1 | /* |
---|
2 | * Copyright (c) 1982, 1986, 1988, 1993 |
---|
3 | * The Regents of the University of California. All rights reserved. |
---|
4 | * |
---|
5 | * Redistribution and use in source and binary forms, with or without |
---|
6 | * modification, are permitted provided that the following conditions |
---|
7 | * are met: |
---|
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. |
---|
20 | * |
---|
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 |
---|
31 | * SUCH DAMAGE. |
---|
32 | * |
---|
33 | * @(#)ip_input.c 8.2 (Berkeley) 1/4/94 |
---|
34 | * $Id$ |
---|
35 | * $ANA: ip_input.c,v 1.5 1996/09/18 14:34:59 wollman Exp $ |
---|
36 | */ |
---|
37 | |
---|
38 | #define _IP_VHL |
---|
39 | |
---|
40 | #include "opt_ipfw.h" |
---|
41 | |
---|
42 | #include <stddef.h> |
---|
43 | |
---|
44 | #include <sys/param.h> |
---|
45 | #include <sys/systm.h> |
---|
46 | #include <sys/malloc.h> |
---|
47 | #include <sys/mbuf.h> |
---|
48 | #include <sys/domain.h> |
---|
49 | #include <sys/protosw.h> |
---|
50 | #include <sys/socket.h> |
---|
51 | #include <sys/errno.h> |
---|
52 | #include <sys/time.h> |
---|
53 | #include <sys/kernel.h> |
---|
54 | #include <sys/syslog.h> |
---|
55 | #include <sys/sysctl.h> |
---|
56 | |
---|
57 | #include <net/if.h> |
---|
58 | #include <net/if_dl.h> |
---|
59 | #include <net/route.h> |
---|
60 | #include <net/netisr.h> |
---|
61 | |
---|
62 | #include <netinet/in.h> |
---|
63 | #include <netinet/in_systm.h> |
---|
64 | #include <netinet/in_var.h> |
---|
65 | #include <netinet/ip.h> |
---|
66 | #include <netinet/in_pcb.h> |
---|
67 | #include <netinet/in_var.h> |
---|
68 | #include <netinet/ip_var.h> |
---|
69 | #include <netinet/ip_icmp.h> |
---|
70 | #include <machine/in_cksum.h> |
---|
71 | |
---|
72 | #include <sys/socketvar.h> |
---|
73 | |
---|
74 | #ifdef IPFIREWALL |
---|
75 | #include <netinet/ip_fw.h> |
---|
76 | #endif |
---|
77 | |
---|
78 | int rsvp_on = 0; |
---|
79 | static int ip_rsvp_on; |
---|
80 | struct socket *ip_rsvpd; |
---|
81 | |
---|
82 | static int ipforwarding = 0; |
---|
83 | SYSCTL_INT(_net_inet_ip, IPCTL_FORWARDING, forwarding, CTLFLAG_RW, |
---|
84 | &ipforwarding, 0, ""); |
---|
85 | |
---|
86 | static int ipsendredirects = 1; /* XXX */ |
---|
87 | SYSCTL_INT(_net_inet_ip, IPCTL_SENDREDIRECTS, redirect, CTLFLAG_RW, |
---|
88 | &ipsendredirects, 0, ""); |
---|
89 | |
---|
90 | int ip_defttl = IPDEFTTL; |
---|
91 | SYSCTL_INT(_net_inet_ip, IPCTL_DEFTTL, ttl, CTLFLAG_RW, |
---|
92 | &ip_defttl, 0, ""); |
---|
93 | |
---|
94 | static int ip_dosourceroute = 0; |
---|
95 | SYSCTL_INT(_net_inet_ip, IPCTL_SOURCEROUTE, sourceroute, CTLFLAG_RW, |
---|
96 | &ip_dosourceroute, 0, ""); |
---|
97 | |
---|
98 | static int ip_acceptsourceroute = 0; |
---|
99 | SYSCTL_INT(_net_inet_ip, IPCTL_ACCEPTSOURCEROUTE, accept_sourceroute, |
---|
100 | CTLFLAG_RW, &ip_acceptsourceroute, 0, ""); |
---|
101 | #ifdef DIAGNOSTIC |
---|
102 | static int ipprintfs = 0; |
---|
103 | #endif |
---|
104 | |
---|
105 | extern struct domain inetdomain; |
---|
106 | extern struct protosw inetsw[]; |
---|
107 | u_char ip_protox[IPPROTO_MAX]; |
---|
108 | static int ipqmaxlen = IFQ_MAXLEN; |
---|
109 | struct in_ifaddr *in_ifaddr; /* first inet address */ |
---|
110 | struct ifqueue ipintrq; |
---|
111 | SYSCTL_INT(_net_inet_ip, IPCTL_INTRQMAXLEN, intr_queue_maxlen, CTLFLAG_RD, |
---|
112 | &ipintrq.ifq_maxlen, 0, ""); |
---|
113 | SYSCTL_INT(_net_inet_ip, IPCTL_INTRQDROPS, intr_queue_drops, CTLFLAG_RD, |
---|
114 | &ipintrq.ifq_drops, 0, ""); |
---|
115 | |
---|
116 | struct ipstat ipstat; |
---|
117 | |
---|
118 | /* Packet reassembly stuff */ |
---|
119 | #define IPREASS_NHASH_LOG2 6 |
---|
120 | #define IPREASS_NHASH (1 << IPREASS_NHASH_LOG2) |
---|
121 | #define IPREASS_HMASK (IPREASS_NHASH - 1) |
---|
122 | #define IPREASS_HASH(x,y) \ |
---|
123 | ((((x) & 0xF | ((((x) >> 8) & 0xF) << 4)) ^ (y)) & IPREASS_HMASK) |
---|
124 | |
---|
125 | static struct ipq ipq[IPREASS_NHASH]; |
---|
126 | static int nipq = 0; /* total # of reass queues */ |
---|
127 | static int maxnipq; |
---|
128 | |
---|
129 | #ifdef IPCTL_DEFMTU |
---|
130 | SYSCTL_INT(_net_inet_ip, IPCTL_DEFMTU, mtu, CTLFLAG_RW, |
---|
131 | &ip_mtu, 0, ""); |
---|
132 | #endif |
---|
133 | |
---|
134 | #if !defined(COMPAT_IPFW) || COMPAT_IPFW == 1 |
---|
135 | #undef COMPAT_IPFW |
---|
136 | #define COMPAT_IPFW 1 |
---|
137 | #else |
---|
138 | #undef COMPAT_IPFW |
---|
139 | #endif |
---|
140 | |
---|
141 | #ifdef COMPAT_IPFW |
---|
142 | /* Firewall hooks */ |
---|
143 | ip_fw_chk_t *ip_fw_chk_ptr; |
---|
144 | ip_fw_ctl_t *ip_fw_ctl_ptr; |
---|
145 | |
---|
146 | /* IP Network Address Translation (NAT) hooks */ |
---|
147 | ip_nat_t *ip_nat_ptr; |
---|
148 | ip_nat_ctl_t *ip_nat_ctl_ptr; |
---|
149 | #endif |
---|
150 | |
---|
151 | /* |
---|
152 | * We need to save the IP options in case a protocol wants to respond |
---|
153 | * to an incoming packet over the same route if the packet got here |
---|
154 | * using IP source routing. This allows connection establishment and |
---|
155 | * maintenance when the remote end is on a network that is not known |
---|
156 | * to us. |
---|
157 | */ |
---|
158 | static int ip_nhops = 0; |
---|
159 | static struct ip_srcrt { |
---|
160 | struct in_addr dst; /* final destination */ |
---|
161 | char nop; /* one NOP to align */ |
---|
162 | char srcopt[IPOPT_OFFSET + 1]; /* OPTVAL, OLEN and OFFSET */ |
---|
163 | struct in_addr route[MAX_IPOPTLEN/sizeof(struct in_addr)]; |
---|
164 | } ip_srcrt; |
---|
165 | |
---|
166 | #ifdef IPDIVERT |
---|
167 | /* |
---|
168 | * Shared variable between ip_input() and ip_reass() to communicate |
---|
169 | * about which packets, once assembled from fragments, get diverted, |
---|
170 | * and to which port. |
---|
171 | */ |
---|
172 | static u_short frag_divert_port; |
---|
173 | #endif |
---|
174 | |
---|
175 | static void save_rte __P((u_char *, struct in_addr)); |
---|
176 | static void ip_deq __P((struct ipasfrag *)); |
---|
177 | static int ip_dooptions __P((struct mbuf *)); |
---|
178 | static void ip_enq __P((struct ipasfrag *, struct ipasfrag *)); |
---|
179 | static void ip_forward __P((struct mbuf *, int)); |
---|
180 | static void ip_freef __P((struct ipq *)); |
---|
181 | static struct ip * |
---|
182 | ip_reass __P((struct ipasfrag *, struct ipq *, struct ipq *)); |
---|
183 | static struct in_ifaddr * |
---|
184 | ip_rtaddr __P((struct in_addr)); |
---|
185 | void ipintr __P((void)); |
---|
186 | /* |
---|
187 | * IP initialization: fill in IP protocol switch table. |
---|
188 | * All protocols not implemented in kernel go to raw IP protocol handler. |
---|
189 | */ |
---|
190 | void |
---|
191 | ip_init() |
---|
192 | { |
---|
193 | register struct protosw *pr; |
---|
194 | register int i; |
---|
195 | |
---|
196 | pr = pffindproto(PF_INET, IPPROTO_RAW, SOCK_RAW); |
---|
197 | if (pr == 0) |
---|
198 | panic("ip_init"); |
---|
199 | for (i = 0; i < IPPROTO_MAX; i++) |
---|
200 | ip_protox[i] = pr - inetsw; |
---|
201 | for (pr = inetdomain.dom_protosw; |
---|
202 | pr < inetdomain.dom_protoswNPROTOSW; pr++) |
---|
203 | if (pr->pr_domain->dom_family == PF_INET && |
---|
204 | pr->pr_protocol && pr->pr_protocol != IPPROTO_RAW) |
---|
205 | ip_protox[pr->pr_protocol] = pr - inetsw; |
---|
206 | |
---|
207 | for (i = 0; i < IPREASS_NHASH; i++) |
---|
208 | ipq[i].next = ipq[i].prev = &ipq[i]; |
---|
209 | |
---|
210 | maxnipq = nmbclusters/4; |
---|
211 | |
---|
212 | ip_id = rtems_bsdnet_seconds_since_boot() & 0xffff; |
---|
213 | ipintrq.ifq_maxlen = ipqmaxlen; |
---|
214 | #ifdef IPFIREWALL |
---|
215 | ip_fw_init(); |
---|
216 | #endif |
---|
217 | #ifdef IPNAT |
---|
218 | ip_nat_init(); |
---|
219 | #endif |
---|
220 | |
---|
221 | } |
---|
222 | |
---|
223 | static struct sockaddr_in ipaddr = { sizeof(ipaddr), AF_INET }; |
---|
224 | static struct route ipforward_rt; |
---|
225 | |
---|
226 | /* |
---|
227 | * Ip input routine. Checksum and byte swap header. If fragmented |
---|
228 | * try to reassemble. Process options. Pass to next level. |
---|
229 | */ |
---|
230 | void |
---|
231 | ip_input(struct mbuf *m) |
---|
232 | { |
---|
233 | struct ip *ip; |
---|
234 | struct ipq *fp; |
---|
235 | struct in_ifaddr *ia; |
---|
236 | int i, hlen; |
---|
237 | u_short sum; |
---|
238 | |
---|
239 | #ifdef DIAGNOSTIC |
---|
240 | if ((m->m_flags & M_PKTHDR) == 0) |
---|
241 | panic("ip_input no HDR"); |
---|
242 | #endif |
---|
243 | /* |
---|
244 | * If no IP addresses have been set yet but the interfaces |
---|
245 | * are receiving, can't do anything with incoming packets yet. |
---|
246 | */ |
---|
247 | if (in_ifaddr == NULL) |
---|
248 | goto bad; |
---|
249 | ipstat.ips_total++; |
---|
250 | |
---|
251 | if (m->m_pkthdr.len < sizeof(struct ip)) |
---|
252 | goto tooshort; |
---|
253 | |
---|
254 | #ifdef DIAGNOSTIC |
---|
255 | if (m->m_len < sizeof(struct ip)) |
---|
256 | panic("ipintr mbuf too short"); |
---|
257 | #endif |
---|
258 | |
---|
259 | if (m->m_len < sizeof (struct ip) && |
---|
260 | (m = m_pullup(m, sizeof (struct ip))) == 0) { |
---|
261 | ipstat.ips_toosmall++; |
---|
262 | return; |
---|
263 | } |
---|
264 | ip = mtod(m, struct ip *); |
---|
265 | |
---|
266 | if (IP_VHL_V(ip->ip_vhl) != IPVERSION) { |
---|
267 | ipstat.ips_badvers++; |
---|
268 | goto bad; |
---|
269 | } |
---|
270 | |
---|
271 | hlen = IP_VHL_HL(ip->ip_vhl) << 2; |
---|
272 | if (hlen < sizeof(struct ip)) { /* minimum header length */ |
---|
273 | ipstat.ips_badhlen++; |
---|
274 | goto bad; |
---|
275 | } |
---|
276 | if (hlen > m->m_len) { |
---|
277 | if ((m = m_pullup(m, hlen)) == 0) { |
---|
278 | ipstat.ips_badhlen++; |
---|
279 | return; |
---|
280 | } |
---|
281 | ip = mtod(m, struct ip *); |
---|
282 | } |
---|
283 | if (hlen == sizeof(struct ip)) { |
---|
284 | sum = in_cksum_hdr(ip); |
---|
285 | } else { |
---|
286 | sum = in_cksum(m, hlen); |
---|
287 | } |
---|
288 | if (sum) { |
---|
289 | ipstat.ips_badsum++; |
---|
290 | goto bad; |
---|
291 | } |
---|
292 | |
---|
293 | /* |
---|
294 | * Convert fields to host representation. |
---|
295 | */ |
---|
296 | NTOHS(ip->ip_len); |
---|
297 | if (ip->ip_len < hlen) { |
---|
298 | ipstat.ips_badlen++; |
---|
299 | goto bad; |
---|
300 | } |
---|
301 | NTOHS(ip->ip_id); |
---|
302 | NTOHS(ip->ip_off); |
---|
303 | |
---|
304 | /* |
---|
305 | * Check that the amount of data in the buffers |
---|
306 | * is as at least much as the IP header would have us expect. |
---|
307 | * Trim mbufs if longer than we expect. |
---|
308 | * Drop packet if shorter than we expect. |
---|
309 | */ |
---|
310 | if (m->m_pkthdr.len < ip->ip_len) { |
---|
311 | tooshort: |
---|
312 | ipstat.ips_tooshort++; |
---|
313 | goto bad; |
---|
314 | } |
---|
315 | if (m->m_pkthdr.len > ip->ip_len) { |
---|
316 | if (m->m_len == m->m_pkthdr.len) { |
---|
317 | m->m_len = ip->ip_len; |
---|
318 | m->m_pkthdr.len = ip->ip_len; |
---|
319 | } else |
---|
320 | m_adj(m, ip->ip_len - m->m_pkthdr.len); |
---|
321 | } |
---|
322 | /* |
---|
323 | * IpHack's section. |
---|
324 | * Right now when no processing on packet has done |
---|
325 | * and it is still fresh out of network we do our black |
---|
326 | * deals with it. |
---|
327 | * - Firewall: deny/allow/divert |
---|
328 | * - Xlate: translate packet's addr/port (NAT). |
---|
329 | * - Wrap: fake packet's addr/port <unimpl.> |
---|
330 | * - Encapsulate: put it in another IP and send out. <unimp.> |
---|
331 | */ |
---|
332 | |
---|
333 | #ifdef COMPAT_IPFW |
---|
334 | if (ip_fw_chk_ptr) { |
---|
335 | #ifdef IPDIVERT |
---|
336 | u_short port; |
---|
337 | |
---|
338 | port = (*ip_fw_chk_ptr)(&ip, hlen, NULL, ip_divert_ignore, &m); |
---|
339 | ip_divert_ignore = 0; |
---|
340 | if (port) { /* Divert packet */ |
---|
341 | frag_divert_port = port; |
---|
342 | goto ours; |
---|
343 | } |
---|
344 | #else |
---|
345 | /* If ipfw says divert, we have to just drop packet */ |
---|
346 | if ((*ip_fw_chk_ptr)(&ip, hlen, NULL, 0, &m)) { |
---|
347 | m_freem(m); |
---|
348 | m = NULL; |
---|
349 | } |
---|
350 | #endif |
---|
351 | if (!m) |
---|
352 | return; |
---|
353 | } |
---|
354 | |
---|
355 | if (ip_nat_ptr && !(*ip_nat_ptr)(&ip, &m, m->m_pkthdr.rcvif, IP_NAT_IN)) |
---|
356 | return; |
---|
357 | #endif |
---|
358 | |
---|
359 | /* |
---|
360 | * Process options and, if not destined for us, |
---|
361 | * ship it on. ip_dooptions returns 1 when an |
---|
362 | * error was detected (causing an icmp message |
---|
363 | * to be sent and the original packet to be freed). |
---|
364 | */ |
---|
365 | ip_nhops = 0; /* for source routed packets */ |
---|
366 | if (hlen > sizeof (struct ip) && ip_dooptions(m)) |
---|
367 | return; |
---|
368 | |
---|
369 | /* greedy RSVP, snatches any PATH packet of the RSVP protocol and no |
---|
370 | * matter if it is destined to another node, or whether it is |
---|
371 | * a multicast one, RSVP wants it! and prevents it from being forwarded |
---|
372 | * anywhere else. Also checks if the rsvp daemon is running before |
---|
373 | * grabbing the packet. |
---|
374 | */ |
---|
375 | if (rsvp_on && ip->ip_p==IPPROTO_RSVP) |
---|
376 | goto ours; |
---|
377 | |
---|
378 | /* |
---|
379 | * Check our list of addresses, to see if the packet is for us. |
---|
380 | */ |
---|
381 | for (ia = in_ifaddr; ia; ia = ia->ia_next) { |
---|
382 | #define satosin(sa) ((struct sockaddr_in *)(sa)) |
---|
383 | |
---|
384 | if (IA_SIN(ia)->sin_addr.s_addr == ip->ip_dst.s_addr) |
---|
385 | goto ours; |
---|
386 | #ifdef BOOTP_COMPAT |
---|
387 | if (IA_SIN(ia)->sin_addr.s_addr == INADDR_ANY) |
---|
388 | goto ours; |
---|
389 | #endif |
---|
390 | if (ia->ia_ifp && ia->ia_ifp->if_flags & IFF_BROADCAST) { |
---|
391 | if (satosin(&ia->ia_broadaddr)->sin_addr.s_addr == |
---|
392 | ip->ip_dst.s_addr) |
---|
393 | goto ours; |
---|
394 | if (ip->ip_dst.s_addr == ia->ia_netbroadcast.s_addr) |
---|
395 | goto ours; |
---|
396 | } |
---|
397 | } |
---|
398 | if (IN_MULTICAST(ntohl(ip->ip_dst.s_addr))) { |
---|
399 | struct in_multi *inm; |
---|
400 | if (ip_mrouter) { |
---|
401 | /* |
---|
402 | * If we are acting as a multicast router, all |
---|
403 | * incoming multicast packets are passed to the |
---|
404 | * kernel-level multicast forwarding function. |
---|
405 | * The packet is returned (relatively) intact; if |
---|
406 | * ip_mforward() returns a non-zero value, the packet |
---|
407 | * must be discarded, else it may be accepted below. |
---|
408 | * |
---|
409 | * (The IP ident field is put in the same byte order |
---|
410 | * as expected when ip_mforward() is called from |
---|
411 | * ip_output().) |
---|
412 | */ |
---|
413 | ip->ip_id = htons(ip->ip_id); |
---|
414 | if (ip_mforward(ip, m->m_pkthdr.rcvif, m, 0) != 0) { |
---|
415 | ipstat.ips_cantforward++; |
---|
416 | m_freem(m); |
---|
417 | return; |
---|
418 | } |
---|
419 | ip->ip_id = ntohs(ip->ip_id); |
---|
420 | |
---|
421 | /* |
---|
422 | * The process-level routing demon needs to receive |
---|
423 | * all multicast IGMP packets, whether or not this |
---|
424 | * host belongs to their destination groups. |
---|
425 | */ |
---|
426 | if (ip->ip_p == IPPROTO_IGMP) |
---|
427 | goto ours; |
---|
428 | ipstat.ips_forward++; |
---|
429 | } |
---|
430 | /* |
---|
431 | * See if we belong to the destination multicast group on the |
---|
432 | * arrival interface. |
---|
433 | */ |
---|
434 | IN_LOOKUP_MULTI(ip->ip_dst, m->m_pkthdr.rcvif, inm); |
---|
435 | if (inm == NULL) { |
---|
436 | ipstat.ips_cantforward++; |
---|
437 | m_freem(m); |
---|
438 | return; |
---|
439 | } |
---|
440 | goto ours; |
---|
441 | } |
---|
442 | if (ip->ip_dst.s_addr == (u_long)INADDR_BROADCAST) |
---|
443 | goto ours; |
---|
444 | if (ip->ip_dst.s_addr == INADDR_ANY) |
---|
445 | goto ours; |
---|
446 | |
---|
447 | /* |
---|
448 | * Not for us; forward if possible and desirable. |
---|
449 | */ |
---|
450 | if (ipforwarding == 0) { |
---|
451 | ipstat.ips_cantforward++; |
---|
452 | m_freem(m); |
---|
453 | } else |
---|
454 | ip_forward(m, 0); |
---|
455 | return; |
---|
456 | |
---|
457 | ours: |
---|
458 | |
---|
459 | /* |
---|
460 | * If offset or IP_MF are set, must reassemble. |
---|
461 | * Otherwise, nothing need be done. |
---|
462 | * (We could look in the reassembly queue to see |
---|
463 | * if the packet was previously fragmented, |
---|
464 | * but it's not worth the time; just let them time out.) |
---|
465 | */ |
---|
466 | if (ip->ip_off &~ (IP_DF | IP_RF)) { |
---|
467 | if (m->m_flags & M_EXT) { /* XXX */ |
---|
468 | if ((m = m_pullup(m, sizeof (struct ip))) == 0) { |
---|
469 | ipstat.ips_toosmall++; |
---|
470 | #ifdef IPDIVERT |
---|
471 | frag_divert_port = 0; |
---|
472 | #endif |
---|
473 | return; |
---|
474 | } |
---|
475 | ip = mtod(m, struct ip *); |
---|
476 | } |
---|
477 | sum = IPREASS_HASH(ip->ip_src.s_addr, ip->ip_id); |
---|
478 | /* |
---|
479 | * Look for queue of fragments |
---|
480 | * of this datagram. |
---|
481 | */ |
---|
482 | for (fp = ipq[sum].next; fp != &ipq[sum]; fp = fp->next) |
---|
483 | if (ip->ip_id == fp->ipq_id && |
---|
484 | ip->ip_src.s_addr == fp->ipq_src.s_addr && |
---|
485 | ip->ip_dst.s_addr == fp->ipq_dst.s_addr && |
---|
486 | ip->ip_p == fp->ipq_p) |
---|
487 | goto found; |
---|
488 | |
---|
489 | fp = 0; |
---|
490 | |
---|
491 | /* check if there's a place for the new queue */ |
---|
492 | if (nipq > maxnipq) { |
---|
493 | /* |
---|
494 | * drop something from the tail of the current queue |
---|
495 | * before proceeding further |
---|
496 | */ |
---|
497 | if (ipq[sum].prev == &ipq[sum]) { /* gak */ |
---|
498 | for (i = 0; i < IPREASS_NHASH; i++) { |
---|
499 | if (ipq[i].prev != &ipq[i]) { |
---|
500 | ip_freef(ipq[i].prev); |
---|
501 | break; |
---|
502 | } |
---|
503 | } |
---|
504 | } else |
---|
505 | ip_freef(ipq[sum].prev); |
---|
506 | } |
---|
507 | found: |
---|
508 | /* |
---|
509 | * Adjust ip_len to not reflect header, |
---|
510 | * set ip_mff if more fragments are expected, |
---|
511 | * convert offset of this to bytes. |
---|
512 | */ |
---|
513 | ip->ip_len -= hlen; |
---|
514 | ((struct ipasfrag *)ip)->ipf_mff &= ~1; |
---|
515 | if (ip->ip_off & IP_MF) |
---|
516 | ((struct ipasfrag *)ip)->ipf_mff |= 1; |
---|
517 | ip->ip_off <<= 3; |
---|
518 | |
---|
519 | /* |
---|
520 | * If datagram marked as having more fragments |
---|
521 | * or if this is not the first fragment, |
---|
522 | * attempt reassembly; if it succeeds, proceed. |
---|
523 | */ |
---|
524 | if (((struct ipasfrag *)ip)->ipf_mff & 1 || ip->ip_off) { |
---|
525 | ipstat.ips_fragments++; |
---|
526 | ip = ip_reass((struct ipasfrag *)ip, fp, &ipq[sum]); |
---|
527 | if (ip == 0) |
---|
528 | return; |
---|
529 | ipstat.ips_reassembled++; |
---|
530 | m = dtom(ip); |
---|
531 | #ifdef IPDIVERT |
---|
532 | if (frag_divert_port) { |
---|
533 | ip->ip_len += hlen; |
---|
534 | HTONS(ip->ip_len); |
---|
535 | HTONS(ip->ip_off); |
---|
536 | HTONS(ip->ip_id); |
---|
537 | ip->ip_sum = 0; |
---|
538 | ip->ip_sum = in_cksum_hdr(ip); |
---|
539 | NTOHS(ip->ip_id); |
---|
540 | NTOHS(ip->ip_off); |
---|
541 | NTOHS(ip->ip_len); |
---|
542 | ip->ip_len -= hlen; |
---|
543 | } |
---|
544 | #endif |
---|
545 | } else |
---|
546 | if (fp) |
---|
547 | ip_freef(fp); |
---|
548 | } else |
---|
549 | ip->ip_len -= hlen; |
---|
550 | |
---|
551 | #ifdef IPDIVERT |
---|
552 | /* |
---|
553 | * Divert reassembled packets to the divert protocol if required |
---|
554 | */ |
---|
555 | if (frag_divert_port) { |
---|
556 | ipstat.ips_delivered++; |
---|
557 | ip_divert_port = frag_divert_port; |
---|
558 | frag_divert_port = 0; |
---|
559 | (*inetsw[ip_protox[IPPROTO_DIVERT]].pr_input)(m, hlen); |
---|
560 | return; |
---|
561 | } |
---|
562 | |
---|
563 | /* Don't let packets divert themselves */ |
---|
564 | if (ip->ip_p == IPPROTO_DIVERT) { |
---|
565 | ipstat.ips_noproto++; |
---|
566 | goto bad; |
---|
567 | } |
---|
568 | #endif |
---|
569 | |
---|
570 | /* |
---|
571 | * Switch out to protocol's input routine. |
---|
572 | */ |
---|
573 | ipstat.ips_delivered++; |
---|
574 | (*inetsw[ip_protox[ip->ip_p]].pr_input)(m, hlen); |
---|
575 | return; |
---|
576 | bad: |
---|
577 | m_freem(m); |
---|
578 | } |
---|
579 | |
---|
580 | /* |
---|
581 | * IP software interrupt routine - to go away sometime soon |
---|
582 | */ |
---|
583 | void |
---|
584 | ipintr(void) |
---|
585 | { |
---|
586 | int s; |
---|
587 | struct mbuf *m; |
---|
588 | |
---|
589 | while(1) { |
---|
590 | s = splimp(); |
---|
591 | IF_DEQUEUE(&ipintrq, m); |
---|
592 | splx(s); |
---|
593 | if (m == 0) |
---|
594 | return; |
---|
595 | ip_input(m); |
---|
596 | } |
---|
597 | } |
---|
598 | |
---|
599 | NETISR_SET(NETISR_IP, ipintr); |
---|
600 | |
---|
601 | /* |
---|
602 | * Take incoming datagram fragment and try to |
---|
603 | * reassemble it into whole datagram. If a chain for |
---|
604 | * reassembly of this datagram already exists, then it |
---|
605 | * is given as fp; otherwise have to make a chain. |
---|
606 | */ |
---|
607 | static struct ip * |
---|
608 | ip_reass(ip, fp, where) |
---|
609 | register struct ipasfrag *ip; |
---|
610 | register struct ipq *fp; |
---|
611 | struct ipq *where; |
---|
612 | { |
---|
613 | register struct mbuf *m = dtom(ip); |
---|
614 | register struct ipasfrag *q; |
---|
615 | struct mbuf *t; |
---|
616 | int hlen = ip->ip_hl << 2; |
---|
617 | int i, next; |
---|
618 | |
---|
619 | /* |
---|
620 | * Presence of header sizes in mbufs |
---|
621 | * would confuse code below. |
---|
622 | */ |
---|
623 | m->m_data += hlen; |
---|
624 | m->m_len -= hlen; |
---|
625 | |
---|
626 | /* |
---|
627 | * If first fragment to arrive, create a reassembly queue. |
---|
628 | */ |
---|
629 | if (fp == 0) { |
---|
630 | if ((t = m_get(M_DONTWAIT, MT_FTABLE)) == NULL) |
---|
631 | goto dropfrag; |
---|
632 | fp = mtod(t, struct ipq *); |
---|
633 | insque(fp, where); |
---|
634 | nipq++; |
---|
635 | fp->ipq_ttl = IPFRAGTTL; |
---|
636 | fp->ipq_p = ip->ip_p; |
---|
637 | fp->ipq_id = ip->ip_id; |
---|
638 | fp->ipq_next = fp->ipq_prev = (struct ipasfrag *)fp; |
---|
639 | fp->ipq_src = ((struct ip *)ip)->ip_src; |
---|
640 | fp->ipq_dst = ((struct ip *)ip)->ip_dst; |
---|
641 | #ifdef IPDIVERT |
---|
642 | fp->ipq_divert = 0; |
---|
643 | #endif |
---|
644 | q = (struct ipasfrag *)fp; |
---|
645 | goto insert; |
---|
646 | } |
---|
647 | |
---|
648 | /* |
---|
649 | * Find a segment which begins after this one does. |
---|
650 | */ |
---|
651 | for (q = fp->ipq_next; q != (struct ipasfrag *)fp; q = q->ipf_next) |
---|
652 | if (q->ip_off > ip->ip_off) |
---|
653 | break; |
---|
654 | |
---|
655 | /* |
---|
656 | * If there is a preceding segment, it may provide some of |
---|
657 | * our data already. If so, drop the data from the incoming |
---|
658 | * segment. If it provides all of our data, drop us. |
---|
659 | */ |
---|
660 | if (q->ipf_prev != (struct ipasfrag *)fp) { |
---|
661 | i = q->ipf_prev->ip_off + q->ipf_prev->ip_len - ip->ip_off; |
---|
662 | if (i > 0) { |
---|
663 | if (i >= ip->ip_len) |
---|
664 | goto dropfrag; |
---|
665 | m_adj(dtom(ip), i); |
---|
666 | ip->ip_off += i; |
---|
667 | ip->ip_len -= i; |
---|
668 | } |
---|
669 | } |
---|
670 | |
---|
671 | /* |
---|
672 | * While we overlap succeeding segments trim them or, |
---|
673 | * if they are completely covered, dequeue them. |
---|
674 | */ |
---|
675 | while (q != (struct ipasfrag *)fp && ip->ip_off + ip->ip_len > q->ip_off) { |
---|
676 | struct mbuf *m0; |
---|
677 | |
---|
678 | i = (ip->ip_off + ip->ip_len) - q->ip_off; |
---|
679 | if (i < q->ip_len) { |
---|
680 | q->ip_len -= i; |
---|
681 | q->ip_off += i; |
---|
682 | m_adj(dtom(q), i); |
---|
683 | break; |
---|
684 | } |
---|
685 | m0 = dtom(q); |
---|
686 | q = q->ipf_next; |
---|
687 | ip_deq(q->ipf_prev); |
---|
688 | m_freem(m0); |
---|
689 | } |
---|
690 | |
---|
691 | insert: |
---|
692 | |
---|
693 | #ifdef IPDIVERT |
---|
694 | /* |
---|
695 | * Any fragment diverting causes the whole packet to divert |
---|
696 | */ |
---|
697 | if (frag_divert_port != 0) |
---|
698 | fp->ipq_divert = frag_divert_port; |
---|
699 | frag_divert_port = 0; |
---|
700 | #endif |
---|
701 | |
---|
702 | /* |
---|
703 | * Stick new segment in its place; |
---|
704 | * check for complete reassembly. |
---|
705 | */ |
---|
706 | ip_enq(ip, q->ipf_prev); |
---|
707 | next = 0; |
---|
708 | for (q = fp->ipq_next; q != (struct ipasfrag *)fp; q = q->ipf_next) { |
---|
709 | if (q->ip_off != next) |
---|
710 | return (0); |
---|
711 | next += q->ip_len; |
---|
712 | } |
---|
713 | if (q->ipf_prev->ipf_mff & 1) |
---|
714 | return (0); |
---|
715 | |
---|
716 | /* |
---|
717 | * Reassembly is complete. Make sure the packet is a sane size. |
---|
718 | */ |
---|
719 | if (next + (IP_VHL_HL(((struct ip *)fp->ipq_next)->ip_vhl) << 2) |
---|
720 | > IP_MAXPACKET) { |
---|
721 | ipstat.ips_toolong++; |
---|
722 | ip_freef(fp); |
---|
723 | return (0); |
---|
724 | } |
---|
725 | |
---|
726 | /* |
---|
727 | * Concatenate fragments. |
---|
728 | */ |
---|
729 | q = fp->ipq_next; |
---|
730 | m = dtom(q); |
---|
731 | t = m->m_next; |
---|
732 | m->m_next = 0; |
---|
733 | m_cat(m, t); |
---|
734 | q = q->ipf_next; |
---|
735 | while (q != (struct ipasfrag *)fp) { |
---|
736 | t = dtom(q); |
---|
737 | q = q->ipf_next; |
---|
738 | m_cat(m, t); |
---|
739 | } |
---|
740 | |
---|
741 | #ifdef IPDIVERT |
---|
742 | /* |
---|
743 | * Record divert port for packet, if any |
---|
744 | */ |
---|
745 | frag_divert_port = fp->ipq_divert; |
---|
746 | #endif |
---|
747 | |
---|
748 | /* |
---|
749 | * Create header for new ip packet by |
---|
750 | * modifying header of first packet; |
---|
751 | * dequeue and discard fragment reassembly header. |
---|
752 | * Make header visible. |
---|
753 | */ |
---|
754 | ip = fp->ipq_next; |
---|
755 | ip->ip_len = next; |
---|
756 | ip->ipf_mff &= ~1; |
---|
757 | ((struct ip *)ip)->ip_src = fp->ipq_src; |
---|
758 | ((struct ip *)ip)->ip_dst = fp->ipq_dst; |
---|
759 | remque(fp); |
---|
760 | nipq--; |
---|
761 | (void) m_free(dtom(fp)); |
---|
762 | m = dtom(ip); |
---|
763 | m->m_len += (ip->ip_hl << 2); |
---|
764 | m->m_data -= (ip->ip_hl << 2); |
---|
765 | /* some debugging cruft by sklower, below, will go away soon */ |
---|
766 | if (m->m_flags & M_PKTHDR) { /* XXX this should be done elsewhere */ |
---|
767 | register int plen = 0; |
---|
768 | for (t = m; m; m = m->m_next) |
---|
769 | plen += m->m_len; |
---|
770 | t->m_pkthdr.len = plen; |
---|
771 | } |
---|
772 | return ((struct ip *)ip); |
---|
773 | |
---|
774 | dropfrag: |
---|
775 | ipstat.ips_fragdropped++; |
---|
776 | m_freem(m); |
---|
777 | return (0); |
---|
778 | } |
---|
779 | |
---|
780 | /* |
---|
781 | * Free a fragment reassembly header and all |
---|
782 | * associated datagrams. |
---|
783 | */ |
---|
784 | static void |
---|
785 | ip_freef(fp) |
---|
786 | struct ipq *fp; |
---|
787 | { |
---|
788 | register struct ipasfrag *q, *p; |
---|
789 | |
---|
790 | for (q = fp->ipq_next; q != (struct ipasfrag *)fp; q = p) { |
---|
791 | p = q->ipf_next; |
---|
792 | ip_deq(q); |
---|
793 | m_freem(dtom(q)); |
---|
794 | } |
---|
795 | remque(fp); |
---|
796 | (void) m_free(dtom(fp)); |
---|
797 | nipq--; |
---|
798 | } |
---|
799 | |
---|
800 | /* |
---|
801 | * Put an ip fragment on a reassembly chain. |
---|
802 | * Like insque, but pointers in middle of structure. |
---|
803 | */ |
---|
804 | static void |
---|
805 | ip_enq(p, prev) |
---|
806 | register struct ipasfrag *p, *prev; |
---|
807 | { |
---|
808 | |
---|
809 | p->ipf_prev = prev; |
---|
810 | p->ipf_next = prev->ipf_next; |
---|
811 | prev->ipf_next->ipf_prev = p; |
---|
812 | prev->ipf_next = p; |
---|
813 | } |
---|
814 | |
---|
815 | /* |
---|
816 | * To ip_enq as remque is to insque. |
---|
817 | */ |
---|
818 | static void |
---|
819 | ip_deq(p) |
---|
820 | register struct ipasfrag *p; |
---|
821 | { |
---|
822 | |
---|
823 | p->ipf_prev->ipf_next = p->ipf_next; |
---|
824 | p->ipf_next->ipf_prev = p->ipf_prev; |
---|
825 | } |
---|
826 | |
---|
827 | /* |
---|
828 | * IP timer processing; |
---|
829 | * if a timer expires on a reassembly |
---|
830 | * queue, discard it. |
---|
831 | */ |
---|
832 | void |
---|
833 | ip_slowtimo() |
---|
834 | { |
---|
835 | register struct ipq *fp; |
---|
836 | int s = splnet(); |
---|
837 | int i; |
---|
838 | |
---|
839 | for (i = 0; i < IPREASS_NHASH; i++) { |
---|
840 | fp = ipq[i].next; |
---|
841 | if (fp == 0) |
---|
842 | continue; |
---|
843 | while (fp != &ipq[i]) { |
---|
844 | --fp->ipq_ttl; |
---|
845 | fp = fp->next; |
---|
846 | if (fp->prev->ipq_ttl == 0) { |
---|
847 | ipstat.ips_fragtimeout++; |
---|
848 | ip_freef(fp->prev); |
---|
849 | } |
---|
850 | } |
---|
851 | } |
---|
852 | splx(s); |
---|
853 | } |
---|
854 | |
---|
855 | /* |
---|
856 | * Drain off all datagram fragments. |
---|
857 | */ |
---|
858 | void |
---|
859 | ip_drain() |
---|
860 | { |
---|
861 | int i; |
---|
862 | |
---|
863 | for (i = 0; i < IPREASS_NHASH; i++) { |
---|
864 | while (ipq[i].next != &ipq[i]) { |
---|
865 | ipstat.ips_fragdropped++; |
---|
866 | ip_freef(ipq[i].next); |
---|
867 | } |
---|
868 | } |
---|
869 | in_rtqdrain(); |
---|
870 | } |
---|
871 | |
---|
872 | /* |
---|
873 | * Do option processing on a datagram, |
---|
874 | * possibly discarding it if bad options are encountered, |
---|
875 | * or forwarding it if source-routed. |
---|
876 | * Returns 1 if packet has been forwarded/freed, |
---|
877 | * 0 if the packet should be processed further. |
---|
878 | */ |
---|
879 | static int |
---|
880 | ip_dooptions(m) |
---|
881 | struct mbuf *m; |
---|
882 | { |
---|
883 | register struct ip *ip = mtod(m, struct ip *); |
---|
884 | register u_char *cp; |
---|
885 | register struct ip_timestamp *ipt; |
---|
886 | register struct in_ifaddr *ia; |
---|
887 | int opt, optlen, cnt, off, code, type = ICMP_PARAMPROB, forward = 0; |
---|
888 | struct in_addr *sin, dst; |
---|
889 | n_time ntime; |
---|
890 | |
---|
891 | dst = ip->ip_dst; |
---|
892 | cp = (u_char *)(ip + 1); |
---|
893 | cnt = (IP_VHL_HL(ip->ip_vhl) << 2) - sizeof (struct ip); |
---|
894 | for (; cnt > 0; cnt -= optlen, cp += optlen) { |
---|
895 | opt = cp[IPOPT_OPTVAL]; |
---|
896 | if (opt == IPOPT_EOL) |
---|
897 | break; |
---|
898 | if (opt == IPOPT_NOP) |
---|
899 | optlen = 1; |
---|
900 | else { |
---|
901 | optlen = cp[IPOPT_OLEN]; |
---|
902 | if (optlen <= 0 || optlen > cnt) { |
---|
903 | code = &cp[IPOPT_OLEN] - (u_char *)ip; |
---|
904 | goto bad; |
---|
905 | } |
---|
906 | } |
---|
907 | switch (opt) { |
---|
908 | |
---|
909 | default: |
---|
910 | break; |
---|
911 | |
---|
912 | /* |
---|
913 | * Source routing with record. |
---|
914 | * Find interface with current destination address. |
---|
915 | * If none on this machine then drop if strictly routed, |
---|
916 | * or do nothing if loosely routed. |
---|
917 | * Record interface address and bring up next address |
---|
918 | * component. If strictly routed make sure next |
---|
919 | * address is on directly accessible net. |
---|
920 | */ |
---|
921 | case IPOPT_LSRR: |
---|
922 | case IPOPT_SSRR: |
---|
923 | if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) { |
---|
924 | code = &cp[IPOPT_OFFSET] - (u_char *)ip; |
---|
925 | goto bad; |
---|
926 | } |
---|
927 | ipaddr.sin_addr = ip->ip_dst; |
---|
928 | ia = (struct in_ifaddr *) |
---|
929 | ifa_ifwithaddr((struct sockaddr *)&ipaddr); |
---|
930 | if (ia == 0) { |
---|
931 | if (opt == IPOPT_SSRR) { |
---|
932 | type = ICMP_UNREACH; |
---|
933 | code = ICMP_UNREACH_SRCFAIL; |
---|
934 | goto bad; |
---|
935 | } |
---|
936 | if (!ip_dosourceroute) |
---|
937 | goto nosourcerouting; |
---|
938 | /* |
---|
939 | * Loose routing, and not at next destination |
---|
940 | * yet; nothing to do except forward. |
---|
941 | */ |
---|
942 | break; |
---|
943 | } |
---|
944 | off--; /* 0 origin */ |
---|
945 | if (off > optlen - sizeof(struct in_addr)) { |
---|
946 | /* |
---|
947 | * End of source route. Should be for us. |
---|
948 | */ |
---|
949 | if (!ip_acceptsourceroute) |
---|
950 | goto nosourcerouting; |
---|
951 | save_rte(cp, ip->ip_src); |
---|
952 | break; |
---|
953 | } |
---|
954 | |
---|
955 | if (!ip_dosourceroute) { |
---|
956 | char buf[4*sizeof "123"]; |
---|
957 | |
---|
958 | nosourcerouting: |
---|
959 | strcpy(buf, inet_ntoa(ip->ip_dst)); |
---|
960 | log(LOG_WARNING, |
---|
961 | "attempted source route from %s to %s\n", |
---|
962 | inet_ntoa(ip->ip_src), buf); |
---|
963 | type = ICMP_UNREACH; |
---|
964 | code = ICMP_UNREACH_SRCFAIL; |
---|
965 | goto bad; |
---|
966 | } |
---|
967 | |
---|
968 | /* |
---|
969 | * locate outgoing interface |
---|
970 | */ |
---|
971 | (void)memcpy(&ipaddr.sin_addr, cp + off, |
---|
972 | sizeof(ipaddr.sin_addr)); |
---|
973 | |
---|
974 | if (opt == IPOPT_SSRR) { |
---|
975 | #define INA struct in_ifaddr * |
---|
976 | #define SA struct sockaddr * |
---|
977 | if ((ia = (INA)ifa_ifwithdstaddr((SA)&ipaddr)) == 0) |
---|
978 | ia = (INA)ifa_ifwithnet((SA)&ipaddr); |
---|
979 | } else |
---|
980 | ia = ip_rtaddr(ipaddr.sin_addr); |
---|
981 | if (ia == 0) { |
---|
982 | type = ICMP_UNREACH; |
---|
983 | code = ICMP_UNREACH_SRCFAIL; |
---|
984 | goto bad; |
---|
985 | } |
---|
986 | ip->ip_dst = ipaddr.sin_addr; |
---|
987 | (void)memcpy(cp + off, &(IA_SIN(ia)->sin_addr), |
---|
988 | sizeof(struct in_addr)); |
---|
989 | cp[IPOPT_OFFSET] += sizeof(struct in_addr); |
---|
990 | /* |
---|
991 | * Let ip_intr's mcast routing check handle mcast pkts |
---|
992 | */ |
---|
993 | forward = !IN_MULTICAST(ntohl(ip->ip_dst.s_addr)); |
---|
994 | break; |
---|
995 | |
---|
996 | case IPOPT_RR: |
---|
997 | if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) { |
---|
998 | code = &cp[IPOPT_OFFSET] - (u_char *)ip; |
---|
999 | goto bad; |
---|
1000 | } |
---|
1001 | /* |
---|
1002 | * If no space remains, ignore. |
---|
1003 | */ |
---|
1004 | off--; /* 0 origin */ |
---|
1005 | if (off > optlen - sizeof(struct in_addr)) |
---|
1006 | break; |
---|
1007 | (void)memcpy(&ipaddr.sin_addr, &ip->ip_dst, |
---|
1008 | sizeof(ipaddr.sin_addr)); |
---|
1009 | /* |
---|
1010 | * locate outgoing interface; if we're the destination, |
---|
1011 | * use the incoming interface (should be same). |
---|
1012 | */ |
---|
1013 | if ((ia = (INA)ifa_ifwithaddr((SA)&ipaddr)) == 0 && |
---|
1014 | (ia = ip_rtaddr(ipaddr.sin_addr)) == 0) { |
---|
1015 | type = ICMP_UNREACH; |
---|
1016 | code = ICMP_UNREACH_HOST; |
---|
1017 | goto bad; |
---|
1018 | } |
---|
1019 | (void)memcpy(cp + off, &(IA_SIN(ia)->sin_addr), |
---|
1020 | sizeof(struct in_addr)); |
---|
1021 | cp[IPOPT_OFFSET] += sizeof(struct in_addr); |
---|
1022 | break; |
---|
1023 | |
---|
1024 | case IPOPT_TS: |
---|
1025 | code = cp - (u_char *)ip; |
---|
1026 | ipt = (struct ip_timestamp *)cp; |
---|
1027 | if (ipt->ipt_len < 5) |
---|
1028 | goto bad; |
---|
1029 | if (ipt->ipt_ptr > ipt->ipt_len - sizeof (long)) { |
---|
1030 | if (++ipt->ipt_oflw == 0) |
---|
1031 | goto bad; |
---|
1032 | break; |
---|
1033 | } |
---|
1034 | sin = (struct in_addr *)(cp + ipt->ipt_ptr - 1); |
---|
1035 | switch (ipt->ipt_flg) { |
---|
1036 | |
---|
1037 | case IPOPT_TS_TSONLY: |
---|
1038 | break; |
---|
1039 | |
---|
1040 | case IPOPT_TS_TSANDADDR: |
---|
1041 | if (ipt->ipt_ptr + sizeof(n_time) + |
---|
1042 | sizeof(struct in_addr) > ipt->ipt_len) |
---|
1043 | goto bad; |
---|
1044 | ipaddr.sin_addr = dst; |
---|
1045 | ia = (INA)ifaof_ifpforaddr((SA)&ipaddr, |
---|
1046 | m->m_pkthdr.rcvif); |
---|
1047 | if (ia == 0) |
---|
1048 | continue; |
---|
1049 | (void)memcpy(sin, &IA_SIN(ia)->sin_addr, |
---|
1050 | sizeof(struct in_addr)); |
---|
1051 | ipt->ipt_ptr += sizeof(struct in_addr); |
---|
1052 | break; |
---|
1053 | |
---|
1054 | case IPOPT_TS_PRESPEC: |
---|
1055 | if (ipt->ipt_ptr + sizeof(n_time) + |
---|
1056 | sizeof(struct in_addr) > ipt->ipt_len) |
---|
1057 | goto bad; |
---|
1058 | (void)memcpy(&ipaddr.sin_addr, sin, |
---|
1059 | sizeof(struct in_addr)); |
---|
1060 | if (ifa_ifwithaddr((SA)&ipaddr) == 0) |
---|
1061 | continue; |
---|
1062 | ipt->ipt_ptr += sizeof(struct in_addr); |
---|
1063 | break; |
---|
1064 | |
---|
1065 | default: |
---|
1066 | goto bad; |
---|
1067 | } |
---|
1068 | ntime = iptime(); |
---|
1069 | (void)memcpy(cp + ipt->ipt_ptr - 1, &ntime, |
---|
1070 | sizeof(n_time)); |
---|
1071 | ipt->ipt_ptr += sizeof(n_time); |
---|
1072 | } |
---|
1073 | } |
---|
1074 | if (forward && ipforwarding) { |
---|
1075 | ip_forward(m, 1); |
---|
1076 | return (1); |
---|
1077 | } |
---|
1078 | return (0); |
---|
1079 | bad: |
---|
1080 | ip->ip_len -= IP_VHL_HL(ip->ip_vhl) << 2; /* XXX icmp_error adds in hdr length */ |
---|
1081 | icmp_error(m, type, code, 0, 0); |
---|
1082 | ipstat.ips_badoptions++; |
---|
1083 | return (1); |
---|
1084 | } |
---|
1085 | |
---|
1086 | /* |
---|
1087 | * Given address of next destination (final or next hop), |
---|
1088 | * return internet address info of interface to be used to get there. |
---|
1089 | */ |
---|
1090 | static struct in_ifaddr * |
---|
1091 | ip_rtaddr(dst) |
---|
1092 | struct in_addr dst; |
---|
1093 | { |
---|
1094 | register struct sockaddr_in *sin; |
---|
1095 | |
---|
1096 | sin = (struct sockaddr_in *) &ipforward_rt.ro_dst; |
---|
1097 | |
---|
1098 | if (ipforward_rt.ro_rt == 0 || dst.s_addr != sin->sin_addr.s_addr) { |
---|
1099 | if (ipforward_rt.ro_rt) { |
---|
1100 | RTFREE(ipforward_rt.ro_rt); |
---|
1101 | ipforward_rt.ro_rt = 0; |
---|
1102 | } |
---|
1103 | sin->sin_family = AF_INET; |
---|
1104 | sin->sin_len = sizeof(*sin); |
---|
1105 | sin->sin_addr = dst; |
---|
1106 | |
---|
1107 | rtalloc_ign(&ipforward_rt, RTF_PRCLONING); |
---|
1108 | } |
---|
1109 | if (ipforward_rt.ro_rt == 0) |
---|
1110 | return ((struct in_ifaddr *)0); |
---|
1111 | return ((struct in_ifaddr *) ipforward_rt.ro_rt->rt_ifa); |
---|
1112 | } |
---|
1113 | |
---|
1114 | /* |
---|
1115 | * Save incoming source route for use in replies, |
---|
1116 | * to be picked up later by ip_srcroute if the receiver is interested. |
---|
1117 | */ |
---|
1118 | void |
---|
1119 | save_rte(option, dst) |
---|
1120 | u_char *option; |
---|
1121 | struct in_addr dst; |
---|
1122 | { |
---|
1123 | unsigned olen; |
---|
1124 | |
---|
1125 | olen = option[IPOPT_OLEN]; |
---|
1126 | #ifdef DIAGNOSTIC |
---|
1127 | if (ipprintfs) |
---|
1128 | printf("save_rte: olen %d\n", olen); |
---|
1129 | #endif |
---|
1130 | if (olen > sizeof(ip_srcrt) - (1 + sizeof(dst))) |
---|
1131 | return; |
---|
1132 | bcopy(option, ip_srcrt.srcopt, olen); |
---|
1133 | ip_nhops = (olen - IPOPT_OFFSET - 1) / sizeof(struct in_addr); |
---|
1134 | ip_srcrt.dst = dst; |
---|
1135 | } |
---|
1136 | |
---|
1137 | /* |
---|
1138 | * Retrieve incoming source route for use in replies, |
---|
1139 | * in the same form used by setsockopt. |
---|
1140 | * The first hop is placed before the options, will be removed later. |
---|
1141 | */ |
---|
1142 | struct mbuf * |
---|
1143 | ip_srcroute() |
---|
1144 | { |
---|
1145 | register struct in_addr *p, *q; |
---|
1146 | register struct mbuf *m; |
---|
1147 | |
---|
1148 | if (ip_nhops == 0) |
---|
1149 | return ((struct mbuf *)0); |
---|
1150 | m = m_get(M_DONTWAIT, MT_SOOPTS); |
---|
1151 | if (m == 0) |
---|
1152 | return ((struct mbuf *)0); |
---|
1153 | |
---|
1154 | #define OPTSIZ (sizeof(ip_srcrt.nop) + sizeof(ip_srcrt.srcopt)) |
---|
1155 | |
---|
1156 | /* length is (nhops+1)*sizeof(addr) + sizeof(nop + srcrt header) */ |
---|
1157 | m->m_len = ip_nhops * sizeof(struct in_addr) + sizeof(struct in_addr) + |
---|
1158 | OPTSIZ; |
---|
1159 | #ifdef DIAGNOSTIC |
---|
1160 | if (ipprintfs) |
---|
1161 | printf("ip_srcroute: nhops %d mlen %d", ip_nhops, m->m_len); |
---|
1162 | #endif |
---|
1163 | |
---|
1164 | /* |
---|
1165 | * First save first hop for return route |
---|
1166 | */ |
---|
1167 | p = &ip_srcrt.route[ip_nhops - 1]; |
---|
1168 | *(mtod(m, struct in_addr *)) = *p--; |
---|
1169 | #ifdef DIAGNOSTIC |
---|
1170 | if (ipprintfs) |
---|
1171 | printf(" hops %lx", ntohl(mtod(m, struct in_addr *)->s_addr)); |
---|
1172 | #endif |
---|
1173 | |
---|
1174 | /* |
---|
1175 | * Copy option fields and padding (nop) to mbuf. |
---|
1176 | */ |
---|
1177 | ip_srcrt.nop = IPOPT_NOP; |
---|
1178 | ip_srcrt.srcopt[IPOPT_OFFSET] = IPOPT_MINOFF; |
---|
1179 | (void)memcpy(mtod(m, caddr_t) + sizeof(struct in_addr), |
---|
1180 | &ip_srcrt.nop, OPTSIZ); |
---|
1181 | q = (struct in_addr *)(mtod(m, caddr_t) + |
---|
1182 | sizeof(struct in_addr) + OPTSIZ); |
---|
1183 | #undef OPTSIZ |
---|
1184 | /* |
---|
1185 | * Record return path as an IP source route, |
---|
1186 | * reversing the path (pointers are now aligned). |
---|
1187 | */ |
---|
1188 | while (p >= ip_srcrt.route) { |
---|
1189 | #ifdef DIAGNOSTIC |
---|
1190 | if (ipprintfs) |
---|
1191 | printf(" %lx", ntohl(q->s_addr)); |
---|
1192 | #endif |
---|
1193 | *q++ = *p--; |
---|
1194 | } |
---|
1195 | /* |
---|
1196 | * Last hop goes to final destination. |
---|
1197 | */ |
---|
1198 | *q = ip_srcrt.dst; |
---|
1199 | #ifdef DIAGNOSTIC |
---|
1200 | if (ipprintfs) |
---|
1201 | printf(" %lx\n", ntohl(q->s_addr)); |
---|
1202 | #endif |
---|
1203 | return (m); |
---|
1204 | } |
---|
1205 | |
---|
1206 | /* |
---|
1207 | * Strip out IP options, at higher |
---|
1208 | * level protocol in the kernel. |
---|
1209 | * Second argument is buffer to which options |
---|
1210 | * will be moved, and return value is their length. |
---|
1211 | * XXX should be deleted; last arg currently ignored. |
---|
1212 | */ |
---|
1213 | void |
---|
1214 | ip_stripoptions(m, mopt) |
---|
1215 | register struct mbuf *m; |
---|
1216 | struct mbuf *mopt; |
---|
1217 | { |
---|
1218 | register int i; |
---|
1219 | struct ip *ip = mtod(m, struct ip *); |
---|
1220 | register caddr_t opts; |
---|
1221 | int olen; |
---|
1222 | |
---|
1223 | olen = (IP_VHL_HL(ip->ip_vhl) << 2) - sizeof (struct ip); |
---|
1224 | opts = (caddr_t)(ip + 1); |
---|
1225 | i = m->m_len - (sizeof (struct ip) + olen); |
---|
1226 | bcopy(opts + olen, opts, (unsigned)i); |
---|
1227 | m->m_len -= olen; |
---|
1228 | if (m->m_flags & M_PKTHDR) |
---|
1229 | m->m_pkthdr.len -= olen; |
---|
1230 | ip->ip_vhl = IP_MAKE_VHL(IPVERSION, sizeof(struct ip) >> 2); |
---|
1231 | } |
---|
1232 | |
---|
1233 | u_char inetctlerrmap[PRC_NCMDS] = { |
---|
1234 | 0, 0, 0, 0, |
---|
1235 | 0, EMSGSIZE, EHOSTDOWN, EHOSTUNREACH, |
---|
1236 | EHOSTUNREACH, EHOSTUNREACH, ECONNREFUSED, ECONNREFUSED, |
---|
1237 | EMSGSIZE, EHOSTUNREACH, 0, 0, |
---|
1238 | 0, 0, 0, 0, |
---|
1239 | ENOPROTOOPT |
---|
1240 | }; |
---|
1241 | |
---|
1242 | /* |
---|
1243 | * Forward a packet. If some error occurs return the sender |
---|
1244 | * an icmp packet. Note we can't always generate a meaningful |
---|
1245 | * icmp message because icmp doesn't have a large enough repertoire |
---|
1246 | * of codes and types. |
---|
1247 | * |
---|
1248 | * If not forwarding, just drop the packet. This could be confusing |
---|
1249 | * if ipforwarding was zero but some routing protocol was advancing |
---|
1250 | * us as a gateway to somewhere. However, we must let the routing |
---|
1251 | * protocol deal with that. |
---|
1252 | * |
---|
1253 | * The srcrt parameter indicates whether the packet is being forwarded |
---|
1254 | * via a source route. |
---|
1255 | */ |
---|
1256 | static void |
---|
1257 | ip_forward(m, srcrt) |
---|
1258 | struct mbuf *m; |
---|
1259 | int srcrt; |
---|
1260 | { |
---|
1261 | register struct ip *ip = mtod(m, struct ip *); |
---|
1262 | register struct sockaddr_in *sin; |
---|
1263 | register struct rtentry *rt; |
---|
1264 | int error, type = 0, code = 0; |
---|
1265 | struct mbuf *mcopy; |
---|
1266 | n_long dest; |
---|
1267 | struct ifnet *destifp; |
---|
1268 | |
---|
1269 | dest = 0; |
---|
1270 | #ifdef DIAGNOSTIC |
---|
1271 | if (ipprintfs) |
---|
1272 | printf("forward: src %lx dst %lx ttl %x\n", |
---|
1273 | ip->ip_src.s_addr, ip->ip_dst.s_addr, ip->ip_ttl); |
---|
1274 | #endif |
---|
1275 | |
---|
1276 | |
---|
1277 | if (m->m_flags & M_BCAST || in_canforward(ip->ip_dst) == 0) { |
---|
1278 | ipstat.ips_cantforward++; |
---|
1279 | m_freem(m); |
---|
1280 | return; |
---|
1281 | } |
---|
1282 | HTONS(ip->ip_id); |
---|
1283 | if (ip->ip_ttl <= IPTTLDEC) { |
---|
1284 | icmp_error(m, ICMP_TIMXCEED, ICMP_TIMXCEED_INTRANS, dest, 0); |
---|
1285 | return; |
---|
1286 | } |
---|
1287 | ip->ip_ttl -= IPTTLDEC; |
---|
1288 | |
---|
1289 | sin = (struct sockaddr_in *)&ipforward_rt.ro_dst; |
---|
1290 | if ((rt = ipforward_rt.ro_rt) == 0 || |
---|
1291 | ip->ip_dst.s_addr != sin->sin_addr.s_addr) { |
---|
1292 | if (ipforward_rt.ro_rt) { |
---|
1293 | RTFREE(ipforward_rt.ro_rt); |
---|
1294 | ipforward_rt.ro_rt = 0; |
---|
1295 | } |
---|
1296 | sin->sin_family = AF_INET; |
---|
1297 | sin->sin_len = sizeof(*sin); |
---|
1298 | sin->sin_addr = ip->ip_dst; |
---|
1299 | |
---|
1300 | rtalloc_ign(&ipforward_rt, RTF_PRCLONING); |
---|
1301 | if (ipforward_rt.ro_rt == 0) { |
---|
1302 | icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_HOST, dest, 0); |
---|
1303 | return; |
---|
1304 | } |
---|
1305 | rt = ipforward_rt.ro_rt; |
---|
1306 | } |
---|
1307 | |
---|
1308 | /* |
---|
1309 | * Save at most 64 bytes of the packet in case |
---|
1310 | * we need to generate an ICMP message to the src. |
---|
1311 | */ |
---|
1312 | mcopy = m_copy(m, 0, imin((int)ip->ip_len, 64)); |
---|
1313 | |
---|
1314 | /* |
---|
1315 | * If forwarding packet using same interface that it came in on, |
---|
1316 | * perhaps should send a redirect to sender to shortcut a hop. |
---|
1317 | * Only send redirect if source is sending directly to us, |
---|
1318 | * and if packet was not source routed (or has any options). |
---|
1319 | * Also, don't send redirect if forwarding using a default route |
---|
1320 | * or a route modified by a redirect. |
---|
1321 | */ |
---|
1322 | #define satosin(sa) ((struct sockaddr_in *)(sa)) |
---|
1323 | if (rt->rt_ifp == m->m_pkthdr.rcvif && |
---|
1324 | (rt->rt_flags & (RTF_DYNAMIC|RTF_MODIFIED)) == 0 && |
---|
1325 | satosin(rt_key(rt))->sin_addr.s_addr != 0 && |
---|
1326 | ipsendredirects && !srcrt) { |
---|
1327 | #define RTA(rt) ((struct in_ifaddr *)(rt->rt_ifa)) |
---|
1328 | u_long src = ntohl(ip->ip_src.s_addr); |
---|
1329 | |
---|
1330 | if (RTA(rt) && |
---|
1331 | (src & RTA(rt)->ia_subnetmask) == RTA(rt)->ia_subnet) { |
---|
1332 | if (rt->rt_flags & RTF_GATEWAY) |
---|
1333 | dest = satosin(rt->rt_gateway)->sin_addr.s_addr; |
---|
1334 | else |
---|
1335 | dest = ip->ip_dst.s_addr; |
---|
1336 | /* Router requirements says to only send host redirects */ |
---|
1337 | type = ICMP_REDIRECT; |
---|
1338 | code = ICMP_REDIRECT_HOST; |
---|
1339 | #ifdef DIAGNOSTIC |
---|
1340 | if (ipprintfs) |
---|
1341 | printf("redirect (%d) to %lx\n", code, (u_long)dest); |
---|
1342 | #endif |
---|
1343 | } |
---|
1344 | } |
---|
1345 | |
---|
1346 | error = ip_output(m, (struct mbuf *)0, &ipforward_rt, |
---|
1347 | IP_FORWARDING, 0); |
---|
1348 | if (error) |
---|
1349 | ipstat.ips_cantforward++; |
---|
1350 | else { |
---|
1351 | ipstat.ips_forward++; |
---|
1352 | if (type) |
---|
1353 | ipstat.ips_redirectsent++; |
---|
1354 | else { |
---|
1355 | if (mcopy) |
---|
1356 | m_freem(mcopy); |
---|
1357 | return; |
---|
1358 | } |
---|
1359 | } |
---|
1360 | if (mcopy == NULL) |
---|
1361 | return; |
---|
1362 | destifp = NULL; |
---|
1363 | |
---|
1364 | switch (error) { |
---|
1365 | |
---|
1366 | case 0: /* forwarded, but need redirect */ |
---|
1367 | /* type, code set above */ |
---|
1368 | break; |
---|
1369 | |
---|
1370 | case ENETUNREACH: /* shouldn't happen, checked above */ |
---|
1371 | case EHOSTUNREACH: |
---|
1372 | case ENETDOWN: |
---|
1373 | case EHOSTDOWN: |
---|
1374 | default: |
---|
1375 | type = ICMP_UNREACH; |
---|
1376 | code = ICMP_UNREACH_HOST; |
---|
1377 | break; |
---|
1378 | |
---|
1379 | case EMSGSIZE: |
---|
1380 | type = ICMP_UNREACH; |
---|
1381 | code = ICMP_UNREACH_NEEDFRAG; |
---|
1382 | if (ipforward_rt.ro_rt) |
---|
1383 | destifp = ipforward_rt.ro_rt->rt_ifp; |
---|
1384 | ipstat.ips_cantfrag++; |
---|
1385 | break; |
---|
1386 | |
---|
1387 | case ENOBUFS: |
---|
1388 | type = ICMP_SOURCEQUENCH; |
---|
1389 | code = 0; |
---|
1390 | break; |
---|
1391 | } |
---|
1392 | icmp_error(mcopy, type, code, dest, destifp); |
---|
1393 | } |
---|
1394 | |
---|
1395 | void |
---|
1396 | ip_savecontrol(inp, mp, ip, m) |
---|
1397 | register struct inpcb *inp; |
---|
1398 | register struct mbuf **mp; |
---|
1399 | register struct ip *ip; |
---|
1400 | register struct mbuf *m; |
---|
1401 | { |
---|
1402 | if (inp->inp_socket->so_options & SO_TIMESTAMP) { |
---|
1403 | struct timeval tv; |
---|
1404 | |
---|
1405 | microtime(&tv); |
---|
1406 | *mp = sbcreatecontrol((caddr_t) &tv, sizeof(tv), |
---|
1407 | SCM_TIMESTAMP, SOL_SOCKET); |
---|
1408 | if (*mp) |
---|
1409 | mp = &(*mp)->m_next; |
---|
1410 | } |
---|
1411 | if (inp->inp_flags & INP_RECVDSTADDR) { |
---|
1412 | *mp = sbcreatecontrol((caddr_t) &ip->ip_dst, |
---|
1413 | sizeof(struct in_addr), IP_RECVDSTADDR, IPPROTO_IP); |
---|
1414 | if (*mp) |
---|
1415 | mp = &(*mp)->m_next; |
---|
1416 | } |
---|
1417 | #ifdef notyet |
---|
1418 | /* XXX |
---|
1419 | * Moving these out of udp_input() made them even more broken |
---|
1420 | * than they already were. |
---|
1421 | */ |
---|
1422 | /* options were tossed already */ |
---|
1423 | if (inp->inp_flags & INP_RECVOPTS) { |
---|
1424 | *mp = sbcreatecontrol((caddr_t) opts_deleted_above, |
---|
1425 | sizeof(struct in_addr), IP_RECVOPTS, IPPROTO_IP); |
---|
1426 | if (*mp) |
---|
1427 | mp = &(*mp)->m_next; |
---|
1428 | } |
---|
1429 | /* ip_srcroute doesn't do what we want here, need to fix */ |
---|
1430 | if (inp->inp_flags & INP_RECVRETOPTS) { |
---|
1431 | *mp = sbcreatecontrol((caddr_t) ip_srcroute(), |
---|
1432 | sizeof(struct in_addr), IP_RECVRETOPTS, IPPROTO_IP); |
---|
1433 | if (*mp) |
---|
1434 | mp = &(*mp)->m_next; |
---|
1435 | } |
---|
1436 | #endif |
---|
1437 | if (inp->inp_flags & INP_RECVIF) { |
---|
1438 | struct sockaddr_dl sdl; |
---|
1439 | |
---|
1440 | sdl.sdl_len = offsetof(struct sockaddr_dl, sdl_data[0]); |
---|
1441 | sdl.sdl_family = AF_LINK; |
---|
1442 | sdl.sdl_index = m->m_pkthdr.rcvif ? |
---|
1443 | m->m_pkthdr.rcvif->if_index : 0; |
---|
1444 | sdl.sdl_nlen = sdl.sdl_alen = sdl.sdl_slen = 0; |
---|
1445 | *mp = sbcreatecontrol((caddr_t) &sdl, sdl.sdl_len, |
---|
1446 | IP_RECVIF, IPPROTO_IP); |
---|
1447 | if (*mp) |
---|
1448 | mp = &(*mp)->m_next; |
---|
1449 | } |
---|
1450 | } |
---|
1451 | |
---|
1452 | int |
---|
1453 | ip_rsvp_init(struct socket *so) |
---|
1454 | { |
---|
1455 | if (so->so_type != SOCK_RAW || |
---|
1456 | so->so_proto->pr_protocol != IPPROTO_RSVP) |
---|
1457 | return EOPNOTSUPP; |
---|
1458 | |
---|
1459 | if (ip_rsvpd != NULL) |
---|
1460 | return EADDRINUSE; |
---|
1461 | |
---|
1462 | ip_rsvpd = so; |
---|
1463 | /* |
---|
1464 | * This may seem silly, but we need to be sure we don't over-increment |
---|
1465 | * the RSVP counter, in case something slips up. |
---|
1466 | */ |
---|
1467 | if (!ip_rsvp_on) { |
---|
1468 | ip_rsvp_on = 1; |
---|
1469 | rsvp_on++; |
---|
1470 | } |
---|
1471 | |
---|
1472 | return 0; |
---|
1473 | } |
---|
1474 | |
---|
1475 | int |
---|
1476 | ip_rsvp_done(void) |
---|
1477 | { |
---|
1478 | ip_rsvpd = NULL; |
---|
1479 | /* |
---|
1480 | * This may seem silly, but we need to be sure we don't over-decrement |
---|
1481 | * the RSVP counter, in case something slips up. |
---|
1482 | */ |
---|
1483 | if (ip_rsvp_on) { |
---|
1484 | ip_rsvp_on = 0; |
---|
1485 | rsvp_on--; |
---|
1486 | } |
---|
1487 | return 0; |
---|
1488 | } |
---|