Annotation of src/sys/netinet/ip_input.c, Revision 1.337.2.5
1.337.2.5! pgoyette 1: /* $NetBSD: ip_input.c,v 1.354 2017/03/31 06:49:44 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:
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>
1.337.2.5! pgoyette 94: __KERNEL_RCSID(0, "$NetBSD: ip_input.c,v 1.354 2017/03/31 06:49:44 ozaki-r Exp $");
1.55 scottr 95:
1.324 pooka 96: #ifdef _KERNEL_OPT
1.184 jonathan 97: #include "opt_inet.h"
1.278 christos 98: #include "opt_compat_netbsd.h"
1.62 matt 99: #include "opt_gateway.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"
1.337.2.2 pgoyette 104: #include "opt_net_mpsafe.h"
1.324 pooka 105: #endif
1.1 cgd 106:
1.325 roy 107: #include "arp.h"
108:
1.5 mycroft 109: #include <sys/param.h>
110: #include <sys/systm.h>
1.318 rmind 111: #include <sys/cpu.h>
1.5 mycroft 112: #include <sys/mbuf.h>
113: #include <sys/domain.h>
114: #include <sys/protosw.h>
115: #include <sys/socket.h>
1.44 thorpej 116: #include <sys/socketvar.h>
1.5 mycroft 117: #include <sys/errno.h>
118: #include <sys/time.h>
119: #include <sys/kernel.h>
1.72 thorpej 120: #include <sys/pool.h>
1.28 christos 121: #include <sys/sysctl.h>
1.230 elad 122: #include <sys/kauth.h>
1.1 cgd 123:
1.5 mycroft 124: #include <net/if.h>
1.44 thorpej 125: #include <net/if_dl.h>
1.5 mycroft 126: #include <net/route.h>
1.318 rmind 127: #include <net/pktqueue.h>
1.45 mrg 128: #include <net/pfil.h>
1.1 cgd 129:
1.5 mycroft 130: #include <netinet/in.h>
131: #include <netinet/in_systm.h>
132: #include <netinet/ip.h>
133: #include <netinet/in_pcb.h>
1.215 yamt 134: #include <netinet/in_proto.h>
1.5 mycroft 135: #include <netinet/in_var.h>
136: #include <netinet/ip_var.h>
1.266 thorpej 137: #include <netinet/ip_private.h>
1.5 mycroft 138: #include <netinet/ip_icmp.h>
1.89 itojun 139: /* just for gif_ttl */
140: #include <netinet/in_gif.h>
141: #include "gif.h"
1.144 martin 142: #include <net/if_gre.h>
143: #include "gre.h"
1.111 jdolecek 144:
145: #ifdef MROUTING
146: #include <netinet/ip_mroute.h>
147: #endif
1.302 christos 148: #include <netinet/portalgo.h>
1.89 itojun 149:
1.304 christos 150: #ifdef IPSEC
1.173 jonathan 151: #include <netipsec/ipsec.h>
1.305 rmind 152: #endif
1.44 thorpej 153:
1.1 cgd 154: #ifndef IPFORWARDING
155: #ifdef GATEWAY
156: #define IPFORWARDING 1 /* forward IP packets not for us */
157: #else /* GATEWAY */
158: #define IPFORWARDING 0 /* don't forward IP packets not for us */
159: #endif /* GATEWAY */
160: #endif /* IPFORWARDING */
161: #ifndef IPSENDREDIRECTS
162: #define IPSENDREDIRECTS 1
163: #endif
1.26 thorpej 164: #ifndef IPFORWSRCRT
1.47 cjs 165: #define IPFORWSRCRT 1 /* forward source-routed packets */
166: #endif
167: #ifndef IPALLOWSRCRT
1.48 mrg 168: #define IPALLOWSRCRT 1 /* allow source-routed packets */
1.26 thorpej 169: #endif
1.53 kml 170: #ifndef IPMTUDISC
1.153 itojun 171: #define IPMTUDISC 1
1.53 kml 172: #endif
1.60 kml 173: #ifndef IPMTUDISCTIMEOUT
1.61 kml 174: #define IPMTUDISCTIMEOUT (10 * 60) /* as per RFC 1191 */
1.60 kml 175: #endif
1.53 kml 176:
1.278 christos 177: #ifdef COMPAT_50
178: #include <compat/sys/time.h>
179: #include <compat/sys/socket.h>
180: #endif
181:
1.27 thorpej 182: /*
183: * Note: DIRECTED_BROADCAST is handled this way so that previous
184: * configuration using this option will Just Work.
185: */
186: #ifndef IPDIRECTEDBCAST
187: #ifdef DIRECTED_BROADCAST
188: #define IPDIRECTEDBCAST 1
189: #else
190: #define IPDIRECTEDBCAST 0
191: #endif /* DIRECTED_BROADCAST */
192: #endif /* IPDIRECTEDBCAST */
1.1 cgd 193: int ipforwarding = IPFORWARDING;
194: int ipsendredirects = IPSENDREDIRECTS;
1.13 mycroft 195: int ip_defttl = IPDEFTTL;
1.26 thorpej 196: int ip_forwsrcrt = IPFORWSRCRT;
1.27 thorpej 197: int ip_directedbcast = IPDIRECTEDBCAST;
1.47 cjs 198: int ip_allowsrcrt = IPALLOWSRCRT;
1.53 kml 199: int ip_mtudisc = IPMTUDISC;
1.156 itojun 200: int ip_mtudisc_timeout = IPMTUDISCTIMEOUT;
1.1 cgd 201: #ifdef DIAGNOSTIC
202: int ipprintfs = 0;
203: #endif
1.184 jonathan 204:
205: int ip_do_randomid = 0;
206:
1.165 christos 207: /*
208: * XXX - Setting ip_checkinterface mostly implements the receive side of
209: * the Strong ES model described in RFC 1122, but since the routing table
210: * and transmit implementation do not implement the Strong ES model,
211: * setting this to 1 results in an odd hybrid.
212: *
213: * XXX - ip_checkinterface currently must be disabled if you use ipnat
214: * to translate the destination address to another local interface.
215: *
216: * XXX - ip_checkinterface must be disabled if you add IP aliases
217: * to the loopback interface instead of the interface where the
218: * packets for those addresses are received.
219: */
1.318 rmind 220: static int ip_checkinterface __read_mostly = 0;
1.165 christos 221:
1.60 kml 222: struct rttimer_queue *ip_mtudisc_timeout_q = NULL;
223:
1.318 rmind 224: pktqueue_t * ip_pktq __read_mostly;
225: pfil_head_t * inet_pfil_hook __read_mostly;
226: ipid_state_t * ip_ids __read_mostly;
227: percpu_t * ipstat_percpu __read_mostly;
228:
229: static struct route ipforward_rt __cacheline_aligned;
1.286 tls 230:
1.183 jonathan 231: uint16_t ip_id;
1.75 thorpej 232:
1.135 thorpej 233: #ifdef INET_CSUM_COUNTERS
234: #include <sys/device.h>
235:
236: struct evcnt ip_hwcsum_bad = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
237: NULL, "inet", "hwcsum bad");
238: struct evcnt ip_hwcsum_ok = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
239: NULL, "inet", "hwcsum ok");
240: struct evcnt ip_swcsum = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
241: NULL, "inet", "swcsum");
242:
243: #define INET_CSUM_COUNTER_INCR(ev) (ev)->ev_count++
244:
1.201 matt 245: EVCNT_ATTACH_STATIC(ip_hwcsum_bad);
246: EVCNT_ATTACH_STATIC(ip_hwcsum_ok);
247: EVCNT_ATTACH_STATIC(ip_swcsum);
248:
1.135 thorpej 249: #else
250:
251: #define INET_CSUM_COUNTER_INCR(ev) /* nothing */
252:
253: #endif /* INET_CSUM_COUNTERS */
254:
1.1 cgd 255: /*
1.337.2.5! pgoyette 256: * Used to save the IP options in case a protocol wants to respond
1.1 cgd 257: * to an incoming packet over the same route if the packet got here
258: * using IP source routing. This allows connection establishment and
259: * maintenance when the remote end is on a network that is not known
260: * to us.
261: */
1.337.2.5! pgoyette 262: struct ip_srcrt {
! 263: int isr_nhops; /* number of hops */
! 264: struct in_addr isr_dst; /* final destination */
! 265: char isr_nop; /* one NOP to align */
! 266: char isr_hdr[IPOPT_OFFSET + 1]; /* OPTVAL, OLEN & OFFSET */
! 267: struct in_addr isr_routes[MAX_IPOPTLEN/sizeof(struct in_addr)];
! 268: };
1.1 cgd 269:
1.295 dyoung 270: static int ip_drainwanted;
271:
1.337.2.5! pgoyette 272: static void save_rte(struct mbuf *, u_char *, struct in_addr);
1.35 mycroft 273:
1.164 matt 274: #ifdef MBUFTRACE
1.234 dogcow 275: struct mowner ip_rx_mowner = MOWNER_INIT("internet", "rx");
276: struct mowner ip_tx_mowner = MOWNER_INIT("internet", "tx");
1.164 matt 277: #endif
278:
1.318 rmind 279: static void ipintr(void *);
1.312 rmind 280: static void ip_input(struct mbuf *);
1.329 ozaki-r 281: static void ip_forward(struct mbuf *, int, struct ifnet *);
1.311 rmind 282: static bool ip_dooptions(struct mbuf *);
1.337.2.3 pgoyette 283: static struct in_ifaddr *ip_rtaddr(struct in_addr, struct psref *);
1.311 rmind 284: static void sysctl_net_inet_ip_setup(struct sysctllog **);
1.284 pooka 285:
1.333 ozaki-r 286: static struct in_ifaddr *ip_match_our_address(struct ifnet *, struct ip *,
287: int *);
288: static struct in_ifaddr *ip_match_our_address_broadcast(struct ifnet *,
289: struct ip *);
290:
1.337.2.2 pgoyette 291: #ifdef NET_MPSAFE
292: #define SOFTNET_LOCK() mutex_enter(softnet_lock)
293: #define SOFTNET_UNLOCK() mutex_exit(softnet_lock)
294: #else
1.318 rmind 295: #define SOFTNET_LOCK() KASSERT(mutex_owned(softnet_lock))
296: #define SOFTNET_UNLOCK() KASSERT(mutex_owned(softnet_lock))
1.337.2.2 pgoyette 297: #endif
1.318 rmind 298:
1.1 cgd 299: /*
300: * IP initialization: fill in IP protocol switch table.
301: * All protocols not implemented in kernel go to raw IP protocol handler.
302: */
1.8 mycroft 303: void
1.211 perry 304: ip_init(void)
1.1 cgd 305: {
1.199 matt 306: const struct protosw *pr;
1.1 cgd 307:
1.311 rmind 308: in_init();
1.284 pooka 309: sysctl_net_inet_ip_setup(NULL);
310:
1.311 rmind 311: pr = pffindproto(PF_INET, IPPROTO_RAW, SOCK_RAW);
312: KASSERT(pr != NULL);
1.275 pooka 313:
1.319 ozaki-r 314: ip_pktq = pktq_create(IFQ_MAXLEN, ipintr, NULL);
1.318 rmind 315: KASSERT(ip_pktq != NULL);
316:
1.328 riastrad 317: for (u_int i = 0; i < IPPROTO_MAX; i++) {
318: ip_protox[i] = pr - inetsw;
1.311 rmind 319: }
1.328 riastrad 320: for (pr = inetdomain.dom_protosw;
321: pr < inetdomain.dom_protoswNPROTOSW; pr++)
1.1 cgd 322: if (pr->pr_domain->dom_family == PF_INET &&
323: pr->pr_protocol && pr->pr_protocol != IPPROTO_RAW)
1.328 riastrad 324: ip_protox[pr->pr_protocol] = pr - inetsw;
1.192 jonathan 325:
1.288 rmind 326: ip_reass_init();
1.190 jonathan 327:
1.291 rmind 328: ip_ids = ip_id_init();
1.323 ozaki-r 329: ip_id = time_uptime & 0xfffff;
1.194 jonathan 330:
1.73 thorpej 331: #ifdef GATEWAY
1.311 rmind 332: ipflow_init();
1.73 thorpej 333: #endif
1.121 thorpej 334:
335: /* Register our Packet Filter hook. */
1.308 rmind 336: inet_pfil_hook = pfil_head_create(PFIL_TYPE_AF, (void *)AF_INET);
337: KASSERT(inet_pfil_hook != NULL);
1.135 thorpej 338:
1.164 matt 339: #ifdef MBUFTRACE
340: MOWNER_ATTACH(&ip_tx_mowner);
341: MOWNER_ATTACH(&ip_rx_mowner);
342: #endif /* MBUFTRACE */
1.266 thorpej 343:
344: ipstat_percpu = percpu_alloc(sizeof(uint64_t) * IP_NSTATS);
1.337.2.4 pgoyette 345:
346: ipforward_rt_percpu = percpu_alloc(sizeof(struct route));
347: if (ipforward_rt_percpu == NULL)
348: panic("failed to allocate ipforward_rt_percpu");
349:
350: ip_mtudisc_timeout_q = rt_timer_queue_create(ip_mtudisc_timeout);
1.1 cgd 351: }
352:
1.333 ozaki-r 353: static struct in_ifaddr *
354: ip_match_our_address(struct ifnet *ifp, struct ip *ip, int *downmatch)
355: {
356: struct in_ifaddr *ia = NULL;
357: int checkif;
358:
359: /*
360: * Enable a consistency check between the destination address
361: * and the arrival interface for a unicast packet (the RFC 1122
362: * strong ES model) if IP forwarding is disabled and the packet
363: * is not locally generated.
364: *
365: * XXX - Checking also should be disabled if the destination
366: * address is ipnat'ed to a different interface.
367: *
368: * XXX - Checking is incompatible with IP aliases added
369: * to the loopback interface instead of the interface where
370: * the packets are received.
371: *
372: * XXX - We need to add a per ifaddr flag for this so that
373: * we get finer grain control.
374: */
375: checkif = ip_checkinterface && (ipforwarding == 0) &&
376: (ifp->if_flags & IFF_LOOPBACK) == 0;
377:
1.334 ozaki-r 378: IN_ADDRHASH_READER_FOREACH(ia, ip->ip_dst.s_addr) {
1.333 ozaki-r 379: if (in_hosteq(ia->ia_addr.sin_addr, ip->ip_dst)) {
380: if (ia->ia4_flags & IN_IFF_NOTREADY)
381: continue;
382: if (checkif && ia->ia_ifp != ifp)
383: continue;
384: if ((ia->ia_ifp->if_flags & IFF_UP) != 0)
385: break;
386: else
1.337.2.1 pgoyette 387: (*downmatch)++;
1.333 ozaki-r 388: }
389: }
390:
391: return ia;
392: }
393:
394: static struct in_ifaddr *
395: ip_match_our_address_broadcast(struct ifnet *ifp, struct ip *ip)
396: {
397: struct in_ifaddr *ia = NULL;
398: struct ifaddr *ifa;
399:
1.336 ozaki-r 400: IFADDR_READER_FOREACH(ifa, ifp) {
1.333 ozaki-r 401: if (ifa->ifa_addr->sa_family != AF_INET)
402: continue;
403: ia = ifatoia(ifa);
404: if (ia->ia4_flags & IN_IFF_NOTREADY)
405: continue;
406: if (in_hosteq(ip->ip_dst, ia->ia_broadaddr.sin_addr) ||
407: in_hosteq(ip->ip_dst, ia->ia_netbroadcast) ||
408: /*
409: * Look for all-0's host part (old broadcast addr),
410: * either for subnet or net.
411: */
412: ip->ip_dst.s_addr == ia->ia_subnet ||
413: ip->ip_dst.s_addr == ia->ia_net)
414: goto matched;
415: /*
416: * An interface with IP address zero accepts
417: * all packets that arrive on that interface.
418: */
419: if (in_nullhost(ia->ia_addr.sin_addr))
420: goto matched;
421: }
422: ia = NULL;
423:
424: matched:
425: return ia;
426: }
427:
1.1 cgd 428: /*
1.318 rmind 429: * IP software interrupt routine.
1.89 itojun 430: */
1.318 rmind 431: static void
432: ipintr(void *arg __unused)
1.89 itojun 433: {
434: struct mbuf *m;
1.286 tls 435:
1.318 rmind 436: KASSERT(cpu_softintr_p());
1.89 itojun 437:
1.337.2.2 pgoyette 438: #ifndef NET_MPSAFE
1.268 ad 439: mutex_enter(softnet_lock);
1.337.2.2 pgoyette 440: #endif
1.318 rmind 441: while ((m = pktq_dequeue(ip_pktq)) != NULL) {
1.89 itojun 442: ip_input(m);
443: }
1.337.2.2 pgoyette 444: #ifndef NET_MPSAFE
1.268 ad 445: mutex_exit(softnet_lock);
1.337.2.2 pgoyette 446: #endif
1.89 itojun 447: }
448:
449: /*
1.318 rmind 450: * IP input routine. Checksum and byte swap header. If fragmented
1.1 cgd 451: * try to reassemble. Process options. Pass to next level.
452: */
1.312 rmind 453: static void
1.89 itojun 454: ip_input(struct mbuf *m)
1.1 cgd 455: {
1.109 augustss 456: struct ip *ip = NULL;
1.337.2.1 pgoyette 457: struct in_ifaddr *ia = NULL;
1.288 rmind 458: int hlen = 0, len;
1.100 itojun 459: int downmatch;
1.169 itojun 460: int srcrt = 0;
1.314 rmind 461: ifnet_t *ifp;
1.329 ozaki-r 462: struct psref psref;
1.337.2.1 pgoyette 463: int s;
1.1 cgd 464:
1.318 rmind 465: KASSERTMSG(cpu_softintr_p(), "ip_input: not in the software "
466: "interrupt handler; synchronization assumptions violated");
467:
1.164 matt 468: MCLAIM(m, &ip_rx_mowner);
1.289 rmind 469: KASSERT((m->m_flags & M_PKTHDR) != 0);
1.329 ozaki-r 470:
471: ifp = m_get_rcvif_psref(m, &psref);
1.330 ozaki-r 472: if (__predict_false(ifp == NULL))
1.332 ozaki-r 473: goto out;
1.164 matt 474:
1.1 cgd 475: /*
476: * If no IP addresses have been set yet but the interfaces
477: * are receiving, can't do anything with incoming packets yet.
1.318 rmind 478: * Note: we pre-check without locks held.
1.1 cgd 479: */
1.335 ozaki-r 480: if (IN_ADDRLIST_READER_EMPTY())
1.332 ozaki-r 481: goto out;
1.266 thorpej 482: IP_STATINC(IP_STAT_TOTAL);
1.318 rmind 483:
1.154 thorpej 484: /*
485: * If the IP header is not aligned, slurp it up into a new
486: * mbuf with space for link headers, in the event we forward
487: * it. Otherwise, if it is aligned, make sure the entire
488: * base IP header is in the first mbuf of the chain.
489: */
1.244 christos 490: if (IP_HDR_ALIGNED_P(mtod(m, void *)) == 0) {
1.154 thorpej 491: if ((m = m_copyup(m, sizeof(struct ip),
492: (max_linkhdr + 3) & ~3)) == NULL) {
493: /* XXXJRT new stat, please */
1.266 thorpej 494: IP_STATINC(IP_STAT_TOOSMALL);
1.329 ozaki-r 495: goto out;
1.154 thorpej 496: }
497: } else if (__predict_false(m->m_len < sizeof (struct ip))) {
498: if ((m = m_pullup(m, sizeof (struct ip))) == NULL) {
1.266 thorpej 499: IP_STATINC(IP_STAT_TOOSMALL);
1.329 ozaki-r 500: goto out;
1.154 thorpej 501: }
1.1 cgd 502: }
503: ip = mtod(m, struct ip *);
1.13 mycroft 504: if (ip->ip_v != IPVERSION) {
1.266 thorpej 505: IP_STATINC(IP_STAT_BADVERS);
1.332 ozaki-r 506: goto out;
1.13 mycroft 507: }
1.1 cgd 508: hlen = ip->ip_hl << 2;
509: if (hlen < sizeof(struct ip)) { /* minimum header length */
1.266 thorpej 510: IP_STATINC(IP_STAT_BADHLEN);
1.332 ozaki-r 511: goto out;
1.1 cgd 512: }
513: if (hlen > m->m_len) {
1.298 liamjfoy 514: if ((m = m_pullup(m, hlen)) == NULL) {
1.266 thorpej 515: IP_STATINC(IP_STAT_BADHLEN);
1.329 ozaki-r 516: goto out;
1.1 cgd 517: }
518: ip = mtod(m, struct ip *);
519: }
1.98 thorpej 520:
1.85 hwr 521: /*
1.99 thorpej 522: * RFC1122: packets with a multicast source address are
1.98 thorpej 523: * not allowed.
1.85 hwr 524: */
525: if (IN_MULTICAST(ip->ip_src.s_addr)) {
1.266 thorpej 526: IP_STATINC(IP_STAT_BADADDR);
1.332 ozaki-r 527: goto out;
1.129 itojun 528: }
529:
530: /* 127/8 must not appear on wire - RFC1122 */
531: if ((ntohl(ip->ip_dst.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET ||
532: (ntohl(ip->ip_src.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET) {
1.314 rmind 533: if ((ifp->if_flags & IFF_LOOPBACK) == 0) {
1.266 thorpej 534: IP_STATINC(IP_STAT_BADADDR);
1.332 ozaki-r 535: goto out;
1.130 itojun 536: }
1.85 hwr 537: }
538:
1.135 thorpej 539: switch (m->m_pkthdr.csum_flags &
1.314 rmind 540: ((ifp->if_csum_flags_rx & M_CSUM_IPv4) |
1.135 thorpej 541: M_CSUM_IPv4_BAD)) {
542: case M_CSUM_IPv4|M_CSUM_IPv4_BAD:
543: INET_CSUM_COUNTER_INCR(&ip_hwcsum_bad);
1.332 ozaki-r 544: IP_STATINC(IP_STAT_BADSUM);
545: goto out;
1.135 thorpej 546:
547: case M_CSUM_IPv4:
548: /* Checksum was okay. */
549: INET_CSUM_COUNTER_INCR(&ip_hwcsum_ok);
550: break;
551:
552: default:
1.206 thorpej 553: /*
554: * Must compute it ourselves. Maybe skip checksum on
555: * loopback interfaces.
556: */
1.314 rmind 557: if (__predict_true(!(ifp->if_flags & IFF_LOOPBACK) ||
558: ip_do_loopback_cksum)) {
1.206 thorpej 559: INET_CSUM_COUNTER_INCR(&ip_swcsum);
1.332 ozaki-r 560: if (in_cksum(m, hlen) != 0) {
561: IP_STATINC(IP_STAT_BADSUM);
562: goto out;
563: }
1.206 thorpej 564: }
1.135 thorpej 565: break;
1.1 cgd 566: }
567:
1.121 thorpej 568: /* Retrieve the packet length. */
569: len = ntohs(ip->ip_len);
1.81 proff 570:
571: /*
572: * Check for additional length bogosity
573: */
1.84 proff 574: if (len < hlen) {
1.266 thorpej 575: IP_STATINC(IP_STAT_BADLEN);
1.332 ozaki-r 576: goto out;
1.81 proff 577: }
1.1 cgd 578:
579: /*
580: * Check that the amount of data in the buffers
581: * is as at least much as the IP header would have us expect.
582: * Trim mbufs if longer than we expect.
583: * Drop packet if shorter than we expect.
584: */
1.35 mycroft 585: if (m->m_pkthdr.len < len) {
1.266 thorpej 586: IP_STATINC(IP_STAT_TOOSHORT);
1.332 ozaki-r 587: goto out;
1.1 cgd 588: }
1.35 mycroft 589: if (m->m_pkthdr.len > len) {
1.1 cgd 590: if (m->m_len == m->m_pkthdr.len) {
1.35 mycroft 591: m->m_len = len;
592: m->m_pkthdr.len = len;
1.1 cgd 593: } else
1.35 mycroft 594: m_adj(m, len - m->m_pkthdr.len);
1.1 cgd 595: }
596:
1.64 thorpej 597: /*
598: * Assume that we can create a fast-forward IP flow entry
599: * based on this packet.
600: */
601: m->m_flags |= M_CANFASTFWD;
602:
1.33 mrg 603: /*
1.64 thorpej 604: * Run through list of hooks for input packets. If there are any
605: * filters which require that additional packets in the flow are
606: * not fast-forwarded, they must clear the M_CANFASTFWD flag.
607: * Note that filters must _never_ set this flag, as another filter
608: * in the list may have previously cleared it.
1.33 mrg 609: */
1.304 christos 610: #if defined(IPSEC)
1.317 christos 611: if (!ipsec_used || !ipsec_indone(m))
1.127 itojun 612: #else
613: if (1)
614: #endif
615: {
1.314 rmind 616: struct in_addr odst = ip->ip_dst;
1.318 rmind 617: bool freed;
1.169 itojun 618:
1.318 rmind 619: freed = pfil_run_hooks(inet_pfil_hook, &m, ifp, PFIL_IN) != 0;
620: if (freed || m == NULL) {
1.337.2.2 pgoyette 621: m = NULL;
1.329 ozaki-r 622: goto out;
1.318 rmind 623: }
1.127 itojun 624: ip = mtod(m, struct ip *);
1.142 darrenr 625: hlen = ip->ip_hl << 2;
1.318 rmind 626:
1.205 darrenr 627: /*
628: * XXX The setting of "srcrt" here is to prevent ip_forward()
629: * from generating ICMP redirects for packets that have
630: * been redirected by a hook back out on to the same LAN that
631: * they came from and is not an indication that the packet
632: * is being inffluenced by source routing options. This
633: * allows things like
634: * "rdr tlp0 0/0 port 80 -> 1.1.1.200 3128 tcp"
635: * where tlp0 is both on the 1.1.1.0/24 network and is the
636: * default route for hosts on 1.1.1.0/24. Of course this
637: * also requires a "map tlp0 ..." to complete the story.
638: * One might argue whether or not this kind of network config.
1.212 perry 639: * should be supported in this manner...
1.205 darrenr 640: */
1.169 itojun 641: srcrt = (odst.s_addr != ip->ip_dst.s_addr);
1.127 itojun 642: }
1.123 thorpej 643:
644: #ifdef ALTQ
645: /* XXX Temporary until ALTQ is changed to use a pfil hook */
1.318 rmind 646: if (altq_input) {
647: SOFTNET_LOCK();
648: if ((*altq_input)(m, AF_INET) == 0) {
649: /* Packet dropped by traffic conditioner. */
650: SOFTNET_UNLOCK();
1.337.2.2 pgoyette 651: m = NULL;
1.329 ozaki-r 652: goto out;
1.318 rmind 653: }
654: SOFTNET_UNLOCK();
1.123 thorpej 655: }
656: #endif
1.121 thorpej 657:
658: /*
1.1 cgd 659: * Process options and, if not destined for us,
660: * ship it on. ip_dooptions returns 1 when an
661: * error was detected (causing an icmp message
662: * to be sent and the original packet to be freed).
663: */
1.337.2.2 pgoyette 664: if (hlen > sizeof (struct ip) && ip_dooptions(m)) {
665: m = NULL;
1.329 ozaki-r 666: goto out;
1.337.2.2 pgoyette 667: }
1.1 cgd 668:
669: /*
670: * Check our list of addresses, to see if the packet is for us.
1.100 itojun 671: *
672: * Traditional 4.4BSD did not consult IFF_UP at all.
673: * The behavior here is to treat addresses on !IFF_UP interface
1.321 roy 674: * or IN_IFF_NOTREADY addresses as not mine.
1.1 cgd 675: */
1.100 itojun 676: downmatch = 0;
1.337.2.1 pgoyette 677: s = pserialize_read_enter();
1.333 ozaki-r 678: ia = ip_match_our_address(ifp, ip, &downmatch);
1.337.2.1 pgoyette 679: if (ia != NULL) {
680: pserialize_read_exit(s);
1.86 thorpej 681: goto ours;
1.337.2.1 pgoyette 682: }
1.333 ozaki-r 683:
1.315 christos 684: if (ifp->if_flags & IFF_BROADCAST) {
1.333 ozaki-r 685: ia = ip_match_our_address_broadcast(ifp, ip);
1.337.2.1 pgoyette 686: if (ia != NULL) {
687: pserialize_read_exit(s);
1.333 ozaki-r 688: goto ours;
1.337.2.1 pgoyette 689: }
1.1 cgd 690: }
1.337.2.1 pgoyette 691: pserialize_read_exit(s);
1.333 ozaki-r 692:
1.18 mycroft 693: if (IN_MULTICAST(ip->ip_dst.s_addr)) {
1.4 hpeyerl 694: #ifdef MROUTING
695: extern struct socket *ip_mrouter;
1.10 brezak 696:
1.4 hpeyerl 697: if (ip_mrouter) {
698: /*
699: * If we are acting as a multicast router, all
700: * incoming multicast packets are passed to the
701: * kernel-level multicast forwarding function.
702: * The packet is returned (relatively) intact; if
703: * ip_mforward() returns a non-zero value, the packet
704: * must be discarded, else it may be accepted below.
705: *
706: * (The IP ident field is put in the same byte order
707: * as expected when ip_mforward() is called from
708: * ip_output().)
709: */
1.318 rmind 710: SOFTNET_LOCK();
1.314 rmind 711: if (ip_mforward(m, ifp) != 0) {
1.318 rmind 712: SOFTNET_UNLOCK();
1.266 thorpej 713: IP_STATINC(IP_STAT_CANTFORWARD);
1.332 ozaki-r 714: goto out;
1.4 hpeyerl 715: }
1.318 rmind 716: SOFTNET_UNLOCK();
1.4 hpeyerl 717:
718: /*
719: * The process-level routing demon needs to receive
720: * all multicast IGMP packets, whether or not this
721: * host belongs to their destination groups.
722: */
1.318 rmind 723: if (ip->ip_p == IPPROTO_IGMP) {
1.4 hpeyerl 724: goto ours;
1.318 rmind 725: }
1.266 thorpej 726: IP_STATINC(IP_STAT_CANTFORWARD);
1.4 hpeyerl 727: }
728: #endif
729: /*
730: * See if we belong to the destination multicast group on the
731: * arrival interface.
732: */
1.316 rmind 733: if (!in_multi_group(ip->ip_dst, ifp, 0)) {
1.266 thorpej 734: IP_STATINC(IP_STAT_CANTFORWARD);
1.332 ozaki-r 735: goto out;
1.4 hpeyerl 736: }
737: goto ours;
738: }
1.19 mycroft 739: if (ip->ip_dst.s_addr == INADDR_BROADCAST ||
1.35 mycroft 740: in_nullhost(ip->ip_dst))
1.1 cgd 741: goto ours;
742:
743: /*
744: * Not for us; forward if possible and desirable.
745: */
746: if (ipforwarding == 0) {
1.329 ozaki-r 747: m_put_rcvif_psref(ifp, &psref);
1.266 thorpej 748: IP_STATINC(IP_STAT_CANTFORWARD);
1.1 cgd 749: m_freem(m);
1.100 itojun 750: } else {
751: /*
752: * If ip_dst matched any of my address on !IFF_UP interface,
753: * and there's no IFF_UP interface that matches ip_dst,
754: * send icmp unreach. Forwarding it will result in in-kernel
755: * forwarding loop till TTL goes to 0.
756: */
757: if (downmatch) {
1.329 ozaki-r 758: m_put_rcvif_psref(ifp, &psref);
1.100 itojun 759: icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_HOST, 0, 0);
1.266 thorpej 760: IP_STATINC(IP_STAT_CANTFORWARD);
1.100 itojun 761: return;
762: }
1.304 christos 763: #ifdef IPSEC
1.305 rmind 764: /* Perform IPsec, if any. */
1.318 rmind 765: if (ipsec_used) {
766: SOFTNET_LOCK();
767: if (ipsec4_input(m, IP_FORWARDING |
768: (ip_directedbcast ? IP_ALLOWBROADCAST : 0)) != 0) {
769: SOFTNET_UNLOCK();
1.332 ozaki-r 770: goto out;
1.318 rmind 771: }
772: SOFTNET_UNLOCK();
1.173 jonathan 773: }
1.305 rmind 774: #endif
1.329 ozaki-r 775: ip_forward(m, srcrt, ifp);
776: m_put_rcvif_psref(ifp, &psref);
1.100 itojun 777: }
1.89 itojun 778: return;
1.1 cgd 779:
780: ours:
1.329 ozaki-r 781: m_put_rcvif_psref(ifp, &psref);
1.331 ozaki-r 782: ifp = NULL;
1.329 ozaki-r 783:
1.1 cgd 784: /*
785: * If offset or IP_MF are set, must reassemble.
786: */
1.155 itojun 787: if (ip->ip_off & ~htons(IP_DF|IP_RF)) {
1.289 rmind 788: /*
789: * Pass to IP reassembly mechanism.
790: */
1.290 rmind 791: if (ip_reass_packet(&m, ip) != 0) {
1.289 rmind 792: /* Failed; invalid fragment(s) or packet. */
1.332 ozaki-r 793: goto out;
1.16 cgd 794: }
1.290 rmind 795: if (m == NULL) {
1.289 rmind 796: /* More fragments should come; silently return. */
1.332 ozaki-r 797: goto out;
1.289 rmind 798: }
1.290 rmind 799: /*
800: * Reassembly is done, we have the final packet.
801: * Updated cached data in local variable(s).
802: */
1.289 rmind 803: ip = mtod(m, struct ip *);
804: hlen = ip->ip_hl << 2;
1.79 mycroft 805: }
1.128 itojun 806:
1.304 christos 807: #ifdef IPSEC
1.173 jonathan 808: /*
1.305 rmind 809: * Enforce IPsec policy checking if we are seeing last header.
810: * Note that we do not visit this with protocols with PCB layer
811: * code - like UDP/TCP/raw IP.
1.173 jonathan 812: */
1.317 christos 813: if (ipsec_used &&
1.328 riastrad 814: (inetsw[ip_protox[ip->ip_p]].pr_flags & PR_LASTHDR) != 0) {
1.318 rmind 815: SOFTNET_LOCK();
1.305 rmind 816: if (ipsec4_input(m, 0) != 0) {
1.318 rmind 817: SOFTNET_UNLOCK();
1.332 ozaki-r 818: goto out;
1.173 jonathan 819: }
1.318 rmind 820: SOFTNET_UNLOCK();
1.173 jonathan 821: }
1.305 rmind 822: #endif
1.1 cgd 823:
824: /*
825: * Switch out to protocol's input routine.
826: */
1.82 aidan 827: #if IFA_STATS
1.337.2.1 pgoyette 828: if (ia && ip) {
829: struct in_ifaddr *_ia;
830: /*
831: * Keep a reference from ip_match_our_address with psref
832: * is expensive, so explore ia here again.
833: */
834: s = pserialize_read_enter();
1.337.2.2 pgoyette 835: _ia = in_get_ia(ip->ip_dst);
1.337.2.1 pgoyette 836: _ia->ia_ifa.ifa_data.ifad_inbytes += ntohs(ip->ip_len);
837: pserialize_read_exit(s);
838: }
1.82 aidan 839: #endif
1.266 thorpej 840: IP_STATINC(IP_STAT_DELIVERED);
1.89 itojun 841:
1.318 rmind 842: const int off = hlen, nh = ip->ip_p;
843:
844: SOFTNET_LOCK();
1.328 riastrad 845: (*inetsw[ip_protox[nh]].pr_input)(m, off, nh);
1.318 rmind 846: SOFTNET_UNLOCK();
1.89 itojun 847: return;
1.135 thorpej 848:
1.329 ozaki-r 849: out:
850: m_put_rcvif_psref(ifp, &psref);
1.332 ozaki-r 851: if (m != NULL)
852: m_freem(m);
1.1 cgd 853: }
854:
855: /*
1.288 rmind 856: * IP timer processing.
1.1 cgd 857: */
1.8 mycroft 858: void
1.211 perry 859: ip_slowtimo(void)
1.1 cgd 860: {
1.268 ad 861:
1.337.2.4 pgoyette 862: #ifndef NET_MPSAFE
1.268 ad 863: mutex_enter(softnet_lock);
864: KERNEL_LOCK(1, NULL);
1.337.2.4 pgoyette 865: #endif
1.1 cgd 866:
1.288 rmind 867: ip_reass_slowtimo();
1.268 ad 868:
1.337.2.4 pgoyette 869: #ifndef NET_MPSAFE
1.268 ad 870: KERNEL_UNLOCK_ONE(NULL);
871: mutex_exit(softnet_lock);
1.337.2.4 pgoyette 872: #endif
1.1 cgd 873: }
874:
875: /*
1.288 rmind 876: * IP drain processing.
1.1 cgd 877: */
1.8 mycroft 878: void
1.211 perry 879: ip_drain(void)
1.1 cgd 880: {
881:
1.268 ad 882: KERNEL_LOCK(1, NULL);
1.288 rmind 883: ip_reass_drain();
1.268 ad 884: KERNEL_UNLOCK_ONE(NULL);
1.1 cgd 885: }
886:
887: /*
1.311 rmind 888: * ip_dooptions: perform option processing on a datagram, possibly discarding
889: * it if bad options are encountered, or forwarding it if source-routed.
890: *
891: * => Returns true if packet has been forwarded/freed.
892: * => Returns false if the packet should be processed further.
1.1 cgd 893: */
1.311 rmind 894: static bool
1.211 perry 895: ip_dooptions(struct mbuf *m)
1.1 cgd 896: {
1.109 augustss 897: struct ip *ip = mtod(m, struct ip *);
898: u_char *cp, *cp0;
899: struct ip_timestamp *ipt;
900: struct in_ifaddr *ia;
1.1 cgd 901: int opt, optlen, cnt, off, code, type = ICMP_PARAMPROB, forward = 0;
1.104 thorpej 902: struct in_addr dst;
1.1 cgd 903: n_time ntime;
1.337.2.4 pgoyette 904: struct ifaddr *ifa = NULL;
1.337.2.1 pgoyette 905: int s;
1.1 cgd 906:
1.13 mycroft 907: dst = ip->ip_dst;
1.1 cgd 908: cp = (u_char *)(ip + 1);
909: cnt = (ip->ip_hl << 2) - sizeof (struct ip);
910: for (; cnt > 0; cnt -= optlen, cp += optlen) {
911: opt = cp[IPOPT_OPTVAL];
912: if (opt == IPOPT_EOL)
913: break;
914: if (opt == IPOPT_NOP)
915: optlen = 1;
916: else {
1.113 itojun 917: if (cnt < IPOPT_OLEN + sizeof(*cp)) {
918: code = &cp[IPOPT_OLEN] - (u_char *)ip;
919: goto bad;
920: }
1.1 cgd 921: optlen = cp[IPOPT_OLEN];
1.114 itojun 922: if (optlen < IPOPT_OLEN + sizeof(*cp) || optlen > cnt) {
1.1 cgd 923: code = &cp[IPOPT_OLEN] - (u_char *)ip;
924: goto bad;
925: }
926: }
927: switch (opt) {
928:
929: default:
930: break;
931:
932: /*
933: * Source routing with record.
934: * Find interface with current destination address.
935: * If none on this machine then drop if strictly routed,
936: * or do nothing if loosely routed.
937: * Record interface address and bring up next address
938: * component. If strictly routed make sure next
939: * address is on directly accessible net.
940: */
941: case IPOPT_LSRR:
1.337.2.3 pgoyette 942: case IPOPT_SSRR: {
943: struct psref psref;
1.337.2.5! pgoyette 944: struct sockaddr_in ipaddr = {
! 945: .sin_len = sizeof(ipaddr),
! 946: .sin_family = AF_INET,
! 947: };
! 948:
1.47 cjs 949: if (ip_allowsrcrt == 0) {
950: type = ICMP_UNREACH;
951: code = ICMP_UNREACH_NET_PROHIB;
952: goto bad;
953: }
1.114 itojun 954: if (optlen < IPOPT_OFFSET + sizeof(*cp)) {
955: code = &cp[IPOPT_OLEN] - (u_char *)ip;
956: goto bad;
957: }
1.1 cgd 958: if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) {
959: code = &cp[IPOPT_OFFSET] - (u_char *)ip;
960: goto bad;
961: }
962: ipaddr.sin_addr = ip->ip_dst;
1.337.2.1 pgoyette 963:
964: s = pserialize_read_enter();
965: ifa = ifa_ifwithaddr(sintosa(&ipaddr));
966: if (ifa == NULL) {
967: pserialize_read_exit(s);
1.1 cgd 968: if (opt == IPOPT_SSRR) {
969: type = ICMP_UNREACH;
970: code = ICMP_UNREACH_SRCFAIL;
971: goto bad;
972: }
973: /*
974: * Loose routing, and not at next destination
975: * yet; nothing to do except forward.
976: */
977: break;
978: }
1.337.2.1 pgoyette 979: pserialize_read_exit(s);
980:
1.1 cgd 981: off--; /* 0 origin */
1.112 sommerfe 982: if ((off + sizeof(struct in_addr)) > optlen) {
1.1 cgd 983: /*
984: * End of source route. Should be for us.
985: */
1.337.2.5! pgoyette 986: save_rte(m, cp, ip->ip_src);
1.1 cgd 987: break;
988: }
989: /*
990: * locate outgoing interface
991: */
1.281 tsutsui 992: memcpy((void *)&ipaddr.sin_addr, (void *)(cp + off),
1.1 cgd 993: sizeof(ipaddr.sin_addr));
1.337.2.1 pgoyette 994: if (opt == IPOPT_SSRR) {
1.337.2.3 pgoyette 995: ifa = ifa_ifwithladdr_psref(sintosa(&ipaddr),
996: &psref);
1.337.2.1 pgoyette 997: if (ifa != NULL)
998: ia = ifatoia(ifa);
999: else
1000: ia = NULL;
1001: } else {
1.337.2.3 pgoyette 1002: ia = ip_rtaddr(ipaddr.sin_addr, &psref);
1.337.2.1 pgoyette 1003: }
1004: if (ia == NULL) {
1.1 cgd 1005: type = ICMP_UNREACH;
1006: code = ICMP_UNREACH_SRCFAIL;
1007: goto bad;
1008: }
1009: ip->ip_dst = ipaddr.sin_addr;
1.244 christos 1010: bcopy((void *)&ia->ia_addr.sin_addr,
1011: (void *)(cp + off), sizeof(struct in_addr));
1.337.2.3 pgoyette 1012: ia4_release(ia, &psref);
1.1 cgd 1013: cp[IPOPT_OFFSET] += sizeof(struct in_addr);
1.13 mycroft 1014: /*
1015: * Let ip_intr's mcast routing check handle mcast pkts
1016: */
1.18 mycroft 1017: forward = !IN_MULTICAST(ip->ip_dst.s_addr);
1.1 cgd 1018: break;
1.337.2.3 pgoyette 1019: }
1.1 cgd 1020:
1.337.2.3 pgoyette 1021: case IPOPT_RR: {
1022: struct psref psref;
1.337.2.5! pgoyette 1023: struct sockaddr_in ipaddr = {
! 1024: .sin_len = sizeof(ipaddr),
! 1025: .sin_family = AF_INET,
! 1026: };
! 1027:
1.114 itojun 1028: if (optlen < IPOPT_OFFSET + sizeof(*cp)) {
1029: code = &cp[IPOPT_OLEN] - (u_char *)ip;
1030: goto bad;
1031: }
1.1 cgd 1032: if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) {
1033: code = &cp[IPOPT_OFFSET] - (u_char *)ip;
1034: goto bad;
1035: }
1036: /*
1037: * If no space remains, ignore.
1038: */
1039: off--; /* 0 origin */
1.112 sommerfe 1040: if ((off + sizeof(struct in_addr)) > optlen)
1.1 cgd 1041: break;
1.281 tsutsui 1042: memcpy((void *)&ipaddr.sin_addr, (void *)(&ip->ip_dst),
1.1 cgd 1043: sizeof(ipaddr.sin_addr));
1044: /*
1045: * locate outgoing interface; if we're the destination,
1046: * use the incoming interface (should be same).
1047: */
1.337.2.3 pgoyette 1048: ifa = ifa_ifwithaddr_psref(sintosa(&ipaddr), &psref);
1.337.2.1 pgoyette 1049: if (ifa == NULL) {
1.337.2.3 pgoyette 1050: ia = ip_rtaddr(ipaddr.sin_addr, &psref);
1.337.2.1 pgoyette 1051: if (ia == NULL) {
1052: type = ICMP_UNREACH;
1053: code = ICMP_UNREACH_HOST;
1054: goto bad;
1055: }
1056: } else {
1057: ia = ifatoia(ifa);
1.1 cgd 1058: }
1.244 christos 1059: bcopy((void *)&ia->ia_addr.sin_addr,
1060: (void *)(cp + off), sizeof(struct in_addr));
1.337.2.3 pgoyette 1061: ia4_release(ia, &psref);
1.1 cgd 1062: cp[IPOPT_OFFSET] += sizeof(struct in_addr);
1063: break;
1.337.2.3 pgoyette 1064: }
1.1 cgd 1065:
1066: case IPOPT_TS:
1067: code = cp - (u_char *)ip;
1068: ipt = (struct ip_timestamp *)cp;
1.114 itojun 1069: if (ipt->ipt_len < 4 || ipt->ipt_len > 40) {
1070: code = (u_char *)&ipt->ipt_len - (u_char *)ip;
1.1 cgd 1071: goto bad;
1.114 itojun 1072: }
1073: if (ipt->ipt_ptr < 5) {
1074: code = (u_char *)&ipt->ipt_ptr - (u_char *)ip;
1075: goto bad;
1076: }
1.15 cgd 1077: if (ipt->ipt_ptr > ipt->ipt_len - sizeof (int32_t)) {
1.114 itojun 1078: if (++ipt->ipt_oflw == 0) {
1079: code = (u_char *)&ipt->ipt_ptr -
1080: (u_char *)ip;
1.1 cgd 1081: goto bad;
1.114 itojun 1082: }
1.1 cgd 1083: break;
1084: }
1.104 thorpej 1085: cp0 = (cp + ipt->ipt_ptr - 1);
1.1 cgd 1086: switch (ipt->ipt_flg) {
1087:
1088: case IPOPT_TS_TSONLY:
1089: break;
1090:
1.329 ozaki-r 1091: case IPOPT_TS_TSANDADDR: {
1092: struct ifnet *rcvif;
1.337.2.1 pgoyette 1093: int _s, _ss;
1.337.2.5! pgoyette 1094: struct sockaddr_in ipaddr = {
! 1095: .sin_len = sizeof(ipaddr),
! 1096: .sin_family = AF_INET,
! 1097: };
1.329 ozaki-r 1098:
1.66 thorpej 1099: if (ipt->ipt_ptr - 1 + sizeof(n_time) +
1.114 itojun 1100: sizeof(struct in_addr) > ipt->ipt_len) {
1101: code = (u_char *)&ipt->ipt_ptr -
1102: (u_char *)ip;
1.1 cgd 1103: goto bad;
1.114 itojun 1104: }
1.13 mycroft 1105: ipaddr.sin_addr = dst;
1.337.2.1 pgoyette 1106: _ss = pserialize_read_enter();
1107: rcvif = m_get_rcvif(m, &_s);
1.337.2.4 pgoyette 1108: if (__predict_true(rcvif != NULL)) {
1109: ifa = ifaof_ifpforaddr(sintosa(&ipaddr),
1110: rcvif);
1111: }
1.337.2.1 pgoyette 1112: m_put_rcvif(rcvif, &_s);
1113: if (ifa == NULL) {
1114: pserialize_read_exit(_ss);
1115: break;
1116: }
1117: ia = ifatoia(ifa);
1.104 thorpej 1118: bcopy(&ia->ia_addr.sin_addr,
1119: cp0, sizeof(struct in_addr));
1.337.2.1 pgoyette 1120: pserialize_read_exit(_ss);
1.1 cgd 1121: ipt->ipt_ptr += sizeof(struct in_addr);
1122: break;
1.329 ozaki-r 1123: }
1.1 cgd 1124:
1.337.2.5! pgoyette 1125: case IPOPT_TS_PRESPEC: {
! 1126: struct sockaddr_in ipaddr = {
! 1127: .sin_len = sizeof(ipaddr),
! 1128: .sin_family = AF_INET,
! 1129: };
! 1130:
1.66 thorpej 1131: if (ipt->ipt_ptr - 1 + sizeof(n_time) +
1.114 itojun 1132: sizeof(struct in_addr) > ipt->ipt_len) {
1133: code = (u_char *)&ipt->ipt_ptr -
1134: (u_char *)ip;
1.1 cgd 1135: goto bad;
1.114 itojun 1136: }
1.281 tsutsui 1137: memcpy(&ipaddr.sin_addr, cp0,
1.1 cgd 1138: sizeof(struct in_addr));
1.337.2.1 pgoyette 1139: s = pserialize_read_enter();
1140: ifa = ifa_ifwithaddr(sintosa(&ipaddr));
1141: if (ifa == NULL) {
1142: pserialize_read_exit(s);
1.1 cgd 1143: continue;
1.337.2.1 pgoyette 1144: }
1145: pserialize_read_exit(s);
1.1 cgd 1146: ipt->ipt_ptr += sizeof(struct in_addr);
1147: break;
1.337.2.5! pgoyette 1148: }
1.1 cgd 1149:
1150: default:
1.114 itojun 1151: /* XXX can't take &ipt->ipt_flg */
1152: code = (u_char *)&ipt->ipt_ptr -
1153: (u_char *)ip + 1;
1.1 cgd 1154: goto bad;
1155: }
1156: ntime = iptime();
1.107 thorpej 1157: cp0 = (u_char *) &ntime; /* XXX grumble, GCC... */
1.244 christos 1158: memmove((char *)cp + ipt->ipt_ptr - 1, cp0,
1.1 cgd 1159: sizeof(n_time));
1160: ipt->ipt_ptr += sizeof(n_time);
1161: }
1162: }
1163: if (forward) {
1.329 ozaki-r 1164: struct ifnet *rcvif;
1.337.2.1 pgoyette 1165: struct psref _psref;
1.329 ozaki-r 1166:
1.26 thorpej 1167: if (ip_forwsrcrt == 0) {
1168: type = ICMP_UNREACH;
1169: code = ICMP_UNREACH_SRCFAIL;
1170: goto bad;
1171: }
1.329 ozaki-r 1172:
1.337.2.1 pgoyette 1173: rcvif = m_get_rcvif_psref(m, &_psref);
1.330 ozaki-r 1174: if (__predict_false(rcvif == NULL)) {
1175: type = ICMP_UNREACH;
1176: code = ICMP_UNREACH_HOST;
1177: goto bad;
1178: }
1.329 ozaki-r 1179: ip_forward(m, 1, rcvif);
1.337.2.1 pgoyette 1180: m_put_rcvif_psref(rcvif, &_psref);
1.311 rmind 1181: return true;
1.13 mycroft 1182: }
1.311 rmind 1183: return false;
1.1 cgd 1184: bad:
1.13 mycroft 1185: icmp_error(m, type, code, 0, 0);
1.266 thorpej 1186: IP_STATINC(IP_STAT_BADOPTIONS);
1.311 rmind 1187: return true;
1.1 cgd 1188: }
1189:
1190: /*
1.311 rmind 1191: * ip_rtaddr: given address of next destination (final or next hop),
1.1 cgd 1192: * return internet address info of interface to be used to get there.
1193: */
1.311 rmind 1194: static struct in_ifaddr *
1.337.2.3 pgoyette 1195: ip_rtaddr(struct in_addr dst, struct psref *psref)
1.1 cgd 1196: {
1.249 dyoung 1197: struct rtentry *rt;
1198: union {
1199: struct sockaddr dst;
1200: struct sockaddr_in dst4;
1201: } u;
1202:
1203: sockaddr_in_init(&u.dst4, &dst, 0);
1204:
1.337.2.3 pgoyette 1205: ro = percpu_getref(ipforward_rt_percpu);
1206: rt = rtcache_lookup(ro, &u.dst);
1207: if (rt == NULL) {
1208: percpu_putref(ipforward_rt_percpu);
1.249 dyoung 1209: return NULL;
1.337.2.3 pgoyette 1210: }
1211:
1212: ia4_acquire(ifatoia(rt->rt_ifa), psref);
1213: rtcache_unref(rt, ro);
1214: percpu_putref(ipforward_rt_percpu);
1.249 dyoung 1215:
1216: return ifatoia(rt->rt_ifa);
1.1 cgd 1217: }
1218:
1219: /*
1.311 rmind 1220: * save_rte: save incoming source route for use in replies, to be picked
1221: * up later by ip_srcroute if the receiver is interested.
1.1 cgd 1222: */
1.311 rmind 1223: static void
1.337.2.5! pgoyette 1224: save_rte(struct mbuf *m, u_char *option, struct in_addr dst)
1.1 cgd 1225: {
1.337.2.5! pgoyette 1226: struct ip_srcrt *isr;
! 1227: struct m_tag *mtag;
1.1 cgd 1228: unsigned olen;
1229:
1230: olen = option[IPOPT_OLEN];
1.337.2.5! pgoyette 1231: if (olen > sizeof(isr->isr_hdr) + sizeof(isr->isr_routes))
! 1232: return;
! 1233:
! 1234: mtag = m_tag_get(PACKET_TAG_SRCROUTE, sizeof(*isr), M_NOWAIT);
! 1235: if (mtag == NULL)
1.1 cgd 1236: return;
1.337.2.5! pgoyette 1237: isr = (struct ip_srcrt *)(mtag + 1);
! 1238:
! 1239: memcpy(isr->isr_hdr, option, olen);
! 1240: isr->isr_nhops = (olen - IPOPT_OFFSET - 1) / sizeof(struct in_addr);
! 1241: isr->isr_dst = dst;
! 1242: m_tag_prepend(m, mtag);
1.1 cgd 1243: }
1244:
1245: /*
1246: * Retrieve incoming source route for use in replies,
1247: * in the same form used by setsockopt.
1248: * The first hop is placed before the options, will be removed later.
1249: */
1250: struct mbuf *
1.337.2.5! pgoyette 1251: ip_srcroute(struct mbuf *m0)
1.1 cgd 1252: {
1.109 augustss 1253: struct in_addr *p, *q;
1254: struct mbuf *m;
1.337.2.5! pgoyette 1255: struct ip_srcrt *isr;
! 1256: struct m_tag *mtag;
1.1 cgd 1257:
1.337.2.5! pgoyette 1258: mtag = m_tag_find(m0, PACKET_TAG_SRCROUTE, NULL);
! 1259: if (mtag == NULL)
1.237 dyoung 1260: return NULL;
1.337.2.5! pgoyette 1261: isr = (struct ip_srcrt *)(mtag + 1);
! 1262:
! 1263: if (isr->isr_nhops == 0)
! 1264: return NULL;
! 1265:
1.1 cgd 1266: m = m_get(M_DONTWAIT, MT_SOOPTS);
1.337.2.5! pgoyette 1267: if (m == NULL)
1.237 dyoung 1268: return NULL;
1.1 cgd 1269:
1.164 matt 1270: MCLAIM(m, &inetdomain.dom_mowner);
1.337.2.5! pgoyette 1271: #define OPTSIZ (sizeof(isr->isr_nop) + sizeof(isr->isr_hdr))
1.1 cgd 1272:
1.337.2.5! pgoyette 1273: /* length is (nhops+1)*sizeof(addr) + sizeof(nop + header) */
! 1274: m->m_len = (isr->isr_nhops + 1) * sizeof(struct in_addr) + OPTSIZ;
1.1 cgd 1275:
1276: /*
1277: * First save first hop for return route
1278: */
1.337.2.5! pgoyette 1279: p = &(isr->isr_routes[isr->isr_nhops - 1]);
1.1 cgd 1280: *(mtod(m, struct in_addr *)) = *p--;
1281:
1282: /*
1283: * Copy option fields and padding (nop) to mbuf.
1284: */
1.337.2.5! pgoyette 1285: isr->isr_nop = IPOPT_NOP;
! 1286: isr->isr_hdr[IPOPT_OFFSET] = IPOPT_MINOFF;
! 1287: memmove(mtod(m, char *) + sizeof(struct in_addr), &isr->isr_nop,
1.244 christos 1288: OPTSIZ);
1289: q = (struct in_addr *)(mtod(m, char *) +
1.1 cgd 1290: sizeof(struct in_addr) + OPTSIZ);
1291: #undef OPTSIZ
1292: /*
1293: * Record return path as an IP source route,
1294: * reversing the path (pointers are now aligned).
1295: */
1.337.2.5! pgoyette 1296: while (p >= isr->isr_routes) {
1.1 cgd 1297: *q++ = *p--;
1298: }
1299: /*
1300: * Last hop goes to final destination.
1301: */
1.337.2.5! pgoyette 1302: *q = isr->isr_dst;
! 1303: m_tag_delete(m0, mtag);
! 1304: return m;
1.1 cgd 1305: }
1306:
1.139 matt 1307: const int inetctlerrmap[PRC_NCMDS] = {
1.256 yamt 1308: [PRC_MSGSIZE] = EMSGSIZE,
1309: [PRC_HOSTDEAD] = EHOSTDOWN,
1310: [PRC_HOSTUNREACH] = EHOSTUNREACH,
1311: [PRC_UNREACH_NET] = EHOSTUNREACH,
1312: [PRC_UNREACH_HOST] = EHOSTUNREACH,
1313: [PRC_UNREACH_PROTOCOL] = ECONNREFUSED,
1314: [PRC_UNREACH_PORT] = ECONNREFUSED,
1315: [PRC_UNREACH_SRCFAIL] = EHOSTUNREACH,
1316: [PRC_PARAMPROB] = ENOPROTOOPT,
1.1 cgd 1317: };
1318:
1.295 dyoung 1319: void
1320: ip_fasttimo(void)
1321: {
1322: if (ip_drainwanted) {
1323: ip_drain();
1324: ip_drainwanted = 0;
1325: }
1326: }
1327:
1328: void
1329: ip_drainstub(void)
1330: {
1331: ip_drainwanted = 1;
1332: }
1333:
1.1 cgd 1334: /*
1335: * Forward a packet. If some error occurs return the sender
1336: * an icmp packet. Note we can't always generate a meaningful
1337: * icmp message because icmp doesn't have a large enough repertoire
1338: * of codes and types.
1339: *
1340: * If not forwarding, just drop the packet. This could be confusing
1341: * if ipforwarding was zero but some routing protocol was advancing
1342: * us as a gateway to somewhere. However, we must let the routing
1343: * protocol deal with that.
1344: *
1345: * The srcrt parameter indicates whether the packet is being forwarded
1346: * via a source route.
1347: */
1.313 rmind 1348: static void
1.329 ozaki-r 1349: ip_forward(struct mbuf *m, int srcrt, struct ifnet *rcvif)
1.1 cgd 1350: {
1.109 augustss 1351: struct ip *ip = mtod(m, struct ip *);
1352: struct rtentry *rt;
1.220 christos 1353: int error, type = 0, code = 0, destmtu = 0;
1.1 cgd 1354: struct mbuf *mcopy;
1.13 mycroft 1355: n_long dest;
1.249 dyoung 1356: union {
1357: struct sockaddr dst;
1358: struct sockaddr_in dst4;
1359: } u;
1.320 ozaki-r 1360: uint64_t *ips;
1.164 matt 1361:
1.318 rmind 1362: KASSERTMSG(cpu_softintr_p(), "ip_forward: not in the software "
1363: "interrupt handler; synchronization assumptions violated");
1364:
1.164 matt 1365: /*
1366: * We are now in the output path.
1367: */
1368: MCLAIM(m, &ip_tx_mowner);
1.135 thorpej 1369:
1370: /*
1371: * Clear any in-bound checksum flags for this packet.
1372: */
1373: m->m_pkthdr.csum_flags = 0;
1.1 cgd 1374:
1.13 mycroft 1375: dest = 0;
1.93 sommerfe 1376: if (m->m_flags & (M_BCAST|M_MCAST) || in_canforward(ip->ip_dst) == 0) {
1.266 thorpej 1377: IP_STATINC(IP_STAT_CANTFORWARD);
1.1 cgd 1378: m_freem(m);
1379: return;
1380: }
1.318 rmind 1381:
1.1 cgd 1382: if (ip->ip_ttl <= IPTTLDEC) {
1.13 mycroft 1383: icmp_error(m, ICMP_TIMXCEED, ICMP_TIMXCEED_INTRANS, dest, 0);
1.1 cgd 1384: return;
1385: }
1386:
1.249 dyoung 1387: sockaddr_in_init(&u.dst4, &ip->ip_dst, 0);
1.318 rmind 1388:
1.337.2.3 pgoyette 1389: ro = percpu_getref(ipforward_rt_percpu);
1390: rt = rtcache_lookup(ro, &u.dst);
1391: if (rt == NULL) {
1392: percpu_putref(ipforward_rt_percpu);
1.249 dyoung 1393: icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_NET, dest, 0);
1394: return;
1.1 cgd 1395: }
1396:
1397: /*
1.34 mycroft 1398: * Save at most 68 bytes of the packet in case
1.1 cgd 1399: * we need to generate an ICMP message to the src.
1.119 itojun 1400: * Pullup to avoid sharing mbuf cluster between m and mcopy.
1.1 cgd 1401: */
1.155 itojun 1402: mcopy = m_copym(m, 0, imin(ntohs(ip->ip_len), 68), M_DONTWAIT);
1.119 itojun 1403: if (mcopy)
1404: mcopy = m_pullup(mcopy, ip->ip_hl << 2);
1.1 cgd 1405:
1.221 christos 1406: ip->ip_ttl -= IPTTLDEC;
1407:
1.1 cgd 1408: /*
1409: * If forwarding packet using same interface that it came in on,
1410: * perhaps should send a redirect to sender to shortcut a hop.
1411: * Only send redirect if source is sending directly to us,
1412: * and if packet was not source routed (or has any options).
1413: * Also, don't send redirect if forwarding using a default route
1414: * or a route modified by a redirect.
1415: */
1.329 ozaki-r 1416: if (rt->rt_ifp == rcvif &&
1.1 cgd 1417: (rt->rt_flags & (RTF_DYNAMIC|RTF_MODIFIED)) == 0 &&
1.250 dyoung 1418: !in_nullhost(satocsin(rt_getkey(rt))->sin_addr) &&
1.1 cgd 1419: ipsendredirects && !srcrt) {
1.19 mycroft 1420: if (rt->rt_ifa &&
1421: (ip->ip_src.s_addr & ifatoia(rt->rt_ifa)->ia_subnetmask) ==
1422: ifatoia(rt->rt_ifa)->ia_subnet) {
1.77 thorpej 1423: if (rt->rt_flags & RTF_GATEWAY)
1424: dest = satosin(rt->rt_gateway)->sin_addr.s_addr;
1425: else
1426: dest = ip->ip_dst.s_addr;
1427: /*
1428: * Router requirements says to only send host
1429: * redirects.
1430: */
1431: type = ICMP_REDIRECT;
1432: code = ICMP_REDIRECT_HOST;
1.1 cgd 1433: }
1434: }
1.337.2.3 pgoyette 1435: rtcache_unref(rt, ro);
1.1 cgd 1436:
1.238 dyoung 1437: error = ip_output(m, NULL, &ipforward_rt,
1.173 jonathan 1438: (IP_FORWARDING | (ip_directedbcast ? IP_ALLOWBROADCAST : 0)),
1.296 plunky 1439: NULL, NULL);
1.173 jonathan 1440:
1.320 ozaki-r 1441: if (error) {
1.266 thorpej 1442: IP_STATINC(IP_STAT_CANTFORWARD);
1.320 ozaki-r 1443: goto error;
1444: }
1445:
1446: ips = IP_STAT_GETREF();
1447: ips[IP_STAT_FORWARD]++;
1448:
1449: if (type) {
1450: ips[IP_STAT_REDIRECTSENT]++;
1451: IP_STAT_PUTREF();
1452: goto redirect;
1453: }
1454:
1455: IP_STAT_PUTREF();
1456: if (mcopy) {
1.63 matt 1457: #ifdef GATEWAY
1.320 ozaki-r 1458: if (mcopy->m_flags & M_CANFASTFWD)
1459: ipflow_create(&ipforward_rt, mcopy);
1.63 matt 1460: #endif
1.320 ozaki-r 1461: m_freem(mcopy);
1.1 cgd 1462: }
1.320 ozaki-r 1463:
1.337.2.2 pgoyette 1464: percpu_putref(ipforward_rt_percpu);
1.320 ozaki-r 1465: return;
1466:
1467: redirect:
1468: error:
1.318 rmind 1469: if (mcopy == NULL) {
1.337.2.2 pgoyette 1470: percpu_putref(ipforward_rt_percpu);
1.1 cgd 1471: return;
1.318 rmind 1472: }
1.13 mycroft 1473:
1.1 cgd 1474: switch (error) {
1475:
1476: case 0: /* forwarded, but need redirect */
1477: /* type, code set above */
1478: break;
1479:
1480: case ENETUNREACH: /* shouldn't happen, checked above */
1481: case EHOSTUNREACH:
1482: case ENETDOWN:
1483: case EHOSTDOWN:
1484: default:
1485: type = ICMP_UNREACH;
1486: code = ICMP_UNREACH_HOST;
1487: break;
1488:
1489: case EMSGSIZE:
1490: type = ICMP_UNREACH;
1491: code = ICMP_UNREACH_NEEDFRAG;
1.263 cube 1492:
1.337.2.3 pgoyette 1493: if ((rt = rtcache_validate(ro)) != NULL) {
1.274 seanb 1494: destmtu = rt->rt_ifp->if_mtu;
1.337.2.3 pgoyette 1495: rtcache_unref(rt, ro);
1496: }
1.305 rmind 1497: #ifdef IPSEC
1.317 christos 1498: if (ipsec_used)
1499: (void)ipsec4_forward(mcopy, &destmtu);
1.305 rmind 1500: #endif
1.266 thorpej 1501: IP_STATINC(IP_STAT_CANTFRAG);
1.1 cgd 1502: break;
1503:
1504: case ENOBUFS:
1.143 itojun 1505: /*
1.311 rmind 1506: * Do not generate ICMP_SOURCEQUENCH as required in RFC 1812,
1507: * Requirements for IP Version 4 Routers. Source quench can
1508: * big problem under DoS attacks or if the underlying
1509: * interface is rate-limited.
1.143 itojun 1510: */
1511: if (mcopy)
1512: m_freem(mcopy);
1.337.2.2 pgoyette 1513: percpu_putref(ipforward_rt_percpu);
1.143 itojun 1514: return;
1.1 cgd 1515: }
1.220 christos 1516: icmp_error(mcopy, type, code, dest, destmtu);
1.337.2.2 pgoyette 1517: percpu_putref(ipforward_rt_percpu);
1.44 thorpej 1518: }
1519:
1520: void
1.211 perry 1521: ip_savecontrol(struct inpcb *inp, struct mbuf **mp, struct ip *ip,
1522: struct mbuf *m)
1.44 thorpej 1523: {
1.311 rmind 1524: struct socket *so = inp->inp_socket;
1.329 ozaki-r 1525: ifnet_t *ifp;
1.311 rmind 1526: int inpflags = inp->inp_flags;
1.329 ozaki-r 1527: struct psref psref;
1.44 thorpej 1528:
1.329 ozaki-r 1529: ifp = m_get_rcvif_psref(m, &psref);
1.330 ozaki-r 1530: if (__predict_false(ifp == NULL))
1531: return; /* XXX should report error? */
1532:
1.311 rmind 1533: if (so->so_options & SO_TIMESTAMP
1.278 christos 1534: #ifdef SO_OTIMESTAMP
1.311 rmind 1535: || so->so_options & SO_OTIMESTAMP
1.278 christos 1536: #endif
1537: ) {
1.44 thorpej 1538: struct timeval tv;
1539:
1540: microtime(&tv);
1.278 christos 1541: #ifdef SO_OTIMESTAMP
1.311 rmind 1542: if (so->so_options & SO_OTIMESTAMP) {
1.278 christos 1543: struct timeval50 tv50;
1544: timeval_to_timeval50(&tv, &tv50);
1545: *mp = sbcreatecontrol((void *) &tv50, sizeof(tv50),
1546: SCM_OTIMESTAMP, SOL_SOCKET);
1547: } else
1548: #endif
1.244 christos 1549: *mp = sbcreatecontrol((void *) &tv, sizeof(tv),
1.44 thorpej 1550: SCM_TIMESTAMP, SOL_SOCKET);
1551: if (*mp)
1552: mp = &(*mp)->m_next;
1553: }
1.311 rmind 1554: if (inpflags & INP_RECVDSTADDR) {
1.244 christos 1555: *mp = sbcreatecontrol((void *) &ip->ip_dst,
1.44 thorpej 1556: sizeof(struct in_addr), IP_RECVDSTADDR, IPPROTO_IP);
1557: if (*mp)
1558: mp = &(*mp)->m_next;
1559: }
1.311 rmind 1560: if (inpflags & INP_RECVPKTINFO) {
1.306 christos 1561: struct in_pktinfo ipi;
1.307 christos 1562: ipi.ipi_addr = ip->ip_src;
1.314 rmind 1563: ipi.ipi_ifindex = ifp->if_index;
1.306 christos 1564: *mp = sbcreatecontrol((void *) &ipi,
1565: sizeof(ipi), IP_RECVPKTINFO, IPPROTO_IP);
1566: if (*mp)
1567: mp = &(*mp)->m_next;
1568: }
1.311 rmind 1569: if (inpflags & INP_PKTINFO) {
1.306 christos 1570: struct in_pktinfo ipi;
1.307 christos 1571: ipi.ipi_addr = ip->ip_dst;
1.314 rmind 1572: ipi.ipi_ifindex = ifp->if_index;
1.306 christos 1573: *mp = sbcreatecontrol((void *) &ipi,
1574: sizeof(ipi), IP_PKTINFO, IPPROTO_IP);
1575: if (*mp)
1576: mp = &(*mp)->m_next;
1577: }
1.311 rmind 1578: if (inpflags & INP_RECVIF) {
1.44 thorpej 1579: struct sockaddr_dl sdl;
1580:
1.337 ozaki-r 1581: sockaddr_dl_init(&sdl, sizeof(sdl), ifp->if_index, 0, NULL, 0,
1582: NULL, 0);
1.251 dyoung 1583: *mp = sbcreatecontrol(&sdl, sdl.sdl_len, IP_RECVIF, IPPROTO_IP);
1.44 thorpej 1584: if (*mp)
1585: mp = &(*mp)->m_next;
1586: }
1.311 rmind 1587: if (inpflags & INP_RECVTTL) {
1.282 minskim 1588: *mp = sbcreatecontrol((void *) &ip->ip_ttl,
1589: sizeof(uint8_t), IP_RECVTTL, IPPROTO_IP);
1590: if (*mp)
1591: mp = &(*mp)->m_next;
1592: }
1.329 ozaki-r 1593: m_put_rcvif_psref(ifp, &psref);
1.13 mycroft 1594: }
1595:
1.189 atatat 1596: /*
1.228 elad 1597: * sysctl helper routine for net.inet.ip.forwsrcrt.
1598: */
1599: static int
1600: sysctl_net_inet_ip_forwsrcrt(SYSCTLFN_ARGS)
1601: {
1602: int error, tmp;
1603: struct sysctlnode node;
1604:
1605: node = *rnode;
1606: tmp = ip_forwsrcrt;
1607: node.sysctl_data = &tmp;
1608: error = sysctl_lookup(SYSCTLFN_CALL(&node));
1609: if (error || newp == NULL)
1610: return (error);
1611:
1.280 elad 1612: error = kauth_authorize_network(l->l_cred, KAUTH_NETWORK_FORWSRCRT,
1613: 0, NULL, NULL, NULL);
1614: if (error)
1615: return (error);
1.228 elad 1616:
1617: ip_forwsrcrt = tmp;
1618:
1619: return (0);
1620: }
1621:
1622: /*
1.189 atatat 1623: * sysctl helper routine for net.inet.ip.mtudisctimeout. checks the
1624: * range of the new value and tweaks timers if it changes.
1625: */
1626: static int
1627: sysctl_net_inet_ip_pmtudto(SYSCTLFN_ARGS)
1.13 mycroft 1628: {
1.189 atatat 1629: int error, tmp;
1630: struct sysctlnode node;
1631:
1.337.2.4 pgoyette 1632: icmp_mtudisc_lock();
1633:
1.189 atatat 1634: node = *rnode;
1635: tmp = ip_mtudisc_timeout;
1636: node.sysctl_data = &tmp;
1637: error = sysctl_lookup(SYSCTLFN_CALL(&node));
1638: if (error || newp == NULL)
1.337.2.4 pgoyette 1639: goto out;
1640: if (tmp < 0) {
1641: error = EINVAL;
1642: goto out;
1643: }
1.273 matt 1644:
1.189 atatat 1645: ip_mtudisc_timeout = tmp;
1646: rt_timer_queue_change(ip_mtudisc_timeout_q, ip_mtudisc_timeout);
1.337.2.4 pgoyette 1647: error = 0;
1648: out:
1649: icmp_mtudisc_unlock();
1650: return error;
1.189 atatat 1651: }
1.54 lukem 1652:
1.266 thorpej 1653: static int
1654: sysctl_net_inet_ip_stats(SYSCTLFN_ARGS)
1655: {
1656:
1.271 thorpej 1657: return (NETSTAT_SYSCTL(ipstat_percpu, IP_NSTATS));
1.266 thorpej 1658: }
1.131 itojun 1659:
1.284 pooka 1660: static void
1661: sysctl_net_inet_ip_setup(struct sysctllog **clog)
1.189 atatat 1662: {
1.197 atatat 1663: sysctl_createv(clog, 0, NULL, NULL,
1664: CTLFLAG_PERMANENT,
1.203 atatat 1665: CTLTYPE_NODE, "inet",
1666: SYSCTL_DESCR("PF_INET related settings"),
1.189 atatat 1667: NULL, 0, NULL, 0,
1668: CTL_NET, PF_INET, CTL_EOL);
1.197 atatat 1669: sysctl_createv(clog, 0, NULL, NULL,
1670: CTLFLAG_PERMANENT,
1.203 atatat 1671: CTLTYPE_NODE, "ip",
1672: SYSCTL_DESCR("IPv4 related settings"),
1.189 atatat 1673: NULL, 0, NULL, 0,
1674: CTL_NET, PF_INET, IPPROTO_IP, CTL_EOL);
1.212 perry 1675:
1.197 atatat 1676: sysctl_createv(clog, 0, NULL, NULL,
1677: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1678: CTLTYPE_INT, "forwarding",
1679: SYSCTL_DESCR("Enable forwarding of INET datagrams"),
1.189 atatat 1680: NULL, 0, &ipforwarding, 0,
1681: CTL_NET, PF_INET, IPPROTO_IP,
1682: IPCTL_FORWARDING, CTL_EOL);
1.197 atatat 1683: sysctl_createv(clog, 0, NULL, NULL,
1684: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1685: CTLTYPE_INT, "redirect",
1686: SYSCTL_DESCR("Enable sending of ICMP redirect messages"),
1.189 atatat 1687: NULL, 0, &ipsendredirects, 0,
1688: CTL_NET, PF_INET, IPPROTO_IP,
1689: IPCTL_SENDREDIRECTS, CTL_EOL);
1.197 atatat 1690: sysctl_createv(clog, 0, NULL, NULL,
1691: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1692: CTLTYPE_INT, "ttl",
1693: SYSCTL_DESCR("Default TTL for an INET datagram"),
1.189 atatat 1694: NULL, 0, &ip_defttl, 0,
1695: CTL_NET, PF_INET, IPPROTO_IP,
1696: IPCTL_DEFTTL, CTL_EOL);
1697: #ifdef IPCTL_DEFMTU
1.197 atatat 1698: sysctl_createv(clog, 0, NULL, NULL,
1699: CTLFLAG_PERMANENT /* |CTLFLAG_READWRITE? */,
1.203 atatat 1700: CTLTYPE_INT, "mtu",
1701: SYSCTL_DESCR("Default MTA for an INET route"),
1.189 atatat 1702: NULL, 0, &ip_mtu, 0,
1703: CTL_NET, PF_INET, IPPROTO_IP,
1704: IPCTL_DEFMTU, CTL_EOL);
1705: #endif /* IPCTL_DEFMTU */
1.197 atatat 1706: sysctl_createv(clog, 0, NULL, NULL,
1.228 elad 1707: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1708: CTLTYPE_INT, "forwsrcrt",
1709: SYSCTL_DESCR("Enable forwarding of source-routed "
1710: "datagrams"),
1.228 elad 1711: sysctl_net_inet_ip_forwsrcrt, 0, &ip_forwsrcrt, 0,
1.189 atatat 1712: CTL_NET, PF_INET, IPPROTO_IP,
1713: IPCTL_FORWSRCRT, CTL_EOL);
1.197 atatat 1714: sysctl_createv(clog, 0, NULL, NULL,
1715: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1716: CTLTYPE_INT, "directed-broadcast",
1717: SYSCTL_DESCR("Enable forwarding of broadcast datagrams"),
1.189 atatat 1718: NULL, 0, &ip_directedbcast, 0,
1719: CTL_NET, PF_INET, IPPROTO_IP,
1720: IPCTL_DIRECTEDBCAST, CTL_EOL);
1.197 atatat 1721: sysctl_createv(clog, 0, NULL, NULL,
1722: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1723: CTLTYPE_INT, "allowsrcrt",
1724: SYSCTL_DESCR("Accept source-routed datagrams"),
1.189 atatat 1725: NULL, 0, &ip_allowsrcrt, 0,
1726: CTL_NET, PF_INET, IPPROTO_IP,
1727: IPCTL_ALLOWSRCRT, CTL_EOL);
1.311 rmind 1728:
1.197 atatat 1729: sysctl_createv(clog, 0, NULL, NULL,
1730: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1731: CTLTYPE_INT, "mtudisc",
1732: SYSCTL_DESCR("Use RFC1191 Path MTU Discovery"),
1.189 atatat 1733: NULL, 0, &ip_mtudisc, 0,
1734: CTL_NET, PF_INET, IPPROTO_IP,
1735: IPCTL_MTUDISC, CTL_EOL);
1.197 atatat 1736: sysctl_createv(clog, 0, NULL, NULL,
1737: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1738: CTLTYPE_INT, "anonportmin",
1739: SYSCTL_DESCR("Lowest ephemeral port number to assign"),
1.189 atatat 1740: sysctl_net_inet_ip_ports, 0, &anonportmin, 0,
1741: CTL_NET, PF_INET, IPPROTO_IP,
1742: IPCTL_ANONPORTMIN, CTL_EOL);
1.197 atatat 1743: sysctl_createv(clog, 0, NULL, NULL,
1744: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1745: CTLTYPE_INT, "anonportmax",
1746: SYSCTL_DESCR("Highest ephemeral port number to assign"),
1.189 atatat 1747: sysctl_net_inet_ip_ports, 0, &anonportmax, 0,
1748: CTL_NET, PF_INET, IPPROTO_IP,
1749: IPCTL_ANONPORTMAX, CTL_EOL);
1.197 atatat 1750: sysctl_createv(clog, 0, NULL, NULL,
1751: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1752: CTLTYPE_INT, "mtudisctimeout",
1753: SYSCTL_DESCR("Lifetime of a Path MTU Discovered route"),
1.300 dsl 1754: sysctl_net_inet_ip_pmtudto, 0, (void *)&ip_mtudisc_timeout, 0,
1.189 atatat 1755: CTL_NET, PF_INET, IPPROTO_IP,
1756: IPCTL_MTUDISCTIMEOUT, CTL_EOL);
1757: #ifndef IPNOPRIVPORTS
1.197 atatat 1758: sysctl_createv(clog, 0, NULL, NULL,
1759: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1760: CTLTYPE_INT, "lowportmin",
1761: SYSCTL_DESCR("Lowest privileged ephemeral port number "
1762: "to assign"),
1.189 atatat 1763: sysctl_net_inet_ip_ports, 0, &lowportmin, 0,
1764: CTL_NET, PF_INET, IPPROTO_IP,
1765: IPCTL_LOWPORTMIN, CTL_EOL);
1.197 atatat 1766: sysctl_createv(clog, 0, NULL, NULL,
1767: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1768: CTLTYPE_INT, "lowportmax",
1769: SYSCTL_DESCR("Highest privileged ephemeral port number "
1770: "to assign"),
1.189 atatat 1771: sysctl_net_inet_ip_ports, 0, &lowportmax, 0,
1772: CTL_NET, PF_INET, IPPROTO_IP,
1773: IPCTL_LOWPORTMAX, CTL_EOL);
1774: #endif /* IPNOPRIVPORTS */
1775: #if NGRE > 0
1.197 atatat 1776: sysctl_createv(clog, 0, NULL, NULL,
1777: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1778: CTLTYPE_INT, "grettl",
1779: SYSCTL_DESCR("Default TTL for a gre tunnel datagram"),
1.189 atatat 1780: NULL, 0, &ip_gre_ttl, 0,
1781: CTL_NET, PF_INET, IPPROTO_IP,
1782: IPCTL_GRE_TTL, CTL_EOL);
1783: #endif /* NGRE */
1.197 atatat 1784: sysctl_createv(clog, 0, NULL, NULL,
1785: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1786: CTLTYPE_INT, "checkinterface",
1787: SYSCTL_DESCR("Enable receive side of Strong ES model "
1788: "from RFC1122"),
1.189 atatat 1789: NULL, 0, &ip_checkinterface, 0,
1790: CTL_NET, PF_INET, IPPROTO_IP,
1791: IPCTL_CHECKINTERFACE, CTL_EOL);
1.197 atatat 1792: sysctl_createv(clog, 0, NULL, NULL,
1793: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203 atatat 1794: CTLTYPE_INT, "random_id",
1795: SYSCTL_DESCR("Assign random ip_id values"),
1.189 atatat 1796: NULL, 0, &ip_do_randomid, 0,
1797: CTL_NET, PF_INET, IPPROTO_IP,
1798: IPCTL_RANDOMID, CTL_EOL);
1.206 thorpej 1799: sysctl_createv(clog, 0, NULL, NULL,
1800: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1801: CTLTYPE_INT, "do_loopback_cksum",
1802: SYSCTL_DESCR("Perform IP checksum on loopback"),
1803: NULL, 0, &ip_do_loopback_cksum, 0,
1804: CTL_NET, PF_INET, IPPROTO_IP,
1805: IPCTL_LOOPBACKCKSUM, CTL_EOL);
1.219 elad 1806: sysctl_createv(clog, 0, NULL, NULL,
1807: CTLFLAG_PERMANENT,
1808: CTLTYPE_STRUCT, "stats",
1809: SYSCTL_DESCR("IP statistics"),
1.266 thorpej 1810: sysctl_net_inet_ip_stats, 0, NULL, 0,
1.219 elad 1811: CTL_NET, PF_INET, IPPROTO_IP, IPCTL_STATS,
1812: CTL_EOL);
1.322 joerg 1813: #if NARP
1.321 roy 1814: sysctl_createv(clog, 0, NULL, NULL,
1815: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1816: CTLTYPE_INT, "dad_count",
1817: SYSCTL_DESCR("Number of Duplicate Address Detection "
1818: "probes to send"),
1819: NULL, 0, &ip_dad_count, 0,
1820: CTL_NET, PF_INET, IPPROTO_IP,
1821: IPCTL_DAD_COUNT, CTL_EOL);
1.322 joerg 1822: #endif
1.301 christos 1823:
1824: /* anonportalgo RFC6056 subtree */
1.302 christos 1825: const struct sysctlnode *portalgo_node;
1826: sysctl_createv(clog, 0, NULL, &portalgo_node,
1.301 christos 1827: CTLFLAG_PERMANENT,
1828: CTLTYPE_NODE, "anonportalgo",
1829: SYSCTL_DESCR("Anonymous Port Algorithm Selection (RFC 6056)"),
1830: NULL, 0, NULL, 0,
1831: CTL_NET, PF_INET, IPPROTO_IP, CTL_CREATE, CTL_EOL);
1.302 christos 1832: sysctl_createv(clog, 0, &portalgo_node, NULL,
1.301 christos 1833: CTLFLAG_PERMANENT,
1834: CTLTYPE_STRING, "available",
1835: SYSCTL_DESCR("available algorithms"),
1.302 christos 1836: sysctl_portalgo_available, 0, NULL, PORTALGO_MAXLEN,
1.301 christos 1837: CTL_CREATE, CTL_EOL);
1.302 christos 1838: sysctl_createv(clog, 0, &portalgo_node, NULL,
1.301 christos 1839: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1840: CTLTYPE_STRING, "selected",
1841: SYSCTL_DESCR("selected algorithm"),
1.303 christos 1842: sysctl_portalgo_selected4, 0, NULL, PORTALGO_MAXLEN,
1843: CTL_CREATE, CTL_EOL);
1844: sysctl_createv(clog, 0, &portalgo_node, NULL,
1845: CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1846: CTLTYPE_STRUCT, "reserve",
1847: SYSCTL_DESCR("bitmap of reserved ports"),
1848: sysctl_portalgo_reserve4, 0, NULL, 0,
1.301 christos 1849: CTL_CREATE, CTL_EOL);
1.1 cgd 1850: }
1.266 thorpej 1851:
1852: void
1853: ip_statinc(u_int stat)
1854: {
1855:
1856: KASSERT(stat < IP_NSTATS);
1857: IP_STATINC(stat);
1858: }
CVSweb <webmaster@jp.NetBSD.org>