raw_ip.c revision 22672 
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 *	$Id$
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
68static struct inpcbhead ripcb;
69static 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 */
84void
85rip_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
97static 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 */
103void
104rip_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 inpcb *last = 0;
111	struct mbuf *opts = 0;
112
113	ripsrc.sin_addr = ip->ip_src;
114	for (inp = ripcb.lh_first; inp != NULL; inp = inp->inp_list.le_next) {
115		if (inp->inp_ip.ip_p && inp->inp_ip.ip_p != ip->ip_p)
116			continue;
117		if (inp->inp_laddr.s_addr &&
118                  inp->inp_laddr.s_addr != ip->ip_dst.s_addr)
119			continue;
120		if (inp->inp_faddr.s_addr &&
121                  inp->inp_faddr.s_addr != ip->ip_src.s_addr)
122			continue;
123		if (last) {
124			struct mbuf *n = m_copy(m, 0, (int)M_COPYALL);
125			if (n) {
126				if (last->inp_flags & INP_CONTROLOPTS ||
127				    last->inp_socket->so_options & SO_TIMESTAMP)
128				    ip_savecontrol(last, &opts, ip, n);
129				if (sbappendaddr(&last->inp_socket->so_rcv,
130				    (struct sockaddr *)&ripsrc, n,
131				    opts) == 0) {
132					/* should notify about lost packet */
133					m_freem(n);
134					if (opts)
135					    m_freem(opts);
136				} else
137					sorwakeup(last->inp_socket);
138				opts = 0;
139			}
140		}
141		last = inp;
142	}
143	if (last) {
144		if (last->inp_flags & INP_CONTROLOPTS ||
145		    last->inp_socket->so_options & SO_TIMESTAMP)
146			ip_savecontrol(last, &opts, ip, m);
147		if (sbappendaddr(&last->inp_socket->so_rcv,
148		    (struct sockaddr *)&ripsrc, m, opts) == 0) {
149			m_freem(m);
150			if (opts)
151			    m_freem(opts);
152		} else
153			sorwakeup(last->inp_socket);
154	} else {
155		m_freem(m);
156              ipstat.ips_noproto++;
157              ipstat.ips_delivered--;
158      }
159}
160
161/*
162 * Generate IP header and pass packet to ip_output.
163 * Tack on options user may have setup with control call.
164 */
165int
166rip_output(m, so, dst)
167	register struct mbuf *m;
168	struct socket *so;
169	u_long dst;
170{
171	register struct ip *ip;
172	register struct inpcb *inp = sotoinpcb(so);
173	int flags = (so->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST;
174
175	/*
176	 * If the user handed us a complete IP packet, use it.
177	 * Otherwise, allocate an mbuf for a header and fill it in.
178	 */
179	if ((inp->inp_flags & INP_HDRINCL) == 0) {
180		if (m->m_pkthdr.len + sizeof(struct ip) > IP_MAXPACKET) {
181			m_freem(m);
182			return(EMSGSIZE);
183		}
184		M_PREPEND(m, sizeof(struct ip), M_WAIT);
185		ip = mtod(m, struct ip *);
186		ip->ip_tos = 0;
187		ip->ip_off = 0;
188		ip->ip_p = inp->inp_ip.ip_p;
189		ip->ip_len = m->m_pkthdr.len;
190		ip->ip_src = inp->inp_laddr;
191		ip->ip_dst.s_addr = dst;
192		ip->ip_ttl = MAXTTL;
193	} else {
194		if (m->m_pkthdr.len > IP_MAXPACKET) {
195			m_freem(m);
196			return(EMSGSIZE);
197		}
198		ip = mtod(m, struct ip *);
199		/* don't allow both user specified and setsockopt options,
200		   and don't allow packet length sizes that will crash */
201		if (((IP_VHL_HL(ip->ip_vhl) != (sizeof (*ip) >> 2))
202		     && inp->inp_options)
203		    || (ip->ip_len > m->m_pkthdr.len)) {
204			m_freem(m);
205			return EINVAL;
206		}
207		if (ip->ip_id == 0)
208			ip->ip_id = htons(ip_id++);
209		/* XXX prevent ip_output from overwriting header fields */
210		flags |= IP_RAWOUTPUT;
211		ipstat.ips_rawout++;
212	}
213	return (ip_output(m, inp->inp_options, &inp->inp_route, flags,
214			  inp->inp_moptions));
215}
216
217/*
218 * Raw IP socket option processing.
219 */
220int
221rip_ctloutput(op, so, level, optname, m)
222	int op;
223	struct socket *so;
224	int level, optname;
225	struct mbuf **m;
226{
227	register struct inpcb *inp = sotoinpcb(so);
228	register int error;
229
230	if (level != IPPROTO_IP) {
231		if (op == PRCO_SETOPT && *m)
232			(void)m_free(*m);
233		return (EINVAL);
234	}
235
236	switch (optname) {
237
238	case IP_HDRINCL:
239		error = 0;
240		if (op == PRCO_SETOPT) {
241			if (m == 0 || *m == 0 || (*m)->m_len < sizeof (int))
242				error = EINVAL;
243			else if (*mtod(*m, int *))
244				inp->inp_flags |= INP_HDRINCL;
245			else
246				inp->inp_flags &= ~INP_HDRINCL;
247			if (*m)
248				(void)m_free(*m);
249		} else {
250			*m = m_get(M_WAIT, MT_SOOPTS);
251			(*m)->m_len = sizeof (int);
252			*mtod(*m, int *) = inp->inp_flags & INP_HDRINCL;
253		}
254		return (error);
255
256#ifdef COMPAT_IPFW
257	case IP_FW_GET:
258		if (ip_fw_ctl_ptr == NULL || op == PRCO_SETOPT) {
259			if (*m) (void)m_free(*m);
260			return(EINVAL);
261		}
262		return (*ip_fw_ctl_ptr)(optname, m);
263
264	case IP_FW_ADD:
265	case IP_FW_DEL:
266	case IP_FW_FLUSH:
267	case IP_FW_ZERO:
268		if (ip_fw_ctl_ptr == NULL || op != PRCO_SETOPT) {
269			if (*m) (void)m_free(*m);
270			return(EINVAL);
271		}
272		return (*ip_fw_ctl_ptr)(optname, m);
273
274	case IP_NAT:
275		if (ip_nat_ctl_ptr == NULL) {
276			if (*m) (void)m_free(*m);
277			return(EINVAL);
278		}
279		return (*ip_nat_ctl_ptr)(op, m);
280
281#endif
282	case IP_RSVP_ON:
283		return ip_rsvp_init(so);
284		break;
285
286	case IP_RSVP_OFF:
287		return ip_rsvp_done();
288		break;
289
290	case IP_RSVP_VIF_ON:
291		return ip_rsvp_vif_init(so, *m);
292
293	case IP_RSVP_VIF_OFF:
294		return ip_rsvp_vif_done(so, *m);
295
296	case MRT_INIT:
297	case MRT_DONE:
298	case MRT_ADD_VIF:
299	case MRT_DEL_VIF:
300	case MRT_ADD_MFC:
301	case MRT_DEL_MFC:
302	case MRT_VERSION:
303	case MRT_ASSERT:
304		if (op == PRCO_SETOPT) {
305			error = ip_mrouter_set(optname, so, *m);
306			if (*m)
307				(void)m_free(*m);
308		} else if (op == PRCO_GETOPT) {
309			error = ip_mrouter_get(optname, so, m);
310		} else
311			error = EINVAL;
312		return (error);
313	}
314	return (ip_ctloutput(op, so, level, optname, m));
315}
316
317/*
318 * This function exists solely to receive the PRC_IFDOWN messages which
319 * are sent by if_down().  It looks for an ifaddr whose ifa_addr is sa,
320 * and calls in_ifadown() to remove all routes corresponding to that address.
321 * It also receives the PRC_IFUP messages from if_up() and reinstalls the
322 * interface routes.
323 */
324void
325rip_ctlinput(cmd, sa, vip)
326	int cmd;
327	struct sockaddr *sa;
328	void *vip;
329{
330	struct in_ifaddr *ia;
331	struct ifnet *ifp;
332	int err;
333	int flags;
334
335	switch(cmd) {
336	case PRC_IFDOWN:
337		for (ia = in_ifaddrhead.tqh_first; ia;
338		     ia = ia->ia_link.tqe_next) {
339			if (ia->ia_ifa.ifa_addr == sa
340			    && (ia->ia_flags & IFA_ROUTE)) {
341				/*
342				 * in_ifscrub kills the interface route.
343				 */
344				in_ifscrub(ia->ia_ifp, ia);
345				/*
346				 * in_ifadown gets rid of all the rest of
347				 * the routes.  This is not quite the right
348				 * thing to do, but at least if we are running
349				 * a routing process they will come back.
350				 */
351				in_ifadown(&ia->ia_ifa);
352				break;
353			}
354		}
355		break;
356
357	case PRC_IFUP:
358		for (ia = in_ifaddrhead.tqh_first; ia;
359		     ia = ia->ia_link.tqe_next) {
360			if (ia->ia_ifa.ifa_addr == sa)
361				break;
362		}
363		if (ia == 0 || (ia->ia_flags & IFA_ROUTE))
364			return;
365		flags = RTF_UP;
366		ifp = ia->ia_ifa.ifa_ifp;
367
368		if ((ifp->if_flags & IFF_LOOPBACK)
369		    || (ifp->if_flags & IFF_POINTOPOINT))
370			flags |= RTF_HOST;
371
372		err = rtinit(&ia->ia_ifa, RTM_ADD, flags);
373		if (err == 0)
374			ia->ia_flags |= IFA_ROUTE;
375		break;
376	}
377}
378
379static u_long	rip_sendspace = RIPSNDQ; /* XXX sysctl ? */
380static u_long	rip_recvspace = RIPRCVQ; /* XXX sysctl ? */
381
382/*ARGSUSED*/
383int
384rip_usrreq(so, req, m, nam, control)
385	register struct socket *so;
386	int req;
387	struct mbuf *m, *nam, *control;
388{
389	register int error = 0;
390	register struct inpcb *inp = sotoinpcb(so);
391
392	if (req == PRU_CONTROL)
393		return (in_control(so, (u_long)m, (caddr_t)nam,
394			(struct ifnet *)control));
395
396	switch (req) {
397
398	case PRU_ATTACH:
399		if (inp)
400			panic("rip_attach");
401		if ((so->so_state & SS_PRIV) == 0) {
402			error = EACCES;
403			break;
404		}
405		if ((error = soreserve(so, rip_sendspace, rip_recvspace)) ||
406		    (error = in_pcballoc(so, &ripcbinfo)))
407			break;
408		inp = (struct inpcb *)so->so_pcb;
409		inp->inp_ip.ip_p = (int)nam;
410		break;
411
412	case PRU_DISCONNECT:
413		if ((so->so_state & SS_ISCONNECTED) == 0) {
414			error = ENOTCONN;
415			break;
416		}
417		/* FALLTHROUGH */
418	case PRU_ABORT:
419		soisdisconnected(so);
420		/* FALLTHROUGH */
421	case PRU_DETACH:
422		if (inp == 0)
423			panic("rip_detach");
424		if (so == ip_mrouter)
425			ip_mrouter_done();
426		ip_rsvp_force_done(so);
427		if (so == ip_rsvpd)
428			ip_rsvp_done();
429		in_pcbdetach(inp);
430		break;
431
432	case PRU_BIND:
433	    {
434		struct sockaddr_in *addr = mtod(nam, struct sockaddr_in *);
435
436		if (nam->m_len != sizeof(*addr)) {
437			error = EINVAL;
438			break;
439		}
440		if (TAILQ_EMPTY(&ifnet) ||
441		    ((addr->sin_family != AF_INET) &&
442		     (addr->sin_family != AF_IMPLINK)) ||
443		    (addr->sin_addr.s_addr &&
444		     ifa_ifwithaddr((struct sockaddr *)addr) == 0)) {
445			error = EADDRNOTAVAIL;
446			break;
447		}
448		inp->inp_laddr = addr->sin_addr;
449		break;
450	    }
451	case PRU_CONNECT:
452	    {
453		struct sockaddr_in *addr = mtod(nam, struct sockaddr_in *);
454
455		if (nam->m_len != sizeof(*addr)) {
456			error = EINVAL;
457			break;
458		}
459		if (TAILQ_EMPTY(&ifnet)) {
460			error = EADDRNOTAVAIL;
461			break;
462		}
463		if ((addr->sin_family != AF_INET) &&
464		     (addr->sin_family != AF_IMPLINK)) {
465			error = EAFNOSUPPORT;
466			break;
467		}
468		inp->inp_faddr = addr->sin_addr;
469		soisconnected(so);
470		break;
471	    }
472
473	case PRU_CONNECT2:
474		error = EOPNOTSUPP;
475		break;
476
477	/*
478	 * Mark the connection as being incapable of further input.
479	 */
480	case PRU_SHUTDOWN:
481		socantsendmore(so);
482		break;
483
484	/*
485	 * Ship a packet out.  The appropriate raw output
486	 * routine handles any massaging necessary.
487	 */
488	case PRU_SEND:
489	    {
490		register u_long dst;
491
492		if (so->so_state & SS_ISCONNECTED) {
493			if (nam) {
494				error = EISCONN;
495				break;
496			}
497			dst = inp->inp_faddr.s_addr;
498		} else {
499			if (nam == NULL) {
500				error = ENOTCONN;
501				break;
502			}
503			dst = mtod(nam, struct sockaddr_in *)->sin_addr.s_addr;
504		}
505		error = rip_output(m, so, dst);
506		m = NULL;
507		break;
508	    }
509
510	case PRU_SENSE:
511		/*
512		 * stat: don't bother with a blocksize.
513		 */
514		return (0);
515
516	/*
517	 * Not supported.
518	 */
519	case PRU_RCVOOB:
520	case PRU_RCVD:
521	case PRU_LISTEN:
522	case PRU_ACCEPT:
523	case PRU_SENDOOB:
524		error = EOPNOTSUPP;
525		break;
526
527	case PRU_SOCKADDR:
528		in_setsockaddr(inp, nam);
529		break;
530
531	case PRU_PEERADDR:
532		in_setpeeraddr(inp, nam);
533		break;
534
535	default:
536		panic("rip_usrreq");
537	}
538	if (m != NULL)
539		m_freem(m);
540	return (error);
541}
542