raw_ip.c revision 27845 
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 1.46 1997/05/22 20:52:56 fenner Exp $
35 */
36
37#include <sys/param.h>
38#include <sys/systm.h>
39#include <sys/kernel.h>
40#include <sys/malloc.h>
41#include <sys/mbuf.h>
42#include <sys/proc.h>
43#include <sys/protosw.h>
44#include <sys/socket.h>
45#include <sys/socketvar.h>
46#include <sys/sysctl.h>
47
48#include <net/if.h>
49#include <net/route.h>
50
51#define _IP_VHL
52#include <netinet/in.h>
53#include <netinet/in_systm.h>
54#include <netinet/ip.h>
55#include <netinet/in_pcb.h>
56#include <netinet/in_var.h>
57#include <netinet/ip_var.h>
58#include <netinet/ip_mroute.h>
59
60#include <netinet/ip_fw.h>
61
62#if !defined(COMPAT_IPFW) || COMPAT_IPFW == 1
63#undef COMPAT_IPFW
64#define COMPAT_IPFW 1
65#else
66#undef COMPAT_IPFW
67#endif
68
69static struct inpcbhead ripcb;
70static struct inpcbinfo ripcbinfo;
71
72/*
73 * Nominal space allocated to a raw ip socket.
74 */
75#define	RIPSNDQ		8192
76#define	RIPRCVQ		8192
77
78/*
79 * Raw interface to IP protocol.
80 */
81
82/*
83 * Initialize raw connection block q.
84 */
85void
86rip_init()
87{
88	LIST_INIT(&ripcb);
89	ripcbinfo.listhead = &ripcb;
90	/*
91	 * XXX We don't use the hash list for raw IP, but it's easier
92	 * to allocate a one entry hash list than it is to check all
93	 * over the place for hashbase == NULL.
94	 */
95	ripcbinfo.hashbase = hashinit(1, M_PCB, &ripcbinfo.hashmask);
96}
97
98static struct	sockaddr_in ripsrc = { sizeof(ripsrc), AF_INET };
99/*
100 * Setup generic address and protocol structures
101 * for raw_input routine, then pass them along with
102 * mbuf chain.
103 */
104void
105rip_input(m, iphlen)
106	struct mbuf *m;
107	int iphlen;
108{
109	register struct ip *ip = mtod(m, struct ip *);
110	register struct inpcb *inp;
111	struct inpcb *last = 0;
112	struct mbuf *opts = 0;
113
114	ripsrc.sin_addr = ip->ip_src;
115	for (inp = ripcb.lh_first; inp != NULL; inp = inp->inp_list.le_next) {
116		if (inp->inp_ip_p && inp->inp_ip_p != ip->ip_p)
117			continue;
118		if (inp->inp_laddr.s_addr &&
119                  inp->inp_laddr.s_addr != ip->ip_dst.s_addr)
120			continue;
121		if (inp->inp_faddr.s_addr &&
122                  inp->inp_faddr.s_addr != ip->ip_src.s_addr)
123			continue;
124		if (last) {
125			struct mbuf *n = m_copy(m, 0, (int)M_COPYALL);
126			if (n) {
127				if (last->inp_flags & INP_CONTROLOPTS ||
128				    last->inp_socket->so_options & SO_TIMESTAMP)
129				    ip_savecontrol(last, &opts, ip, n);
130				if (sbappendaddr(&last->inp_socket->so_rcv,
131				    (struct sockaddr *)&ripsrc, n,
132				    opts) == 0) {
133					/* should notify about lost packet */
134					m_freem(n);
135					if (opts)
136					    m_freem(opts);
137				} else
138					sorwakeup(last->inp_socket);
139				opts = 0;
140			}
141		}
142		last = inp;
143	}
144	if (last) {
145		if (last->inp_flags & INP_CONTROLOPTS ||
146		    last->inp_socket->so_options & SO_TIMESTAMP)
147			ip_savecontrol(last, &opts, ip, m);
148		if (sbappendaddr(&last->inp_socket->so_rcv,
149		    (struct sockaddr *)&ripsrc, m, opts) == 0) {
150			m_freem(m);
151			if (opts)
152			    m_freem(opts);
153		} else
154			sorwakeup(last->inp_socket);
155	} else {
156		m_freem(m);
157              ipstat.ips_noproto++;
158              ipstat.ips_delivered--;
159      }
160}
161
162/*
163 * Generate IP header and pass packet to ip_output.
164 * Tack on options user may have setup with control call.
165 */
166int
167rip_output(m, so, dst)
168	register struct mbuf *m;
169	struct socket *so;
170	u_long dst;
171{
172	register struct ip *ip;
173	register struct inpcb *inp = sotoinpcb(so);
174	int flags = (so->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST;
175
176	/*
177	 * If the user handed us a complete IP packet, use it.
178	 * Otherwise, allocate an mbuf for a header and fill it in.
179	 */
180	if ((inp->inp_flags & INP_HDRINCL) == 0) {
181		if (m->m_pkthdr.len + sizeof(struct ip) > IP_MAXPACKET) {
182			m_freem(m);
183			return(EMSGSIZE);
184		}
185		M_PREPEND(m, sizeof(struct ip), M_WAIT);
186		ip = mtod(m, struct ip *);
187		ip->ip_tos = 0;
188		ip->ip_off = 0;
189		ip->ip_p = inp->inp_ip_p;
190		ip->ip_len = m->m_pkthdr.len;
191		ip->ip_src = inp->inp_laddr;
192		ip->ip_dst.s_addr = dst;
193		ip->ip_ttl = MAXTTL;
194	} else {
195		if (m->m_pkthdr.len > IP_MAXPACKET) {
196			m_freem(m);
197			return(EMSGSIZE);
198		}
199		ip = mtod(m, struct ip *);
200		/* don't allow both user specified and setsockopt options,
201		   and don't allow packet length sizes that will crash */
202		if (((IP_VHL_HL(ip->ip_vhl) != (sizeof (*ip) >> 2))
203		     && inp->inp_options)
204		    || (ip->ip_len > m->m_pkthdr.len)
205		    || (ip->ip_len < (IP_VHL_HL(ip->ip_vhl) << 2))) {
206			m_freem(m);
207			return EINVAL;
208		}
209		if (ip->ip_id == 0)
210			ip->ip_id = htons(ip_id++);
211		/* XXX prevent ip_output from overwriting header fields */
212		flags |= IP_RAWOUTPUT;
213		ipstat.ips_rawout++;
214	}
215	return (ip_output(m, inp->inp_options, &inp->inp_route, flags,
216			  inp->inp_moptions));
217}
218
219/*
220 * Raw IP socket option processing.
221 */
222int
223rip_ctloutput(op, so, level, optname, m, p)
224	int op;
225	struct socket *so;
226	int level, optname;
227	struct mbuf **m;
228	struct proc *p;
229{
230	register struct inpcb *inp = sotoinpcb(so);
231	register int error;
232
233	if (level != IPPROTO_IP) {
234		if (op == PRCO_SETOPT && *m)
235			(void)m_free(*m);
236		return (EINVAL);
237	}
238
239	switch (optname) {
240
241	case IP_HDRINCL:
242		error = 0;
243		if (op == PRCO_SETOPT) {
244			if (m == 0 || *m == 0 || (*m)->m_len < sizeof (int))
245				error = EINVAL;
246			else if (*mtod(*m, int *))
247				inp->inp_flags |= INP_HDRINCL;
248			else
249				inp->inp_flags &= ~INP_HDRINCL;
250			if (*m)
251				(void)m_free(*m);
252		} else {
253			*m = m_get(M_WAIT, MT_SOOPTS);
254			(*m)->m_len = sizeof (int);
255			*mtod(*m, int *) = inp->inp_flags & INP_HDRINCL;
256		}
257		return (error);
258
259#ifdef COMPAT_IPFW
260	case IP_FW_GET:
261		if (ip_fw_ctl_ptr == NULL || op == PRCO_SETOPT) {
262			if (*m) (void)m_free(*m);
263			return(EINVAL);
264		}
265		return (*ip_fw_ctl_ptr)(optname, m);
266
267	case IP_FW_ADD:
268	case IP_FW_DEL:
269	case IP_FW_FLUSH:
270	case IP_FW_ZERO:
271		if (ip_fw_ctl_ptr == NULL || op != PRCO_SETOPT) {
272			if (*m) (void)m_free(*m);
273			return(EINVAL);
274		}
275		return (*ip_fw_ctl_ptr)(optname, m);
276
277	case IP_NAT:
278		if (ip_nat_ctl_ptr == NULL) {
279			if (*m) (void)m_free(*m);
280			return(EINVAL);
281		}
282		return (*ip_nat_ctl_ptr)(op, m);
283
284#endif
285	case IP_RSVP_ON:
286		return ip_rsvp_init(so);
287		break;
288
289	case IP_RSVP_OFF:
290		return ip_rsvp_done();
291		break;
292
293	case IP_RSVP_VIF_ON:
294		return ip_rsvp_vif_init(so, *m);
295
296	case IP_RSVP_VIF_OFF:
297		return ip_rsvp_vif_done(so, *m);
298
299	case MRT_INIT:
300	case MRT_DONE:
301	case MRT_ADD_VIF:
302	case MRT_DEL_VIF:
303	case MRT_ADD_MFC:
304	case MRT_DEL_MFC:
305	case MRT_VERSION:
306	case MRT_ASSERT:
307		if (op == PRCO_SETOPT) {
308			error = ip_mrouter_set(optname, so, *m);
309			if (*m)
310				(void)m_free(*m);
311		} else if (op == PRCO_GETOPT) {
312			error = ip_mrouter_get(optname, so, m);
313		} else
314			error = EINVAL;
315		return (error);
316	}
317	return (ip_ctloutput(op, so, level, optname, m, p));
318}
319
320/*
321 * This function exists solely to receive the PRC_IFDOWN messages which
322 * are sent by if_down().  It looks for an ifaddr whose ifa_addr is sa,
323 * and calls in_ifadown() to remove all routes corresponding to that address.
324 * It also receives the PRC_IFUP messages from if_up() and reinstalls the
325 * interface routes.
326 */
327void
328rip_ctlinput(cmd, sa, vip)
329	int cmd;
330	struct sockaddr *sa;
331	void *vip;
332{
333	struct in_ifaddr *ia;
334	struct ifnet *ifp;
335	int err;
336	int flags;
337
338	switch(cmd) {
339	case PRC_IFDOWN:
340		for (ia = in_ifaddrhead.tqh_first; ia;
341		     ia = ia->ia_link.tqe_next) {
342			if (ia->ia_ifa.ifa_addr == sa
343			    && (ia->ia_flags & IFA_ROUTE)) {
344				/*
345				 * in_ifscrub kills the interface route.
346				 */
347				in_ifscrub(ia->ia_ifp, ia);
348				/*
349				 * in_ifadown gets rid of all the rest of
350				 * the routes.  This is not quite the right
351				 * thing to do, but at least if we are running
352				 * a routing process they will come back.
353				 */
354				in_ifadown(&ia->ia_ifa);
355				break;
356			}
357		}
358		break;
359
360	case PRC_IFUP:
361		for (ia = in_ifaddrhead.tqh_first; ia;
362		     ia = ia->ia_link.tqe_next) {
363			if (ia->ia_ifa.ifa_addr == sa)
364				break;
365		}
366		if (ia == 0 || (ia->ia_flags & IFA_ROUTE))
367			return;
368		flags = RTF_UP;
369		ifp = ia->ia_ifa.ifa_ifp;
370
371		if ((ifp->if_flags & IFF_LOOPBACK)
372		    || (ifp->if_flags & IFF_POINTOPOINT))
373			flags |= RTF_HOST;
374
375		err = rtinit(&ia->ia_ifa, RTM_ADD, flags);
376		if (err == 0)
377			ia->ia_flags |= IFA_ROUTE;
378		break;
379	}
380}
381
382static u_long	rip_sendspace = RIPSNDQ;
383static u_long	rip_recvspace = RIPRCVQ;
384
385SYSCTL_INT(_net_inet_raw, OID_AUTO, maxdgram, CTLFLAG_RW, &rip_sendspace,
386	   0, "");
387SYSCTL_INT(_net_inet_raw, OID_AUTO, recvspace, CTLFLAG_RW, &rip_recvspace,
388	   0, "");
389
390static int
391rip_attach(struct socket *so, int proto, struct proc *p)
392{
393	struct inpcb *inp;
394	int error;
395
396	inp = sotoinpcb(so);
397	if (inp)
398		panic("rip_attach");
399	if (p && (error = suser(p->p_ucred, &p->p_acflag)) != 0)
400		return error;
401
402	if ((error = soreserve(so, rip_sendspace, rip_recvspace)) ||
403	    (error = in_pcballoc(so, &ripcbinfo, p)))
404		return error;
405	inp = (struct inpcb *)so->so_pcb;
406	inp->inp_ip_p = proto;
407	return 0;
408}
409
410static int
411rip_detach(struct socket *so)
412{
413	struct inpcb *inp;
414
415	inp = sotoinpcb(so);
416	if (inp == 0)
417		panic("rip_detach");
418	if (so == ip_mrouter)
419		ip_mrouter_done();
420	ip_rsvp_force_done(so);
421	if (so == ip_rsvpd)
422		ip_rsvp_done();
423	in_pcbdetach(inp);
424	return 0;
425}
426
427static int
428rip_abort(struct socket *so)
429{
430	soisdisconnected(so);
431	return rip_detach(so);
432}
433
434static int
435rip_disconnect(struct socket *so)
436{
437	if ((so->so_state & SS_ISCONNECTED) == 0)
438		return ENOTCONN;
439	return rip_abort(so);
440}
441
442static int
443rip_bind(struct socket *so, struct mbuf *nam, struct proc *p)
444{
445	struct inpcb *inp = sotoinpcb(so);
446	struct sockaddr_in *addr = mtod(nam, struct sockaddr_in *);
447
448	if (nam->m_len != sizeof(*addr))
449		return EINVAL;
450
451	if (TAILQ_EMPTY(&ifnet) || ((addr->sin_family != AF_INET) &&
452				    (addr->sin_family != AF_IMPLINK)) ||
453	    (addr->sin_addr.s_addr &&
454	     ifa_ifwithaddr((struct sockaddr *)addr) == 0))
455		return EADDRNOTAVAIL;
456	inp->inp_laddr = addr->sin_addr;
457	return 0;
458}
459
460static int
461rip_connect(struct socket *so, struct mbuf *nam, struct proc *p)
462{
463	struct inpcb *inp = sotoinpcb(so);
464	struct sockaddr_in *addr = mtod(nam, struct sockaddr_in *);
465
466	if (nam->m_len != sizeof(*addr))
467		return EINVAL;
468	if (TAILQ_EMPTY(&ifnet))
469		return EADDRNOTAVAIL;
470	if ((addr->sin_family != AF_INET) &&
471	    (addr->sin_family != AF_IMPLINK))
472		return EAFNOSUPPORT;
473	inp->inp_faddr = addr->sin_addr;
474	soisconnected(so);
475	return 0;
476}
477
478static int
479rip_shutdown(struct socket *so)
480{
481	socantsendmore(so);
482	return 0;
483}
484
485static int
486rip_send(struct socket *so, int flags, struct mbuf *m, struct mbuf *nam,
487	 struct mbuf *control, struct proc *p)
488{
489	struct inpcb *inp = sotoinpcb(so);
490	register u_long dst;
491
492	if (so->so_state & SS_ISCONNECTED) {
493		if (nam) {
494			m_freem(m);
495			return EISCONN;
496		}
497		dst = inp->inp_faddr.s_addr;
498	} else {
499		if (nam == NULL) {
500			m_freem(m);
501			return ENOTCONN;
502		}
503		dst = mtod(nam, struct sockaddr_in *)->sin_addr.s_addr;
504	}
505	return rip_output(m, so, dst);
506}
507
508struct pr_usrreqs rip_usrreqs = {
509	rip_abort, pru_accept_notsupp, rip_attach, rip_bind, rip_connect,
510	pru_connect2_notsupp, in_control, rip_detach, rip_disconnect,
511	pru_listen_notsupp, in_setpeeraddr, pru_rcvd_notsupp,
512	pru_rcvoob_notsupp, rip_send, pru_sense_null, rip_shutdown,
513	in_setsockaddr, sosend, soreceive, soselect
514};
515