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.
41 #include "qemu/osdep.h"
43 #include <qemu/osdep.h>
46 static struct ip *ip_reass(Slirp *slirp, struct ip *ip, struct ipq *fp);
47 static void ip_freef(Slirp *slirp, struct ipq *fp);
48 static void ip_enq(register struct ipasfrag *p,
49 register struct ipasfrag *prev);
50 static void ip_deq(register struct ipasfrag *p);
53 * IP initialization: fill in IP protocol switch table.
54 * All protocols not implemented in kernel go to raw IP protocol handler.
59 slirp->ipq.ip_link.next = slirp->ipq.ip_link.prev = &slirp->ipq.ip_link;
65 void ip_cleanup(Slirp *slirp)
73 * Ip input routine. Checksum and byte swap header. If fragmented
74 * try to reassemble. Process options. Pass to next level.
77 ip_input(struct mbuf *m)
79 Slirp *slirp = m->slirp;
80 register struct ip *ip;
83 if (!slirp->in_enabled) {
87 DEBUG_CALL("ip_input");
88 DEBUG_ARG("m = %p", m);
89 DEBUG_ARG("m_len = %d", m->m_len);
91 if (m->m_len < sizeof (struct ip)) {
95 ip = mtod(m, struct ip *);
97 if (ip->ip_v != IPVERSION) {
101 hlen = ip->ip_hl << 2;
102 if (hlen<sizeof(struct ip ) || hlen>m->m_len) {/* min header length */
103 goto bad; /* or packet too short */
106 /* keep ip header intact for ICMP reply
107 * ip->ip_sum = cksum(m, hlen);
115 * Convert fields to host representation.
118 if (ip->ip_len < hlen) {
125 * Check that the amount of data in the buffers
126 * is as at least much as the IP header would have us expect.
127 * Trim mbufs if longer than we expect.
128 * Drop packet if shorter than we expect.
130 if (m->m_len < ip->ip_len) {
134 /* Should drop packet if mbuf too long? hmmm... */
135 if (m->m_len > ip->ip_len)
136 m_adj(m, ip->ip_len - m->m_len);
138 /* check ip_ttl for a correct ICMP reply */
139 if (ip->ip_ttl == 0) {
140 icmp_send_error(m, ICMP_TIMXCEED, ICMP_TIMXCEED_INTRANS, 0, "ttl");
145 * If offset or IP_MF are set, must reassemble.
146 * Otherwise, nothing need be done.
147 * (We could look in the reassembly queue to see
148 * if the packet was previously fragmented,
149 * but it's not worth the time; just let them time out.)
151 * XXX This should fail, don't fragment yet
153 if (ip->ip_off &~ IP_DF) {
154 register struct ipq *fp;
157 * Look for queue of fragments
160 for (l = slirp->ipq.ip_link.next; l != &slirp->ipq.ip_link;
162 fp = container_of(l, struct ipq, ip_link);
163 if (ip->ip_id == fp->ipq_id &&
164 ip->ip_src.s_addr == fp->ipq_src.s_addr &&
165 ip->ip_dst.s_addr == fp->ipq_dst.s_addr &&
166 ip->ip_p == fp->ipq_p)
173 * Adjust ip_len to not reflect header,
174 * set ip_mff if more fragments are expected,
175 * convert offset of this to bytes.
178 if (ip->ip_off & IP_MF)
186 * If datagram marked as having more fragments
187 * or if this is not the first fragment,
188 * attempt reassembly; if it succeeds, proceed.
190 if (ip->ip_tos & 1 || ip->ip_off) {
191 ip = ip_reass(slirp, ip, fp);
203 * Switch out to protocol's input routine.
207 tcp_input(m, hlen, (struct socket *)NULL, AF_INET);
223 #define iptofrag(P) ((struct ipasfrag *)(((char*)(P)) - sizeof(struct qlink)))
224 #define fragtoip(P) ((struct ip*)(((char*)(P)) + sizeof(struct qlink)))
226 * Take incoming datagram fragment and try to
227 * reassemble it into whole datagram. If a chain for
228 * reassembly of this datagram already exists, then it
229 * is given as fp; otherwise have to make a chain.
232 ip_reass(Slirp *slirp, struct ip *ip, struct ipq *fp)
234 register struct mbuf *m = dtom(slirp, ip);
235 register struct ipasfrag *q;
236 int hlen = ip->ip_hl << 2;
239 DEBUG_CALL("ip_reass");
240 DEBUG_ARG("ip = %p", ip);
241 DEBUG_ARG("fp = %p", fp);
242 DEBUG_ARG("m = %p", m);
245 * Presence of header sizes in mbufs
246 * would confuse code below.
247 * Fragment m_data is concatenated.
253 * If first fragment to arrive, create a reassembly queue.
256 struct mbuf *t = m_get(slirp);
261 fp = mtod(t, struct ipq *);
262 insque(&fp->ip_link, &slirp->ipq.ip_link);
263 fp->ipq_ttl = IPFRAGTTL;
264 fp->ipq_p = ip->ip_p;
265 fp->ipq_id = ip->ip_id;
266 fp->frag_link.next = fp->frag_link.prev = &fp->frag_link;
267 fp->ipq_src = ip->ip_src;
268 fp->ipq_dst = ip->ip_dst;
269 q = (struct ipasfrag *)fp;
274 * Find a segment which begins after this one does.
276 for (q = fp->frag_link.next; q != (struct ipasfrag *)&fp->frag_link;
278 if (q->ipf_off > ip->ip_off)
282 * If there is a preceding segment, it may provide some of
283 * our data already. If so, drop the data from the incoming
284 * segment. If it provides all of our data, drop us.
286 if (q->ipf_prev != &fp->frag_link) {
287 struct ipasfrag *pq = q->ipf_prev;
288 i = pq->ipf_off + pq->ipf_len - ip->ip_off;
292 m_adj(dtom(slirp, ip), i);
299 * While we overlap succeeding segments trim them or,
300 * if they are completely covered, dequeue them.
302 while (q != (struct ipasfrag*)&fp->frag_link &&
303 ip->ip_off + ip->ip_len > q->ipf_off) {
304 i = (ip->ip_off + ip->ip_len) - q->ipf_off;
305 if (i < q->ipf_len) {
308 m_adj(dtom(slirp, q), i);
312 m_free(dtom(slirp, q->ipf_prev));
318 * Stick new segment in its place;
319 * check for complete reassembly.
321 ip_enq(iptofrag(ip), q->ipf_prev);
323 for (q = fp->frag_link.next; q != (struct ipasfrag*)&fp->frag_link;
325 if (q->ipf_off != next)
329 if (((struct ipasfrag *)(q->ipf_prev))->ipf_tos & 1)
333 * Reassembly is complete; concatenate fragments.
335 q = fp->frag_link.next;
338 q = (struct ipasfrag *) q->ipf_next;
339 while (q != (struct ipasfrag*)&fp->frag_link) {
340 struct mbuf *t = dtom(slirp, q);
341 q = (struct ipasfrag *) q->ipf_next;
346 * Create header for new ip packet by
347 * modifying header of first packet;
348 * dequeue and discard fragment reassembly header.
349 * Make header visible.
351 q = fp->frag_link.next;
354 * If the fragments concatenated to an mbuf that's
355 * bigger than the total size of the fragment, then and
356 * m_ext buffer was alloced. But fp->ipq_next points to
357 * the old buffer (in the mbuf), so we must point ip
358 * into the new buffer.
360 if (m->m_flags & M_EXT) {
361 int delta = (char *)q - m->m_dat;
362 q = (struct ipasfrag *)(m->m_ext + delta);
368 ip->ip_src = fp->ipq_src;
369 ip->ip_dst = fp->ipq_dst;
370 remque(&fp->ip_link);
371 (void) m_free(dtom(slirp, fp));
372 m->m_len += (ip->ip_hl << 2);
373 m->m_data -= (ip->ip_hl << 2);
383 * Free a fragment reassembly header and all
384 * associated datagrams.
387 ip_freef(Slirp *slirp, struct ipq *fp)
389 register struct ipasfrag *q, *p;
391 for (q = fp->frag_link.next; q != (struct ipasfrag*)&fp->frag_link; q = p) {
394 m_free(dtom(slirp, q));
396 remque(&fp->ip_link);
397 (void) m_free(dtom(slirp, fp));
401 * Put an ip fragment on a reassembly chain.
402 * Like insque, but pointers in middle of structure.
405 ip_enq(register struct ipasfrag *p, register struct ipasfrag *prev)
407 DEBUG_CALL("ip_enq");
408 DEBUG_ARG("prev = %p", prev);
410 p->ipf_next = prev->ipf_next;
411 ((struct ipasfrag *)(prev->ipf_next))->ipf_prev = p;
416 * To ip_enq as remque is to insque.
419 ip_deq(register struct ipasfrag *p)
421 ((struct ipasfrag *)(p->ipf_prev))->ipf_next = p->ipf_next;
422 ((struct ipasfrag *)(p->ipf_next))->ipf_prev = p->ipf_prev;
426 * IP timer processing;
427 * if a timer expires on a reassembly
431 ip_slowtimo(Slirp *slirp)
435 DEBUG_CALL("ip_slowtimo");
437 l = slirp->ipq.ip_link.next;
442 while (l != &slirp->ipq.ip_link) {
443 struct ipq *fp = container_of(l, struct ipq, ip_link);
445 if (--fp->ipq_ttl == 0) {
452 * Do option processing on a datagram,
453 * possibly discarding it if bad options are encountered,
454 * or forwarding it if source-routed.
455 * Returns 1 if packet has been forwarded/freed,
456 * 0 if the packet should be processed further.
465 register struct ip *ip = mtod(m, struct ip *);
467 register struct ip_timestamp *ipt;
468 register struct in_ifaddr *ia;
469 int opt, optlen, cnt, off, code, type, forward = 0;
470 struct in_addr *sin, dst;
471 typedef uint32_t n_time;
475 cp = (u_char *)(ip + 1);
476 cnt = (ip->ip_hl << 2) - sizeof (struct ip);
477 for (; cnt > 0; cnt -= optlen, cp += optlen) {
478 opt = cp[IPOPT_OPTVAL];
479 if (opt == IPOPT_EOL)
481 if (opt == IPOPT_NOP)
484 optlen = cp[IPOPT_OLEN];
485 if (optlen <= 0 || optlen > cnt) {
486 code = &cp[IPOPT_OLEN] - (u_char *)ip;
496 * Source routing with record.
497 * Find interface with current destination address.
498 * If none on this machine then drop if strictly routed,
499 * or do nothing if loosely routed.
500 * Record interface address and bring up next address
501 * component. If strictly routed make sure next
502 * address is on directly accessible net.
506 if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) {
507 code = &cp[IPOPT_OFFSET] - (u_char *)ip;
510 ipaddr.sin_addr = ip->ip_dst;
511 ia = (struct in_ifaddr *)
512 ifa_ifwithaddr((struct sockaddr *)&ipaddr);
514 if (opt == IPOPT_SSRR) {
516 code = ICMP_UNREACH_SRCFAIL;
520 * Loose routing, and not at next destination
521 * yet; nothing to do except forward.
525 off--; /* 0 origin */
526 if (off > optlen - sizeof(struct in_addr)) {
528 * End of source route. Should be for us.
530 save_rte(cp, ip->ip_src);
534 * locate outgoing interface
536 bcopy((caddr_t)(cp + off), (caddr_t)&ipaddr.sin_addr,
537 sizeof(ipaddr.sin_addr));
538 if (opt == IPOPT_SSRR) {
539 #define INA struct in_ifaddr *
540 #define SA struct sockaddr *
541 if ((ia = (INA)ifa_ifwithdstaddr((SA)&ipaddr)) == 0)
542 ia = (INA)ifa_ifwithnet((SA)&ipaddr);
544 ia = ip_rtaddr(ipaddr.sin_addr);
547 code = ICMP_UNREACH_SRCFAIL;
550 ip->ip_dst = ipaddr.sin_addr;
551 bcopy((caddr_t)&(IA_SIN(ia)->sin_addr),
552 (caddr_t)(cp + off), sizeof(struct in_addr));
553 cp[IPOPT_OFFSET] += sizeof(struct in_addr);
555 * Let ip_intr's mcast routing check handle mcast pkts
557 forward = !IN_MULTICAST(ntohl(ip->ip_dst.s_addr));
561 if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) {
562 code = &cp[IPOPT_OFFSET] - (u_char *)ip;
566 * If no space remains, ignore.
568 off--; /* 0 origin */
569 if (off > optlen - sizeof(struct in_addr))
571 bcopy((caddr_t)(&ip->ip_dst), (caddr_t)&ipaddr.sin_addr,
572 sizeof(ipaddr.sin_addr));
574 * locate outgoing interface; if we're the destination,
575 * use the incoming interface (should be same).
577 if ((ia = (INA)ifa_ifwithaddr((SA)&ipaddr)) == 0 &&
578 (ia = ip_rtaddr(ipaddr.sin_addr)) == 0) {
580 code = ICMP_UNREACH_HOST;
583 bcopy((caddr_t)&(IA_SIN(ia)->sin_addr),
584 (caddr_t)(cp + off), sizeof(struct in_addr));
585 cp[IPOPT_OFFSET] += sizeof(struct in_addr);
589 code = cp - (u_char *)ip;
590 ipt = (struct ip_timestamp *)cp;
591 if (ipt->ipt_len < 5)
593 if (ipt->ipt_ptr > ipt->ipt_len - sizeof (int32_t)) {
594 if (++ipt->ipt_oflw == 0)
598 sin = (struct in_addr *)(cp + ipt->ipt_ptr - 1);
599 switch (ipt->ipt_flg) {
601 case IPOPT_TS_TSONLY:
604 case IPOPT_TS_TSANDADDR:
605 if (ipt->ipt_ptr + sizeof(n_time) +
606 sizeof(struct in_addr) > ipt->ipt_len)
608 ipaddr.sin_addr = dst;
609 ia = (INA)ifaof_ i f p foraddr((SA)&ipaddr,
613 bcopy((caddr_t)&IA_SIN(ia)->sin_addr,
614 (caddr_t)sin, sizeof(struct in_addr));
615 ipt->ipt_ptr += sizeof(struct in_addr);
618 case IPOPT_TS_PRESPEC:
619 if (ipt->ipt_ptr + sizeof(n_time) +
620 sizeof(struct in_addr) > ipt->ipt_len)
622 bcopy((caddr_t)sin, (caddr_t)&ipaddr.sin_addr,
623 sizeof(struct in_addr));
624 if (ifa_ifwithaddr((SA)&ipaddr) == 0)
626 ipt->ipt_ptr += sizeof(struct in_addr);
633 bcopy((caddr_t)&ntime, (caddr_t)cp + ipt->ipt_ptr - 1,
635 ipt->ipt_ptr += sizeof(n_time);
644 icmp_send_error(m, type, code, 0, 0);
652 * Strip out IP options, at higher
653 * level protocol in the kernel.
654 * Second argument is buffer to which options
655 * will be moved, and return value is their length.
656 * (XXX) should be deleted; last arg currently ignored.
659 ip_stripoptions(register struct mbuf *m, struct mbuf *mopt)
662 struct ip *ip = mtod(m, struct ip *);
663 register caddr_t opts;
666 olen = (ip->ip_hl<<2) - sizeof (struct ip);
667 opts = (caddr_t)(ip + 1);
668 i = m->m_len - (sizeof (struct ip) + olen);
669 memcpy(opts, opts + olen, (unsigned)i);
672 ip->ip_hl = sizeof(struct ip) >> 2;