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
;
64 * Ip input routine. Checksum and byte swap header. If fragmented
65 * try to reassemble. Process options. Pass to next level.
68 ip_input(struct mbuf
*m
)
70 Slirp
*slirp
= m
->slirp
;
71 register struct ip
*ip
;
74 DEBUG_CALL("ip_input");
75 DEBUG_ARG("m = %lx", (long)m
);
76 DEBUG_ARG("m_len = %d", m
->m_len
);
78 if (m
->m_len
< sizeof (struct ip
)) {
82 ip
= mtod(m
, struct ip
*);
84 if (ip
->ip_v
!= IPVERSION
) {
88 hlen
= ip
->ip_hl
<< 2;
89 if (hlen
<sizeof(struct ip
) || hlen
>m
->m_len
) {/* min header length */
90 goto bad
; /* or packet too short */
93 /* keep ip header intact for ICMP reply
94 * ip->ip_sum = cksum(m, hlen);
102 * Convert fields to host representation.
105 if (ip
->ip_len
< hlen
) {
112 * Check that the amount of data in the buffers
113 * is as at least much as the IP header would have us expect.
114 * Trim mbufs if longer than we expect.
115 * Drop packet if shorter than we expect.
117 if (m
->m_len
< ip
->ip_len
) {
121 if (slirp
->restricted
) {
122 if ((ip
->ip_dst
.s_addr
& slirp
->vnetwork_mask
.s_addr
) ==
123 slirp
->vnetwork_addr
.s_addr
) {
124 if (ip
->ip_dst
.s_addr
== 0xffffffff && ip
->ip_p
!= IPPROTO_UDP
)
127 uint32_t inv_mask
= ~slirp
->vnetwork_mask
.s_addr
;
128 struct ex_list
*ex_ptr
;
130 if ((ip
->ip_dst
.s_addr
& inv_mask
) == inv_mask
) {
133 for (ex_ptr
= slirp
->exec_list
; ex_ptr
; ex_ptr
= ex_ptr
->ex_next
)
134 if (ex_ptr
->ex_addr
.s_addr
== ip
->ip_dst
.s_addr
)
142 /* Should drop packet if mbuf too long? hmmm... */
143 if (m
->m_len
> ip
->ip_len
)
144 m_adj(m
, ip
->ip_len
- m
->m_len
);
146 /* check ip_ttl for a correct ICMP reply */
147 if(ip
->ip_ttl
==0 || ip
->ip_ttl
==1) {
148 icmp_error(m
, ICMP_TIMXCEED
,ICMP_TIMXCEED_INTRANS
, 0,"ttl");
153 * If offset or IP_MF are set, must reassemble.
154 * Otherwise, nothing need be done.
155 * (We could look in the reassembly queue to see
156 * if the packet was previously fragmented,
157 * but it's not worth the time; just let them time out.)
159 * XXX This should fail, don't fragment yet
161 if (ip
->ip_off
&~ IP_DF
) {
162 register struct ipq
*fp
;
165 * Look for queue of fragments
168 for (l
= slirp
->ipq
.ip_link
.next
; l
!= &slirp
->ipq
.ip_link
;
170 fp
= container_of(l
, struct ipq
, ip_link
);
171 if (ip
->ip_id
== fp
->ipq_id
&&
172 ip
->ip_src
.s_addr
== fp
->ipq_src
.s_addr
&&
173 ip
->ip_dst
.s_addr
== fp
->ipq_dst
.s_addr
&&
174 ip
->ip_p
== fp
->ipq_p
)
181 * Adjust ip_len to not reflect header,
182 * set ip_mff if more fragments are expected,
183 * convert offset of this to bytes.
186 if (ip
->ip_off
& IP_MF
)
194 * If datagram marked as having more fragments
195 * or if this is not the first fragment,
196 * attempt reassembly; if it succeeds, proceed.
198 if (ip
->ip_tos
& 1 || ip
->ip_off
) {
199 ip
= ip_reass(slirp
, ip
, fp
);
211 * Switch out to protocol's input routine.
215 tcp_input(m
, hlen
, (struct socket
*)NULL
);
232 #define iptofrag(P) ((struct ipasfrag *)(((char*)(P)) - sizeof(struct qlink)))
233 #define fragtoip(P) ((struct ip*)(((char*)(P)) + sizeof(struct qlink)))
235 * Take incoming datagram fragment and try to
236 * reassemble it into whole datagram. If a chain for
237 * reassembly of this datagram already exists, then it
238 * is given as fp; otherwise have to make a chain.
241 ip_reass(Slirp
*slirp
, struct ip
*ip
, struct ipq
*fp
)
243 register struct mbuf
*m
= dtom(slirp
, 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.
265 struct mbuf
*t
= m_get(slirp
);
270 fp
= mtod(t
, struct ipq
*);
271 insque(&fp
->ip_link
, &slirp
->ipq
.ip_link
);
272 fp
->ipq_ttl
= IPFRAGTTL
;
273 fp
->ipq_p
= ip
->ip_p
;
274 fp
->ipq_id
= ip
->ip_id
;
275 fp
->frag_link
.next
= fp
->frag_link
.prev
= &fp
->frag_link
;
276 fp
->ipq_src
= ip
->ip_src
;
277 fp
->ipq_dst
= ip
->ip_dst
;
278 q
= (struct ipasfrag
*)fp
;
283 * Find a segment which begins after this one does.
285 for (q
= fp
->frag_link
.next
; q
!= (struct ipasfrag
*)&fp
->frag_link
;
287 if (q
->ipf_off
> ip
->ip_off
)
291 * If there is a preceding segment, it may provide some of
292 * our data already. If so, drop the data from the incoming
293 * segment. If it provides all of our data, drop us.
295 if (q
->ipf_prev
!= &fp
->frag_link
) {
296 struct ipasfrag
*pq
= q
->ipf_prev
;
297 i
= pq
->ipf_off
+ pq
->ipf_len
- ip
->ip_off
;
301 m_adj(dtom(slirp
, ip
), i
);
308 * While we overlap succeeding segments trim them or,
309 * if they are completely covered, dequeue them.
311 while (q
!= (struct ipasfrag
*)&fp
->frag_link
&&
312 ip
->ip_off
+ ip
->ip_len
> q
->ipf_off
) {
313 i
= (ip
->ip_off
+ ip
->ip_len
) - q
->ipf_off
;
314 if (i
< q
->ipf_len
) {
317 m_adj(dtom(slirp
, q
), i
);
321 m_freem(dtom(slirp
, q
->ipf_prev
));
327 * Stick new segment in its place;
328 * check for complete reassembly.
330 ip_enq(iptofrag(ip
), q
->ipf_prev
);
332 for (q
= fp
->frag_link
.next
; q
!= (struct ipasfrag
*)&fp
->frag_link
;
334 if (q
->ipf_off
!= next
)
338 if (((struct ipasfrag
*)(q
->ipf_prev
))->ipf_tos
& 1)
342 * Reassembly is complete; concatenate fragments.
344 q
= fp
->frag_link
.next
;
347 q
= (struct ipasfrag
*) q
->ipf_next
;
348 while (q
!= (struct ipasfrag
*)&fp
->frag_link
) {
349 struct mbuf
*t
= dtom(slirp
, q
);
350 q
= (struct ipasfrag
*) q
->ipf_next
;
355 * Create header for new ip packet by
356 * modifying header of first packet;
357 * dequeue and discard fragment reassembly header.
358 * Make header visible.
360 q
= fp
->frag_link
.next
;
363 * If the fragments concatenated to an mbuf that's
364 * bigger than the total size of the fragment, then and
365 * m_ext buffer was alloced. But fp->ipq_next points to
366 * the old buffer (in the mbuf), so we must point ip
367 * into the new buffer.
369 if (m
->m_flags
& M_EXT
) {
370 int delta
= (char *)q
- m
->m_dat
;
371 q
= (struct ipasfrag
*)(m
->m_ext
+ delta
);
377 ip
->ip_src
= fp
->ipq_src
;
378 ip
->ip_dst
= fp
->ipq_dst
;
379 remque(&fp
->ip_link
);
380 (void) m_free(dtom(slirp
, fp
));
381 m
->m_len
+= (ip
->ip_hl
<< 2);
382 m
->m_data
-= (ip
->ip_hl
<< 2);
392 * Free a fragment reassembly header and all
393 * associated datagrams.
396 ip_freef(Slirp
*slirp
, struct ipq
*fp
)
398 register struct ipasfrag
*q
, *p
;
400 for (q
= fp
->frag_link
.next
; q
!= (struct ipasfrag
*)&fp
->frag_link
; q
= p
) {
403 m_freem(dtom(slirp
, q
));
405 remque(&fp
->ip_link
);
406 (void) m_free(dtom(slirp
, fp
));
410 * Put an ip fragment on a reassembly chain.
411 * Like insque, but pointers in middle of structure.
414 ip_enq(register struct ipasfrag
*p
, register struct ipasfrag
*prev
)
416 DEBUG_CALL("ip_enq");
417 DEBUG_ARG("prev = %lx", (long)prev
);
419 p
->ipf_next
= prev
->ipf_next
;
420 ((struct ipasfrag
*)(prev
->ipf_next
))->ipf_prev
= p
;
425 * To ip_enq as remque is to insque.
428 ip_deq(register struct ipasfrag
*p
)
430 ((struct ipasfrag
*)(p
->ipf_prev
))->ipf_next
= p
->ipf_next
;
431 ((struct ipasfrag
*)(p
->ipf_next
))->ipf_prev
= p
->ipf_prev
;
435 * IP timer processing;
436 * if a timer expires on a reassembly
440 ip_slowtimo(Slirp
*slirp
)
444 DEBUG_CALL("ip_slowtimo");
446 l
= slirp
->ipq
.ip_link
.next
;
451 while (l
!= &slirp
->ipq
.ip_link
) {
452 struct ipq
*fp
= container_of(l
, struct ipq
, ip_link
);
454 if (--fp
->ipq_ttl
== 0) {
461 * Do option processing on a datagram,
462 * possibly discarding it if bad options are encountered,
463 * or forwarding it if source-routed.
464 * Returns 1 if packet has been forwarded/freed,
465 * 0 if the packet should be processed further.
474 register struct ip
*ip
= mtod(m
, struct ip
*);
476 register struct ip_timestamp
*ipt
;
477 register struct in_ifaddr
*ia
;
478 int opt
, optlen
, cnt
, off
, code
, type
, forward
= 0;
479 struct in_addr
*sin
, dst
;
480 typedef uint32_t n_time
;
484 cp
= (u_char
*)(ip
+ 1);
485 cnt
= (ip
->ip_hl
<< 2) - sizeof (struct ip
);
486 for (; cnt
> 0; cnt
-= optlen
, cp
+= optlen
) {
487 opt
= cp
[IPOPT_OPTVAL
];
488 if (opt
== IPOPT_EOL
)
490 if (opt
== IPOPT_NOP
)
493 optlen
= cp
[IPOPT_OLEN
];
494 if (optlen
<= 0 || optlen
> cnt
) {
495 code
= &cp
[IPOPT_OLEN
] - (u_char
*)ip
;
505 * Source routing with record.
506 * Find interface with current destination address.
507 * If none on this machine then drop if strictly routed,
508 * or do nothing if loosely routed.
509 * Record interface address and bring up next address
510 * component. If strictly routed make sure next
511 * address is on directly accessible net.
515 if ((off
= cp
[IPOPT_OFFSET
]) < IPOPT_MINOFF
) {
516 code
= &cp
[IPOPT_OFFSET
] - (u_char
*)ip
;
519 ipaddr
.sin_addr
= ip
->ip_dst
;
520 ia
= (struct in_ifaddr
*)
521 ifa_ifwithaddr((struct sockaddr
*)&ipaddr
);
523 if (opt
== IPOPT_SSRR
) {
525 code
= ICMP_UNREACH_SRCFAIL
;
529 * Loose routing, and not at next destination
530 * yet; nothing to do except forward.
534 off
--; / * 0 origin
* /
535 if (off
> optlen
- sizeof(struct in_addr
)) {
537 * End of source route. Should be for us.
539 save_rte(cp
, ip
->ip_src
);
543 * locate outgoing interface
545 bcopy((caddr_t
)(cp
+ off
), (caddr_t
)&ipaddr
.sin_addr
,
546 sizeof(ipaddr
.sin_addr
));
547 if (opt
== IPOPT_SSRR
) {
548 #define INA struct in_ifaddr *
549 #define SA struct sockaddr *
550 if ((ia
= (INA
)ifa_ifwithdstaddr((SA
)&ipaddr
)) == 0)
551 ia
= (INA
)ifa_ifwithnet((SA
)&ipaddr
);
553 ia
= ip_rtaddr(ipaddr
.sin_addr
);
556 code
= ICMP_UNREACH_SRCFAIL
;
559 ip
->ip_dst
= ipaddr
.sin_addr
;
560 bcopy((caddr_t
)&(IA_SIN(ia
)->sin_addr
),
561 (caddr_t
)(cp
+ off
), sizeof(struct in_addr
));
562 cp
[IPOPT_OFFSET
] += sizeof(struct in_addr
);
564 * Let ip_intr's mcast routing check handle mcast pkts
566 forward
= !IN_MULTICAST(ntohl(ip
->ip_dst
.s_addr
));
570 if ((off
= cp
[IPOPT_OFFSET
]) < IPOPT_MINOFF
) {
571 code
= &cp
[IPOPT_OFFSET
] - (u_char
*)ip
;
575 * If no space remains, ignore.
578 if (off
> optlen
- sizeof(struct in_addr
))
580 bcopy((caddr_t
)(&ip
->ip_dst
), (caddr_t
)&ipaddr
.sin_addr
,
581 sizeof(ipaddr
.sin_addr
));
583 * locate outgoing interface; if we're the destination,
584 * use the incoming interface (should be same).
586 if ((ia
= (INA
)ifa_ifwithaddr((SA
)&ipaddr
)) == 0 &&
587 (ia
= ip_rtaddr(ipaddr
.sin_addr
)) == 0) {
589 code
= ICMP_UNREACH_HOST
;
592 bcopy((caddr_t
)&(IA_SIN(ia
)->sin_addr
),
593 (caddr_t
)(cp
+ off
), sizeof(struct in_addr
));
594 cp
[IPOPT_OFFSET
] += sizeof(struct in_addr
);
598 code
= cp
- (u_char
*)ip
;
599 ipt
= (struct ip_timestamp
*)cp
;
600 if (ipt
->ipt_len
< 5)
602 if (ipt
->ipt_ptr
> ipt
->ipt_len
- sizeof (int32_t)) {
603 if (++ipt
->ipt_oflw
== 0)
607 sin
= (struct in_addr
*)(cp
+ ipt
->ipt_ptr
- 1);
608 switch (ipt
->ipt_flg
) {
610 case IPOPT_TS_TSONLY
:
613 case IPOPT_TS_TSANDADDR
:
614 if (ipt
->ipt_ptr
+ sizeof(n_time
) +
615 sizeof(struct in_addr
) > ipt
->ipt_len
)
617 ipaddr
.sin_addr
= dst
;
618 ia
= (INA
)ifaof_ i f p
foraddr((SA
)&ipaddr
,
622 bcopy((caddr_t
)&IA_SIN(ia
)->sin_addr
,
623 (caddr_t
)sin
, sizeof(struct in_addr
));
624 ipt
->ipt_ptr
+= sizeof(struct in_addr
);
627 case IPOPT_TS_PRESPEC
:
628 if (ipt
->ipt_ptr
+ sizeof(n_time
) +
629 sizeof(struct in_addr
) > ipt
->ipt_len
)
631 bcopy((caddr_t
)sin
, (caddr_t
)&ipaddr
.sin_addr
,
632 sizeof(struct in_addr
));
633 if (ifa_ifwithaddr((SA
)&ipaddr
) == 0)
635 ipt
->ipt_ptr
+= sizeof(struct in_addr
);
642 bcopy((caddr_t
)&ntime
, (caddr_t
)cp
+ ipt
->ipt_ptr
- 1,
644 ipt
->ipt_ptr
+= sizeof(n_time
);
653 icmp_error(m
, type
, code
, 0, 0);
661 * Strip out IP options, at higher
662 * level protocol in the kernel.
663 * Second argument is buffer to which options
664 * will be moved, and return value is their length.
665 * (XXX) should be deleted; last arg currently ignored.
668 ip_stripoptions(register struct mbuf
*m
, struct mbuf
*mopt
)
671 struct ip
*ip
= mtod(m
, struct ip
*);
672 register caddr_t opts
;
675 olen
= (ip
->ip_hl
<<2) - sizeof (struct ip
);
676 opts
= (caddr_t
)(ip
+ 1);
677 i
= m
->m_len
- (sizeof (struct ip
) + olen
);
678 memcpy(opts
, opts
+ olen
, (unsigned)i
);
681 ip
->ip_hl
= sizeof(struct ip
) >> 2;