2 * Copyright (c) 1982, 1986, 1988, 1993
3 * The Regents of the University of California. All rights reserved.
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
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.
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
33 * @(#)ip_input.c 8.2 (Berkeley) 1/4/94
34 * ip_input.c,v 1.11 1994/11/16 10:17:08 jkh Exp
38 * Changes and additions relating to SLiRP are
39 * Copyright (c) 1995 Danny Gasparovski.
41 * Please read the file COPYRIGHT for the
42 * terms and conditions of the copyright.
53 * IP initialization: fill in IP protocol switch table.
54 * All protocols not implemented in kernel go to raw IP protocol handler.
59 ipq
.next
= ipq
.prev
= (ipqp_32
)&ipq
;
60 ip_id
= tt
.tv_sec
& 0xffff;
67 * Ip input routine. Checksum and byte swap header. If fragmented
68 * try to reassemble. Process options. Pass to next level.
74 register struct ip
*ip
;
77 DEBUG_CALL("ip_input");
78 DEBUG_ARG("m = %lx", (long)m
);
79 DEBUG_ARG("m_len = %d", m
->m_len
);
83 if (m
->m_len
< sizeof (struct ip
)) {
84 ipstat
.ips_toosmall
++;
88 ip
= mtod(m
, struct ip
*);
90 if (ip
->ip_v
!= IPVERSION
) {
95 hlen
= ip
->ip_hl
<< 2;
96 if (hlen
<sizeof(struct ip
) || hlen
>m
->m_len
) {/* min header length */
97 ipstat
.ips_badhlen
++; /* or packet too short */
101 /* keep ip header intact for ICMP reply
102 * ip->ip_sum = cksum(m, hlen);
111 * Convert fields to host representation.
114 if (ip
->ip_len
< hlen
) {
122 * Check that the amount of data in the buffers
123 * is as at least much as the IP header would have us expect.
124 * Trim mbufs if longer than we expect.
125 * Drop packet if shorter than we expect.
127 if (m
->m_len
< ip
->ip_len
) {
128 ipstat
.ips_tooshort
++;
131 /* Should drop packet if mbuf too long? hmmm... */
132 if (m
->m_len
> ip
->ip_len
)
133 m_adj(m
, ip
->ip_len
- m
->m_len
);
135 /* check ip_ttl for a correct ICMP reply */
136 if(ip
->ip_ttl
==0 || ip
->ip_ttl
==1) {
137 icmp_error(m
, ICMP_TIMXCEED
,ICMP_TIMXCEED_INTRANS
, 0,"ttl");
142 * Process options and, if not destined for us,
143 * ship it on. ip_dooptions returns 1 when an
144 * error was detected (causing an icmp message
145 * to be sent and the original packet to be freed).
147 /* We do no IP options */
148 /* if (hlen > sizeof (struct ip) && ip_dooptions(m))
152 * If offset or IP_MF are set, must reassemble.
153 * Otherwise, nothing need be done.
154 * (We could look in the reassembly queue to see
155 * if the packet was previously fragmented,
156 * but it's not worth the time; just let them time out.)
158 * XXX This should fail, don't fragment yet
160 if (ip
->ip_off
&~ IP_DF
) {
161 register struct ipq
*fp
;
163 * Look for queue of fragments
166 for (fp
= (struct ipq
*) ipq
.next
; fp
!= &ipq
;
167 fp
= (struct ipq
*) fp
->next
)
168 if (ip
->ip_id
== fp
->ipq_id
&&
169 ip
->ip_src
.s_addr
== fp
->ipq_src
.s_addr
&&
170 ip
->ip_dst
.s_addr
== fp
->ipq_dst
.s_addr
&&
171 ip
->ip_p
== fp
->ipq_p
)
177 * Adjust ip_len to not reflect header,
178 * set ip_mff if more fragments are expected,
179 * convert offset of this to bytes.
182 if (ip
->ip_off
& IP_MF
)
183 ((struct ipasfrag
*)ip
)->ipf_mff
|= 1;
185 ((struct ipasfrag
*)ip
)->ipf_mff
&= ~1;
190 * If datagram marked as having more fragments
191 * or if this is not the first fragment,
192 * attempt reassembly; if it succeeds, proceed.
194 if (((struct ipasfrag
*)ip
)->ipf_mff
& 1 || ip
->ip_off
) {
195 ipstat
.ips_fragments
++;
196 ip
= ip_reass((struct ipasfrag
*)ip
, fp
);
199 ipstat
.ips_reassembled
++;
209 * Switch out to protocol's input routine.
211 ipstat
.ips_delivered
++;
214 tcp_input(m
, hlen
, (struct socket
*)NULL
);
223 ipstat
.ips_noproto
++;
233 * Take incoming datagram fragment and try to
234 * reassemble it into whole datagram. If a chain for
235 * reassembly of this datagram already exists, then it
236 * is given as fp; otherwise have to make a chain.
240 register struct ipasfrag
*ip
;
241 register struct ipq
*fp
;
243 register struct mbuf
*m
= dtom(ip
);
244 register struct ipasfrag
*q
;
245 int hlen
= ip
->ip_hl
<< 2;
248 DEBUG_CALL("ip_reass");
249 DEBUG_ARG("ip = %lx", (long)ip
);
250 DEBUG_ARG("fp = %lx", (long)fp
);
251 DEBUG_ARG("m = %lx", (long)m
);
254 * Presence of header sizes in mbufs
255 * would confuse code below.
256 * Fragment m_data is concatenated.
262 * If first fragment to arrive, create a reassembly queue.
266 if ((t
= m_get()) == NULL
) goto dropfrag
;
267 fp
= mtod(t
, struct ipq
*);
269 fp
->ipq_ttl
= IPFRAGTTL
;
270 fp
->ipq_p
= ip
->ip_p
;
271 fp
->ipq_id
= ip
->ip_id
;
272 fp
->ipq_next
= fp
->ipq_prev
= (ipasfragp_32
)fp
;
273 fp
->ipq_src
= ((struct ip
*)ip
)->ip_src
;
274 fp
->ipq_dst
= ((struct ip
*)ip
)->ip_dst
;
275 q
= (struct ipasfrag
*)fp
;
280 * Find a segment which begins after this one does.
282 for (q
= (struct ipasfrag
*)fp
->ipq_next
; q
!= (struct ipasfrag
*)fp
;
283 q
= (struct ipasfrag
*)q
->ipf_next
)
284 if (q
->ip_off
> ip
->ip_off
)
288 * If there is a preceding segment, it may provide some of
289 * our data already. If so, drop the data from the incoming
290 * segment. If it provides all of our data, drop us.
292 if (q
->ipf_prev
!= (ipasfragp_32
)fp
) {
293 i
= ((struct ipasfrag
*)(q
->ipf_prev
))->ip_off
+
294 ((struct ipasfrag
*)(q
->ipf_prev
))->ip_len
- ip
->ip_off
;
305 * While we overlap succeeding segments trim them or,
306 * if they are completely covered, dequeue them.
308 while (q
!= (struct ipasfrag
*)fp
&& ip
->ip_off
+ ip
->ip_len
> q
->ip_off
) {
309 i
= (ip
->ip_off
+ ip
->ip_len
) - q
->ip_off
;
316 q
= (struct ipasfrag
*) q
->ipf_next
;
317 m_freem(dtom((struct ipasfrag
*) q
->ipf_prev
));
318 ip_deq((struct ipasfrag
*) q
->ipf_prev
);
323 * Stick new segment in its place;
324 * check for complete reassembly.
326 ip_enq(ip
, (struct ipasfrag
*) q
->ipf_prev
);
328 for (q
= (struct ipasfrag
*) fp
->ipq_next
; q
!= (struct ipasfrag
*)fp
;
329 q
= (struct ipasfrag
*) q
->ipf_next
) {
330 if (q
->ip_off
!= next
)
334 if (((struct ipasfrag
*)(q
->ipf_prev
))->ipf_mff
& 1)
338 * Reassembly is complete; concatenate fragments.
340 q
= (struct ipasfrag
*) fp
->ipq_next
;
343 q
= (struct ipasfrag
*) q
->ipf_next
;
344 while (q
!= (struct ipasfrag
*)fp
) {
348 q
= (struct ipasfrag
*) q
->ipf_next
;
352 * Create header for new ip packet by
353 * modifying header of first packet;
354 * dequeue and discard fragment reassembly header.
355 * Make header visible.
357 ip
= (struct ipasfrag
*) fp
->ipq_next
;
360 * If the fragments concatenated to an mbuf that's
361 * bigger than the total size of the fragment, then and
362 * m_ext buffer was alloced. But fp->ipq_next points to
363 * the old buffer (in the mbuf), so we must point ip
364 * into the new buffer.
366 if (m
->m_flags
& M_EXT
) {
368 delta
= (char *)ip
- m
->m_dat
;
369 ip
= (struct ipasfrag
*)(m
->m_ext
+ delta
);
372 /* DEBUG_ARG("ip = %lx", (long)ip);
373 * ip=(struct ipasfrag *)m->m_data; */
377 ((struct ip
*)ip
)->ip_src
= fp
->ipq_src
;
378 ((struct ip
*)ip
)->ip_dst
= fp
->ipq_dst
;
380 (void) m_free(dtom(fp
));
382 m
->m_len
+= (ip
->ip_hl
<< 2);
383 m
->m_data
-= (ip
->ip_hl
<< 2);
385 return ((struct ip
*)ip
);
388 ipstat
.ips_fragdropped
++;
394 * Free a fragment reassembly header and all
395 * associated datagrams.
401 register struct ipasfrag
*q
, *p
;
403 for (q
= (struct ipasfrag
*) fp
->ipq_next
; q
!= (struct ipasfrag
*)fp
;
405 p
= (struct ipasfrag
*) q
->ipf_next
;
410 (void) m_free(dtom(fp
));
414 * Put an ip fragment on a reassembly chain.
415 * Like insque, but pointers in middle of structure.
419 register struct ipasfrag
*p
, *prev
;
421 DEBUG_CALL("ip_enq");
422 DEBUG_ARG("prev = %lx", (long)prev
);
423 p
->ipf_prev
= (ipasfragp_32
) prev
;
424 p
->ipf_next
= prev
->ipf_next
;
425 ((struct ipasfrag
*)(prev
->ipf_next
))->ipf_prev
= (ipasfragp_32
) p
;
426 prev
->ipf_next
= (ipasfragp_32
) p
;
430 * To ip_enq as remque is to insque.
434 register struct ipasfrag
*p
;
436 ((struct ipasfrag
*)(p
->ipf_prev
))->ipf_next
= p
->ipf_next
;
437 ((struct ipasfrag
*)(p
->ipf_next
))->ipf_prev
= p
->ipf_prev
;
441 * IP timer processing;
442 * if a timer expires on a reassembly
448 register struct ipq
*fp
;
450 DEBUG_CALL("ip_slowtimo");
452 fp
= (struct ipq
*) ipq
.next
;
458 fp
= (struct ipq
*) fp
->next
;
459 if (((struct ipq
*)(fp
->prev
))->ipq_ttl
== 0) {
460 ipstat
.ips_fragtimeout
++;
461 ip_freef((struct ipq
*) fp
->prev
);
467 * Do option processing on a datagram,
468 * possibly discarding it if bad options are encountered,
469 * or forwarding it if source-routed.
470 * Returns 1 if packet has been forwarded/freed,
471 * 0 if the packet should be processed further.
480 register struct ip
*ip
= mtod(m
, struct ip
*);
482 register struct ip_timestamp
*ipt
;
483 register struct in_ifaddr
*ia
;
484 /* int opt, optlen, cnt, off, code, type = ICMP_PARAMPROB, forward = 0; */
485 int opt
, optlen
, cnt
, off
, code
, type
, forward
= 0;
486 struct in_addr
*sin
, dst
;
487 typedef u_int32_t n_time
;
491 cp
= (u_char
*)(ip
+ 1);
492 cnt
= (ip
->ip_hl
<< 2) - sizeof (struct ip
);
493 for (; cnt
> 0; cnt
-= optlen
, cp
+= optlen
) {
494 opt
= cp
[IPOPT_OPTVAL
];
495 if (opt
== IPOPT_EOL
)
497 if (opt
== IPOPT_NOP
)
500 optlen
= cp
[IPOPT_OLEN
];
501 if (optlen
<= 0 || optlen
> cnt
) {
502 code
= &cp
[IPOPT_OLEN
] - (u_char
*)ip
;
512 * Source routing with record.
513 * Find interface with current destination address.
514 * If none on this machine then drop if strictly routed,
515 * or do nothing if loosely routed.
516 * Record interface address and bring up next address
517 * component. If strictly routed make sure next
518 * address is on directly accessible net.
522 if ((off
= cp
[IPOPT_OFFSET
]) < IPOPT_MINOFF
) {
523 code
= &cp
[IPOPT_OFFSET
] - (u_char
*)ip
;
526 ipaddr
.sin_addr
= ip
->ip_dst
;
527 ia
= (struct in_ifaddr
*)
528 ifa_ifwithaddr((struct sockaddr
*)&ipaddr
);
530 if (opt
== IPOPT_SSRR
) {
532 code
= ICMP_UNREACH_SRCFAIL
;
536 * Loose routing, and not at next destination
537 * yet; nothing to do except forward.
541 off
--; / * 0 origin
* /
542 if (off
> optlen
- sizeof(struct in_addr
)) {
544 * End of source route. Should be for us.
546 save_rte(cp
, ip
->ip_src
);
550 * locate outgoing interface
552 bcopy((caddr_t
)(cp
+ off
), (caddr_t
)&ipaddr
.sin_addr
,
553 sizeof(ipaddr
.sin_addr
));
554 if (opt
== IPOPT_SSRR
) {
555 #define INA struct in_ifaddr *
556 #define SA struct sockaddr *
557 if ((ia
= (INA
)ifa_ifwithdstaddr((SA
)&ipaddr
)) == 0)
558 ia
= (INA
)ifa_ifwithnet((SA
)&ipaddr
);
560 ia
= ip_rtaddr(ipaddr
.sin_addr
);
563 code
= ICMP_UNREACH_SRCFAIL
;
566 ip
->ip_dst
= ipaddr
.sin_addr
;
567 bcopy((caddr_t
)&(IA_SIN(ia
)->sin_addr
),
568 (caddr_t
)(cp
+ off
), sizeof(struct in_addr
));
569 cp
[IPOPT_OFFSET
] += sizeof(struct in_addr
);
571 * Let ip_intr's mcast routing check handle mcast pkts
573 forward
= !IN_MULTICAST(ntohl(ip
->ip_dst
.s_addr
));
577 if ((off
= cp
[IPOPT_OFFSET
]) < IPOPT_MINOFF
) {
578 code
= &cp
[IPOPT_OFFSET
] - (u_char
*)ip
;
582 * If no space remains, ignore.
585 if (off
> optlen
- sizeof(struct in_addr
))
587 bcopy((caddr_t
)(&ip
->ip_dst
), (caddr_t
)&ipaddr
.sin_addr
,
588 sizeof(ipaddr
.sin_addr
));
590 * locate outgoing interface; if we're the destination,
591 * use the incoming interface (should be same).
593 if ((ia
= (INA
)ifa_ifwithaddr((SA
)&ipaddr
)) == 0 &&
594 (ia
= ip_rtaddr(ipaddr
.sin_addr
)) == 0) {
596 code
= ICMP_UNREACH_HOST
;
599 bcopy((caddr_t
)&(IA_SIN(ia
)->sin_addr
),
600 (caddr_t
)(cp
+ off
), sizeof(struct in_addr
));
601 cp
[IPOPT_OFFSET
] += sizeof(struct in_addr
);
605 code
= cp
- (u_char
*)ip
;
606 ipt
= (struct ip_timestamp
*)cp
;
607 if (ipt
->ipt_len
< 5)
609 if (ipt
->ipt_ptr
> ipt
->ipt_len
- sizeof (int32_t)) {
610 if (++ipt
->ipt_oflw
== 0)
614 sin
= (struct in_addr
*)(cp
+ ipt
->ipt_ptr
- 1);
615 switch (ipt
->ipt_flg
) {
617 case IPOPT_TS_TSONLY
:
620 case IPOPT_TS_TSANDADDR
:
621 if (ipt
->ipt_ptr
+ sizeof(n_time
) +
622 sizeof(struct in_addr
) > ipt
->ipt_len
)
624 ipaddr
.sin_addr
= dst
;
625 ia
= (INA
)ifaof_ i f p
foraddr((SA
)&ipaddr
,
629 bcopy((caddr_t
)&IA_SIN(ia
)->sin_addr
,
630 (caddr_t
)sin
, sizeof(struct in_addr
));
631 ipt
->ipt_ptr
+= sizeof(struct in_addr
);
634 case IPOPT_TS_PRESPEC
:
635 if (ipt
->ipt_ptr
+ sizeof(n_time
) +
636 sizeof(struct in_addr
) > ipt
->ipt_len
)
638 bcopy((caddr_t
)sin
, (caddr_t
)&ipaddr
.sin_addr
,
639 sizeof(struct in_addr
));
640 if (ifa_ifwithaddr((SA
)&ipaddr
) == 0)
642 ipt
->ipt_ptr
+= sizeof(struct in_addr
);
649 bcopy((caddr_t
)&ntime
, (caddr_t
)cp
+ ipt
->ipt_ptr
- 1,
651 ipt
->ipt_ptr
+= sizeof(n_time
);
662 /* ip->ip_len -= ip->ip_hl << 2; XXX icmp_error adds in hdr length */
665 icmp_error(m
, type
, code
, 0, 0);
667 ipstat
.ips_badoptions
++;
674 * Strip out IP options, at higher
675 * level protocol in the kernel.
676 * Second argument is buffer to which options
677 * will be moved, and return value is their length.
678 * (XXX) should be deleted; last arg currently ignored.
681 ip_stripoptions(m
, mopt
)
682 register struct mbuf
*m
;
686 struct ip
*ip
= mtod(m
, struct ip
*);
687 register caddr_t opts
;
690 olen
= (ip
->ip_hl
<<2) - sizeof (struct ip
);
691 opts
= (caddr_t
)(ip
+ 1);
692 i
= m
->m_len
- (sizeof (struct ip
) + olen
);
693 memcpy(opts
, opts
+ olen
, (unsigned)i
);
696 ip
->ip_hl
= sizeof(struct ip
) >> 2;