xref: /netbsd-src/sys/netinet/raw_ip.c (revision fa28c6faa16e0b00edee7acdcaf4899797043def)
1 /*	$NetBSD: raw_ip.c,v 1.153 2015/08/24 22:21:26 pooka Exp $	*/
2 
3 /*
4  * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
5  * All rights reserved.
6  *
7  * Redistribution and use in source and binary forms, with or without
8  * modification, are permitted provided that the following conditions
9  * are met:
10  * 1. Redistributions of source code must retain the above copyright
11  *    notice, this list of conditions and the following disclaimer.
12  * 2. Redistributions in binary form must reproduce the above copyright
13  *    notice, this list of conditions and the following disclaimer in the
14  *    documentation and/or other materials provided with the distribution.
15  * 3. Neither the name of the project nor the names of its contributors
16  *    may be used to endorse or promote products derived from this software
17  *    without specific prior written permission.
18  *
19  * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
20  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22  * ARE DISCLAIMED.  IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
23  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
24  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
25  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
26  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
28  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29  * SUCH DAMAGE.
30  */
31 
32 /*
33  * Copyright (c) 1982, 1986, 1988, 1993
34  *	The Regents of the University of California.  All rights reserved.
35  *
36  * Redistribution and use in source and binary forms, with or without
37  * modification, are permitted provided that the following conditions
38  * are met:
39  * 1. Redistributions of source code must retain the above copyright
40  *    notice, this list of conditions and the following disclaimer.
41  * 2. Redistributions in binary form must reproduce the above copyright
42  *    notice, this list of conditions and the following disclaimer in the
43  *    documentation and/or other materials provided with the distribution.
44  * 3. Neither the name of the University nor the names of its contributors
45  *    may be used to endorse or promote products derived from this software
46  *    without specific prior written permission.
47  *
48  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
49  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
50  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
51  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
52  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
53  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
54  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
55  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
56  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
57  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
58  * SUCH DAMAGE.
59  *
60  *	@(#)raw_ip.c	8.7 (Berkeley) 5/15/95
61  */
62 
63 /*
64  * Raw interface to IP protocol.
65  */
66 
67 #include <sys/cdefs.h>
68 __KERNEL_RCSID(0, "$NetBSD: raw_ip.c,v 1.153 2015/08/24 22:21:26 pooka Exp $");
69 
70 #ifdef _KERNEL_OPT
71 #include "opt_inet.h"
72 #include "opt_compat_netbsd.h"
73 #include "opt_ipsec.h"
74 #include "opt_mrouting.h"
75 #endif
76 
77 #include <sys/param.h>
78 #include <sys/sysctl.h>
79 #include <sys/mbuf.h>
80 #include <sys/socket.h>
81 #include <sys/protosw.h>
82 #include <sys/socketvar.h>
83 #include <sys/errno.h>
84 #include <sys/systm.h>
85 #include <sys/proc.h>
86 #include <sys/kauth.h>
87 
88 #include <net/if.h>
89 #include <net/route.h>
90 
91 #include <netinet/in.h>
92 #include <netinet/in_systm.h>
93 #include <netinet/ip.h>
94 #include <netinet/ip_var.h>
95 #include <netinet/ip_private.h>
96 #include <netinet/ip_mroute.h>
97 #include <netinet/ip_icmp.h>
98 #include <netinet/in_pcb.h>
99 #include <netinet/in_proto.h>
100 #include <netinet/in_var.h>
101 
102 #ifdef IPSEC
103 #include <netipsec/ipsec.h>
104 #include <netipsec/ipsec_var.h>
105 #include <netipsec/ipsec_private.h>
106 #endif	/* IPSEC */
107 
108 #ifdef COMPAT_50
109 #include <compat/sys/socket.h>
110 #endif
111 
112 struct inpcbtable rawcbtable;
113 
114 int	 rip_pcbnotify(struct inpcbtable *, struct in_addr,
115     struct in_addr, int, int, void (*)(struct inpcb *, int));
116 static int	 rip_connect_pcb(struct inpcb *, struct sockaddr_in *);
117 static void	 rip_disconnect1(struct inpcb *);
118 
119 static void sysctl_net_inet_raw_setup(struct sysctllog **);
120 
121 /*
122  * Nominal space allocated to a raw ip socket.
123  */
124 #define	RIPSNDQ		8192
125 #define	RIPRCVQ		8192
126 
127 static u_long		rip_sendspace = RIPSNDQ;
128 static u_long		rip_recvspace = RIPRCVQ;
129 
130 /*
131  * Raw interface to IP protocol.
132  */
133 
134 /*
135  * Initialize raw connection block q.
136  */
137 void
138 rip_init(void)
139 {
140 
141 	sysctl_net_inet_raw_setup(NULL);
142 	in_pcbinit(&rawcbtable, 1, 1);
143 }
144 
145 static void
146 rip_sbappendaddr(struct inpcb *last, struct ip *ip, const struct sockaddr *sa,
147     int hlen, struct mbuf *opts, struct mbuf *n)
148 {
149 	if (last->inp_flags & INP_NOHEADER)
150 		m_adj(n, hlen);
151 	if (last->inp_flags & INP_CONTROLOPTS
152 #ifdef SO_OTIMESTAMP
153 	    || last->inp_socket->so_options & SO_OTIMESTAMP
154 #endif
155 	    || last->inp_socket->so_options & SO_TIMESTAMP)
156 		ip_savecontrol(last, &opts, ip, n);
157 	if (sbappendaddr(&last->inp_socket->so_rcv, sa, n, opts) == 0) {
158 		/* should notify about lost packet */
159 		m_freem(n);
160 		if (opts)
161 			m_freem(opts);
162 	} else
163 		sorwakeup(last->inp_socket);
164 }
165 
166 /*
167  * Setup generic address and protocol structures
168  * for raw_input routine, then pass them along with
169  * mbuf chain.
170  */
171 void
172 rip_input(struct mbuf *m, ...)
173 {
174 	int hlen, proto;
175 	struct ip *ip = mtod(m, struct ip *);
176 	struct inpcb_hdr *inph;
177 	struct inpcb *inp;
178 	struct inpcb *last = NULL;
179 	struct mbuf *n, *opts = NULL;
180 	struct sockaddr_in ripsrc;
181 	va_list ap;
182 
183 	va_start(ap, m);
184 	(void)va_arg(ap, int);		/* ignore value, advance ap */
185 	proto = va_arg(ap, int);
186 	va_end(ap);
187 
188 	sockaddr_in_init(&ripsrc, &ip->ip_src, 0);
189 
190 	/*
191 	 * XXX Compatibility: programs using raw IP expect ip_len
192 	 * XXX to have the header length subtracted, and in host order.
193 	 * XXX ip_off is also expected to be host order.
194 	 */
195 	hlen = ip->ip_hl << 2;
196 	ip->ip_len = ntohs(ip->ip_len) - hlen;
197 	NTOHS(ip->ip_off);
198 
199 	TAILQ_FOREACH(inph, &rawcbtable.inpt_queue, inph_queue) {
200 		inp = (struct inpcb *)inph;
201 		if (inp->inp_af != AF_INET)
202 			continue;
203 		if (inp->inp_ip.ip_p && inp->inp_ip.ip_p != proto)
204 			continue;
205 		if (!in_nullhost(inp->inp_laddr) &&
206 		    !in_hosteq(inp->inp_laddr, ip->ip_dst))
207 			continue;
208 		if (!in_nullhost(inp->inp_faddr) &&
209 		    !in_hosteq(inp->inp_faddr, ip->ip_src))
210 			continue;
211 		if (last == NULL)
212 			;
213 #if defined(IPSEC)
214 		/* check AH/ESP integrity. */
215 		else if (ipsec_used &&
216 		    ipsec4_in_reject_so(m, last->inp_socket)) {
217 			IPSEC_STATINC(IPSEC_STAT_IN_POLVIO);
218 			/* do not inject data to pcb */
219 		}
220 #endif /*IPSEC*/
221 		else if ((n = m_copypacket(m, M_DONTWAIT)) != NULL) {
222 			rip_sbappendaddr(last, ip, sintosa(&ripsrc), hlen, opts,
223 			    n);
224 			opts = NULL;
225 		}
226 		last = inp;
227 	}
228 #if defined(IPSEC)
229 	/* check AH/ESP integrity. */
230 	if (ipsec_used && last != NULL
231 	    && ipsec4_in_reject_so(m, last->inp_socket)) {
232 		m_freem(m);
233 		IPSEC_STATINC(IPSEC_STAT_IN_POLVIO);
234 		IP_STATDEC(IP_STAT_DELIVERED);
235 		/* do not inject data to pcb */
236 	} else
237 #endif /*IPSEC*/
238 	if (last != NULL)
239 		rip_sbappendaddr(last, ip, sintosa(&ripsrc), hlen, opts, m);
240 	else if (inetsw[ip_protox[ip->ip_p]].pr_input == rip_input) {
241 		uint64_t *ips;
242 
243 		icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_PROTOCOL,
244 		    0, 0);
245 		ips = IP_STAT_GETREF();
246 		ips[IP_STAT_NOPROTO]++;
247 		ips[IP_STAT_DELIVERED]--;
248 		IP_STAT_PUTREF();
249 	} else
250 		m_freem(m);
251 	return;
252 }
253 
254 int
255 rip_pcbnotify(struct inpcbtable *table,
256     struct in_addr faddr, struct in_addr laddr, int proto, int errno,
257     void (*notify)(struct inpcb *, int))
258 {
259 	struct inpcb_hdr *inph, *ninph;
260 	int nmatch;
261 
262 	nmatch = 0;
263 	TAILQ_FOREACH_SAFE(inph, &table->inpt_queue, inph_queue, ninph) {
264 		struct inpcb *inp = (struct inpcb *)inph;
265 		if (inp->inp_af != AF_INET)
266 			continue;
267 		if (inp->inp_ip.ip_p && inp->inp_ip.ip_p != proto)
268 			continue;
269 		if (in_hosteq(inp->inp_faddr, faddr) &&
270 		    in_hosteq(inp->inp_laddr, laddr)) {
271 			(*notify)(inp, errno);
272 			nmatch++;
273 		}
274 	}
275 
276 	return nmatch;
277 }
278 
279 void *
280 rip_ctlinput(int cmd, const struct sockaddr *sa, void *v)
281 {
282 	struct ip *ip = v;
283 	void (*notify)(struct inpcb *, int) = in_rtchange;
284 	int errno;
285 
286 	if (sa->sa_family != AF_INET ||
287 	    sa->sa_len != sizeof(struct sockaddr_in))
288 		return NULL;
289 	if ((unsigned)cmd >= PRC_NCMDS)
290 		return NULL;
291 	errno = inetctlerrmap[cmd];
292 	if (PRC_IS_REDIRECT(cmd))
293 		notify = in_rtchange, ip = 0;
294 	else if (cmd == PRC_HOSTDEAD)
295 		ip = 0;
296 	else if (errno == 0)
297 		return NULL;
298 	if (ip) {
299 		rip_pcbnotify(&rawcbtable, satocsin(sa)->sin_addr,
300 		    ip->ip_src, ip->ip_p, errno, notify);
301 
302 		/* XXX mapped address case */
303 	} else
304 		in_pcbnotifyall(&rawcbtable, satocsin(sa)->sin_addr, errno,
305 		    notify);
306 	return NULL;
307 }
308 
309 /*
310  * Generate IP header and pass packet to ip_output.
311  * Tack on options user may have setup with control call.
312  */
313 int
314 rip_output(struct mbuf *m, ...)
315 {
316 	struct inpcb *inp;
317 	struct ip *ip;
318 	struct mbuf *opts;
319 	int flags;
320 	va_list ap;
321 
322 	va_start(ap, m);
323 	inp = va_arg(ap, struct inpcb *);
324 	va_end(ap);
325 
326 	flags =
327 	    (inp->inp_socket->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST
328 	    | IP_RETURNMTU;
329 
330 	/*
331 	 * If the user handed us a complete IP packet, use it.
332 	 * Otherwise, allocate an mbuf for a header and fill it in.
333 	 */
334 	if ((inp->inp_flags & INP_HDRINCL) == 0) {
335 		if ((m->m_pkthdr.len + sizeof(struct ip)) > IP_MAXPACKET) {
336 			m_freem(m);
337 			return (EMSGSIZE);
338 		}
339 		M_PREPEND(m, sizeof(struct ip), M_DONTWAIT);
340 		if (!m)
341 			return (ENOBUFS);
342 		ip = mtod(m, struct ip *);
343 		ip->ip_tos = 0;
344 		ip->ip_off = htons(0);
345 		ip->ip_p = inp->inp_ip.ip_p;
346 		ip->ip_len = htons(m->m_pkthdr.len);
347 		ip->ip_src = inp->inp_laddr;
348 		ip->ip_dst = inp->inp_faddr;
349 		ip->ip_ttl = MAXTTL;
350 		opts = inp->inp_options;
351 	} else {
352 		if (m->m_pkthdr.len > IP_MAXPACKET) {
353 			m_freem(m);
354 			return (EMSGSIZE);
355 		}
356 		ip = mtod(m, struct ip *);
357 
358 		/*
359 		 * If the mbuf is read-only, we need to allocate
360 		 * a new mbuf for the header, since we need to
361 		 * modify the header.
362 		 */
363 		if (M_READONLY(m)) {
364 			int hlen = ip->ip_hl << 2;
365 
366 			m = m_copyup(m, hlen, (max_linkhdr + 3) & ~3);
367 			if (m == NULL)
368 				return (ENOMEM);	/* XXX */
369 			ip = mtod(m, struct ip *);
370 		}
371 
372 		/* XXX userland passes ip_len and ip_off in host order */
373 		if (m->m_pkthdr.len != ip->ip_len) {
374 			m_freem(m);
375 			return (EINVAL);
376 		}
377 		HTONS(ip->ip_len);
378 		HTONS(ip->ip_off);
379 		if (ip->ip_id != 0 || m->m_pkthdr.len < IP_MINFRAGSIZE)
380 			flags |= IP_NOIPNEWID;
381 		opts = NULL;
382 		/* XXX prevent ip_output from overwriting header fields */
383 		flags |= IP_RAWOUTPUT;
384 		IP_STATINC(IP_STAT_RAWOUT);
385 	}
386 
387 	/*
388 	 * IP output.  Note: if IP_RETURNMTU flag is set, the MTU size
389 	 * will be stored in inp_errormtu.
390 	 */
391 	return ip_output(m, opts, &inp->inp_route, flags, inp->inp_moptions,
392 	     inp->inp_socket);
393 }
394 
395 /*
396  * Raw IP socket option processing.
397  */
398 int
399 rip_ctloutput(int op, struct socket *so, struct sockopt *sopt)
400 {
401 	struct inpcb *inp = sotoinpcb(so);
402 	int error = 0;
403 	int optval;
404 
405 	if (sopt->sopt_level == SOL_SOCKET && sopt->sopt_name == SO_NOHEADER) {
406 		if (op == PRCO_GETOPT) {
407 			optval = (inp->inp_flags & INP_NOHEADER) ? 1 : 0;
408 			error = sockopt_set(sopt, &optval, sizeof(optval));
409 		} else if (op == PRCO_SETOPT) {
410 			error = sockopt_getint(sopt, &optval);
411 			if (error)
412 				goto out;
413 			if (optval) {
414 				inp->inp_flags &= ~INP_HDRINCL;
415 				inp->inp_flags |= INP_NOHEADER;
416 			} else
417 				inp->inp_flags &= ~INP_NOHEADER;
418 		}
419 		goto out;
420 	} else if (sopt->sopt_level != IPPROTO_IP)
421 		return ip_ctloutput(op, so, sopt);
422 
423 	switch (op) {
424 
425 	case PRCO_SETOPT:
426 		switch (sopt->sopt_name) {
427 		case IP_HDRINCL:
428 			error = sockopt_getint(sopt, &optval);
429 			if (error)
430 				break;
431 			if (optval)
432 				inp->inp_flags |= INP_HDRINCL;
433 			else
434 				inp->inp_flags &= ~INP_HDRINCL;
435 			break;
436 
437 #ifdef MROUTING
438 		case MRT_INIT:
439 		case MRT_DONE:
440 		case MRT_ADD_VIF:
441 		case MRT_DEL_VIF:
442 		case MRT_ADD_MFC:
443 		case MRT_DEL_MFC:
444 		case MRT_ASSERT:
445 		case MRT_API_CONFIG:
446 		case MRT_ADD_BW_UPCALL:
447 		case MRT_DEL_BW_UPCALL:
448 			error = ip_mrouter_set(so, sopt);
449 			break;
450 #endif
451 
452 		default:
453 			error = ip_ctloutput(op, so, sopt);
454 			break;
455 		}
456 		break;
457 
458 	case PRCO_GETOPT:
459 		switch (sopt->sopt_name) {
460 		case IP_HDRINCL:
461 			optval = inp->inp_flags & INP_HDRINCL;
462 			error = sockopt_set(sopt, &optval, sizeof(optval));
463 			break;
464 
465 #ifdef MROUTING
466 		case MRT_VERSION:
467 		case MRT_ASSERT:
468 		case MRT_API_SUPPORT:
469 		case MRT_API_CONFIG:
470 			error = ip_mrouter_get(so, sopt);
471 			break;
472 #endif
473 
474 		default:
475 			error = ip_ctloutput(op, so, sopt);
476 			break;
477 		}
478 		break;
479 	}
480  out:
481 	return error;
482 }
483 
484 int
485 rip_connect_pcb(struct inpcb *inp, struct sockaddr_in *addr)
486 {
487 
488 	if (IFNET_EMPTY())
489 		return (EADDRNOTAVAIL);
490 	if (addr->sin_family != AF_INET)
491 		return (EAFNOSUPPORT);
492 	inp->inp_faddr = addr->sin_addr;
493 	return (0);
494 }
495 
496 static void
497 rip_disconnect1(struct inpcb *inp)
498 {
499 
500 	inp->inp_faddr = zeroin_addr;
501 }
502 
503 static int
504 rip_attach(struct socket *so, int proto)
505 {
506 	struct inpcb *inp;
507 	int error;
508 
509 	KASSERT(sotoinpcb(so) == NULL);
510 	sosetlock(so);
511 
512 	if (so->so_snd.sb_hiwat == 0 || so->so_rcv.sb_hiwat == 0) {
513 		error = soreserve(so, rip_sendspace, rip_recvspace);
514 		if (error) {
515 			return error;
516 		}
517 	}
518 
519 	error = in_pcballoc(so, &rawcbtable);
520 	if (error) {
521 		return error;
522 	}
523 	inp = sotoinpcb(so);
524 	inp->inp_ip.ip_p = proto;
525 	KASSERT(solocked(so));
526 
527 	return 0;
528 }
529 
530 static void
531 rip_detach(struct socket *so)
532 {
533 	struct inpcb *inp;
534 
535 	KASSERT(solocked(so));
536 	inp = sotoinpcb(so);
537 	KASSERT(inp != NULL);
538 
539 #ifdef MROUTING
540 	extern struct socket *ip_mrouter;
541 	if (so == ip_mrouter) {
542 		ip_mrouter_done();
543 	}
544 #endif
545 	in_pcbdetach(inp);
546 }
547 
548 static int
549 rip_accept(struct socket *so, struct sockaddr *nam)
550 {
551 	KASSERT(solocked(so));
552 
553 	panic("rip_accept");
554 
555 	return EOPNOTSUPP;
556 }
557 
558 static int
559 rip_bind(struct socket *so, struct sockaddr *nam, struct lwp *l)
560 {
561 	struct inpcb *inp = sotoinpcb(so);
562 	struct sockaddr_in *addr = (struct sockaddr_in *)nam;
563 	int error = 0;
564 	int s;
565 	struct ifaddr *ia;
566 
567 	KASSERT(solocked(so));
568 	KASSERT(inp != NULL);
569 	KASSERT(nam != NULL);
570 
571 	if (addr->sin_len != sizeof(*addr))
572 		return EINVAL;
573 
574 	s = splsoftnet();
575 	if (IFNET_EMPTY()) {
576 		error = EADDRNOTAVAIL;
577 		goto release;
578 	}
579 	if (addr->sin_family != AF_INET) {
580 		error = EAFNOSUPPORT;
581 		goto release;
582 	}
583 	if ((ia = ifa_ifwithaddr(sintosa(addr))) == 0 &&
584 	    !in_nullhost(addr->sin_addr))
585 	{
586 		error = EADDRNOTAVAIL;
587 		goto release;
588 	}
589         if (ia && ((struct in_ifaddr *)ia)->ia4_flags &
590 	            (IN6_IFF_NOTREADY | IN_IFF_DETACHED))
591 	{
592 		error = EADDRNOTAVAIL;
593 		goto release;
594 	}
595 
596 	inp->inp_laddr = addr->sin_addr;
597 
598 release:
599 	splx(s);
600 	return error;
601 }
602 
603 static int
604 rip_listen(struct socket *so, struct lwp *l)
605 {
606 	KASSERT(solocked(so));
607 
608 	return EOPNOTSUPP;
609 }
610 
611 static int
612 rip_connect(struct socket *so, struct sockaddr *nam, struct lwp *l)
613 {
614 	struct inpcb *inp = sotoinpcb(so);
615 	int error = 0;
616 	int s;
617 
618 	KASSERT(solocked(so));
619 	KASSERT(inp != NULL);
620 	KASSERT(nam != NULL);
621 
622 	s = splsoftnet();
623 	error = rip_connect_pcb(inp, (struct sockaddr_in *)nam);
624 	if (! error)
625 		soisconnected(so);
626 	splx(s);
627 
628 	return error;
629 }
630 
631 static int
632 rip_connect2(struct socket *so, struct socket *so2)
633 {
634 	KASSERT(solocked(so));
635 
636 	return EOPNOTSUPP;
637 }
638 
639 static int
640 rip_disconnect(struct socket *so)
641 {
642 	struct inpcb *inp = sotoinpcb(so);
643 	int s;
644 
645 	KASSERT(solocked(so));
646 	KASSERT(inp != NULL);
647 
648 	s = splsoftnet();
649 	soisdisconnected(so);
650 	rip_disconnect1(inp);
651 	splx(s);
652 
653 	return 0;
654 }
655 
656 static int
657 rip_shutdown(struct socket *so)
658 {
659 	int s;
660 
661 	KASSERT(solocked(so));
662 
663 	/*
664 	 * Mark the connection as being incapable of further input.
665 	 */
666 	s = splsoftnet();
667 	socantsendmore(so);
668 	splx(s);
669 
670 	return 0;
671 }
672 
673 static int
674 rip_abort(struct socket *so)
675 {
676 	KASSERT(solocked(so));
677 
678 	panic("rip_abort");
679 
680 	return EOPNOTSUPP;
681 }
682 
683 static int
684 rip_ioctl(struct socket *so, u_long cmd, void *nam, struct ifnet *ifp)
685 {
686 	return in_control(so, cmd, nam, ifp);
687 }
688 
689 static int
690 rip_stat(struct socket *so, struct stat *ub)
691 {
692 	KASSERT(solocked(so));
693 
694 	/* stat: don't bother with a blocksize. */
695 	return 0;
696 }
697 
698 static int
699 rip_peeraddr(struct socket *so, struct sockaddr *nam)
700 {
701 	int s;
702 
703 	KASSERT(solocked(so));
704 	KASSERT(sotoinpcb(so) != NULL);
705 	KASSERT(nam != NULL);
706 
707 	s = splsoftnet();
708 	in_setpeeraddr(sotoinpcb(so), (struct sockaddr_in *)nam);
709 	splx(s);
710 
711 	return 0;
712 }
713 
714 static int
715 rip_sockaddr(struct socket *so, struct sockaddr *nam)
716 {
717 	int s;
718 
719 	KASSERT(solocked(so));
720 	KASSERT(sotoinpcb(so) != NULL);
721 	KASSERT(nam != NULL);
722 
723 	s = splsoftnet();
724 	in_setsockaddr(sotoinpcb(so), (struct sockaddr_in *)nam);
725 	splx(s);
726 
727 	return 0;
728 }
729 
730 static int
731 rip_rcvd(struct socket *so, int flags, struct lwp *l)
732 {
733 	KASSERT(solocked(so));
734 
735 	return EOPNOTSUPP;
736 }
737 
738 static int
739 rip_recvoob(struct socket *so, struct mbuf *m, int flags)
740 {
741 	KASSERT(solocked(so));
742 
743 	return EOPNOTSUPP;
744 }
745 
746 static int
747 rip_send(struct socket *so, struct mbuf *m, struct sockaddr *nam,
748     struct mbuf *control, struct lwp *l)
749 {
750 	struct inpcb *inp = sotoinpcb(so);
751 	int error = 0;
752 	int s;
753 
754 	KASSERT(solocked(so));
755 	KASSERT(inp != NULL);
756 	KASSERT(m != NULL);
757 
758 	/*
759 	 * Ship a packet out.  The appropriate raw output
760 	 * routine handles any massaging necessary.
761 	 */
762 	if (control && control->m_len) {
763 		m_freem(control);
764 		m_freem(m);
765 		return EINVAL;
766 	}
767 
768 	s = splsoftnet();
769 	if (nam) {
770 		if ((so->so_state & SS_ISCONNECTED) != 0) {
771 			error = EISCONN;
772 			goto die;
773 		}
774 		error = rip_connect_pcb(inp, (struct sockaddr_in *)nam);
775 		if (error) {
776 		die:
777 			m_freem(m);
778 			splx(s);
779 			return error;
780 		}
781 	} else {
782 		if ((so->so_state & SS_ISCONNECTED) == 0) {
783 			error = ENOTCONN;
784 			goto die;
785 		}
786 	}
787 	error = rip_output(m, inp);
788 	if (nam)
789 		rip_disconnect1(inp);
790 
791 	splx(s);
792 	return error;
793 }
794 
795 static int
796 rip_sendoob(struct socket *so, struct mbuf *m, struct mbuf *control)
797 {
798 	KASSERT(solocked(so));
799 
800 	m_freem(m);
801 	m_freem(control);
802 
803 	return EOPNOTSUPP;
804 }
805 
806 static int
807 rip_purgeif(struct socket *so, struct ifnet *ifp)
808 {
809 	int s;
810 
811 	s = splsoftnet();
812 	mutex_enter(softnet_lock);
813 	in_pcbpurgeif0(&rawcbtable, ifp);
814 	in_purgeif(ifp);
815 	in_pcbpurgeif(&rawcbtable, ifp);
816 	mutex_exit(softnet_lock);
817 	splx(s);
818 
819 	return 0;
820 }
821 
822 PR_WRAP_USRREQS(rip)
823 #define	rip_attach	rip_attach_wrapper
824 #define	rip_detach	rip_detach_wrapper
825 #define	rip_accept	rip_accept_wrapper
826 #define	rip_bind	rip_bind_wrapper
827 #define	rip_listen	rip_listen_wrapper
828 #define	rip_connect	rip_connect_wrapper
829 #define	rip_connect2	rip_connect2_wrapper
830 #define	rip_disconnect	rip_disconnect_wrapper
831 #define	rip_shutdown	rip_shutdown_wrapper
832 #define	rip_abort	rip_abort_wrapper
833 #define	rip_ioctl	rip_ioctl_wrapper
834 #define	rip_stat	rip_stat_wrapper
835 #define	rip_peeraddr	rip_peeraddr_wrapper
836 #define	rip_sockaddr	rip_sockaddr_wrapper
837 #define	rip_rcvd	rip_rcvd_wrapper
838 #define	rip_recvoob	rip_recvoob_wrapper
839 #define	rip_send	rip_send_wrapper
840 #define	rip_sendoob	rip_sendoob_wrapper
841 #define	rip_purgeif	rip_purgeif_wrapper
842 
843 const struct pr_usrreqs rip_usrreqs = {
844 	.pr_attach	= rip_attach,
845 	.pr_detach	= rip_detach,
846 	.pr_accept	= rip_accept,
847 	.pr_bind	= rip_bind,
848 	.pr_listen	= rip_listen,
849 	.pr_connect	= rip_connect,
850 	.pr_connect2	= rip_connect2,
851 	.pr_disconnect	= rip_disconnect,
852 	.pr_shutdown	= rip_shutdown,
853 	.pr_abort	= rip_abort,
854 	.pr_ioctl	= rip_ioctl,
855 	.pr_stat	= rip_stat,
856 	.pr_peeraddr	= rip_peeraddr,
857 	.pr_sockaddr	= rip_sockaddr,
858 	.pr_rcvd	= rip_rcvd,
859 	.pr_recvoob	= rip_recvoob,
860 	.pr_send	= rip_send,
861 	.pr_sendoob	= rip_sendoob,
862 	.pr_purgeif	= rip_purgeif,
863 };
864 
865 static void
866 sysctl_net_inet_raw_setup(struct sysctllog **clog)
867 {
868 
869 	sysctl_createv(clog, 0, NULL, NULL,
870 		       CTLFLAG_PERMANENT,
871 		       CTLTYPE_NODE, "inet", NULL,
872 		       NULL, 0, NULL, 0,
873 		       CTL_NET, PF_INET, CTL_EOL);
874 	sysctl_createv(clog, 0, NULL, NULL,
875 		       CTLFLAG_PERMANENT,
876 		       CTLTYPE_NODE, "raw",
877 		       SYSCTL_DESCR("Raw IPv4 settings"),
878 		       NULL, 0, NULL, 0,
879 		       CTL_NET, PF_INET, IPPROTO_RAW, CTL_EOL);
880 
881 	sysctl_createv(clog, 0, NULL, NULL,
882 		       CTLFLAG_PERMANENT,
883 		       CTLTYPE_STRUCT, "pcblist",
884 		       SYSCTL_DESCR("Raw IPv4 control block list"),
885 		       sysctl_inpcblist, 0, &rawcbtable, 0,
886 		       CTL_NET, PF_INET, IPPROTO_RAW,
887 		       CTL_CREATE, CTL_EOL);
888 }
889