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. Neither the name of the University nor the names of its contributors
14 * may be used to endorse or promote products derived from this software
15 * without specific prior written permission.
17 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29 * @(#)ip_input.c 8.2 (Berkeley) 1/4/94
30 * ip_input.c,v 1.11 1994/11/16 10:17:08 jkh Exp
34 * Changes and additions relating to SLiRP are
35 * Copyright (c) 1995 Danny Gasparovski.
37 * Please read the file COPYRIGHT for the
38 * terms and conditions of the copyright.
45 static struct ip
*ip_reass(Slirp
*slirp
, struct ip
*ip
, struct ipq
*fp
);
46 static void ip_freef(Slirp
*slirp
, struct ipq
*fp
);
47 static void ip_enq(register struct ipasfrag
*p
,
48 register struct ipasfrag
*prev
);
49 static void ip_deq(register struct ipasfrag
*p
);
52 * IP initialization: fill in IP protocol switch table.
53 * All protocols not implemented in kernel go to raw IP protocol handler.
58 slirp
->ipq
.ip_link
.next
= slirp
->ipq
.ip_link
.prev
= &slirp
->ipq
.ip_link
;
65 * Ip input routine. Checksum and byte swap header. If fragmented
66 * try to reassemble. Process options. Pass to next level.
69 ip_input(struct mbuf
*m
)
71 Slirp
*slirp
= m
->slirp
;
72 register struct ip
*ip
;
75 DEBUG_CALL("ip_input");
76 DEBUG_ARG("m = %lx", (long)m
);
77 DEBUG_ARG("m_len = %d", m
->m_len
);
79 if (m
->m_len
< sizeof (struct ip
)) {
83 ip
= mtod(m
, struct ip
*);
85 if (ip
->ip_v
!= IPVERSION
) {
89 hlen
= ip
->ip_hl
<< 2;
90 if (hlen
<sizeof(struct ip
) || hlen
>m
->m_len
) {/* min header length */
91 goto bad
; /* or packet too short */
94 /* keep ip header intact for ICMP reply
95 * ip->ip_sum = cksum(m, hlen);
103 * Convert fields to host representation.
106 if (ip
->ip_len
< hlen
) {
113 * Check that the amount of data in the buffers
114 * is as at least much as the IP header would have us expect.
115 * Trim mbufs if longer than we expect.
116 * Drop packet if shorter than we expect.
118 if (m
->m_len
< ip
->ip_len
) {
122 /* Should drop packet if mbuf too long? hmmm... */
123 if (m
->m_len
> ip
->ip_len
)
124 m_adj(m
, ip
->ip_len
- m
->m_len
);
126 /* check ip_ttl for a correct ICMP reply */
128 icmp_error(m
, ICMP_TIMXCEED
,ICMP_TIMXCEED_INTRANS
, 0,"ttl");
133 * If offset or IP_MF are set, must reassemble.
134 * Otherwise, nothing need be done.
135 * (We could look in the reassembly queue to see
136 * if the packet was previously fragmented,
137 * but it's not worth the time; just let them time out.)
139 * XXX This should fail, don't fragment yet
141 if (ip
->ip_off
&~ IP_DF
) {
142 register struct ipq
*fp
;
145 * Look for queue of fragments
148 for (l
= slirp
->ipq
.ip_link
.next
; l
!= &slirp
->ipq
.ip_link
;
150 fp
= container_of(l
, struct ipq
, ip_link
);
151 if (ip
->ip_id
== fp
->ipq_id
&&
152 ip
->ip_src
.s_addr
== fp
->ipq_src
.s_addr
&&
153 ip
->ip_dst
.s_addr
== fp
->ipq_dst
.s_addr
&&
154 ip
->ip_p
== fp
->ipq_p
)
161 * Adjust ip_len to not reflect header,
162 * set ip_mff if more fragments are expected,
163 * convert offset of this to bytes.
166 if (ip
->ip_off
& IP_MF
)
174 * If datagram marked as having more fragments
175 * or if this is not the first fragment,
176 * attempt reassembly; if it succeeds, proceed.
178 if (ip
->ip_tos
& 1 || ip
->ip_off
) {
179 ip
= ip_reass(slirp
, ip
, fp
);
191 * Switch out to protocol's input routine.
195 tcp_input(m
, hlen
, (struct socket
*)NULL
);
212 #define iptofrag(P) ((struct ipasfrag *)(((char*)(P)) - sizeof(struct qlink)))
213 #define fragtoip(P) ((struct ip*)(((char*)(P)) + sizeof(struct qlink)))
215 * Take incoming datagram fragment and try to
216 * reassemble it into whole datagram. If a chain for
217 * reassembly of this datagram already exists, then it
218 * is given as fp; otherwise have to make a chain.
221 ip_reass(Slirp
*slirp
, struct ip
*ip
, struct ipq
*fp
)
223 register struct mbuf
*m
= dtom(slirp
, ip
);
224 register struct ipasfrag
*q
;
225 int hlen
= ip
->ip_hl
<< 2;
228 DEBUG_CALL("ip_reass");
229 DEBUG_ARG("ip = %lx", (long)ip
);
230 DEBUG_ARG("fp = %lx", (long)fp
);
231 DEBUG_ARG("m = %lx", (long)m
);
234 * Presence of header sizes in mbufs
235 * would confuse code below.
236 * Fragment m_data is concatenated.
242 * If first fragment to arrive, create a reassembly queue.
245 struct mbuf
*t
= m_get(slirp
);
250 fp
= mtod(t
, struct ipq
*);
251 insque(&fp
->ip_link
, &slirp
->ipq
.ip_link
);
252 fp
->ipq_ttl
= IPFRAGTTL
;
253 fp
->ipq_p
= ip
->ip_p
;
254 fp
->ipq_id
= ip
->ip_id
;
255 fp
->frag_link
.next
= fp
->frag_link
.prev
= &fp
->frag_link
;
256 fp
->ipq_src
= ip
->ip_src
;
257 fp
->ipq_dst
= ip
->ip_dst
;
258 q
= (struct ipasfrag
*)fp
;
263 * Find a segment which begins after this one does.
265 for (q
= fp
->frag_link
.next
; q
!= (struct ipasfrag
*)&fp
->frag_link
;
267 if (q
->ipf_off
> ip
->ip_off
)
271 * If there is a preceding segment, it may provide some of
272 * our data already. If so, drop the data from the incoming
273 * segment. If it provides all of our data, drop us.
275 if (q
->ipf_prev
!= &fp
->frag_link
) {
276 struct ipasfrag
*pq
= q
->ipf_prev
;
277 i
= pq
->ipf_off
+ pq
->ipf_len
- ip
->ip_off
;
281 m_adj(dtom(slirp
, ip
), i
);
288 * While we overlap succeeding segments trim them or,
289 * if they are completely covered, dequeue them.
291 while (q
!= (struct ipasfrag
*)&fp
->frag_link
&&
292 ip
->ip_off
+ ip
->ip_len
> q
->ipf_off
) {
293 i
= (ip
->ip_off
+ ip
->ip_len
) - q
->ipf_off
;
294 if (i
< q
->ipf_len
) {
297 m_adj(dtom(slirp
, q
), i
);
301 m_free(dtom(slirp
, q
->ipf_prev
));
307 * Stick new segment in its place;
308 * check for complete reassembly.
310 ip_enq(iptofrag(ip
), q
->ipf_prev
);
312 for (q
= fp
->frag_link
.next
; q
!= (struct ipasfrag
*)&fp
->frag_link
;
314 if (q
->ipf_off
!= next
)
318 if (((struct ipasfrag
*)(q
->ipf_prev
))->ipf_tos
& 1)
322 * Reassembly is complete; concatenate fragments.
324 q
= fp
->frag_link
.next
;
327 q
= (struct ipasfrag
*) q
->ipf_next
;
328 while (q
!= (struct ipasfrag
*)&fp
->frag_link
) {
329 struct mbuf
*t
= dtom(slirp
, q
);
330 q
= (struct ipasfrag
*) q
->ipf_next
;
335 * Create header for new ip packet by
336 * modifying header of first packet;
337 * dequeue and discard fragment reassembly header.
338 * Make header visible.
340 q
= fp
->frag_link
.next
;
343 * If the fragments concatenated to an mbuf that's
344 * bigger than the total size of the fragment, then and
345 * m_ext buffer was alloced. But fp->ipq_next points to
346 * the old buffer (in the mbuf), so we must point ip
347 * into the new buffer.
349 if (m
->m_flags
& M_EXT
) {
350 int delta
= (char *)q
- m
->m_dat
;
351 q
= (struct ipasfrag
*)(m
->m_ext
+ delta
);
357 ip
->ip_src
= fp
->ipq_src
;
358 ip
->ip_dst
= fp
->ipq_dst
;
359 remque(&fp
->ip_link
);
360 (void) m_free(dtom(slirp
, fp
));
361 m
->m_len
+= (ip
->ip_hl
<< 2);
362 m
->m_data
-= (ip
->ip_hl
<< 2);
372 * Free a fragment reassembly header and all
373 * associated datagrams.
376 ip_freef(Slirp
*slirp
, struct ipq
*fp
)
378 register struct ipasfrag
*q
, *p
;
380 for (q
= fp
->frag_link
.next
; q
!= (struct ipasfrag
*)&fp
->frag_link
; q
= p
) {
383 m_free(dtom(slirp
, q
));
385 remque(&fp
->ip_link
);
386 (void) m_free(dtom(slirp
, fp
));
390 * Put an ip fragment on a reassembly chain.
391 * Like insque, but pointers in middle of structure.
394 ip_enq(register struct ipasfrag
*p
, register struct ipasfrag
*prev
)
396 DEBUG_CALL("ip_enq");
397 DEBUG_ARG("prev = %lx", (long)prev
);
399 p
->ipf_next
= prev
->ipf_next
;
400 ((struct ipasfrag
*)(prev
->ipf_next
))->ipf_prev
= p
;
405 * To ip_enq as remque is to insque.
408 ip_deq(register struct ipasfrag
*p
)
410 ((struct ipasfrag
*)(p
->ipf_prev
))->ipf_next
= p
->ipf_next
;
411 ((struct ipasfrag
*)(p
->ipf_next
))->ipf_prev
= p
->ipf_prev
;
415 * IP timer processing;
416 * if a timer expires on a reassembly
420 ip_slowtimo(Slirp
*slirp
)
424 DEBUG_CALL("ip_slowtimo");
426 l
= slirp
->ipq
.ip_link
.next
;
431 while (l
!= &slirp
->ipq
.ip_link
) {
432 struct ipq
*fp
= container_of(l
, struct ipq
, ip_link
);
434 if (--fp
->ipq_ttl
== 0) {
441 * Do option processing on a datagram,
442 * possibly discarding it if bad options are encountered,
443 * or forwarding it if source-routed.
444 * Returns 1 if packet has been forwarded/freed,
445 * 0 if the packet should be processed further.
454 register struct ip
*ip
= mtod(m
, struct ip
*);
456 register struct ip_timestamp
*ipt
;
457 register struct in_ifaddr
*ia
;
458 int opt
, optlen
, cnt
, off
, code
, type
, forward
= 0;
459 struct in_addr
*sin
, dst
;
460 typedef uint32_t n_time
;
464 cp
= (u_char
*)(ip
+ 1);
465 cnt
= (ip
->ip_hl
<< 2) - sizeof (struct ip
);
466 for (; cnt
> 0; cnt
-= optlen
, cp
+= optlen
) {
467 opt
= cp
[IPOPT_OPTVAL
];
468 if (opt
== IPOPT_EOL
)
470 if (opt
== IPOPT_NOP
)
473 optlen
= cp
[IPOPT_OLEN
];
474 if (optlen
<= 0 || optlen
> cnt
) {
475 code
= &cp
[IPOPT_OLEN
] - (u_char
*)ip
;
485 * Source routing with record.
486 * Find interface with current destination address.
487 * If none on this machine then drop if strictly routed,
488 * or do nothing if loosely routed.
489 * Record interface address and bring up next address
490 * component. If strictly routed make sure next
491 * address is on directly accessible net.
495 if ((off
= cp
[IPOPT_OFFSET
]) < IPOPT_MINOFF
) {
496 code
= &cp
[IPOPT_OFFSET
] - (u_char
*)ip
;
499 ipaddr
.sin_addr
= ip
->ip_dst
;
500 ia
= (struct in_ifaddr
*)
501 ifa_ifwithaddr((struct sockaddr
*)&ipaddr
);
503 if (opt
== IPOPT_SSRR
) {
505 code
= ICMP_UNREACH_SRCFAIL
;
509 * Loose routing, and not at next destination
510 * yet; nothing to do except forward.
514 off
--; / * 0 origin
* /
515 if (off
> optlen
- sizeof(struct in_addr
)) {
517 * End of source route. Should be for us.
519 save_rte(cp
, ip
->ip_src
);
523 * locate outgoing interface
525 bcopy((caddr_t
)(cp
+ off
), (caddr_t
)&ipaddr
.sin_addr
,
526 sizeof(ipaddr
.sin_addr
));
527 if (opt
== IPOPT_SSRR
) {
528 #define INA struct in_ifaddr *
529 #define SA struct sockaddr *
530 if ((ia
= (INA
)ifa_ifwithdstaddr((SA
)&ipaddr
)) == 0)
531 ia
= (INA
)ifa_ifwithnet((SA
)&ipaddr
);
533 ia
= ip_rtaddr(ipaddr
.sin_addr
);
536 code
= ICMP_UNREACH_SRCFAIL
;
539 ip
->ip_dst
= ipaddr
.sin_addr
;
540 bcopy((caddr_t
)&(IA_SIN(ia
)->sin_addr
),
541 (caddr_t
)(cp
+ off
), sizeof(struct in_addr
));
542 cp
[IPOPT_OFFSET
] += sizeof(struct in_addr
);
544 * Let ip_intr's mcast routing check handle mcast pkts
546 forward
= !IN_MULTICAST(ntohl(ip
->ip_dst
.s_addr
));
550 if ((off
= cp
[IPOPT_OFFSET
]) < IPOPT_MINOFF
) {
551 code
= &cp
[IPOPT_OFFSET
] - (u_char
*)ip
;
555 * If no space remains, ignore.
558 if (off
> optlen
- sizeof(struct in_addr
))
560 bcopy((caddr_t
)(&ip
->ip_dst
), (caddr_t
)&ipaddr
.sin_addr
,
561 sizeof(ipaddr
.sin_addr
));
563 * locate outgoing interface; if we're the destination,
564 * use the incoming interface (should be same).
566 if ((ia
= (INA
)ifa_ifwithaddr((SA
)&ipaddr
)) == 0 &&
567 (ia
= ip_rtaddr(ipaddr
.sin_addr
)) == 0) {
569 code
= ICMP_UNREACH_HOST
;
572 bcopy((caddr_t
)&(IA_SIN(ia
)->sin_addr
),
573 (caddr_t
)(cp
+ off
), sizeof(struct in_addr
));
574 cp
[IPOPT_OFFSET
] += sizeof(struct in_addr
);
578 code
= cp
- (u_char
*)ip
;
579 ipt
= (struct ip_timestamp
*)cp
;
580 if (ipt
->ipt_len
< 5)
582 if (ipt
->ipt_ptr
> ipt
->ipt_len
- sizeof (int32_t)) {
583 if (++ipt
->ipt_oflw
== 0)
587 sin
= (struct in_addr
*)(cp
+ ipt
->ipt_ptr
- 1);
588 switch (ipt
->ipt_flg
) {
590 case IPOPT_TS_TSONLY
:
593 case IPOPT_TS_TSANDADDR
:
594 if (ipt
->ipt_ptr
+ sizeof(n_time
) +
595 sizeof(struct in_addr
) > ipt
->ipt_len
)
597 ipaddr
.sin_addr
= dst
;
598 ia
= (INA
)ifaof_ i f p
foraddr((SA
)&ipaddr
,
602 bcopy((caddr_t
)&IA_SIN(ia
)->sin_addr
,
603 (caddr_t
)sin
, sizeof(struct in_addr
));
604 ipt
->ipt_ptr
+= sizeof(struct in_addr
);
607 case IPOPT_TS_PRESPEC
:
608 if (ipt
->ipt_ptr
+ sizeof(n_time
) +
609 sizeof(struct in_addr
) > ipt
->ipt_len
)
611 bcopy((caddr_t
)sin
, (caddr_t
)&ipaddr
.sin_addr
,
612 sizeof(struct in_addr
));
613 if (ifa_ifwithaddr((SA
)&ipaddr
) == 0)
615 ipt
->ipt_ptr
+= sizeof(struct in_addr
);
622 bcopy((caddr_t
)&ntime
, (caddr_t
)cp
+ ipt
->ipt_ptr
- 1,
624 ipt
->ipt_ptr
+= sizeof(n_time
);
633 icmp_error(m
, type
, code
, 0, 0);
641 * Strip out IP options, at higher
642 * level protocol in the kernel.
643 * Second argument is buffer to which options
644 * will be moved, and return value is their length.
645 * (XXX) should be deleted; last arg currently ignored.
648 ip_stripoptions(register struct mbuf
*m
, struct mbuf
*mopt
)
651 struct ip
*ip
= mtod(m
, struct ip
*);
652 register caddr_t opts
;
655 olen
= (ip
->ip_hl
<<2) - sizeof (struct ip
);
656 opts
= (caddr_t
)(ip
+ 1);
657 i
= m
->m_len
- (sizeof (struct ip
) + olen
);
658 memcpy(opts
, opts
+ olen
, (unsigned)i
);
661 ip
->ip_hl
= sizeof(struct ip
) >> 2;