xref: /openbsd/sys/netinet/raw_ip.c (revision 17df1aa7)
1 /*	$OpenBSD: raw_ip.c,v 1.48 2009/11/03 10:59:04 claudio Exp $	*/
2 /*	$NetBSD: raw_ip.c,v 1.25 1996/02/18 18:58:33 christos Exp $	*/
3 
4 /*
5  * Copyright (c) 1982, 1986, 1988, 1993
6  *	The Regents of the University of California.  All rights reserved.
7  *
8  * Redistribution and use in source and binary forms, with or without
9  * modification, are permitted provided that the following conditions
10  * are met:
11  * 1. Redistributions of source code must retain the above copyright
12  *    notice, this list of conditions and the following disclaimer.
13  * 2. Redistributions in binary form must reproduce the above copyright
14  *    notice, this list of conditions and the following disclaimer in the
15  *    documentation and/or other materials provided with the distribution.
16  * 3. Neither the name of the University nor the names of its contributors
17  *    may be used to endorse or promote products derived from this software
18  *    without specific prior written permission.
19  *
20  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
21  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
22  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
23  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
24  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
25  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
26  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
28  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
29  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
30  * SUCH DAMAGE.
31  *
32  *	@(#)COPYRIGHT	1.1 (NRL) 17 January 1995
33  *
34  * NRL grants permission for redistribution and use in source and binary
35  * forms, with or without modification, of the software and documentation
36  * created at NRL provided that the following conditions are met:
37  *
38  * 1. Redistributions of source code must retain the above copyright
39  *    notice, this list of conditions and the following disclaimer.
40  * 2. Redistributions in binary form must reproduce the above copyright
41  *    notice, this list of conditions and the following disclaimer in the
42  *    documentation and/or other materials provided with the distribution.
43  * 3. All advertising materials mentioning features or use of this software
44  *    must display the following acknowledgements:
45  * 	This product includes software developed by the University of
46  * 	California, Berkeley and its contributors.
47  * 	This product includes software developed at the Information
48  * 	Technology Division, US Naval Research Laboratory.
49  * 4. Neither the name of the NRL nor the names of its contributors
50  *    may be used to endorse or promote products derived from this software
51  *    without specific prior written permission.
52  *
53  * THE SOFTWARE PROVIDED BY NRL IS PROVIDED BY NRL AND CONTRIBUTORS ``AS
54  * IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
55  * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
56  * PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL NRL OR
57  * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
58  * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
59  * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
60  * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
61  * LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
62  * NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
63  * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
64  *
65  * The views and conclusions contained in the software and documentation
66  * are those of the authors and should not be interpreted as representing
67  * official policies, either expressed or implied, of the US Naval
68  * Research Laboratory (NRL).
69  */
70 
71 #include <sys/param.h>
72 #include <sys/systm.h>
73 #include <sys/mbuf.h>
74 #include <sys/socket.h>
75 #include <sys/protosw.h>
76 #include <sys/socketvar.h>
77 
78 #include <net/if.h>
79 #include <net/route.h>
80 #include <net/pfvar.h>
81 
82 #include <netinet/in.h>
83 #include <netinet/in_systm.h>
84 #include <netinet/ip.h>
85 #include <netinet/ip_mroute.h>
86 #include <netinet/ip_var.h>
87 #include <netinet/in_pcb.h>
88 #include <netinet/in_var.h>
89 #include <netinet/ip_icmp.h>
90 
91 #include "pf.h"
92 
93 struct inpcbtable rawcbtable;
94 
95 /*
96  * Nominal space allocated to a raw ip socket.
97  */
98 #define	RIPSNDQ		8192
99 #define	RIPRCVQ		8192
100 
101 /*
102  * Raw interface to IP protocol.
103  */
104 
105 /*
106  * Initialize raw connection block q.
107  */
108 void
109 rip_init()
110 {
111 
112 	in_pcbinit(&rawcbtable, 1);
113 }
114 
115 struct sockaddr_in ripsrc = { sizeof(ripsrc), AF_INET };
116 
117 /*
118  * Setup generic address and protocol structures
119  * for raw_input routine, then pass them along with
120  * mbuf chain.
121  */
122 void
123 rip_input(struct mbuf *m, ...)
124 {
125 	struct ip *ip = mtod(m, struct ip *);
126 	struct inpcb *inp, *last = NULL;
127 	struct mbuf *opts = NULL;
128 
129 	ripsrc.sin_addr = ip->ip_src;
130 	CIRCLEQ_FOREACH(inp, &rawcbtable.inpt_queue, inp_queue) {
131 #ifdef INET6
132 		if (inp->inp_flags & INP_IPV6)
133 			continue;
134 #endif
135 		if (inp->inp_rdomain != rtable_l2(m->m_pkthdr.rdomain))
136 			continue;
137 
138 		if (inp->inp_ip.ip_p && inp->inp_ip.ip_p != ip->ip_p)
139 			continue;
140 #if NPF > 0
141 		if (m->m_pkthdr.pf.flags & PF_TAG_DIVERTED) {
142 			struct pf_divert *divert;
143 
144 			/* XXX rdomain support */
145 			if ((divert = pf_find_divert(m)) == NULL)
146 				continue;
147 			if (inp->inp_laddr.s_addr != divert->addr.ipv4.s_addr)
148 				continue;
149 		} else
150 #endif
151 		if (inp->inp_laddr.s_addr &&
152 		    inp->inp_laddr.s_addr != ip->ip_dst.s_addr)
153 			continue;
154 		if (inp->inp_faddr.s_addr &&
155 		    inp->inp_faddr.s_addr != ip->ip_src.s_addr)
156 			continue;
157 		if (last) {
158 			struct mbuf *n;
159 
160 			if ((n = m_copy(m, 0, (int)M_COPYALL)) != NULL) {
161 				if (last->inp_flags & INP_CONTROLOPTS ||
162 				    last->inp_socket->so_options & SO_TIMESTAMP)
163 					ip_savecontrol(last, &opts, ip, n);
164 				if (sbappendaddr(&last->inp_socket->so_rcv,
165 				    sintosa(&ripsrc), n, opts) == 0) {
166 					/* should notify about lost packet */
167 					m_freem(n);
168 					if (opts)
169 						m_freem(opts);
170 				} else
171 					sorwakeup(last->inp_socket);
172 				opts = NULL;
173 			}
174 		}
175 		last = inp;
176 	}
177 	if (last) {
178 		if (last->inp_flags & INP_CONTROLOPTS ||
179 		    last->inp_socket->so_options & SO_TIMESTAMP)
180 			ip_savecontrol(last, &opts, ip, m);
181 		if (sbappendaddr(&last->inp_socket->so_rcv, sintosa(&ripsrc), m,
182 		    opts) == 0) {
183 			m_freem(m);
184 			if (opts)
185 				m_freem(opts);
186 		} else
187 			sorwakeup(last->inp_socket);
188 	} else {
189 		if (ip->ip_p != IPPROTO_ICMP)
190 			icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_PROTOCOL, 0, 0);
191 		else
192 			m_freem(m);
193 		ipstat.ips_noproto++;
194 		ipstat.ips_delivered--;
195 	}
196 }
197 
198 /*
199  * Generate IP header and pass packet to ip_output.
200  * Tack on options user may have setup with control call.
201  */
202 int
203 rip_output(struct mbuf *m, ...)
204 {
205 	struct socket *so;
206 	u_long dst;
207 	struct ip *ip;
208 	struct inpcb *inp;
209 	int flags;
210 	va_list ap;
211 
212 	va_start(ap, m);
213 	so = va_arg(ap, struct socket *);
214 	dst = va_arg(ap, u_long);
215 	va_end(ap);
216 
217 	inp = sotoinpcb(so);
218 	flags = (so->so_options & (SO_DONTROUTE|SO_JUMBO)) | IP_ALLOWBROADCAST;
219 
220 	/*
221 	 * If the user handed us a complete IP packet, use it.
222 	 * Otherwise, allocate an mbuf for a header and fill it in.
223 	 */
224 	if ((inp->inp_flags & INP_HDRINCL) == 0) {
225 		if ((m->m_pkthdr.len + sizeof(struct ip)) > IP_MAXPACKET) {
226 			m_freem(m);
227 			return (EMSGSIZE);
228 		}
229 		M_PREPEND(m, sizeof(struct ip), M_DONTWAIT);
230 		if (!m)
231 			return (ENOBUFS);
232 		ip = mtod(m, struct ip *);
233 		ip->ip_tos = inp->inp_ip.ip_tos;
234 		ip->ip_off = htons(0);
235 		ip->ip_p = inp->inp_ip.ip_p;
236 		ip->ip_len = htons(m->m_pkthdr.len);
237 		ip->ip_src = inp->inp_laddr;
238 		ip->ip_dst.s_addr = dst;
239 		ip->ip_ttl = inp->inp_ip.ip_ttl ? inp->inp_ip.ip_ttl : MAXTTL;
240 	} else {
241 		if (m->m_pkthdr.len > IP_MAXPACKET) {
242 			m_freem(m);
243 			return (EMSGSIZE);
244 		}
245 		if (m->m_pkthdr.len < sizeof(struct ip)) {
246 			m_freem(m);
247 			return (EINVAL);
248 		}
249 		ip = mtod(m, struct ip *);
250 		/*
251 		 * don't allow both user specified and setsockopt options,
252 		 * and don't allow packet length sizes that will crash
253 		 */
254 		if ((ip->ip_hl != (sizeof (*ip) >> 2) && inp->inp_options) ||
255 		    ntohs(ip->ip_len) > m->m_pkthdr.len ||
256 		    ntohs(ip->ip_len) < ip->ip_hl << 2) {
257 			m_freem(m);
258 			return (EINVAL);
259 		}
260 		if (ip->ip_id == 0) {
261 			ip->ip_id = htons(ip_randomid());
262 		}
263 		/* XXX prevent ip_output from overwriting header fields */
264 		flags |= IP_RAWOUTPUT;
265 		ipstat.ips_rawout++;
266 	}
267 #ifdef INET6
268 	/*
269 	 * A thought:  Even though raw IP shouldn't be able to set IPv6
270 	 *             multicast options, if it does, the last parameter to
271 	 *             ip_output should be guarded against v6/v4 problems.
272 	 */
273 #endif
274 	/* force routing domain */
275 	m->m_pkthdr.rdomain = inp->inp_rdomain;
276 
277 	return (ip_output(m, inp->inp_options, &inp->inp_route, flags,
278 	    inp->inp_moptions, inp));
279 }
280 
281 /*
282  * Raw IP socket option processing.
283  */
284 int
285 rip_ctloutput(int op, struct socket *so, int level, int optname,
286     struct mbuf **m)
287 {
288 	struct inpcb *inp = sotoinpcb(so);
289 	int error;
290 
291 	if (level != IPPROTO_IP) {
292 		if (op == PRCO_SETOPT && *m)
293 			(void) m_free(*m);
294 		return (EINVAL);
295 	}
296 
297 	switch (optname) {
298 
299 	case IP_HDRINCL:
300 		error = 0;
301 		if (op == PRCO_SETOPT) {
302 			if (*m == 0 || (*m)->m_len < sizeof (int))
303 				error = EINVAL;
304 			else if (*mtod(*m, int *))
305 				inp->inp_flags |= INP_HDRINCL;
306 			else
307 				inp->inp_flags &= ~INP_HDRINCL;
308 			if (*m)
309 				(void)m_free(*m);
310 		} else {
311 			*m = m_get(M_WAIT, M_SOOPTS);
312 			(*m)->m_len = sizeof(int);
313 			*mtod(*m, int *) = inp->inp_flags & INP_HDRINCL;
314 		}
315 		return (error);
316 
317 	case MRT_INIT:
318 	case MRT_DONE:
319 	case MRT_ADD_VIF:
320 	case MRT_DEL_VIF:
321 	case MRT_ADD_MFC:
322 	case MRT_DEL_MFC:
323 	case MRT_VERSION:
324 	case MRT_ASSERT:
325 	case MRT_API_SUPPORT:
326 	case MRT_API_CONFIG:
327 	case MRT_ADD_BW_UPCALL:
328 	case MRT_DEL_BW_UPCALL:
329 #ifdef MROUTING
330 		switch (op) {
331 		case PRCO_SETOPT:
332 			error = ip_mrouter_set(so, optname, m);
333 			break;
334 		case PRCO_GETOPT:
335 			error = ip_mrouter_get(so, optname, m);
336 			break;
337 		default:
338 			error = EINVAL;
339 			break;
340 		}
341 		return (error);
342 #else
343 		if (op == PRCO_SETOPT && *m)
344 			m_free(*m);
345 		return (EOPNOTSUPP);
346 #endif
347 	}
348 	return (ip_ctloutput(op, so, level, optname, m));
349 }
350 
351 u_long	rip_sendspace = RIPSNDQ;
352 u_long	rip_recvspace = RIPRCVQ;
353 
354 /*ARGSUSED*/
355 int
356 rip_usrreq(struct socket *so, int req, struct mbuf *m, struct mbuf *nam,
357     struct mbuf *control, struct proc *p)
358 {
359 	int error = 0;
360 	struct inpcb *inp = sotoinpcb(so);
361 #ifdef MROUTING
362 	extern struct socket *ip_mrouter;
363 #endif
364 	if (req == PRU_CONTROL)
365 		return (in_control(so, (u_long)m, (caddr_t)nam,
366 		    (struct ifnet *)control));
367 
368 	if (inp == NULL && req != PRU_ATTACH) {
369 		error = EINVAL;
370 		goto release;
371 	}
372 
373 	switch (req) {
374 
375 	case PRU_ATTACH:
376 		if (inp)
377 			panic("rip_attach");
378 		if ((so->so_state & SS_PRIV) == 0) {
379 			error = EACCES;
380 			break;
381 		}
382 		if ((error = soreserve(so, rip_sendspace, rip_recvspace)) ||
383 		    (error = in_pcballoc(so, &rawcbtable)))
384 			break;
385 		inp = (struct inpcb *)so->so_pcb;
386 		inp->inp_ip.ip_p = (long)nam;
387 		break;
388 
389 	case PRU_DISCONNECT:
390 		if ((so->so_state & SS_ISCONNECTED) == 0) {
391 			error = ENOTCONN;
392 			break;
393 		}
394 		/* FALLTHROUGH */
395 	case PRU_ABORT:
396 		soisdisconnected(so);
397 		/* FALLTHROUGH */
398 	case PRU_DETACH:
399 		if (inp == 0)
400 			panic("rip_detach");
401 #ifdef MROUTING
402 		if (so == ip_mrouter)
403 			ip_mrouter_done();
404 #endif
405 		in_pcbdetach(inp);
406 		break;
407 
408 	case PRU_BIND:
409 	    {
410 		struct sockaddr_in *addr = mtod(nam, struct sockaddr_in *);
411 
412 		if (nam->m_len != sizeof(*addr)) {
413 			error = EINVAL;
414 			break;
415 		}
416 		if ((TAILQ_EMPTY(&ifnet)) ||
417 		    ((addr->sin_family != AF_INET) &&
418 		     (addr->sin_family != AF_IMPLINK)) ||
419 		    (addr->sin_addr.s_addr &&
420 		     (!(so->so_options & SO_BINDANY) &&
421 		     in_iawithaddr(addr->sin_addr, NULL, inp->inp_rdomain) ==
422 		     0))) {
423 			error = EADDRNOTAVAIL;
424 			break;
425 		}
426 		inp->inp_laddr = addr->sin_addr;
427 		break;
428 	    }
429 	case PRU_CONNECT:
430 	    {
431 		struct sockaddr_in *addr = mtod(nam, struct sockaddr_in *);
432 
433 		if (nam->m_len != sizeof(*addr)) {
434 			error = EINVAL;
435 			break;
436 		}
437 		if (TAILQ_EMPTY(&ifnet)) {
438 			error = EADDRNOTAVAIL;
439 			break;
440 		}
441 		if ((addr->sin_family != AF_INET) &&
442 		     (addr->sin_family != AF_IMPLINK)) {
443 			error = EAFNOSUPPORT;
444 			break;
445 		}
446 		inp->inp_faddr = addr->sin_addr;
447 		soisconnected(so);
448 		break;
449 	    }
450 
451 	case PRU_CONNECT2:
452 		error = EOPNOTSUPP;
453 		break;
454 
455 	/*
456 	 * Mark the connection as being incapable of further input.
457 	 */
458 	case PRU_SHUTDOWN:
459 		socantsendmore(so);
460 		break;
461 
462 	/*
463 	 * Ship a packet out.  The appropriate raw output
464 	 * routine handles any massaging necessary.
465 	 */
466 	case PRU_SEND:
467 	    {
468 		u_int32_t dst;
469 
470 		if (so->so_state & SS_ISCONNECTED) {
471 			if (nam) {
472 				error = EISCONN;
473 				break;
474 			}
475 			dst = inp->inp_faddr.s_addr;
476 		} else {
477 			if (nam == NULL) {
478 				error = ENOTCONN;
479 				break;
480 			}
481 			dst = mtod(nam, struct sockaddr_in *)->sin_addr.s_addr;
482 		}
483 #ifdef IPSEC
484 		/* XXX Find an IPsec TDB */
485 #endif
486 		error = rip_output(m, so, dst);
487 		m = NULL;
488 		break;
489 	    }
490 
491 	case PRU_SENSE:
492 		/*
493 		 * stat: don't bother with a blocksize.
494 		 */
495 		return (0);
496 
497 	/*
498 	 * Not supported.
499 	 */
500 	case PRU_RCVOOB:
501 	case PRU_RCVD:
502 	case PRU_LISTEN:
503 	case PRU_ACCEPT:
504 	case PRU_SENDOOB:
505 		error = EOPNOTSUPP;
506 		break;
507 
508 	case PRU_SOCKADDR:
509 		in_setsockaddr(inp, nam);
510 		break;
511 
512 	case PRU_PEERADDR:
513 		in_setpeeraddr(inp, nam);
514 		break;
515 
516 	default:
517 		panic("rip_usrreq");
518 	}
519 release:
520 	if (m != NULL)
521 		m_freem(m);
522 	return (error);
523 }
524