[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 1.3.4.1

1.1       cgd         1: /*
                      2:  * Copyright (c) 1982, 1986, 1988 Regents of the University of California.
                      3:  * All rights reserved.
                      4:  *
                      5:  * Redistribution and use in source and binary forms, with or without
                      6:  * modification, are permitted provided that the following conditions
                      7:  * are met:
                      8:  * 1. Redistributions of source code must retain the above copyright
                      9:  *    notice, this list of conditions and the following disclaimer.
                     10:  * 2. Redistributions in binary form must reproduce the above copyright
                     11:  *    notice, this list of conditions and the following disclaimer in the
                     12:  *    documentation and/or other materials provided with the distribution.
                     13:  * 3. All advertising materials mentioning features or use of this software
                     14:  *    must display the following acknowledgement:
                     15:  *     This product includes software developed by the University of
                     16:  *     California, Berkeley and its contributors.
                     17:  * 4. Neither the name of the University nor the names of its contributors
                     18:  *    may be used to endorse or promote products derived from this software
                     19:  *    without specific prior written permission.
                     20:  *
                     21:  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
                     22:  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
                     23:  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
                     24:  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
                     25:  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
                     26:  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
                     27:  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
                     28:  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
                     29:  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
                     30:  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
                     31:  * SUCH DAMAGE.
                     32:  *
1.3       cgd        33:  *     from: @(#)ip_input.c    7.19 (Berkeley) 5/25/91
1.3.4.1 ! mycroft    34:  *     $Id: ip_input.c,v 1.3 1993/05/20 03:50:25 cgd Exp $
1.1       cgd        35:  */
                     36:
                     37: #include "param.h"
                     38: #include "systm.h"
                     39: #include "malloc.h"
                     40: #include "mbuf.h"
                     41: #include "domain.h"
                     42: #include "protosw.h"
                     43: #include "socket.h"
                     44: #include "errno.h"
                     45: #include "time.h"
                     46: #include "kernel.h"
1.3.4.1 ! mycroft    47:
        !            48: #include "machine/cpu.h"
1.1       cgd        49:
                     50: #include "../net/if.h"
                     51: #include "../net/route.h"
                     52:
                     53: #include "in.h"
                     54: #include "in_systm.h"
                     55: #include "ip.h"
                     56: #include "in_pcb.h"
                     57: #include "in_var.h"
                     58: #include "ip_var.h"
                     59: #include "ip_icmp.h"
                     60:
                     61: #ifndef        IPFORWARDING
                     62: #ifdef GATEWAY
                     63: #define        IPFORWARDING    1       /* forward IP packets not for us */
                     64: #else /* GATEWAY */
                     65: #define        IPFORWARDING    0       /* don't forward IP packets not for us */
                     66: #endif /* GATEWAY */
                     67: #endif /* IPFORWARDING */
                     68: #ifndef        IPSENDREDIRECTS
                     69: #define        IPSENDREDIRECTS 1
                     70: #endif
                     71: int    ipforwarding = IPFORWARDING;
                     72: int    ipsendredirects = IPSENDREDIRECTS;
                     73: #ifdef DIAGNOSTIC
                     74: int    ipprintfs = 0;
                     75: #endif
                     76:
                     77: extern struct domain inetdomain;
                     78: extern struct protosw inetsw[];
                     79: u_char ip_protox[IPPROTO_MAX];
                     80: int    ipqmaxlen = IFQ_MAXLEN;
                     81: struct in_ifaddr *in_ifaddr;                   /* first inet address */
                     82:
                     83: /*
                     84:  * We need to save the IP options in case a protocol wants to respond
                     85:  * to an incoming packet over the same route if the packet got here
                     86:  * using IP source routing.  This allows connection establishment and
                     87:  * maintenance when the remote end is on a network that is not known
                     88:  * to us.
                     89:  */
                     90: int    ip_nhops = 0;
                     91: static struct ip_srcrt {
                     92:        struct  in_addr dst;                    /* final destination */
                     93:        char    nop;                            /* one NOP to align */
                     94:        char    srcopt[IPOPT_OFFSET + 1];       /* OPTVAL, OLEN and OFFSET */
                     95:        struct  in_addr route[MAX_IPOPTLEN/sizeof(struct in_addr)];
                     96: } ip_srcrt;
                     97:
                     98: #ifdef GATEWAY
                     99: extern int if_index;
                    100: u_long *ip_ifmatrix;
                    101: #endif
                    102:
                    103: /*
                    104:  * IP initialization: fill in IP protocol switch table.
                    105:  * All protocols not implemented in kernel go to raw IP protocol handler.
                    106:  */
                    107: ip_init()
                    108: {
                    109:        register struct protosw *pr;
                    110:        register int i;
                    111:
                    112:        pr = pffindproto(PF_INET, IPPROTO_RAW, SOCK_RAW);
                    113:        if (pr == 0)
                    114:                panic("ip_init");
                    115:        for (i = 0; i < IPPROTO_MAX; i++)
                    116:                ip_protox[i] = pr - inetsw;
                    117:        for (pr = inetdomain.dom_protosw;
                    118:            pr < inetdomain.dom_protoswNPROTOSW; pr++)
                    119:                if (pr->pr_domain->dom_family == PF_INET &&
                    120:                    pr->pr_protocol && pr->pr_protocol != IPPROTO_RAW)
                    121:                        ip_protox[pr->pr_protocol] = pr - inetsw;
                    122:        ipq.next = ipq.prev = &ipq;
                    123:        ip_id = time.tv_sec & 0xffff;
                    124:        ipintrq.ifq_maxlen = ipqmaxlen;
                    125: #ifdef GATEWAY
                    126:        i = (if_index + 1) * (if_index + 1) * sizeof (u_long);
                    127:        if ((ip_ifmatrix = (u_long *) malloc(i, M_RTABLE, M_WAITOK)) == 0)
                    128:                panic("no memory for ip_ifmatrix");
                    129: #endif
                    130: }
                    131:
                    132: struct ip *ip_reass();
                    133: struct sockaddr_in ipaddr = { sizeof(ipaddr), AF_INET };
                    134: struct route ipforward_rt;
                    135:
                    136: /*
                    137:  * Ip input routine.  Checksum and byte swap header.  If fragmented
                    138:  * try to reassemble.  Process options.  Pass to next level.
                    139:  */
                    140: ipintr()
                    141: {
                    142:        register struct ip *ip;
                    143:        register struct mbuf *m;
                    144:        register struct ipq *fp;
                    145:        register struct in_ifaddr *ia;
                    146:        int hlen, s;
1.2       cgd       147: #ifdef PARANOID
                    148:        static int busy = 0;
1.1       cgd       149:
1.2       cgd       150:        if (busy)
                    151:                panic("ipintr: called recursively\n");
                    152:        ++busy;
                    153: #endif
1.1       cgd       154: next:
                    155:        /*
                    156:         * Get next datagram off input queue and get IP header
                    157:         * in first mbuf.
                    158:         */
                    159:        s = splimp();
                    160:        IF_DEQUEUE(&ipintrq, m);
                    161:        splx(s);
1.2       cgd       162:        if (m == 0) {
                    163: #ifdef PARANOID
                    164:                --busy;
                    165: #endif
1.1       cgd       166:                return;
1.2       cgd       167:        }
1.1       cgd       168: #ifdef DIAGNOSTIC
                    169:        if ((m->m_flags & M_PKTHDR) == 0)
                    170:                panic("ipintr no HDR");
                    171: #endif
                    172:        /*
                    173:         * If no IP addresses have been set yet but the interfaces
                    174:         * are receiving, can't do anything with incoming packets yet.
                    175:         */
                    176:        if (in_ifaddr == NULL)
                    177:                goto bad;
                    178:        ipstat.ips_total++;
                    179:        if (m->m_len < sizeof (struct ip) &&
                    180:            (m = m_pullup(m, sizeof (struct ip))) == 0) {
                    181:                ipstat.ips_toosmall++;
                    182:                goto next;
                    183:        }
                    184:        ip = mtod(m, struct ip *);
                    185:        hlen = ip->ip_hl << 2;
                    186:        if (hlen < sizeof(struct ip)) { /* minimum header length */
                    187:                ipstat.ips_badhlen++;
                    188:                goto bad;
                    189:        }
                    190:        if (hlen > m->m_len) {
                    191:                if ((m = m_pullup(m, hlen)) == 0) {
                    192:                        ipstat.ips_badhlen++;
                    193:                        goto next;
                    194:                }
                    195:                ip = mtod(m, struct ip *);
                    196:        }
                    197:        if (ip->ip_sum = in_cksum(m, hlen)) {
                    198:                ipstat.ips_badsum++;
                    199:                goto bad;
                    200:        }
                    201:
                    202:        /*
                    203:         * Convert fields to host representation.
                    204:         */
                    205:        NTOHS(ip->ip_len);
                    206:        if (ip->ip_len < hlen) {
                    207:                ipstat.ips_badlen++;
                    208:                goto bad;
                    209:        }
                    210:        NTOHS(ip->ip_id);
                    211:        NTOHS(ip->ip_off);
                    212:
                    213:        /*
                    214:         * Check that the amount of data in the buffers
                    215:         * is as at least much as the IP header would have us expect.
                    216:         * Trim mbufs if longer than we expect.
                    217:         * Drop packet if shorter than we expect.
                    218:         */
                    219:        if (m->m_pkthdr.len < ip->ip_len) {
                    220:                ipstat.ips_tooshort++;
                    221:                goto bad;
                    222:        }
                    223:        if (m->m_pkthdr.len > ip->ip_len) {
                    224:                if (m->m_len == m->m_pkthdr.len) {
                    225:                        m->m_len = ip->ip_len;
                    226:                        m->m_pkthdr.len = ip->ip_len;
                    227:                } else
                    228:                        m_adj(m, ip->ip_len - m->m_pkthdr.len);
                    229:        }
                    230:
                    231:        /*
                    232:         * Process options and, if not destined for us,
                    233:         * ship it on.  ip_dooptions returns 1 when an
                    234:         * error was detected (causing an icmp message
                    235:         * to be sent and the original packet to be freed).
                    236:         */
                    237:        ip_nhops = 0;           /* for source routed packets */
                    238:        if (hlen > sizeof (struct ip) && ip_dooptions(m))
                    239:                goto next;
                    240:
                    241:        /*
                    242:         * Check our list of addresses, to see if the packet is for us.
                    243:         */
                    244:        for (ia = in_ifaddr; ia; ia = ia->ia_next) {
                    245: #define        satosin(sa)     ((struct sockaddr_in *)(sa))
                    246:
                    247:                if (IA_SIN(ia)->sin_addr.s_addr == ip->ip_dst.s_addr)
                    248:                        goto ours;
                    249:                if (
                    250: #ifdef DIRECTED_BROADCAST
                    251:                    ia->ia_ifp == m->m_pkthdr.rcvif &&
                    252: #endif
                    253:                    (ia->ia_ifp->if_flags & IFF_BROADCAST)) {
                    254:                        u_long t;
                    255:
                    256:                        if (satosin(&ia->ia_broadaddr)->sin_addr.s_addr ==
                    257:                            ip->ip_dst.s_addr)
                    258:                                goto ours;
                    259:                        if (ip->ip_dst.s_addr == ia->ia_netbroadcast.s_addr)
                    260:                                goto ours;
                    261:                        /*
                    262:                         * Look for all-0's host part (old broadcast addr),
                    263:                         * either for subnet or net.
                    264:                         */
                    265:                        t = ntohl(ip->ip_dst.s_addr);
                    266:                        if (t == ia->ia_subnet)
                    267:                                goto ours;
                    268:                        if (t == ia->ia_net)
                    269:                                goto ours;
                    270:                }
                    271:        }
                    272:        if (ip->ip_dst.s_addr == (u_long)INADDR_BROADCAST)
                    273:                goto ours;
                    274:        if (ip->ip_dst.s_addr == INADDR_ANY)
                    275:                goto ours;
                    276:
                    277:        /*
                    278:         * Not for us; forward if possible and desirable.
                    279:         */
                    280:        if (ipforwarding == 0) {
                    281:                ipstat.ips_cantforward++;
                    282:                m_freem(m);
                    283:        } else
                    284:                ip_forward(m, 0);
                    285:        goto next;
                    286:
                    287: ours:
                    288:        /*
                    289:         * If offset or IP_MF are set, must reassemble.
                    290:         * Otherwise, nothing need be done.
                    291:         * (We could look in the reassembly queue to see
                    292:         * if the packet was previously fragmented,
                    293:         * but it's not worth the time; just let them time out.)
                    294:         */
                    295:        if (ip->ip_off &~ IP_DF) {
                    296:                if (m->m_flags & M_EXT) {               /* XXX */
                    297:                        if ((m = m_pullup(m, sizeof (struct ip))) == 0) {
                    298:                                ipstat.ips_toosmall++;
                    299:                                goto next;
                    300:                        }
                    301:                        ip = mtod(m, struct ip *);
                    302:                }
                    303:                /*
                    304:                 * Look for queue of fragments
                    305:                 * of this datagram.
                    306:                 */
                    307:                for (fp = ipq.next; fp != &ipq; fp = fp->next)
                    308:                        if (ip->ip_id == fp->ipq_id &&
                    309:                            ip->ip_src.s_addr == fp->ipq_src.s_addr &&
                    310:                            ip->ip_dst.s_addr == fp->ipq_dst.s_addr &&
                    311:                            ip->ip_p == fp->ipq_p)
                    312:                                goto found;
                    313:                fp = 0;
                    314: found:
                    315:
                    316:                /*
                    317:                 * Adjust ip_len to not reflect header,
                    318:                 * set ip_mff if more fragments are expected,
                    319:                 * convert offset of this to bytes.
                    320:                 */
                    321:                ip->ip_len -= hlen;
                    322:                ((struct ipasfrag *)ip)->ipf_mff = 0;
                    323:                if (ip->ip_off & IP_MF)
                    324:                        ((struct ipasfrag *)ip)->ipf_mff = 1;
                    325:                ip->ip_off <<= 3;
                    326:
                    327:                /*
                    328:                 * If datagram marked as having more fragments
                    329:                 * or if this is not the first fragment,
                    330:                 * attempt reassembly; if it succeeds, proceed.
                    331:                 */
                    332:                if (((struct ipasfrag *)ip)->ipf_mff || ip->ip_off) {
                    333:                        ipstat.ips_fragments++;
                    334:                        ip = ip_reass((struct ipasfrag *)ip, fp);
                    335:                        if (ip == 0)
                    336:                                goto next;
                    337:                        else
                    338:                                ipstat.ips_reassembled++;
                    339:                        m = dtom(ip);
                    340:                } else
                    341:                        if (fp)
                    342:                                ip_freef(fp);
                    343:        } else
                    344:                ip->ip_len -= hlen;
                    345:
                    346:        /*
                    347:         * Switch out to protocol's input routine.
                    348:         */
                    349:        ipstat.ips_delivered++;
                    350:        (*inetsw[ip_protox[ip->ip_p]].pr_input)(m, hlen);
                    351:        goto next;
                    352: bad:
                    353:        m_freem(m);
                    354:        goto next;
                    355: }
                    356:
                    357: /*
                    358:  * Take incoming datagram fragment and try to
                    359:  * reassemble it into whole datagram.  If a chain for
                    360:  * reassembly of this datagram already exists, then it
                    361:  * is given as fp; otherwise have to make a chain.
                    362:  */
                    363: struct ip *
                    364: ip_reass(ip, fp)
                    365:        register struct ipasfrag *ip;
                    366:        register struct ipq *fp;
                    367: {
                    368:        register struct mbuf *m = dtom(ip);
                    369:        register struct ipasfrag *q;
                    370:        struct mbuf *t;
                    371:        int hlen = ip->ip_hl << 2;
                    372:        int i, next;
                    373:
                    374:        /*
                    375:         * Presence of header sizes in mbufs
                    376:         * would confuse code below.
                    377:         */
                    378:        m->m_data += hlen;
                    379:        m->m_len -= hlen;
                    380:
                    381:        /*
                    382:         * If first fragment to arrive, create a reassembly queue.
                    383:         */
                    384:        if (fp == 0) {
                    385:                if ((t = m_get(M_DONTWAIT, MT_FTABLE)) == NULL)
                    386:                        goto dropfrag;
                    387:                fp = mtod(t, struct ipq *);
                    388:                insque(fp, &ipq);
                    389:                fp->ipq_ttl = IPFRAGTTL;
                    390:                fp->ipq_p = ip->ip_p;
                    391:                fp->ipq_id = ip->ip_id;
                    392:                fp->ipq_next = fp->ipq_prev = (struct ipasfrag *)fp;
                    393:                fp->ipq_src = ((struct ip *)ip)->ip_src;
                    394:                fp->ipq_dst = ((struct ip *)ip)->ip_dst;
                    395:                q = (struct ipasfrag *)fp;
                    396:                goto insert;
                    397:        }
                    398:
                    399:        /*
                    400:         * Find a segment which begins after this one does.
                    401:         */
                    402:        for (q = fp->ipq_next; q != (struct ipasfrag *)fp; q = q->ipf_next)
                    403:                if (q->ip_off > ip->ip_off)
                    404:                        break;
                    405:
                    406:        /*
                    407:         * If there is a preceding segment, it may provide some of
                    408:         * our data already.  If so, drop the data from the incoming
                    409:         * segment.  If it provides all of our data, drop us.
                    410:         */
                    411:        if (q->ipf_prev != (struct ipasfrag *)fp) {
                    412:                i = q->ipf_prev->ip_off + q->ipf_prev->ip_len - ip->ip_off;
                    413:                if (i > 0) {
                    414:                        if (i >= ip->ip_len)
                    415:                                goto dropfrag;
                    416:                        m_adj(dtom(ip), i);
                    417:                        ip->ip_off += i;
                    418:                        ip->ip_len -= i;
                    419:                }
                    420:        }
                    421:
                    422:        /*
                    423:         * While we overlap succeeding segments trim them or,
                    424:         * if they are completely covered, dequeue them.
                    425:         */
                    426:        while (q != (struct ipasfrag *)fp && ip->ip_off + ip->ip_len > q->ip_off) {
                    427:                i = (ip->ip_off + ip->ip_len) - q->ip_off;
                    428:                if (i < q->ip_len) {
                    429:                        q->ip_len -= i;
                    430:                        q->ip_off += i;
                    431:                        m_adj(dtom(q), i);
                    432:                        break;
                    433:                }
                    434:                q = q->ipf_next;
                    435:                m_freem(dtom(q->ipf_prev));
                    436:                ip_deq(q->ipf_prev);
                    437:        }
                    438:
                    439: insert:
                    440:        /*
                    441:         * Stick new segment in its place;
                    442:         * check for complete reassembly.
                    443:         */
                    444:        ip_enq(ip, q->ipf_prev);
                    445:        next = 0;
                    446:        for (q = fp->ipq_next; q != (struct ipasfrag *)fp; q = q->ipf_next) {
                    447:                if (q->ip_off != next)
                    448:                        return (0);
                    449:                next += q->ip_len;
                    450:        }
                    451:        if (q->ipf_prev->ipf_mff)
                    452:                return (0);
                    453:
                    454:        /*
                    455:         * Reassembly is complete; concatenate fragments.
                    456:         */
                    457:        q = fp->ipq_next;
                    458:        m = dtom(q);
                    459:        t = m->m_next;
                    460:        m->m_next = 0;
                    461:        m_cat(m, t);
                    462:        q = q->ipf_next;
                    463:        while (q != (struct ipasfrag *)fp) {
                    464:                t = dtom(q);
                    465:                q = q->ipf_next;
                    466:                m_cat(m, t);
                    467:        }
                    468:
                    469:        /*
                    470:         * Create header for new ip packet by
                    471:         * modifying header of first packet;
                    472:         * dequeue and discard fragment reassembly header.
                    473:         * Make header visible.
                    474:         */
                    475:        ip = fp->ipq_next;
                    476:        ip->ip_len = next;
                    477:        ((struct ip *)ip)->ip_src = fp->ipq_src;
                    478:        ((struct ip *)ip)->ip_dst = fp->ipq_dst;
                    479:        remque(fp);
                    480:        (void) m_free(dtom(fp));
                    481:        m = dtom(ip);
                    482:        m->m_len += (ip->ip_hl << 2);
                    483:        m->m_data -= (ip->ip_hl << 2);
                    484:        /* some debugging cruft by sklower, below, will go away soon */
                    485:        if (m->m_flags & M_PKTHDR) { /* XXX this should be done elsewhere */
                    486:                register int plen = 0;
                    487:                for (t = m; m; m = m->m_next)
                    488:                        plen += m->m_len;
                    489:                t->m_pkthdr.len = plen;
                    490:        }
                    491:        return ((struct ip *)ip);
                    492:
                    493: dropfrag:
                    494:        ipstat.ips_fragdropped++;
                    495:        m_freem(m);
                    496:        return (0);
                    497: }
                    498:
                    499: /*
                    500:  * Free a fragment reassembly header and all
                    501:  * associated datagrams.
                    502:  */
                    503: ip_freef(fp)
                    504:        struct ipq *fp;
                    505: {
                    506:        register struct ipasfrag *q, *p;
                    507:
                    508:        for (q = fp->ipq_next; q != (struct ipasfrag *)fp; q = p) {
                    509:                p = q->ipf_next;
                    510:                ip_deq(q);
                    511:                m_freem(dtom(q));
                    512:        }
                    513:        remque(fp);
                    514:        (void) m_free(dtom(fp));
                    515: }
                    516:
                    517: /*
                    518:  * Put an ip fragment on a reassembly chain.
                    519:  * Like insque, but pointers in middle of structure.
                    520:  */
                    521: ip_enq(p, prev)
                    522:        register struct ipasfrag *p, *prev;
                    523: {
                    524:
                    525:        p->ipf_prev = prev;
                    526:        p->ipf_next = prev->ipf_next;
                    527:        prev->ipf_next->ipf_prev = p;
                    528:        prev->ipf_next = p;
                    529: }
                    530:
                    531: /*
                    532:  * To ip_enq as remque is to insque.
                    533:  */
                    534: ip_deq(p)
                    535:        register struct ipasfrag *p;
                    536: {
                    537:
                    538:        p->ipf_prev->ipf_next = p->ipf_next;
                    539:        p->ipf_next->ipf_prev = p->ipf_prev;
                    540: }
                    541:
                    542: /*
                    543:  * IP timer processing;
                    544:  * if a timer expires on a reassembly
                    545:  * queue, discard it.
                    546:  */
                    547: ip_slowtimo()
                    548: {
                    549:        register struct ipq *fp;
                    550:        int s = splnet();
                    551:
                    552:        fp = ipq.next;
                    553:        if (fp == 0) {
                    554:                splx(s);
                    555:                return;
                    556:        }
                    557:        while (fp != &ipq) {
                    558:                --fp->ipq_ttl;
                    559:                fp = fp->next;
                    560:                if (fp->prev->ipq_ttl == 0) {
                    561:                        ipstat.ips_fragtimeout++;
                    562:                        ip_freef(fp->prev);
                    563:                }
                    564:        }
                    565:        splx(s);
                    566: }
                    567:
                    568: /*
                    569:  * Drain off all datagram fragments.
                    570:  */
                    571: ip_drain()
                    572: {
                    573:
                    574:        while (ipq.next != &ipq) {
                    575:                ipstat.ips_fragdropped++;
                    576:                ip_freef(ipq.next);
                    577:        }
                    578: }
                    579:
                    580: extern struct in_ifaddr *ifptoia();
                    581: struct in_ifaddr *ip_rtaddr();
                    582:
                    583: /*
                    584:  * Do option processing on a datagram,
                    585:  * possibly discarding it if bad options are encountered,
                    586:  * or forwarding it if source-routed.
                    587:  * Returns 1 if packet has been forwarded/freed,
                    588:  * 0 if the packet should be processed further.
                    589:  */
                    590: ip_dooptions(m)
                    591:        struct mbuf *m;
                    592: {
                    593:        register struct ip *ip = mtod(m, struct ip *);
                    594:        register u_char *cp;
                    595:        register struct ip_timestamp *ipt;
                    596:        register struct in_ifaddr *ia;
                    597:        int opt, optlen, cnt, off, code, type = ICMP_PARAMPROB, forward = 0;
                    598:        struct in_addr *sin;
                    599:        n_time ntime;
                    600:
                    601:        cp = (u_char *)(ip + 1);
                    602:        cnt = (ip->ip_hl << 2) - sizeof (struct ip);
                    603:        for (; cnt > 0; cnt -= optlen, cp += optlen) {
                    604:                opt = cp[IPOPT_OPTVAL];
                    605:                if (opt == IPOPT_EOL)
                    606:                        break;
                    607:                if (opt == IPOPT_NOP)
                    608:                        optlen = 1;
                    609:                else {
                    610:                        optlen = cp[IPOPT_OLEN];
                    611:                        if (optlen <= 0 || optlen > cnt) {
                    612:                                code = &cp[IPOPT_OLEN] - (u_char *)ip;
                    613:                                goto bad;
                    614:                        }
                    615:                }
                    616:                switch (opt) {
                    617:
                    618:                default:
                    619:                        break;
                    620:
                    621:                /*
                    622:                 * Source routing with record.
                    623:                 * Find interface with current destination address.
                    624:                 * If none on this machine then drop if strictly routed,
                    625:                 * or do nothing if loosely routed.
                    626:                 * Record interface address and bring up next address
                    627:                 * component.  If strictly routed make sure next
                    628:                 * address is on directly accessible net.
                    629:                 */
                    630:                case IPOPT_LSRR:
                    631:                case IPOPT_SSRR:
                    632:                        if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) {
                    633:                                code = &cp[IPOPT_OFFSET] - (u_char *)ip;
                    634:                                goto bad;
                    635:                        }
                    636:                        ipaddr.sin_addr = ip->ip_dst;
                    637:                        ia = (struct in_ifaddr *)
                    638:                                ifa_ifwithaddr((struct sockaddr *)&ipaddr);
                    639:                        if (ia == 0) {
                    640:                                if (opt == IPOPT_SSRR) {
                    641:                                        type = ICMP_UNREACH;
                    642:                                        code = ICMP_UNREACH_SRCFAIL;
                    643:                                        goto bad;
                    644:                                }
                    645:                                /*
                    646:                                 * Loose routing, and not at next destination
                    647:                                 * yet; nothing to do except forward.
                    648:                                 */
                    649:                                break;
                    650:                        }
                    651:                        off--;                  /* 0 origin */
                    652:                        if (off > optlen - sizeof(struct in_addr)) {
                    653:                                /*
                    654:                                 * End of source route.  Should be for us.
                    655:                                 */
                    656:                                save_rte(cp, ip->ip_src);
                    657:                                break;
                    658:                        }
                    659:                        /*
                    660:                         * locate outgoing interface
                    661:                         */
                    662:                        bcopy((caddr_t)(cp + off), (caddr_t)&ipaddr.sin_addr,
                    663:                            sizeof(ipaddr.sin_addr));
                    664:                        if (opt == IPOPT_SSRR) {
                    665: #define        INA     struct in_ifaddr *
                    666: #define        SA      struct sockaddr *
                    667:                            if ((ia = (INA)ifa_ifwithdstaddr((SA)&ipaddr)) == 0)
                    668:                                ia = in_iaonnetof(in_netof(ipaddr.sin_addr));
                    669:                        } else
                    670:                                ia = ip_rtaddr(ipaddr.sin_addr);
                    671:                        if (ia == 0) {
                    672:                                type = ICMP_UNREACH;
                    673:                                code = ICMP_UNREACH_SRCFAIL;
                    674:                                goto bad;
                    675:                        }
                    676:                        ip->ip_dst = ipaddr.sin_addr;
                    677:                        bcopy((caddr_t)&(IA_SIN(ia)->sin_addr),
                    678:                            (caddr_t)(cp + off), sizeof(struct in_addr));
                    679:                        cp[IPOPT_OFFSET] += sizeof(struct in_addr);
                    680:                        forward = 1;
                    681:                        break;
                    682:
                    683:                case IPOPT_RR:
                    684:                        if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) {
                    685:                                code = &cp[IPOPT_OFFSET] - (u_char *)ip;
                    686:                                goto bad;
                    687:                        }
                    688:                        /*
                    689:                         * If no space remains, ignore.
                    690:                         */
                    691:                        off--;                  /* 0 origin */
                    692:                        if (off > optlen - sizeof(struct in_addr))
                    693:                                break;
                    694:                        bcopy((caddr_t)(&ip->ip_dst), (caddr_t)&ipaddr.sin_addr,
                    695:                            sizeof(ipaddr.sin_addr));
                    696:                        /*
                    697:                         * locate outgoing interface; if we're the destination,
                    698:                         * use the incoming interface (should be same).
                    699:                         */
                    700:                        if ((ia = (INA)ifa_ifwithaddr((SA)&ipaddr)) == 0 &&
                    701:                            (ia = ip_rtaddr(ipaddr.sin_addr)) == 0) {
                    702:                                type = ICMP_UNREACH;
                    703:                                code = ICMP_UNREACH_HOST;
                    704:                                goto bad;
                    705:                        }
                    706:                        bcopy((caddr_t)&(IA_SIN(ia)->sin_addr),
                    707:                            (caddr_t)(cp + off), sizeof(struct in_addr));
                    708:                        cp[IPOPT_OFFSET] += sizeof(struct in_addr);
                    709:                        break;
                    710:
                    711:                case IPOPT_TS:
                    712:                        code = cp - (u_char *)ip;
                    713:                        ipt = (struct ip_timestamp *)cp;
                    714:                        if (ipt->ipt_len < 5)
                    715:                                goto bad;
                    716:                        if (ipt->ipt_ptr > ipt->ipt_len - sizeof (long)) {
                    717:                                if (++ipt->ipt_oflw == 0)
                    718:                                        goto bad;
                    719:                                break;
                    720:                        }
                    721:                        sin = (struct in_addr *)(cp + ipt->ipt_ptr - 1);
                    722:                        switch (ipt->ipt_flg) {
                    723:
                    724:                        case IPOPT_TS_TSONLY:
                    725:                                break;
                    726:
                    727:                        case IPOPT_TS_TSANDADDR:
                    728:                                if (ipt->ipt_ptr + sizeof(n_time) +
                    729:                                    sizeof(struct in_addr) > ipt->ipt_len)
                    730:                                        goto bad;
                    731:                                ia = ifptoia(m->m_pkthdr.rcvif);
                    732:                                bcopy((caddr_t)&IA_SIN(ia)->sin_addr,
                    733:                                    (caddr_t)sin, sizeof(struct in_addr));
                    734:                                ipt->ipt_ptr += sizeof(struct in_addr);
                    735:                                break;
                    736:
                    737:                        case IPOPT_TS_PRESPEC:
                    738:                                if (ipt->ipt_ptr + sizeof(n_time) +
                    739:                                    sizeof(struct in_addr) > ipt->ipt_len)
                    740:                                        goto bad;
                    741:                                bcopy((caddr_t)sin, (caddr_t)&ipaddr.sin_addr,
                    742:                                    sizeof(struct in_addr));
                    743:                                if (ifa_ifwithaddr((SA)&ipaddr) == 0)
                    744:                                        continue;
                    745:                                ipt->ipt_ptr += sizeof(struct in_addr);
                    746:                                break;
                    747:
                    748:                        default:
                    749:                                goto bad;
                    750:                        }
                    751:                        ntime = iptime();
                    752:                        bcopy((caddr_t)&ntime, (caddr_t)cp + ipt->ipt_ptr - 1,
                    753:                            sizeof(n_time));
                    754:                        ipt->ipt_ptr += sizeof(n_time);
                    755:                }
                    756:        }
                    757:        if (forward) {
                    758:                ip_forward(m, 1);
                    759:                return (1);
                    760:        } else
                    761:                return (0);
                    762: bad:
                    763:        icmp_error(m, type, code);
                    764:        return (1);
                    765: }
                    766:
                    767: /*
                    768:  * Given address of next destination (final or next hop),
                    769:  * return internet address info of interface to be used to get there.
                    770:  */
                    771: struct in_ifaddr *
                    772: ip_rtaddr(dst)
                    773:         struct in_addr dst;
                    774: {
                    775:        register struct sockaddr_in *sin;
                    776:
                    777:        sin = (struct sockaddr_in *) &ipforward_rt.ro_dst;
                    778:
                    779:        if (ipforward_rt.ro_rt == 0 || dst.s_addr != sin->sin_addr.s_addr) {
                    780:                if (ipforward_rt.ro_rt) {
                    781:                        RTFREE(ipforward_rt.ro_rt);
                    782:                        ipforward_rt.ro_rt = 0;
                    783:                }
                    784:                sin->sin_family = AF_INET;
                    785:                sin->sin_len = sizeof(*sin);
                    786:                sin->sin_addr = dst;
                    787:
                    788:                rtalloc(&ipforward_rt);
                    789:        }
                    790:        if (ipforward_rt.ro_rt == 0)
                    791:                return ((struct in_ifaddr *)0);
                    792:        return ((struct in_ifaddr *) ipforward_rt.ro_rt->rt_ifa);
                    793: }
                    794:
                    795: /*
                    796:  * Save incoming source route for use in replies,
                    797:  * to be picked up later by ip_srcroute if the receiver is interested.
                    798:  */
                    799: save_rte(option, dst)
                    800:        u_char *option;
                    801:        struct in_addr dst;
                    802: {
                    803:        unsigned olen;
                    804:
                    805:        olen = option[IPOPT_OLEN];
                    806: #ifdef DIAGNOSTIC
                    807:        if (ipprintfs)
                    808:                printf("save_rte: olen %d\n", olen);
                    809: #endif
                    810:        if (olen > sizeof(ip_srcrt) - (1 + sizeof(dst)))
                    811:                return;
                    812:        bcopy((caddr_t)option, (caddr_t)ip_srcrt.srcopt, olen);
                    813:        ip_nhops = (olen - IPOPT_OFFSET - 1) / sizeof(struct in_addr);
                    814:        ip_srcrt.dst = dst;
                    815: }
                    816:
                    817: /*
                    818:  * Retrieve incoming source route for use in replies,
                    819:  * in the same form used by setsockopt.
                    820:  * The first hop is placed before the options, will be removed later.
                    821:  */
                    822: struct mbuf *
                    823: ip_srcroute()
                    824: {
                    825:        register struct in_addr *p, *q;
                    826:        register struct mbuf *m;
                    827:
                    828:        if (ip_nhops == 0)
                    829:                return ((struct mbuf *)0);
                    830:        m = m_get(M_DONTWAIT, MT_SOOPTS);
                    831:        if (m == 0)
                    832:                return ((struct mbuf *)0);
                    833:
                    834: #define OPTSIZ (sizeof(ip_srcrt.nop) + sizeof(ip_srcrt.srcopt))
                    835:
                    836:        /* length is (nhops+1)*sizeof(addr) + sizeof(nop + srcrt header) */
                    837:        m->m_len = ip_nhops * sizeof(struct in_addr) + sizeof(struct in_addr) +
                    838:            OPTSIZ;
                    839: #ifdef DIAGNOSTIC
                    840:        if (ipprintfs)
                    841:                printf("ip_srcroute: nhops %d mlen %d", ip_nhops, m->m_len);
                    842: #endif
                    843:
                    844:        /*
                    845:         * First save first hop for return route
                    846:         */
                    847:        p = &ip_srcrt.route[ip_nhops - 1];
                    848:        *(mtod(m, struct in_addr *)) = *p--;
                    849: #ifdef DIAGNOSTIC
                    850:        if (ipprintfs)
                    851:                printf(" hops %lx", ntohl(mtod(m, struct in_addr *)->s_addr));
                    852: #endif
                    853:
                    854:        /*
                    855:         * Copy option fields and padding (nop) to mbuf.
                    856:         */
                    857:        ip_srcrt.nop = IPOPT_NOP;
                    858:        ip_srcrt.srcopt[IPOPT_OFFSET] = IPOPT_MINOFF;
                    859:        bcopy((caddr_t)&ip_srcrt.nop,
                    860:            mtod(m, caddr_t) + sizeof(struct in_addr), OPTSIZ);
                    861:        q = (struct in_addr *)(mtod(m, caddr_t) +
                    862:            sizeof(struct in_addr) + OPTSIZ);
                    863: #undef OPTSIZ
                    864:        /*
                    865:         * Record return path as an IP source route,
                    866:         * reversing the path (pointers are now aligned).
                    867:         */
                    868:        while (p >= ip_srcrt.route) {
                    869: #ifdef DIAGNOSTIC
                    870:                if (ipprintfs)
                    871:                        printf(" %lx", ntohl(q->s_addr));
                    872: #endif
                    873:                *q++ = *p--;
                    874:        }
                    875:        /*
                    876:         * Last hop goes to final destination.
                    877:         */
                    878:        *q = ip_srcrt.dst;
                    879: #ifdef DIAGNOSTIC
                    880:        if (ipprintfs)
                    881:                printf(" %lx\n", ntohl(q->s_addr));
                    882: #endif
                    883:        return (m);
                    884: }
                    885:
                    886: /*
                    887:  * Strip out IP options, at higher
                    888:  * level protocol in the kernel.
                    889:  * Second argument is buffer to which options
                    890:  * will be moved, and return value is their length.
                    891:  * XXX should be deleted; last arg currently ignored.
                    892:  */
                    893: ip_stripoptions(m, mopt)
                    894:        register struct mbuf *m;
                    895:        struct mbuf *mopt;
                    896: {
                    897:        register int i;
                    898:        struct ip *ip = mtod(m, struct ip *);
                    899:        register caddr_t opts;
                    900:        int olen;
                    901:
                    902:        olen = (ip->ip_hl<<2) - sizeof (struct ip);
                    903:        opts = (caddr_t)(ip + 1);
                    904:        i = m->m_len - (sizeof (struct ip) + olen);
                    905:        bcopy(opts  + olen, opts, (unsigned)i);
                    906:        m->m_len -= olen;
                    907:        if (m->m_flags & M_PKTHDR)
                    908:                m->m_pkthdr.len -= olen;
                    909:        ip->ip_hl = sizeof(struct ip) >> 2;
                    910: }
                    911:
                    912: u_char inetctlerrmap[PRC_NCMDS] = {
                    913:        0,              0,              0,              0,
                    914:        0,              EMSGSIZE,       EHOSTDOWN,      EHOSTUNREACH,
                    915:        EHOSTUNREACH,   EHOSTUNREACH,   ECONNREFUSED,   ECONNREFUSED,
                    916:        EMSGSIZE,       EHOSTUNREACH,   0,              0,
                    917:        0,              0,              0,              0,
                    918:        ENOPROTOOPT
                    919: };
                    920:
                    921: /*
                    922:  * Forward a packet.  If some error occurs return the sender
                    923:  * an icmp packet.  Note we can't always generate a meaningful
                    924:  * icmp message because icmp doesn't have a large enough repertoire
                    925:  * of codes and types.
                    926:  *
                    927:  * If not forwarding, just drop the packet.  This could be confusing
                    928:  * if ipforwarding was zero but some routing protocol was advancing
                    929:  * us as a gateway to somewhere.  However, we must let the routing
                    930:  * protocol deal with that.
                    931:  *
                    932:  * The srcrt parameter indicates whether the packet is being forwarded
                    933:  * via a source route.
                    934:  */
                    935: ip_forward(m, srcrt)
                    936:        struct mbuf *m;
                    937:        int srcrt;
                    938: {
                    939:        register struct ip *ip = mtod(m, struct ip *);
                    940:        register struct sockaddr_in *sin;
                    941:        register struct rtentry *rt;
                    942:        int error, type = 0, code;
                    943:        struct mbuf *mcopy;
                    944:        struct in_addr dest;
                    945:
                    946:        dest.s_addr = 0;
                    947: #ifdef DIAGNOSTIC
                    948:        if (ipprintfs)
                    949:                printf("forward: src %x dst %x ttl %x\n", ip->ip_src,
                    950:                        ip->ip_dst, ip->ip_ttl);
                    951: #endif
                    952:        if (m->m_flags & M_BCAST || in_canforward(ip->ip_dst) == 0) {
                    953:                ipstat.ips_cantforward++;
                    954:                m_freem(m);
                    955:                return;
                    956:        }
                    957:        HTONS(ip->ip_id);
                    958:        if (ip->ip_ttl <= IPTTLDEC) {
                    959:                icmp_error(m, ICMP_TIMXCEED, ICMP_TIMXCEED_INTRANS, dest);
                    960:                return;
                    961:        }
                    962:        ip->ip_ttl -= IPTTLDEC;
                    963:
                    964:        sin = (struct sockaddr_in *)&ipforward_rt.ro_dst;
                    965:        if ((rt = ipforward_rt.ro_rt) == 0 ||
                    966:            ip->ip_dst.s_addr != sin->sin_addr.s_addr) {
                    967:                if (ipforward_rt.ro_rt) {
                    968:                        RTFREE(ipforward_rt.ro_rt);
                    969:                        ipforward_rt.ro_rt = 0;
                    970:                }
                    971:                sin->sin_family = AF_INET;
                    972:                sin->sin_len = sizeof(*sin);
                    973:                sin->sin_addr = ip->ip_dst;
                    974:
                    975:                rtalloc(&ipforward_rt);
                    976:                if (ipforward_rt.ro_rt == 0) {
                    977:                        icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_HOST, dest);
                    978:                        return;
                    979:                }
                    980:                rt = ipforward_rt.ro_rt;
                    981:        }
                    982:
                    983:        /*
                    984:         * Save at most 64 bytes of the packet in case
                    985:         * we need to generate an ICMP message to the src.
                    986:         */
                    987:        mcopy = m_copy(m, 0, imin((int)ip->ip_len, 64));
                    988:
                    989: #ifdef GATEWAY
                    990:        ip_ifmatrix[rt->rt_ifp->if_index +
                    991:             if_index * m->m_pkthdr.rcvif->if_index]++;
                    992: #endif
                    993:        /*
                    994:         * If forwarding packet using same interface that it came in on,
                    995:         * perhaps should send a redirect to sender to shortcut a hop.
                    996:         * Only send redirect if source is sending directly to us,
                    997:         * and if packet was not source routed (or has any options).
                    998:         * Also, don't send redirect if forwarding using a default route
                    999:         * or a route modified by a redirect.
                   1000:         */
                   1001: #define        satosin(sa)     ((struct sockaddr_in *)(sa))
                   1002:        if (rt->rt_ifp == m->m_pkthdr.rcvif &&
                   1003:            (rt->rt_flags & (RTF_DYNAMIC|RTF_MODIFIED)) == 0 &&
                   1004:            satosin(rt_key(rt))->sin_addr.s_addr != 0 &&
                   1005:            ipsendredirects && !srcrt) {
                   1006:                struct in_ifaddr *ia;
                   1007:                u_long src = ntohl(ip->ip_src.s_addr);
                   1008:                u_long dst = ntohl(ip->ip_dst.s_addr);
                   1009:
                   1010:                if ((ia = ifptoia(m->m_pkthdr.rcvif)) &&
                   1011:                   (src & ia->ia_subnetmask) == ia->ia_subnet) {
                   1012:                    if (rt->rt_flags & RTF_GATEWAY)
                   1013:                        dest = satosin(rt->rt_gateway)->sin_addr;
                   1014:                    else
                   1015:                        dest = ip->ip_dst;
                   1016:                    /*
                   1017:                     * If the destination is reached by a route to host,
                   1018:                     * is on a subnet of a local net, or is directly
                   1019:                     * on the attached net (!), use host redirect.
                   1020:                     * (We may be the correct first hop for other subnets.)
                   1021:                     */
                   1022: #define        RTA(rt) ((struct in_ifaddr *)(rt->rt_ifa))
                   1023:                    type = ICMP_REDIRECT;
                   1024:                    if ((rt->rt_flags & RTF_HOST) ||
                   1025:                        (rt->rt_flags & RTF_GATEWAY) == 0)
                   1026:                            code = ICMP_REDIRECT_HOST;
                   1027:                    else if (RTA(rt)->ia_subnetmask != RTA(rt)->ia_netmask &&
                   1028:                        (dst & RTA(rt)->ia_netmask) ==  RTA(rt)->ia_net)
                   1029:                            code = ICMP_REDIRECT_HOST;
                   1030:                    else
                   1031:                            code = ICMP_REDIRECT_NET;
                   1032: #ifdef DIAGNOSTIC
                   1033:                    if (ipprintfs)
                   1034:                        printf("redirect (%d) to %x\n", code, dest.s_addr);
                   1035: #endif
                   1036:                }
                   1037:        }
                   1038:
                   1039:        error = ip_output(m, (struct mbuf *)0, &ipforward_rt, IP_FORWARDING);
                   1040:        if (error)
                   1041:                ipstat.ips_cantforward++;
                   1042:        else {
                   1043:                ipstat.ips_forward++;
                   1044:                if (type)
                   1045:                        ipstat.ips_redirectsent++;
                   1046:                else {
                   1047:                        if (mcopy)
                   1048:                                m_freem(mcopy);
                   1049:                        return;
                   1050:                }
                   1051:        }
                   1052:        if (mcopy == NULL)
                   1053:                return;
                   1054:        switch (error) {
                   1055:
                   1056:        case 0:                         /* forwarded, but need redirect */
                   1057:                /* type, code set above */
                   1058:                break;
                   1059:
                   1060:        case ENETUNREACH:               /* shouldn't happen, checked above */
                   1061:        case EHOSTUNREACH:
                   1062:        case ENETDOWN:
                   1063:        case EHOSTDOWN:
                   1064:        default:
                   1065:                type = ICMP_UNREACH;
                   1066:                code = ICMP_UNREACH_HOST;
                   1067:                break;
                   1068:
                   1069:        case EMSGSIZE:
                   1070:                type = ICMP_UNREACH;
                   1071:                code = ICMP_UNREACH_NEEDFRAG;
                   1072:                ipstat.ips_cantfrag++;
                   1073:                break;
                   1074:
                   1075:        case ENOBUFS:
                   1076:                type = ICMP_SOURCEQUENCH;
                   1077:                code = 0;
                   1078:                break;
                   1079:        }
                   1080:        icmp_error(mcopy, type, code, dest);
                   1081: }

CVSweb <webmaster@jp.NetBSD.org>