|
|
1.1 root 1: /*
2: * Copyright (c) 1999 Apple Computer, Inc. All rights reserved.
3: *
4: * @APPLE_LICENSE_HEADER_START@
5: *
6: * Portions Copyright (c) 1999 Apple Computer, Inc. All Rights
7: * Reserved. This file contains Original Code and/or Modifications of
8: * Original Code as defined in and that are subject to the Apple Public
9: * Source License Version 1.1 (the "License"). You may not use this file
10: * except in compliance with the License. Please obtain a copy of the
11: * License at http://www.apple.com/publicsource and read it before using
12: * this file.
13: *
14: * The Original Code and all software distributed under the License are
15: * distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY KIND, EITHER
16: * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
17: * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
18: * FITNESS FOR A PARTICULAR PURPOSE OR NON-INFRINGEMENT. Please see the
19: * License for the specific language governing rights and limitations
20: * under the License.
21: *
22: * @APPLE_LICENSE_HEADER_END@
23: */
24:
25: /* Copyright (c) 1995 NeXT Computer, Inc. All Rights Reserved */
26: /*
27: * Copyright (c) 1982, 1986, 1988, 1993
28: * The Regents of the University of California. All rights reserved.
29: *
30: * Redistribution and use in source and binary forms, with or without
31: * modification, are permitted provided that the following conditions
32: * are met:
33: * 1. Redistributions of source code must retain the above copyright
34: * notice, this list of conditions and the following disclaimer.
35: * 2. Redistributions in binary form must reproduce the above copyright
36: * notice, this list of conditions and the following disclaimer in the
37: * documentation and/or other materials provided with the distribution.
38: * 3. All advertising materials mentioning features or use of this software
39: * must display the following acknowledgement:
40: * This product includes software developed by the University of
41: * California, Berkeley and its contributors.
42: * 4. Neither the name of the University nor the names of its contributors
43: * may be used to endorse or promote products derived from this software
44: * without specific prior written permission.
45: *
46: * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
47: * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
48: * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
49: * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
50: * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
51: * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
52: * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
53: * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
54: * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
55: * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
56: * SUCH DAMAGE.
57: *
58: * @(#)ip_input.c 8.2 (Berkeley) 1/4/94
59: */
60:
61: #include <sys/param.h>
62: #include <sys/systm.h>
63: #include <sys/malloc.h>
64: #include <sys/mbuf.h>
65: #include <sys/domain.h>
66: #include <sys/protosw.h>
67: #include <sys/socket.h>
68: #include <sys/errno.h>
69: #include <sys/time.h>
70: #include <sys/kernel.h>
71:
72: #include <net/if.h>
73: #include <net/route.h>
74:
75: #include <netinet/in.h>
76: #include <netinet/in_systm.h>
77: #include <netinet/ip.h>
78: #include <netinet/in_pcb.h>
79: #include <netinet/in_var.h>
80: #include <netinet/ip_var.h>
81: #include <netinet/ip_icmp.h>
82:
83:
84: #if NEXT
85: #import <netinet/if_ether.h>
86: #import <netinet/udp.h>
87: #import <netinet/udp_var.h>
88: #import <netinet/bootp.h>
89: #import <kern/kdebug.h>
90:
91: #if KDEBUG
92:
93: #define DBG_LAYER_BEG NETDBG_CODE(DBG_NETIP, 0)
94: #define DBG_LAYER_END NETDBG_CODE(DBG_NETIP, 2)
95: #define DBG_FNC_IP_INPUT NETDBG_CODE(DBG_NETIP, (2 << 8))
96:
97: #endif
98:
99: #endif NEXT
100:
101: #ifndef IPFORWARDING
102: #if GATEWAY
103: #define IPFORWARDING 1 /* forward IP packets not for us */
104: #else /* GATEWAY */
105: #define IPFORWARDING 0 /* don't forward IP packets not for us */
106: #endif /* GATEWAY */
107: #endif /* IPFORWARDING */
108: #ifndef IPSENDREDIRECTS
109: #define IPSENDREDIRECTS 1
110: #endif
111: int ipforwarding = IPFORWARDING;
112: int ipsendredirects = IPSENDREDIRECTS;
113: int ip_defttl = IPDEFTTL;
114: #if DIAGNOSTIC
115: int ipprintfs = 0;
116: #endif
117:
118: struct ipstat ipstat;
119: struct ipq ipq; /* ip reass. queue */
120: u_short ip_id; /* ip packet ctr, for ids */
121: int ip_defttl; /* default IP ttl */
122:
123: extern struct domain inetdomain;
124: extern struct protosw inetsw[];
125: u_char ip_protox[IPPROTO_MAX];
126: int ipqmaxlen = IFQ_MAXLEN;
127: struct in_ifaddr *in_ifaddr; /* first inet address */
128: struct ifqueue ipintrq;
129:
130:
131:
132: /*
133: * We need to save the IP options in case a protocol wants to respond
134: * to an incoming packet over the same route if the packet got here
135: * using IP source routing. This allows connection establishment and
136: * maintenance when the remote end is on a network that is not known
137: * to us.
138: */
139: int ip_nhops = 0;
140: static struct ip_srcrt {
141: struct in_addr dst; /* final destination */
142: char nop; /* one NOP to align */
143: char srcopt[IPOPT_OFFSET + 1]; /* OPTVAL, OLEN and OFFSET */
144: struct in_addr route[MAX_IPOPTLEN/sizeof(struct in_addr)];
145: } ip_srcrt;
146:
147: #if GATEWAY
148: extern int if_index;
149: u_long *ip_ifmatrix;
150: #endif
151:
152: static void save_rte __P((u_char *, struct in_addr));
153: /*
154: * IP initialization: fill in IP protocol switch table.
155: * All protocols not implemented in kernel go to raw IP protocol handler.
156: */
157: void
158: ip_init()
159: {
160: register struct protosw *pr;
161: register int i;
162:
163: pr = pffindproto(PF_INET, IPPROTO_RAW, SOCK_RAW);
164: if (pr == 0)
165: panic("ip_init");
166: for (i = 0; i < IPPROTO_MAX; i++)
167: ip_protox[i] = pr - inetsw;
168: for (pr = inetdomain.dom_protosw;
169: pr < inetdomain.dom_protoswNPROTOSW; pr++)
170: if (pr->pr_domain->dom_family == PF_INET &&
171: pr->pr_protocol && pr->pr_protocol != IPPROTO_RAW)
172: ip_protox[pr->pr_protocol] = pr - inetsw;
173: ipq.next = ipq.prev = &ipq;
174: ip_id = time.tv_sec & 0xffff;
175: ipintrq.ifq_maxlen = ipqmaxlen;
176: #if GATEWAY
177: i = (if_index + 1) * (if_index + 1) * sizeof (u_long);
178: ip_ifmatrix = (u_long *) _MALLOC(i, M_RTABLE, M_WAITOK);
179: bzero((char *)ip_ifmatrix, i);
180: #endif
181: }
182:
183: struct sockaddr_in ipaddr = { sizeof(ipaddr), AF_INET };
184: struct route ipforward_rt;
185:
186: /*
187: * Ip input routine. Checksum and byte swap header. If fragmented
188: * try to reassemble. Process options. Pass to next level.
189: */
190: void
191: ipintr()
192: {
193: register struct ip *ip;
194: register struct mbuf *m;
195: register struct ipq *fp;
196: register struct in_ifaddr *ia;
197: struct ifnet *ifp;
198: int hlen, s;
199:
200:
201: KERNEL_DEBUG(DBG_FNC_IP_INPUT | DBG_FUNC_START, 0,0,0,0,0);
202:
203: next:
204: /*
205: * Get next datagram off input queue and get IP header
206: * in first mbuf.
207: */
208: s = splimp();
209: IF_DEQUEUE(&ipintrq, m);
210: splx(s);
211: if (m == 0)
212: {
213: KERNEL_DEBUG(DBG_FNC_IP_INPUT | DBG_FUNC_END, 0,0,0,0,0);
214: return;
215: }
216:
217: #if DIAGNOSTIC
218: if ((m->m_flags & M_PKTHDR) == 0)
219: panic("ipintr no HDR");
220: #endif
221: ifp = m->m_pkthdr.rcvif;
222:
223: /*
224: * If no IP addresses have been set yet but the interfaces
225: * are receiving, can't do anything with incoming packets yet.
226: */
227: if (in_ifaddr == NULL)
228: goto bad;
229: ipstat.ips_total++;
230:
231: if (m->m_len < sizeof (struct ip) &&
232: (m = m_pullup(m, sizeof (struct ip))) == 0) {
233: ipstat.ips_toosmall++;
234: goto next;
235: }
236: ip = mtod(m, struct ip *);
237:
238: KERNEL_DEBUG(DBG_LAYER_BEG, ip->ip_dst.s_addr,
239: ip->ip_src.s_addr, ip->ip_p, ip->ip_off, ip->ip_len);
240:
241: if (ip->ip_v != IPVERSION) {
242: ipstat.ips_badvers++;
243: goto bad;
244: }
245: hlen = ip->ip_hl << 2;
246: if (hlen < sizeof(struct ip)) { /* minimum header length */
247: ipstat.ips_badhlen++;
248: goto bad;
249: }
250: if (hlen > m->m_len) {
251: if ((m = m_pullup(m, hlen)) == 0) {
252: ipstat.ips_badhlen++;
253: goto next;
254: }
255: ip = mtod(m, struct ip *);
256: }
257: if (ip->ip_sum = in_cksum(m, hlen)) {
258: ipstat.ips_badsum++;
259: goto bad;
260: }
261:
262: /*
263: * Convert fields to host representation.
264: */
265: NTOHS(ip->ip_len);
266: if (ip->ip_len < hlen) {
267: ipstat.ips_badlen++;
268: goto bad;
269: }
270: NTOHS(ip->ip_id);
271: NTOHS(ip->ip_off);
272:
273: /*
274: * Check that the amount of data in the buffers
275: * is as at least much as the IP header would have us expect.
276: * Trim mbufs if longer than we expect.
277: * Drop packet if shorter than we expect.
278: */
279: if (m->m_pkthdr.len < ip->ip_len) {
280: ipstat.ips_tooshort++;
281: goto bad;
282: }
283: if (m->m_pkthdr.len > ip->ip_len) {
284: if (m->m_len == m->m_pkthdr.len) {
285: m->m_len = ip->ip_len;
286: m->m_pkthdr.len = ip->ip_len;
287: } else
288: m_adj(m, ip->ip_len - m->m_pkthdr.len);
289: }
290:
291: /*
292: * Process options and, if not destined for us,
293: * ship it on. ip_dooptions returns 1 when an
294: * error was detected (causing an icmp message
295: * to be sent and the original packet to be freed).
296: */
297: ip_nhops = 0; /* for source routed packets */
298: if (hlen > sizeof (struct ip) && ip_dooptions(m))
299: goto next;
300:
301: #if NEXT
302: /*
303: * Let BOOTP packets through on autoconfiguring interfaces
304: * regardless of the address.
305: */
306: if (ifp && (ifp->if_eflags & IFEF_AUTOCONF)) {
307: struct udpiphdr *ui;
308:
309: /*
310: * Get IP and UDP header together in first mbuf.
311: */
312: if ((m->m_len >= sizeof (struct udpiphdr))
313: || (m = m_pullup(m, sizeof (struct udpiphdr))) != 0) {
314: ui = mtod(m, struct udpiphdr *);
315: if (ui->ui_pr == IPPROTO_UDP &&
316: ui->ui_dport == ntohs(IPPORT_BOOTPC))
317: goto ours;
318: }
319: }
320: #endif NEXT
321:
322: /*
323: * Check our list of addresses, to see if the packet is for us.
324: */
325: for (ia = in_ifaddr; ia; ia = ia->ia_next) {
326: #define satosin(sa) ((struct sockaddr_in *)(sa))
327:
328: if (IA_SIN(ia)->sin_addr.s_addr == ip->ip_dst.s_addr)
329: goto ours;
330: if (
331: #ifdef DIRECTED_BROADCAST
332: ia->ia_ifp == m->m_pkthdr.rcvif &&
333: #endif
334: (ia->ia_ifp->if_flags & IFF_BROADCAST)) {
335: u_long t;
336:
337: if (satosin(&ia->ia_broadaddr)->sin_addr.s_addr ==
338: ip->ip_dst.s_addr)
339: goto ours;
340: if (ip->ip_dst.s_addr == ia->ia_netbroadcast.s_addr)
341: goto ours;
342: /*
343: * Look for all-0's host part (old broadcast addr),
344: * either for subnet or net.
345: */
346: t = ntohl(ip->ip_dst.s_addr);
347: if (t == ia->ia_subnet)
348: goto ours;
349: if (t == ia->ia_net)
350: goto ours;
351: }
352: }
353: if (IN_MULTICAST(ntohl(ip->ip_dst.s_addr))) {
354: struct in_multi *inm;
355: #if MROUTING
356: extern struct socket *ip_mrouter;
357:
358: if (m->m_flags & M_EXT) {
359: if ((m = m_pullup(m, hlen)) == 0) {
360: ipstat.ips_toosmall++;
361: goto next;
362: }
363: ip = mtod(m, struct ip *);
364: }
365:
366: if (ip_mrouter) {
367: /*
368: * If we are acting as a multicast router, all
369: * incoming multicast packets are passed to the
370: * kernel-level multicast forwarding function.
371: * The packet is returned (relatively) intact; if
372: * ip_mforward() returns a non-zero value, the packet
373: * must be discarded, else it may be accepted below.
374: *
375: * (The IP ident field is put in the same byte order
376: * as expected when ip_mforward() is called from
377: * ip_output().)
378: */
379: ip->ip_id = htons(ip->ip_id);
380: if (ip_mforward(m, m->m_pkthdr.rcvif) != 0) {
381: ipstat.ips_cantforward++;
382: m_freem(m);
383: goto next;
384: }
385: ip->ip_id = ntohs(ip->ip_id);
386:
387: /*
388: * The process-level routing demon needs to receive
389: * all multicast IGMP packets, whether or not this
390: * host belongs to their destination groups.
391: */
392: if (ip->ip_p == IPPROTO_IGMP)
393: goto ours;
394: ipstat.ips_forward++;
395: }
396: #endif
397: /*
398: * See if we belong to the destination multicast group on the
399: * arrival interface.
400: */
401: IN_LOOKUP_MULTI(ip->ip_dst, m->m_pkthdr.rcvif, inm);
402: if (inm == NULL) {
403: ipstat.ips_cantforward++;
404: m_freem(m);
405: goto next;
406: }
407: goto ours;
408: }
409: if (ip->ip_dst.s_addr == (u_long)INADDR_BROADCAST)
410: goto ours;
411: if (ip->ip_dst.s_addr == INADDR_ANY)
412: goto ours;
413:
414: /*
415: * Not for us; forward if possible and desirable.
416: */
417: if (ipforwarding == 0) {
418: ipstat.ips_cantforward++;
419: m_freem(m);
420: } else
421: ip_forward(m, 0);
422: goto next;
423:
424: ours:
425: /*
426: * If offset or IP_MF are set, must reassemble.
427: * Otherwise, nothing need be done.
428: * (We could look in the reassembly queue to see
429: * if the packet was previously fragmented,
430: * but it's not worth the time; just let them time out.)
431: */
432: if (ip->ip_off &~ IP_DF) {
433: if (m->m_flags & M_EXT) { /* XXX */
434: if ((m = m_pullup(m, sizeof (struct ip))) == 0) {
435: ipstat.ips_toosmall++;
436: goto next;
437: }
438: ip = mtod(m, struct ip *);
439: }
440: /*
441: * Look for queue of fragments
442: * of this datagram.
443: */
444: for (fp = ipq.next; fp != &ipq; fp = fp->next)
445: if (ip->ip_id == fp->ipq_id &&
446: ip->ip_src.s_addr == fp->ipq_src.s_addr &&
447: ip->ip_dst.s_addr == fp->ipq_dst.s_addr &&
448: ip->ip_p == fp->ipq_p)
449: goto found;
450: fp = 0;
451: found:
452:
453: /*
454: * Adjust ip_len to not reflect header,
455: * set ip_mff if more fragments are expected,
456: * convert offset of this to bytes.
457: */
458: ip->ip_len -= hlen;
459: ((struct ipasfrag *)ip)->ipf_mff &= ~1;
460: if (ip->ip_off & IP_MF)
461: ((struct ipasfrag *)ip)->ipf_mff |= 1;
462: ip->ip_off <<= 3;
463:
464: /*
465: * If datagram marked as having more fragments
466: * or if this is not the first fragment,
467: * attempt reassembly; if it succeeds, proceed.
468: */
469: if (((struct ipasfrag *)ip)->ipf_mff & 1 || ip->ip_off) {
470: ipstat.ips_fragments++;
471: ip = ip_reass((struct ipasfrag *)ip, fp);
472: if (ip == 0)
473: goto next;
474: ipstat.ips_reassembled++;
475: m = dtom(ip);
476: } else
477: if (fp)
478: ip_freef(fp);
479: } else
480: ip->ip_len -= hlen;
481:
482: /*
483: * Switch out to protocol's input routine.
484: */
485: ipstat.ips_delivered++;
486:
487: KERNEL_DEBUG(DBG_LAYER_END, ip->ip_dst.s_addr,
488: ip->ip_src.s_addr, ip->ip_p, ip->ip_off, ip->ip_len);
489:
490: (*inetsw[ip_protox[ip->ip_p]].pr_input)(m, hlen);
491: goto next;
492: bad:
493: m_freem(m);
494: goto next;
495: }
496:
497: /*
498: * Take incoming datagram fragment and try to
499: * reassemble it into whole datagram. If a chain for
500: * reassembly of this datagram already exists, then it
501: * is given as fp; otherwise have to make a chain.
502: */
503: struct ip *
504: ip_reass(ip, fp)
505: register struct ipasfrag *ip;
506: register struct ipq *fp;
507: {
508: register struct mbuf *m = dtom(ip);
509: register struct ipasfrag *q;
510: struct mbuf *t;
511: int hlen = ip->ip_hl << 2;
512: int i, next;
513:
514: /*
515: * Presence of header sizes in mbufs
516: * would confuse code below.
517: */
518: m->m_data += hlen;
519: m->m_len -= hlen;
520:
521: /*
522: * If first fragment to arrive, create a reassembly queue.
523: */
524: if (fp == 0) {
525: if ((t = m_get(M_DONTWAIT, MT_FTABLE)) == NULL)
526: goto dropfrag;
527: fp = mtod(t, struct ipq *);
528: insque((queue_t)fp, (queue_t)&ipq);
529: fp->ipq_ttl = IPFRAGTTL;
530: fp->ipq_p = ip->ip_p;
531: fp->ipq_id = ip->ip_id;
532: fp->ipq_next = fp->ipq_prev = (struct ipasfrag *)fp;
533: fp->ipq_src = ((struct ip *)ip)->ip_src;
534: fp->ipq_dst = ((struct ip *)ip)->ip_dst;
535: q = (struct ipasfrag *)fp;
536: goto insert;
537: }
538:
539: /*
540: * Find a segment which begins after this one does.
541: */
542: for (q = fp->ipq_next; q != (struct ipasfrag *)fp; q = q->ipf_next)
543: if (q->ip_off > ip->ip_off)
544: break;
545:
546: /*
547: * If there is a preceding segment, it may provide some of
548: * our data already. If so, drop the data from the incoming
549: * segment. If it provides all of our data, drop us.
550: */
551: if (q->ipf_prev != (struct ipasfrag *)fp) {
552: i = q->ipf_prev->ip_off + q->ipf_prev->ip_len - ip->ip_off;
553: if (i > 0) {
554: if (i >= ip->ip_len)
555: goto dropfrag;
556: m_adj(dtom(ip), i);
557: ip->ip_off += i;
558: ip->ip_len -= i;
559: }
560: }
561:
562: /*
563: * While we overlap succeeding segments trim them or,
564: * if they are completely covered, dequeue them.
565: */
566: while (q != (struct ipasfrag *)fp && ip->ip_off + ip->ip_len > q->ip_off) {
567: i = (ip->ip_off + ip->ip_len) - q->ip_off;
568: if (i < q->ip_len) {
569: q->ip_len -= i;
570: q->ip_off += i;
571: m_adj(dtom(q), i);
572: break;
573: }
574: q = q->ipf_next;
575: m_freem(dtom(q->ipf_prev));
576: ip_deq(q->ipf_prev);
577: }
578:
579: insert:
580: /*
581: * Stick new segment in its place;
582: * check for complete reassembly.
583: */
584: ip_enq(ip, q->ipf_prev);
585: next = 0;
586: for (q = fp->ipq_next; q != (struct ipasfrag *)fp; q = q->ipf_next) {
587: if (q->ip_off != next)
588: return (0);
589: next += q->ip_len;
590: }
591: if (q->ipf_prev->ipf_mff & 1)
592: return (0);
593:
594: /*
595: * Reassembly is complete; concatenate fragments.
596: */
597: q = fp->ipq_next;
598: m = dtom(q);
599: t = m->m_next;
600: m->m_next = 0;
601: m_cat(m, t);
602: q = q->ipf_next;
603: while (q != (struct ipasfrag *)fp) {
604: t = dtom(q);
605: q = q->ipf_next;
606: m_cat(m, t);
607: }
608:
609: /*
610: * Create header for new ip packet by
611: * modifying header of first packet;
612: * dequeue and discard fragment reassembly header.
613: * Make header visible.
614: */
615: ip = fp->ipq_next;
616: ip->ip_len = next;
617: ip->ipf_mff &= ~1;
618: ((struct ip *)ip)->ip_src = fp->ipq_src;
619: ((struct ip *)ip)->ip_dst = fp->ipq_dst;
620: remque((queue_t)fp);
621: (void) m_free(dtom(fp));
622: m = dtom(ip);
623: m->m_len += (ip->ip_hl << 2);
624: m->m_data -= (ip->ip_hl << 2);
625: /* some debugging cruft by sklower, below, will go away soon */
626: if (m->m_flags & M_PKTHDR) { /* XXX this should be done elsewhere */
627: register int plen = 0;
628: for (t = m; m; m = m->m_next)
629: plen += m->m_len;
630: t->m_pkthdr.len = plen;
631: }
632: return ((struct ip *)ip);
633:
634: dropfrag:
635: ipstat.ips_fragdropped++;
636: m_freem(m);
637: return (0);
638: }
639:
640: /*
641: * Free a fragment reassembly header and all
642: * associated datagrams.
643: */
644: void
645: ip_freef(fp)
646: struct ipq *fp;
647: {
648: register struct ipasfrag *q, *p;
649:
650: for (q = fp->ipq_next; q != (struct ipasfrag *)fp; q = p) {
651: p = q->ipf_next;
652: ip_deq(q);
653: m_freem(dtom(q));
654: }
655: remque((queue_t)fp);
656: (void) m_free(dtom(fp));
657: }
658:
659: /*
660: * Put an ip fragment on a reassembly chain.
661: * Like insque, but pointers in middle of structure.
662: */
663: void
664: ip_enq(p, prev)
665: register struct ipasfrag *p, *prev;
666: {
667:
668: p->ipf_prev = prev;
669: p->ipf_next = prev->ipf_next;
670: prev->ipf_next->ipf_prev = p;
671: prev->ipf_next = p;
672: }
673:
674: /*
675: * To ip_enq as remque is to insque.
676: */
677: void
678: ip_deq(p)
679: register struct ipasfrag *p;
680: {
681:
682: p->ipf_prev->ipf_next = p->ipf_next;
683: p->ipf_next->ipf_prev = p->ipf_prev;
684: }
685:
686: /*
687: * IP timer processing;
688: * if a timer expires on a reassembly
689: * queue, discard it.
690: */
691: void
692: ip_slowtimo()
693: {
694: register struct ipq *fp;
695: int s = splnet();
696:
697: fp = ipq.next;
698: if (fp == 0) {
699: splx(s);
700: return;
701: }
702: while (fp != &ipq) {
703: --fp->ipq_ttl;
704: fp = fp->next;
705: if (fp->prev->ipq_ttl == 0) {
706: ipstat.ips_fragtimeout++;
707: ip_freef(fp->prev);
708: }
709: }
710: splx(s);
711: }
712:
713: /*
714: * Drain off all datagram fragments.
715: */
716: void
717: ip_drain()
718: {
719:
720: while (ipq.next != &ipq) {
721: ipstat.ips_fragdropped++;
722: ip_freef(ipq.next);
723: }
724: }
725:
726: /*
727: * Do option processing on a datagram,
728: * possibly discarding it if bad options are encountered,
729: * or forwarding it if source-routed.
730: * Returns 1 if packet has been forwarded/freed,
731: * 0 if the packet should be processed further.
732: */
733: int
734: ip_dooptions(m)
735: struct mbuf *m;
736: {
737: register struct ip *ip = mtod(m, struct ip *);
738: register u_char *cp;
739: register struct ip_timestamp *ipt;
740: register struct in_ifaddr *ia;
741: int opt, optlen, cnt, off, code, type = ICMP_PARAMPROB, forward = 0;
742: struct in_addr *sin, dst;
743: n_time ntime;
744:
745: dst = ip->ip_dst;
746: cp = (u_char *)(ip + 1);
747: cnt = (ip->ip_hl << 2) - sizeof (struct ip);
748: for (; cnt > 0; cnt -= optlen, cp += optlen) {
749: opt = cp[IPOPT_OPTVAL];
750: if (opt == IPOPT_EOL)
751: break;
752: if (opt == IPOPT_NOP)
753: optlen = 1;
754: else {
755: optlen = cp[IPOPT_OLEN];
756: if (optlen <= 0 || optlen > cnt) {
757: code = &cp[IPOPT_OLEN] - (u_char *)ip;
758: goto bad;
759: }
760: }
761: switch (opt) {
762:
763: default:
764: break;
765:
766: /*
767: * Source routing with record.
768: * Find interface with current destination address.
769: * If none on this machine then drop if strictly routed,
770: * or do nothing if loosely routed.
771: * Record interface address and bring up next address
772: * component. If strictly routed make sure next
773: * address is on directly accessible net.
774: */
775: case IPOPT_LSRR:
776: case IPOPT_SSRR:
777: if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) {
778: code = &cp[IPOPT_OFFSET] - (u_char *)ip;
779: goto bad;
780: }
781: ipaddr.sin_addr = ip->ip_dst;
782: ia = (struct in_ifaddr *)
783: ifa_ifwithaddr((struct sockaddr *)&ipaddr);
784: if (ia == 0) {
785: if (opt == IPOPT_SSRR) {
786: type = ICMP_UNREACH;
787: code = ICMP_UNREACH_SRCFAIL;
788: goto bad;
789: }
790: /*
791: * Loose routing, and not at next destination
792: * yet; nothing to do except forward.
793: */
794: break;
795: }
796: off--; /* 0 origin */
797: if (off > optlen - sizeof(struct in_addr)) {
798: /*
799: * End of source route. Should be for us.
800: */
801: save_rte(cp, ip->ip_src);
802: break;
803: }
804: /*
805: * locate outgoing interface
806: */
807: bcopy((caddr_t)(cp + off), (caddr_t)&ipaddr.sin_addr,
808: sizeof(ipaddr.sin_addr));
809: if (opt == IPOPT_SSRR) {
810: #define INA struct in_ifaddr *
811: #define SA struct sockaddr *
812: if ((ia = (INA)ifa_ifwithdstaddr((SA)&ipaddr)) == 0)
813: ia = (INA)ifa_ifwithnet((SA)&ipaddr);
814: } else
815: ia = ip_rtaddr(ipaddr.sin_addr);
816: if (ia == 0) {
817: type = ICMP_UNREACH;
818: code = ICMP_UNREACH_SRCFAIL;
819: goto bad;
820: }
821: ip->ip_dst = ipaddr.sin_addr;
822: bcopy((caddr_t)&(IA_SIN(ia)->sin_addr),
823: (caddr_t)(cp + off), sizeof(struct in_addr));
824: cp[IPOPT_OFFSET] += sizeof(struct in_addr);
825: /*
826: * Let ip_intr's mcast routing check handle mcast pkts
827: */
828: forward = !IN_MULTICAST(ntohl(ip->ip_dst.s_addr));
829: break;
830:
831: case IPOPT_RR:
832: if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) {
833: code = &cp[IPOPT_OFFSET] - (u_char *)ip;
834: goto bad;
835: }
836: /*
837: * If no space remains, ignore.
838: */
839: off--; /* 0 origin */
840: if (off > optlen - sizeof(struct in_addr))
841: break;
842: bcopy((caddr_t)(&ip->ip_dst), (caddr_t)&ipaddr.sin_addr,
843: sizeof(ipaddr.sin_addr));
844: /*
845: * locate outgoing interface; if we're the destination,
846: * use the incoming interface (should be same).
847: */
848: if ((ia = (INA)ifa_ifwithaddr((SA)&ipaddr)) == 0 &&
849: (ia = ip_rtaddr(ipaddr.sin_addr)) == 0) {
850: type = ICMP_UNREACH;
851: code = ICMP_UNREACH_HOST;
852: goto bad;
853: }
854: bcopy((caddr_t)&(IA_SIN(ia)->sin_addr),
855: (caddr_t)(cp + off), sizeof(struct in_addr));
856: cp[IPOPT_OFFSET] += sizeof(struct in_addr);
857: break;
858:
859: case IPOPT_TS:
860: code = cp - (u_char *)ip;
861: ipt = (struct ip_timestamp *)cp;
862: if (ipt->ipt_len < 5)
863: goto bad;
864: if (ipt->ipt_ptr > ipt->ipt_len - sizeof (long)) {
865: if (++ipt->ipt_oflw == 0)
866: goto bad;
867: break;
868: }
869: sin = (struct in_addr *)(cp + ipt->ipt_ptr - 1);
870: switch (ipt->ipt_flg) {
871:
872: case IPOPT_TS_TSONLY:
873: break;
874:
875: case IPOPT_TS_TSANDADDR:
876: if (ipt->ipt_ptr + sizeof(n_time) +
877: sizeof(struct in_addr) > ipt->ipt_len)
878: goto bad;
879: ipaddr.sin_addr = dst;
880: ia = (INA)ifaof_ifpforaddr((SA)&ipaddr,
881: m->m_pkthdr.rcvif);
882: if (ia == 0)
883: continue;
884: bcopy((caddr_t)&IA_SIN(ia)->sin_addr,
885: (caddr_t)sin, sizeof(struct in_addr));
886: ipt->ipt_ptr += sizeof(struct in_addr);
887: break;
888:
889: case IPOPT_TS_PRESPEC:
890: if (ipt->ipt_ptr + sizeof(n_time) +
891: sizeof(struct in_addr) > ipt->ipt_len)
892: goto bad;
893: bcopy((caddr_t)sin, (caddr_t)&ipaddr.sin_addr,
894: sizeof(struct in_addr));
895: if (ifa_ifwithaddr((SA)&ipaddr) == 0)
896: continue;
897: ipt->ipt_ptr += sizeof(struct in_addr);
898: break;
899:
900: default:
901: goto bad;
902: }
903: ntime = iptime();
904: bcopy((caddr_t)&ntime, (caddr_t)cp + ipt->ipt_ptr - 1,
905: sizeof(n_time));
906: ipt->ipt_ptr += sizeof(n_time);
907: }
908: }
909: if (forward) {
910: ip_forward(m, 1);
911: return (1);
912: }
913: return (0);
914: bad:
915: ip->ip_len -= ip->ip_hl << 2; /* XXX icmp_error adds in hdr length */
916: icmp_error(m, type, code, 0, 0);
917: ipstat.ips_badoptions++;
918: return (1);
919: }
920:
921: /*
922: * Given address of next destination (final or next hop),
923: * return internet address info of interface to be used to get there.
924: */
925: struct in_ifaddr *
926: ip_rtaddr(dst)
927: struct in_addr dst;
928: {
929: register struct sockaddr_in *sin;
930:
931: sin = (struct sockaddr_in *) &ipforward_rt.ro_dst;
932:
933: if (ipforward_rt.ro_rt == 0 || dst.s_addr != sin->sin_addr.s_addr) {
934: if (ipforward_rt.ro_rt) {
935: RTFREE(ipforward_rt.ro_rt);
936: ipforward_rt.ro_rt = 0;
937: }
938: sin->sin_family = AF_INET;
939: sin->sin_len = sizeof(*sin);
940: sin->sin_addr = dst;
941:
942: rtalloc(&ipforward_rt);
943: }
944: if (ipforward_rt.ro_rt == 0)
945: return ((struct in_ifaddr *)0);
946: return ((struct in_ifaddr *) ipforward_rt.ro_rt->rt_ifa);
947: }
948:
949: /*
950: * Save incoming source route for use in replies,
951: * to be picked up later by ip_srcroute if the receiver is interested.
952: */
953: void
954: save_rte(option, dst)
955: u_char *option;
956: struct in_addr dst;
957: {
958: unsigned olen;
959:
960: olen = option[IPOPT_OLEN];
961: #if DIAGNOSTIC
962: if (ipprintfs)
963: printf("save_rte: olen %d\n", olen);
964: #endif
965: if (olen > sizeof(ip_srcrt) - (1 + sizeof(dst)))
966: return;
967: bcopy((caddr_t)option, (caddr_t)ip_srcrt.srcopt, olen);
968: ip_nhops = (olen - IPOPT_OFFSET - 1) / sizeof(struct in_addr);
969: ip_srcrt.dst = dst;
970: }
971:
972: /*
973: * Retrieve incoming source route for use in replies,
974: * in the same form used by setsockopt.
975: * The first hop is placed before the options, will be removed later.
976: */
977: struct mbuf *
978: ip_srcroute()
979: {
980: register struct in_addr *p, *q;
981: register struct mbuf *m;
982:
983: if (ip_nhops == 0)
984: return ((struct mbuf *)0);
985: m = m_get(M_DONTWAIT, MT_SOOPTS);
986: if (m == 0)
987: return ((struct mbuf *)0);
988:
989: #define OPTSIZ (sizeof(ip_srcrt.nop) + sizeof(ip_srcrt.srcopt))
990:
991: /* length is (nhops+1)*sizeof(addr) + sizeof(nop + srcrt header) */
992: m->m_len = ip_nhops * sizeof(struct in_addr) + sizeof(struct in_addr) +
993: OPTSIZ;
994: #if DIAGNOSTIC
995: if (ipprintfs)
996: printf("ip_srcroute: nhops %d mlen %d", ip_nhops, m->m_len);
997: #endif
998:
999: /*
1000: * First save first hop for return route
1001: */
1002: p = &ip_srcrt.route[ip_nhops - 1];
1003: *(mtod(m, struct in_addr *)) = *p--;
1004: #if DIAGNOSTIC
1005: if (ipprintfs)
1006: printf(" hops %lx", ntohl(mtod(m, struct in_addr *)->s_addr));
1007: #endif
1008:
1009: /*
1010: * Copy option fields and padding (nop) to mbuf.
1011: */
1012: ip_srcrt.nop = IPOPT_NOP;
1013: ip_srcrt.srcopt[IPOPT_OFFSET] = IPOPT_MINOFF;
1014: bcopy((caddr_t)&ip_srcrt.nop,
1015: mtod(m, caddr_t) + sizeof(struct in_addr), OPTSIZ);
1016: q = (struct in_addr *)(mtod(m, caddr_t) +
1017: sizeof(struct in_addr) + OPTSIZ);
1018: #undef OPTSIZ
1019: /*
1020: * Record return path as an IP source route,
1021: * reversing the path (pointers are now aligned).
1022: */
1023: while (p >= ip_srcrt.route) {
1024: #if DIAGNOSTIC
1025: if (ipprintfs)
1026: printf(" %lx", ntohl(q->s_addr));
1027: #endif
1028: *q++ = *p--;
1029: }
1030: /*
1031: * Last hop goes to final destination.
1032: */
1033: *q = ip_srcrt.dst;
1034: #if DIAGNOSTIC
1035: if (ipprintfs)
1036: printf(" %lx\n", ntohl(q->s_addr));
1037: #endif
1038: return (m);
1039: }
1040:
1041: /*
1042: * Strip out IP options, at higher
1043: * level protocol in the kernel.
1044: * Second argument is buffer to which options
1045: * will be moved, and return value is their length.
1046: * XXX should be deleted; last arg currently ignored.
1047: */
1048: void
1049: ip_stripoptions(m, mopt)
1050: register struct mbuf *m;
1051: struct mbuf *mopt;
1052: {
1053: register int i;
1054: struct ip *ip = mtod(m, struct ip *);
1055: register caddr_t opts;
1056: int olen;
1057:
1058: olen = (ip->ip_hl<<2) - sizeof (struct ip);
1059: opts = (caddr_t)(ip + 1);
1060: i = m->m_len - (sizeof (struct ip) + olen);
1061: bcopy(opts + olen, opts, (unsigned)i);
1062: m->m_len -= olen;
1063: if (m->m_flags & M_PKTHDR)
1064: m->m_pkthdr.len -= olen;
1065: ip->ip_hl = sizeof(struct ip) >> 2;
1066: }
1067:
1068: u_char inetctlerrmap[PRC_NCMDS] = {
1069: 0, 0, 0, 0,
1070: 0, EMSGSIZE, EHOSTDOWN, EHOSTUNREACH,
1071: EHOSTUNREACH, EHOSTUNREACH, ECONNREFUSED, ECONNREFUSED,
1072: EMSGSIZE, EHOSTUNREACH, 0, 0,
1073: 0, 0, 0, 0,
1074: ENOPROTOOPT
1075: };
1076:
1077: /*
1078: * Forward a packet. If some error occurs return the sender
1079: * an icmp packet. Note we can't always generate a meaningful
1080: * icmp message because icmp doesn't have a large enough repertoire
1081: * of codes and types.
1082: *
1083: * If not forwarding, just drop the packet. This could be confusing
1084: * if ipforwarding was zero but some routing protocol was advancing
1085: * us as a gateway to somewhere. However, we must let the routing
1086: * protocol deal with that.
1087: *
1088: * The srcrt parameter indicates whether the packet is being forwarded
1089: * via a source route.
1090: */
1091: void
1092: ip_forward(m, srcrt)
1093: struct mbuf *m;
1094: int srcrt;
1095: {
1096: register struct ip *ip = mtod(m, struct ip *);
1097: register struct sockaddr_in *sin;
1098: register struct rtentry *rt;
1099: int error, type = 0, code;
1100: struct mbuf *mcopy;
1101: n_long dest;
1102: struct ifnet *destifp;
1103:
1104: dest = 0;
1105: #if DIAGNOSTIC
1106: if (ipprintfs)
1107: printf("forward: src %x dst %x ttl %x\n", ip->ip_src,
1108: ip->ip_dst, ip->ip_ttl);
1109: #endif
1110: if (m->m_flags & M_BCAST || in_canforward(ip->ip_dst) == 0) {
1111: ipstat.ips_cantforward++;
1112: m_freem(m);
1113: return;
1114: }
1115: HTONS(ip->ip_id);
1116: if (ip->ip_ttl <= IPTTLDEC) {
1117: icmp_error(m, ICMP_TIMXCEED, ICMP_TIMXCEED_INTRANS, dest, 0);
1118: return;
1119: }
1120: ip->ip_ttl -= IPTTLDEC;
1121:
1122: sin = (struct sockaddr_in *)&ipforward_rt.ro_dst;
1123: if ((rt = ipforward_rt.ro_rt) == 0 ||
1124: ip->ip_dst.s_addr != sin->sin_addr.s_addr) {
1125: if (ipforward_rt.ro_rt) {
1126: RTFREE(ipforward_rt.ro_rt);
1127: ipforward_rt.ro_rt = 0;
1128: }
1129: sin->sin_family = AF_INET;
1130: sin->sin_len = sizeof(*sin);
1131: sin->sin_addr = ip->ip_dst;
1132:
1133: rtalloc(&ipforward_rt);
1134: if (ipforward_rt.ro_rt == 0) {
1135: icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_HOST, dest, 0);
1136: return;
1137: }
1138: rt = ipforward_rt.ro_rt;
1139: }
1140:
1141: /*
1142: * Save at most 64 bytes of the packet in case
1143: * we need to generate an ICMP message to the src.
1144: */
1145: mcopy = m_copy(m, 0, imin((int)ip->ip_len, 64));
1146:
1147: #if GATEWAY
1148: ip_ifmatrix[rt->rt_ifp->if_index +
1149: if_index * m->m_pkthdr.rcvif->if_index]++;
1150: #endif
1151: /*
1152: * If forwarding packet using same interface that it came in on,
1153: * perhaps should send a redirect to sender to shortcut a hop.
1154: * Only send redirect if source is sending directly to us,
1155: * and if packet was not source routed (or has any options).
1156: * Also, don't send redirect if forwarding using a default route
1157: * or a route modified by a redirect.
1158: */
1159: #define satosin(sa) ((struct sockaddr_in *)(sa))
1160: if (rt->rt_ifp == m->m_pkthdr.rcvif &&
1161: (rt->rt_flags & (RTF_DYNAMIC|RTF_MODIFIED)) == 0 &&
1162: satosin(rt_key(rt))->sin_addr.s_addr != 0 &&
1163: ipsendredirects && !srcrt) {
1164: #define RTA(rt) ((struct in_ifaddr *)(rt->rt_ifa))
1165: u_long src = ntohl(ip->ip_src.s_addr);
1166:
1167: if (RTA(rt) &&
1168: (src & RTA(rt)->ia_subnetmask) == RTA(rt)->ia_subnet) {
1169: if (rt->rt_flags & RTF_GATEWAY)
1170: dest = satosin(rt->rt_gateway)->sin_addr.s_addr;
1171: else
1172: dest = ip->ip_dst.s_addr;
1173: /* Router requirements says to only send host redirects */
1174: type = ICMP_REDIRECT;
1175: code = ICMP_REDIRECT_HOST;
1176: #if DIAGNOSTIC
1177: if (ipprintfs)
1178: printf("redirect (%d) to %lx\n", code, (u_long)dest);
1179: #endif
1180: }
1181: }
1182:
1183: error = ip_output(m, (struct mbuf *)0, &ipforward_rt, IP_FORWARDING
1184: #if DIRECTED_BROADCAST
1185: | IP_ALLOWBROADCAST
1186: #endif
1187: , 0);
1188: if (error)
1189: ipstat.ips_cantforward++;
1190: else {
1191: ipstat.ips_forward++;
1192: if (type)
1193: ipstat.ips_redirectsent++;
1194: else {
1195: if (mcopy)
1196: m_freem(mcopy);
1197: return;
1198: }
1199: }
1200: if (mcopy == NULL)
1201: return;
1202: destifp = NULL;
1203:
1204: switch (error) {
1205:
1206: case 0: /* forwarded, but need redirect */
1207: /* type, code set above */
1208: break;
1209:
1210: case ENETUNREACH: /* shouldn't happen, checked above */
1211: case EHOSTUNREACH:
1212: case ENETDOWN:
1213: case EHOSTDOWN:
1214: default:
1215: type = ICMP_UNREACH;
1216: code = ICMP_UNREACH_HOST;
1217: break;
1218:
1219: case EMSGSIZE:
1220: type = ICMP_UNREACH;
1221: code = ICMP_UNREACH_NEEDFRAG;
1222: if (ipforward_rt.ro_rt)
1223: destifp = ipforward_rt.ro_rt->rt_ifp;
1224: ipstat.ips_cantfrag++;
1225: break;
1226:
1227: case ENOBUFS:
1228: type = ICMP_SOURCEQUENCH;
1229: code = 0;
1230: break;
1231: }
1232: icmp_error(mcopy, type, code, dest, destifp);
1233: }
1234:
1235: int
1236: ip_sysctl(name, namelen, oldp, oldlenp, newp, newlen)
1237: int *name;
1238: u_int namelen;
1239: void *oldp;
1240: size_t *oldlenp;
1241: void *newp;
1242: size_t newlen;
1243: {
1244: /* All sysctl names at this level are terminal. */
1245: if (namelen != 1)
1246: return (ENOTDIR);
1247:
1248: switch (name[0]) {
1249: case IPCTL_FORWARDING:
1250: return (sysctl_int(oldp, oldlenp, newp, newlen, &ipforwarding));
1251: case IPCTL_SENDREDIRECTS:
1252: return (sysctl_int(oldp, oldlenp, newp, newlen,
1253: &ipsendredirects));
1254: case IPCTL_DEFTTL:
1255: return (sysctl_int(oldp, oldlenp, newp, newlen, &ip_defttl));
1256: #ifdef notyet
1257: case IPCTL_DEFMTU:
1258: return (sysctl_int(oldp, oldlenp, newp, newlen, &ip_mtu));
1259: #endif
1260: default:
1261: return (EOPNOTSUPP);
1262: }
1263: /* NOTREACHED */
1264: }
This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.