */
1.8 mycroft 892: void
1.211 perry 893: ip_drain(void)
1.1 cgd 894: {
895:
1.268 ad 896: KERNEL_LOCK(1, NULL);
1.288 rmind 897: ip_reass_drain();
1.268 ad 898: KERNEL_UNLOCK_ONE(NULL);
1.1 cgd 899: }
900:
901: /*
902: * Do option processing on a datagram,
903: * possibly discarding it if bad options are encountered,
904: * or forwarding it if source-routed.
905: * Returns 1 if packet has been forwarded/freed,
906: * 0 if the packet should be processed further.
907: */
1.8 mycroft 908: int
1.211 perry 909: ip_dooptions(struct mbuf *m)
1.1 cgd 910: {
1.109 augustss 911: struct ip *ip = mtod(m, struct ip *);
912: u_char *cp, *cp0;
913: struct ip_timestamp *ipt;
914: struct in_ifaddr *ia;
1.1 cgd 915: int opt, optlen, cnt, off, code, type = ICMP_PARAMPROB, forward = 0;
1.104 thorpej 916: struct in_addr dst;
1.1 cgd 917: n_time ntime;
918:
1.13 mycroft 919: dst = ip->ip_dst;
1.1 cgd 920: cp = (u_char *)(ip + 1);
921: cnt = (ip->ip_hl << 2) - sizeof (struct ip);
922: for (; cnt > 0; cnt -= optlen, cp += optlen) {
923: opt = cp[IPOPT_OPTVAL];
924: if (opt == IPOPT_EOL)
925: break;
926: if (opt == IPOPT_NOP)
927: optlen = 1;
928: else {
1.113 itojun 929: if (cnt < IPOPT_OLEN + sizeof(*cp)) {
930: code = &cp[IPOPT_OLEN] - (u_char *)ip;
931: goto bad;
932: }
1.1 cgd 933: optlen = cp[IPOPT_OLEN];
1.114 itojun 934: if (optlen < IPOPT_OLEN + sizeof(*cp) || optlen > cnt) {
1.1 cgd 935: code = &cp[IPOPT_OLEN] - (u_char *)ip;
936: goto bad;
937: }
938: }
939: switch (opt) {
940:
941: default:
942: break;
943:
944: /*
945: * Source routing with record.
946: * Find interface with current destination address.
947: * If none on this machine then drop if strictly routed,
948: * or do nothing if loosely routed.
949: * Record interface address and bring up next address
950: * component. If strictly routed make sure next
951: * address is on directly accessible net.
952: */
953: case IPOPT_LSRR:
954: case IPOPT_SSRR:
1.47 cjs 955: if (ip_allowsrcrt == 0) {
956: type = ICMP_UNREACH;
957: code = ICMP_UNREACH_NET_PROHIB;
958: goto bad;
959: }
1.114 itojun 960: if (optlen < IPOPT_OFFSET + sizeof(*cp)) {
961: code = &cp[IPOPT_OLEN] - (u_char *)ip;
962: goto bad;
963: }
1.1 cgd 964: if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) {
965: code = &cp[IPOPT_OFFSET] - (u_char *)ip;
966: goto bad;
967: }
968: ipaddr.sin_addr = ip->ip_dst;
1.19 mycroft 969: ia = ifatoia(ifa_ifwithaddr(sintosa(&ipaddr)));
1.1 cgd 970: if (ia == 0) {
971: if (opt == IPOPT_SSRR) {
972: type = ICMP_UNREACH;
973: code = ICMP_UNREACH_SRCFAIL;
974: goto bad;
975: }
976: /*
977: * Loose routing, and not at next destination
978: * yet; nothing to do except forward.
979: */
980: break;
981: }
982: off--; /* 0 origin */
1.112 sommerfe 983: if ((off + sizeof(struct in_addr)) > optlen) {
1.1 cgd 984: /*
985: * End of source route. Should be for us.
986: */
987: save_rte(cp, ip->ip_src);
988: break;
989: }
990: /*
991: * locate outgoing interface
992: */
1.281 tsutsui 993: memcpy((void *)&ipaddr.sin_addr, (void *)(cp + off),
1.1 cgd 994: sizeof(ipaddr.sin_addr));
1.96 thorpej 995: if (opt == IPOPT_SSRR)
1.196 itojun 996: ia = ifatoia(ifa_ifwithladdr(sintosa(&ipaddr)));
1.96 thorpej 997: else
1.1 cgd 998: ia = ip_rtaddr(ipaddr.sin_addr);
999: if (ia == 0) {
1000: type = ICMP_UNREACH;
1001: code = ICMP_UNREACH_SRCFAIL;
1002: goto bad;
1003: }
1004: ip->ip_dst = ipaddr.sin_addr;
1.244 christos 1005: bcopy((void *)&ia->ia_addr.sin_addr,
1006: (void *)(cp + off), sizeof(struct in_addr));
1.1 cgd 1007: cp[IPOPT_OFFSET] += sizeof(struct in_addr);
1.13 mycroft 1008: /*
1009: * Let ip_intr's mcast routing check handle mcast pkts
1010: */
1.18 mycroft 1011: forward = !IN_MULTICAST(ip->ip_dst.s_addr);
1.1 cgd 1012: break;
1013:
1014: case IPOPT_RR:
1.114 itojun 1015: if (optlen < IPOPT_OFFSET + sizeof(*cp)) {
1016: code = &cp[IPOPT_OLEN] - (u_char *)ip;
1017: goto bad;
1018: }
1.1 cgd 1019: if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) {
1020: code = &cp[IPOPT_OFFSET] - (u_char *)ip;
1021: goto bad;
1022: }
1023: /*
1024: * If no space remains, ignore.
1025: */
1026: off--; /* 0 origin */
1.112 sommerfe 1027: if ((off + sizeof(struct in_addr)) > optlen)
1.1 cgd 1028: break;
1.281 tsutsui 1029: memcpy((void *)&ipaddr.sin_addr, (void *)(&ip->ip_dst),
1.1 cgd 1030: sizeof(ipaddr.sin_addr));
1031: /*
1032: * locate outgoing interface; if we're the destination,
1033: * use the incoming interface (should be same).
1034: */
1.96 thorpej 1035: if ((ia = ifatoia(ifa_ifwithaddr(sintosa(&ipaddr))))
1036: == NULL &&
1037: (ia = ip_rtaddr(ipaddr.sin_addr)) == NULL) {
1.1 cgd 1038: type = ICMP_UNREACH;
1039: code = ICMP_UNREACH_HOST;
1040: goto bad;
1041: }
1.244 christos 1042: bcopy((void *)&ia->ia_addr.sin_addr,
1043: (void *)(cp + off), sizeof(struct in_addr));
1.1 cgd 1044: cp[IPOPT_OFFSET] += sizeof(struct in_addr);
1045: break;
1046:
1047: case IPOPT_TS:
1048: code = cp - (u_char *)ip;
1049: ipt = (struct ip_timestamp *)cp;
1.114 itojun 1050: if (ipt->ipt_len < 4 || ipt->ipt_len > 40) {
1051: code = (u_char *)&ipt->ipt_len - (u_char *)ip;
1.1 cgd 1052: goto bad;
1.114 itojun 1053: }
1054: if (ipt->ipt_ptr < 5) {
1055: code = (u_char *)&ipt->ipt_ptr - (u_char *)ip;
1056: goto bad;
1057: }
1.15 cgd 1058: if (ipt->ipt_ptr > ipt->ipt_len - sizeof (int32_t)) {
1.114 itojun 1059: if (++ipt->ipt_oflw == 0) {
1060: code = (u_char *)&ipt->ipt_ptr -
1061: (u_char *)ip;
1.1 cgd 1062: goto bad;
1.114 itojun 1063: }
1.1 cgd 1064: break;
1065: }
1.104 thorpej 1066: cp0 = (cp + ipt->ipt_ptr - 1);
1.1 cgd 1067: switch (ipt->ipt_flg) {
1068:
1069: case IPOPT_TS_TSONLY:
1070: break;
1071:
1072: case IPOPT_TS_TSANDADDR:
1.66 thorpej 1073: if (ipt->ipt_ptr - 1 + sizeof(n_time) +
1.114 itojun 1074: sizeof(struct in_addr) > ipt->ipt_len) {
1075: code = (u_char *)&ipt->ipt_ptr -
1076: (u_char *)ip;
1.1 cgd 1077: goto bad;
1.114 itojun 1078: }
1.13 mycroft 1079: ipaddr.sin_addr = dst;
1.96 thorpej 1080: ia = ifatoia(ifaof_ifpforaddr(sintosa(&ipaddr),
1081: m->m_pkthdr.rcvif));
1.13 mycroft 1082: if (ia == 0)
1083: continue;
1.104 thorpej 1084: bcopy(&ia->ia_addr.sin_addr,
1085: cp0, sizeof(struct in_addr));
1.1 cgd 1086: ipt->ipt_ptr += sizeof(struct in_addr);
1087: break;
1088:
1089: case IPOPT_TS_PRESPEC:
1.66 thorpej 1090: if (ipt->ipt_ptr - 1 + sizeof(n_time) +
1.114 itojun 1091: sizeof(struct in_addr) > ipt->ipt_len) {
1092: code = (u_char *)&ipt->ipt_ptr -
1093: (u_char *)ip;
1.1 cgd 1094: goto bad;
1.114 itojun 1095: }
1.281 tsutsui 1096: memcpy(&ipaddr.sin_addr, cp0,
1.1 cgd 1097: sizeof(struct in_addr));
1.96 thorpej 1098: if (ifatoia(ifa_ifwithaddr(sintosa(&ipaddr)))
1099: == NULL)
1.1 cgd 1100: continue;
1101: ipt->ipt_ptr += sizeof(struct in_addr);
1102: break;
1103:
1104: default:
1.114 itojun 1105: /* XXX can't take &ipt->ipt_flg */
1106: code = (u_char *)&ipt->ipt_ptr -
1107: (u_char *)ip + 1;
1.1 cgd 1108: goto bad;
1109: }
1110: ntime = iptime();
1.107 thorpej 1111: cp0 = (u_char *) &ntime; /* XXX grumble, GCC... */
1.244 christos 1112: memmove((char *)cp + ipt->ipt_ptr - 1, cp0,
1.1 cgd 1113: sizeof(n_time));
1114: ipt->ipt_ptr += sizeof(n_time);
1115: }
1116: }
1117: if (forward) {
1.26 thorpej 1118: if (ip_forwsrcrt == 0) {
1119: type = ICMP_UNREACH;
1120: code = ICMP_UNREACH_SRCFAIL;
1121: goto bad;
1122: }
1.1 cgd 1123: ip_forward(m, 1);
1124: return (1);
1.13 mycroft 1125: }
1126: return (0);
1.1 cgd 1127: bad:
1.13 mycroft 1128: icmp_error(m, type, code, 0, 0);
1.266 thorpej 1129: IP_STATINC(IP_STAT_BADOPTIONS);
1.1 cgd 1130: return (1);
1131: }
1132:
1133: /*
1134: * Given address of next destination (final or next hop),
1135: * return internet address info of interface to be used to get there.
1136: */
1137: struct in_ifaddr *
1.211 perry 1138: ip_rtaddr(struct in_addr dst)
1.1 cgd 1139: {
1.249 dyoung 1140: struct rtentry *rt;
1141: union {
1142: struct sockaddr dst;
1143: struct sockaddr_in dst4;
1144: } u;
1145:
1146: sockaddr_in_init(&u.dst4, &dst, 0);
1147:
1148: if ((rt = rtcache_lookup(&ipforward_rt, &u.dst)) == NULL)
1149: return NULL;
1150:
1151: return ifatoia(rt->rt_ifa);
1.1 cgd 1152: }
1153:
1154: /*
1155: * Save incoming source route for use in replies,
1156: * to be picked up later by ip_srcroute if the receiver is interested.
1157: */
1.13 mycroft 1158: void
1.211 perry 1159: save_rte(u_char *option, struct in_addr dst)
1.1 cgd 1160: {
1161: unsigned olen;
1162:
1163: olen = option[IPOPT_OLEN];
1164: #ifdef DIAGNOSTIC
1165: if (ipprintfs)
1.39 christos 1166: printf("save_rte: olen %d\n", olen);
1.89 itojun 1167: #endif /* 0 */
1.1 cgd 1168: if (olen > sizeof(ip_srcrt) - (1 + sizeof(dst)))
1169: return;
1.281 tsutsui 1170: memcpy((void *)ip_srcrt.srcopt, (void *)option, olen);
1.1 cgd 1171: ip_nhops = (olen - IPOPT_OFFSET - 1) / sizeof(struct in_addr);
1172: ip_srcrt.dst = dst;
1173: }
1174:
1175: /*
1176: * Retrieve incoming source route for use in replies,
1177: * in the same form used by setsockopt.
1178: * The first hop is placed before the options, will be removed later.
1179: */
1180: struct mbuf *
1.211 perry 1181: ip_srcroute(void)
1.1 cgd 1182: {
1.109 augustss 1183: struct in_addr *p, *q;
1184: struct mbuf *m;
1.1 cgd 1185:
1186: if (ip_nhops == 0)
1.237 dyoung 1187: return NULL;
1.1 cgd 1188: m = m_get(M_DONTWAIT, MT_SOOPTS);
1189: if (m == 0)
1.237 dyoung 1190: return NULL;
1.1 cgd 1191:
1.164 matt 1192: MCLAIM(m, &inetdomain.dom_mowner);
1.13 mycroft 1193: #define OPTSIZ (sizeof(ip_srcrt.nop) + sizeof(ip_srcrt.srcopt))
1.1 cgd 1194:
1195: /* length is (nhops+1)*sizeof(addr) + sizeof(nop + srcrt header) */
1196: m->m_len = ip_nhops * sizeof(struct in_addr) + sizeof(struct in_addr) +
1197: OPTSIZ;
1198: #ifdef DIAGNOSTIC
1199: if (ipprintfs)
1.39 christos 1200: printf("ip_srcroute: nhops %d mlen %d", ip_nhops, m->m_len);
1.1 cgd 1201: #endif
1202:
1203: /*
1204: * First save first hop for return route
1205: */
1206: p = &ip_srcrt.route[ip_nhops - 1];
1207: *(mtod(m, struct in_addr *)) = *p--;
1208: #ifdef DIAGNOSTIC
1209: if (ipprintfs)
1.39 christos 1210: printf(" hops %x", ntohl(mtod(m, struct in_addr *)->s_addr));
1.1 cgd 1211: #endif
1212:
1213: /*
1214: * Copy option fields and padding (nop) to mbuf.
1215: */
1216: ip_srcrt.nop = IPOPT_NOP;
1217: ip_srcrt.srcopt[IPOPT_OFFSET] = IPOPT_MINOFF;
1.244 christos 1218: memmove(mtod(m, char *) + sizeof(struct in_addr), &ip_srcrt.nop,
1219: OPTSIZ);
1220: q = (struct in_addr *)(mtod(m, char *) +
1.1 cgd 1221: sizeof(struct in_addr) + OPTSIZ);
1222: #undef OPTSIZ
1223: /*
1224: * Record return path as an IP source route,
1225: * reversing the path (pointers are now aligned).
1226: */
1227: while (p >= ip_srcrt.route) {
1228: #ifdef DIAGNOSTIC
1229: if (ipprintfs)
1.39 christos 1230: printf(" %x", ntohl(q->s_addr));
1.1 cgd 1231: #endif
1232: *q++ = *p--;
1233: }
1234: /*
1235: * Last hop goes to final destination.
1236: */
1237: *q = ip_srcrt.dst;
1238: #ifdef DIAGNOSTIC
1239: if (ipprintfs)
1.39 christos 1240: printf(" %x\n", ntohl(q->s_addr));
1.1 cgd 1241: #endif
1242: return (m);
1243: }
1244:
1.139 matt 1245: const int inetctlerrmap[PRC_NCMDS] = {
1.256 yamt 1246: [PRC_MSGSIZE] = EMSGSIZE,
1247: [PRC_HOSTDEAD] = EHOSTDOWN,
1248: [PRC_HOSTUNREACH] = EHOSTUNREACH,
1249: [PRC_UNREACH_NET] = EHOSTUNREACH,
1250: [PRC_UNREACH_HOST] = EHOSTUNREACH,
1251: [PRC_UNREACH_PROTOCOL] = ECONNREFUSED,
1252: [PRC_UNREACH_PORT] = ECONNREFUSED,
1253: [PRC_UNREACH_SRCFAIL] = EHOSTUNREACH,
1254: [PRC_PARAMPROB] = ENOPROTOOPT,
1.1 cgd 1255: };
1256:
1.295 dyoung 1257: void
1258: ip_fasttimo(void)
1259: {
1260: if (ip_drainwanted) {
1261: ip_drain();
1262: ip_drainwanted = 0;
1263: }
1264: }
1265:
1266: void
1267: ip_drainstub(void)
1268: {
1269: ip_drainwanted = 1;
1270: }
1271:
1.1 cgd 1272: /*
1273: * Forward a packet. If some error occurs return the sender
1274: * an icmp packet. Note we can't always generate a meaningful
1275: * icmp message because icmp doesn't have a large enough repertoire
1276: * of codes and types.
1277: *
1278: * If not forwarding, just drop the packet. This could be confusing
1279: * if ipforwarding was zero but some routing protocol was advancing
1280: * us as a gateway to somewhere. However, we must let the routing
1281: * protocol deal with that.
1282: *
1283: * The srcrt parameter indicates whether the packet is being forwarded
1284: * via a source route.
1285: */
1.13 mycroft 1286: void
1.211 perry 1287: ip_forward(struct mbuf *m, int srcrt)
1.1 cgd 1288: {
1.109 augustss 1289: struct ip *ip = mtod(m, struct ip *);
1290: struct rtentry *rt;
1.220 christos 1291: int error, type = 0, code = 0, destmtu = 0;
1.1 cgd 1292: struct mbuf *mcopy;
1.13 mycroft 1293: n_long dest;
1.249 dyoung 1294: union {
1295: struct sockaddr dst;
1296: struct sockaddr_in dst4;
1297: } u;
1.164 matt 1298:
1299: /*
1300: * We are now in the output path.
1301: */
1302: MCLAIM(m, &ip_tx_mowner);
1.135 thorpej 1303:
1304: /*
1305: * Clear any in-bound checksum flags for this packet.
1306: */
1307: m->m_pkthdr.csum_flags = 0;
1.1 cgd 1308:
1.13 mycroft 1309: dest = 0;
1.1 cgd 1310: #ifdef DIAGNOSTIC
1.224 joerg 1311: if (ipprintfs) {
1312: printf("forward: src %s ", inet_ntoa(ip->ip_src));
1313: printf("dst %s ttl %x\n", inet_ntoa(ip->ip_dst), ip->ip_ttl);
1314: }
1.1 cgd 1315: #endif
1.93 sommerfe 1316: if (m->m_flags & (M_BCAST|M_MCAST) || in_canforward(ip->ip_dst) == 0) {
1.266 thorpej 1317: IP_STATINC(IP_STAT_CANTFORWARD);
1.1 cgd 1318: m_freem(m);
1319: return;
1320: }
1321: if (ip->ip_ttl <= IPTTLDEC) {
1.13 mycroft 1322: icmp_error(m, ICMP_TIMXCEED, ICMP_TIMXCEED_INTRANS, dest, 0);
1.1 cgd 1323: return;
1324: }
1325:
1.249 dyoung 1326: sockaddr_in_init(&u.dst4, &ip->ip_dst, 0);
1327: if ((rt = rtcache_lookup(&ipforward_rt, &u.dst)) == NULL) {
1328: icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_NET, dest, 0);
1329: return;
1.1 cgd 1330: }
1331:
1332: /*
1.34 mycroft 1333: * Save at most 68 bytes of the packet in case
1.1 cgd 1334: * we need to generate an ICMP message to the src.
1.119 itojun 1335: * Pullup to avoid sharing mbuf cluster between m and mcopy.
1.1 cgd 1336: */
1.155 itojun 1337: mcopy = m_copym(m, 0, imin(ntohs(ip->ip_len), 68), M_DONTWAIT);
1.119 itojun 1338: if (mcopy)
1339: mcopy = m_pullup(mcopy, ip->ip_hl << 2);
1.1 cgd 1340:
1.221 christos 1341: ip->ip_ttl -= IPTTLDEC;
1342:
1.1 cgd 1343: /*
1344: * If forwarding packet using same interface that it came in on,
1345: * perhaps should send a redirect to sender to shortcut a hop.
1346: * Only send redirect if source is sending directly to us,
1347: * and if packet was not source routed (or has any options).
1348: * Also, don't send redirect if forwarding using a default route
1349: * or a route modified by a redirect.
1350: */
1351: if (rt->rt_ifp == m->m_pkthdr.rcvif &&
1352: (rt->rt_flags & (RTF_DYNAMIC|RTF_MODIFIED)) == 0 &&
1.250 dyoung 1353: !in_nullhost(satocsin(rt_getkey(rt))->sin_addr) &&
1.1 cgd 1354: ipsendredirects && !srcrt) {
1.19 mycroft 1355: if (rt->rt_ifa &&
1356: (ip->ip_src.s_addr & ifatoia(rt->rt_ifa)->ia_subnetmask) ==
1357: ifatoia(rt->rt_ifa)->ia_subnet) {
1.77 thorpej 1358: if (rt->rt_flags & RTF_GATEWAY)
1359: dest = satosin(rt->rt_gateway)->sin_addr.s_addr;
1360: else
1361: dest = ip->ip_dst.s_addr;
1362: /*
1363: * Router requirements says to only send host
1364: * redirects.
1365: */
1366: type = ICMP_REDIRECT;
1367: code = ICMP_REDIRECT_HOST;
1.1 cgd 1368: #ifdef DIAGNOSTIC
1.77 thorpej 1369: if (ipprintfs)
1370: printf("redirect (%d) to %x\n", code,
1371: (u_int32_t)dest);
1.1 cgd 1372: #endif
1373: }
1374: }
1375:
1.238 dyoung 1376: error = ip_output(m, NULL, &ipforward_rt,
1.173 jonathan 1377: (IP_FORWARDING | (ip_directedbcast ? IP_ALLOWBROADCAST : 0)),
1.296 plunky 1378: NULL, NULL);
1.173 jonathan 1379:
1.1 cgd 1380: if (error)
1.266 thorpej 1381: IP_STATINC(IP_STAT_CANTFORWARD);
1.1 cgd 1382: else {
1.266 thorpej 1383: uint64_t *ips = IP_STAT_GETREF();
1384: ips[IP_STAT_FORWARD]++;
1385: if (type) {
1386: ips[IP_STAT_REDIRECTSENT]++;
1387: IP_STAT_PUTREF();
1388: } else {
1389: IP_STAT_PUTREF();
1.63 matt 1390: if (mcopy) {
1391: #ifdef GATEWAY
1.64 thorpej 1392: if (mcopy->m_flags & M_CANFASTFWD)
1393: ipflow_create(&ipforward_rt, mcopy);
1.63 matt 1394: #endif
1.1 cgd 1395: m_freem(mcopy);
1.63 matt 1396: }
1.1 cgd 1397: return;
1398: }
1399: }
1400: if (mcopy == NULL)
1401: return;
1.13 mycroft 1402:
1.1 cgd 1403: switch (error) {
1404:
1405: case 0: /* forwarded, but need redirect */
1406: /* type, code set above */
1407: break;
1408:
1409: case ENETUNREACH: /* shouldn't happen, checked above */
1410: case EHOSTUNREACH:
1411: case ENETDOWN:
1412: case EHOSTDOWN:
1413: default:
1414: type = ICMP_UNREACH;
1415: code = ICMP_UNREACH_HOST;
1416: break;
1417:
1418: case EMSGSIZE:
1419: type = ICMP_UNREACH;
1420: code = ICMP_UNREACH_NEEDFRAG;
1.263 cube 1421:
1.274 seanb 1422: if ((rt = rtcache_validate(&ipforward_rt)) != NULL)
1423: destmtu = rt->rt_ifp->if_mtu;
1.263 cube 1424:
1.299 ! drochner 1425: #if defined(FAST_IPSEC)
1.274 seanb 1426: {
1.263 cube 1427: /*
1428: * If the packet is routed over IPsec tunnel, tell the
1429: * originator the tunnel MTU.
1430: * tunnel MTU = if MTU - sizeof(IP) - ESP/AH hdrsiz
1431: * XXX quickhack!!!
1432: */
1433:
1.89 itojun 1434: struct secpolicy *sp;
1435: int ipsecerror;
1.95 itojun 1436: size_t ipsechdr;
1.89 itojun 1437: struct route *ro;
1438:
1439: sp = ipsec4_getpolicybyaddr(mcopy,
1.170 itojun 1440: IPSEC_DIR_OUTBOUND, IP_FORWARDING,
1441: &ipsecerror);
1.89 itojun 1442:
1.263 cube 1443: if (sp != NULL) {
1.89 itojun 1444: /* count IPsec header size */
1.95 itojun 1445: ipsechdr = ipsec4_hdrsiz(mcopy,
1.170 itojun 1446: IPSEC_DIR_OUTBOUND, NULL);
1.89 itojun 1447:
1448: /*
1449: * find the correct route for outer IPv4
1450: * header, compute tunnel MTU.
1451: */
1.220 christos 1452:
1.89 itojun 1453: if (sp->req != NULL
1.95 itojun 1454: && sp->req->sav != NULL
1455: && sp->req->sav->sah != NULL) {
1456: ro = &sp->req->sav->sah->sa_route;
1.274 seanb 1457: rt = rtcache_validate(ro);
1.257 dyoung 1458: if (rt && rt->rt_ifp) {
1.220 christos 1459: destmtu =
1.257 dyoung 1460: rt->rt_rmx.rmx_mtu ?
1461: rt->rt_rmx.rmx_mtu :
1462: rt->rt_ifp->if_mtu;
1.220 christos 1463: destmtu -= ipsechdr;
1.89 itojun 1464: }
1465: }
1466:
1.173 jonathan 1467: KEY_FREESP(&sp);
1.89 itojun 1468: }
1.274 seanb 1469: }
1.299 ! drochner 1470: #endif /*defined(FAST_IPSEC)*/
1.266 thorpej 1471: IP_STATINC(IP_STAT_CANTFRAG);
1.1 cgd 1472: break;
1473:
1474: case ENOBUFS:
1.143 itojun 1475: #if 1
1476: /*
1477: * a router should not generate ICMP_SOURCEQUENCH as
1478: * required in RFC1812 Requirements for IP Version 4 Routers.
1479: * source quench could be a big problem under DoS attacks,
1.149 wiz 1480: * or if the underlying interface is rate-limited.
1.143 itojun 1481: */
1482: if (mcopy)
1483: m_freem(mcopy);
1484: return;
1485: #else
1.1 cgd 1486: type = ICMP_SOURCEQUENCH;
1487: code = 0;
1488: break;
1.143 itojun 1489: #endif
1.1 cgd 1490: }
1.220 christos 1491: icmp_error(mcopy, type, code, dest, destmtu);
1.44 thorpej 1492: }
1493:
1494: void
1.211 perry 1495: ip_savecontrol(struct inpcb *inp, struct mbuf **mp, struct ip *ip,
1496: struct mbuf *m)
1.44 thorpej 1497: {
1498:
1.283 minskim 1499: if (inp->inp_socket->so_options & SO_TIMESTAMP
1.278 christos 1500: #ifdef SO_OTIMESTAMP
1.283 minskim 1501: || inp->inp_socket->so_options & SO_OTIMESTAMP
1.278 christos 1502: #endif
1503: ) {
1.44 thorpej 1504: struct timeval tv;
1505:
1506: microtime(&tv);
1.278 christos 1507: #ifdef SO_OTIMESTAMP
1508: if (inp->inp_socket->so_options & SO_OTIMESTAMP) {
1509: struct timeval50 tv50;
1510: timeval_to_timeval50(&tv, &tv50);
1511: *mp = sbcreatecontrol((void *) &tv50, sizeof(tv50),
1512: SCM_OTIMESTAMP, SOL_SOCKET);
1513: } else
1514: #endif
1.244 christos 1515: *mp = sbcreatecontrol((void *) &tv, sizeof(tv),
1.44 thorpej 1516: SCM_TIMESTAMP, SOL_SOCKET);
1517: if (*mp)
1518: mp = &(*mp)->m_next;
1519: }
1520: if (inp->inp_flags & INP_RECVDSTADDR) {
1.244 christos 1521: *mp = sbcreatecontrol((void *) &ip->ip_dst,
1.44 thorpej 1522: sizeof(struct in_addr), IP_RECVDSTADDR, IPPROTO_IP);
1523: if (*mp)
1524: mp = &(*mp)->m_next;
1525: }
1526: #ifdef notyet
1527: /*
1528: * XXX
1529: * Moving these out of udp_input() made them even more broken
1530: * than they already were.
1531: * - fenner@parc.xerox.com
1532: */
1533: /* options were tossed already */
1534: if (inp->inp_flags & INP_RECVOPTS) {
1.244 christos 1535: *mp = sbcreatecontrol((void *) opts_deleted_above,
1.44 thorpej 1536: sizeof(struct in_addr), IP_RECVOPTS, IPPROTO_IP);
1537: if (*mp)
1538: mp = &(*mp)->m_next;
1539: }
1540: /* ip_srcroute doesn't do what we want here, need to fix */
1541: if (inp->inp_flags & INP_RECVRETOPTS) {
1.244 christos 1542: *mp = sbcreatecontrol((void *) ip_srcroute(),
1.44 thorpej 1543: sizeof(struct in_addr), IP_RECVRETOPTS, IPPROTO_IP);
1544: if (*mp)
1545: mp = &(*mp)->m_next;
1546: }
1547: #endif
1548: if (inp->inp_flags & INP_RECVIF) {
1549: struct sockaddr_dl sdl;
1550:
1.252 dyoung 1551: sockaddr_dl_init(&sdl, sizeof(sdl),
1552: (m->m_pkthdr.rcvif != NULL)
1553: ? m->m_pkthdr.rcvif->if_index
1554: : 0,
1555: 0, NULL, 0, NULL, 0);
1.251 dyoung 1556: *mp = sbcreatecontrol(&sdl, sdl.sdl_len, IP_RECVIF, IPPROTO_IP);
1.44 thorpej 1557: if (*mp)
1558: mp = &(*mp)->m_next;
1559: }
1.282 minskim 1560: if (inp->inp_flags & INP_RECVTTL) {
1561: *mp = sbcreatecontrol((void *) &ip->ip_ttl,
1562: sizeof(uint8_t), IP_RECVTTL, IPPROTO_IP);
1563: if (*mp)
1564: mp = &(*mp)->m_next;
1565: }
1.13 mycroft 1566: }
1567:
1.189 atatat 1568: /*
1.228 elad 1569: * sysctl helper routine for net.inet.ip.forwsrcrt.
1570: */
1571: static int
1572: sysctl_net_inet_ip_forwsrcrt(SYSCTLFN_ARGS)
1573: {
1574: int error, tmp;
1575: struct sysctlnode node;
1576:
1577: node = *rnode;
1578: tmp = ip_forwsrcrt;
1579: node.sysctl_data = &tmp;
1580: error = sysctl_lookup(SYSCTLFN_CALL(&node));
1581: if (error || newp == NULL)
1582: return (error);
1583:
1.280 elad 1584: error = kauth_authorize_network(l->l_cred, KAUTH_NETWORK_FORWSRCRT,
1585: 0, NULL, NULL, NULL);
1586: if (error)
1587: return (error);
1.228 elad 1588:
1589: ip_forwsrcrt = tmp;
1590:
1591: return (0);
1592: }
1593:
1594: /*
1.189 atatat 1595: * sysctl helper routine for net.inet.ip.mtudisctimeout. checks the
1596: * range of the new value and tweaks timers if it changes.
1597: */
1598: static int
1599: sysctl_net_inet_ip_pmtudto(SYSCTLFN_ARGS)
1.13 mycroft 1600: {
1.189 atatat 1601: int error, tmp;
1602: struct sysctlnode node;
1603:
1604: node = *rnode;
1605: tmp = ip_mtudisc_timeout;
1606: node.sysctl_data = &tmp;
1607: error = sysctl_lookup(SYSCTLFN_CALL(&node));
1608: if (error || newp == NULL)
1609: return (error);
1610: if (tmp < 0)
1611: return (EINVAL);
1.52 thorpej 1612:
1.273 matt 1613: mutex_enter(softnet_lock);
1614:
1.189 atatat 1615: ip_mtudisc_timeout = tmp;
1616: rt_timer_queue_change(ip_mtudisc_timeout_q, ip_mtudisc_timeout);
1617:
1.273 matt 1618: mutex_exit(softnet_lock);
1619:
1.189 atatat 1620: return (0);
1621: }
1.54 lukem 1622:
1.65 matt 1623: #ifdef GATEWAY
1.189 atatat 1624: /*
1.247 liamjfoy 1625: * sysctl helper routine for net.inet.ip.maxflows.
1.189 atatat 1626: */
1627: static int
1628: sysctl_net_inet_ip_maxflows(SYSCTLFN_ARGS)
1629: {
1.273 matt 1630: int error;
1.67 thorpej 1631:
1.273 matt 1632: error = sysctl_lookup(SYSCTLFN_CALL(rnode));
1633: if (error || newp == NULL)
1634: return (error);
1635:
1636: mutex_enter(softnet_lock);
1637: KERNEL_LOCK(1, NULL);
1.212 perry 1638:
1.265 thorpej 1639: ipflow_prune();
1.273 matt 1640:
1641: KERNEL_UNLOCK_ONE(NULL);
1642: mutex_exit(softnet_lock);
1.144 martin 1643:
1.189 atatat 1644: return (0);
1645: }
1.248 liamjfoy 1646:
1647: static int
1648: sysctl_net_inet_ip_hashsize(SYSCTLFN_ARGS)
1.283 minskim 1649: {
1.248 liamjfoy 1650: int error, tmp;
1651: struct sysctlnode node;
1652:
1653: node = *rnode;
1654: tmp = ip_hashsize;
1655: node.sysctl_data = &tmp;
1656: error = sysctl_lookup(SYSCTLFN_CALL(&node));
1657: if (error || newp == NULL)
1658: return (error);
1659:
1660: if ((tmp & (tmp - 1)) == 0 && tmp != 0) {
1661: /*
1662: * Can only fail due to malloc()
1663: */
1.273 matt 1664: mutex_enter(softnet_lock);
1665: KERNEL_LOCK(1, NULL);
1666:
1667: error = ipflow_invalidate_all(tmp);
1668:
1669: KERNEL_UNLOCK_ONE(NULL);
1670: mutex_exit(softnet_lock);
1671:
1.248 liamjfoy 1672: } else {
1673: /*
1674: * EINVAL if not a power of 2
1675: */
1.273 matt 1676: error = EINVAL;
1.283 minskim 1677: }
1.248 liamjfoy 1678:
1.273 matt 1679: return error;
1.248 liamjfoy 1680: }
1.189 atatat 1681: #endif /* GATEWAY */
1.117 tron 1682:
1.266 thorpej 1683: static int
1684: sysctl_net_inet_ip_stats(SYSCTLFN_ARGS)
1685: {
1686:
1.271 thorpej 1687: return (NETSTAT_SYSCTL(ipstat_percpu, IP_NSTATS));
1.266 thorpej 1688: }
1.131 itojun 1689:
1.284 pooka 1690: static void
1691: sysctl_net_inet_ip_setup(struct sysctllog **clog)
1.189 atatat 1692: {
1693: extern int subnetsarelocal, hostzeroisbroadcast;
1.180 jonathan 1694:
1.197 atatat 1695: sysctl_createv(clog, 0, NULL, NULL,
1696: CTLFLAG_PERMANENT,
1.189 atatat 1697: CTLTYPE_NODE, "net", NULL,
1698: NULL, 0, NULL, 0,
1699: CTL_NET, CTL_EOL);
1.197 atatat 1700: sysctl_createv(clog, 0, NULL, NULL,
1701: CTLFLAG_PERMANENT,
1.203 atatat 1702: CTLTYPE_NODE, "inet",
1703: SYSCTL_DESCR("PF_INET related settings"),
1.189 atatat 1704: NULL, 0, NULL, 0,
1705: CTL_NET, PF_INET, CTL_EOL);
1.197 atatat 1706: sysctl_createv(clog, 0, NULL, NULL,
1707: CTLFLAG_PERMANENT,
1.203 atatat 1708: CTLTYPE_NODE, "ip",
1709: SYSCTL_DESCR("IPv4 related settings"),
1.189 atatat 1710: NULL, 0, NULL, 0,
1711: CTL_NET, PF_INET, IPPROTO_IP, CTL_EOL);
1.212 perry 1712:
1.197 atatat 1713: sysctl_createv(clog, 0, NULL, NULL,
1714: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1715: CTLTYPE_INT, "forwarding",
1716: SYSCTL_DESCR("Enable forwarding of INET datagrams"),
1.189 atatat 1717: NULL, 0, &ipforwarding, 0,
1718: CTL_NET, PF_INET, IPPROTO_IP,
1719: IPCTL_FORWARDING, CTL_EOL);
1.197 atatat 1720: sysctl_createv(clog, 0, NULL, NULL,
1721: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1722: CTLTYPE_INT, "redirect",
1723: SYSCTL_DESCR("Enable sending of ICMP redirect messages"),
1.189 atatat 1724: NULL, 0, &ipsendredirects, 0,
1725: CTL_NET, PF_INET, IPPROTO_IP,
1726: IPCTL_SENDREDIRECTS, CTL_EOL);
1.197 atatat 1727: sysctl_createv(clog, 0, NULL, NULL,
1728: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1729: CTLTYPE_INT, "ttl",
1730: SYSCTL_DESCR("Default TTL for an INET datagram"),
1.189 atatat 1731: NULL, 0, &ip_defttl, 0,
1732: CTL_NET, PF_INET, IPPROTO_IP,
1733: IPCTL_DEFTTL, CTL_EOL);
1734: #ifdef IPCTL_DEFMTU
1.197 atatat 1735: sysctl_createv(clog, 0, NULL, NULL,
1736: CTLFLAG_PERMANENT /* |CTLFLAG_READWRITE? */,
1.203 atatat 1737: CTLTYPE_INT, "mtu",
1738: SYSCTL_DESCR("Default MTA for an INET route"),
1.189 atatat 1739: NULL, 0, &ip_mtu, 0,
1740: CTL_NET, PF_INET, IPPROTO_IP,
1741: IPCTL_DEFMTU, CTL_EOL);
1742: #endif /* IPCTL_DEFMTU */
1.197 atatat 1743: sysctl_createv(clog, 0, NULL, NULL,
1.228 elad 1744: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1745: CTLTYPE_INT, "forwsrcrt",
1746: SYSCTL_DESCR("Enable forwarding of source-routed "
1747: "datagrams"),
1.228 elad 1748: sysctl_net_inet_ip_forwsrcrt, 0, &ip_forwsrcrt, 0,
1.189 atatat 1749: CTL_NET, PF_INET, IPPROTO_IP,
1750: IPCTL_FORWSRCRT, CTL_EOL);
1.197 atatat 1751: sysctl_createv(clog, 0, NULL, NULL,
1752: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1753: CTLTYPE_INT, "directed-broadcast",
1754: SYSCTL_DESCR("Enable forwarding of broadcast datagrams"),
1.189 atatat 1755: NULL, 0, &ip_directedbcast, 0,
1756: CTL_NET, PF_INET, IPPROTO_IP,
1757: IPCTL_DIRECTEDBCAST, CTL_EOL);
1.197 atatat 1758: sysctl_createv(clog, 0, NULL, NULL,
1759: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1760: CTLTYPE_INT, "allowsrcrt",
1761: SYSCTL_DESCR("Accept source-routed datagrams"),
1.189 atatat 1762: NULL, 0, &ip_allowsrcrt, 0,
1763: CTL_NET, PF_INET, IPPROTO_IP,
1764: IPCTL_ALLOWSRCRT, CTL_EOL);
1.197 atatat 1765: sysctl_createv(clog, 0, NULL, NULL,
1766: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1767: CTLTYPE_INT, "subnetsarelocal",
1768: SYSCTL_DESCR("Whether logical subnets are considered "
1769: "local"),
1.189 atatat 1770: NULL, 0, &subnetsarelocal, 0,
1771: CTL_NET, PF_INET, IPPROTO_IP,
1772: IPCTL_SUBNETSARELOCAL, CTL_EOL);
1.197 atatat 1773: sysctl_createv(clog, 0, NULL, NULL,
1774: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1775: CTLTYPE_INT, "mtudisc",
1776: SYSCTL_DESCR("Use RFC1191 Path MTU Discovery"),
1.189 atatat 1777: NULL, 0, &ip_mtudisc, 0,
1778: CTL_NET, PF_INET, IPPROTO_IP,
1779: IPCTL_MTUDISC, CTL_EOL);
1.197 atatat 1780: sysctl_createv(clog, 0, NULL, NULL,
1781: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1782: CTLTYPE_INT, "anonportmin",
1783: SYSCTL_DESCR("Lowest ephemeral port number to assign"),
1.189 atatat 1784: sysctl_net_inet_ip_ports, 0, &anonportmin, 0,
1785: CTL_NET, PF_INET, IPPROTO_IP,
1786: IPCTL_ANONPORTMIN, CTL_EOL);
1.197 atatat 1787: sysctl_createv(clog, 0, NULL, NULL,
1788: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1789: CTLTYPE_INT, "anonportmax",
1790: SYSCTL_DESCR("Highest ephemeral port number to assign"),
1.189 atatat 1791: sysctl_net_inet_ip_ports, 0, &anonportmax, 0,
1792: CTL_NET, PF_INET, IPPROTO_IP,
1793: IPCTL_ANONPORTMAX, CTL_EOL);
1.197 atatat 1794: sysctl_createv(clog, 0, NULL, NULL,
1795: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1796: CTLTYPE_INT, "mtudisctimeout",
1797: SYSCTL_DESCR("Lifetime of a Path MTU Discovered route"),
1.189 atatat 1798: sysctl_net_inet_ip_pmtudto, 0, &ip_mtudisc_timeout, 0,
1799: CTL_NET, PF_INET, IPPROTO_IP,
1800: IPCTL_MTUDISCTIMEOUT, CTL_EOL);
1801: #ifdef GATEWAY
1.197 atatat 1802: sysctl_createv(clog, 0, NULL, NULL,
1803: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1804: CTLTYPE_INT, "maxflows",
1805: SYSCTL_DESCR("Number of flows for fast forwarding"),
1.189 atatat 1806: sysctl_net_inet_ip_maxflows, 0, &ip_maxflows, 0,
1807: CTL_NET, PF_INET, IPPROTO_IP,
1808: IPCTL_MAXFLOWS, CTL_EOL);
1.248 liamjfoy 1809: sysctl_createv(clog, 0, NULL, NULL,
1810: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1811: CTLTYPE_INT, "hashsize",
1812: SYSCTL_DESCR("Size of hash table for fast forwarding (IPv4)"),
1813: sysctl_net_inet_ip_hashsize, 0, &ip_hashsize, 0,
1814: CTL_NET, PF_INET, IPPROTO_IP,
1815: CTL_CREATE, CTL_EOL);
1.189 atatat 1816: #endif /* GATEWAY */
1.197 atatat 1817: sysctl_createv(clog, 0, NULL, NULL,
1818: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1819: CTLTYPE_INT, "hostzerobroadcast",
1820: SYSCTL_DESCR("All zeroes address is broadcast address"),
1.189 atatat 1821: NULL, 0, &hostzeroisbroadcast, 0,
1822: CTL_NET, PF_INET, IPPROTO_IP,
1823: IPCTL_HOSTZEROBROADCAST, CTL_EOL);
1824: #if NGIF > 0
1.197 atatat 1825: sysctl_createv(clog, 0, NULL, NULL,
1826: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1827: CTLTYPE_INT, "gifttl",
1828: SYSCTL_DESCR("Default TTL for a gif tunnel datagram"),
1.189 atatat 1829: NULL, 0, &ip_gif_ttl, 0,
1830: CTL_NET, PF_INET, IPPROTO_IP,
1831: IPCTL_GIF_TTL, CTL_EOL);
1832: #endif /* NGIF */
1833: #ifndef IPNOPRIVPORTS
1.197 atatat 1834: sysctl_createv(clog, 0, NULL, NULL,
1835: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1836: CTLTYPE_INT, "lowportmin",
1837: SYSCTL_DESCR("Lowest privileged ephemeral port number "
1838: "to assign"),
1.189 atatat 1839: sysctl_net_inet_ip_ports, 0, &lowportmin, 0,
1840: CTL_NET, PF_INET, IPPROTO_IP,
1841: IPCTL_LOWPORTMIN, CTL_EOL);
1.197 atatat 1842: sysctl_createv(clog, 0, NULL, NULL,
1843: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1844: CTLTYPE_INT, "lowportmax",
1845: SYSCTL_DESCR("Highest privileged ephemeral port number "
1846: "to assign"),
1.189 atatat 1847: sysctl_net_inet_ip_ports, 0, &lowportmax, 0,
1848: CTL_NET, PF_INET, IPPROTO_IP,
1849: IPCTL_LOWPORTMAX, CTL_EOL);
1850: #endif /* IPNOPRIVPORTS */
1851: #if NGRE > 0
1.197 atatat 1852: sysctl_createv(clog, 0, NULL, NULL,
1853: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1854: CTLTYPE_INT, "grettl",
1855: SYSCTL_DESCR("Default TTL for a gre tunnel datagram"),
1.189 atatat 1856: NULL, 0, &ip_gre_ttl, 0,
1857: CTL_NET, PF_INET, IPPROTO_IP,
1858: IPCTL_GRE_TTL, CTL_EOL);
1859: #endif /* NGRE */
1.197 atatat 1860: sysctl_createv(clog, 0, NULL, NULL,
1861: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1862: CTLTYPE_INT, "checkinterface",
1863: SYSCTL_DESCR("Enable receive side of Strong ES model "
1864: "from RFC1122"),
1.189 atatat 1865: NULL, 0, &ip_checkinterface, 0,
1866: CTL_NET, PF_INET, IPPROTO_IP,
1867: IPCTL_CHECKINTERFACE, CTL_EOL);
1.197 atatat 1868: sysctl_createv(clog, 0, NULL, NULL,
1869: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1870: CTLTYPE_INT, "random_id",
1871: SYSCTL_DESCR("Assign random ip_id values"),
1.189 atatat 1872: NULL, 0, &ip_do_randomid, 0,
1873: CTL_NET, PF_INET, IPPROTO_IP,
1874: IPCTL_RANDOMID, CTL_EOL);
1.206 thorpej 1875: sysctl_createv(clog, 0, NULL, NULL,
1876: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1877: CTLTYPE_INT, "do_loopback_cksum",
1878: SYSCTL_DESCR("Perform IP checksum on loopback"),
1879: NULL, 0, &ip_do_loopback_cksum, 0,
1880: CTL_NET, PF_INET, IPPROTO_IP,
1881: IPCTL_LOOPBACKCKSUM, CTL_EOL);
1.219 elad 1882: sysctl_createv(clog, 0, NULL, NULL,
1883: CTLFLAG_PERMANENT,
1884: CTLTYPE_STRUCT, "stats",
1885: SYSCTL_DESCR("IP statistics"),
1.266 thorpej 1886: sysctl_net_inet_ip_stats, 0, NULL, 0,
1.219 elad 1887: CTL_NET, PF_INET, IPPROTO_IP, IPCTL_STATS,
1888: CTL_EOL);
1.1 cgd 1889: }
1.266 thorpej 1890:
1891: void
1892: ip_statinc(u_int stat)
1893: {
1894:
1895: KASSERT(stat < IP_NSTATS);
1896: IP_STATINC(stat);
1897: }
CVSweb <webmaster@jp.NetBSD.org>