[BACK]Return to ip_input.c CVS log [TXT][DIR] Up to [cvs.NetBSD.org] / src / sys / netinet

Annotation of src/sys/netinet/ip_input.c, Revision 1.268

1.268   ! ad          1: /*     $NetBSD: ip_input.c,v 1.267 2008/04/23 06:09:04 thorpej 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:  * 3. All advertising materials mentioning features or use of this software
                     49:  *    must display the following acknowledgement:
                     50:  *     This product includes software developed by the NetBSD
                     51:  *     Foundation, Inc. and its contributors.
                     52:  * 4. Neither the name of The NetBSD Foundation nor the names of its
                     53:  *    contributors may be used to endorse or promote products derived
                     54:  *    from this software without specific prior written permission.
                     55:  *
                     56:  * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
                     57:  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
                     58:  * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
                     59:  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
                     60:  * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
                     61:  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
                     62:  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
                     63:  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
                     64:  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
                     65:  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
                     66:  * POSSIBILITY OF SUCH DAMAGE.
                     67:  */
1.14      cgd        68:
1.1       cgd        69: /*
1.13      mycroft    70:  * Copyright (c) 1982, 1986, 1988, 1993
                     71:  *     The Regents of the University of California.  All rights reserved.
1.1       cgd        72:  *
                     73:  * Redistribution and use in source and binary forms, with or without
                     74:  * modification, are permitted provided that the following conditions
                     75:  * are met:
                     76:  * 1. Redistributions of source code must retain the above copyright
                     77:  *    notice, this list of conditions and the following disclaimer.
                     78:  * 2. Redistributions in binary form must reproduce the above copyright
                     79:  *    notice, this list of conditions and the following disclaimer in the
                     80:  *    documentation and/or other materials provided with the distribution.
1.172     agc        81:  * 3. Neither the name of the University nor the names of its contributors
1.1       cgd        82:  *    may be used to endorse or promote products derived from this software
                     83:  *    without specific prior written permission.
                     84:  *
                     85:  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
                     86:  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
                     87:  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
                     88:  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
                     89:  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
                     90:  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
                     91:  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
                     92:  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
                     93:  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
                     94:  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
                     95:  * SUCH DAMAGE.
                     96:  *
1.14      cgd        97:  *     @(#)ip_input.c  8.2 (Berkeley) 1/4/94
1.1       cgd        98:  */
1.141     lukem      99:
                    100: #include <sys/cdefs.h>
1.268   ! ad        101: __KERNEL_RCSID(0, "$NetBSD: ip_input.c,v 1.267 2008/04/23 06:09:04 thorpej Exp $");
1.55      scottr    102:
1.184     jonathan  103: #include "opt_inet.h"
1.62      matt      104: #include "opt_gateway.h"
1.69      mrg       105: #include "opt_pfil_hooks.h"
1.91      thorpej   106: #include "opt_ipsec.h"
1.55      scottr    107: #include "opt_mrouting.h"
1.167     martin    108: #include "opt_mbuftrace.h"
1.135     thorpej   109: #include "opt_inet_csum.h"
1.1       cgd       110:
1.5       mycroft   111: #include <sys/param.h>
                    112: #include <sys/systm.h>
                    113: #include <sys/malloc.h>
                    114: #include <sys/mbuf.h>
                    115: #include <sys/domain.h>
                    116: #include <sys/protosw.h>
                    117: #include <sys/socket.h>
1.44      thorpej   118: #include <sys/socketvar.h>
1.5       mycroft   119: #include <sys/errno.h>
                    120: #include <sys/time.h>
                    121: #include <sys/kernel.h>
1.72      thorpej   122: #include <sys/pool.h>
1.28      christos  123: #include <sys/sysctl.h>
1.230     elad      124: #include <sys/kauth.h>
1.1       cgd       125:
1.5       mycroft   126: #include <net/if.h>
1.44      thorpej   127: #include <net/if_dl.h>
1.5       mycroft   128: #include <net/route.h>
1.45      mrg       129: #include <net/pfil.h>
1.1       cgd       130:
1.5       mycroft   131: #include <netinet/in.h>
                    132: #include <netinet/in_systm.h>
                    133: #include <netinet/ip.h>
                    134: #include <netinet/in_pcb.h>
1.215     yamt      135: #include <netinet/in_proto.h>
1.5       mycroft   136: #include <netinet/in_var.h>
                    137: #include <netinet/ip_var.h>
1.266     thorpej   138: #include <netinet/ip_private.h>
1.5       mycroft   139: #include <netinet/ip_icmp.h>
1.89      itojun    140: /* just for gif_ttl */
                    141: #include <netinet/in_gif.h>
                    142: #include "gif.h"
1.144     martin    143: #include <net/if_gre.h>
                    144: #include "gre.h"
1.111     jdolecek  145:
                    146: #ifdef MROUTING
                    147: #include <netinet/ip_mroute.h>
                    148: #endif
1.89      itojun    149:
                    150: #ifdef IPSEC
                    151: #include <netinet6/ipsec.h>
1.267     thorpej   152: #include <netinet6/ipsec_private.h>
1.89      itojun    153: #include <netkey/key.h>
                    154: #endif
1.173     jonathan  155: #ifdef FAST_IPSEC
                    156: #include <netipsec/ipsec.h>
                    157: #include <netipsec/key.h>
                    158: #endif /* FAST_IPSEC*/
1.44      thorpej   159:
1.1       cgd       160: #ifndef        IPFORWARDING
                    161: #ifdef GATEWAY
                    162: #define        IPFORWARDING    1       /* forward IP packets not for us */
                    163: #else /* GATEWAY */
                    164: #define        IPFORWARDING    0       /* don't forward IP packets not for us */
                    165: #endif /* GATEWAY */
                    166: #endif /* IPFORWARDING */
                    167: #ifndef        IPSENDREDIRECTS
                    168: #define        IPSENDREDIRECTS 1
                    169: #endif
1.26      thorpej   170: #ifndef IPFORWSRCRT
1.47      cjs       171: #define        IPFORWSRCRT     1       /* forward source-routed packets */
                    172: #endif
                    173: #ifndef IPALLOWSRCRT
1.48      mrg       174: #define        IPALLOWSRCRT    1       /* allow source-routed packets */
1.26      thorpej   175: #endif
1.53      kml       176: #ifndef IPMTUDISC
1.153     itojun    177: #define IPMTUDISC      1
1.53      kml       178: #endif
1.60      kml       179: #ifndef IPMTUDISCTIMEOUT
1.61      kml       180: #define IPMTUDISCTIMEOUT (10 * 60)     /* as per RFC 1191 */
1.60      kml       181: #endif
1.53      kml       182:
1.27      thorpej   183: /*
                    184:  * Note: DIRECTED_BROADCAST is handled this way so that previous
                    185:  * configuration using this option will Just Work.
                    186:  */
                    187: #ifndef IPDIRECTEDBCAST
                    188: #ifdef DIRECTED_BROADCAST
                    189: #define IPDIRECTEDBCAST        1
                    190: #else
                    191: #define        IPDIRECTEDBCAST 0
                    192: #endif /* DIRECTED_BROADCAST */
                    193: #endif /* IPDIRECTEDBCAST */
1.1       cgd       194: int    ipforwarding = IPFORWARDING;
                    195: int    ipsendredirects = IPSENDREDIRECTS;
1.13      mycroft   196: int    ip_defttl = IPDEFTTL;
1.26      thorpej   197: int    ip_forwsrcrt = IPFORWSRCRT;
1.27      thorpej   198: int    ip_directedbcast = IPDIRECTEDBCAST;
1.47      cjs       199: int    ip_allowsrcrt = IPALLOWSRCRT;
1.53      kml       200: int    ip_mtudisc = IPMTUDISC;
1.156     itojun    201: int    ip_mtudisc_timeout = IPMTUDISCTIMEOUT;
1.1       cgd       202: #ifdef DIAGNOSTIC
                    203: int    ipprintfs = 0;
                    204: #endif
1.184     jonathan  205:
                    206: int    ip_do_randomid = 0;
                    207:
1.165     christos  208: /*
                    209:  * XXX - Setting ip_checkinterface mostly implements the receive side of
                    210:  * the Strong ES model described in RFC 1122, but since the routing table
                    211:  * and transmit implementation do not implement the Strong ES model,
                    212:  * setting this to 1 results in an odd hybrid.
                    213:  *
                    214:  * XXX - ip_checkinterface currently must be disabled if you use ipnat
                    215:  * to translate the destination address to another local interface.
                    216:  *
                    217:  * XXX - ip_checkinterface must be disabled if you add IP aliases
                    218:  * to the loopback interface instead of the interface where the
                    219:  * packets for those addresses are received.
                    220:  */
                    221: int    ip_checkinterface = 0;
                    222:
1.1       cgd       223:
1.60      kml       224: struct rttimer_queue *ip_mtudisc_timeout_q = NULL;
                    225:
1.1       cgd       226: int    ipqmaxlen = IFQ_MAXLEN;
1.150     matt      227: u_long in_ifaddrhash;                          /* size of hash table - 1 */
                    228: int    in_ifaddrentries;                       /* total number of addrs */
1.212     perry     229: struct in_ifaddrhead in_ifaddrhead;
1.57      tls       230: struct in_ifaddrhashhead *in_ifaddrhashtbl;
1.166     matt      231: u_long in_multihash;                           /* size of hash table - 1 */
                    232: int    in_multientries;                        /* total number of addrs */
                    233: struct in_multihashhead *in_multihashtbl;
1.13      mycroft   234: struct ifqueue ipintrq;
1.183     jonathan  235: uint16_t ip_id;
1.75      thorpej   236:
1.266     thorpej   237: percpu_t *ipstat_percpu;
                    238:
1.121     thorpej   239: #ifdef PFIL_HOOKS
                    240: struct pfil_head inet_pfil_hook;
                    241: #endif
                    242:
1.194     jonathan  243: /*
                    244:  * Cached copy of nmbclusters. If nbclusters is different,
                    245:  * recalculate IP parameters derived from nmbclusters.
                    246:  */
                    247: static int     ip_nmbclusters;                 /* copy of nmbclusters */
1.210     perry     248: static void    ip_nmbclusters_changed(void);   /* recalc limits */
1.194     jonathan  249:
1.195     thorpej   250: #define CHECK_NMBCLUSTER_PARAMS()                              \
                    251: do {                                                           \
                    252:        if (__predict_false(ip_nmbclusters != nmbclusters))     \
                    253:                ip_nmbclusters_changed();                       \
                    254: } while (/*CONSTCOND*/0)
1.194     jonathan  255:
1.190     jonathan  256: /* IP datagram reassembly queues (hashed) */
                    257: #define IPREASS_NHASH_LOG2      6
                    258: #define IPREASS_NHASH           (1 << IPREASS_NHASH_LOG2)
                    259: #define IPREASS_HMASK           (IPREASS_NHASH - 1)
                    260: #define IPREASS_HASH(x,y) \
                    261:        (((((x) & 0xF) | ((((x) >> 8) & 0xF) << 4)) ^ (y)) & IPREASS_HMASK)
                    262: struct ipqhead ipq[IPREASS_NHASH];
1.75      thorpej   263: int    ipq_locked;
1.212     perry     264: static int     ip_nfragpackets;        /* packets in reass queue */
1.194     jonathan  265: static int     ip_nfrags;              /* total fragments in reass queues */
                    266:
                    267: int    ip_maxfragpackets = 200;        /* limit on packets. XXX sysctl */
                    268: int    ip_maxfrags;                    /* limit on fragments. XXX sysctl */
                    269:
                    270:
                    271: /*
                    272:  * Additive-Increase/Multiplicative-Decrease (AIMD) strategy for
                    273:  * IP reassembly queue buffer managment.
1.212     perry     274:  *
1.194     jonathan  275:  * We keep a count of total IP fragments (NB: not fragmented packets!)
                    276:  * awaiting reassembly (ip_nfrags) and a limit (ip_maxfrags) on fragments.
                    277:  * If ip_nfrags exceeds ip_maxfrags the limit, we drop half the
                    278:  * total fragments in  reassembly queues.This AIMD policy avoids
                    279:  * repeatedly deleting single packets under heavy fragmentation load
                    280:  * (e.g., from lossy NFS peers).
                    281:  */
1.212     perry     282: static u_int   ip_reass_ttl_decr(u_int ticks);
1.210     perry     283: static void    ip_reass_drophalf(void);
1.194     jonathan  284:
1.75      thorpej   285:
1.223     perry     286: static inline int ipq_lock_try(void);
                    287: static inline void ipq_unlock(void);
1.75      thorpej   288:
1.223     perry     289: static inline int
1.211     perry     290: ipq_lock_try(void)
1.75      thorpej   291: {
                    292:        int s;
                    293:
1.132     thorpej   294:        /*
1.149     wiz       295:         * Use splvm() -- we're blocking things that would cause
1.132     thorpej   296:         * mbuf allocation.
                    297:         */
                    298:        s = splvm();
1.75      thorpej   299:        if (ipq_locked) {
                    300:                splx(s);
                    301:                return (0);
                    302:        }
                    303:        ipq_locked = 1;
                    304:        splx(s);
                    305:        return (1);
                    306: }
                    307:
1.223     perry     308: static inline void
1.211     perry     309: ipq_unlock(void)
1.75      thorpej   310: {
                    311:        int s;
                    312:
1.132     thorpej   313:        s = splvm();
1.75      thorpej   314:        ipq_locked = 0;
                    315:        splx(s);
                    316: }
                    317:
                    318: #ifdef DIAGNOSTIC
                    319: #define        IPQ_LOCK()                                                      \
                    320: do {                                                                   \
                    321:        if (ipq_lock_try() == 0) {                                      \
                    322:                printf("%s:%d: ipq already locked\n", __FILE__, __LINE__); \
                    323:                panic("ipq_lock");                                      \
                    324:        }                                                               \
1.159     perry     325: } while (/*CONSTCOND*/ 0)
1.75      thorpej   326: #define        IPQ_LOCK_CHECK()                                                \
                    327: do {                                                                   \
                    328:        if (ipq_locked == 0) {                                          \
                    329:                printf("%s:%d: ipq lock not held\n", __FILE__, __LINE__); \
                    330:                panic("ipq lock check");                                \
                    331:        }                                                               \
1.159     perry     332: } while (/*CONSTCOND*/ 0)
1.75      thorpej   333: #else
                    334: #define        IPQ_LOCK()              (void) ipq_lock_try()
                    335: #define        IPQ_LOCK_CHECK()        /* nothing */
                    336: #endif
                    337:
                    338: #define        IPQ_UNLOCK()            ipq_unlock()
1.1       cgd       339:
1.246     ad        340: POOL_INIT(inmulti_pool, sizeof(struct in_multi), 0, 0, 0, "inmltpl", NULL,
                    341:     IPL_SOFTNET);
                    342: POOL_INIT(ipqent_pool, sizeof(struct ipqent), 0, 0, 0, "ipqepl", NULL,
                    343:     IPL_VM);
1.72      thorpej   344:
1.135     thorpej   345: #ifdef INET_CSUM_COUNTERS
                    346: #include <sys/device.h>
                    347:
                    348: struct evcnt ip_hwcsum_bad = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
                    349:     NULL, "inet", "hwcsum bad");
                    350: struct evcnt ip_hwcsum_ok = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
                    351:     NULL, "inet", "hwcsum ok");
                    352: struct evcnt ip_swcsum = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
                    353:     NULL, "inet", "swcsum");
                    354:
                    355: #define        INET_CSUM_COUNTER_INCR(ev)      (ev)->ev_count++
                    356:
1.201     matt      357: EVCNT_ATTACH_STATIC(ip_hwcsum_bad);
                    358: EVCNT_ATTACH_STATIC(ip_hwcsum_ok);
                    359: EVCNT_ATTACH_STATIC(ip_swcsum);
                    360:
1.135     thorpej   361: #else
                    362:
                    363: #define        INET_CSUM_COUNTER_INCR(ev)      /* nothing */
                    364:
                    365: #endif /* INET_CSUM_COUNTERS */
                    366:
1.1       cgd       367: /*
                    368:  * We need to save the IP options in case a protocol wants to respond
                    369:  * to an incoming packet over the same route if the packet got here
                    370:  * using IP source routing.  This allows connection establishment and
                    371:  * maintenance when the remote end is on a network that is not known
                    372:  * to us.
                    373:  */
                    374: int    ip_nhops = 0;
                    375: static struct ip_srcrt {
                    376:        struct  in_addr dst;                    /* final destination */
                    377:        char    nop;                            /* one NOP to align */
                    378:        char    srcopt[IPOPT_OFFSET + 1];       /* OPTVAL, OLEN and OFFSET */
                    379:        struct  in_addr route[MAX_IPOPTLEN/sizeof(struct in_addr)];
                    380: } ip_srcrt;
                    381:
1.210     perry     382: static void save_rte(u_char *, struct in_addr);
1.35      mycroft   383:
1.164     matt      384: #ifdef MBUFTRACE
1.234     dogcow    385: struct mowner ip_rx_mowner = MOWNER_INIT("internet", "rx");
                    386: struct mowner ip_tx_mowner = MOWNER_INIT("internet", "tx");
1.164     matt      387: #endif
                    388:
1.1       cgd       389: /*
1.194     jonathan  390:  * Compute IP limits derived from the value of nmbclusters.
                    391:  */
                    392: static void
                    393: ip_nmbclusters_changed(void)
                    394: {
                    395:        ip_maxfrags = nmbclusters / 4;
                    396:        ip_nmbclusters =  nmbclusters;
                    397: }
                    398:
                    399: /*
1.1       cgd       400:  * IP initialization: fill in IP protocol switch table.
                    401:  * All protocols not implemented in kernel go to raw IP protocol handler.
                    402:  */
1.8       mycroft   403: void
1.211     perry     404: ip_init(void)
1.1       cgd       405: {
1.199     matt      406:        const struct protosw *pr;
1.109     augustss  407:        int i;
1.1       cgd       408:
                    409:        pr = pffindproto(PF_INET, IPPROTO_RAW, SOCK_RAW);
                    410:        if (pr == 0)
                    411:                panic("ip_init");
                    412:        for (i = 0; i < IPPROTO_MAX; i++)
                    413:                ip_protox[i] = pr - inetsw;
                    414:        for (pr = inetdomain.dom_protosw;
                    415:            pr < inetdomain.dom_protoswNPROTOSW; pr++)
                    416:                if (pr->pr_domain->dom_family == PF_INET &&
                    417:                    pr->pr_protocol && pr->pr_protocol != IPPROTO_RAW)
                    418:                        ip_protox[pr->pr_protocol] = pr - inetsw;
1.192     jonathan  419:
1.190     jonathan  420:        for (i = 0; i < IPREASS_NHASH; i++)
                    421:                LIST_INIT(&ipq[i]);
                    422:
1.262     matt      423:        ip_initid();
1.227     kardel    424:        ip_id = time_second & 0xfffff;
1.194     jonathan  425:
1.1       cgd       426:        ipintrq.ifq_maxlen = ipqmaxlen;
1.194     jonathan  427:        ip_nmbclusters_changed();
                    428:
1.181     jonathan  429:        TAILQ_INIT(&in_ifaddrhead);
1.120     ad        430:        in_ifaddrhashtbl = hashinit(IN_IFADDR_HASH_SIZE, HASH_LIST, M_IFADDR,
                    431:            M_WAITOK, &in_ifaddrhash);
1.166     matt      432:        in_multihashtbl = hashinit(IN_IFADDR_HASH_SIZE, HASH_LIST, M_IPMADDR,
                    433:            M_WAITOK, &in_multihash);
1.160     itojun    434:        ip_mtudisc_timeout_q = rt_timer_queue_create(ip_mtudisc_timeout);
1.73      thorpej   435: #ifdef GATEWAY
1.248     liamjfoy  436:        ipflow_init(ip_hashsize);
1.73      thorpej   437: #endif
1.121     thorpej   438:
                    439: #ifdef PFIL_HOOKS
                    440:        /* Register our Packet Filter hook. */
1.126     thorpej   441:        inet_pfil_hook.ph_type = PFIL_TYPE_AF;
                    442:        inet_pfil_hook.ph_af   = AF_INET;
1.121     thorpej   443:        i = pfil_head_register(&inet_pfil_hook);
                    444:        if (i != 0)
                    445:                printf("ip_init: WARNING: unable to register pfil hook, "
                    446:                    "error %d\n", i);
                    447: #endif /* PFIL_HOOKS */
1.135     thorpej   448:
1.164     matt      449: #ifdef MBUFTRACE
                    450:        MOWNER_ATTACH(&ip_tx_mowner);
                    451:        MOWNER_ATTACH(&ip_rx_mowner);
                    452: #endif /* MBUFTRACE */
1.266     thorpej   453:
                    454:        ipstat_percpu = percpu_alloc(sizeof(uint64_t) * IP_NSTATS);
1.1       cgd       455: }
                    456:
1.229     christos  457: struct sockaddr_in ipaddr = {
                    458:        .sin_len = sizeof(ipaddr),
                    459:        .sin_family = AF_INET,
                    460: };
1.1       cgd       461: struct route ipforward_rt;
                    462:
                    463: /*
1.89      itojun    464:  * IP software interrupt routine
                    465:  */
                    466: void
1.211     perry     467: ipintr(void)
1.89      itojun    468: {
                    469:        int s;
                    470:        struct mbuf *m;
                    471:
1.268   ! ad        472:        mutex_enter(softnet_lock);
        !           473:        KERNEL_LOCK(1, NULL);
1.241     ad        474:        while (!IF_IS_EMPTY(&ipintrq)) {
1.132     thorpej   475:                s = splnet();
1.89      itojun    476:                IF_DEQUEUE(&ipintrq, m);
                    477:                splx(s);
1.268   ! ad        478:                if (m == NULL)
        !           479:                        break;
1.89      itojun    480:                ip_input(m);
                    481:        }
1.268   ! ad        482:        KERNEL_UNLOCK_ONE(NULL);
        !           483:        mutex_exit(softnet_lock);
1.89      itojun    484: }
                    485:
                    486: /*
1.1       cgd       487:  * Ip input routine.  Checksum and byte swap header.  If fragmented
                    488:  * try to reassemble.  Process options.  Pass to next level.
                    489:  */
1.8       mycroft   490: void
1.89      itojun    491: ip_input(struct mbuf *m)
1.1       cgd       492: {
1.109     augustss  493:        struct ip *ip = NULL;
                    494:        struct ipq *fp;
                    495:        struct in_ifaddr *ia;
                    496:        struct ifaddr *ifa;
1.25      cgd       497:        struct ipqent *ipqe;
1.89      itojun    498:        int hlen = 0, mff, len;
1.100     itojun    499:        int downmatch;
1.165     christos  500:        int checkif;
1.169     itojun    501:        int srcrt = 0;
1.233     tls       502:        int s;
1.190     jonathan  503:        u_int hash;
1.173     jonathan  504: #ifdef FAST_IPSEC
                    505:        struct m_tag *mtag;
                    506:        struct tdb_ident *tdbi;
                    507:        struct secpolicy *sp;
1.233     tls       508:        int error;
1.173     jonathan  509: #endif /* FAST_IPSEC */
1.1       cgd       510:
1.164     matt      511:        MCLAIM(m, &ip_rx_mowner);
1.1       cgd       512: #ifdef DIAGNOSTIC
                    513:        if ((m->m_flags & M_PKTHDR) == 0)
                    514:                panic("ipintr no HDR");
1.89      itojun    515: #endif
1.164     matt      516:
1.1       cgd       517:        /*
                    518:         * If no IP addresses have been set yet but the interfaces
                    519:         * are receiving, can't do anything with incoming packets yet.
                    520:         */
1.181     jonathan  521:        if (TAILQ_FIRST(&in_ifaddrhead) == 0)
1.1       cgd       522:                goto bad;
1.266     thorpej   523:        IP_STATINC(IP_STAT_TOTAL);
1.154     thorpej   524:        /*
                    525:         * If the IP header is not aligned, slurp it up into a new
                    526:         * mbuf with space for link headers, in the event we forward
                    527:         * it.  Otherwise, if it is aligned, make sure the entire
                    528:         * base IP header is in the first mbuf of the chain.
                    529:         */
1.244     christos  530:        if (IP_HDR_ALIGNED_P(mtod(m, void *)) == 0) {
1.154     thorpej   531:                if ((m = m_copyup(m, sizeof(struct ip),
                    532:                                  (max_linkhdr + 3) & ~3)) == NULL) {
                    533:                        /* XXXJRT new stat, please */
1.266     thorpej   534:                        IP_STATINC(IP_STAT_TOOSMALL);
1.154     thorpej   535:                        return;
                    536:                }
                    537:        } else if (__predict_false(m->m_len < sizeof (struct ip))) {
                    538:                if ((m = m_pullup(m, sizeof (struct ip))) == NULL) {
1.266     thorpej   539:                        IP_STATINC(IP_STAT_TOOSMALL);
1.154     thorpej   540:                        return;
                    541:                }
1.1       cgd       542:        }
                    543:        ip = mtod(m, struct ip *);
1.13      mycroft   544:        if (ip->ip_v != IPVERSION) {
1.266     thorpej   545:                IP_STATINC(IP_STAT_BADVERS);
1.13      mycroft   546:                goto bad;
                    547:        }
1.1       cgd       548:        hlen = ip->ip_hl << 2;
                    549:        if (hlen < sizeof(struct ip)) { /* minimum header length */
1.266     thorpej   550:                IP_STATINC(IP_STAT_BADHLEN);
1.1       cgd       551:                goto bad;
                    552:        }
                    553:        if (hlen > m->m_len) {
                    554:                if ((m = m_pullup(m, hlen)) == 0) {
1.266     thorpej   555:                        IP_STATINC(IP_STAT_BADHLEN);
1.89      itojun    556:                        return;
1.1       cgd       557:                }
                    558:                ip = mtod(m, struct ip *);
                    559:        }
1.98      thorpej   560:
1.85      hwr       561:        /*
1.99      thorpej   562:         * RFC1122: packets with a multicast source address are
1.98      thorpej   563:         * not allowed.
1.85      hwr       564:         */
                    565:        if (IN_MULTICAST(ip->ip_src.s_addr)) {
1.266     thorpej   566:                IP_STATINC(IP_STAT_BADADDR);
1.85      hwr       567:                goto bad;
1.129     itojun    568:        }
                    569:
                    570:        /* 127/8 must not appear on wire - RFC1122 */
                    571:        if ((ntohl(ip->ip_dst.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET ||
                    572:            (ntohl(ip->ip_src.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET) {
1.130     itojun    573:                if ((m->m_pkthdr.rcvif->if_flags & IFF_LOOPBACK) == 0) {
1.266     thorpej   574:                        IP_STATINC(IP_STAT_BADADDR);
1.129     itojun    575:                        goto bad;
1.130     itojun    576:                }
1.85      hwr       577:        }
                    578:
1.135     thorpej   579:        switch (m->m_pkthdr.csum_flags &
1.137     thorpej   580:                ((m->m_pkthdr.rcvif->if_csum_flags_rx & M_CSUM_IPv4) |
1.135     thorpej   581:                 M_CSUM_IPv4_BAD)) {
                    582:        case M_CSUM_IPv4|M_CSUM_IPv4_BAD:
                    583:                INET_CSUM_COUNTER_INCR(&ip_hwcsum_bad);
                    584:                goto badcsum;
                    585:
                    586:        case M_CSUM_IPv4:
                    587:                /* Checksum was okay. */
                    588:                INET_CSUM_COUNTER_INCR(&ip_hwcsum_ok);
                    589:                break;
                    590:
                    591:        default:
1.206     thorpej   592:                /*
                    593:                 * Must compute it ourselves.  Maybe skip checksum on
                    594:                 * loopback interfaces.
                    595:                 */
                    596:                if (__predict_true(!(m->m_pkthdr.rcvif->if_flags &
                    597:                                     IFF_LOOPBACK) || ip_do_loopback_cksum)) {
                    598:                        INET_CSUM_COUNTER_INCR(&ip_swcsum);
                    599:                        if (in_cksum(m, hlen) != 0)
                    600:                                goto badcsum;
                    601:                }
1.135     thorpej   602:                break;
1.1       cgd       603:        }
                    604:
1.121     thorpej   605:        /* Retrieve the packet length. */
                    606:        len = ntohs(ip->ip_len);
1.81      proff     607:
                    608:        /*
                    609:         * Check for additional length bogosity
                    610:         */
1.84      proff     611:        if (len < hlen) {
1.266     thorpej   612:                IP_STATINC(IP_STAT_BADLEN);
1.81      proff     613:                goto bad;
                    614:        }
1.1       cgd       615:
                    616:        /*
                    617:         * Check that the amount of data in the buffers
                    618:         * is as at least much as the IP header would have us expect.
                    619:         * Trim mbufs if longer than we expect.
                    620:         * Drop packet if shorter than we expect.
                    621:         */
1.35      mycroft   622:        if (m->m_pkthdr.len < len) {
1.266     thorpej   623:                IP_STATINC(IP_STAT_TOOSHORT);
1.1       cgd       624:                goto bad;
                    625:        }
1.35      mycroft   626:        if (m->m_pkthdr.len > len) {
1.1       cgd       627:                if (m->m_len == m->m_pkthdr.len) {
1.35      mycroft   628:                        m->m_len = len;
                    629:                        m->m_pkthdr.len = len;
1.1       cgd       630:                } else
1.35      mycroft   631:                        m_adj(m, len - m->m_pkthdr.len);
1.1       cgd       632:        }
                    633:
1.193     scw       634: #if defined(IPSEC)
1.149     wiz       635:        /* ipflow (IP fast forwarding) is not compatible with IPsec. */
1.94      itojun    636:        m->m_flags &= ~M_CANFASTFWD;
                    637: #else
1.64      thorpej   638:        /*
                    639:         * Assume that we can create a fast-forward IP flow entry
                    640:         * based on this packet.
                    641:         */
                    642:        m->m_flags |= M_CANFASTFWD;
1.94      itojun    643: #endif
1.64      thorpej   644:
1.36      mrg       645: #ifdef PFIL_HOOKS
1.33      mrg       646:        /*
1.64      thorpej   647:         * Run through list of hooks for input packets.  If there are any
                    648:         * filters which require that additional packets in the flow are
                    649:         * not fast-forwarded, they must clear the M_CANFASTFWD flag.
                    650:         * Note that filters must _never_ set this flag, as another filter
                    651:         * in the list may have previously cleared it.
1.33      mrg       652:         */
1.127     itojun    653:        /*
                    654:         * let ipfilter look at packet on the wire,
                    655:         * not the decapsulated packet.
                    656:         */
                    657: #ifdef IPSEC
1.136     itojun    658:        if (!ipsec_getnhist(m))
1.186     scw       659: #elif defined(FAST_IPSEC)
                    660:        if (!ipsec_indone(m))
1.127     itojun    661: #else
                    662:        if (1)
                    663: #endif
                    664:        {
1.169     itojun    665:                struct in_addr odst;
                    666:
                    667:                odst = ip->ip_dst;
1.127     itojun    668:                if (pfil_run_hooks(&inet_pfil_hook, &m, m->m_pkthdr.rcvif,
1.168     itojun    669:                    PFIL_IN) != 0)
                    670:                        return;
1.127     itojun    671:                if (m == NULL)
                    672:                        return;
                    673:                ip = mtod(m, struct ip *);
1.142     darrenr   674:                hlen = ip->ip_hl << 2;
1.205     darrenr   675:                /*
                    676:                 * XXX The setting of "srcrt" here is to prevent ip_forward()
                    677:                 * from generating ICMP redirects for packets that have
                    678:                 * been redirected by a hook back out on to the same LAN that
                    679:                 * they came from and is not an indication that the packet
                    680:                 * is being inffluenced by source routing options.  This
                    681:                 * allows things like
                    682:                 * "rdr tlp0 0/0 port 80 -> 1.1.1.200 3128 tcp"
                    683:                 * where tlp0 is both on the 1.1.1.0/24 network and is the
                    684:                 * default route for hosts on 1.1.1.0/24.  Of course this
                    685:                 * also requires a "map tlp0 ..." to complete the story.
                    686:                 * One might argue whether or not this kind of network config.
1.212     perry     687:                 * should be supported in this manner...
1.205     darrenr   688:                 */
1.169     itojun    689:                srcrt = (odst.s_addr != ip->ip_dst.s_addr);
1.127     itojun    690:        }
1.36      mrg       691: #endif /* PFIL_HOOKS */
1.123     thorpej   692:
                    693: #ifdef ALTQ
                    694:        /* XXX Temporary until ALTQ is changed to use a pfil hook */
                    695:        if (altq_input != NULL && (*altq_input)(m, AF_INET) == 0) {
                    696:                /* packet dropped by traffic conditioner */
                    697:                return;
                    698:        }
                    699: #endif
1.121     thorpej   700:
                    701:        /*
1.1       cgd       702:         * Process options and, if not destined for us,
                    703:         * ship it on.  ip_dooptions returns 1 when an
                    704:         * error was detected (causing an icmp message
                    705:         * to be sent and the original packet to be freed).
                    706:         */
                    707:        ip_nhops = 0;           /* for source routed packets */
                    708:        if (hlen > sizeof (struct ip) && ip_dooptions(m))
1.89      itojun    709:                return;
1.1       cgd       710:
                    711:        /*
1.165     christos  712:         * Enable a consistency check between the destination address
                    713:         * and the arrival interface for a unicast packet (the RFC 1122
                    714:         * strong ES model) if IP forwarding is disabled and the packet
                    715:         * is not locally generated.
                    716:         *
                    717:         * XXX - Checking also should be disabled if the destination
                    718:         * address is ipnat'ed to a different interface.
                    719:         *
                    720:         * XXX - Checking is incompatible with IP aliases added
                    721:         * to the loopback interface instead of the interface where
                    722:         * the packets are received.
                    723:         *
                    724:         * XXX - We need to add a per ifaddr flag for this so that
                    725:         * we get finer grain control.
                    726:         */
                    727:        checkif = ip_checkinterface && (ipforwarding == 0) &&
                    728:            (m->m_pkthdr.rcvif != NULL) &&
                    729:            ((m->m_pkthdr.rcvif->if_flags & IFF_LOOPBACK) == 0);
                    730:
                    731:        /*
1.1       cgd       732:         * Check our list of addresses, to see if the packet is for us.
1.100     itojun    733:         *
                    734:         * Traditional 4.4BSD did not consult IFF_UP at all.
                    735:         * The behavior here is to treat addresses on !IFF_UP interface
                    736:         * as not mine.
1.1       cgd       737:         */
1.100     itojun    738:        downmatch = 0;
1.140     matt      739:        LIST_FOREACH(ia, &IN_IFADDR_HASH(ip->ip_dst.s_addr), ia_hash) {
1.97      itojun    740:                if (in_hosteq(ia->ia_addr.sin_addr, ip->ip_dst)) {
1.165     christos  741:                        if (checkif && ia->ia_ifp != m->m_pkthdr.rcvif)
                    742:                                continue;
1.97      itojun    743:                        if ((ia->ia_ifp->if_flags & IFF_UP) != 0)
                    744:                                break;
1.100     itojun    745:                        else
                    746:                                downmatch++;
1.97      itojun    747:                }
                    748:        }
1.86      thorpej   749:        if (ia != NULL)
                    750:                goto ours;
1.225     christos  751:        if (m->m_pkthdr.rcvif && m->m_pkthdr.rcvif->if_flags & IFF_BROADCAST) {
1.209     matt      752:                IFADDR_FOREACH(ifa, m->m_pkthdr.rcvif) {
1.140     matt      753:                        if (ifa->ifa_addr->sa_family != AF_INET)
                    754:                                continue;
1.57      tls       755:                        ia = ifatoia(ifa);
1.35      mycroft   756:                        if (in_hosteq(ip->ip_dst, ia->ia_broadaddr.sin_addr) ||
                    757:                            in_hosteq(ip->ip_dst, ia->ia_netbroadcast) ||
1.20      mycroft   758:                            /*
                    759:                             * Look for all-0's host part (old broadcast addr),
                    760:                             * either for subnet or net.
                    761:                             */
                    762:                            ip->ip_dst.s_addr == ia->ia_subnet ||
1.18      mycroft   763:                            ip->ip_dst.s_addr == ia->ia_net)
1.1       cgd       764:                                goto ours;
1.57      tls       765:                        /*
                    766:                         * An interface with IP address zero accepts
                    767:                         * all packets that arrive on that interface.
                    768:                         */
                    769:                        if (in_nullhost(ia->ia_addr.sin_addr))
                    770:                                goto ours;
1.1       cgd       771:                }
                    772:        }
1.18      mycroft   773:        if (IN_MULTICAST(ip->ip_dst.s_addr)) {
1.4       hpeyerl   774:                struct in_multi *inm;
                    775: #ifdef MROUTING
                    776:                extern struct socket *ip_mrouter;
1.10      brezak    777:
1.4       hpeyerl   778:                if (ip_mrouter) {
                    779:                        /*
                    780:                         * If we are acting as a multicast router, all
                    781:                         * incoming multicast packets are passed to the
                    782:                         * kernel-level multicast forwarding function.
                    783:                         * The packet is returned (relatively) intact; if
                    784:                         * ip_mforward() returns a non-zero value, the packet
                    785:                         * must be discarded, else it may be accepted below.
                    786:                         *
                    787:                         * (The IP ident field is put in the same byte order
                    788:                         * as expected when ip_mforward() is called from
                    789:                         * ip_output().)
                    790:                         */
1.13      mycroft   791:                        if (ip_mforward(m, m->m_pkthdr.rcvif) != 0) {
1.266     thorpej   792:                                IP_STATINC(IP_STAT_CANTFORWARD);
1.4       hpeyerl   793:                                m_freem(m);
1.89      itojun    794:                                return;
1.4       hpeyerl   795:                        }
                    796:
                    797:                        /*
                    798:                         * The process-level routing demon needs to receive
                    799:                         * all multicast IGMP packets, whether or not this
                    800:                         * host belongs to their destination groups.
                    801:                         */
                    802:                        if (ip->ip_p == IPPROTO_IGMP)
                    803:                                goto ours;
1.266     thorpej   804:                        IP_STATINC(IP_STAT_CANTFORWARD);
1.4       hpeyerl   805:                }
                    806: #endif
                    807:                /*
                    808:                 * See if we belong to the destination multicast group on the
                    809:                 * arrival interface.
                    810:                 */
                    811:                IN_LOOKUP_MULTI(ip->ip_dst, m->m_pkthdr.rcvif, inm);
                    812:                if (inm == NULL) {
1.266     thorpej   813:                        IP_STATINC(IP_STAT_CANTFORWARD);
1.4       hpeyerl   814:                        m_freem(m);
1.89      itojun    815:                        return;
1.4       hpeyerl   816:                }
                    817:                goto ours;
                    818:        }
1.19      mycroft   819:        if (ip->ip_dst.s_addr == INADDR_BROADCAST ||
1.35      mycroft   820:            in_nullhost(ip->ip_dst))
1.1       cgd       821:                goto ours;
                    822:
                    823:        /*
                    824:         * Not for us; forward if possible and desirable.
                    825:         */
                    826:        if (ipforwarding == 0) {
1.266     thorpej   827:                IP_STATINC(IP_STAT_CANTFORWARD);
1.1       cgd       828:                m_freem(m);
1.100     itojun    829:        } else {
                    830:                /*
                    831:                 * If ip_dst matched any of my address on !IFF_UP interface,
                    832:                 * and there's no IFF_UP interface that matches ip_dst,
                    833:                 * send icmp unreach.  Forwarding it will result in in-kernel
                    834:                 * forwarding loop till TTL goes to 0.
                    835:                 */
                    836:                if (downmatch) {
                    837:                        icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_HOST, 0, 0);
1.266     thorpej   838:                        IP_STATINC(IP_STAT_CANTFORWARD);
1.100     itojun    839:                        return;
                    840:                }
1.145     itojun    841: #ifdef IPSEC
                    842:                if (ipsec4_in_reject(m, NULL)) {
1.267     thorpej   843:                        IPSEC_STATINC(IPSEC_STAT_IN_POLVIO);
1.145     itojun    844:                        goto bad;
                    845:                }
                    846: #endif
1.173     jonathan  847: #ifdef FAST_IPSEC
                    848:                mtag = m_tag_find(m, PACKET_TAG_IPSEC_IN_DONE, NULL);
                    849:                s = splsoftnet();
                    850:                if (mtag != NULL) {
                    851:                        tdbi = (struct tdb_ident *)(mtag + 1);
                    852:                        sp = ipsec_getpolicy(tdbi, IPSEC_DIR_INBOUND);
                    853:                } else {
                    854:                        sp = ipsec_getpolicybyaddr(m, IPSEC_DIR_INBOUND,
1.212     perry     855:                                                   IP_FORWARDING, &error);
1.173     jonathan  856:                }
                    857:                if (sp == NULL) {       /* NB: can happen if error */
                    858:                        splx(s);
                    859:                        /*XXX error stat???*/
                    860:                        DPRINTF(("ip_input: no SP for forwarding\n"));  /*XXX*/
                    861:                        goto bad;
                    862:                }
                    863:
                    864:                /*
                    865:                 * Check security policy against packet attributes.
                    866:                 */
                    867:                error = ipsec_in_reject(sp, m);
                    868:                KEY_FREESP(&sp);
                    869:                splx(s);
                    870:                if (error) {
1.266     thorpej   871:                        IP_STATINC(IP_STAT_CANTFORWARD);
1.173     jonathan  872:                        goto bad;
1.193     scw       873:                }
                    874:
                    875:                /*
                    876:                 * Peek at the outbound SP for this packet to determine if
                    877:                 * it's a Fast Forward candidate.
                    878:                 */
                    879:                mtag = m_tag_find(m, PACKET_TAG_IPSEC_PENDING_TDB, NULL);
                    880:                if (mtag != NULL)
                    881:                        m->m_flags &= ~M_CANFASTFWD;
                    882:                else {
                    883:                        s = splsoftnet();
                    884:                        sp = ipsec4_checkpolicy(m, IPSEC_DIR_OUTBOUND,
                    885:                            (IP_FORWARDING |
                    886:                             (ip_directedbcast ? IP_ALLOWBROADCAST : 0)),
                    887:                            &error, NULL);
                    888:                        if (sp != NULL) {
                    889:                                m->m_flags &= ~M_CANFASTFWD;
                    890:                                KEY_FREESP(&sp);
                    891:                        }
                    892:                        splx(s);
1.173     jonathan  893:                }
                    894: #endif /* FAST_IPSEC */
1.145     itojun    895:
1.169     itojun    896:                ip_forward(m, srcrt);
1.100     itojun    897:        }
1.89      itojun    898:        return;
1.1       cgd       899:
                    900: ours:
                    901:        /*
                    902:         * If offset or IP_MF are set, must reassemble.
                    903:         * Otherwise, nothing need be done.
                    904:         * (We could look in the reassembly queue to see
                    905:         * if the packet was previously fragmented,
                    906:         * but it's not worth the time; just let them time out.)
                    907:         */
1.155     itojun    908:        if (ip->ip_off & ~htons(IP_DF|IP_RF)) {
1.258     matt      909:                uint16_t off;
                    910:                /*
                    911:                 * Prevent TCP blind data attacks by not allowing non-initial
                    912:                 * fragments to start at less than 68 bytes (minimal fragment
1.259     matt      913:                 * size) and making sure the first fragment is at least 68
                    914:                 * bytes.
1.258     matt      915:                 */
1.260     matt      916:                off = (ntohs(ip->ip_off) & IP_OFFMASK) << 3;
1.259     matt      917:                if ((off > 0 ? off + hlen : len) < IP_MINFRAGSIZE - 1) {
1.266     thorpej   918:                        IP_STATINC(IP_STAT_BADFRAGS);
1.258     matt      919:                        goto bad;
                    920:                }
1.1       cgd       921:                /*
                    922:                 * Look for queue of fragments
                    923:                 * of this datagram.
                    924:                 */
1.75      thorpej   925:                IPQ_LOCK();
1.190     jonathan  926:                hash = IPREASS_HASH(ip->ip_src.s_addr, ip->ip_id);
1.250     dyoung    927:                LIST_FOREACH(fp, &ipq[hash], ipq_q) {
1.1       cgd       928:                        if (ip->ip_id == fp->ipq_id &&
1.35      mycroft   929:                            in_hosteq(ip->ip_src, fp->ipq_src) &&
                    930:                            in_hosteq(ip->ip_dst, fp->ipq_dst) &&
1.260     matt      931:                            ip->ip_p == fp->ipq_p) {
                    932:                                /*
                    933:                                 * Make sure the TOS is matches previous
                    934:                                 * fragments.
                    935:                                 */
                    936:                                if (ip->ip_tos != fp->ipq_tos) {
1.266     thorpej   937:                                        IP_STATINC(IP_STAT_BADFRAGS);
1.260     matt      938:                                        goto bad;
                    939:                                }
1.1       cgd       940:                                goto found;
1.260     matt      941:                        }
1.190     jonathan  942:                }
1.1       cgd       943:                fp = 0;
                    944: found:
                    945:
                    946:                /*
                    947:                 * Adjust ip_len to not reflect header,
1.25      cgd       948:                 * set ipqe_mff if more fragments are expected,
1.1       cgd       949:                 * convert offset of this to bytes.
                    950:                 */
1.155     itojun    951:                ip->ip_len = htons(ntohs(ip->ip_len) - hlen);
                    952:                mff = (ip->ip_off & htons(IP_MF)) != 0;
1.25      cgd       953:                if (mff) {
1.16      cgd       954:                        /*
                    955:                         * Make sure that fragments have a data length
                    956:                         * that's a non-zero multiple of 8 bytes.
                    957:                         */
1.155     itojun    958:                        if (ntohs(ip->ip_len) == 0 ||
                    959:                            (ntohs(ip->ip_len) & 0x7) != 0) {
1.266     thorpej   960:                                IP_STATINC(IP_STAT_BADFRAGS);
1.75      thorpej   961:                                IPQ_UNLOCK();
1.16      cgd       962:                                goto bad;
                    963:                        }
                    964:                }
1.155     itojun    965:                ip->ip_off = htons((ntohs(ip->ip_off) & IP_OFFMASK) << 3);
1.1       cgd       966:
                    967:                /*
                    968:                 * If datagram marked as having more fragments
                    969:                 * or if this is not the first fragment,
                    970:                 * attempt reassembly; if it succeeds, proceed.
                    971:                 */
1.155     itojun    972:                if (mff || ip->ip_off != htons(0)) {
1.266     thorpej   973:                        IP_STATINC(IP_STAT_FRAGMENTS);
1.233     tls       974:                        s = splvm();
1.72      thorpej   975:                        ipqe = pool_get(&ipqent_pool, PR_NOWAIT);
1.233     tls       976:                        splx(s);
1.25      cgd       977:                        if (ipqe == NULL) {
1.266     thorpej   978:                                IP_STATINC(IP_STAT_RCVMEMDROP);
1.75      thorpej   979:                                IPQ_UNLOCK();
1.25      cgd       980:                                goto bad;
                    981:                        }
                    982:                        ipqe->ipqe_mff = mff;
1.50      thorpej   983:                        ipqe->ipqe_m = m;
1.25      cgd       984:                        ipqe->ipqe_ip = ip;
1.190     jonathan  985:                        m = ip_reass(ipqe, fp, &ipq[hash]);
1.75      thorpej   986:                        if (m == 0) {
                    987:                                IPQ_UNLOCK();
1.89      itojun    988:                                return;
1.75      thorpej   989:                        }
1.266     thorpej   990:                        IP_STATINC(IP_STAT_REASSEMBLED);
1.50      thorpej   991:                        ip = mtod(m, struct ip *);
1.74      thorpej   992:                        hlen = ip->ip_hl << 2;
1.155     itojun    993:                        ip->ip_len = htons(ntohs(ip->ip_len) + hlen);
1.1       cgd       994:                } else
                    995:                        if (fp)
                    996:                                ip_freef(fp);
1.75      thorpej   997:                IPQ_UNLOCK();
1.79      mycroft   998:        }
1.128     itojun    999:
1.173     jonathan 1000: #if defined(IPSEC)
1.128     itojun   1001:        /*
                   1002:         * enforce IPsec policy checking if we are seeing last header.
                   1003:         * note that we do not visit this with protocols with pcb layer
                   1004:         * code - like udp/tcp/raw ip.
                   1005:         */
                   1006:        if ((inetsw[ip_protox[ip->ip_p]].pr_flags & PR_LASTHDR) != 0 &&
                   1007:            ipsec4_in_reject(m, NULL)) {
1.267     thorpej  1008:                IPSEC_STATINC(IPSEC_STAT_IN_POLVIO);
1.128     itojun   1009:                goto bad;
                   1010:        }
                   1011: #endif
1.226     liamjfoy 1012: #ifdef FAST_IPSEC
1.173     jonathan 1013:        /*
                   1014:         * enforce IPsec policy checking if we are seeing last header.
                   1015:         * note that we do not visit this with protocols with pcb layer
                   1016:         * code - like udp/tcp/raw ip.
                   1017:         */
                   1018:        if ((inetsw[ip_protox[ip->ip_p]].pr_flags & PR_LASTHDR) != 0) {
                   1019:                /*
                   1020:                 * Check if the packet has already had IPsec processing
                   1021:                 * done.  If so, then just pass it along.  This tag gets
                   1022:                 * set during AH, ESP, etc. input handling, before the
                   1023:                 * packet is returned to the ip input queue for delivery.
1.212     perry    1024:                 */
1.173     jonathan 1025:                mtag = m_tag_find(m, PACKET_TAG_IPSEC_IN_DONE, NULL);
                   1026:                s = splsoftnet();
                   1027:                if (mtag != NULL) {
                   1028:                        tdbi = (struct tdb_ident *)(mtag + 1);
                   1029:                        sp = ipsec_getpolicy(tdbi, IPSEC_DIR_INBOUND);
                   1030:                } else {
                   1031:                        sp = ipsec_getpolicybyaddr(m, IPSEC_DIR_INBOUND,
1.212     perry    1032:                                                   IP_FORWARDING, &error);
1.173     jonathan 1033:                }
                   1034:                if (sp != NULL) {
                   1035:                        /*
                   1036:                         * Check security policy against packet attributes.
                   1037:                         */
                   1038:                        error = ipsec_in_reject(sp, m);
                   1039:                        KEY_FREESP(&sp);
                   1040:                } else {
                   1041:                        /* XXX error stat??? */
                   1042:                        error = EINVAL;
                   1043: DPRINTF(("ip_input: no SP, packet discarded\n"));/*XXX*/
                   1044:                }
                   1045:                splx(s);
                   1046:                if (error)
                   1047:                        goto bad;
                   1048:        }
                   1049: #endif /* FAST_IPSEC */
1.1       cgd      1050:
                   1051:        /*
                   1052:         * Switch out to protocol's input routine.
                   1053:         */
1.82      aidan    1054: #if IFA_STATS
1.122     itojun   1055:        if (ia && ip)
1.155     itojun   1056:                ia->ia_ifa.ifa_data.ifad_inbytes += ntohs(ip->ip_len);
1.82      aidan    1057: #endif
1.266     thorpej  1058:        IP_STATINC(IP_STAT_DELIVERED);
1.89      itojun   1059:     {
                   1060:        int off = hlen, nh = ip->ip_p;
                   1061:
                   1062:        (*inetsw[ip_protox[nh]].pr_input)(m, off, nh);
                   1063:        return;
                   1064:     }
1.1       cgd      1065: bad:
                   1066:        m_freem(m);
1.135     thorpej  1067:        return;
                   1068:
                   1069: badcsum:
1.266     thorpej  1070:        IP_STATINC(IP_STAT_BADSUM);
1.135     thorpej  1071:        m_freem(m);
1.1       cgd      1072: }
                   1073:
                   1074: /*
                   1075:  * Take incoming datagram fragment and try to
                   1076:  * reassemble it into whole datagram.  If a chain for
                   1077:  * reassembly of this datagram already exists, then it
                   1078:  * is given as fp; otherwise have to make a chain.
                   1079:  */
1.50      thorpej  1080: struct mbuf *
1.211     perry    1081: ip_reass(struct ipqent *ipqe, struct ipq *fp, struct ipqhead *ipqhead)
1.1       cgd      1082: {
1.109     augustss 1083:        struct mbuf *m = ipqe->ipqe_m;
                   1084:        struct ipqent *nq, *p, *q;
1.25      cgd      1085:        struct ip *ip;
1.1       cgd      1086:        struct mbuf *t;
1.25      cgd      1087:        int hlen = ipqe->ipqe_ip->ip_hl << 2;
1.233     tls      1088:        int i, next, s;
1.1       cgd      1089:
1.75      thorpej  1090:        IPQ_LOCK_CHECK();
                   1091:
1.1       cgd      1092:        /*
                   1093:         * Presence of header sizes in mbufs
                   1094:         * would confuse code below.
                   1095:         */
                   1096:        m->m_data += hlen;
                   1097:        m->m_len -= hlen;
                   1098:
1.194     jonathan 1099: #ifdef notyet
                   1100:        /* make sure fragment limit is up-to-date */
                   1101:        CHECK_NMBCLUSTER_PARAMS();
                   1102:
                   1103:        /* If we have too many fragments, drop the older half. */
                   1104:        if (ip_nfrags >= ip_maxfrags)
                   1105:                ip_reass_drophalf(void);
                   1106: #endif
                   1107:
1.1       cgd      1108:        /*
1.192     jonathan 1109:         * We are about to add a fragment; increment frag count.
                   1110:         */
                   1111:        ip_nfrags++;
1.212     perry    1112:
1.192     jonathan 1113:        /*
1.1       cgd      1114:         * If first fragment to arrive, create a reassembly queue.
                   1115:         */
                   1116:        if (fp == 0) {
1.131     itojun   1117:                /*
                   1118:                 * Enforce upper bound on number of fragmented packets
                   1119:                 * for which we attempt reassembly;
                   1120:                 * If maxfrag is 0, never accept fragments.
                   1121:                 * If maxfrag is -1, accept all fragments without limitation.
                   1122:                 */
                   1123:                if (ip_maxfragpackets < 0)
                   1124:                        ;
                   1125:                else if (ip_nfragpackets >= ip_maxfragpackets)
                   1126:                        goto dropfrag;
                   1127:                ip_nfragpackets++;
1.50      thorpej  1128:                MALLOC(fp, struct ipq *, sizeof (struct ipq),
                   1129:                    M_FTABLE, M_NOWAIT);
                   1130:                if (fp == NULL)
1.1       cgd      1131:                        goto dropfrag;
1.190     jonathan 1132:                LIST_INSERT_HEAD(ipqhead, fp, ipq_q);
1.192     jonathan 1133:                fp->ipq_nfrags = 1;
1.1       cgd      1134:                fp->ipq_ttl = IPFRAGTTL;
1.25      cgd      1135:                fp->ipq_p = ipqe->ipqe_ip->ip_p;
                   1136:                fp->ipq_id = ipqe->ipqe_ip->ip_id;
1.260     matt     1137:                fp->ipq_tos = ipqe->ipqe_ip->ip_tos;
1.148     matt     1138:                TAILQ_INIT(&fp->ipq_fragq);
1.25      cgd      1139:                fp->ipq_src = ipqe->ipqe_ip->ip_src;
                   1140:                fp->ipq_dst = ipqe->ipqe_ip->ip_dst;
                   1141:                p = NULL;
1.1       cgd      1142:                goto insert;
1.192     jonathan 1143:        } else {
                   1144:                fp->ipq_nfrags++;
1.1       cgd      1145:        }
                   1146:
                   1147:        /*
                   1148:         * Find a segment which begins after this one does.
                   1149:         */
1.148     matt     1150:        for (p = NULL, q = TAILQ_FIRST(&fp->ipq_fragq); q != NULL;
                   1151:            p = q, q = TAILQ_NEXT(q, ipqe_q))
1.155     itojun   1152:                if (ntohs(q->ipqe_ip->ip_off) > ntohs(ipqe->ipqe_ip->ip_off))
1.1       cgd      1153:                        break;
                   1154:
                   1155:        /*
                   1156:         * If there is a preceding segment, it may provide some of
                   1157:         * our data already.  If so, drop the data from the incoming
                   1158:         * segment.  If it provides all of our data, drop us.
                   1159:         */
1.25      cgd      1160:        if (p != NULL) {
1.155     itojun   1161:                i = ntohs(p->ipqe_ip->ip_off) + ntohs(p->ipqe_ip->ip_len) -
                   1162:                    ntohs(ipqe->ipqe_ip->ip_off);
1.1       cgd      1163:                if (i > 0) {
1.155     itojun   1164:                        if (i >= ntohs(ipqe->ipqe_ip->ip_len))
1.1       cgd      1165:                                goto dropfrag;
1.50      thorpej  1166:                        m_adj(ipqe->ipqe_m, i);
1.155     itojun   1167:                        ipqe->ipqe_ip->ip_off =
                   1168:                            htons(ntohs(ipqe->ipqe_ip->ip_off) + i);
                   1169:                        ipqe->ipqe_ip->ip_len =
                   1170:                            htons(ntohs(ipqe->ipqe_ip->ip_len) - i);
1.1       cgd      1171:                }
                   1172:        }
                   1173:
                   1174:        /*
                   1175:         * While we overlap succeeding segments trim them or,
                   1176:         * if they are completely covered, dequeue them.
                   1177:         */
1.155     itojun   1178:        for (; q != NULL &&
                   1179:            ntohs(ipqe->ipqe_ip->ip_off) + ntohs(ipqe->ipqe_ip->ip_len) >
                   1180:            ntohs(q->ipqe_ip->ip_off); q = nq) {
                   1181:                i = (ntohs(ipqe->ipqe_ip->ip_off) +
                   1182:                    ntohs(ipqe->ipqe_ip->ip_len)) - ntohs(q->ipqe_ip->ip_off);
                   1183:                if (i < ntohs(q->ipqe_ip->ip_len)) {
                   1184:                        q->ipqe_ip->ip_len =
                   1185:                            htons(ntohs(q->ipqe_ip->ip_len) - i);
                   1186:                        q->ipqe_ip->ip_off =
                   1187:                            htons(ntohs(q->ipqe_ip->ip_off) + i);
1.50      thorpej  1188:                        m_adj(q->ipqe_m, i);
1.1       cgd      1189:                        break;
                   1190:                }
1.148     matt     1191:                nq = TAILQ_NEXT(q, ipqe_q);
1.50      thorpej  1192:                m_freem(q->ipqe_m);
1.148     matt     1193:                TAILQ_REMOVE(&fp->ipq_fragq, q, ipqe_q);
1.233     tls      1194:                s = splvm();
1.72      thorpej  1195:                pool_put(&ipqent_pool, q);
1.233     tls      1196:                splx(s);
1.192     jonathan 1197:                fp->ipq_nfrags--;
                   1198:                ip_nfrags--;
1.1       cgd      1199:        }
                   1200:
                   1201: insert:
                   1202:        /*
                   1203:         * Stick new segment in its place;
                   1204:         * check for complete reassembly.
                   1205:         */
1.25      cgd      1206:        if (p == NULL) {
1.148     matt     1207:                TAILQ_INSERT_HEAD(&fp->ipq_fragq, ipqe, ipqe_q);
1.25      cgd      1208:        } else {
1.148     matt     1209:                TAILQ_INSERT_AFTER(&fp->ipq_fragq, p, ipqe, ipqe_q);
1.25      cgd      1210:        }
1.1       cgd      1211:        next = 0;
1.148     matt     1212:        for (p = NULL, q = TAILQ_FIRST(&fp->ipq_fragq); q != NULL;
                   1213:            p = q, q = TAILQ_NEXT(q, ipqe_q)) {
1.155     itojun   1214:                if (ntohs(q->ipqe_ip->ip_off) != next)
1.1       cgd      1215:                        return (0);
1.155     itojun   1216:                next += ntohs(q->ipqe_ip->ip_len);
1.1       cgd      1217:        }
1.25      cgd      1218:        if (p->ipqe_mff)
1.1       cgd      1219:                return (0);
                   1220:
                   1221:        /*
1.41      thorpej  1222:         * Reassembly is complete.  Check for a bogus message size and
                   1223:         * concatenate fragments.
1.1       cgd      1224:         */
1.148     matt     1225:        q = TAILQ_FIRST(&fp->ipq_fragq);
1.25      cgd      1226:        ip = q->ipqe_ip;
1.41      thorpej  1227:        if ((next + (ip->ip_hl << 2)) > IP_MAXPACKET) {
1.266     thorpej  1228:                IP_STATINC(IP_STAT_TOOLONG);
1.41      thorpej  1229:                ip_freef(fp);
                   1230:                return (0);
                   1231:        }
1.50      thorpej  1232:        m = q->ipqe_m;
1.1       cgd      1233:        t = m->m_next;
                   1234:        m->m_next = 0;
                   1235:        m_cat(m, t);
1.148     matt     1236:        nq = TAILQ_NEXT(q, ipqe_q);
1.233     tls      1237:        s = splvm();
1.72      thorpej  1238:        pool_put(&ipqent_pool, q);
1.233     tls      1239:        splx(s);
1.25      cgd      1240:        for (q = nq; q != NULL; q = nq) {
1.50      thorpej  1241:                t = q->ipqe_m;
1.148     matt     1242:                nq = TAILQ_NEXT(q, ipqe_q);
1.233     tls      1243:                s = splvm();
1.72      thorpej  1244:                pool_put(&ipqent_pool, q);
1.233     tls      1245:                splx(s);
1.1       cgd      1246:                m_cat(m, t);
                   1247:        }
1.192     jonathan 1248:        ip_nfrags -= fp->ipq_nfrags;
1.1       cgd      1249:
                   1250:        /*
                   1251:         * Create header for new ip packet by
                   1252:         * modifying header of first packet;
                   1253:         * dequeue and discard fragment reassembly header.
                   1254:         * Make header visible.
                   1255:         */
1.155     itojun   1256:        ip->ip_len = htons(next);
1.25      cgd      1257:        ip->ip_src = fp->ipq_src;
                   1258:        ip->ip_dst = fp->ipq_dst;
                   1259:        LIST_REMOVE(fp, ipq_q);
1.50      thorpej  1260:        FREE(fp, M_FTABLE);
1.131     itojun   1261:        ip_nfragpackets--;
1.1       cgd      1262:        m->m_len += (ip->ip_hl << 2);
                   1263:        m->m_data -= (ip->ip_hl << 2);
                   1264:        /* some debugging cruft by sklower, below, will go away soon */
                   1265:        if (m->m_flags & M_PKTHDR) { /* XXX this should be done elsewhere */
1.109     augustss 1266:                int plen = 0;
1.50      thorpej  1267:                for (t = m; t; t = t->m_next)
                   1268:                        plen += t->m_len;
                   1269:                m->m_pkthdr.len = plen;
1.213     yamt     1270:                m->m_pkthdr.csum_flags = 0;
1.1       cgd      1271:        }
1.50      thorpej  1272:        return (m);
1.1       cgd      1273:
                   1274: dropfrag:
1.192     jonathan 1275:        if (fp != 0)
                   1276:                fp->ipq_nfrags--;
                   1277:        ip_nfrags--;
1.266     thorpej  1278:        IP_STATINC(IP_STAT_FRAGDROPPED);
1.1       cgd      1279:        m_freem(m);
1.233     tls      1280:        s = splvm();
1.72      thorpej  1281:        pool_put(&ipqent_pool, ipqe);
1.233     tls      1282:        splx(s);
1.1       cgd      1283:        return (0);
                   1284: }
                   1285:
                   1286: /*
                   1287:  * Free a fragment reassembly header and all
                   1288:  * associated datagrams.
                   1289:  */
1.8       mycroft  1290: void
1.211     perry    1291: ip_freef(struct ipq *fp)
1.1       cgd      1292: {
1.109     augustss 1293:        struct ipqent *q, *p;
1.192     jonathan 1294:        u_int nfrags = 0;
1.233     tls      1295:        int s;
1.1       cgd      1296:
1.75      thorpej  1297:        IPQ_LOCK_CHECK();
                   1298:
1.148     matt     1299:        for (q = TAILQ_FIRST(&fp->ipq_fragq); q != NULL; q = p) {
                   1300:                p = TAILQ_NEXT(q, ipqe_q);
1.50      thorpej  1301:                m_freem(q->ipqe_m);
1.192     jonathan 1302:                nfrags++;
1.148     matt     1303:                TAILQ_REMOVE(&fp->ipq_fragq, q, ipqe_q);
1.233     tls      1304:                s = splvm();
1.72      thorpej  1305:                pool_put(&ipqent_pool, q);
1.233     tls      1306:                splx(s);
1.1       cgd      1307:        }
1.192     jonathan 1308:
                   1309:        if (nfrags != fp->ipq_nfrags)
                   1310:            printf("ip_freef: nfrags %d != %d\n", fp->ipq_nfrags, nfrags);
                   1311:        ip_nfrags -= nfrags;
1.25      cgd      1312:        LIST_REMOVE(fp, ipq_q);
1.50      thorpej  1313:        FREE(fp, M_FTABLE);
1.131     itojun   1314:        ip_nfragpackets--;
1.1       cgd      1315: }
                   1316:
                   1317: /*
1.194     jonathan 1318:  * IP reassembly TTL machinery for  multiplicative drop.
                   1319:  */
                   1320: static u_int   fragttl_histo[(IPFRAGTTL+1)];
                   1321:
                   1322:
                   1323: /*
                   1324:  * Decrement TTL of all reasembly queue entries by `ticks'.
                   1325:  * Count number of distinct fragments (as opposed to partial, fragmented
                   1326:  * datagrams) in the reassembly queue.  While we  traverse the entire
                   1327:  * reassembly queue, compute and return the median TTL over all fragments.
                   1328:  */
                   1329: static u_int
                   1330: ip_reass_ttl_decr(u_int ticks)
                   1331: {
1.198     matt     1332:        u_int nfrags, median, dropfraction, keepfraction;
1.194     jonathan 1333:        struct ipq *fp, *nfp;
1.198     matt     1334:        int i;
1.212     perry    1335:
1.194     jonathan 1336:        nfrags = 0;
                   1337:        memset(fragttl_histo, 0, sizeof fragttl_histo);
1.212     perry    1338:
1.194     jonathan 1339:        for (i = 0; i < IPREASS_NHASH; i++) {
                   1340:                for (fp = LIST_FIRST(&ipq[i]); fp != NULL; fp = nfp) {
                   1341:                        fp->ipq_ttl = ((fp->ipq_ttl  <= ticks) ?
                   1342:                                       0 : fp->ipq_ttl - ticks);
                   1343:                        nfp = LIST_NEXT(fp, ipq_q);
                   1344:                        if (fp->ipq_ttl == 0) {
1.266     thorpej  1345:                                IP_STATINC(IP_STAT_FRAGTIMEOUT);
1.194     jonathan 1346:                                ip_freef(fp);
                   1347:                        } else {
                   1348:                                nfrags += fp->ipq_nfrags;
                   1349:                                fragttl_histo[fp->ipq_ttl] += fp->ipq_nfrags;
                   1350:                        }
                   1351:                }
                   1352:        }
                   1353:
                   1354:        KASSERT(ip_nfrags == nfrags);
                   1355:
                   1356:        /* Find median (or other drop fraction) in histogram. */
                   1357:        dropfraction = (ip_nfrags / 2);
                   1358:        keepfraction = ip_nfrags - dropfraction;
                   1359:        for (i = IPFRAGTTL, median = 0; i >= 0; i--) {
                   1360:                median +=  fragttl_histo[i];
                   1361:                if (median >= keepfraction)
                   1362:                        break;
                   1363:        }
                   1364:
                   1365:        /* Return TTL of median (or other fraction). */
                   1366:        return (u_int)i;
                   1367: }
                   1368:
                   1369: void
                   1370: ip_reass_drophalf(void)
                   1371: {
                   1372:
                   1373:        u_int median_ticks;
                   1374:        /*
                   1375:         * Compute median TTL of all fragments, and count frags
                   1376:         * with that TTL or lower (roughly half of all fragments).
                   1377:         */
                   1378:        median_ticks = ip_reass_ttl_decr(0);
                   1379:
                   1380:        /* Drop half. */
                   1381:        median_ticks = ip_reass_ttl_decr(median_ticks);
                   1382:
                   1383: }
                   1384:
                   1385: /*
1.1       cgd      1386:  * IP timer processing;
                   1387:  * if a timer expires on a reassembly
                   1388:  * queue, discard it.
                   1389:  */
1.8       mycroft  1390: void
1.211     perry    1391: ip_slowtimo(void)
1.1       cgd      1392: {
1.191     jonathan 1393:        static u_int dropscanidx = 0;
                   1394:        u_int i;
1.194     jonathan 1395:        u_int median_ttl;
1.268   ! ad       1396:
        !          1397:        mutex_enter(softnet_lock);
        !          1398:        KERNEL_LOCK(1, NULL);
1.1       cgd      1399:
1.75      thorpej  1400:        IPQ_LOCK();
1.194     jonathan 1401:
                   1402:        /* Age TTL of all fragments by 1 tick .*/
                   1403:        median_ttl = ip_reass_ttl_decr(1);
                   1404:
                   1405:        /* make sure fragment limit is up-to-date */
                   1406:        CHECK_NMBCLUSTER_PARAMS();
                   1407:
                   1408:        /* If we have too many fragments, drop the older half. */
                   1409:        if (ip_nfrags > ip_maxfrags)
                   1410:                ip_reass_ttl_decr(median_ttl);
                   1411:
1.131     itojun   1412:        /*
1.194     jonathan 1413:         * If we are over the maximum number of fragmented packets
1.131     itojun   1414:         * (due to the limit being lowered), drain off
1.190     jonathan 1415:         * enough to get down to the new limit. Start draining
                   1416:         * from the reassembly hashqueue most recently drained.
1.131     itojun   1417:         */
                   1418:        if (ip_maxfragpackets < 0)
                   1419:                ;
                   1420:        else {
1.190     jonathan 1421:                int wrapped = 0;
                   1422:
                   1423:                i = dropscanidx;
                   1424:                while (ip_nfragpackets > ip_maxfragpackets && wrapped == 0) {
                   1425:                        while (LIST_FIRST(&ipq[i]) != NULL)
                   1426:                                ip_freef(LIST_FIRST(&ipq[i]));
                   1427:                        if (++i >= IPREASS_NHASH) {
                   1428:                                i = 0;
                   1429:                        }
                   1430:                        /*
                   1431:                         * Dont scan forever even if fragment counters are
                   1432:                         * wrong: stop after scanning entire reassembly queue.
                   1433:                         */
                   1434:                        if (i == dropscanidx)
                   1435:                            wrapped = 1;
                   1436:                }
                   1437:                dropscanidx = i;
1.131     itojun   1438:        }
1.75      thorpej  1439:        IPQ_UNLOCK();
1.268   ! ad       1440:
        !          1441:        KERNEL_UNLOCK_ONE(NULL);
        !          1442:        mutex_exit(softnet_lock);
1.1       cgd      1443: }
                   1444:
                   1445: /*
                   1446:  * Drain off all datagram fragments.
                   1447:  */
1.8       mycroft  1448: void
1.211     perry    1449: ip_drain(void)
1.1       cgd      1450: {
                   1451:
1.268   ! ad       1452:        mutex_enter(softnet_lock);
        !          1453:        KERNEL_LOCK(1, NULL);
        !          1454:
1.75      thorpej  1455:        /*
                   1456:         * We may be called from a device's interrupt context.  If
                   1457:         * the ipq is already busy, just bail out now.
                   1458:         */
1.268   ! ad       1459:        if (ipq_lock_try() != 0) {
        !          1460:                /*
        !          1461:                 * Drop half the total fragments now. If more mbufs are
        !          1462:                 * needed, we will be called again soon.
        !          1463:                 */
        !          1464:                ip_reass_drophalf();
        !          1465:                IPQ_UNLOCK();
        !          1466:        }
1.75      thorpej  1467:
1.268   ! ad       1468:        KERNEL_UNLOCK_ONE(NULL);
        !          1469:        mutex_exit(softnet_lock);
1.1       cgd      1470: }
                   1471:
                   1472: /*
                   1473:  * Do option processing on a datagram,
                   1474:  * possibly discarding it if bad options are encountered,
                   1475:  * or forwarding it if source-routed.
                   1476:  * Returns 1 if packet has been forwarded/freed,
                   1477:  * 0 if the packet should be processed further.
                   1478:  */
1.8       mycroft  1479: int
1.211     perry    1480: ip_dooptions(struct mbuf *m)
1.1       cgd      1481: {
1.109     augustss 1482:        struct ip *ip = mtod(m, struct ip *);
                   1483:        u_char *cp, *cp0;
                   1484:        struct ip_timestamp *ipt;
                   1485:        struct in_ifaddr *ia;
1.1       cgd      1486:        int opt, optlen, cnt, off, code, type = ICMP_PARAMPROB, forward = 0;
1.104     thorpej  1487:        struct in_addr dst;
1.1       cgd      1488:        n_time ntime;
                   1489:
1.13      mycroft  1490:        dst = ip->ip_dst;
1.1       cgd      1491:        cp = (u_char *)(ip + 1);
                   1492:        cnt = (ip->ip_hl << 2) - sizeof (struct ip);
                   1493:        for (; cnt > 0; cnt -= optlen, cp += optlen) {
                   1494:                opt = cp[IPOPT_OPTVAL];
                   1495:                if (opt == IPOPT_EOL)
                   1496:                        break;
                   1497:                if (opt == IPOPT_NOP)
                   1498:                        optlen = 1;
                   1499:                else {
1.113     itojun   1500:                        if (cnt < IPOPT_OLEN + sizeof(*cp)) {
                   1501:                                code = &cp[IPOPT_OLEN] - (u_char *)ip;
                   1502:                                goto bad;
                   1503:                        }
1.1       cgd      1504:                        optlen = cp[IPOPT_OLEN];
1.114     itojun   1505:                        if (optlen < IPOPT_OLEN + sizeof(*cp) || optlen > cnt) {
1.1       cgd      1506:                                code = &cp[IPOPT_OLEN] - (u_char *)ip;
                   1507:                                goto bad;
                   1508:                        }
                   1509:                }
                   1510:                switch (opt) {
                   1511:
                   1512:                default:
                   1513:                        break;
                   1514:
                   1515:                /*
                   1516:                 * Source routing with record.
                   1517:                 * Find interface with current destination address.
                   1518:                 * If none on this machine then drop if strictly routed,
                   1519:                 * or do nothing if loosely routed.
                   1520:                 * Record interface address and bring up next address
                   1521:                 * component.  If strictly routed make sure next
                   1522:                 * address is on directly accessible net.
                   1523:                 */
                   1524:                case IPOPT_LSRR:
                   1525:                case IPOPT_SSRR:
1.47      cjs      1526:                        if (ip_allowsrcrt == 0) {
                   1527:                                type = ICMP_UNREACH;
                   1528:                                code = ICMP_UNREACH_NET_PROHIB;
                   1529:                                goto bad;
                   1530:                        }
1.114     itojun   1531:                        if (optlen < IPOPT_OFFSET + sizeof(*cp)) {
                   1532:                                code = &cp[IPOPT_OLEN] - (u_char *)ip;
                   1533:                                goto bad;
                   1534:                        }
1.1       cgd      1535:                        if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) {
                   1536:                                code = &cp[IPOPT_OFFSET] - (u_char *)ip;
                   1537:                                goto bad;
                   1538:                        }
                   1539:                        ipaddr.sin_addr = ip->ip_dst;
1.19      mycroft  1540:                        ia = ifatoia(ifa_ifwithaddr(sintosa(&ipaddr)));
1.1       cgd      1541:                        if (ia == 0) {
                   1542:                                if (opt == IPOPT_SSRR) {
                   1543:                                        type = ICMP_UNREACH;
                   1544:                                        code = ICMP_UNREACH_SRCFAIL;
                   1545:                                        goto bad;
                   1546:                                }
                   1547:                                /*
                   1548:                                 * Loose routing, and not at next destination
                   1549:                                 * yet; nothing to do except forward.
                   1550:                                 */
                   1551:                                break;
                   1552:                        }
                   1553:                        off--;                  /* 0 origin */
1.112     sommerfe 1554:                        if ((off + sizeof(struct in_addr)) > optlen) {
1.1       cgd      1555:                                /*
                   1556:                                 * End of source route.  Should be for us.
                   1557:                                 */
                   1558:                                save_rte(cp, ip->ip_src);
                   1559:                                break;
                   1560:                        }
                   1561:                        /*
                   1562:                         * locate outgoing interface
                   1563:                         */
1.244     christos 1564:                        bcopy((void *)(cp + off), (void *)&ipaddr.sin_addr,
1.1       cgd      1565:                            sizeof(ipaddr.sin_addr));
1.96      thorpej  1566:                        if (opt == IPOPT_SSRR)
1.196     itojun   1567:                                ia = ifatoia(ifa_ifwithladdr(sintosa(&ipaddr)));
1.96      thorpej  1568:                        else
1.1       cgd      1569:                                ia = ip_rtaddr(ipaddr.sin_addr);
                   1570:                        if (ia == 0) {
                   1571:                                type = ICMP_UNREACH;
                   1572:                                code = ICMP_UNREACH_SRCFAIL;
                   1573:                                goto bad;
                   1574:                        }
                   1575:                        ip->ip_dst = ipaddr.sin_addr;
1.244     christos 1576:                        bcopy((void *)&ia->ia_addr.sin_addr,
                   1577:                            (void *)(cp + off), sizeof(struct in_addr));
1.1       cgd      1578:                        cp[IPOPT_OFFSET] += sizeof(struct in_addr);
1.13      mycroft  1579:                        /*
                   1580:                         * Let ip_intr's mcast routing check handle mcast pkts
                   1581:                         */
1.18      mycroft  1582:                        forward = !IN_MULTICAST(ip->ip_dst.s_addr);
1.1       cgd      1583:                        break;
                   1584:
                   1585:                case IPOPT_RR:
1.114     itojun   1586:                        if (optlen < IPOPT_OFFSET + sizeof(*cp)) {
                   1587:                                code = &cp[IPOPT_OLEN] - (u_char *)ip;
                   1588:                                goto bad;
                   1589:                        }
1.1       cgd      1590:                        if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) {
                   1591:                                code = &cp[IPOPT_OFFSET] - (u_char *)ip;
                   1592:                                goto bad;
                   1593:                        }
                   1594:                        /*
                   1595:                         * If no space remains, ignore.
                   1596:                         */
                   1597:                        off--;                  /* 0 origin */
1.112     sommerfe 1598:                        if ((off + sizeof(struct in_addr)) > optlen)
1.1       cgd      1599:                                break;
1.244     christos 1600:                        bcopy((void *)(&ip->ip_dst), (void *)&ipaddr.sin_addr,
1.1       cgd      1601:                            sizeof(ipaddr.sin_addr));
                   1602:                        /*
                   1603:                         * locate outgoing interface; if we're the destination,
                   1604:                         * use the incoming interface (should be same).
                   1605:                         */
1.96      thorpej  1606:                        if ((ia = ifatoia(ifa_ifwithaddr(sintosa(&ipaddr))))
                   1607:                            == NULL &&
                   1608:                            (ia = ip_rtaddr(ipaddr.sin_addr)) == NULL) {
1.1       cgd      1609:                                type = ICMP_UNREACH;
                   1610:                                code = ICMP_UNREACH_HOST;
                   1611:                                goto bad;
                   1612:                        }
1.244     christos 1613:                        bcopy((void *)&ia->ia_addr.sin_addr,
                   1614:                            (void *)(cp + off), sizeof(struct in_addr));
1.1       cgd      1615:                        cp[IPOPT_OFFSET] += sizeof(struct in_addr);
                   1616:                        break;
                   1617:
                   1618:                case IPOPT_TS:
                   1619:                        code = cp - (u_char *)ip;
                   1620:                        ipt = (struct ip_timestamp *)cp;
1.114     itojun   1621:                        if (ipt->ipt_len < 4 || ipt->ipt_len > 40) {
                   1622:                                code = (u_char *)&ipt->ipt_len - (u_char *)ip;
1.1       cgd      1623:                                goto bad;
1.114     itojun   1624:                        }
                   1625:                        if (ipt->ipt_ptr < 5) {
                   1626:                                code = (u_char *)&ipt->ipt_ptr - (u_char *)ip;
                   1627:                                goto bad;
                   1628:                        }
1.15      cgd      1629:                        if (ipt->ipt_ptr > ipt->ipt_len - sizeof (int32_t)) {
1.114     itojun   1630:                                if (++ipt->ipt_oflw == 0) {
                   1631:                                        code = (u_char *)&ipt->ipt_ptr -
                   1632:                                            (u_char *)ip;
1.1       cgd      1633:                                        goto bad;
1.114     itojun   1634:                                }
1.1       cgd      1635:                                break;
                   1636:                        }
1.104     thorpej  1637:                        cp0 = (cp + ipt->ipt_ptr - 1);
1.1       cgd      1638:                        switch (ipt->ipt_flg) {
                   1639:
                   1640:                        case IPOPT_TS_TSONLY:
                   1641:                                break;
                   1642:
                   1643:                        case IPOPT_TS_TSANDADDR:
1.66      thorpej  1644:                                if (ipt->ipt_ptr - 1 + sizeof(n_time) +
1.114     itojun   1645:                                    sizeof(struct in_addr) > ipt->ipt_len) {
                   1646:                                        code = (u_char *)&ipt->ipt_ptr -
                   1647:                                            (u_char *)ip;
1.1       cgd      1648:                                        goto bad;
1.114     itojun   1649:                                }
1.13      mycroft  1650:                                ipaddr.sin_addr = dst;
1.96      thorpej  1651:                                ia = ifatoia(ifaof_ifpforaddr(sintosa(&ipaddr),
                   1652:                                    m->m_pkthdr.rcvif));
1.13      mycroft  1653:                                if (ia == 0)
                   1654:                                        continue;
1.104     thorpej  1655:                                bcopy(&ia->ia_addr.sin_addr,
                   1656:                                    cp0, sizeof(struct in_addr));
1.1       cgd      1657:                                ipt->ipt_ptr += sizeof(struct in_addr);
                   1658:                                break;
                   1659:
                   1660:                        case IPOPT_TS_PRESPEC:
1.66      thorpej  1661:                                if (ipt->ipt_ptr - 1 + sizeof(n_time) +
1.114     itojun   1662:                                    sizeof(struct in_addr) > ipt->ipt_len) {
                   1663:                                        code = (u_char *)&ipt->ipt_ptr -
                   1664:                                            (u_char *)ip;
1.1       cgd      1665:                                        goto bad;
1.114     itojun   1666:                                }
1.104     thorpej  1667:                                bcopy(cp0, &ipaddr.sin_addr,
1.1       cgd      1668:                                    sizeof(struct in_addr));
1.96      thorpej  1669:                                if (ifatoia(ifa_ifwithaddr(sintosa(&ipaddr)))
                   1670:                                    == NULL)
1.1       cgd      1671:                                        continue;
                   1672:                                ipt->ipt_ptr += sizeof(struct in_addr);
                   1673:                                break;
                   1674:
                   1675:                        default:
1.114     itojun   1676:                                /* XXX can't take &ipt->ipt_flg */
                   1677:                                code = (u_char *)&ipt->ipt_ptr -
                   1678:                                    (u_char *)ip + 1;
1.1       cgd      1679:                                goto bad;
                   1680:                        }
                   1681:                        ntime = iptime();
1.107     thorpej  1682:                        cp0 = (u_char *) &ntime; /* XXX grumble, GCC... */
1.244     christos 1683:                        memmove((char *)cp + ipt->ipt_ptr - 1, cp0,
1.1       cgd      1684:                            sizeof(n_time));
                   1685:                        ipt->ipt_ptr += sizeof(n_time);
                   1686:                }
                   1687:        }
                   1688:        if (forward) {
1.26      thorpej  1689:                if (ip_forwsrcrt == 0) {
                   1690:                        type = ICMP_UNREACH;
                   1691:                        code = ICMP_UNREACH_SRCFAIL;
                   1692:                        goto bad;
                   1693:                }
1.1       cgd      1694:                ip_forward(m, 1);
                   1695:                return (1);
1.13      mycroft  1696:        }
                   1697:        return (0);
1.1       cgd      1698: bad:
1.13      mycroft  1699:        icmp_error(m, type, code, 0, 0);
1.266     thorpej  1700:        IP_STATINC(IP_STAT_BADOPTIONS);
1.1       cgd      1701:        return (1);
                   1702: }
                   1703:
                   1704: /*
                   1705:  * Given address of next destination (final or next hop),
                   1706:  * return internet address info of interface to be used to get there.
                   1707:  */
                   1708: struct in_ifaddr *
1.211     perry    1709: ip_rtaddr(struct in_addr dst)
1.1       cgd      1710: {
1.249     dyoung   1711:        struct rtentry *rt;
                   1712:        union {
                   1713:                struct sockaddr         dst;
                   1714:                struct sockaddr_in      dst4;
                   1715:        } u;
                   1716:
                   1717:        sockaddr_in_init(&u.dst4, &dst, 0);
                   1718:
                   1719:        if ((rt = rtcache_lookup(&ipforward_rt, &u.dst)) == NULL)
                   1720:                return NULL;
                   1721:
                   1722:        return ifatoia(rt->rt_ifa);
1.1       cgd      1723: }
                   1724:
                   1725: /*
                   1726:  * Save incoming source route for use in replies,
                   1727:  * to be picked up later by ip_srcroute if the receiver is interested.
                   1728:  */
1.13      mycroft  1729: void
1.211     perry    1730: save_rte(u_char *option, struct in_addr dst)
1.1       cgd      1731: {
                   1732:        unsigned olen;
                   1733:
                   1734:        olen = option[IPOPT_OLEN];
                   1735: #ifdef DIAGNOSTIC
                   1736:        if (ipprintfs)
1.39      christos 1737:                printf("save_rte: olen %d\n", olen);
1.89      itojun   1738: #endif /* 0 */
1.1       cgd      1739:        if (olen > sizeof(ip_srcrt) - (1 + sizeof(dst)))
                   1740:                return;
1.244     christos 1741:        bcopy((void *)option, (void *)ip_srcrt.srcopt, olen);
1.1       cgd      1742:        ip_nhops = (olen - IPOPT_OFFSET - 1) / sizeof(struct in_addr);
                   1743:        ip_srcrt.dst = dst;
                   1744: }
                   1745:
                   1746: /*
                   1747:  * Retrieve incoming source route for use in replies,
                   1748:  * in the same form used by setsockopt.
                   1749:  * The first hop is placed before the options, will be removed later.
                   1750:  */
                   1751: struct mbuf *
1.211     perry    1752: ip_srcroute(void)
1.1       cgd      1753: {
1.109     augustss 1754:        struct in_addr *p, *q;
                   1755:        struct mbuf *m;
1.1       cgd      1756:
                   1757:        if (ip_nhops == 0)
1.237     dyoung   1758:                return NULL;
1.1       cgd      1759:        m = m_get(M_DONTWAIT, MT_SOOPTS);
                   1760:        if (m == 0)
1.237     dyoung   1761:                return NULL;
1.1       cgd      1762:
1.164     matt     1763:        MCLAIM(m, &inetdomain.dom_mowner);
1.13      mycroft  1764: #define OPTSIZ (sizeof(ip_srcrt.nop) + sizeof(ip_srcrt.srcopt))
1.1       cgd      1765:
                   1766:        /* length is (nhops+1)*sizeof(addr) + sizeof(nop + srcrt header) */
                   1767:        m->m_len = ip_nhops * sizeof(struct in_addr) + sizeof(struct in_addr) +
                   1768:            OPTSIZ;
                   1769: #ifdef DIAGNOSTIC
                   1770:        if (ipprintfs)
1.39      christos 1771:                printf("ip_srcroute: nhops %d mlen %d", ip_nhops, m->m_len);
1.1       cgd      1772: #endif
                   1773:
                   1774:        /*
                   1775:         * First save first hop for return route
                   1776:         */
                   1777:        p = &ip_srcrt.route[ip_nhops - 1];
                   1778:        *(mtod(m, struct in_addr *)) = *p--;
                   1779: #ifdef DIAGNOSTIC
                   1780:        if (ipprintfs)
1.39      christos 1781:                printf(" hops %x", ntohl(mtod(m, struct in_addr *)->s_addr));
1.1       cgd      1782: #endif
                   1783:
                   1784:        /*
                   1785:         * Copy option fields and padding (nop) to mbuf.
                   1786:         */
                   1787:        ip_srcrt.nop = IPOPT_NOP;
                   1788:        ip_srcrt.srcopt[IPOPT_OFFSET] = IPOPT_MINOFF;
1.244     christos 1789:        memmove(mtod(m, char *) + sizeof(struct in_addr), &ip_srcrt.nop,
                   1790:            OPTSIZ);
                   1791:        q = (struct in_addr *)(mtod(m, char *) +
1.1       cgd      1792:            sizeof(struct in_addr) + OPTSIZ);
                   1793: #undef OPTSIZ
                   1794:        /*
                   1795:         * Record return path as an IP source route,
                   1796:         * reversing the path (pointers are now aligned).
                   1797:         */
                   1798:        while (p >= ip_srcrt.route) {
                   1799: #ifdef DIAGNOSTIC
                   1800:                if (ipprintfs)
1.39      christos 1801:                        printf(" %x", ntohl(q->s_addr));
1.1       cgd      1802: #endif
                   1803:                *q++ = *p--;
                   1804:        }
                   1805:        /*
                   1806:         * Last hop goes to final destination.
                   1807:         */
                   1808:        *q = ip_srcrt.dst;
                   1809: #ifdef DIAGNOSTIC
                   1810:        if (ipprintfs)
1.39      christos 1811:                printf(" %x\n", ntohl(q->s_addr));
1.1       cgd      1812: #endif
                   1813:        return (m);
                   1814: }
                   1815:
1.139     matt     1816: const int inetctlerrmap[PRC_NCMDS] = {
1.256     yamt     1817:        [PRC_MSGSIZE] = EMSGSIZE,
                   1818:        [PRC_HOSTDEAD] = EHOSTDOWN,
                   1819:        [PRC_HOSTUNREACH] = EHOSTUNREACH,
                   1820:        [PRC_UNREACH_NET] = EHOSTUNREACH,
                   1821:        [PRC_UNREACH_HOST] = EHOSTUNREACH,
                   1822:        [PRC_UNREACH_PROTOCOL] = ECONNREFUSED,
                   1823:        [PRC_UNREACH_PORT] = ECONNREFUSED,
                   1824:        [PRC_UNREACH_SRCFAIL] = EHOSTUNREACH,
                   1825:        [PRC_PARAMPROB] = ENOPROTOOPT,
1.1       cgd      1826: };
                   1827:
                   1828: /*
                   1829:  * Forward a packet.  If some error occurs return the sender
                   1830:  * an icmp packet.  Note we can't always generate a meaningful
                   1831:  * icmp message because icmp doesn't have a large enough repertoire
                   1832:  * of codes and types.
                   1833:  *
                   1834:  * If not forwarding, just drop the packet.  This could be confusing
                   1835:  * if ipforwarding was zero but some routing protocol was advancing
                   1836:  * us as a gateway to somewhere.  However, we must let the routing
                   1837:  * protocol deal with that.
                   1838:  *
                   1839:  * The srcrt parameter indicates whether the packet is being forwarded
                   1840:  * via a source route.
                   1841:  */
1.13      mycroft  1842: void
1.211     perry    1843: ip_forward(struct mbuf *m, int srcrt)
1.1       cgd      1844: {
1.109     augustss 1845:        struct ip *ip = mtod(m, struct ip *);
                   1846:        struct rtentry *rt;
1.220     christos 1847:        int error, type = 0, code = 0, destmtu = 0;
1.1       cgd      1848:        struct mbuf *mcopy;
1.13      mycroft  1849:        n_long dest;
1.249     dyoung   1850:        union {
                   1851:                struct sockaddr         dst;
                   1852:                struct sockaddr_in      dst4;
                   1853:        } u;
1.164     matt     1854:
                   1855:        /*
                   1856:         * We are now in the output path.
                   1857:         */
                   1858:        MCLAIM(m, &ip_tx_mowner);
1.135     thorpej  1859:
                   1860:        /*
                   1861:         * Clear any in-bound checksum flags for this packet.
                   1862:         */
                   1863:        m->m_pkthdr.csum_flags = 0;
1.1       cgd      1864:
1.13      mycroft  1865:        dest = 0;
1.1       cgd      1866: #ifdef DIAGNOSTIC
1.224     joerg    1867:        if (ipprintfs) {
                   1868:                printf("forward: src %s ", inet_ntoa(ip->ip_src));
                   1869:                printf("dst %s ttl %x\n", inet_ntoa(ip->ip_dst), ip->ip_ttl);
                   1870:        }
1.1       cgd      1871: #endif
1.93      sommerfe 1872:        if (m->m_flags & (M_BCAST|M_MCAST) || in_canforward(ip->ip_dst) == 0) {
1.266     thorpej  1873:                IP_STATINC(IP_STAT_CANTFORWARD);
1.1       cgd      1874:                m_freem(m);
                   1875:                return;
                   1876:        }
                   1877:        if (ip->ip_ttl <= IPTTLDEC) {
1.13      mycroft  1878:                icmp_error(m, ICMP_TIMXCEED, ICMP_TIMXCEED_INTRANS, dest, 0);
1.1       cgd      1879:                return;
                   1880:        }
                   1881:
1.249     dyoung   1882:        sockaddr_in_init(&u.dst4, &ip->ip_dst, 0);
                   1883:        if ((rt = rtcache_lookup(&ipforward_rt, &u.dst)) == NULL) {
                   1884:                icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_NET, dest, 0);
                   1885:                return;
1.1       cgd      1886:        }
                   1887:
                   1888:        /*
1.34      mycroft  1889:         * Save at most 68 bytes of the packet in case
1.1       cgd      1890:         * we need to generate an ICMP message to the src.
1.119     itojun   1891:         * Pullup to avoid sharing mbuf cluster between m and mcopy.
1.1       cgd      1892:         */
1.155     itojun   1893:        mcopy = m_copym(m, 0, imin(ntohs(ip->ip_len), 68), M_DONTWAIT);
1.119     itojun   1894:        if (mcopy)
                   1895:                mcopy = m_pullup(mcopy, ip->ip_hl << 2);
1.1       cgd      1896:
1.221     christos 1897:        ip->ip_ttl -= IPTTLDEC;
                   1898:
1.1       cgd      1899:        /*
                   1900:         * If forwarding packet using same interface that it came in on,
                   1901:         * perhaps should send a redirect to sender to shortcut a hop.
                   1902:         * Only send redirect if source is sending directly to us,
                   1903:         * and if packet was not source routed (or has any options).
                   1904:         * Also, don't send redirect if forwarding using a default route
                   1905:         * or a route modified by a redirect.
                   1906:         */
                   1907:        if (rt->rt_ifp == m->m_pkthdr.rcvif &&
                   1908:            (rt->rt_flags & (RTF_DYNAMIC|RTF_MODIFIED)) == 0 &&
1.250     dyoung   1909:            !in_nullhost(satocsin(rt_getkey(rt))->sin_addr) &&
1.1       cgd      1910:            ipsendredirects && !srcrt) {
1.19      mycroft  1911:                if (rt->rt_ifa &&
                   1912:                    (ip->ip_src.s_addr & ifatoia(rt->rt_ifa)->ia_subnetmask) ==
                   1913:                    ifatoia(rt->rt_ifa)->ia_subnet) {
1.77      thorpej  1914:                        if (rt->rt_flags & RTF_GATEWAY)
                   1915:                                dest = satosin(rt->rt_gateway)->sin_addr.s_addr;
                   1916:                        else
                   1917:                                dest = ip->ip_dst.s_addr;
                   1918:                        /*
                   1919:                         * Router requirements says to only send host
                   1920:                         * redirects.
                   1921:                         */
                   1922:                        type = ICMP_REDIRECT;
                   1923:                        code = ICMP_REDIRECT_HOST;
1.1       cgd      1924: #ifdef DIAGNOSTIC
1.77      thorpej  1925:                        if (ipprintfs)
                   1926:                                printf("redirect (%d) to %x\n", code,
                   1927:                                    (u_int32_t)dest);
1.1       cgd      1928: #endif
                   1929:                }
                   1930:        }
                   1931:
1.238     dyoung   1932:        error = ip_output(m, NULL, &ipforward_rt,
1.173     jonathan 1933:            (IP_FORWARDING | (ip_directedbcast ? IP_ALLOWBROADCAST : 0)),
1.174     itojun   1934:            (struct ip_moptions *)NULL, (struct socket *)NULL);
1.173     jonathan 1935:
1.1       cgd      1936:        if (error)
1.266     thorpej  1937:                IP_STATINC(IP_STAT_CANTFORWARD);
1.1       cgd      1938:        else {
1.266     thorpej  1939:                uint64_t *ips = IP_STAT_GETREF();
                   1940:                ips[IP_STAT_FORWARD]++;
                   1941:                if (type) {
                   1942:                        ips[IP_STAT_REDIRECTSENT]++;
                   1943:                        IP_STAT_PUTREF();
                   1944:                } else {
                   1945:                        IP_STAT_PUTREF();
1.63      matt     1946:                        if (mcopy) {
                   1947: #ifdef GATEWAY
1.64      thorpej  1948:                                if (mcopy->m_flags & M_CANFASTFWD)
                   1949:                                        ipflow_create(&ipforward_rt, mcopy);
1.63      matt     1950: #endif
1.1       cgd      1951:                                m_freem(mcopy);
1.63      matt     1952:                        }
1.1       cgd      1953:                        return;
                   1954:                }
                   1955:        }
                   1956:        if (mcopy == NULL)
                   1957:                return;
1.13      mycroft  1958:
1.1       cgd      1959:        switch (error) {
                   1960:
                   1961:        case 0:                         /* forwarded, but need redirect */
                   1962:                /* type, code set above */
                   1963:                break;
                   1964:
                   1965:        case ENETUNREACH:               /* shouldn't happen, checked above */
                   1966:        case EHOSTUNREACH:
                   1967:        case ENETDOWN:
                   1968:        case EHOSTDOWN:
                   1969:        default:
                   1970:                type = ICMP_UNREACH;
                   1971:                code = ICMP_UNREACH_HOST;
                   1972:                break;
                   1973:
                   1974:        case EMSGSIZE:
                   1975:                type = ICMP_UNREACH;
                   1976:                code = ICMP_UNREACH_NEEDFRAG;
1.263     cube     1977:
1.261     dyoung   1978:                if ((rt = rtcache_validate(&ipforward_rt)) != NULL) {
1.263     cube     1979:
                   1980: #if defined(IPSEC) || defined(FAST_IPSEC)
                   1981:                        /*
                   1982:                         * If the packet is routed over IPsec tunnel, tell the
                   1983:                         * originator the tunnel MTU.
                   1984:                         *      tunnel MTU = if MTU - sizeof(IP) - ESP/AH hdrsiz
                   1985:                         * XXX quickhack!!!
                   1986:                         */
                   1987:
1.89      itojun   1988:                        struct secpolicy *sp;
                   1989:                        int ipsecerror;
1.95      itojun   1990:                        size_t ipsechdr;
1.89      itojun   1991:                        struct route *ro;
                   1992:
                   1993:                        sp = ipsec4_getpolicybyaddr(mcopy,
1.170     itojun   1994:                            IPSEC_DIR_OUTBOUND, IP_FORWARDING,
                   1995:                            &ipsecerror);
1.263     cube     1996: #endif
1.89      itojun   1997:
1.263     cube     1998:                        destmtu = rt->rt_ifp->if_mtu;
                   1999: #if defined(IPSEC) || defined(FAST_IPSEC)
                   2000:                        if (sp != NULL) {
1.89      itojun   2001:                                /* count IPsec header size */
1.95      itojun   2002:                                ipsechdr = ipsec4_hdrsiz(mcopy,
1.170     itojun   2003:                                    IPSEC_DIR_OUTBOUND, NULL);
1.89      itojun   2004:
                   2005:                                /*
                   2006:                                 * find the correct route for outer IPv4
                   2007:                                 * header, compute tunnel MTU.
                   2008:                                 */
1.220     christos 2009:
1.89      itojun   2010:                                if (sp->req != NULL
1.95      itojun   2011:                                 && sp->req->sav != NULL
                   2012:                                 && sp->req->sav->sah != NULL) {
                   2013:                                        ro = &sp->req->sav->sah->sa_route;
1.257     dyoung   2014:                                        if (rt && rt->rt_ifp) {
1.220     christos 2015:                                                destmtu =
1.257     dyoung   2016:                                                    rt->rt_rmx.rmx_mtu ?
                   2017:                                                    rt->rt_rmx.rmx_mtu :
                   2018:                                                    rt->rt_ifp->if_mtu;
1.220     christos 2019:                                                destmtu -= ipsechdr;
1.89      itojun   2020:                                        }
                   2021:                                }
                   2022:
1.173     jonathan 2023: #ifdef IPSEC
1.89      itojun   2024:                                key_freesp(sp);
1.173     jonathan 2025: #else
                   2026:                                KEY_FREESP(&sp);
                   2027: #endif
1.89      itojun   2028:                        }
1.263     cube     2029: #endif /*defined(IPSEC) || defined(FAST_IPSEC)*/
1.89      itojun   2030:                }
1.266     thorpej  2031:                IP_STATINC(IP_STAT_CANTFRAG);
1.1       cgd      2032:                break;
                   2033:
                   2034:        case ENOBUFS:
1.143     itojun   2035: #if 1
                   2036:                /*
                   2037:                 * a router should not generate ICMP_SOURCEQUENCH as
                   2038:                 * required in RFC1812 Requirements for IP Version 4 Routers.
                   2039:                 * source quench could be a big problem under DoS attacks,
1.149     wiz      2040:                 * or if the underlying interface is rate-limited.
1.143     itojun   2041:                 */
                   2042:                if (mcopy)
                   2043:                        m_freem(mcopy);
                   2044:                return;
                   2045: #else
1.1       cgd      2046:                type = ICMP_SOURCEQUENCH;
                   2047:                code = 0;
                   2048:                break;
1.143     itojun   2049: #endif
1.1       cgd      2050:        }
1.220     christos 2051:        icmp_error(mcopy, type, code, dest, destmtu);
1.44      thorpej  2052: }
                   2053:
                   2054: void
1.211     perry    2055: ip_savecontrol(struct inpcb *inp, struct mbuf **mp, struct ip *ip,
                   2056:     struct mbuf *m)
1.44      thorpej  2057: {
                   2058:
                   2059:        if (inp->inp_socket->so_options & SO_TIMESTAMP) {
                   2060:                struct timeval tv;
                   2061:
                   2062:                microtime(&tv);
1.244     christos 2063:                *mp = sbcreatecontrol((void *) &tv, sizeof(tv),
1.44      thorpej  2064:                    SCM_TIMESTAMP, SOL_SOCKET);
                   2065:                if (*mp)
                   2066:                        mp = &(*mp)->m_next;
                   2067:        }
                   2068:        if (inp->inp_flags & INP_RECVDSTADDR) {
1.244     christos 2069:                *mp = sbcreatecontrol((void *) &ip->ip_dst,
1.44      thorpej  2070:                    sizeof(struct in_addr), IP_RECVDSTADDR, IPPROTO_IP);
                   2071:                if (*mp)
                   2072:                        mp = &(*mp)->m_next;
                   2073:        }
                   2074: #ifdef notyet
                   2075:        /*
                   2076:         * XXX
                   2077:         * Moving these out of udp_input() made them even more broken
                   2078:         * than they already were.
                   2079:         *      - fenner@parc.xerox.com
                   2080:         */
                   2081:        /* options were tossed already */
                   2082:        if (inp->inp_flags & INP_RECVOPTS) {
1.244     christos 2083:                *mp = sbcreatecontrol((void *) opts_deleted_above,
1.44      thorpej  2084:                    sizeof(struct in_addr), IP_RECVOPTS, IPPROTO_IP);
                   2085:                if (*mp)
                   2086:                        mp = &(*mp)->m_next;
                   2087:        }
                   2088:        /* ip_srcroute doesn't do what we want here, need to fix */
                   2089:        if (inp->inp_flags & INP_RECVRETOPTS) {
1.244     christos 2090:                *mp = sbcreatecontrol((void *) ip_srcroute(),
1.44      thorpej  2091:                    sizeof(struct in_addr), IP_RECVRETOPTS, IPPROTO_IP);
                   2092:                if (*mp)
                   2093:                        mp = &(*mp)->m_next;
                   2094:        }
                   2095: #endif
                   2096:        if (inp->inp_flags & INP_RECVIF) {
                   2097:                struct sockaddr_dl sdl;
                   2098:
1.252     dyoung   2099:                sockaddr_dl_init(&sdl, sizeof(sdl),
                   2100:                    (m->m_pkthdr.rcvif != NULL)
                   2101:                        ?  m->m_pkthdr.rcvif->if_index
                   2102:                        : 0,
                   2103:                        0, NULL, 0, NULL, 0);
1.251     dyoung   2104:                *mp = sbcreatecontrol(&sdl, sdl.sdl_len, IP_RECVIF, IPPROTO_IP);
1.44      thorpej  2105:                if (*mp)
                   2106:                        mp = &(*mp)->m_next;
                   2107:        }
1.13      mycroft  2108: }
                   2109:
1.189     atatat   2110: /*
1.228     elad     2111:  * sysctl helper routine for net.inet.ip.forwsrcrt.
                   2112:  */
                   2113: static int
                   2114: sysctl_net_inet_ip_forwsrcrt(SYSCTLFN_ARGS)
                   2115: {
                   2116:        int error, tmp;
                   2117:        struct sysctlnode node;
                   2118:
                   2119:        node = *rnode;
                   2120:        tmp = ip_forwsrcrt;
                   2121:        node.sysctl_data = &tmp;
                   2122:        error = sysctl_lookup(SYSCTLFN_CALL(&node));
                   2123:        if (error || newp == NULL)
                   2124:                return (error);
                   2125:
1.230     elad     2126:        if (kauth_authorize_network(l->l_cred, KAUTH_NETWORK_FORWSRCRT,
1.232     elad     2127:            0, NULL, NULL, NULL))
1.228     elad     2128:                return (EPERM);
                   2129:
                   2130:        ip_forwsrcrt = tmp;
                   2131:
                   2132:        return (0);
                   2133: }
                   2134:
                   2135: /*
1.189     atatat   2136:  * sysctl helper routine for net.inet.ip.mtudisctimeout.  checks the
                   2137:  * range of the new value and tweaks timers if it changes.
                   2138:  */
                   2139: static int
                   2140: sysctl_net_inet_ip_pmtudto(SYSCTLFN_ARGS)
1.13      mycroft  2141: {
1.189     atatat   2142:        int error, tmp;
                   2143:        struct sysctlnode node;
                   2144:
                   2145:        node = *rnode;
                   2146:        tmp = ip_mtudisc_timeout;
                   2147:        node.sysctl_data = &tmp;
                   2148:        error = sysctl_lookup(SYSCTLFN_CALL(&node));
                   2149:        if (error || newp == NULL)
                   2150:                return (error);
                   2151:        if (tmp < 0)
                   2152:                return (EINVAL);
1.52      thorpej  2153:
1.189     atatat   2154:        ip_mtudisc_timeout = tmp;
                   2155:        rt_timer_queue_change(ip_mtudisc_timeout_q, ip_mtudisc_timeout);
                   2156:
                   2157:        return (0);
                   2158: }
1.54      lukem    2159:
1.65      matt     2160: #ifdef GATEWAY
1.189     atatat   2161: /*
1.247     liamjfoy 2162:  * sysctl helper routine for net.inet.ip.maxflows.
1.189     atatat   2163:  */
                   2164: static int
                   2165: sysctl_net_inet_ip_maxflows(SYSCTLFN_ARGS)
                   2166: {
                   2167:        int s;
1.67      thorpej  2168:
1.217     atatat   2169:        s = sysctl_lookup(SYSCTLFN_CALL(rnode));
1.247     liamjfoy 2170:        if (s || newp == NULL)
1.189     atatat   2171:                return (s);
1.212     perry    2172:
1.189     atatat   2173:        s = splsoftnet();
1.265     thorpej  2174:        ipflow_prune();
1.189     atatat   2175:        splx(s);
1.144     martin   2176:
1.189     atatat   2177:        return (0);
                   2178: }
1.248     liamjfoy 2179:
                   2180: static int
                   2181: sysctl_net_inet_ip_hashsize(SYSCTLFN_ARGS)
                   2182: {
                   2183:        int error, tmp;
                   2184:        struct sysctlnode node;
                   2185:
                   2186:        node = *rnode;
                   2187:        tmp = ip_hashsize;
                   2188:        node.sysctl_data = &tmp;
                   2189:        error = sysctl_lookup(SYSCTLFN_CALL(&node));
                   2190:        if (error || newp == NULL)
                   2191:                return (error);
                   2192:
                   2193:        if ((tmp & (tmp - 1)) == 0 && tmp != 0) {
                   2194:                /*
                   2195:                 * Can only fail due to malloc()
                   2196:                 */
                   2197:                if (ipflow_invalidate_all(tmp))
                   2198:                        return ENOMEM;
                   2199:        } else {
                   2200:                /*
                   2201:                 * EINVAL if not a power of 2
                   2202:                 */
                   2203:                return EINVAL;
                   2204:        }
                   2205:
                   2206:        return (0);
                   2207: }
1.189     atatat   2208: #endif /* GATEWAY */
1.117     tron     2209:
1.266     thorpej  2210: static int
                   2211: sysctl_net_inet_ip_stats(SYSCTLFN_ARGS)
                   2212: {
1.267     thorpej  2213:        netstat_sysctl_context ctx;
1.266     thorpej  2214:        uint64_t ips[IP_NSTATS];
                   2215:
1.267     thorpej  2216:        ctx.ctx_stat = ipstat_percpu;
                   2217:        ctx.ctx_counters = ips;
                   2218:        ctx.ctx_ncounters = IP_NSTATS;
                   2219:        return (NETSTAT_SYSCTL(&ctx));
1.266     thorpej  2220: }
1.131     itojun   2221:
1.189     atatat   2222: SYSCTL_SETUP(sysctl_net_inet_ip_setup, "sysctl net.inet.ip subtree setup")
                   2223: {
                   2224:        extern int subnetsarelocal, hostzeroisbroadcast;
1.180     jonathan 2225:
1.197     atatat   2226:        sysctl_createv(clog, 0, NULL, NULL,
                   2227:                       CTLFLAG_PERMANENT,
1.189     atatat   2228:                       CTLTYPE_NODE, "net", NULL,
                   2229:                       NULL, 0, NULL, 0,
                   2230:                       CTL_NET, CTL_EOL);
1.197     atatat   2231:        sysctl_createv(clog, 0, NULL, NULL,
                   2232:                       CTLFLAG_PERMANENT,
1.203     atatat   2233:                       CTLTYPE_NODE, "inet",
                   2234:                       SYSCTL_DESCR("PF_INET related settings"),
1.189     atatat   2235:                       NULL, 0, NULL, 0,
                   2236:                       CTL_NET, PF_INET, CTL_EOL);
1.197     atatat   2237:        sysctl_createv(clog, 0, NULL, NULL,
                   2238:                       CTLFLAG_PERMANENT,
1.203     atatat   2239:                       CTLTYPE_NODE, "ip",
                   2240:                       SYSCTL_DESCR("IPv4 related settings"),
1.189     atatat   2241:                       NULL, 0, NULL, 0,
                   2242:                       CTL_NET, PF_INET, IPPROTO_IP, CTL_EOL);
1.212     perry    2243:
1.197     atatat   2244:        sysctl_createv(clog, 0, NULL, NULL,
                   2245:                       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203     atatat   2246:                       CTLTYPE_INT, "forwarding",
                   2247:                       SYSCTL_DESCR("Enable forwarding of INET datagrams"),
1.189     atatat   2248:                       NULL, 0, &ipforwarding, 0,
                   2249:                       CTL_NET, PF_INET, IPPROTO_IP,
                   2250:                       IPCTL_FORWARDING, CTL_EOL);
1.197     atatat   2251:        sysctl_createv(clog, 0, NULL, NULL,
                   2252:                       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203     atatat   2253:                       CTLTYPE_INT, "redirect",
                   2254:                       SYSCTL_DESCR("Enable sending of ICMP redirect messages"),
1.189     atatat   2255:                       NULL, 0, &ipsendredirects, 0,
                   2256:                       CTL_NET, PF_INET, IPPROTO_IP,
                   2257:                       IPCTL_SENDREDIRECTS, CTL_EOL);
1.197     atatat   2258:        sysctl_createv(clog, 0, NULL, NULL,
                   2259:                       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203     atatat   2260:                       CTLTYPE_INT, "ttl",
                   2261:                       SYSCTL_DESCR("Default TTL for an INET datagram"),
1.189     atatat   2262:                       NULL, 0, &ip_defttl, 0,
                   2263:                       CTL_NET, PF_INET, IPPROTO_IP,
                   2264:                       IPCTL_DEFTTL, CTL_EOL);
                   2265: #ifdef IPCTL_DEFMTU
1.197     atatat   2266:        sysctl_createv(clog, 0, NULL, NULL,
                   2267:                       CTLFLAG_PERMANENT /* |CTLFLAG_READWRITE? */,
1.203     atatat   2268:                       CTLTYPE_INT, "mtu",
                   2269:                       SYSCTL_DESCR("Default MTA for an INET route"),
1.189     atatat   2270:                       NULL, 0, &ip_mtu, 0,
                   2271:                       CTL_NET, PF_INET, IPPROTO_IP,
                   2272:                       IPCTL_DEFMTU, CTL_EOL);
                   2273: #endif /* IPCTL_DEFMTU */
1.197     atatat   2274:        sysctl_createv(clog, 0, NULL, NULL,
1.228     elad     2275:                       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203     atatat   2276:                       CTLTYPE_INT, "forwsrcrt",
                   2277:                       SYSCTL_DESCR("Enable forwarding of source-routed "
                   2278:                                    "datagrams"),
1.228     elad     2279:                       sysctl_net_inet_ip_forwsrcrt, 0, &ip_forwsrcrt, 0,
1.189     atatat   2280:                       CTL_NET, PF_INET, IPPROTO_IP,
                   2281:                       IPCTL_FORWSRCRT, CTL_EOL);
1.197     atatat   2282:        sysctl_createv(clog, 0, NULL, NULL,
                   2283:                       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203     atatat   2284:                       CTLTYPE_INT, "directed-broadcast",
                   2285:                       SYSCTL_DESCR("Enable forwarding of broadcast datagrams"),
1.189     atatat   2286:                       NULL, 0, &ip_directedbcast, 0,
                   2287:                       CTL_NET, PF_INET, IPPROTO_IP,
                   2288:                       IPCTL_DIRECTEDBCAST, CTL_EOL);
1.197     atatat   2289:        sysctl_createv(clog, 0, NULL, NULL,
                   2290:                       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203     atatat   2291:                       CTLTYPE_INT, "allowsrcrt",
                   2292:                       SYSCTL_DESCR("Accept source-routed datagrams"),
1.189     atatat   2293:                       NULL, 0, &ip_allowsrcrt, 0,
                   2294:                       CTL_NET, PF_INET, IPPROTO_IP,
                   2295:                       IPCTL_ALLOWSRCRT, CTL_EOL);
1.197     atatat   2296:        sysctl_createv(clog, 0, NULL, NULL,
                   2297:                       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203     atatat   2298:                       CTLTYPE_INT, "subnetsarelocal",
                   2299:                       SYSCTL_DESCR("Whether logical subnets are considered "
                   2300:                                    "local"),
1.189     atatat   2301:                       NULL, 0, &subnetsarelocal, 0,
                   2302:                       CTL_NET, PF_INET, IPPROTO_IP,
                   2303:                       IPCTL_SUBNETSARELOCAL, CTL_EOL);
1.197     atatat   2304:        sysctl_createv(clog, 0, NULL, NULL,
                   2305:                       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203     atatat   2306:                       CTLTYPE_INT, "mtudisc",
                   2307:                       SYSCTL_DESCR("Use RFC1191 Path MTU Discovery"),
1.189     atatat   2308:                       NULL, 0, &ip_mtudisc, 0,
                   2309:                       CTL_NET, PF_INET, IPPROTO_IP,
                   2310:                       IPCTL_MTUDISC, CTL_EOL);
1.197     atatat   2311:        sysctl_createv(clog, 0, NULL, NULL,
                   2312:                       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203     atatat   2313:                       CTLTYPE_INT, "anonportmin",
                   2314:                       SYSCTL_DESCR("Lowest ephemeral port number to assign"),
1.189     atatat   2315:                       sysctl_net_inet_ip_ports, 0, &anonportmin, 0,
                   2316:                       CTL_NET, PF_INET, IPPROTO_IP,
                   2317:                       IPCTL_ANONPORTMIN, CTL_EOL);
1.197     atatat   2318:        sysctl_createv(clog, 0, NULL, NULL,
                   2319:                       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203     atatat   2320:                       CTLTYPE_INT, "anonportmax",
                   2321:                       SYSCTL_DESCR("Highest ephemeral port number to assign"),
1.189     atatat   2322:                       sysctl_net_inet_ip_ports, 0, &anonportmax, 0,
                   2323:                       CTL_NET, PF_INET, IPPROTO_IP,
                   2324:                       IPCTL_ANONPORTMAX, CTL_EOL);
1.197     atatat   2325:        sysctl_createv(clog, 0, NULL, NULL,
                   2326:                       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203     atatat   2327:                       CTLTYPE_INT, "mtudisctimeout",
                   2328:                       SYSCTL_DESCR("Lifetime of a Path MTU Discovered route"),
1.189     atatat   2329:                       sysctl_net_inet_ip_pmtudto, 0, &ip_mtudisc_timeout, 0,
                   2330:                       CTL_NET, PF_INET, IPPROTO_IP,
                   2331:                       IPCTL_MTUDISCTIMEOUT, CTL_EOL);
                   2332: #ifdef GATEWAY
1.197     atatat   2333:        sysctl_createv(clog, 0, NULL, NULL,
                   2334:                       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203     atatat   2335:                       CTLTYPE_INT, "maxflows",
                   2336:                       SYSCTL_DESCR("Number of flows for fast forwarding"),
1.189     atatat   2337:                       sysctl_net_inet_ip_maxflows, 0, &ip_maxflows, 0,
                   2338:                       CTL_NET, PF_INET, IPPROTO_IP,
                   2339:                       IPCTL_MAXFLOWS, CTL_EOL);
1.248     liamjfoy 2340:        sysctl_createv(clog, 0, NULL, NULL,
                   2341:                        CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
                   2342:                        CTLTYPE_INT, "hashsize",
                   2343:                        SYSCTL_DESCR("Size of hash table for fast forwarding (IPv4)"),
                   2344:                        sysctl_net_inet_ip_hashsize, 0, &ip_hashsize, 0,
                   2345:                        CTL_NET, PF_INET, IPPROTO_IP,
                   2346:                        CTL_CREATE, CTL_EOL);
1.189     atatat   2347: #endif /* GATEWAY */
1.197     atatat   2348:        sysctl_createv(clog, 0, NULL, NULL,
                   2349:                       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203     atatat   2350:                       CTLTYPE_INT, "hostzerobroadcast",
                   2351:                       SYSCTL_DESCR("All zeroes address is broadcast address"),
1.189     atatat   2352:                       NULL, 0, &hostzeroisbroadcast, 0,
                   2353:                       CTL_NET, PF_INET, IPPROTO_IP,
                   2354:                       IPCTL_HOSTZEROBROADCAST, CTL_EOL);
                   2355: #if NGIF > 0
1.197     atatat   2356:        sysctl_createv(clog, 0, NULL, NULL,
                   2357:                       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203     atatat   2358:                       CTLTYPE_INT, "gifttl",
                   2359:                       SYSCTL_DESCR("Default TTL for a gif tunnel datagram"),
1.189     atatat   2360:                       NULL, 0, &ip_gif_ttl, 0,
                   2361:                       CTL_NET, PF_INET, IPPROTO_IP,
                   2362:                       IPCTL_GIF_TTL, CTL_EOL);
                   2363: #endif /* NGIF */
                   2364: #ifndef IPNOPRIVPORTS
1.197     atatat   2365:        sysctl_createv(clog, 0, NULL, NULL,
                   2366:                       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203     atatat   2367:                       CTLTYPE_INT, "lowportmin",
                   2368:                       SYSCTL_DESCR("Lowest privileged ephemeral port number "
                   2369:                                    "to assign"),
1.189     atatat   2370:                       sysctl_net_inet_ip_ports, 0, &lowportmin, 0,
                   2371:                       CTL_NET, PF_INET, IPPROTO_IP,
                   2372:                       IPCTL_LOWPORTMIN, CTL_EOL);
1.197     atatat   2373:        sysctl_createv(clog, 0, NULL, NULL,
                   2374:                       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203     atatat   2375:                       CTLTYPE_INT, "lowportmax",
                   2376:                       SYSCTL_DESCR("Highest privileged ephemeral port number "
                   2377:                                    "to assign"),
1.189     atatat   2378:                       sysctl_net_inet_ip_ports, 0, &lowportmax, 0,
                   2379:                       CTL_NET, PF_INET, IPPROTO_IP,
                   2380:                       IPCTL_LOWPORTMAX, CTL_EOL);
                   2381: #endif /* IPNOPRIVPORTS */
1.197     atatat   2382:        sysctl_createv(clog, 0, NULL, NULL,
                   2383:                       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203     atatat   2384:                       CTLTYPE_INT, "maxfragpackets",
                   2385:                       SYSCTL_DESCR("Maximum number of fragments to retain for "
                   2386:                                    "possible reassembly"),
1.189     atatat   2387:                       NULL, 0, &ip_maxfragpackets, 0,
                   2388:                       CTL_NET, PF_INET, IPPROTO_IP,
                   2389:                       IPCTL_MAXFRAGPACKETS, CTL_EOL);
                   2390: #if NGRE > 0
1.197     atatat   2391:        sysctl_createv(clog, 0, NULL, NULL,
                   2392:                       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203     atatat   2393:                       CTLTYPE_INT, "grettl",
                   2394:                       SYSCTL_DESCR("Default TTL for a gre tunnel datagram"),
1.189     atatat   2395:                       NULL, 0, &ip_gre_ttl, 0,
                   2396:                       CTL_NET, PF_INET, IPPROTO_IP,
                   2397:                       IPCTL_GRE_TTL, CTL_EOL);
                   2398: #endif /* NGRE */
1.197     atatat   2399:        sysctl_createv(clog, 0, NULL, NULL,
                   2400:                       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203     atatat   2401:                       CTLTYPE_INT, "checkinterface",
                   2402:                       SYSCTL_DESCR("Enable receive side of Strong ES model "
                   2403:                                    "from RFC1122"),
1.189     atatat   2404:                       NULL, 0, &ip_checkinterface, 0,
                   2405:                       CTL_NET, PF_INET, IPPROTO_IP,
                   2406:                       IPCTL_CHECKINTERFACE, CTL_EOL);
1.197     atatat   2407:        sysctl_createv(clog, 0, NULL, NULL,
                   2408:                       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1.203     atatat   2409:                       CTLTYPE_INT, "random_id",
                   2410:                       SYSCTL_DESCR("Assign random ip_id values"),
1.189     atatat   2411:                       NULL, 0, &ip_do_randomid, 0,
                   2412:                       CTL_NET, PF_INET, IPPROTO_IP,
                   2413:                       IPCTL_RANDOMID, CTL_EOL);
1.206     thorpej  2414:        sysctl_createv(clog, 0, NULL, NULL,
                   2415:                       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
                   2416:                       CTLTYPE_INT, "do_loopback_cksum",
                   2417:                       SYSCTL_DESCR("Perform IP checksum on loopback"),
                   2418:                       NULL, 0, &ip_do_loopback_cksum, 0,
                   2419:                       CTL_NET, PF_INET, IPPROTO_IP,
                   2420:                       IPCTL_LOOPBACKCKSUM, CTL_EOL);
1.219     elad     2421:        sysctl_createv(clog, 0, NULL, NULL,
                   2422:                       CTLFLAG_PERMANENT,
                   2423:                       CTLTYPE_STRUCT, "stats",
                   2424:                       SYSCTL_DESCR("IP statistics"),
1.266     thorpej  2425:                       sysctl_net_inet_ip_stats, 0, NULL, 0,
1.219     elad     2426:                       CTL_NET, PF_INET, IPPROTO_IP, IPCTL_STATS,
                   2427:                       CTL_EOL);
1.1       cgd      2428: }
1.266     thorpej  2429:
                   2430: void
                   2431: ip_statinc(u_int stat)
                   2432: {
                   2433:
                   2434:        KASSERT(stat < IP_NSTATS);
                   2435:        IP_STATINC(stat);
                   2436: }

CVSweb <webmaster@jp.NetBSD.org>