xref: /openbsd-src/sys/netinet/raw_ip.c (revision 4c1e55dc91edd6e69ccc60ce855900fbc12cf34f)
1 /*	$OpenBSD: raw_ip.c,v 1.61 2012/03/17 10:16:41 dlg 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 		if (inp->inp_socket->so_state & SS_CANTRCVMORE)
132 			continue;
133 #ifdef INET6
134 		if (inp->inp_flags & INP_IPV6)
135 			continue;
136 #endif
137 		if (rtable_l2(inp->inp_rtableid) !=
138 		    rtable_l2(m->m_pkthdr.rdomain))
139 			continue;
140 
141 		if (inp->inp_ip.ip_p && inp->inp_ip.ip_p != ip->ip_p)
142 			continue;
143 #if NPF > 0
144 		if (m->m_pkthdr.pf.flags & PF_TAG_DIVERTED) {
145 			struct pf_divert *divert;
146 
147 			/* XXX rdomain support */
148 			if ((divert = pf_find_divert(m)) == NULL)
149 				continue;
150 			if (inp->inp_laddr.s_addr != divert->addr.v4.s_addr)
151 				continue;
152 		} else
153 #endif
154 		if (inp->inp_laddr.s_addr &&
155 		    inp->inp_laddr.s_addr != ip->ip_dst.s_addr)
156 			continue;
157 		if (inp->inp_faddr.s_addr &&
158 		    inp->inp_faddr.s_addr != ip->ip_src.s_addr)
159 			continue;
160 		if (last) {
161 			struct mbuf *n;
162 
163 			if ((n = m_copy(m, 0, (int)M_COPYALL)) != NULL) {
164 				if (last->inp_flags & INP_CONTROLOPTS ||
165 				    last->inp_socket->so_options & SO_TIMESTAMP)
166 					ip_savecontrol(last, &opts, ip, n);
167 				if (sbappendaddr(&last->inp_socket->so_rcv,
168 				    sintosa(&ripsrc), n, opts) == 0) {
169 					/* should notify about lost packet */
170 					m_freem(n);
171 					if (opts)
172 						m_freem(opts);
173 				} else
174 					sorwakeup(last->inp_socket);
175 				opts = NULL;
176 			}
177 		}
178 		last = inp;
179 	}
180 	if (last) {
181 		if (last->inp_flags & INP_CONTROLOPTS ||
182 		    last->inp_socket->so_options & SO_TIMESTAMP)
183 			ip_savecontrol(last, &opts, ip, m);
184 		if (sbappendaddr(&last->inp_socket->so_rcv, sintosa(&ripsrc), m,
185 		    opts) == 0) {
186 			m_freem(m);
187 			if (opts)
188 				m_freem(opts);
189 		} else
190 			sorwakeup(last->inp_socket);
191 	} else {
192 		if (ip->ip_p != IPPROTO_ICMP)
193 			icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_PROTOCOL, 0, 0);
194 		else
195 			m_freem(m);
196 		ipstat.ips_noproto++;
197 		ipstat.ips_delivered--;
198 	}
199 }
200 
201 /*
202  * Generate IP header and pass packet to ip_output.
203  * Tack on options user may have setup with control call.
204  */
205 int
206 rip_output(struct mbuf *m, ...)
207 {
208 	struct socket *so;
209 	u_long dst;
210 	struct ip *ip;
211 	struct inpcb *inp;
212 	int flags, error;
213 	va_list ap;
214 
215 	va_start(ap, m);
216 	so = va_arg(ap, struct socket *);
217 	dst = va_arg(ap, u_long);
218 	va_end(ap);
219 
220 	inp = sotoinpcb(so);
221 	flags = (so->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST;
222 
223 	/*
224 	 * If the user handed us a complete IP packet, use it.
225 	 * Otherwise, allocate an mbuf for a header and fill it in.
226 	 */
227 	if ((inp->inp_flags & INP_HDRINCL) == 0) {
228 		if ((m->m_pkthdr.len + sizeof(struct ip)) > IP_MAXPACKET) {
229 			m_freem(m);
230 			return (EMSGSIZE);
231 		}
232 		M_PREPEND(m, sizeof(struct ip), M_DONTWAIT);
233 		if (!m)
234 			return (ENOBUFS);
235 		ip = mtod(m, struct ip *);
236 		ip->ip_tos = inp->inp_ip.ip_tos;
237 		ip->ip_off = htons(0);
238 		ip->ip_p = inp->inp_ip.ip_p;
239 		ip->ip_len = htons(m->m_pkthdr.len);
240 		ip->ip_src = inp->inp_laddr;
241 		ip->ip_dst.s_addr = dst;
242 		ip->ip_ttl = inp->inp_ip.ip_ttl ? inp->inp_ip.ip_ttl : MAXTTL;
243 	} else {
244 		if (m->m_pkthdr.len > IP_MAXPACKET) {
245 			m_freem(m);
246 			return (EMSGSIZE);
247 		}
248 		if (m->m_pkthdr.len < sizeof(struct ip)) {
249 			m_freem(m);
250 			return (EINVAL);
251 		}
252 		ip = mtod(m, struct ip *);
253 		/*
254 		 * don't allow both user specified and setsockopt options,
255 		 * and don't allow packet length sizes that will crash
256 		 */
257 		if ((ip->ip_hl != (sizeof (*ip) >> 2) && inp->inp_options) ||
258 		    ntohs(ip->ip_len) > m->m_pkthdr.len ||
259 		    ntohs(ip->ip_len) < ip->ip_hl << 2) {
260 			m_freem(m);
261 			return (EINVAL);
262 		}
263 		if (ip->ip_id == 0) {
264 			ip->ip_id = htons(ip_randomid());
265 		}
266 		/* XXX prevent ip_output from overwriting header fields */
267 		flags |= IP_RAWOUTPUT;
268 		ipstat.ips_rawout++;
269 	}
270 #ifdef INET6
271 	/*
272 	 * A thought:  Even though raw IP shouldn't be able to set IPv6
273 	 *             multicast options, if it does, the last parameter to
274 	 *             ip_output should be guarded against v6/v4 problems.
275 	 */
276 #endif
277 	/* force routing domain */
278 	m->m_pkthdr.rdomain = inp->inp_rtableid;
279 
280 	error = ip_output(m, inp->inp_options, &inp->inp_route, flags,
281 	    inp->inp_moptions, inp);
282 	if (error == EACCES)	/* translate pf(4) error for userland */
283 		error = EHOSTUNREACH;
284 	return (error);
285 }
286 
287 /*
288  * Raw IP socket option processing.
289  */
290 int
291 rip_ctloutput(int op, struct socket *so, int level, int optname,
292     struct mbuf **m)
293 {
294 	struct inpcb *inp = sotoinpcb(so);
295 	int error;
296 
297 	if (level != IPPROTO_IP) {
298 		if (op == PRCO_SETOPT && *m)
299 			(void) m_free(*m);
300 		return (EINVAL);
301 	}
302 
303 	switch (optname) {
304 
305 	case IP_HDRINCL:
306 		error = 0;
307 		if (op == PRCO_SETOPT) {
308 			if (*m == 0 || (*m)->m_len < sizeof (int))
309 				error = EINVAL;
310 			else if (*mtod(*m, int *))
311 				inp->inp_flags |= INP_HDRINCL;
312 			else
313 				inp->inp_flags &= ~INP_HDRINCL;
314 			if (*m)
315 				(void)m_free(*m);
316 		} else {
317 			*m = m_get(M_WAIT, M_SOOPTS);
318 			(*m)->m_len = sizeof(int);
319 			*mtod(*m, int *) = inp->inp_flags & INP_HDRINCL;
320 		}
321 		return (error);
322 
323 	case MRT_INIT:
324 	case MRT_DONE:
325 	case MRT_ADD_VIF:
326 	case MRT_DEL_VIF:
327 	case MRT_ADD_MFC:
328 	case MRT_DEL_MFC:
329 	case MRT_VERSION:
330 	case MRT_ASSERT:
331 	case MRT_API_SUPPORT:
332 	case MRT_API_CONFIG:
333 	case MRT_ADD_BW_UPCALL:
334 	case MRT_DEL_BW_UPCALL:
335 #ifdef MROUTING
336 		switch (op) {
337 		case PRCO_SETOPT:
338 			error = ip_mrouter_set(so, optname, m);
339 			break;
340 		case PRCO_GETOPT:
341 			error = ip_mrouter_get(so, optname, m);
342 			break;
343 		default:
344 			error = EINVAL;
345 			break;
346 		}
347 		return (error);
348 #else
349 		if (op == PRCO_SETOPT && *m)
350 			m_free(*m);
351 		return (EOPNOTSUPP);
352 #endif
353 	}
354 	return (ip_ctloutput(op, so, level, optname, m));
355 }
356 
357 u_long	rip_sendspace = RIPSNDQ;
358 u_long	rip_recvspace = RIPRCVQ;
359 
360 /*ARGSUSED*/
361 int
362 rip_usrreq(struct socket *so, int req, struct mbuf *m, struct mbuf *nam,
363     struct mbuf *control, struct proc *p)
364 {
365 	int error = 0;
366 	struct inpcb *inp = sotoinpcb(so);
367 #ifdef MROUTING
368 	extern struct socket *ip_mrouter;
369 #endif
370 	if (req == PRU_CONTROL)
371 		return (in_control(so, (u_long)m, (caddr_t)nam,
372 		    (struct ifnet *)control));
373 
374 	if (inp == NULL && req != PRU_ATTACH) {
375 		error = EINVAL;
376 		goto release;
377 	}
378 
379 	switch (req) {
380 
381 	case PRU_ATTACH:
382 		if (inp)
383 			panic("rip_attach");
384 		if ((so->so_state & SS_PRIV) == 0) {
385 			error = EACCES;
386 			break;
387 		}
388 		if ((error = soreserve(so, rip_sendspace, rip_recvspace)) ||
389 		    (error = in_pcballoc(so, &rawcbtable)))
390 			break;
391 		inp = (struct inpcb *)so->so_pcb;
392 		inp->inp_ip.ip_p = (long)nam;
393 		break;
394 
395 	case PRU_DISCONNECT:
396 		if ((so->so_state & SS_ISCONNECTED) == 0) {
397 			error = ENOTCONN;
398 			break;
399 		}
400 		/* FALLTHROUGH */
401 	case PRU_ABORT:
402 		soisdisconnected(so);
403 		/* FALLTHROUGH */
404 	case PRU_DETACH:
405 		if (inp == 0)
406 			panic("rip_detach");
407 #ifdef MROUTING
408 		if (so == ip_mrouter)
409 			ip_mrouter_done();
410 #endif
411 		in_pcbdetach(inp);
412 		break;
413 
414 	case PRU_BIND:
415 	    {
416 		struct sockaddr_in *addr = mtod(nam, struct sockaddr_in *);
417 
418 		if (nam->m_len != sizeof(*addr)) {
419 			error = EINVAL;
420 			break;
421 		}
422 		if (TAILQ_EMPTY(&ifnet) || addr->sin_family != AF_INET) {
423 			error = EADDRNOTAVAIL;
424 			break;
425 		}
426 		if (!((so->so_options & SO_BINDANY) ||
427 		    addr->sin_addr.s_addr == 0 ||
428 		    in_iawithaddr(addr->sin_addr, inp->inp_rtableid) ||
429 		    in_broadcast(addr->sin_addr, NULL, inp->inp_rtableid))) {
430 			error = EADDRNOTAVAIL;
431 			break;
432 		}
433 		inp->inp_laddr = addr->sin_addr;
434 		break;
435 	    }
436 	case PRU_CONNECT:
437 	    {
438 		struct sockaddr_in *addr = mtod(nam, struct sockaddr_in *);
439 
440 		if (nam->m_len != sizeof(*addr)) {
441 			error = EINVAL;
442 			break;
443 		}
444 		if (TAILQ_EMPTY(&ifnet)) {
445 			error = EADDRNOTAVAIL;
446 			break;
447 		}
448 		if (addr->sin_family != AF_INET) {
449 			error = EAFNOSUPPORT;
450 			break;
451 		}
452 		inp->inp_faddr = addr->sin_addr;
453 		soisconnected(so);
454 		break;
455 	    }
456 
457 	case PRU_CONNECT2:
458 		error = EOPNOTSUPP;
459 		break;
460 
461 	/*
462 	 * Mark the connection as being incapable of further input.
463 	 */
464 	case PRU_SHUTDOWN:
465 		socantsendmore(so);
466 		break;
467 
468 	/*
469 	 * Ship a packet out.  The appropriate raw output
470 	 * routine handles any massaging necessary.
471 	 */
472 	case PRU_SEND:
473 	    {
474 		u_int32_t dst;
475 
476 		if (so->so_state & SS_ISCONNECTED) {
477 			if (nam) {
478 				error = EISCONN;
479 				break;
480 			}
481 			dst = inp->inp_faddr.s_addr;
482 		} else {
483 			if (nam == NULL) {
484 				error = ENOTCONN;
485 				break;
486 			}
487 			dst = mtod(nam, struct sockaddr_in *)->sin_addr.s_addr;
488 		}
489 #ifdef IPSEC
490 		/* XXX Find an IPsec TDB */
491 #endif
492 		error = rip_output(m, so, dst);
493 		m = NULL;
494 		break;
495 	    }
496 
497 	case PRU_SENSE:
498 		/*
499 		 * stat: don't bother with a blocksize.
500 		 */
501 		return (0);
502 
503 	/*
504 	 * Not supported.
505 	 */
506 	case PRU_RCVOOB:
507 	case PRU_RCVD:
508 	case PRU_LISTEN:
509 	case PRU_ACCEPT:
510 	case PRU_SENDOOB:
511 		error = EOPNOTSUPP;
512 		break;
513 
514 	case PRU_SOCKADDR:
515 		in_setsockaddr(inp, nam);
516 		break;
517 
518 	case PRU_PEERADDR:
519 		in_setpeeraddr(inp, nam);
520 		break;
521 
522 	default:
523 		panic("rip_usrreq");
524 	}
525 release:
526 	if (m != NULL)
527 		m_freem(m);
528 	return (error);
529 }
530