xref: /netbsd-src/sys/netinet/raw_ip.c (revision 37afb7eb6895c833050f8bfb1d1bb2f99f332539)
1 /*	$NetBSD: raw_ip.c,v 1.152 2015/05/02 17:18:03 rtr 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 1.152 2015/05/02 17:18:03 rtr Exp $");
69 
70 #include "opt_inet.h"
71 #include "opt_compat_netbsd.h"
72 #include "opt_ipsec.h"
73 #include "opt_mrouting.h"
74 
75 #include <sys/param.h>
76 #include <sys/sysctl.h>
77 #include <sys/mbuf.h>
78 #include <sys/socket.h>
79 #include <sys/protosw.h>
80 #include <sys/socketvar.h>
81 #include <sys/errno.h>
82 #include <sys/systm.h>
83 #include <sys/proc.h>
84 #include <sys/kauth.h>
85 
86 #include <net/if.h>
87 #include <net/route.h>
88 
89 #include <netinet/in.h>
90 #include <netinet/in_systm.h>
91 #include <netinet/ip.h>
92 #include <netinet/ip_var.h>
93 #include <netinet/ip_private.h>
94 #include <netinet/ip_mroute.h>
95 #include <netinet/ip_icmp.h>
96 #include <netinet/in_pcb.h>
97 #include <netinet/in_proto.h>
98 #include <netinet/in_var.h>
99 
100 #ifdef IPSEC
101 #include <netipsec/ipsec.h>
102 #include <netipsec/ipsec_var.h>
103 #include <netipsec/ipsec_private.h>
104 #endif	/* IPSEC */
105 
106 #ifdef COMPAT_50
107 #include <compat/sys/socket.h>
108 #endif
109 
110 struct inpcbtable rawcbtable;
111 
112 int	 rip_pcbnotify(struct inpcbtable *, struct in_addr,
113     struct in_addr, int, int, void (*)(struct inpcb *, int));
114 static int	 rip_connect_pcb(struct inpcb *, struct sockaddr_in *);
115 static void	 rip_disconnect1(struct inpcb *);
116 
117 static void sysctl_net_inet_raw_setup(struct sysctllog **);
118 
119 /*
120  * Nominal space allocated to a raw ip socket.
121  */
122 #define	RIPSNDQ		8192
123 #define	RIPRCVQ		8192
124 
125 static u_long		rip_sendspace = RIPSNDQ;
126 static u_long		rip_recvspace = RIPRCVQ;
127 
128 /*
129  * Raw interface to IP protocol.
130  */
131 
132 /*
133  * Initialize raw connection block q.
134  */
135 void
136 rip_init(void)
137 {
138 
139 	sysctl_net_inet_raw_setup(NULL);
140 	in_pcbinit(&rawcbtable, 1, 1);
141 }
142 
143 static void
144 rip_sbappendaddr(struct inpcb *last, struct ip *ip, const struct sockaddr *sa,
145     int hlen, struct mbuf *opts, struct mbuf *n)
146 {
147 	if (last->inp_flags & INP_NOHEADER)
148 		m_adj(n, hlen);
149 	if (last->inp_flags & INP_CONTROLOPTS
150 #ifdef SO_OTIMESTAMP
151 	    || last->inp_socket->so_options & SO_OTIMESTAMP
152 #endif
153 	    || last->inp_socket->so_options & SO_TIMESTAMP)
154 		ip_savecontrol(last, &opts, ip, n);
155 	if (sbappendaddr(&last->inp_socket->so_rcv, sa, n, opts) == 0) {
156 		/* should notify about lost packet */
157 		m_freem(n);
158 		if (opts)
159 			m_freem(opts);
160 	} else
161 		sorwakeup(last->inp_socket);
162 }
163 
164 /*
165  * Setup generic address and protocol structures
166  * for raw_input routine, then pass them along with
167  * mbuf chain.
168  */
169 void
170 rip_input(struct mbuf *m, ...)
171 {
172 	int hlen, proto;
173 	struct ip *ip = mtod(m, struct ip *);
174 	struct inpcb_hdr *inph;
175 	struct inpcb *inp;
176 	struct inpcb *last = NULL;
177 	struct mbuf *n, *opts = NULL;
178 	struct sockaddr_in ripsrc;
179 	va_list ap;
180 
181 	va_start(ap, m);
182 	(void)va_arg(ap, int);		/* ignore value, advance ap */
183 	proto = va_arg(ap, int);
184 	va_end(ap);
185 
186 	sockaddr_in_init(&ripsrc, &ip->ip_src, 0);
187 
188 	/*
189 	 * XXX Compatibility: programs using raw IP expect ip_len
190 	 * XXX to have the header length subtracted, and in host order.
191 	 * XXX ip_off is also expected to be host order.
192 	 */
193 	hlen = ip->ip_hl << 2;
194 	ip->ip_len = ntohs(ip->ip_len) - hlen;
195 	NTOHS(ip->ip_off);
196 
197 	TAILQ_FOREACH(inph, &rawcbtable.inpt_queue, inph_queue) {
198 		inp = (struct inpcb *)inph;
199 		if (inp->inp_af != AF_INET)
200 			continue;
201 		if (inp->inp_ip.ip_p && inp->inp_ip.ip_p != proto)
202 			continue;
203 		if (!in_nullhost(inp->inp_laddr) &&
204 		    !in_hosteq(inp->inp_laddr, ip->ip_dst))
205 			continue;
206 		if (!in_nullhost(inp->inp_faddr) &&
207 		    !in_hosteq(inp->inp_faddr, ip->ip_src))
208 			continue;
209 		if (last == NULL)
210 			;
211 #if defined(IPSEC)
212 		/* check AH/ESP integrity. */
213 		else if (ipsec_used &&
214 		    ipsec4_in_reject_so(m, last->inp_socket)) {
215 			IPSEC_STATINC(IPSEC_STAT_IN_POLVIO);
216 			/* do not inject data to pcb */
217 		}
218 #endif /*IPSEC*/
219 		else if ((n = m_copypacket(m, M_DONTWAIT)) != NULL) {
220 			rip_sbappendaddr(last, ip, sintosa(&ripsrc), hlen, opts,
221 			    n);
222 			opts = NULL;
223 		}
224 		last = inp;
225 	}
226 #if defined(IPSEC)
227 	/* check AH/ESP integrity. */
228 	if (ipsec_used && last != NULL
229 	    && ipsec4_in_reject_so(m, last->inp_socket)) {
230 		m_freem(m);
231 		IPSEC_STATINC(IPSEC_STAT_IN_POLVIO);
232 		IP_STATDEC(IP_STAT_DELIVERED);
233 		/* do not inject data to pcb */
234 	} else
235 #endif /*IPSEC*/
236 	if (last != NULL)
237 		rip_sbappendaddr(last, ip, sintosa(&ripsrc), hlen, opts, m);
238 	else if (inetsw[ip_protox[ip->ip_p]].pr_input == rip_input) {
239 		uint64_t *ips;
240 
241 		icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_PROTOCOL,
242 		    0, 0);
243 		ips = IP_STAT_GETREF();
244 		ips[IP_STAT_NOPROTO]++;
245 		ips[IP_STAT_DELIVERED]--;
246 		IP_STAT_PUTREF();
247 	} else
248 		m_freem(m);
249 	return;
250 }
251 
252 int
253 rip_pcbnotify(struct inpcbtable *table,
254     struct in_addr faddr, struct in_addr laddr, int proto, int errno,
255     void (*notify)(struct inpcb *, int))
256 {
257 	struct inpcb_hdr *inph, *ninph;
258 	int nmatch;
259 
260 	nmatch = 0;
261 	TAILQ_FOREACH_SAFE(inph, &table->inpt_queue, inph_queue, ninph) {
262 		struct inpcb *inp = (struct inpcb *)inph;
263 		if (inp->inp_af != AF_INET)
264 			continue;
265 		if (inp->inp_ip.ip_p && inp->inp_ip.ip_p != proto)
266 			continue;
267 		if (in_hosteq(inp->inp_faddr, faddr) &&
268 		    in_hosteq(inp->inp_laddr, laddr)) {
269 			(*notify)(inp, errno);
270 			nmatch++;
271 		}
272 	}
273 
274 	return nmatch;
275 }
276 
277 void *
278 rip_ctlinput(int cmd, const struct sockaddr *sa, void *v)
279 {
280 	struct ip *ip = v;
281 	void (*notify)(struct inpcb *, int) = in_rtchange;
282 	int errno;
283 
284 	if (sa->sa_family != AF_INET ||
285 	    sa->sa_len != sizeof(struct sockaddr_in))
286 		return NULL;
287 	if ((unsigned)cmd >= PRC_NCMDS)
288 		return NULL;
289 	errno = inetctlerrmap[cmd];
290 	if (PRC_IS_REDIRECT(cmd))
291 		notify = in_rtchange, ip = 0;
292 	else if (cmd == PRC_HOSTDEAD)
293 		ip = 0;
294 	else if (errno == 0)
295 		return NULL;
296 	if (ip) {
297 		rip_pcbnotify(&rawcbtable, satocsin(sa)->sin_addr,
298 		    ip->ip_src, ip->ip_p, errno, notify);
299 
300 		/* XXX mapped address case */
301 	} else
302 		in_pcbnotifyall(&rawcbtable, satocsin(sa)->sin_addr, errno,
303 		    notify);
304 	return NULL;
305 }
306 
307 /*
308  * Generate IP header and pass packet to ip_output.
309  * Tack on options user may have setup with control call.
310  */
311 int
312 rip_output(struct mbuf *m, ...)
313 {
314 	struct inpcb *inp;
315 	struct ip *ip;
316 	struct mbuf *opts;
317 	int flags;
318 	va_list ap;
319 
320 	va_start(ap, m);
321 	inp = va_arg(ap, struct inpcb *);
322 	va_end(ap);
323 
324 	flags =
325 	    (inp->inp_socket->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST
326 	    | IP_RETURNMTU;
327 
328 	/*
329 	 * If the user handed us a complete IP packet, use it.
330 	 * Otherwise, allocate an mbuf for a header and fill it in.
331 	 */
332 	if ((inp->inp_flags & INP_HDRINCL) == 0) {
333 		if ((m->m_pkthdr.len + sizeof(struct ip)) > IP_MAXPACKET) {
334 			m_freem(m);
335 			return (EMSGSIZE);
336 		}
337 		M_PREPEND(m, sizeof(struct ip), M_DONTWAIT);
338 		if (!m)
339 			return (ENOBUFS);
340 		ip = mtod(m, struct ip *);
341 		ip->ip_tos = 0;
342 		ip->ip_off = htons(0);
343 		ip->ip_p = inp->inp_ip.ip_p;
344 		ip->ip_len = htons(m->m_pkthdr.len);
345 		ip->ip_src = inp->inp_laddr;
346 		ip->ip_dst = inp->inp_faddr;
347 		ip->ip_ttl = MAXTTL;
348 		opts = inp->inp_options;
349 	} else {
350 		if (m->m_pkthdr.len > IP_MAXPACKET) {
351 			m_freem(m);
352 			return (EMSGSIZE);
353 		}
354 		ip = mtod(m, struct ip *);
355 
356 		/*
357 		 * If the mbuf is read-only, we need to allocate
358 		 * a new mbuf for the header, since we need to
359 		 * modify the header.
360 		 */
361 		if (M_READONLY(m)) {
362 			int hlen = ip->ip_hl << 2;
363 
364 			m = m_copyup(m, hlen, (max_linkhdr + 3) & ~3);
365 			if (m == NULL)
366 				return (ENOMEM);	/* XXX */
367 			ip = mtod(m, struct ip *);
368 		}
369 
370 		/* XXX userland passes ip_len and ip_off in host order */
371 		if (m->m_pkthdr.len != ip->ip_len) {
372 			m_freem(m);
373 			return (EINVAL);
374 		}
375 		HTONS(ip->ip_len);
376 		HTONS(ip->ip_off);
377 		if (ip->ip_id != 0 || m->m_pkthdr.len < IP_MINFRAGSIZE)
378 			flags |= IP_NOIPNEWID;
379 		opts = NULL;
380 		/* XXX prevent ip_output from overwriting header fields */
381 		flags |= IP_RAWOUTPUT;
382 		IP_STATINC(IP_STAT_RAWOUT);
383 	}
384 
385 	/*
386 	 * IP output.  Note: if IP_RETURNMTU flag is set, the MTU size
387 	 * will be stored in inp_errormtu.
388 	 */
389 	return ip_output(m, opts, &inp->inp_route, flags, inp->inp_moptions,
390 	     inp->inp_socket);
391 }
392 
393 /*
394  * Raw IP socket option processing.
395  */
396 int
397 rip_ctloutput(int op, struct socket *so, struct sockopt *sopt)
398 {
399 	struct inpcb *inp = sotoinpcb(so);
400 	int error = 0;
401 	int optval;
402 
403 	if (sopt->sopt_level == SOL_SOCKET && sopt->sopt_name == SO_NOHEADER) {
404 		if (op == PRCO_GETOPT) {
405 			optval = (inp->inp_flags & INP_NOHEADER) ? 1 : 0;
406 			error = sockopt_set(sopt, &optval, sizeof(optval));
407 		} else if (op == PRCO_SETOPT) {
408 			error = sockopt_getint(sopt, &optval);
409 			if (error)
410 				goto out;
411 			if (optval) {
412 				inp->inp_flags &= ~INP_HDRINCL;
413 				inp->inp_flags |= INP_NOHEADER;
414 			} else
415 				inp->inp_flags &= ~INP_NOHEADER;
416 		}
417 		goto out;
418 	} else if (sopt->sopt_level != IPPROTO_IP)
419 		return ip_ctloutput(op, so, sopt);
420 
421 	switch (op) {
422 
423 	case PRCO_SETOPT:
424 		switch (sopt->sopt_name) {
425 		case IP_HDRINCL:
426 			error = sockopt_getint(sopt, &optval);
427 			if (error)
428 				break;
429 			if (optval)
430 				inp->inp_flags |= INP_HDRINCL;
431 			else
432 				inp->inp_flags &= ~INP_HDRINCL;
433 			break;
434 
435 #ifdef MROUTING
436 		case MRT_INIT:
437 		case MRT_DONE:
438 		case MRT_ADD_VIF:
439 		case MRT_DEL_VIF:
440 		case MRT_ADD_MFC:
441 		case MRT_DEL_MFC:
442 		case MRT_ASSERT:
443 		case MRT_API_CONFIG:
444 		case MRT_ADD_BW_UPCALL:
445 		case MRT_DEL_BW_UPCALL:
446 			error = ip_mrouter_set(so, sopt);
447 			break;
448 #endif
449 
450 		default:
451 			error = ip_ctloutput(op, so, sopt);
452 			break;
453 		}
454 		break;
455 
456 	case PRCO_GETOPT:
457 		switch (sopt->sopt_name) {
458 		case IP_HDRINCL:
459 			optval = inp->inp_flags & INP_HDRINCL;
460 			error = sockopt_set(sopt, &optval, sizeof(optval));
461 			break;
462 
463 #ifdef MROUTING
464 		case MRT_VERSION:
465 		case MRT_ASSERT:
466 		case MRT_API_SUPPORT:
467 		case MRT_API_CONFIG:
468 			error = ip_mrouter_get(so, sopt);
469 			break;
470 #endif
471 
472 		default:
473 			error = ip_ctloutput(op, so, sopt);
474 			break;
475 		}
476 		break;
477 	}
478  out:
479 	return error;
480 }
481 
482 int
483 rip_connect_pcb(struct inpcb *inp, struct sockaddr_in *addr)
484 {
485 
486 	if (IFNET_EMPTY())
487 		return (EADDRNOTAVAIL);
488 	if (addr->sin_family != AF_INET)
489 		return (EAFNOSUPPORT);
490 	inp->inp_faddr = addr->sin_addr;
491 	return (0);
492 }
493 
494 static void
495 rip_disconnect1(struct inpcb *inp)
496 {
497 
498 	inp->inp_faddr = zeroin_addr;
499 }
500 
501 static int
502 rip_attach(struct socket *so, int proto)
503 {
504 	struct inpcb *inp;
505 	int error;
506 
507 	KASSERT(sotoinpcb(so) == NULL);
508 	sosetlock(so);
509 
510 	if (so->so_snd.sb_hiwat == 0 || so->so_rcv.sb_hiwat == 0) {
511 		error = soreserve(so, rip_sendspace, rip_recvspace);
512 		if (error) {
513 			return error;
514 		}
515 	}
516 
517 	error = in_pcballoc(so, &rawcbtable);
518 	if (error) {
519 		return error;
520 	}
521 	inp = sotoinpcb(so);
522 	inp->inp_ip.ip_p = proto;
523 	KASSERT(solocked(so));
524 
525 	return 0;
526 }
527 
528 static void
529 rip_detach(struct socket *so)
530 {
531 	struct inpcb *inp;
532 
533 	KASSERT(solocked(so));
534 	inp = sotoinpcb(so);
535 	KASSERT(inp != NULL);
536 
537 #ifdef MROUTING
538 	extern struct socket *ip_mrouter;
539 	if (so == ip_mrouter) {
540 		ip_mrouter_done();
541 	}
542 #endif
543 	in_pcbdetach(inp);
544 }
545 
546 static int
547 rip_accept(struct socket *so, struct sockaddr *nam)
548 {
549 	KASSERT(solocked(so));
550 
551 	panic("rip_accept");
552 
553 	return EOPNOTSUPP;
554 }
555 
556 static int
557 rip_bind(struct socket *so, struct sockaddr *nam, struct lwp *l)
558 {
559 	struct inpcb *inp = sotoinpcb(so);
560 	struct sockaddr_in *addr = (struct sockaddr_in *)nam;
561 	int error = 0;
562 	int s;
563 	struct ifaddr *ia;
564 
565 	KASSERT(solocked(so));
566 	KASSERT(inp != NULL);
567 	KASSERT(nam != NULL);
568 
569 	if (addr->sin_len != sizeof(*addr))
570 		return EINVAL;
571 
572 	s = splsoftnet();
573 	if (IFNET_EMPTY()) {
574 		error = EADDRNOTAVAIL;
575 		goto release;
576 	}
577 	if (addr->sin_family != AF_INET) {
578 		error = EAFNOSUPPORT;
579 		goto release;
580 	}
581 	if ((ia = ifa_ifwithaddr(sintosa(addr))) == 0 &&
582 	    !in_nullhost(addr->sin_addr))
583 	{
584 		error = EADDRNOTAVAIL;
585 		goto release;
586 	}
587         if (ia && ((struct in_ifaddr *)ia)->ia4_flags &
588 	            (IN6_IFF_NOTREADY | IN_IFF_DETACHED))
589 	{
590 		error = EADDRNOTAVAIL;
591 		goto release;
592 	}
593 
594 	inp->inp_laddr = addr->sin_addr;
595 
596 release:
597 	splx(s);
598 	return error;
599 }
600 
601 static int
602 rip_listen(struct socket *so, struct lwp *l)
603 {
604 	KASSERT(solocked(so));
605 
606 	return EOPNOTSUPP;
607 }
608 
609 static int
610 rip_connect(struct socket *so, struct sockaddr *nam, struct lwp *l)
611 {
612 	struct inpcb *inp = sotoinpcb(so);
613 	int error = 0;
614 	int s;
615 
616 	KASSERT(solocked(so));
617 	KASSERT(inp != NULL);
618 	KASSERT(nam != NULL);
619 
620 	s = splsoftnet();
621 	error = rip_connect_pcb(inp, (struct sockaddr_in *)nam);
622 	if (! error)
623 		soisconnected(so);
624 	splx(s);
625 
626 	return error;
627 }
628 
629 static int
630 rip_connect2(struct socket *so, struct socket *so2)
631 {
632 	KASSERT(solocked(so));
633 
634 	return EOPNOTSUPP;
635 }
636 
637 static int
638 rip_disconnect(struct socket *so)
639 {
640 	struct inpcb *inp = sotoinpcb(so);
641 	int s;
642 
643 	KASSERT(solocked(so));
644 	KASSERT(inp != NULL);
645 
646 	s = splsoftnet();
647 	soisdisconnected(so);
648 	rip_disconnect1(inp);
649 	splx(s);
650 
651 	return 0;
652 }
653 
654 static int
655 rip_shutdown(struct socket *so)
656 {
657 	int s;
658 
659 	KASSERT(solocked(so));
660 
661 	/*
662 	 * Mark the connection as being incapable of further input.
663 	 */
664 	s = splsoftnet();
665 	socantsendmore(so);
666 	splx(s);
667 
668 	return 0;
669 }
670 
671 static int
672 rip_abort(struct socket *so)
673 {
674 	KASSERT(solocked(so));
675 
676 	panic("rip_abort");
677 
678 	return EOPNOTSUPP;
679 }
680 
681 static int
682 rip_ioctl(struct socket *so, u_long cmd, void *nam, struct ifnet *ifp)
683 {
684 	return in_control(so, cmd, nam, ifp);
685 }
686 
687 static int
688 rip_stat(struct socket *so, struct stat *ub)
689 {
690 	KASSERT(solocked(so));
691 
692 	/* stat: don't bother with a blocksize. */
693 	return 0;
694 }
695 
696 static int
697 rip_peeraddr(struct socket *so, struct sockaddr *nam)
698 {
699 	int s;
700 
701 	KASSERT(solocked(so));
702 	KASSERT(sotoinpcb(so) != NULL);
703 	KASSERT(nam != NULL);
704 
705 	s = splsoftnet();
706 	in_setpeeraddr(sotoinpcb(so), (struct sockaddr_in *)nam);
707 	splx(s);
708 
709 	return 0;
710 }
711 
712 static int
713 rip_sockaddr(struct socket *so, struct sockaddr *nam)
714 {
715 	int s;
716 
717 	KASSERT(solocked(so));
718 	KASSERT(sotoinpcb(so) != NULL);
719 	KASSERT(nam != NULL);
720 
721 	s = splsoftnet();
722 	in_setsockaddr(sotoinpcb(so), (struct sockaddr_in *)nam);
723 	splx(s);
724 
725 	return 0;
726 }
727 
728 static int
729 rip_rcvd(struct socket *so, int flags, struct lwp *l)
730 {
731 	KASSERT(solocked(so));
732 
733 	return EOPNOTSUPP;
734 }
735 
736 static int
737 rip_recvoob(struct socket *so, struct mbuf *m, int flags)
738 {
739 	KASSERT(solocked(so));
740 
741 	return EOPNOTSUPP;
742 }
743 
744 static int
745 rip_send(struct socket *so, struct mbuf *m, struct sockaddr *nam,
746     struct mbuf *control, struct lwp *l)
747 {
748 	struct inpcb *inp = sotoinpcb(so);
749 	int error = 0;
750 	int s;
751 
752 	KASSERT(solocked(so));
753 	KASSERT(inp != NULL);
754 	KASSERT(m != NULL);
755 
756 	/*
757 	 * Ship a packet out.  The appropriate raw output
758 	 * routine handles any massaging necessary.
759 	 */
760 	if (control && control->m_len) {
761 		m_freem(control);
762 		m_freem(m);
763 		return EINVAL;
764 	}
765 
766 	s = splsoftnet();
767 	if (nam) {
768 		if ((so->so_state & SS_ISCONNECTED) != 0) {
769 			error = EISCONN;
770 			goto die;
771 		}
772 		error = rip_connect_pcb(inp, (struct sockaddr_in *)nam);
773 		if (error) {
774 		die:
775 			m_freem(m);
776 			splx(s);
777 			return error;
778 		}
779 	} else {
780 		if ((so->so_state & SS_ISCONNECTED) == 0) {
781 			error = ENOTCONN;
782 			goto die;
783 		}
784 	}
785 	error = rip_output(m, inp);
786 	if (nam)
787 		rip_disconnect1(inp);
788 
789 	splx(s);
790 	return error;
791 }
792 
793 static int
794 rip_sendoob(struct socket *so, struct mbuf *m, struct mbuf *control)
795 {
796 	KASSERT(solocked(so));
797 
798 	m_freem(m);
799 	m_freem(control);
800 
801 	return EOPNOTSUPP;
802 }
803 
804 static int
805 rip_purgeif(struct socket *so, struct ifnet *ifp)
806 {
807 	int s;
808 
809 	s = splsoftnet();
810 	mutex_enter(softnet_lock);
811 	in_pcbpurgeif0(&rawcbtable, ifp);
812 	in_purgeif(ifp);
813 	in_pcbpurgeif(&rawcbtable, ifp);
814 	mutex_exit(softnet_lock);
815 	splx(s);
816 
817 	return 0;
818 }
819 
820 PR_WRAP_USRREQS(rip)
821 #define	rip_attach	rip_attach_wrapper
822 #define	rip_detach	rip_detach_wrapper
823 #define	rip_accept	rip_accept_wrapper
824 #define	rip_bind	rip_bind_wrapper
825 #define	rip_listen	rip_listen_wrapper
826 #define	rip_connect	rip_connect_wrapper
827 #define	rip_connect2	rip_connect2_wrapper
828 #define	rip_disconnect	rip_disconnect_wrapper
829 #define	rip_shutdown	rip_shutdown_wrapper
830 #define	rip_abort	rip_abort_wrapper
831 #define	rip_ioctl	rip_ioctl_wrapper
832 #define	rip_stat	rip_stat_wrapper
833 #define	rip_peeraddr	rip_peeraddr_wrapper
834 #define	rip_sockaddr	rip_sockaddr_wrapper
835 #define	rip_rcvd	rip_rcvd_wrapper
836 #define	rip_recvoob	rip_recvoob_wrapper
837 #define	rip_send	rip_send_wrapper
838 #define	rip_sendoob	rip_sendoob_wrapper
839 #define	rip_purgeif	rip_purgeif_wrapper
840 
841 const struct pr_usrreqs rip_usrreqs = {
842 	.pr_attach	= rip_attach,
843 	.pr_detach	= rip_detach,
844 	.pr_accept	= rip_accept,
845 	.pr_bind	= rip_bind,
846 	.pr_listen	= rip_listen,
847 	.pr_connect	= rip_connect,
848 	.pr_connect2	= rip_connect2,
849 	.pr_disconnect	= rip_disconnect,
850 	.pr_shutdown	= rip_shutdown,
851 	.pr_abort	= rip_abort,
852 	.pr_ioctl	= rip_ioctl,
853 	.pr_stat	= rip_stat,
854 	.pr_peeraddr	= rip_peeraddr,
855 	.pr_sockaddr	= rip_sockaddr,
856 	.pr_rcvd	= rip_rcvd,
857 	.pr_recvoob	= rip_recvoob,
858 	.pr_send	= rip_send,
859 	.pr_sendoob	= rip_sendoob,
860 	.pr_purgeif	= rip_purgeif,
861 };
862 
863 static void
864 sysctl_net_inet_raw_setup(struct sysctllog **clog)
865 {
866 
867 	sysctl_createv(clog, 0, NULL, NULL,
868 		       CTLFLAG_PERMANENT,
869 		       CTLTYPE_NODE, "inet", NULL,
870 		       NULL, 0, NULL, 0,
871 		       CTL_NET, PF_INET, CTL_EOL);
872 	sysctl_createv(clog, 0, NULL, NULL,
873 		       CTLFLAG_PERMANENT,
874 		       CTLTYPE_NODE, "raw",
875 		       SYSCTL_DESCR("Raw IPv4 settings"),
876 		       NULL, 0, NULL, 0,
877 		       CTL_NET, PF_INET, IPPROTO_RAW, CTL_EOL);
878 
879 	sysctl_createv(clog, 0, NULL, NULL,
880 		       CTLFLAG_PERMANENT,
881 		       CTLTYPE_STRUCT, "pcblist",
882 		       SYSCTL_DESCR("Raw IPv4 control block list"),
883 		       sysctl_inpcblist, 0, &rawcbtable, 0,
884 		       CTL_NET, PF_INET, IPPROTO_RAW,
885 		       CTL_CREATE, CTL_EOL);
886 }
887