Annotation of src/sys/netinet/ip_input.c, Revision 1.396
1.396 ! ozaki-r 1: /* $NetBSD: ip_input.c,v 1.395 2020/08/28 06:20:44 ozaki-r 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:
1.365 maxv 32: /*
1.76 thorpej 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>
1.396 ! ozaki-r 94: __KERNEL_RCSID(0, "$NetBSD: ip_input.c,v 1.395 2020/08/28 06:20:44 ozaki-r Exp $");
1.55 scottr 95:
1.324 pooka 96: #ifdef _KERNEL_OPT
1.184 jonathan 97: #include "opt_inet.h"
1.62 matt 98: #include "opt_gateway.h"
1.91 thorpej 99: #include "opt_ipsec.h"
1.55 scottr 100: #include "opt_mrouting.h"
1.167 martin 101: #include "opt_mbuftrace.h"
1.135 thorpej 102: #include "opt_inet_csum.h"
1.344 ozaki-r 103: #include "opt_net_mpsafe.h"
1.324 pooka 104: #endif
1.1 cgd 105:
1.325 roy 106: #include "arp.h"
107:
1.5 mycroft 108: #include <sys/param.h>
109: #include <sys/systm.h>
1.318 rmind 110: #include <sys/cpu.h>
1.5 mycroft 111: #include <sys/mbuf.h>
112: #include <sys/domain.h>
113: #include <sys/protosw.h>
114: #include <sys/socket.h>
1.44 thorpej 115: #include <sys/socketvar.h>
1.5 mycroft 116: #include <sys/errno.h>
117: #include <sys/time.h>
118: #include <sys/kernel.h>
1.72 thorpej 119: #include <sys/pool.h>
1.28 christos 120: #include <sys/sysctl.h>
1.230 elad 121: #include <sys/kauth.h>
1.1 cgd 122:
1.5 mycroft 123: #include <net/if.h>
1.44 thorpej 124: #include <net/if_dl.h>
1.5 mycroft 125: #include <net/route.h>
1.318 rmind 126: #include <net/pktqueue.h>
1.45 mrg 127: #include <net/pfil.h>
1.1 cgd 128:
1.5 mycroft 129: #include <netinet/in.h>
130: #include <netinet/in_systm.h>
131: #include <netinet/ip.h>
132: #include <netinet/in_pcb.h>
1.215 yamt 133: #include <netinet/in_proto.h>
1.5 mycroft 134: #include <netinet/in_var.h>
135: #include <netinet/ip_var.h>
1.266 thorpej 136: #include <netinet/ip_private.h>
1.5 mycroft 137: #include <netinet/ip_icmp.h>
1.89 itojun 138: /* just for gif_ttl */
139: #include <netinet/in_gif.h>
140: #include "gif.h"
1.144 martin 141: #include <net/if_gre.h>
142: #include "gre.h"
1.111 jdolecek 143:
144: #ifdef MROUTING
145: #include <netinet/ip_mroute.h>
146: #endif
1.302 christos 147: #include <netinet/portalgo.h>
1.89 itojun 148:
1.304 christos 149: #ifdef IPSEC
1.173 jonathan 150: #include <netipsec/ipsec.h>
1.305 rmind 151: #endif
1.44 thorpej 152:
1.1 cgd 153: #ifndef IPFORWARDING
154: #ifdef GATEWAY
155: #define IPFORWARDING 1 /* forward IP packets not for us */
1.370 maxv 156: #else
1.1 cgd 157: #define IPFORWARDING 0 /* don't forward IP packets not for us */
1.370 maxv 158: #endif
159: #endif
160:
1.61 kml 161: #define IPMTUDISCTIMEOUT (10 * 60) /* as per RFC 1191 */
1.53 kml 162:
1.370 maxv 163: int ipforwarding = IPFORWARDING;
1.372 maxv 164: int ipsendredirects = 1;
1.370 maxv 165: int ip_defttl = IPDEFTTL;
166: int ip_forwsrcrt = 0;
1.371 maxv 167: int ip_directedbcast = 0;
1.370 maxv 168: int ip_allowsrcrt = 0;
169: int ip_mtudisc = 1;
170: int ip_mtudisc_timeout = IPMTUDISCTIMEOUT;
171: int ip_do_randomid = 0;
1.184 jonathan 172:
1.165 christos 173: /*
174: * XXX - Setting ip_checkinterface mostly implements the receive side of
175: * the Strong ES model described in RFC 1122, but since the routing table
176: * and transmit implementation do not implement the Strong ES model,
177: * setting this to 1 results in an odd hybrid.
178: *
1.386 maxv 179: * XXX - ip_checkinterface currently must be disabled if you use NAT
1.165 christos 180: * to translate the destination address to another local interface.
181: *
182: * XXX - ip_checkinterface must be disabled if you add IP aliases
183: * to the loopback interface instead of the interface where the
184: * packets for those addresses are received.
185: */
1.318 rmind 186: static int ip_checkinterface __read_mostly = 0;
1.165 christos 187:
1.60 kml 188: struct rttimer_queue *ip_mtudisc_timeout_q = NULL;
189:
1.318 rmind 190: pktqueue_t * ip_pktq __read_mostly;
191: pfil_head_t * inet_pfil_hook __read_mostly;
192: ipid_state_t * ip_ids __read_mostly;
193: percpu_t * ipstat_percpu __read_mostly;
194:
1.340 ozaki-r 195: static percpu_t *ipforward_rt_percpu __cacheline_aligned;
1.286 tls 196:
1.183 jonathan 197: uint16_t ip_id;
1.75 thorpej 198:
1.135 thorpej 199: #ifdef INET_CSUM_COUNTERS
200: #include <sys/device.h>
201:
202: struct evcnt ip_hwcsum_bad = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
203: NULL, "inet", "hwcsum bad");
204: struct evcnt ip_hwcsum_ok = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
205: NULL, "inet", "hwcsum ok");
206: struct evcnt ip_swcsum = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
207: NULL, "inet", "swcsum");
208:
209: #define INET_CSUM_COUNTER_INCR(ev) (ev)->ev_count++
210:
1.201 matt 211: EVCNT_ATTACH_STATIC(ip_hwcsum_bad);
212: EVCNT_ATTACH_STATIC(ip_hwcsum_ok);
213: EVCNT_ATTACH_STATIC(ip_swcsum);
214:
1.135 thorpej 215: #else
216:
217: #define INET_CSUM_COUNTER_INCR(ev) /* nothing */
218:
219: #endif /* INET_CSUM_COUNTERS */
220:
1.1 cgd 221: /*
1.354 ozaki-r 222: * Used to save the IP options in case a protocol wants to respond
1.1 cgd 223: * to an incoming packet over the same route if the packet got here
224: * using IP source routing. This allows connection establishment and
225: * maintenance when the remote end is on a network that is not known
226: * to us.
227: */
1.354 ozaki-r 228: struct ip_srcrt {
229: int isr_nhops; /* number of hops */
230: struct in_addr isr_dst; /* final destination */
231: char isr_nop; /* one NOP to align */
232: char isr_hdr[IPOPT_OFFSET + 1]; /* OPTVAL, OLEN & OFFSET */
233: struct in_addr isr_routes[MAX_IPOPTLEN/sizeof(struct in_addr)];
234: };
1.1 cgd 235:
1.295 dyoung 236: static int ip_drainwanted;
237:
1.354 ozaki-r 238: static void save_rte(struct mbuf *, u_char *, struct in_addr);
1.35 mycroft 239:
1.164 matt 240: #ifdef MBUFTRACE
1.234 dogcow 241: struct mowner ip_rx_mowner = MOWNER_INIT("internet", "rx");
242: struct mowner ip_tx_mowner = MOWNER_INIT("internet", "tx");
1.164 matt 243: #endif
244:
1.318 rmind 245: static void ipintr(void *);
1.396 ! ozaki-r 246: static void ip_input(struct mbuf *, struct ifnet *);
1.329 ozaki-r 247: static void ip_forward(struct mbuf *, int, struct ifnet *);
1.311 rmind 248: static bool ip_dooptions(struct mbuf *);
1.346 ozaki-r 249: static struct in_ifaddr *ip_rtaddr(struct in_addr, struct psref *);
1.311 rmind 250: static void sysctl_net_inet_ip_setup(struct sysctllog **);
1.284 pooka 251:
1.333 ozaki-r 252: static struct in_ifaddr *ip_match_our_address(struct ifnet *, struct ip *,
253: int *);
254: static struct in_ifaddr *ip_match_our_address_broadcast(struct ifnet *,
255: struct ip *);
256:
1.344 ozaki-r 257: #ifdef NET_MPSAFE
258: #define SOFTNET_LOCK() mutex_enter(softnet_lock)
259: #define SOFTNET_UNLOCK() mutex_exit(softnet_lock)
260: #else
1.318 rmind 261: #define SOFTNET_LOCK() KASSERT(mutex_owned(softnet_lock))
262: #define SOFTNET_UNLOCK() KASSERT(mutex_owned(softnet_lock))
1.344 ozaki-r 263: #endif
1.318 rmind 264:
1.1 cgd 265: /*
266: * IP initialization: fill in IP protocol switch table.
267: * All protocols not implemented in kernel go to raw IP protocol handler.
268: */
1.8 mycroft 269: void
1.211 perry 270: ip_init(void)
1.1 cgd 271: {
1.199 matt 272: const struct protosw *pr;
1.1 cgd 273:
1.311 rmind 274: in_init();
1.284 pooka 275: sysctl_net_inet_ip_setup(NULL);
276:
1.311 rmind 277: pr = pffindproto(PF_INET, IPPROTO_RAW, SOCK_RAW);
278: KASSERT(pr != NULL);
1.275 pooka 279:
1.319 ozaki-r 280: ip_pktq = pktq_create(IFQ_MAXLEN, ipintr, NULL);
1.318 rmind 281: KASSERT(ip_pktq != NULL);
282:
1.328 riastrad 283: for (u_int i = 0; i < IPPROTO_MAX; i++) {
284: ip_protox[i] = pr - inetsw;
1.311 rmind 285: }
1.328 riastrad 286: for (pr = inetdomain.dom_protosw;
287: pr < inetdomain.dom_protoswNPROTOSW; pr++)
1.1 cgd 288: if (pr->pr_domain->dom_family == PF_INET &&
289: pr->pr_protocol && pr->pr_protocol != IPPROTO_RAW)
1.328 riastrad 290: ip_protox[pr->pr_protocol] = pr - inetsw;
1.192 jonathan 291:
1.288 rmind 292: ip_reass_init();
1.190 jonathan 293:
1.291 rmind 294: ip_ids = ip_id_init();
1.323 ozaki-r 295: ip_id = time_uptime & 0xfffff;
1.194 jonathan 296:
1.73 thorpej 297: #ifdef GATEWAY
1.311 rmind 298: ipflow_init();
1.73 thorpej 299: #endif
1.121 thorpej 300:
301: /* Register our Packet Filter hook. */
1.308 rmind 302: inet_pfil_hook = pfil_head_create(PFIL_TYPE_AF, (void *)AF_INET);
303: KASSERT(inet_pfil_hook != NULL);
1.135 thorpej 304:
1.164 matt 305: #ifdef MBUFTRACE
306: MOWNER_ATTACH(&ip_tx_mowner);
307: MOWNER_ATTACH(&ip_rx_mowner);
1.373 maxv 308: #endif
1.266 thorpej 309:
310: ipstat_percpu = percpu_alloc(sizeof(uint64_t) * IP_NSTATS);
1.391 ozaki-r 311: ipforward_rt_percpu = rtcache_percpu_alloc();
1.352 ozaki-r 312: ip_mtudisc_timeout_q = rt_timer_queue_create(ip_mtudisc_timeout);
1.1 cgd 313: }
314:
1.333 ozaki-r 315: static struct in_ifaddr *
316: ip_match_our_address(struct ifnet *ifp, struct ip *ip, int *downmatch)
317: {
318: struct in_ifaddr *ia = NULL;
319: int checkif;
320:
321: /*
322: * Enable a consistency check between the destination address
323: * and the arrival interface for a unicast packet (the RFC 1122
324: * strong ES model) if IP forwarding is disabled and the packet
325: * is not locally generated.
326: *
327: * XXX - We need to add a per ifaddr flag for this so that
328: * we get finer grain control.
329: */
330: checkif = ip_checkinterface && (ipforwarding == 0) &&
331: (ifp->if_flags & IFF_LOOPBACK) == 0;
332:
1.334 ozaki-r 333: IN_ADDRHASH_READER_FOREACH(ia, ip->ip_dst.s_addr) {
1.333 ozaki-r 334: if (in_hosteq(ia->ia_addr.sin_addr, ip->ip_dst)) {
335: if (ia->ia4_flags & IN_IFF_NOTREADY)
336: continue;
337: if (checkif && ia->ia_ifp != ifp)
338: continue;
1.363 roy 339: if ((ia->ia_ifp->if_flags & IFF_UP) == 0) {
1.338 ozaki-r 340: (*downmatch)++;
1.363 roy 341: continue;
342: }
343: if (ia->ia4_flags & IN_IFF_DETACHED &&
344: (ifp->if_flags & IFF_LOOPBACK) == 0)
345: continue;
346: break;
1.333 ozaki-r 347: }
348: }
349:
350: return ia;
351: }
352:
353: static struct in_ifaddr *
354: ip_match_our_address_broadcast(struct ifnet *ifp, struct ip *ip)
355: {
356: struct in_ifaddr *ia = NULL;
357: struct ifaddr *ifa;
358:
1.336 ozaki-r 359: IFADDR_READER_FOREACH(ifa, ifp) {
1.333 ozaki-r 360: if (ifa->ifa_addr->sa_family != AF_INET)
361: continue;
362: ia = ifatoia(ifa);
1.363 roy 363: if (ia->ia4_flags & IN_IFF_NOTREADY)
364: continue;
365: if (ia->ia4_flags & IN_IFF_DETACHED &&
366: (ifp->if_flags & IFF_LOOPBACK) == 0)
1.333 ozaki-r 367: continue;
368: if (in_hosteq(ip->ip_dst, ia->ia_broadaddr.sin_addr) ||
369: in_hosteq(ip->ip_dst, ia->ia_netbroadcast) ||
370: /*
371: * Look for all-0's host part (old broadcast addr),
372: * either for subnet or net.
373: */
374: ip->ip_dst.s_addr == ia->ia_subnet ||
375: ip->ip_dst.s_addr == ia->ia_net)
376: goto matched;
377: /*
378: * An interface with IP address zero accepts
379: * all packets that arrive on that interface.
380: */
381: if (in_nullhost(ia->ia_addr.sin_addr))
382: goto matched;
383: }
384: ia = NULL;
385:
386: matched:
387: return ia;
388: }
389:
1.1 cgd 390: /*
1.318 rmind 391: * IP software interrupt routine.
1.89 itojun 392: */
1.318 rmind 393: static void
394: ipintr(void *arg __unused)
1.89 itojun 395: {
396: struct mbuf *m;
1.286 tls 397:
1.318 rmind 398: KASSERT(cpu_softintr_p());
1.89 itojun 399:
1.376 ozaki-r 400: SOFTNET_KERNEL_LOCK_UNLESS_NET_MPSAFE();
1.318 rmind 401: while ((m = pktq_dequeue(ip_pktq)) != NULL) {
1.396 ! ozaki-r 402: struct ifnet *ifp;
! 403: struct psref psref;
! 404:
! 405: ifp = m_get_rcvif_psref(m, &psref);
! 406: if (__predict_false(ifp == NULL)) {
! 407: m_freem(m);
! 408: continue;
! 409: }
! 410:
! 411: ip_input(m, ifp);
! 412:
! 413: m_put_rcvif_psref(ifp, &psref);
1.89 itojun 414: }
1.376 ozaki-r 415: SOFTNET_KERNEL_UNLOCK_UNLESS_NET_MPSAFE();
1.89 itojun 416: }
417:
418: /*
1.318 rmind 419: * IP input routine. Checksum and byte swap header. If fragmented
1.1 cgd 420: * try to reassemble. Process options. Pass to next level.
421: */
1.312 rmind 422: static void
1.396 ! ozaki-r 423: ip_input(struct mbuf *m, struct ifnet *ifp)
1.1 cgd 424: {
1.109 augustss 425: struct ip *ip = NULL;
1.339 ozaki-r 426: struct in_ifaddr *ia = NULL;
1.288 rmind 427: int hlen = 0, len;
1.100 itojun 428: int downmatch;
1.169 itojun 429: int srcrt = 0;
1.339 ozaki-r 430: int s;
1.1 cgd 431:
1.318 rmind 432: KASSERTMSG(cpu_softintr_p(), "ip_input: not in the software "
433: "interrupt handler; synchronization assumptions violated");
434:
1.164 matt 435: MCLAIM(m, &ip_rx_mowner);
1.289 rmind 436: KASSERT((m->m_flags & M_PKTHDR) != 0);
1.329 ozaki-r 437:
1.1 cgd 438: /*
439: * If no IP addresses have been set yet but the interfaces
440: * are receiving, can't do anything with incoming packets yet.
1.318 rmind 441: * Note: we pre-check without locks held.
1.1 cgd 442: */
1.335 ozaki-r 443: if (IN_ADDRLIST_READER_EMPTY())
1.332 ozaki-r 444: goto out;
1.266 thorpej 445: IP_STATINC(IP_STAT_TOTAL);
1.318 rmind 446:
1.154 thorpej 447: /*
448: * If the IP header is not aligned, slurp it up into a new
449: * mbuf with space for link headers, in the event we forward
450: * it. Otherwise, if it is aligned, make sure the entire
451: * base IP header is in the first mbuf of the chain.
452: */
1.244 christos 453: if (IP_HDR_ALIGNED_P(mtod(m, void *)) == 0) {
1.154 thorpej 454: if ((m = m_copyup(m, sizeof(struct ip),
1.365 maxv 455: (max_linkhdr + 3) & ~3)) == NULL) {
1.154 thorpej 456: /* XXXJRT new stat, please */
1.266 thorpej 457: IP_STATINC(IP_STAT_TOOSMALL);
1.329 ozaki-r 458: goto out;
1.154 thorpej 459: }
1.365 maxv 460: } else if (__predict_false(m->m_len < sizeof(struct ip))) {
461: if ((m = m_pullup(m, sizeof(struct ip))) == NULL) {
1.266 thorpej 462: IP_STATINC(IP_STAT_TOOSMALL);
1.329 ozaki-r 463: goto out;
1.154 thorpej 464: }
1.1 cgd 465: }
466: ip = mtod(m, struct ip *);
1.13 mycroft 467: if (ip->ip_v != IPVERSION) {
1.266 thorpej 468: IP_STATINC(IP_STAT_BADVERS);
1.332 ozaki-r 469: goto out;
1.13 mycroft 470: }
1.1 cgd 471: hlen = ip->ip_hl << 2;
472: if (hlen < sizeof(struct ip)) { /* minimum header length */
1.266 thorpej 473: IP_STATINC(IP_STAT_BADHLEN);
1.332 ozaki-r 474: goto out;
1.1 cgd 475: }
476: if (hlen > m->m_len) {
1.298 liamjfoy 477: if ((m = m_pullup(m, hlen)) == NULL) {
1.266 thorpej 478: IP_STATINC(IP_STAT_BADHLEN);
1.329 ozaki-r 479: goto out;
1.1 cgd 480: }
481: ip = mtod(m, struct ip *);
482: }
1.98 thorpej 483:
1.85 hwr 484: /*
1.99 thorpej 485: * RFC1122: packets with a multicast source address are
1.98 thorpej 486: * not allowed.
1.85 hwr 487: */
488: if (IN_MULTICAST(ip->ip_src.s_addr)) {
1.266 thorpej 489: IP_STATINC(IP_STAT_BADADDR);
1.332 ozaki-r 490: goto out;
1.129 itojun 491: }
492:
493: /* 127/8 must not appear on wire - RFC1122 */
494: if ((ntohl(ip->ip_dst.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET ||
495: (ntohl(ip->ip_src.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET) {
1.314 rmind 496: if ((ifp->if_flags & IFF_LOOPBACK) == 0) {
1.266 thorpej 497: IP_STATINC(IP_STAT_BADADDR);
1.332 ozaki-r 498: goto out;
1.130 itojun 499: }
1.85 hwr 500: }
501:
1.135 thorpej 502: switch (m->m_pkthdr.csum_flags &
1.365 maxv 503: ((ifp->if_csum_flags_rx & M_CSUM_IPv4) | M_CSUM_IPv4_BAD)) {
1.135 thorpej 504: case M_CSUM_IPv4|M_CSUM_IPv4_BAD:
505: INET_CSUM_COUNTER_INCR(&ip_hwcsum_bad);
1.332 ozaki-r 506: IP_STATINC(IP_STAT_BADSUM);
507: goto out;
1.135 thorpej 508:
509: case M_CSUM_IPv4:
510: /* Checksum was okay. */
511: INET_CSUM_COUNTER_INCR(&ip_hwcsum_ok);
512: break;
513:
514: default:
1.206 thorpej 515: /*
516: * Must compute it ourselves. Maybe skip checksum on
517: * loopback interfaces.
518: */
1.314 rmind 519: if (__predict_true(!(ifp->if_flags & IFF_LOOPBACK) ||
520: ip_do_loopback_cksum)) {
1.206 thorpej 521: INET_CSUM_COUNTER_INCR(&ip_swcsum);
1.332 ozaki-r 522: if (in_cksum(m, hlen) != 0) {
523: IP_STATINC(IP_STAT_BADSUM);
524: goto out;
525: }
1.206 thorpej 526: }
1.135 thorpej 527: break;
1.1 cgd 528: }
529:
1.121 thorpej 530: /* Retrieve the packet length. */
531: len = ntohs(ip->ip_len);
1.81 proff 532:
533: /*
534: * Check for additional length bogosity
535: */
1.84 proff 536: if (len < hlen) {
1.266 thorpej 537: IP_STATINC(IP_STAT_BADLEN);
1.332 ozaki-r 538: goto out;
1.81 proff 539: }
1.1 cgd 540:
541: /*
1.365 maxv 542: * Check that the amount of data in the buffers is at least as much
543: * as the IP header would have us expect. Trim mbufs if longer than
544: * we expect. Drop packet if shorter than we expect.
1.1 cgd 545: */
1.35 mycroft 546: if (m->m_pkthdr.len < len) {
1.266 thorpej 547: IP_STATINC(IP_STAT_TOOSHORT);
1.332 ozaki-r 548: goto out;
1.1 cgd 549: }
1.35 mycroft 550: if (m->m_pkthdr.len > len) {
1.1 cgd 551: if (m->m_len == m->m_pkthdr.len) {
1.35 mycroft 552: m->m_len = len;
553: m->m_pkthdr.len = len;
1.1 cgd 554: } else
1.35 mycroft 555: m_adj(m, len - m->m_pkthdr.len);
1.1 cgd 556: }
557:
1.64 thorpej 558: /*
559: * Assume that we can create a fast-forward IP flow entry
560: * based on this packet.
561: */
562: m->m_flags |= M_CANFASTFWD;
563:
1.33 mrg 564: /*
1.64 thorpej 565: * Run through list of hooks for input packets. If there are any
566: * filters which require that additional packets in the flow are
567: * not fast-forwarded, they must clear the M_CANFASTFWD flag.
568: * Note that filters must _never_ set this flag, as another filter
569: * in the list may have previously cleared it.
1.378 maxv 570: *
571: * Don't call hooks if the packet has already been processed by
572: * IPsec (encapsulated, tunnel mode).
1.33 mrg 573: */
1.304 christos 574: #if defined(IPSEC)
1.388 knakahar 575: if (!ipsec_used || !ipsec_skip_pfil(m))
1.127 itojun 576: #else
577: if (1)
578: #endif
579: {
1.314 rmind 580: struct in_addr odst = ip->ip_dst;
1.318 rmind 581: bool freed;
1.169 itojun 582:
1.318 rmind 583: freed = pfil_run_hooks(inet_pfil_hook, &m, ifp, PFIL_IN) != 0;
584: if (freed || m == NULL) {
1.343 ozaki-r 585: m = NULL;
1.389 ozaki-r 586: IP_STATINC(IP_STAT_PFILDROP_IN);
1.329 ozaki-r 587: goto out;
1.318 rmind 588: }
1.390 bouyer 589: if (__predict_false(m->m_len < sizeof(struct ip))) {
590: if ((m = m_pullup(m, sizeof(struct ip))) == NULL) {
591: IP_STATINC(IP_STAT_TOOSMALL);
592: goto out;
593: }
594: }
1.127 itojun 595: ip = mtod(m, struct ip *);
1.142 darrenr 596: hlen = ip->ip_hl << 2;
1.390 bouyer 597: if (hlen < sizeof(struct ip)) { /* minimum header length */
598: IP_STATINC(IP_STAT_BADHLEN);
599: goto out;
600: }
601: if (hlen > m->m_len) {
602: if ((m = m_pullup(m, hlen)) == NULL) {
603: IP_STATINC(IP_STAT_BADHLEN);
604: goto out;
605: }
606: ip = mtod(m, struct ip *);
607: }
1.318 rmind 608:
1.205 darrenr 609: /*
610: * XXX The setting of "srcrt" here is to prevent ip_forward()
611: * from generating ICMP redirects for packets that have
612: * been redirected by a hook back out on to the same LAN that
613: * they came from and is not an indication that the packet
1.373 maxv 614: * is being influenced by source routing options. This
1.205 darrenr 615: * allows things like
616: * "rdr tlp0 0/0 port 80 -> 1.1.1.200 3128 tcp"
617: * where tlp0 is both on the 1.1.1.0/24 network and is the
618: * default route for hosts on 1.1.1.0/24. Of course this
619: * also requires a "map tlp0 ..." to complete the story.
620: * One might argue whether or not this kind of network config.
1.212 perry 621: * should be supported in this manner...
1.205 darrenr 622: */
1.169 itojun 623: srcrt = (odst.s_addr != ip->ip_dst.s_addr);
1.127 itojun 624: }
1.123 thorpej 625:
626: #ifdef ALTQ
627: /* XXX Temporary until ALTQ is changed to use a pfil hook */
1.318 rmind 628: if (altq_input) {
629: SOFTNET_LOCK();
630: if ((*altq_input)(m, AF_INET) == 0) {
631: /* Packet dropped by traffic conditioner. */
632: SOFTNET_UNLOCK();
1.343 ozaki-r 633: m = NULL;
1.329 ozaki-r 634: goto out;
1.318 rmind 635: }
636: SOFTNET_UNLOCK();
1.123 thorpej 637: }
638: #endif
1.121 thorpej 639:
640: /*
1.1 cgd 641: * Process options and, if not destined for us,
642: * ship it on. ip_dooptions returns 1 when an
643: * error was detected (causing an icmp message
644: * to be sent and the original packet to be freed).
645: */
1.365 maxv 646: if (hlen > sizeof(struct ip) && ip_dooptions(m)) {
1.343 ozaki-r 647: m = NULL;
1.329 ozaki-r 648: goto out;
1.343 ozaki-r 649: }
1.1 cgd 650:
651: /*
652: * Check our list of addresses, to see if the packet is for us.
1.100 itojun 653: *
654: * Traditional 4.4BSD did not consult IFF_UP at all.
655: * The behavior here is to treat addresses on !IFF_UP interface
1.321 roy 656: * or IN_IFF_NOTREADY addresses as not mine.
1.1 cgd 657: */
1.100 itojun 658: downmatch = 0;
1.339 ozaki-r 659: s = pserialize_read_enter();
1.333 ozaki-r 660: ia = ip_match_our_address(ifp, ip, &downmatch);
1.339 ozaki-r 661: if (ia != NULL) {
662: pserialize_read_exit(s);
1.86 thorpej 663: goto ours;
1.339 ozaki-r 664: }
1.333 ozaki-r 665:
1.315 christos 666: if (ifp->if_flags & IFF_BROADCAST) {
1.333 ozaki-r 667: ia = ip_match_our_address_broadcast(ifp, ip);
1.339 ozaki-r 668: if (ia != NULL) {
669: pserialize_read_exit(s);
1.333 ozaki-r 670: goto ours;
1.339 ozaki-r 671: }
1.1 cgd 672: }
1.339 ozaki-r 673: pserialize_read_exit(s);
1.333 ozaki-r 674:
1.18 mycroft 675: if (IN_MULTICAST(ip->ip_dst.s_addr)) {
1.4 hpeyerl 676: #ifdef MROUTING
677: extern struct socket *ip_mrouter;
1.10 brezak 678:
1.4 hpeyerl 679: if (ip_mrouter) {
680: /*
681: * If we are acting as a multicast router, all
682: * incoming multicast packets are passed to the
683: * kernel-level multicast forwarding function.
684: * The packet is returned (relatively) intact; if
685: * ip_mforward() returns a non-zero value, the packet
686: * must be discarded, else it may be accepted below.
687: *
688: * (The IP ident field is put in the same byte order
689: * as expected when ip_mforward() is called from
690: * ip_output().)
691: */
1.318 rmind 692: SOFTNET_LOCK();
1.314 rmind 693: if (ip_mforward(m, ifp) != 0) {
1.318 rmind 694: SOFTNET_UNLOCK();
1.266 thorpej 695: IP_STATINC(IP_STAT_CANTFORWARD);
1.332 ozaki-r 696: goto out;
1.4 hpeyerl 697: }
1.318 rmind 698: SOFTNET_UNLOCK();
1.4 hpeyerl 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: */
1.318 rmind 705: if (ip->ip_p == IPPROTO_IGMP) {
1.4 hpeyerl 706: goto ours;
1.318 rmind 707: }
1.266 thorpej 708: IP_STATINC(IP_STAT_CANTFORWARD);
1.4 hpeyerl 709: }
710: #endif
711: /*
712: * See if we belong to the destination multicast group on the
713: * arrival interface.
714: */
1.316 rmind 715: if (!in_multi_group(ip->ip_dst, ifp, 0)) {
1.266 thorpej 716: IP_STATINC(IP_STAT_CANTFORWARD);
1.332 ozaki-r 717: goto out;
1.4 hpeyerl 718: }
719: goto ours;
720: }
1.19 mycroft 721: if (ip->ip_dst.s_addr == INADDR_BROADCAST ||
1.35 mycroft 722: in_nullhost(ip->ip_dst))
1.1 cgd 723: goto ours;
724:
725: /*
726: * Not for us; forward if possible and desirable.
727: */
728: if (ipforwarding == 0) {
1.266 thorpej 729: IP_STATINC(IP_STAT_CANTFORWARD);
1.1 cgd 730: m_freem(m);
1.100 itojun 731: } else {
732: /*
733: * If ip_dst matched any of my address on !IFF_UP interface,
734: * and there's no IFF_UP interface that matches ip_dst,
735: * send icmp unreach. Forwarding it will result in in-kernel
736: * forwarding loop till TTL goes to 0.
737: */
738: if (downmatch) {
739: icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_HOST, 0, 0);
1.266 thorpej 740: IP_STATINC(IP_STAT_CANTFORWARD);
1.100 itojun 741: return;
742: }
1.304 christos 743: #ifdef IPSEC
1.359 ozaki-r 744: /* Check the security policy (SP) for the packet */
1.318 rmind 745: if (ipsec_used) {
1.395 ozaki-r 746: if (ipsec_ip_input_checkpolicy(m, true) != 0) {
1.394 ozaki-r 747: IP_STATINC(IP_STAT_IPSECDROP_IN);
1.332 ozaki-r 748: goto out;
1.318 rmind 749: }
1.173 jonathan 750: }
1.305 rmind 751: #endif
1.329 ozaki-r 752: ip_forward(m, srcrt, ifp);
1.100 itojun 753: }
1.89 itojun 754: return;
1.1 cgd 755:
756: ours:
757: /*
758: * If offset or IP_MF are set, must reassemble.
759: */
1.155 itojun 760: if (ip->ip_off & ~htons(IP_DF|IP_RF)) {
1.289 rmind 761: /*
762: * Pass to IP reassembly mechanism.
763: */
1.385 maxv 764: if (ip_reass_packet(&m) != 0) {
1.289 rmind 765: /* Failed; invalid fragment(s) or packet. */
1.332 ozaki-r 766: goto out;
1.16 cgd 767: }
1.290 rmind 768: if (m == NULL) {
1.289 rmind 769: /* More fragments should come; silently return. */
1.332 ozaki-r 770: goto out;
1.289 rmind 771: }
1.290 rmind 772: /*
773: * Reassembly is done, we have the final packet.
1.365 maxv 774: * Update cached data in local variable(s).
1.290 rmind 775: */
1.289 rmind 776: ip = mtod(m, struct ip *);
777: hlen = ip->ip_hl << 2;
1.79 mycroft 778: }
1.128 itojun 779:
1.380 maxv 780: M_VERIFY_PACKET(m);
781:
1.304 christos 782: #ifdef IPSEC
1.173 jonathan 783: /*
1.305 rmind 784: * Enforce IPsec policy checking if we are seeing last header.
785: * Note that we do not visit this with protocols with PCB layer
786: * code - like UDP/TCP/raw IP.
1.173 jonathan 787: */
1.317 christos 788: if (ipsec_used &&
1.328 riastrad 789: (inetsw[ip_protox[ip->ip_p]].pr_flags & PR_LASTHDR) != 0) {
1.395 ozaki-r 790: if (ipsec_ip_input_checkpolicy(m, false) != 0) {
1.394 ozaki-r 791: IP_STATINC(IP_STAT_IPSECDROP_IN);
1.332 ozaki-r 792: goto out;
1.173 jonathan 793: }
794: }
1.305 rmind 795: #endif
1.1 cgd 796:
797: /*
798: * Switch out to protocol's input routine.
799: */
1.82 aidan 800: #if IFA_STATS
1.374 maxv 801: if (ia) {
1.339 ozaki-r 802: struct in_ifaddr *_ia;
803: /*
804: * Keep a reference from ip_match_our_address with psref
805: * is expensive, so explore ia here again.
806: */
807: s = pserialize_read_enter();
1.342 ozaki-r 808: _ia = in_get_ia(ip->ip_dst);
1.339 ozaki-r 809: _ia->ia_ifa.ifa_data.ifad_inbytes += ntohs(ip->ip_len);
810: pserialize_read_exit(s);
811: }
1.82 aidan 812: #endif
1.266 thorpej 813: IP_STATINC(IP_STAT_DELIVERED);
1.89 itojun 814:
1.318 rmind 815: const int off = hlen, nh = ip->ip_p;
816:
1.328 riastrad 817: (*inetsw[ip_protox[nh]].pr_input)(m, off, nh);
1.89 itojun 818: return;
1.135 thorpej 819:
1.329 ozaki-r 820: out:
1.332 ozaki-r 821: if (m != NULL)
822: m_freem(m);
1.1 cgd 823: }
824:
825: /*
1.288 rmind 826: * IP timer processing.
1.1 cgd 827: */
1.8 mycroft 828: void
1.211 perry 829: ip_slowtimo(void)
1.1 cgd 830: {
1.268 ad 831:
1.362 ozaki-r 832: SOFTNET_KERNEL_LOCK_UNLESS_NET_MPSAFE();
1.1 cgd 833:
1.288 rmind 834: ip_reass_slowtimo();
1.268 ad 835:
1.362 ozaki-r 836: SOFTNET_KERNEL_UNLOCK_UNLESS_NET_MPSAFE();
1.1 cgd 837: }
838:
839: /*
1.288 rmind 840: * IP drain processing.
1.1 cgd 841: */
1.8 mycroft 842: void
1.211 perry 843: ip_drain(void)
1.1 cgd 844: {
845:
1.268 ad 846: KERNEL_LOCK(1, NULL);
1.288 rmind 847: ip_reass_drain();
1.268 ad 848: KERNEL_UNLOCK_ONE(NULL);
1.1 cgd 849: }
850:
851: /*
1.311 rmind 852: * ip_dooptions: perform option processing on a datagram, possibly discarding
853: * it if bad options are encountered, or forwarding it if source-routed.
854: *
855: * => Returns true if packet has been forwarded/freed.
856: * => Returns false if the packet should be processed further.
1.1 cgd 857: */
1.311 rmind 858: static bool
1.211 perry 859: ip_dooptions(struct mbuf *m)
1.1 cgd 860: {
1.109 augustss 861: struct ip *ip = mtod(m, struct ip *);
862: u_char *cp, *cp0;
863: struct ip_timestamp *ipt;
864: struct in_ifaddr *ia;
1.1 cgd 865: int opt, optlen, cnt, off, code, type = ICMP_PARAMPROB, forward = 0;
1.369 maxv 866: int srr_present, rr_present, ts_present;
1.104 thorpej 867: struct in_addr dst;
1.1 cgd 868: n_time ntime;
1.349 ozaki-r 869: struct ifaddr *ifa = NULL;
1.339 ozaki-r 870: int s;
1.1 cgd 871:
1.369 maxv 872: srr_present = 0;
1.368 maxv 873: rr_present = 0;
874: ts_present = 0;
875:
1.13 mycroft 876: dst = ip->ip_dst;
1.1 cgd 877: cp = (u_char *)(ip + 1);
1.365 maxv 878: cnt = (ip->ip_hl << 2) - sizeof(struct ip);
1.1 cgd 879: for (; cnt > 0; cnt -= optlen, cp += optlen) {
880: opt = cp[IPOPT_OPTVAL];
881: if (opt == IPOPT_EOL)
882: break;
883: if (opt == IPOPT_NOP)
884: optlen = 1;
885: else {
1.113 itojun 886: if (cnt < IPOPT_OLEN + sizeof(*cp)) {
887: code = &cp[IPOPT_OLEN] - (u_char *)ip;
888: goto bad;
889: }
1.1 cgd 890: optlen = cp[IPOPT_OLEN];
1.114 itojun 891: if (optlen < IPOPT_OLEN + sizeof(*cp) || optlen > cnt) {
1.1 cgd 892: code = &cp[IPOPT_OLEN] - (u_char *)ip;
893: goto bad;
894: }
895: }
896: switch (opt) {
897:
898: default:
899: break;
900:
901: /*
902: * Source routing with record.
903: * Find interface with current destination address.
904: * If none on this machine then drop if strictly routed,
905: * or do nothing if loosely routed.
906: * Record interface address and bring up next address
907: * component. If strictly routed make sure next
908: * address is on directly accessible net.
909: */
910: case IPOPT_LSRR:
1.346 ozaki-r 911: case IPOPT_SSRR: {
912: struct psref psref;
1.353 ozaki-r 913: struct sockaddr_in ipaddr = {
914: .sin_len = sizeof(ipaddr),
915: .sin_family = AF_INET,
916: };
917:
1.47 cjs 918: if (ip_allowsrcrt == 0) {
919: type = ICMP_UNREACH;
920: code = ICMP_UNREACH_NET_PROHIB;
921: goto bad;
922: }
1.369 maxv 923: if (srr_present++) {
924: code = &cp[IPOPT_OPTVAL] - (u_char *)ip;
925: goto bad;
1.368 maxv 926: }
1.114 itojun 927: if (optlen < IPOPT_OFFSET + sizeof(*cp)) {
928: code = &cp[IPOPT_OLEN] - (u_char *)ip;
929: goto bad;
930: }
1.1 cgd 931: if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) {
932: code = &cp[IPOPT_OFFSET] - (u_char *)ip;
933: goto bad;
934: }
935: ipaddr.sin_addr = ip->ip_dst;
1.339 ozaki-r 936:
937: s = pserialize_read_enter();
938: ifa = ifa_ifwithaddr(sintosa(&ipaddr));
939: if (ifa == NULL) {
940: pserialize_read_exit(s);
1.1 cgd 941: if (opt == IPOPT_SSRR) {
942: type = ICMP_UNREACH;
943: code = ICMP_UNREACH_SRCFAIL;
944: goto bad;
945: }
946: /*
947: * Loose routing, and not at next destination
948: * yet; nothing to do except forward.
949: */
950: break;
951: }
1.339 ozaki-r 952: pserialize_read_exit(s);
953:
1.1 cgd 954: off--; /* 0 origin */
1.112 sommerfe 955: if ((off + sizeof(struct in_addr)) > optlen) {
1.1 cgd 956: /*
957: * End of source route. Should be for us.
958: */
1.354 ozaki-r 959: save_rte(m, cp, ip->ip_src);
1.1 cgd 960: break;
961: }
962: /*
963: * locate outgoing interface
964: */
1.281 tsutsui 965: memcpy((void *)&ipaddr.sin_addr, (void *)(cp + off),
1.1 cgd 966: sizeof(ipaddr.sin_addr));
1.339 ozaki-r 967: if (opt == IPOPT_SSRR) {
1.346 ozaki-r 968: ifa = ifa_ifwithladdr_psref(sintosa(&ipaddr),
969: &psref);
1.339 ozaki-r 970: if (ifa != NULL)
971: ia = ifatoia(ifa);
972: else
973: ia = NULL;
974: } else {
1.346 ozaki-r 975: ia = ip_rtaddr(ipaddr.sin_addr, &psref);
1.339 ozaki-r 976: }
977: if (ia == NULL) {
1.1 cgd 978: type = ICMP_UNREACH;
979: code = ICMP_UNREACH_SRCFAIL;
980: goto bad;
981: }
982: ip->ip_dst = ipaddr.sin_addr;
1.377 maxv 983: memcpy(cp + off, &ia->ia_addr.sin_addr,
984: sizeof(struct in_addr));
1.346 ozaki-r 985: ia4_release(ia, &psref);
1.1 cgd 986: cp[IPOPT_OFFSET] += sizeof(struct in_addr);
1.13 mycroft 987: /*
988: * Let ip_intr's mcast routing check handle mcast pkts
989: */
1.18 mycroft 990: forward = !IN_MULTICAST(ip->ip_dst.s_addr);
1.1 cgd 991: break;
1.346 ozaki-r 992: }
1.1 cgd 993:
1.346 ozaki-r 994: case IPOPT_RR: {
995: struct psref psref;
1.353 ozaki-r 996: struct sockaddr_in ipaddr = {
997: .sin_len = sizeof(ipaddr),
998: .sin_family = AF_INET,
999: };
1000:
1.368 maxv 1001: if (rr_present++) {
1002: code = &cp[IPOPT_OPTVAL] - (u_char *)ip;
1003: goto bad;
1004: }
1.114 itojun 1005: if (optlen < IPOPT_OFFSET + sizeof(*cp)) {
1006: code = &cp[IPOPT_OLEN] - (u_char *)ip;
1007: goto bad;
1008: }
1.1 cgd 1009: if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) {
1010: code = &cp[IPOPT_OFFSET] - (u_char *)ip;
1011: goto bad;
1012: }
1013: /*
1014: * If no space remains, ignore.
1015: */
1016: off--; /* 0 origin */
1.112 sommerfe 1017: if ((off + sizeof(struct in_addr)) > optlen)
1.1 cgd 1018: break;
1.377 maxv 1019: memcpy((void *)&ipaddr.sin_addr, (void *)&ip->ip_dst,
1.1 cgd 1020: sizeof(ipaddr.sin_addr));
1021: /*
1022: * locate outgoing interface; if we're the destination,
1023: * use the incoming interface (should be same).
1024: */
1.346 ozaki-r 1025: ifa = ifa_ifwithaddr_psref(sintosa(&ipaddr), &psref);
1.339 ozaki-r 1026: if (ifa == NULL) {
1.346 ozaki-r 1027: ia = ip_rtaddr(ipaddr.sin_addr, &psref);
1.339 ozaki-r 1028: if (ia == NULL) {
1029: type = ICMP_UNREACH;
1030: code = ICMP_UNREACH_HOST;
1031: goto bad;
1032: }
1033: } else {
1034: ia = ifatoia(ifa);
1.1 cgd 1035: }
1.377 maxv 1036: memcpy(cp + off, &ia->ia_addr.sin_addr,
1037: sizeof(struct in_addr));
1.346 ozaki-r 1038: ia4_release(ia, &psref);
1.1 cgd 1039: cp[IPOPT_OFFSET] += sizeof(struct in_addr);
1040: break;
1.346 ozaki-r 1041: }
1.1 cgd 1042:
1043: case IPOPT_TS:
1044: code = cp - (u_char *)ip;
1045: ipt = (struct ip_timestamp *)cp;
1.368 maxv 1046: if (ts_present++) {
1047: code = &cp[IPOPT_OPTVAL] - (u_char *)ip;
1048: goto bad;
1049: }
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: }
1.365 maxv 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:
1.329 ozaki-r 1072: case IPOPT_TS_TSANDADDR: {
1073: struct ifnet *rcvif;
1.339 ozaki-r 1074: int _s, _ss;
1.353 ozaki-r 1075: struct sockaddr_in ipaddr = {
1076: .sin_len = sizeof(ipaddr),
1077: .sin_family = AF_INET,
1078: };
1.329 ozaki-r 1079:
1.66 thorpej 1080: if (ipt->ipt_ptr - 1 + sizeof(n_time) +
1.114 itojun 1081: sizeof(struct in_addr) > ipt->ipt_len) {
1082: code = (u_char *)&ipt->ipt_ptr -
1083: (u_char *)ip;
1.1 cgd 1084: goto bad;
1.114 itojun 1085: }
1.13 mycroft 1086: ipaddr.sin_addr = dst;
1.339 ozaki-r 1087: _ss = pserialize_read_enter();
1088: rcvif = m_get_rcvif(m, &_s);
1.349 ozaki-r 1089: if (__predict_true(rcvif != NULL)) {
1090: ifa = ifaof_ifpforaddr(sintosa(&ipaddr),
1091: rcvif);
1092: }
1.339 ozaki-r 1093: m_put_rcvif(rcvif, &_s);
1094: if (ifa == NULL) {
1095: pserialize_read_exit(_ss);
1096: break;
1097: }
1098: ia = ifatoia(ifa);
1.377 maxv 1099: memcpy(cp0, &ia->ia_addr.sin_addr,
1100: sizeof(struct in_addr));
1.339 ozaki-r 1101: pserialize_read_exit(_ss);
1.1 cgd 1102: ipt->ipt_ptr += sizeof(struct in_addr);
1103: break;
1.329 ozaki-r 1104: }
1.1 cgd 1105:
1.353 ozaki-r 1106: case IPOPT_TS_PRESPEC: {
1107: struct sockaddr_in ipaddr = {
1108: .sin_len = sizeof(ipaddr),
1109: .sin_family = AF_INET,
1110: };
1111:
1.66 thorpej 1112: if (ipt->ipt_ptr - 1 + sizeof(n_time) +
1.114 itojun 1113: sizeof(struct in_addr) > ipt->ipt_len) {
1114: code = (u_char *)&ipt->ipt_ptr -
1115: (u_char *)ip;
1.1 cgd 1116: goto bad;
1.114 itojun 1117: }
1.281 tsutsui 1118: memcpy(&ipaddr.sin_addr, cp0,
1.1 cgd 1119: sizeof(struct in_addr));
1.339 ozaki-r 1120: s = pserialize_read_enter();
1121: ifa = ifa_ifwithaddr(sintosa(&ipaddr));
1122: if (ifa == NULL) {
1123: pserialize_read_exit(s);
1.1 cgd 1124: continue;
1.339 ozaki-r 1125: }
1126: pserialize_read_exit(s);
1.1 cgd 1127: ipt->ipt_ptr += sizeof(struct in_addr);
1128: break;
1.353 ozaki-r 1129: }
1.1 cgd 1130:
1131: default:
1.114 itojun 1132: /* XXX can't take &ipt->ipt_flg */
1133: code = (u_char *)&ipt->ipt_ptr -
1134: (u_char *)ip + 1;
1.1 cgd 1135: goto bad;
1136: }
1137: ntime = iptime();
1.107 thorpej 1138: cp0 = (u_char *) &ntime; /* XXX grumble, GCC... */
1.244 christos 1139: memmove((char *)cp + ipt->ipt_ptr - 1, cp0,
1.1 cgd 1140: sizeof(n_time));
1141: ipt->ipt_ptr += sizeof(n_time);
1142: }
1143: }
1144: if (forward) {
1.329 ozaki-r 1145: struct ifnet *rcvif;
1.339 ozaki-r 1146: struct psref _psref;
1.329 ozaki-r 1147:
1.26 thorpej 1148: if (ip_forwsrcrt == 0) {
1149: type = ICMP_UNREACH;
1150: code = ICMP_UNREACH_SRCFAIL;
1151: goto bad;
1152: }
1.329 ozaki-r 1153:
1.339 ozaki-r 1154: rcvif = m_get_rcvif_psref(m, &_psref);
1.330 ozaki-r 1155: if (__predict_false(rcvif == NULL)) {
1156: type = ICMP_UNREACH;
1157: code = ICMP_UNREACH_HOST;
1158: goto bad;
1159: }
1.329 ozaki-r 1160: ip_forward(m, 1, rcvif);
1.339 ozaki-r 1161: m_put_rcvif_psref(rcvif, &_psref);
1.311 rmind 1162: return true;
1.13 mycroft 1163: }
1.311 rmind 1164: return false;
1.1 cgd 1165: bad:
1.13 mycroft 1166: icmp_error(m, type, code, 0, 0);
1.266 thorpej 1167: IP_STATINC(IP_STAT_BADOPTIONS);
1.311 rmind 1168: return true;
1.1 cgd 1169: }
1170:
1171: /*
1.311 rmind 1172: * ip_rtaddr: given address of next destination (final or next hop),
1.1 cgd 1173: * return internet address info of interface to be used to get there.
1174: */
1.311 rmind 1175: static struct in_ifaddr *
1.346 ozaki-r 1176: ip_rtaddr(struct in_addr dst, struct psref *psref)
1.1 cgd 1177: {
1.249 dyoung 1178: struct rtentry *rt;
1179: union {
1180: struct sockaddr dst;
1181: struct sockaddr_in dst4;
1182: } u;
1.340 ozaki-r 1183: struct route *ro;
1.249 dyoung 1184:
1185: sockaddr_in_init(&u.dst4, &dst, 0);
1186:
1.391 ozaki-r 1187: ro = rtcache_percpu_getref(ipforward_rt_percpu);
1.340 ozaki-r 1188: rt = rtcache_lookup(ro, &u.dst);
1.345 ozaki-r 1189: if (rt == NULL) {
1.391 ozaki-r 1190: rtcache_percpu_putref(ipforward_rt_percpu);
1.345 ozaki-r 1191: return NULL;
1192: }
1193:
1.346 ozaki-r 1194: ia4_acquire(ifatoia(rt->rt_ifa), psref);
1.345 ozaki-r 1195: rtcache_unref(rt, ro);
1.391 ozaki-r 1196: rtcache_percpu_putref(ipforward_rt_percpu);
1.249 dyoung 1197:
1198: return ifatoia(rt->rt_ifa);
1.1 cgd 1199: }
1200:
1201: /*
1.311 rmind 1202: * save_rte: save incoming source route for use in replies, to be picked
1203: * up later by ip_srcroute if the receiver is interested.
1.1 cgd 1204: */
1.311 rmind 1205: static void
1.354 ozaki-r 1206: save_rte(struct mbuf *m, u_char *option, struct in_addr dst)
1.1 cgd 1207: {
1.354 ozaki-r 1208: struct ip_srcrt *isr;
1209: struct m_tag *mtag;
1.1 cgd 1210: unsigned olen;
1211:
1212: olen = option[IPOPT_OLEN];
1.354 ozaki-r 1213: if (olen > sizeof(isr->isr_hdr) + sizeof(isr->isr_routes))
1214: return;
1215:
1216: mtag = m_tag_get(PACKET_TAG_SRCROUTE, sizeof(*isr), M_NOWAIT);
1217: if (mtag == NULL)
1.1 cgd 1218: return;
1.354 ozaki-r 1219: isr = (struct ip_srcrt *)(mtag + 1);
1220:
1221: memcpy(isr->isr_hdr, option, olen);
1222: isr->isr_nhops = (olen - IPOPT_OFFSET - 1) / sizeof(struct in_addr);
1223: isr->isr_dst = dst;
1224: m_tag_prepend(m, mtag);
1.1 cgd 1225: }
1226:
1227: /*
1228: * Retrieve incoming source route for use in replies,
1229: * in the same form used by setsockopt.
1230: * The first hop is placed before the options, will be removed later.
1231: */
1232: struct mbuf *
1.354 ozaki-r 1233: ip_srcroute(struct mbuf *m0)
1.1 cgd 1234: {
1.109 augustss 1235: struct in_addr *p, *q;
1236: struct mbuf *m;
1.354 ozaki-r 1237: struct ip_srcrt *isr;
1238: struct m_tag *mtag;
1239:
1.387 maxv 1240: mtag = m_tag_find(m0, PACKET_TAG_SRCROUTE);
1.354 ozaki-r 1241: if (mtag == NULL)
1242: return NULL;
1243: isr = (struct ip_srcrt *)(mtag + 1);
1.1 cgd 1244:
1.354 ozaki-r 1245: if (isr->isr_nhops == 0)
1.237 dyoung 1246: return NULL;
1.354 ozaki-r 1247:
1.1 cgd 1248: m = m_get(M_DONTWAIT, MT_SOOPTS);
1.354 ozaki-r 1249: if (m == NULL)
1.237 dyoung 1250: return NULL;
1.1 cgd 1251:
1.164 matt 1252: MCLAIM(m, &inetdomain.dom_mowner);
1.354 ozaki-r 1253: #define OPTSIZ (sizeof(isr->isr_nop) + sizeof(isr->isr_hdr))
1.1 cgd 1254:
1.354 ozaki-r 1255: /* length is (nhops+1)*sizeof(addr) + sizeof(nop + header) */
1256: m->m_len = (isr->isr_nhops + 1) * sizeof(struct in_addr) + OPTSIZ;
1.1 cgd 1257:
1258: /*
1259: * First save first hop for return route
1260: */
1.354 ozaki-r 1261: p = &(isr->isr_routes[isr->isr_nhops - 1]);
1.1 cgd 1262: *(mtod(m, struct in_addr *)) = *p--;
1263:
1264: /*
1265: * Copy option fields and padding (nop) to mbuf.
1266: */
1.354 ozaki-r 1267: isr->isr_nop = IPOPT_NOP;
1268: isr->isr_hdr[IPOPT_OFFSET] = IPOPT_MINOFF;
1269: memmove(mtod(m, char *) + sizeof(struct in_addr), &isr->isr_nop,
1.244 christos 1270: OPTSIZ);
1271: q = (struct in_addr *)(mtod(m, char *) +
1.1 cgd 1272: sizeof(struct in_addr) + OPTSIZ);
1273: #undef OPTSIZ
1274: /*
1275: * Record return path as an IP source route,
1276: * reversing the path (pointers are now aligned).
1277: */
1.354 ozaki-r 1278: while (p >= isr->isr_routes) {
1.1 cgd 1279: *q++ = *p--;
1280: }
1281: /*
1282: * Last hop goes to final destination.
1283: */
1.354 ozaki-r 1284: *q = isr->isr_dst;
1285: m_tag_delete(m0, mtag);
1286: return m;
1.1 cgd 1287: }
1288:
1.139 matt 1289: const int inetctlerrmap[PRC_NCMDS] = {
1.256 yamt 1290: [PRC_MSGSIZE] = EMSGSIZE,
1291: [PRC_HOSTDEAD] = EHOSTDOWN,
1292: [PRC_HOSTUNREACH] = EHOSTUNREACH,
1293: [PRC_UNREACH_NET] = EHOSTUNREACH,
1294: [PRC_UNREACH_HOST] = EHOSTUNREACH,
1295: [PRC_UNREACH_PROTOCOL] = ECONNREFUSED,
1296: [PRC_UNREACH_PORT] = ECONNREFUSED,
1297: [PRC_UNREACH_SRCFAIL] = EHOSTUNREACH,
1298: [PRC_PARAMPROB] = ENOPROTOOPT,
1.1 cgd 1299: };
1300:
1.295 dyoung 1301: void
1302: ip_fasttimo(void)
1303: {
1304: if (ip_drainwanted) {
1305: ip_drain();
1306: ip_drainwanted = 0;
1307: }
1308: }
1309:
1310: void
1311: ip_drainstub(void)
1312: {
1313: ip_drainwanted = 1;
1314: }
1315:
1.1 cgd 1316: /*
1317: * Forward a packet. If some error occurs return the sender
1318: * an icmp packet. Note we can't always generate a meaningful
1319: * icmp message because icmp doesn't have a large enough repertoire
1320: * of codes and types.
1321: *
1322: * If not forwarding, just drop the packet. This could be confusing
1323: * if ipforwarding was zero but some routing protocol was advancing
1324: * us as a gateway to somewhere. However, we must let the routing
1325: * protocol deal with that.
1326: *
1327: * The srcrt parameter indicates whether the packet is being forwarded
1328: * via a source route.
1329: */
1.313 rmind 1330: static void
1.329 ozaki-r 1331: ip_forward(struct mbuf *m, int srcrt, struct ifnet *rcvif)
1.1 cgd 1332: {
1.109 augustss 1333: struct ip *ip = mtod(m, struct ip *);
1334: struct rtentry *rt;
1.220 christos 1335: int error, type = 0, code = 0, destmtu = 0;
1.1 cgd 1336: struct mbuf *mcopy;
1.13 mycroft 1337: n_long dest;
1.249 dyoung 1338: union {
1339: struct sockaddr dst;
1340: struct sockaddr_in dst4;
1341: } u;
1.320 ozaki-r 1342: uint64_t *ips;
1.340 ozaki-r 1343: struct route *ro;
1.164 matt 1344:
1.318 rmind 1345: KASSERTMSG(cpu_softintr_p(), "ip_forward: not in the software "
1346: "interrupt handler; synchronization assumptions violated");
1347:
1.164 matt 1348: /*
1349: * We are now in the output path.
1350: */
1351: MCLAIM(m, &ip_tx_mowner);
1.135 thorpej 1352:
1353: /*
1354: * Clear any in-bound checksum flags for this packet.
1355: */
1356: m->m_pkthdr.csum_flags = 0;
1.1 cgd 1357:
1.13 mycroft 1358: dest = 0;
1.93 sommerfe 1359: if (m->m_flags & (M_BCAST|M_MCAST) || in_canforward(ip->ip_dst) == 0) {
1.266 thorpej 1360: IP_STATINC(IP_STAT_CANTFORWARD);
1.1 cgd 1361: m_freem(m);
1362: return;
1363: }
1.318 rmind 1364:
1.1 cgd 1365: if (ip->ip_ttl <= IPTTLDEC) {
1.13 mycroft 1366: icmp_error(m, ICMP_TIMXCEED, ICMP_TIMXCEED_INTRANS, dest, 0);
1.1 cgd 1367: return;
1368: }
1369:
1.249 dyoung 1370: sockaddr_in_init(&u.dst4, &ip->ip_dst, 0);
1.318 rmind 1371:
1.392 ozaki-r 1372: ro = rtcache_percpu_getref(ipforward_rt_percpu);
1.345 ozaki-r 1373: rt = rtcache_lookup(ro, &u.dst);
1374: if (rt == NULL) {
1.391 ozaki-r 1375: rtcache_percpu_putref(ipforward_rt_percpu);
1.249 dyoung 1376: icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_NET, dest, 0);
1377: return;
1.1 cgd 1378: }
1379:
1380: /*
1.34 mycroft 1381: * Save at most 68 bytes of the packet in case
1.1 cgd 1382: * we need to generate an ICMP message to the src.
1.119 itojun 1383: * Pullup to avoid sharing mbuf cluster between m and mcopy.
1.1 cgd 1384: */
1.155 itojun 1385: mcopy = m_copym(m, 0, imin(ntohs(ip->ip_len), 68), M_DONTWAIT);
1.119 itojun 1386: if (mcopy)
1387: mcopy = m_pullup(mcopy, ip->ip_hl << 2);
1.1 cgd 1388:
1.221 christos 1389: ip->ip_ttl -= IPTTLDEC;
1390:
1.1 cgd 1391: /*
1392: * If forwarding packet using same interface that it came in on,
1393: * perhaps should send a redirect to sender to shortcut a hop.
1394: * Only send redirect if source is sending directly to us,
1395: * and if packet was not source routed (or has any options).
1396: * Also, don't send redirect if forwarding using a default route
1397: * or a route modified by a redirect.
1398: */
1.329 ozaki-r 1399: if (rt->rt_ifp == rcvif &&
1.1 cgd 1400: (rt->rt_flags & (RTF_DYNAMIC|RTF_MODIFIED)) == 0 &&
1.250 dyoung 1401: !in_nullhost(satocsin(rt_getkey(rt))->sin_addr) &&
1.1 cgd 1402: ipsendredirects && !srcrt) {
1.393 ozaki-r 1403: if ((ip->ip_src.s_addr & ifatoia(rt->rt_ifa)->ia_subnetmask) ==
1.19 mycroft 1404: ifatoia(rt->rt_ifa)->ia_subnet) {
1.77 thorpej 1405: if (rt->rt_flags & RTF_GATEWAY)
1406: dest = satosin(rt->rt_gateway)->sin_addr.s_addr;
1407: else
1408: dest = ip->ip_dst.s_addr;
1409: /*
1410: * Router requirements says to only send host
1411: * redirects.
1412: */
1413: type = ICMP_REDIRECT;
1414: code = ICMP_REDIRECT_HOST;
1.1 cgd 1415: }
1416: }
1.345 ozaki-r 1417: rtcache_unref(rt, ro);
1.1 cgd 1418:
1.340 ozaki-r 1419: error = ip_output(m, NULL, ro,
1.173 jonathan 1420: (IP_FORWARDING | (ip_directedbcast ? IP_ALLOWBROADCAST : 0)),
1.296 plunky 1421: NULL, NULL);
1.173 jonathan 1422:
1.320 ozaki-r 1423: if (error) {
1.266 thorpej 1424: IP_STATINC(IP_STAT_CANTFORWARD);
1.320 ozaki-r 1425: goto error;
1426: }
1427:
1428: ips = IP_STAT_GETREF();
1429: ips[IP_STAT_FORWARD]++;
1430:
1431: if (type) {
1432: ips[IP_STAT_REDIRECTSENT]++;
1433: IP_STAT_PUTREF();
1434: goto redirect;
1435: }
1436:
1437: IP_STAT_PUTREF();
1438: if (mcopy) {
1.63 matt 1439: #ifdef GATEWAY
1.320 ozaki-r 1440: if (mcopy->m_flags & M_CANFASTFWD)
1.340 ozaki-r 1441: ipflow_create(ro, mcopy);
1.63 matt 1442: #endif
1.320 ozaki-r 1443: m_freem(mcopy);
1.1 cgd 1444: }
1.320 ozaki-r 1445:
1.391 ozaki-r 1446: rtcache_percpu_putref(ipforward_rt_percpu);
1.320 ozaki-r 1447: return;
1448:
1449: redirect:
1450: error:
1.318 rmind 1451: if (mcopy == NULL) {
1.391 ozaki-r 1452: rtcache_percpu_putref(ipforward_rt_percpu);
1.1 cgd 1453: return;
1.318 rmind 1454: }
1.13 mycroft 1455:
1.1 cgd 1456: switch (error) {
1457:
1458: case 0: /* forwarded, but need redirect */
1459: /* type, code set above */
1460: break;
1461:
1462: case ENETUNREACH: /* shouldn't happen, checked above */
1463: case EHOSTUNREACH:
1464: case ENETDOWN:
1465: case EHOSTDOWN:
1466: default:
1467: type = ICMP_UNREACH;
1468: code = ICMP_UNREACH_HOST;
1469: break;
1470:
1471: case EMSGSIZE:
1472: type = ICMP_UNREACH;
1473: code = ICMP_UNREACH_NEEDFRAG;
1.263 cube 1474:
1.345 ozaki-r 1475: if ((rt = rtcache_validate(ro)) != NULL) {
1.274 seanb 1476: destmtu = rt->rt_ifp->if_mtu;
1.345 ozaki-r 1477: rtcache_unref(rt, ro);
1478: }
1.305 rmind 1479: #ifdef IPSEC
1.317 christos 1480: if (ipsec_used)
1.382 maxv 1481: ipsec_mtu(mcopy, &destmtu);
1.305 rmind 1482: #endif
1.266 thorpej 1483: IP_STATINC(IP_STAT_CANTFRAG);
1.1 cgd 1484: break;
1485:
1486: case ENOBUFS:
1.143 itojun 1487: /*
1.311 rmind 1488: * Do not generate ICMP_SOURCEQUENCH as required in RFC 1812,
1489: * Requirements for IP Version 4 Routers. Source quench can
1.373 maxv 1490: * be a big problem under DoS attacks or if the underlying
1.311 rmind 1491: * interface is rate-limited.
1.143 itojun 1492: */
1493: if (mcopy)
1494: m_freem(mcopy);
1.391 ozaki-r 1495: rtcache_percpu_putref(ipforward_rt_percpu);
1.143 itojun 1496: return;
1.1 cgd 1497: }
1.220 christos 1498: icmp_error(mcopy, type, code, dest, destmtu);
1.392 ozaki-r 1499: rtcache_percpu_putref(ipforward_rt_percpu);
1.44 thorpej 1500: }
1501:
1502: void
1.211 perry 1503: ip_savecontrol(struct inpcb *inp, struct mbuf **mp, struct ip *ip,
1504: struct mbuf *m)
1.44 thorpej 1505: {
1.311 rmind 1506: struct socket *so = inp->inp_socket;
1507: int inpflags = inp->inp_flags;
1.330 ozaki-r 1508:
1.356 christos 1509: if (SOOPT_TIMESTAMP(so->so_options))
1.381 maxv 1510: mp = sbsavetimestamp(so->so_options, mp);
1.356 christos 1511:
1.311 rmind 1512: if (inpflags & INP_RECVDSTADDR) {
1.357 christos 1513: *mp = sbcreatecontrol(&ip->ip_dst,
1.44 thorpej 1514: sizeof(struct in_addr), IP_RECVDSTADDR, IPPROTO_IP);
1515: if (*mp)
1516: mp = &(*mp)->m_next;
1517: }
1.358 christos 1518:
1519: if (inpflags & INP_RECVTTL) {
1520: *mp = sbcreatecontrol(&ip->ip_ttl,
1521: sizeof(uint8_t), IP_RECVTTL, IPPROTO_IP);
1522: if (*mp)
1523: mp = &(*mp)->m_next;
1524: }
1525:
1526: struct psref psref;
1527: ifnet_t *ifp = m_get_rcvif_psref(m, &psref);
1528: if (__predict_false(ifp == NULL)) {
1529: #ifdef DIAGNOSTIC
1530: printf("%s: missing receive interface\n", __func__);
1531: #endif
1532: return; /* XXX should report error? */
1533: }
1534:
1.311 rmind 1535: if (inpflags & INP_RECVPKTINFO) {
1.306 christos 1536: struct in_pktinfo ipi;
1.307 christos 1537: ipi.ipi_addr = ip->ip_dst;
1.314 rmind 1538: ipi.ipi_ifindex = ifp->if_index;
1.357 christos 1539: *mp = sbcreatecontrol(&ipi,
1.306 christos 1540: sizeof(ipi), IP_PKTINFO, IPPROTO_IP);
1541: if (*mp)
1542: mp = &(*mp)->m_next;
1543: }
1.311 rmind 1544: if (inpflags & INP_RECVIF) {
1.44 thorpej 1545: struct sockaddr_dl sdl;
1546:
1.337 ozaki-r 1547: sockaddr_dl_init(&sdl, sizeof(sdl), ifp->if_index, 0, NULL, 0,
1548: NULL, 0);
1.251 dyoung 1549: *mp = sbcreatecontrol(&sdl, sdl.sdl_len, IP_RECVIF, IPPROTO_IP);
1.44 thorpej 1550: if (*mp)
1551: mp = &(*mp)->m_next;
1552: }
1.329 ozaki-r 1553: m_put_rcvif_psref(ifp, &psref);
1.13 mycroft 1554: }
1555:
1.189 atatat 1556: /*
1.228 elad 1557: * sysctl helper routine for net.inet.ip.forwsrcrt.
1558: */
1559: static int
1560: sysctl_net_inet_ip_forwsrcrt(SYSCTLFN_ARGS)
1561: {
1562: int error, tmp;
1563: struct sysctlnode node;
1564:
1565: node = *rnode;
1566: tmp = ip_forwsrcrt;
1567: node.sysctl_data = &tmp;
1568: error = sysctl_lookup(SYSCTLFN_CALL(&node));
1569: if (error || newp == NULL)
1570: return (error);
1571:
1.280 elad 1572: error = kauth_authorize_network(l->l_cred, KAUTH_NETWORK_FORWSRCRT,
1573: 0, NULL, NULL, NULL);
1574: if (error)
1575: return (error);
1.228 elad 1576:
1577: ip_forwsrcrt = tmp;
1578:
1579: return (0);
1580: }
1581:
1582: /*
1.189 atatat 1583: * sysctl helper routine for net.inet.ip.mtudisctimeout. checks the
1584: * range of the new value and tweaks timers if it changes.
1585: */
1586: static int
1587: sysctl_net_inet_ip_pmtudto(SYSCTLFN_ARGS)
1.13 mycroft 1588: {
1.189 atatat 1589: int error, tmp;
1590: struct sysctlnode node;
1591:
1.350 ozaki-r 1592: icmp_mtudisc_lock();
1593:
1.189 atatat 1594: node = *rnode;
1595: tmp = ip_mtudisc_timeout;
1596: node.sysctl_data = &tmp;
1597: error = sysctl_lookup(SYSCTLFN_CALL(&node));
1598: if (error || newp == NULL)
1.350 ozaki-r 1599: goto out;
1600: if (tmp < 0) {
1601: error = EINVAL;
1602: goto out;
1603: }
1.273 matt 1604:
1.189 atatat 1605: ip_mtudisc_timeout = tmp;
1606: rt_timer_queue_change(ip_mtudisc_timeout_q, ip_mtudisc_timeout);
1.350 ozaki-r 1607: error = 0;
1608: out:
1609: icmp_mtudisc_unlock();
1.351 ozaki-r 1610: return error;
1.189 atatat 1611: }
1.54 lukem 1612:
1.266 thorpej 1613: static int
1614: sysctl_net_inet_ip_stats(SYSCTLFN_ARGS)
1615: {
1616:
1.271 thorpej 1617: return (NETSTAT_SYSCTL(ipstat_percpu, IP_NSTATS));
1.266 thorpej 1618: }
1.131 itojun 1619:
1.284 pooka 1620: static void
1621: sysctl_net_inet_ip_setup(struct sysctllog **clog)
1.189 atatat 1622: {
1.197 atatat 1623: sysctl_createv(clog, 0, NULL, NULL,
1624: CTLFLAG_PERMANENT,
1.203 atatat 1625: CTLTYPE_NODE, "inet",
1626: SYSCTL_DESCR("PF_INET related settings"),
1.189 atatat 1627: NULL, 0, NULL, 0,
1628: CTL_NET, PF_INET, CTL_EOL);
1.197 atatat 1629: sysctl_createv(clog, 0, NULL, NULL,
1630: CTLFLAG_PERMANENT,
1.203 atatat 1631: CTLTYPE_NODE, "ip",
1632: SYSCTL_DESCR("IPv4 related settings"),
1.189 atatat 1633: NULL, 0, NULL, 0,
1634: CTL_NET, PF_INET, IPPROTO_IP, CTL_EOL);
1.212 perry 1635:
1.197 atatat 1636: sysctl_createv(clog, 0, NULL, NULL,
1637: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1638: CTLTYPE_INT, "forwarding",
1639: SYSCTL_DESCR("Enable forwarding of INET datagrams"),
1.189 atatat 1640: NULL, 0, &ipforwarding, 0,
1641: CTL_NET, PF_INET, IPPROTO_IP,
1642: IPCTL_FORWARDING, CTL_EOL);
1.197 atatat 1643: sysctl_createv(clog, 0, NULL, NULL,
1644: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1645: CTLTYPE_INT, "redirect",
1646: SYSCTL_DESCR("Enable sending of ICMP redirect messages"),
1.189 atatat 1647: NULL, 0, &ipsendredirects, 0,
1648: CTL_NET, PF_INET, IPPROTO_IP,
1649: IPCTL_SENDREDIRECTS, CTL_EOL);
1.197 atatat 1650: sysctl_createv(clog, 0, NULL, NULL,
1651: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1652: CTLTYPE_INT, "ttl",
1653: SYSCTL_DESCR("Default TTL for an INET datagram"),
1.189 atatat 1654: NULL, 0, &ip_defttl, 0,
1655: CTL_NET, PF_INET, IPPROTO_IP,
1656: IPCTL_DEFTTL, CTL_EOL);
1.197 atatat 1657: sysctl_createv(clog, 0, NULL, NULL,
1.228 elad 1658: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1659: CTLTYPE_INT, "forwsrcrt",
1660: SYSCTL_DESCR("Enable forwarding of source-routed "
1661: "datagrams"),
1.228 elad 1662: sysctl_net_inet_ip_forwsrcrt, 0, &ip_forwsrcrt, 0,
1.189 atatat 1663: CTL_NET, PF_INET, IPPROTO_IP,
1664: IPCTL_FORWSRCRT, CTL_EOL);
1.197 atatat 1665: sysctl_createv(clog, 0, NULL, NULL,
1666: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1667: CTLTYPE_INT, "directed-broadcast",
1668: SYSCTL_DESCR("Enable forwarding of broadcast datagrams"),
1.189 atatat 1669: NULL, 0, &ip_directedbcast, 0,
1670: CTL_NET, PF_INET, IPPROTO_IP,
1671: IPCTL_DIRECTEDBCAST, CTL_EOL);
1.197 atatat 1672: sysctl_createv(clog, 0, NULL, NULL,
1673: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1674: CTLTYPE_INT, "allowsrcrt",
1675: SYSCTL_DESCR("Accept source-routed datagrams"),
1.189 atatat 1676: NULL, 0, &ip_allowsrcrt, 0,
1677: CTL_NET, PF_INET, IPPROTO_IP,
1678: IPCTL_ALLOWSRCRT, CTL_EOL);
1.311 rmind 1679:
1.197 atatat 1680: sysctl_createv(clog, 0, NULL, NULL,
1681: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1682: CTLTYPE_INT, "mtudisc",
1683: SYSCTL_DESCR("Use RFC1191 Path MTU Discovery"),
1.189 atatat 1684: NULL, 0, &ip_mtudisc, 0,
1685: CTL_NET, PF_INET, IPPROTO_IP,
1686: IPCTL_MTUDISC, CTL_EOL);
1.197 atatat 1687: sysctl_createv(clog, 0, NULL, NULL,
1688: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1689: CTLTYPE_INT, "anonportmin",
1690: SYSCTL_DESCR("Lowest ephemeral port number to assign"),
1.189 atatat 1691: sysctl_net_inet_ip_ports, 0, &anonportmin, 0,
1692: CTL_NET, PF_INET, IPPROTO_IP,
1693: IPCTL_ANONPORTMIN, CTL_EOL);
1.197 atatat 1694: sysctl_createv(clog, 0, NULL, NULL,
1695: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1696: CTLTYPE_INT, "anonportmax",
1697: SYSCTL_DESCR("Highest ephemeral port number to assign"),
1.189 atatat 1698: sysctl_net_inet_ip_ports, 0, &anonportmax, 0,
1699: CTL_NET, PF_INET, IPPROTO_IP,
1700: IPCTL_ANONPORTMAX, CTL_EOL);
1.197 atatat 1701: sysctl_createv(clog, 0, NULL, NULL,
1702: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1703: CTLTYPE_INT, "mtudisctimeout",
1704: SYSCTL_DESCR("Lifetime of a Path MTU Discovered route"),
1.300 dsl 1705: sysctl_net_inet_ip_pmtudto, 0, (void *)&ip_mtudisc_timeout, 0,
1.189 atatat 1706: CTL_NET, PF_INET, IPPROTO_IP,
1707: IPCTL_MTUDISCTIMEOUT, CTL_EOL);
1708: #ifndef IPNOPRIVPORTS
1.197 atatat 1709: sysctl_createv(clog, 0, NULL, NULL,
1710: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1711: CTLTYPE_INT, "lowportmin",
1712: SYSCTL_DESCR("Lowest privileged ephemeral port number "
1713: "to assign"),
1.189 atatat 1714: sysctl_net_inet_ip_ports, 0, &lowportmin, 0,
1715: CTL_NET, PF_INET, IPPROTO_IP,
1716: IPCTL_LOWPORTMIN, CTL_EOL);
1.197 atatat 1717: sysctl_createv(clog, 0, NULL, NULL,
1718: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1719: CTLTYPE_INT, "lowportmax",
1720: SYSCTL_DESCR("Highest privileged ephemeral port number "
1721: "to assign"),
1.189 atatat 1722: sysctl_net_inet_ip_ports, 0, &lowportmax, 0,
1723: CTL_NET, PF_INET, IPPROTO_IP,
1724: IPCTL_LOWPORTMAX, CTL_EOL);
1725: #endif /* IPNOPRIVPORTS */
1726: #if NGRE > 0
1.197 atatat 1727: sysctl_createv(clog, 0, NULL, NULL,
1728: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1729: CTLTYPE_INT, "grettl",
1730: SYSCTL_DESCR("Default TTL for a gre tunnel datagram"),
1.189 atatat 1731: NULL, 0, &ip_gre_ttl, 0,
1732: CTL_NET, PF_INET, IPPROTO_IP,
1733: IPCTL_GRE_TTL, CTL_EOL);
1734: #endif /* NGRE */
1.197 atatat 1735: sysctl_createv(clog, 0, NULL, NULL,
1736: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1737: CTLTYPE_INT, "checkinterface",
1738: SYSCTL_DESCR("Enable receive side of Strong ES model "
1739: "from RFC1122"),
1.189 atatat 1740: NULL, 0, &ip_checkinterface, 0,
1741: CTL_NET, PF_INET, IPPROTO_IP,
1742: IPCTL_CHECKINTERFACE, CTL_EOL);
1.197 atatat 1743: sysctl_createv(clog, 0, NULL, NULL,
1744: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1745: CTLTYPE_INT, "random_id",
1746: SYSCTL_DESCR("Assign random ip_id values"),
1.189 atatat 1747: NULL, 0, &ip_do_randomid, 0,
1748: CTL_NET, PF_INET, IPPROTO_IP,
1749: IPCTL_RANDOMID, CTL_EOL);
1.206 thorpej 1750: sysctl_createv(clog, 0, NULL, NULL,
1751: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1752: CTLTYPE_INT, "do_loopback_cksum",
1753: SYSCTL_DESCR("Perform IP checksum on loopback"),
1754: NULL, 0, &ip_do_loopback_cksum, 0,
1755: CTL_NET, PF_INET, IPPROTO_IP,
1756: IPCTL_LOOPBACKCKSUM, CTL_EOL);
1.219 elad 1757: sysctl_createv(clog, 0, NULL, NULL,
1758: CTLFLAG_PERMANENT,
1759: CTLTYPE_STRUCT, "stats",
1760: SYSCTL_DESCR("IP statistics"),
1.266 thorpej 1761: sysctl_net_inet_ip_stats, 0, NULL, 0,
1.219 elad 1762: CTL_NET, PF_INET, IPPROTO_IP, IPCTL_STATS,
1763: CTL_EOL);
1.322 joerg 1764: #if NARP
1.321 roy 1765: sysctl_createv(clog, 0, NULL, NULL,
1766: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1767: CTLTYPE_INT, "dad_count",
1768: SYSCTL_DESCR("Number of Duplicate Address Detection "
1769: "probes to send"),
1770: NULL, 0, &ip_dad_count, 0,
1771: CTL_NET, PF_INET, IPPROTO_IP,
1772: IPCTL_DAD_COUNT, CTL_EOL);
1.322 joerg 1773: #endif
1.301 christos 1774:
1775: /* anonportalgo RFC6056 subtree */
1.302 christos 1776: const struct sysctlnode *portalgo_node;
1777: sysctl_createv(clog, 0, NULL, &portalgo_node,
1.301 christos 1778: CTLFLAG_PERMANENT,
1779: CTLTYPE_NODE, "anonportalgo",
1780: SYSCTL_DESCR("Anonymous Port Algorithm Selection (RFC 6056)"),
1781: NULL, 0, NULL, 0,
1782: CTL_NET, PF_INET, IPPROTO_IP, CTL_CREATE, CTL_EOL);
1.302 christos 1783: sysctl_createv(clog, 0, &portalgo_node, NULL,
1.301 christos 1784: CTLFLAG_PERMANENT,
1785: CTLTYPE_STRING, "available",
1786: SYSCTL_DESCR("available algorithms"),
1.302 christos 1787: sysctl_portalgo_available, 0, NULL, PORTALGO_MAXLEN,
1.301 christos 1788: CTL_CREATE, CTL_EOL);
1.302 christos 1789: sysctl_createv(clog, 0, &portalgo_node, NULL,
1.301 christos 1790: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1791: CTLTYPE_STRING, "selected",
1792: SYSCTL_DESCR("selected algorithm"),
1.303 christos 1793: sysctl_portalgo_selected4, 0, NULL, PORTALGO_MAXLEN,
1794: CTL_CREATE, CTL_EOL);
1795: sysctl_createv(clog, 0, &portalgo_node, NULL,
1796: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1797: CTLTYPE_STRUCT, "reserve",
1798: SYSCTL_DESCR("bitmap of reserved ports"),
1799: sysctl_portalgo_reserve4, 0, NULL, 0,
1.301 christos 1800: CTL_CREATE, CTL_EOL);
1.1 cgd 1801: }
1.266 thorpej 1802:
1803: void
1804: ip_statinc(u_int stat)
1805: {
1806:
1807: KASSERT(stat < IP_NSTATS);
1808: IP_STATINC(stat);
1809: }
CVSweb <webmaster@jp.NetBSD.org>