xref: /freebsd/sys/netinet/raw_ip.c (revision e627b39baccd1ec9129690167cf5e6d860509655)
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  *	@(#)raw_ip.c	8.7 (Berkeley) 5/15/95
34  *	$Id: raw_ip.c,v 1.34 1996/08/21 21:37:07 sos Exp $
35  */
36 
37 #include <sys/param.h>
38 #include <sys/queue.h>
39 #include <sys/malloc.h>
40 #include <sys/mbuf.h>
41 #include <sys/socket.h>
42 #include <sys/protosw.h>
43 #include <sys/socketvar.h>
44 #include <sys/errno.h>
45 #include <sys/systm.h>
46 
47 #include <net/if.h>
48 #include <net/route.h>
49 
50 #define _IP_VHL
51 #include <netinet/in.h>
52 #include <netinet/in_systm.h>
53 #include <netinet/ip.h>
54 #include <netinet/in_pcb.h>
55 #include <netinet/in_var.h>
56 #include <netinet/ip_var.h>
57 #include <netinet/ip_mroute.h>
58 
59 #include <netinet/ip_fw.h>
60 
61 #if !defined(COMPAT_IPFW) || COMPAT_IPFW == 1
62 #undef COMPAT_IPFW
63 #define COMPAT_IPFW 1
64 #else
65 #undef COMPAT_IPFW
66 #endif
67 
68 static struct inpcbhead ripcb;
69 static struct inpcbinfo ripcbinfo;
70 
71 /*
72  * Nominal space allocated to a raw ip socket.
73  */
74 #define	RIPSNDQ		8192
75 #define	RIPRCVQ		8192
76 
77 /*
78  * Raw interface to IP protocol.
79  */
80 
81 /*
82  * Initialize raw connection block q.
83  */
84 void
85 rip_init()
86 {
87 	LIST_INIT(&ripcb);
88 	ripcbinfo.listhead = &ripcb;
89 	/*
90 	 * XXX We don't use the hash list for raw IP, but it's easier
91 	 * to allocate a one entry hash list than it is to check all
92 	 * over the place for hashbase == NULL.
93 	 */
94 	ripcbinfo.hashbase = phashinit(1, M_PCB, &ripcbinfo.hashsize);
95 }
96 
97 static struct	sockaddr_in ripsrc = { sizeof(ripsrc), AF_INET };
98 /*
99  * Setup generic address and protocol structures
100  * for raw_input routine, then pass them along with
101  * mbuf chain.
102  */
103 void
104 rip_input(m, iphlen)
105 	struct mbuf *m;
106 	int iphlen;
107 {
108 	register struct ip *ip = mtod(m, struct ip *);
109 	register struct inpcb *inp;
110 	struct socket *last = 0;
111 
112 	ripsrc.sin_addr = ip->ip_src;
113 	for (inp = ripcb.lh_first; inp != NULL; inp = inp->inp_list.le_next) {
114 		if (inp->inp_ip.ip_p && inp->inp_ip.ip_p != ip->ip_p)
115 			continue;
116 		if (inp->inp_laddr.s_addr &&
117                   inp->inp_laddr.s_addr != ip->ip_dst.s_addr)
118 			continue;
119 		if (inp->inp_faddr.s_addr &&
120                   inp->inp_faddr.s_addr != ip->ip_src.s_addr)
121 			continue;
122 		if (last) {
123 			struct mbuf *n = m_copy(m, 0, (int)M_COPYALL);
124 			if (n) {
125 				if (sbappendaddr(&last->so_rcv,
126 				    (struct sockaddr *)&ripsrc, n,
127 				    (struct mbuf *)0) == 0)
128 					/* should notify about lost packet */
129 					m_freem(n);
130 				else
131 					sorwakeup(last);
132 			}
133 		}
134 		last = inp->inp_socket;
135 	}
136 	if (last) {
137 		if (sbappendaddr(&last->so_rcv, (struct sockaddr *)&ripsrc,
138 		    m, (struct mbuf *)0) == 0)
139 			m_freem(m);
140 		else
141 			sorwakeup(last);
142 	} else {
143 		m_freem(m);
144               ipstat.ips_noproto++;
145               ipstat.ips_delivered--;
146       }
147 }
148 
149 /*
150  * Generate IP header and pass packet to ip_output.
151  * Tack on options user may have setup with control call.
152  */
153 int
154 rip_output(m, so, dst)
155 	register struct mbuf *m;
156 	struct socket *so;
157 	u_long dst;
158 {
159 	register struct ip *ip;
160 	register struct inpcb *inp = sotoinpcb(so);
161 	int flags = (so->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST;
162 
163 	/*
164 	 * If the user handed us a complete IP packet, use it.
165 	 * Otherwise, allocate an mbuf for a header and fill it in.
166 	 */
167 	if ((inp->inp_flags & INP_HDRINCL) == 0) {
168 		M_PREPEND(m, sizeof(struct ip), M_WAIT);
169 		ip = mtod(m, struct ip *);
170 		ip->ip_tos = 0;
171 		ip->ip_off = 0;
172 		ip->ip_p = inp->inp_ip.ip_p;
173 		ip->ip_len = m->m_pkthdr.len;
174 		ip->ip_src = inp->inp_laddr;
175 		ip->ip_dst.s_addr = dst;
176 		ip->ip_ttl = MAXTTL;
177 	} else {
178 		ip = mtod(m, struct ip *);
179 		/* don't allow both user specified and setsockopt options,
180 		   and don't allow packet length sizes that will crash */
181 		if (((IP_VHL_HL(ip->ip_vhl) != (sizeof (*ip) >> 2))
182 		     && inp->inp_options)
183 		    || (ip->ip_len > m->m_pkthdr.len)) {
184 			m_freem(m);
185 			return EINVAL;
186 		}
187 		if (ip->ip_id == 0)
188 			ip->ip_id = htons(ip_id++);
189 		/* XXX prevent ip_output from overwriting header fields */
190 		flags |= IP_RAWOUTPUT;
191 		ipstat.ips_rawout++;
192 	}
193 	return (ip_output(m, inp->inp_options, &inp->inp_route, flags,
194 			  inp->inp_moptions));
195 }
196 
197 /*
198  * Raw IP socket option processing.
199  */
200 int
201 rip_ctloutput(op, so, level, optname, m)
202 	int op;
203 	struct socket *so;
204 	int level, optname;
205 	struct mbuf **m;
206 {
207 	register struct inpcb *inp = sotoinpcb(so);
208 	register int error;
209 
210 	if (level != IPPROTO_IP) {
211 		if (op == PRCO_SETOPT && *m)
212 			(void)m_free(*m);
213 		return (EINVAL);
214 	}
215 
216 	switch (optname) {
217 
218 	case IP_HDRINCL:
219 		error = 0;
220 		if (op == PRCO_SETOPT) {
221 			if (m == 0 || *m == 0 || (*m)->m_len < sizeof (int))
222 				error = EINVAL;
223 			else if (*mtod(*m, int *))
224 				inp->inp_flags |= INP_HDRINCL;
225 			else
226 				inp->inp_flags &= ~INP_HDRINCL;
227 			if (*m)
228 				(void)m_free(*m);
229 		} else {
230 			*m = m_get(M_WAIT, MT_SOOPTS);
231 			(*m)->m_len = sizeof (int);
232 			*mtod(*m, int *) = inp->inp_flags & INP_HDRINCL;
233 		}
234 		return (error);
235 
236 #ifdef COMPAT_IPFW
237 	case IP_FW_GET:
238 		if (ip_fw_ctl_ptr == NULL || op == PRCO_SETOPT) {
239 			if (*m) (void)m_free(*m);
240 			return(EINVAL);
241 		}
242 		return (*ip_fw_ctl_ptr)(optname, m);
243 
244 	case IP_FW_ADD:
245 	case IP_FW_DEL:
246 	case IP_FW_FLUSH:
247 	case IP_FW_ZERO:
248 		if (ip_fw_ctl_ptr == NULL || op != PRCO_SETOPT) {
249 			if (*m) (void)m_free(*m);
250 			return(EINVAL);
251 		}
252 		return (*ip_fw_ctl_ptr)(optname, m);
253 #endif
254 
255 	case IP_NAT:
256 		if (ip_nat_ctl_ptr == NULL) {
257 			if (*m) (void)m_free(*m);
258 			return(EINVAL);
259 		}
260 		return (*ip_nat_ctl_ptr)(op, m);
261 
262 	case IP_RSVP_ON:
263 		return ip_rsvp_init(so);
264 		break;
265 
266 	case IP_RSVP_OFF:
267 		return ip_rsvp_done();
268 		break;
269 
270 	case IP_RSVP_VIF_ON:
271 		return ip_rsvp_vif_init(so, *m);
272 
273 	case IP_RSVP_VIF_OFF:
274 		return ip_rsvp_vif_done(so, *m);
275 
276 	case MRT_INIT:
277 	case MRT_DONE:
278 	case MRT_ADD_VIF:
279 	case MRT_DEL_VIF:
280 	case MRT_ADD_MFC:
281 	case MRT_DEL_MFC:
282 	case MRT_VERSION:
283 	case MRT_ASSERT:
284 		if (op == PRCO_SETOPT) {
285 			error = ip_mrouter_set(optname, so, *m);
286 			if (*m)
287 				(void)m_free(*m);
288 		} else if (op == PRCO_GETOPT) {
289 			error = ip_mrouter_get(optname, so, m);
290 		} else
291 			error = EINVAL;
292 		return (error);
293 	}
294 	return (ip_ctloutput(op, so, level, optname, m));
295 }
296 
297 static u_long	rip_sendspace = RIPSNDQ; /* XXX sysctl ? */
298 static u_long	rip_recvspace = RIPRCVQ; /* XXX sysctl ? */
299 
300 /*ARGSUSED*/
301 int
302 rip_usrreq(so, req, m, nam, control)
303 	register struct socket *so;
304 	int req;
305 	struct mbuf *m, *nam, *control;
306 {
307 	register int error = 0;
308 	register struct inpcb *inp = sotoinpcb(so);
309 
310 	if (req == PRU_CONTROL)
311 		return (in_control(so, (u_long)m, (caddr_t)nam,
312 			(struct ifnet *)control));
313 
314 	switch (req) {
315 
316 	case PRU_ATTACH:
317 		if (inp)
318 			panic("rip_attach");
319 		if ((so->so_state & SS_PRIV) == 0) {
320 			error = EACCES;
321 			break;
322 		}
323 		if ((error = soreserve(so, rip_sendspace, rip_recvspace)) ||
324 		    (error = in_pcballoc(so, &ripcbinfo)))
325 			break;
326 		inp = (struct inpcb *)so->so_pcb;
327 		inp->inp_ip.ip_p = (int)nam;
328 		break;
329 
330 	case PRU_DISCONNECT:
331 		if ((so->so_state & SS_ISCONNECTED) == 0) {
332 			error = ENOTCONN;
333 			break;
334 		}
335 		/* FALLTHROUGH */
336 	case PRU_ABORT:
337 		soisdisconnected(so);
338 		/* FALLTHROUGH */
339 	case PRU_DETACH:
340 		if (inp == 0)
341 			panic("rip_detach");
342 		if (so == ip_mrouter)
343 			ip_mrouter_done();
344 		ip_rsvp_force_done(so);
345 		if (so == ip_rsvpd)
346 			ip_rsvp_done();
347 		in_pcbdetach(inp);
348 		break;
349 
350 	case PRU_BIND:
351 	    {
352 		struct sockaddr_in *addr = mtod(nam, struct sockaddr_in *);
353 
354 		if (nam->m_len != sizeof(*addr)) {
355 			error = EINVAL;
356 			break;
357 		}
358 		if ((ifnet == 0) ||
359 		    ((addr->sin_family != AF_INET) &&
360 		     (addr->sin_family != AF_IMPLINK)) ||
361 		    (addr->sin_addr.s_addr &&
362 		     ifa_ifwithaddr((struct sockaddr *)addr) == 0)) {
363 			error = EADDRNOTAVAIL;
364 			break;
365 		}
366 		inp->inp_laddr = addr->sin_addr;
367 		break;
368 	    }
369 	case PRU_CONNECT:
370 	    {
371 		struct sockaddr_in *addr = mtod(nam, struct sockaddr_in *);
372 
373 		if (nam->m_len != sizeof(*addr)) {
374 			error = EINVAL;
375 			break;
376 		}
377 		if (ifnet == 0) {
378 			error = EADDRNOTAVAIL;
379 			break;
380 		}
381 		if ((addr->sin_family != AF_INET) &&
382 		     (addr->sin_family != AF_IMPLINK)) {
383 			error = EAFNOSUPPORT;
384 			break;
385 		}
386 		inp->inp_faddr = addr->sin_addr;
387 		soisconnected(so);
388 		break;
389 	    }
390 
391 	case PRU_CONNECT2:
392 		error = EOPNOTSUPP;
393 		break;
394 
395 	/*
396 	 * Mark the connection as being incapable of further input.
397 	 */
398 	case PRU_SHUTDOWN:
399 		socantsendmore(so);
400 		break;
401 
402 	/*
403 	 * Ship a packet out.  The appropriate raw output
404 	 * routine handles any massaging necessary.
405 	 */
406 	case PRU_SEND:
407 	    {
408 		register u_long dst;
409 
410 		if (so->so_state & SS_ISCONNECTED) {
411 			if (nam) {
412 				error = EISCONN;
413 				break;
414 			}
415 			dst = inp->inp_faddr.s_addr;
416 		} else {
417 			if (nam == NULL) {
418 				error = ENOTCONN;
419 				break;
420 			}
421 			dst = mtod(nam, struct sockaddr_in *)->sin_addr.s_addr;
422 		}
423 		error = rip_output(m, so, dst);
424 		m = NULL;
425 		break;
426 	    }
427 
428 	case PRU_SENSE:
429 		/*
430 		 * stat: don't bother with a blocksize.
431 		 */
432 		return (0);
433 
434 	/*
435 	 * Not supported.
436 	 */
437 	case PRU_RCVOOB:
438 	case PRU_RCVD:
439 	case PRU_LISTEN:
440 	case PRU_ACCEPT:
441 	case PRU_SENDOOB:
442 		error = EOPNOTSUPP;
443 		break;
444 
445 	case PRU_SOCKADDR:
446 		in_setsockaddr(inp, nam);
447 		break;
448 
449 	case PRU_PEERADDR:
450 		in_setpeeraddr(inp, nam);
451 		break;
452 
453 	default:
454 		panic("rip_usrreq");
455 	}
456 	if (m != NULL)
457 		m_freem(m);
458 	return (error);
459 }
460